Financial Diaries with Smallholder Families

PERSPECTIVES | No. 2 | February 2016 SMALLHOLDER DIARIES Building the Evidence Base with Farming Families in Mozambique, Tanzania, and Pakistan JAMIE ANDERSON AND WAJIHA AHMED

ACKNOWLEDGMENTS The authors wish to thank Mwombeki Baregu (Financial Sector Deepening Trust Tanzania), Juan Buchenau (World Bank Group), Wendy Chamberlin (Bill & Melinda Gates Foundation), Carlos E. Cuevas (Evans School of Public Policy and Governance, University of Washington), Janine Firpo (Bill & Melinda Gates Foundation), Scott Gardner (InterMedia), Michael Hamp (International Fund for Agricultural Development), Sosthenes Kewe (Financial Sector Deepening Trust Tanzania), Colleen Learch (Inter- Media), Richard Meyer (Ohio State University), Benedito Murambire (Financial Sector Deepening Trust Mozambique), John Ndunguru (Bill & Melinda Gates Foundation), Jessie Wild Sneller (Ideo.org), Rossana Ramirez, Stuart Rutherford (SafeSave), and Jason Wendle (Rural and Agricultural Finance Learning Lab) for their valuable guidance in this research and the production of this paper. © 2016, Consultative Group to Assist the Poor (CGAP) 1818 H Street NW, MSN IS7-700 Washington DC 20433 Internet: www.cgap.org Email: [email protected] Telephone: +1 202 473 9594 Rights and Permissions This work is available under the Creative Commons Attribution 3.0 Unported license (CC BY 3.0) http:// creativecommons .org/licenses/by/3.0. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Cite the work as follows: Anderson, Jamie, and Wajiha Ahmed. 2016. “Smallholder Diaries: Building the Evidence Base with Farming Families in Mozambique, Tanzania, and Pakistan.” Perspectives 2. Washington, D.C.: CGAP. License: Creative Commons Attribution CC BY 3.0 Translations—If you create a translation of this work, add the following disclaimer along with the attribution: This translation was not created by CGAP and should not be considered an ofﬁcial translation. CGAP shall not be liable for any content or error in this translation. All queries on rights and licenses should be addressed to CGAP Publications, 1818 H Street, NW, MSN IS7-700, Washington, DC 20433 USA; e-mail: cgap@world bank.org.

SMALLHOLDER DIARIES Building the Evidence Base with Farming Families in Mozambique, Tanzania, and Pakistan JAMIE ANDERSON AND WAJIHA AHMED

Photo by Erin Scronce.

CONTENTS Executive Summary 1 Introduction and methodology 13 Sample demographics and agricultural characteristics 21 Chapter 1 Income sources and the role of in-kind consumption 29 Case 1 Juggling multiple income sources to repay debts: Bertha (Tanzania) 34 Chapter 2 Patterns of agricultural production and sales 37 Case 2 The beneﬁts and burdens of a tight value chain: The Maliks (Pakistan) 41 Chapter 3 Relationships between income volatility and agricultural production 43 Case 3 Coping with income volatility and anticipating the hunger season: Alina (Mozambique) 50 Chapter 4 Risk mitigation and coping strategies 53 Case 4 Diversifying crops, diet, and income, but constrained from doing more: The Namuacas (Mozambique) 60 Chapter 5 Household ﬁnancial portfolios 63 Case 5 Working with a moderate ﬁnancial portfolio and struggling to make agricultural investments: The Bitungwas (Tanzania) 70 Chapter 6 Access to mobile phones and use of digital ﬁnancial tools 73 Case 6 Relying on credit in months between harvest: The Bhattis (Pakistan) 76 Chapter 7 Implications for ﬁnancial solutions: Translating the evidence into ﬁnancial tools tailored to each smallholder household proﬁle 79 Annex 1. Average value of assets per household 88 Annex 2. Deﬁnitions of income sources in the Smallholder Diaries 89 Annex 3. Deﬁnitions of ﬁnancial tools in the Smallholder Diaries 90 Annex 4. Use of ﬁnancial tools for savings and credit among Smallholder Diaries households 92 References 95 Notes 97

Financial Diaries with Smallholder Families - Page 8

Mozambique. Photo by Erin Scronce

EXECUTIVE SUMMARY How do small-scale agricultural producers manage their money, and what do their strategies tell us about their need for ﬁnancial tools? Smallholder house- holds, cultivating approximately two hectares (ﬁve acres) or less, present one of the most challenging client segments for ﬁnancial service providers (Ss) due to three uniue aspects of their ﬁnancial lives (i) their income from agricultural production is often erratic and infreuent (ii) their reuired investments can be signiﬁcant and must be made at speciﬁc times of the year and (iii) their riss, whose incidence and covariance across the agricultural sector can be diﬃcult to mitigate €n addition, their ﬁnancial needs extend beyond agricultural produc- tion to a variety of nonagricultural, oﬀ-farm enterprises ƒie any family, small- holder households also need a range of ﬁnancial tools to meet regular expenses, respond to emergencies, and ﬁnance milestones such as weddings and funerals urthermore, improvements to the ﬁnancial portfolios of smallholder house- holds represent only one contribution to their overall well-being, and ma„or … challenges related to health, infrastructure, and education persist †‡ˆ launched the year-long inancial ‰iaries with Smallholder amilies (the “Smallholder ‰iaries”) to elucidate the ﬁnancial lives of smallholder households and build the evidence base on this important client group Œhe study, conducted between Žune ‘’…“ and Žuly ‘’…”, captured the ﬁnancial and in-ind transactions of ‘•’ households in impoverished northern –o—ambiue, the fertile farmlands of western Œan—ania, and the un„ab province, the bread- baset of aistan ˜early all adults in the Smallholder ‰iaries sample were born into farming households Œhey began woring in agriculture at a young age and self-identiﬁed as being part of an agricultural household ™hen the Smallholder ‰iaries began, participants indicated that their agricultural activi- ties were their most important sources of income Œhe data, however, told a more nuanced story Œhe Smallholder ‰iaries provide a deep view of how smallholders are aﬀected by the agricultural cycle and manage their money in response to its ebbs and ﬂows, as well as point to ways that Ss might better meet smallhold- ers’ needs ™hile the Smallholder ‰iaries methodology and sample si—e are not statistically representative of all smallholder families in a given country, the ﬁndings from the Smallholder ‰iaries have global implications for the small- holder household sector Œhe sample of smallholder households from each study country has characteristics that are broadly representative of the types of smallholder segments identiﬁed in countries around the world, which presents an opportunity to discuss the types of ﬁnancial tools that these segments demand regardless of their location | 1

Financial Diaries with Smallholder Families - Page 11

2 | SMALLHOLDER DIARIES Income sources and the role of in-kind consumption Smallholder Diaries families had numerous sources of cash income, which tended to fall into three categories: (i) agricultural production, (ii) casual labor (often related to agriculture), and (iii) other oﬀ-farm, nonagricultural sources such as managing a small business, receiving remittances, or engaging in regular or waged employment ˆt the median, households had a total of eight income sources in –o—ambiue, …… in Œan—ania, and nine in aistan (see igure ‘ œS-…) –uch of their casual labor was on neighboring farms, and thus still lined to agriculture €n the Œan—ania sample, more than half of the income sources classiﬁed as casual labor was related to agriculture Families in the Smallholder Diaries also earned the majority of their house- hold net cash income from their numerous nonagricultural production acti- ities ˆmong the sample families, the median proportion of household cash income from nonagricultural production sources was žŸ percent in –o—am- biue, •“ percent in Œan—ania, and ”¡ percent in aistan (see igure œS-…) FIGURE ES-1: Smallholder Diaries: Household income from agricultural and nonagricultural production JUNE 2014–JULY 2015 Pakistan Income sources 4 5 Proportion of 42% 58% income Income sources 5 5 w/in-kind Proportion of income 53% 47% w/in-kind Tanzania Income sources 2 9 Proportion of 26% 74% income Income sources 3 9 w/in-kind Proportion of income 46% 54% w/in-kind Mozambique Income sources 2 6 Proportion of 7% 93% income Income sources 6 6 w/in-kind Proportion of income 49% 51% w/in-kind Agricultural production income Nonagricultural production income (1) Median number of household income sources (2) Median proportion of total household net cash income (3) Median number of household income sources, factoring in in-kind consumption (4) Median proportion of total household income factoring in in-kind consumption Note: Each crop or livestock byproduct (e.g., milk, eggs) that was sold at least once in the Smallholder Diaries is considered a distinct source of agricultural production income. When tracking in-kind consumption, the Smallholder Diaries recorded only activities or transactions related to crops, not livestock byproducts.

EXECUTIVE SUMMARY | 3 ut focusing only on cash income underplays the importance of agriculture to smallholder households Œhe relative importance of agricultural production income increases maredly when household consumption of crops is included in the analysis (see igure œS-…) €n the sample in Œan—ania and aistan, the median proportion of household income derived from crop production increased from ‘¢ percent to “¢ percent and “‘ percent to ”Ÿ percent, respec- tively, when also considering in-ind consumption £ut the diﬀerence is most dramatic in –o—ambiue, „umping from • percent to “ž percent Patterns of agricultural production and sales he degree to which smallholders sold their production and the pace of this monetiation aries sharply across the three country samples Œhis echoes the segmentation framewor proposed by †‡ˆ that diﬀerentiates among (i) noncommercial smallholders, (ii) commercial smallholders in loose value chains, and (iii) commercial smallholders in tight value chains according to what they grow, how they engage with marets as buyers and¤or sellers, and how those marets are organi—ed (†hristen and ˆnderson ‘’…Ÿ) €f and how smallholder households consume or sell their crops and the nature of their con- nections to value chains, for example, have important implications for the roles that ﬁnancial tools can play in their lives and how they are tailored to household circumstances • Smallholder families in the oambi ue sample were “net consumers” of their agricultural production (ie, they consumed more of their agricul- tural output than they sold) Œhese were largely noncommercial smallholder households, with limited sales of crops and livestoc ˆnd many households did not sell anything over the entire year of data collection Œhey were able to consume smoothly what they produced, but they were unable to do much else with their harvest Œhey did not sell their crops for the cash needed to buy other foods, diversify their diet, or meet other household needs • Smallholder households in the anania sample were “net sellers” in loose alue chains (ie, they sold more of their agricultural output than they consumed) Œhese households typically had one ma„or harvest of a cash crop (eg, rice, potatoes) each year, which would typically be sold for cash to vil- lage-level agents and¤or larger aggregating buyers ˆ few sold direct to mar- et as well ¦verall, sales of their agricultural production were fairly lumpy (ie, they occurred in distinct periods over the year, not continuously) and spied during the main harvest Households also stored crops, consuming some proportion of some of them over time, and moneti—ing them when they needed cash • Smallholder households in the ‚aƒistan sample were net sellers in tight alue chains, consuming an even smaller proportion of their production than in the Œan—ania sample ˆfter each ma„or harvest in ˜ovember and in –ay, they were usually obliged to sell their output immediately bac to the middleman to repay debts for the costly agricultural inputs they had ﬁnanced

4 | SMALLHOLDER DIARIES Relationships between income volatility and agricultural production „gricultural production income was marƒedly more olatile than other sources of income in all three samples, and a household’s oerall income olatility depended on the balance between agricultural production and nonagricultural production income (see igure œS-‘) Œhe wide range of income sources outside of crop and livestoc production did dampen the eﬀects of the agricultural cycle on the sample households, but only to a point Œhe vol- atility of agricultural production and its inherent riss still exerted a strong inﬂuence over the ﬁnancial lives of smallholders €n terms of household ﬁnances and health, respondents in all three countries struggled most in the months between harvests urthermore, smallholders most often pointed to agriculture as causes of ﬁnancial hardship, citing reasons such as “waiting to harvest crops” and “expenses related to agriculture” (see igure œS-Ÿ) FIGURE ES-2: Volatility of income: Median standard deviation of monthly income relative to average monthly income, a JULY 2014– JUNE 2015 500 458% 400 300 281% 222% 213% 200 120% 100 89% 84% 101% 76% 0 Mozambique Tanzania Pakistan Agricultural production volatility Nonagricultural production volatility Total income volatility a. Relative standard deviation of income = (Standard deviation of monthly income * 100)/ Average monthly income Standard deviation of monthly income represents the amount by which a household’s income deviates from the average monthly income of that household. FIGURE ES-3: “[In the months when the family struggled most with money] what happened to cause this difficulty?” Multiple responses allowed; percentage of household 100% 90% 80% 70% 30% 2% 60% 50% 44% 12% 41% 23% 40% 7% 2% 30% 5% 57% 18% 11% 4% 20% 41% 51% 1% 30% 15% 10% 28% 23% 13% 26% 20% 1% 17% 5% 4% 6% 10% 9% 0% 1% Waiting to Expenses Already sold Harvests Price of Extensive Price of No casual Health No specific Repayment Expenses Theft harvest crops related to all crops were lower crops rose crop loss crops work or emergency reason of debt related to agriculture than at time to dropped at income life events expected buy food selling time activities Mozambique Tanzania Pakistan

EXECUTIVE SUMMARY | 5 or the sample in aistan, month-to-month net income from agricultural pro- duction was highly volatile (“”¡ percent relative to average income) €t swung from lows well below —ero, when there were ma„or expenditures on inputs, to signiﬁcant highs, when they sold their output immediately after harvest €n the –o—ambiue sample, given the low level of crop sales and high reliance on other sources of cash income, families experienced less severe ﬂuctuations in overall income †‡penses were smoother than income, but still ﬂuctuated to some e‡tent with income in all three samples €n –o—ambiue and Œan—ania, sample households tended to spend money as it came in, with expenses more closely tied to cash income (ie, a “spend-as-you-go” expenditure pattern) (see igures œS-“ and œS-”) Œhe aistan sample experienced the largest swings in income, FIGURE ES-4: Mozambique Smallholder Diaries: Net income and household expenditures all sample level, JULY 2014–JUNE 2015 (US$)a $4,500 $4,000 $3,500 $3,000 Hardship months/ hunger season $2,500 $2,000 $1,500 $1,000 $500 $0 Jul Aug Sep OctNov Dec Jan Feb Mar Apr May Jun Household expenditures Net-income a. The green income line refers to net income. For agricultural production, and small businesses in particular, income refers to revenue less related expenditures. The red expenses line refers to operational expenses of the household separate from income or financial transactions (e.g., spending on groceries, clothes, education, transportation). FIGURE ES-5: Tanzania Smallholder Diaries, rice production village: Net income and household expenditures all sample level, JULY 2014–MAY 2015 (US$)a $3,500 $3,000 $2,500 $2,000 $1,500 Hardship months $1,000 $500 $0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Household expenditures Net income a. The green income line refers to net income. For agricultural production, and small businesses in particular, income refers to revenue less related expenditures. The red expenses line refers to operational expenses of the household separate from income or financial transactions (e.g., spending on groceries, clothes, education, transportation).

6 | SMALLHOLDER DIARIES but they were able to maintain a consistent level of expenditures thans to their greater access to credit options Risk mitigation and coping strategies ‰ess commercialied smallholders e‡perienced more production-related shocƒs, such as bad weather, drought, and pests, while more commercial- ied households faced greater marƒet-related ones, such as ﬂuctuations in input and crop prices Œhese challenges came on top of the shocs commonly experienced by households (eg, sicness, death of a family member, „ob loss) • Šn the largely noncommercial oambi ue sample, production risƒs were paramount, largely because the sample was not engaged in marƒets Households in the –o—ambiue sample needed to store their harvest for many months to cover their own food consumption and relied on bags ept in the house for crop storage £ut their stored crops were vulnerable nearly two-thirds of sample households had lost crops while in storage (see igure œS-¢) €mprovements in post-harvest handling and crop storage would bet- ter preserve their agricultural production and maintain its value as both food and a form of savings • he sample in anania, largely commercial smallholders in loose alue chains, e‡perienced the full spectrum of agricultural shocƒs, both pro- duction and marƒet related, though at somewhat lower levels than the other two samples ¦ver one-third of the sample had experienced signiﬁcant crop loss due to weather shocs (Ÿ¢ percent) and pests (“’ percent) and decreases in the prices for their own agricultural production (Ÿž percent) Œhis range of riss calls for a variety of mitigating mechanisms, involving both agricultural techniues and ﬁnancial tools • For commercial smallholders in ‚aƒistan embedded in tight alue chains, marƒet risƒ was more pressing than production risƒ §irtually every house- hold was aﬀected by increases in the price of inputs and decreases in the pur- FIGURE ES-6: Households that experienced selected agricultural shocks at least once in the past five years, JULY 2015 (Percentage) 100% 96% 99% 90% 80% 72% 70% 61% 61% 60% 51% 50% 43% 38% 40% 41% 40% 36% 32% 30% 26% 23% 26% 21% 18% 20% 10% 8% 7% 10% 5% 2% 0% Lost land Crops destroyed Crops destroyed Crops destroyed Crops stolen Price of own Price of inputs Could not Could not rent by weather by pests in by pests in crops decreased increased rent enough enough land the field storage significantly significantly land due to due to price increase availability Mozambique Tanzania Pakistan

EXECUTIVE SUMMARY | 7 chase prices for their agricultural production ˆ wider range of potential buyers and more ﬁnancial tools to mitigate these riss could be beneﬁcial any Smallholder Diaries households had no speciﬁc response to an agri- cultural shocƒ, signaling a lacƒ of tools with which to cope Œhe diﬀerences between the samples revealed varying degrees of access to ﬁnancial tools and safety nets, as well as their degree of maret engagement ™hen their crops were destroyed by weather, for example, many Œan—anian households in the Smallholder ‰iaries sample did nothing (•‘ percent), reﬂecting an apparent lac of perceived fallbac options (see igure œS-•) ™hen the sample in ai- stan faced the same situation, some smallholder households borrowed money (Ÿ¡ percent) about one-third also had no speciﬁc coping response (Ÿ“ percent) Household ﬁnancial portfolios he degree to which Smallholder Diaries households could sustain their consumption leels and cope with shocƒs during lean periods between har- Ÿ ests depended heaily on the range of tools in their ﬁnancial portfolios –ost Smallholder ‰iaries households had access to only a thin scattering of informal ﬁnancial tools—borrowing from friends and family, credit from a store or agent, saving at home or with a money guard—and each mechanism had its limitations (eg, amount available, proximity, timing) ©se of formal ﬁnancial tools and digital ﬁnance remains limited Households new when to anticipate cash ﬂow problems from past experience, but they laced the ﬁnancial tools to eﬀectively and suﬃciently smooth their consumption • Smallholder households in the oambi ue sample used only three ﬁnan- cial instruments at the median Œheir very narrow ﬁnancial portfolio was mostly limited to savings at home and only a fraction of the –o—ambiue sam- ple was engaged in informal savings and credit groups …‘ percent used rotat- ing savings and credit associations (ª¦S†ˆs), ž percent used accumulating savings and credit associations (ˆS†ˆs), and ” percent used a money guard to “ save Œhough almost half had a mobile phone (“” percent) and some had FIGURE ES-7: Households that used these coping mechanisms when crops were destroyed by weather (Percentage; multiple answers allowed) 80% 72% 70% 60% 50% 40% 37% 38% 34% 35% 30% 28% 20% 10% 9% 10% 9% 6% 6% 4% 7% 5% 3% 3% 3% 3% 2% 0% Nothing specialBorrowed money Used savings Reduced Worked more Sold crops Sold other assets Received help consumption from friends and expenses and family Mozambique Tanzania Pakistan

8 | SMALLHOLDER DIARIES heard of mobile money products (‘… percent), use of mobile money was non- existent ™ith limited savings and credit options, sample households looed to casual labor to get through the hunger season, though the timing of this income did not always match the timing of their needs and it was insuﬃcient to carry families through this diﬃcult period • „t the median, smallholder households in the anania sample used ŒŽ diﬀerent ﬁnancial tools Œhey relied most heavily on current income and short-term savings for both farming and nonfarming expenses Œhe sample also used stored crops as a ind of “term deposit” ‘… percent considered crop storage their most important form of savings, with crops tending to gain “interest” as the price increases over time Œhose who borrowed had a num- ber of small loans with informal groups and from those in their social net- wors, and many found casual wor to generate income when they needed cash ˆlmost everyone in the sample had heard of mobile money (ž¡ per- cent), but only …ž percent had used it to store money or send and receive money –any of the Œan—anian respondents planned strategically to mae investments in their farm, but with few opportunities to borrow, most relied heavily on short-term savings, including crop storage, and earnings from casual labor to buy agricultural inputs • ‘orƒing with the broadest, most robust ﬁnancial portfolio, the sample of smallholder households in ‚aƒistan used Œ’ ﬁnancial tools at the median Households used various forms of credit to get through the months when spending on agricultural inputs was high and revenue from farming was low Œhe great ma„ority of smallholders in aistan had heard of mobile money (¡‘ percent), but no families used it during the Smallholder ‰iaries study However, ”• percent said they would liely use it to send or receive money, indicating an aspiration to use the service Access to mobile phones and use of digital ﬁnancial tools he limited capability of the Smallholder Diaries sample to send and receie SS te‡ts points to a crucial gap between basic access to a phone, which itself remains a barrier, and the ability to perform ﬁnancial transactions with it Œhe ma„ority of respondents among the sample in aistan and Œan—ania owned a mobile phone (•’ percent and ”¢ percent, respectively), but less than half of the respondents in –o—ambiue did (“” percent) (see Œable œS-…) €n fact, only ”” percent of the –o—ambican respondents had used a phone at all in the prior year ˆmong the sample that had access to a mobile phone, the ability to do relatively more complex tass was limited ¢¡ percent in Œan—ania new how to send and receive an S–S text (but not access the internet), but this dropped to ‘” percent in –o—ambiue and ‘“ percent in aistan “se of digital ﬁnancial tools in the Smallholder Diaries sample was ery lim- ited (and only in the anania sample), despite arying leels of awareness of and aspiration to use these tools ‡eneral awareness of mobile money (deﬁned in this study as a transfer of funds using a mobile wallet) as a ﬁnancial tool ranged greatly across Smallholder ‰iaries respondents, from a low of ‘… percent in the –o—ambiue sample to near complete awareness among the sample in Œan—ania (see Œable œS-…), which is expected given the strength of the

EXECUTIVE SUMMARY | 9 a TABLE ES-1: Mobile phones and mobile money among Smallholder Diaries households (percent), NOVEMBER 2014 MOZAMBIQUE TANZANIA PAKISTAN Access to mobile phones and use of mobile money Had a mobile phone 45 56 70 Had a SIM card 48 56 65 Other household members had a mobile phone 38 43 21 Other household members had a SIM 57 35 22 Had used a phone, even a borrowed one, in the past year 55 77 73 Had heard of mobile money 21 98 82 Selected “Mobile money” as one response when asked “What would you likely use to send or receive money?” (multiple answers allowed) 0 74 57 Had used mobile money (for transfers and transactions on a mobile wallet based on actual cash ﬂows from June 2014 to June 2015) 0 19 0 Self-reported capability with mobile phones “I cannot initiate or receive a call, or send or receive an SMS.” 0 3 1 “I can only receive calls.” 45 9 7 “I can only dial and initiate calls.” 2 0 19 “I can dial and initiate calls and receive calls.” 27 15 37 “I can dial and initiate calls, receive calls, and send and receive SMS.” 25 68 24 “I can dial and initiate a call, receive calls, send and receive SMS, and access the internet.” 0 1 7 a. Responses are from a Smallholder Diaries module administered to the most economically active member of the sample household. Œan—anian digital infrastructure Œhis was reﬂected in the perception of mobile money among the sample as a ﬁnancial tool relevant to their needs ™hen ased what ﬁnancial mechanisms they might use to send or receive money, “mobile money” was selected as one answer by almost three-uarters of the sample in Œan—ania and more than half of the sample in aistan «et, despite its perceived relevance, only …ž percent of Smallholder ‰iaries families in Œan—ania used mobile money and no smallholder households in either the –o—ambiue or aistan sample used mobile money at all ”onsidering the potential of digital ﬁnancial serices (DFS) to sere small- holder households in areas that traditional bricƒ-and-mortar FS‚s hae failed to reach, DFS remain important tools to e‡plore and e‡pand for purposes of ﬁnancial inclusion, and they must be carefully targeted to each customer pro- ﬁle ‰igital savings and credit products could provide more compelling use cases than payments, as many households in the sample laced access to even informal ﬁnancial services such as savings groups ˆdditionally, some smallhold- ers may be well-served by digital payment services that facilitate transactions such as bill payments and school fee payments (ie, person-to-business ¬‘£® and person-to-government ¬‘‡® payments), though smallholders may prefer over-the-counter (¦Œ†) payment methods over self-initiated mobile transac- tions from their own wallet

10 | SMALLHOLDER DIARIES Implications for ﬁnancial solutions: Translating the evidence into ﬁnancial tools tailored to each smallholder household proﬁle he Smallholder Diaries e‡plored the relationship between the leel of agri- cultural commercialiation and the breadth of ﬁnancial portfolios in each country sample, and releant ﬁnancial tools must address the uni ue chal- lenges and needs of each of these general proﬁles of smallholder households (see igure œS-¡) • •oncommercial smallholders in the oambi ue sample had a ery nar- row portfolio of ﬁnancial tools ˆdvances in their ﬁnancial inclusion will ” liely emphasi—e savings and layaway products, through digital channels where possible, and the better management of agricultural mechanisms of ﬁnance such as crop storage as opposed to pure, standalone credit products €mproved agronomic practices and better agricultural ris management (eg, post-harvest storage, water catchment, drought-tolerant crop varieties) are also important ƒarger buyers and agricultural processors would need to bundle a meaningful pacage of agronomic support with ﬁnancial tools to reach more smallholders in this proﬁle ˆdditional ﬁnancial tools to help families store and stretch the small amounts of income earned on a daily basis would also be beneﬁcial, especially as a type of safety net during the hunger season • Smallholders in loose alue chains in the anania sample need greater capacity to store money across a dierse set of saings instruments, as well as access to higher leels of credit to maƒe desired inestments in agricultural production ¦verwhelmingly, the Smallholder ‰iaries sample in Œan—ania ept its savings in-ind or under the mattress, presenting a clear opportunity for Ss to oﬀer more avenues to store money ˆ deeper under- FIGURE ES-8: Three smallholder profiles based on degree of agricultural commercialization and breadth of financial portfolio Rapid Tight value chains monetization Smallholders in loose value chains Narrow Broad portfolio portfolio/ credit availability Noncommercial smallholders Non- monetization/ self-consumption

EXECUTIVE SUMMARY | 11 standing of the dynamics of household cash ﬂows in this proﬁle could also provide comfort to both Ss and borrowers alie by demonstrating that cer- tain forms of agricultural lending may not be as risy as previously perceived, or that they are at least mitigated by other less volatile sources of income outside of agricultural production ªelatively higher-income (or somewhat less poor) households lie those in the Œan—ania sample can also more suc- cessfully postpone crop sales to wait for a better price or purposefully use their stored crops to “save” for a lump sum of money †loser connections to buyers and aggregators in the value chain could also beneﬁt this proﬁle €n a country with a robust digital infrastructure lie Œan—ania, these relation- ships and services could be enabled via digital channels Such services could facilitate the creation of purchase agreements or formal contracts, for exam- ple, against which smallholders could borrow for fertili—er, an oft-cited need among the Œan—ania sample households • Smallholders in the relatiely tight alue chains in the ‚aƒistan sample need ﬁnancial tools that facilitate their relationships with middlemen, as well as a range of other mechanisms to reduce their dependence on them Œhe sample in aistan faced ma„or agricultural spending at the beginning of each season and relied on one ma„or buyer to ﬁnance these inputs and also purchase their production Œheir longstanding connection to these middle- men did facilitate a range of other ﬁnancial services, including holding sav- ings and ﬁnancing family milestones and emergencies, and allowed them to reﬁnance and bounce bac after a bad harvest £ut the general terms of their agreement reuired repayment immediately after harvest, forcing small- holders to sell when prices were lowest €n eﬀorts to create a paper trail of transactions and purchase agreements to improve the transparency of these relationships, as well as build a credit history of interest to formal ﬁnancial institutions, digital solutions could play a role ¦ver time, such a system could help smallholders ﬁnd alternatives to middleman as sources of credit, thereby allowing them to wait longer to sell their agricultural production at higher prices Œo compete, however, other service providers would need to emulate the ﬂexibility and proximity of middlemen while oﬀering improved terms Ss might also focus on middlemen as a maret in need of expanded ﬁnancial tools

Mozambique. Photo by Erin Scronce.

INTRODUCTION AND METHODOLOGY BUILDING THE EVIDENCE BASE ON SMALLHOLDER HOUSEHOLDS Œhe sheer numbers of smallholders (estimated at “•” million to ”’’ million smallholder farmers, with …” billion to ‘” billion people living in smallholder households worldwide), their signiﬁcant share of the world’s poor, and their role in food security in low-income countries mae ﬁnancial inclusion of small- • holder households a development priority ƒimited ﬁnancial inclusion among smallholder households is driven by a range of factors Œhese include, inter alia, the high transaction costs of reaching rural households the low population density of rural areas, which maes reaching scale economies challenging and the systemic (and perceived) riss of agricultural production that deter ﬁnan- cial service providers (Ss) from lending for agricultural production or other- ¡ wise engaging with farming households ‰espite the renewed appreciation for the role that smallholder households can play in driving ﬁnancial inclusion, little is nown about this uniue yet mas- sive client group €nformation about how they manage their ﬁnancial lives and the ﬁnancial tools they demand is even more diﬃcult to ﬁnd Œhough no ﬁnancial inclusion statistics have speciﬁcally traced smallholder households, smallholder families are liely over-represented among the ﬁnan- cially excluded ªural inclusion, often taen as a proxy for smallholder inclu- ž sion, is lower than urban inclusion Œhe urban¤rural disparities reported in index ‘’…… for adults holding an account in sub-Saharan ˆfrica are Ÿ¡ percent urban¤‘… percent rural, and for South ˆsia are Ÿ• percent urban¤Ÿ… percent rural…’ ƒooing more closely at individual countries, estimates using index ‘’…“ data for aistan, Œan—ania, and ©ganda show a similar urban¤rural divide, with estimated account penetration in rural areas of Œan—ania, for example, less …… than half of that in urban areas (see Œable …) †‡ˆ has been woring to build the evidence base on smallholder house- holds as a ﬁrst step in improving smallholders access to and uptae of a range of relevant ﬁnancial tools Œo address the dearth of relevant data to guide ﬁnancial inclusion interventions that target smallholder households, †‡ˆ launched two ma„or demand-side research initiatives with smallholder households to better understand their ﬁnancial lives, agricultural activities, and household livelihood strategies () ﬁnancial diaries with smallholder households (“Small- | 13

14 | SMALLHOLDER DIARIES …‘ holder ‰iaries”) and (ii) nationally representative household surveys and seg- …Ÿ mentations of the smallholder sector Œhese are two complementary research methods Œhe national surveys, which administer a household uestionnaire to a nationally representative sam- ple of about Ÿ,’’’ smallholder households in each country, are wide, while the Smallholder ‰iaries are deep, with a year of very detailed data collection on all income sources, expenditures, and shocs with approximately ‘•’ smallholder households across three countries Œogether, these two lenses on smallholder households detail the granularity of the smallholder sector and identify distinct groups of agricultural households and their needs Œhe results also point to opportunities for service providers to improve ﬁnancial tools for each segment and oﬀer maret information that informs the business case for doing so ¦pening a uniue window onto the livelihoods of smallholder families and how they manage their money, this paper focuses on the Smallholder ‰iaries and presents its methodology and ﬁndings Œhis wor builds on the more gen- eral ﬁndings about low-income households in other ﬁnancial diaries research— TABLE 1: Mobile phones and mobile money among smallholder diaries households (percent),a NOVEMBER 2014 COUNTRY URBAN ADULTS RURAL ADULTS (percent) (percent) Pakistan 11 8 Tanzania 33 14 Uganda 27 19 a. The World Bank Group Global Findex database does not include Mozambique. Data from Uganda was shared to add another perspective from eastern and southern Africa and one of the countries in which CGAP is conducting national surveys of smallholder households. “The 2014 Global Findex database deﬁnes account ownership as having an account either at a ﬁnancial institution or through a mobile money provider” (see Demirguc-Kunt et al. 2014). Source: Author estimates using Global Findex database (2014) and urban/rural population shares (IFAD 2010). BOX 1 RELEVANT FINANCIAL OR AGRICULTURAL SURVEYS OF SMALLHOLDER HOUSEHOLDS Other survey research has included smallholder house- given country source income and manage their ﬁnancial holds in the sample but tends to explore either their ﬁnan- lives and to explore attitudes and perceptions regarding cial or agricultural lives, but not both. ﬁnancial products and services. It is designed to be repre- The Living Standards Measurement Study-Integrated sentative of the adult population of a country and does not Surveys on Agriculture (LSMS-ISA) are designed to focus only on those engaged in agriculture, thus it provides improve the understanding of development in Africa with a only a broad sense of the attributes of adults that generate particular focus on agriculture and the linkages between an income through agricultural activities, and is not sufﬁ- farm and nonfarm activities. Its design does not necessarily cient to facilitate targeted interventions in ﬁnancing small- result in a subsample of smallholder farmers that is repre- holder households. sentative of the population of smallholder famers in a coun- The Agricultural Financial Markets Scoping (AgFiMS) try (although most of the participants in the survey might be diagnostic on ﬁnancial services in the agricultural sector smallholder farmers) and though the questionnaire gener- includes a comprehensive, nationally representative survey ally asks about access to credit, it does not provide a thor- tool focusing on potentially commercially viable agricultural ough understanding of the access to or use of a wider enterprises, including producers, processors, and service spectrum of ﬁnancial services or delve deeply into small- providers, which orients the survey sample toward a focus holder attitudes and ﬁnancial decision-making. on the top-end of the agribusiness market (in terms of The FinScope consumer survey developed by FinMark annual turnover) and therefore excludes the majority of Trust is designed to provide insights into how adults in a smallholder farmers.

INTRODUCTION AND METHODOLOGY | 15 notably Portfolios of the Poor and the ﬁnancial diaries in ¯enya—with a …“ particular focus on smallholder households Œhe next section presents the methodology of the Smallholder ‰iaries, and the subseuent sections report and analy—e its ﬁndings œach section is complemented by a case study on the experiences of a family participating in the Smallholder ‰iaries SMALLHOLDER DIARIES METHODOLOGY ƒaunched by †‡ˆ in ‘’…“, the Smallholder ‰iaries is a research study that col- lected detailed ﬁnancial data from approximately ‘•’ small-scale farming households in –o—ambiue, Œan—ania, and aistan over the course of a year…” Œhe ‰iaries methodology combines in-depth uantitative and ualitative research ªesearch teams met participating families every two wees to collect granular data on cash ﬂows in and out of the household, ﬁnancial tools, assets, …¢ ma„or life events, and attitudes toward agriculture and ﬁnancial services ªesearchers captured the data using a tailored survey instrument ˆt the start, they guided households through three initial uestionnaires that recorded their demographic information as well as income sources, assets, and ﬁnancial tools Œhis baseline information then generated a tailored Smallholder ‰iaries uestionnaire for each family (see igure …) €n all three countries, the research teams visited the Smallholder ‰iaries households using tablet computers to record their ﬁnancial information ‰uring their regular visits every two wees, interviewers captured a com- plete set of individual cash ﬂows from the preceding two-wee period ¦ver the course of the discussion, interviewers ased household members about all of their various income sources, expense categories, ﬁnancial tools, and transac- FIGURE 1: The Smallholder Diaries process Questionnaire 1: Questionnaire 2: Questionnaire 3: household roster; member income sources; financial instruments demographics; living standards physical assets Start/stop Data from 3 initial financial instruments questionnaires used and income sources to generate on-going as needed diaries questionnaire Ongoing Diaries Questionnaire 5 qualitation modules Acquisition/ to capture cash on risk, aspirations, loss/sale of inflows and agriculture, financial physical assets outflows tools, labor decisions from income, household expenditures, and financial transactions Crop tracker Crop tracker to record harvests, (weddings, births, etc.) sale, consumption, and loss of each crop

16 | SMALLHOLDER DIARIES TABLE 2: Sample summary of sources and uses of income: Tanzania Smallholder Diaries household SOURCES USD USES USD Selling beans $5.5 Clothes and shoes $5.8 Self-employment selling chips—income $71.4 Housekeeping supplies $1.2 Resources received from brother $14.4 Posho mill/grinding machine $3.5 Cash at home (withdrawal) $34.5 Public TV/movies $1.7 Supplier credit (borrowing) $5.2 Groceries $1.4 Personal care $1.2 Prepaid phone credit $0.3 Self-employment selling chips— expenses $58.7 Cash at home (deposits) $51.8 Supplier credit (repayment) $5.2 Total sources $131 Total uses $131 tions in the prior two wees ¦ne of the goals of each conversation was to bal- ance the sources and uses of money (see Œable ‘) €f, for example, after the interviewer had ased about the sources and uses of money, the respondent was to then mention that he or she had purchased some fertili—er, the interviewer would follow up with more uestions to understand where the money had come from to mae that purchase, woring to bridge the gap between the uses of money and their sources Œhese gaps were monitored on a monthly basis, with an aim to eep them below ” percent Since in-ind transactions mae meaningful contributions to smallholder household well-being, interviewers also recorded the amount of select in-ind …• transactions and their approximate (self-reported) value Œhe crop tracer in the Smallholder ‰iaries allowed the research team to trac each crop grown by the household and record what was harvested, consumed, sold, lost, and given away Œhis information painted a picture of crop ﬂuctuations in households over the course of year, and clearly illuminated the extent of household depen- dence on the in-ind consumption of its production and the magnitude and implications of any crop loss it experienced Œhe crop tracer data comple- mented the data on sales of livestoc byproducts (eg, eggs, mil) and changes in livestoc assets Œhe Smallholder ‰iaries also traced all separate income sources, such as “agricultural production income” and “odd „ob income” (also called casual income), and carefully noted their stops and starts over time €f a given income source temporarily stopped for a family, for example, and then restarted later in the year, it continued to be traced but was not double-counted as a new income source €n addition, each diﬀerent crop and type of livestoc was counted and traced as a separate source of agricultural production income, given that each may have its own distinct production costs, timing considerations (eg, prepara- tion, planting, harvest), marets, payment modalities, and riss Œhe Smallholder ‰iaries also collected information on the ma„or income, health, and other shocs that households faced, and the related strategies they used to cope with them ¦ver time, when households experienced change—for example in their employment, when they started using new ﬁnancial tools, and when their family experienced important life events such as births and adop- tions—interviewers recorded “change uestionnaires” Œhis information then

INTRODUCTION AND METHODOLOGY | 17 Mozambique. Photo by Erin Scronce. updated the regular ﬁnancial diaries uestionnaires, which were used for that household going forward £uilding on the standard biweely data collection, separate interview modules were used to explore ey topics more thoroughly Œhe ﬁve topical modules explored (i) decision-maing in income choices (eg, agricultural production versus nonagricultural production) (ii) the aspirations of small- holder households, including both farmers’ aspirations and their hopes for their children (iii) the use of and preference for various ﬁnancial tools (based on nowledge of each household’s ﬁnancial portfolio) (iv) decision-maing in agricultural production (eg, why smallholders grow some crops and not others, how and when they decide to sell) and (v) riss, including how they are perceived and prioriti—ed and how families coped with the shocs they experienced Œhe methodology and sample si—e of the Smallholder ‰iaries were designed to generate a rich pool of detailed information and insights on a …¡ targeted population Œhe Smallholder ‰iaries are not statistically represen- tative of smallholder families in participating countries €nstead, through intensive, biweely interviews about the sources and uses of household income and life events over the course of a year, the Smallholder ‰iaries paint a rich picture of the ﬁnancial lives of a speciﬁc group of smallholder house- holds in –o—ambiue, Œan—ania, and aistan œlucidating these diﬀerences is an important step in better understanding the ﬁnancial management chal- lenges facing smallholder households

18 | SMALLHOLDER DIARIES SMALLHOLDER DIARIES SITE AND HOUSEHOLD SELECTION †‡ˆ chose to implement the Smallholder ‰iaries in –o—ambiue, Œan—ania, and aistan because of the variation in their agricultural sectors and the potential to wor with smallholder households with diﬀerent mixes of crops and livestoc, degrees of engagement in agriculture, and maret relationships, all within the varied mobile ﬁnancial services ecosystems across the three countries ™hen selecting the speciﬁc research sites in each of the three Small- holder ‰iaries countries, the goal was to locate at least two villages no more than “’ ilometers apart Œhe villages needed to be accessible by a passable road so researchers could reach them throughout the year, and they also needed to present distinctions in at least a few important aspects, such as prev- alent varieties of crops and livestoc, access to irrigation, and average house- hold income levels • €n –o—ambiue, three villages in ªapale district of northern ˜ampula rov- ince were selected based on strong recommendations from local staehold- ers ™hile some large companies buy cash crops in the province, smallholders tend to practice the subsistence, rain-fed agriculture that is more commonly found throughout –o—ambiue • €n Œan—ania, the Smallholder ‰iaries sites included two villages located in the region of –beya, home to one of the largest farming populations in Œan- —ania –beya sits within the Southern ˆgricultural ‡rowth †orridor of Œan- —ania (Sˆ‡†¦Œ), a region nown for a productive agroecological climate and an array of crops and livestoc armers in the region most commonly produce mai—e, as well as coﬀee and tea, rice, potatoes, pyrethrum, and cas- sava Œo explore the diversity within this region, Smallholder ‰iaries sites were selected in two diﬀerent districts Œhe two selected villages exhibit important diﬀerences in available economic activities, climate, harvest sea- sons, crops, and use of agricultural inputs • €n aistan, the Smallholder ‰iaries were conducted in southern un„ab, within the country’s breadbaset ªice, wheat, and cotton are commonly grown and typically sold through a networ of local commission agents (nown as arthis) and village traders ‡iven the dominance of agricultural middlemen in aistan, two villages in the district of £ahawalnagar were selected as representative of an area with relatively looser connections to agricultural value chains and middlemen ¦nce the villages for the Smallholder ‰iaries were selected, the research teams used a screening process to help identify a range of families with diverse income sources, access to agricultural inputs, wealth levels, crops, and livestoc to par- …ž ticipate in the research • €n Œan—ania and –o—ambiue, households were selected using a participa- tory rural appraisal wealth-raning techniue ™oring with committees of village representatives, the research teams conducted wealth-raning exer- cises to assess the relative wealth of households in village hamlets or subar- eas ©sing the wealth raning, eligible households were selected based on their landholdings, number of crops and harvests per year, use of inputs, and integration with local marets

INTRODUCTION AND METHODOLOGY | 19 • €n aistan, the sample was selected using a traditional screener survey with uestions related to household demographics, crops and livestoc, main income sources, and wealth indicators ˆs a supplement to this pro- cess, village leaders and community representatives were consulted to help ensure local ownership and eliminate households with large landholdings Œhe speciﬁc sites selected for the Smallholder ‰iaries within each country also broadly reﬂect †‡ˆ’s global segmentation of smallholder households noncommercial smallholders, commercial smallholders in loose value chains, and commercial smallholders in tight value chains †‡ˆ’s focus is primarily on the lower-income, more vulnerable groups of farming families (ie, non- commercial smallholders and commercial smallholders in loose value chains), as the relatively smaller percentage of commercial smallholders in tight value chains are liely to beneﬁt from a number of interventions and developments in value chain ﬁnance ˜ampula rovince in –o—ambiue is the least devel- oped of the three regions in terms of agriculture, with almost no use of agricul- tural inputs and limited access to irrigation and marets, unlie the –beya region in Œan—ania Southern un„ab, aistan, has the most robust agricul- tural maret, even in the relatively less-developed (compared with ˜orth of the province) Smallholder ‰iaries site

Tanzania. Photo by Erin Scronce.

SAMPLE DEMOGRAPHICS AND AGRICULTURAL CHARACTERISTICS Œhis section provides more detail on the demographics and agricultural charac- teristics of each Smallholder ‰iaries sample, setting the stage for the results that follow €n terms of income, smallholder households across the samples in –o—ambiue, Œan—ania, and aistan reported average monthly per capita income of ©S±””¢, ©S±…’, and ©S±‘”žž, respectively (see igure ‘), with added income diversity within each country sample †haracteristics of agricultural production also varied across the samples in the three countries, with mared diﬀerences in input use, access to irrigation, and crop and livestoc mix (see Œable Ÿ), as well as their physical assets and ﬁnancial net worth (see igure Ÿ) a FIGURE 2: Average monthly income and consumption (operational expenditures) per capita b (US$) $30.00 $25.99 $25.00 $22.49 $20.00 $15.00 $10.00 $10.00 $7.33 $5.56 $4.18 $5.00 0 Mozambique Tanzania Pakistan Average monthly net income per capita Average monthly household expenditures per capita a. Consumption here includes all operational expenses. It does not include stock purchases for self-owned businesses or farming expenses related to agricultural production. Consumption here does not include in-kind consumption of production from their own farms (these data are covered later in the report). Per capita calculations were derived by dividing by the number of people in the household. b. The timeframes for observations are not exactly the same for every sample and household. In some areas and with some households, the Smallholder Diaries started slightly earlier than others. | 21

22 | SMALLHOLDER DIARIES TABLE 3: Agricultural proﬁle of households in the Smallholder Diaries MOZAMBIQUE SAMPLE— TANZANIA SAMPLE— PAKISTAN SAMPLE— NAMPULA PROVINCE MBEYA REGION BAHAWALPUR DISTRICT (North) (West) (Central) a National GDP per capita (US$) (2014) 602 998 1,334 Monthly per capita net income in sample (US$) 5.56 10.00 25.99 Total number of households in sample 93 86 94 Average number of household members in sample 6 5 7 Female-headed households in sample (%) 17 29 1 Agricultural indicators Average land holdings (hectares) 1.31 0.8 1.18 Use pesticides (%) 13 47 100 Use fertilizer (%) 4 64 100 Use irrigation (%) 2 27 53 (canal) b Use diesel tube wells (%) — — 83 Sample households growing various crops (Percent) Cotton 0 0 37 Wheat 0 7 100 Rice 49 35 96 Peanuts 96 0 0 Potatoes 0 73 0 Maize 65 100 0 Cassava 100 6 0 Beans 97 43 0 Other vegetables 41 15 12 Sample households raising various livestock (Percent) Poultry 53 64 33 Goats 10 26 55 Pigs 3 12 0 Buffalo 0 0 89 Cattle 0 21 39 Donkeys 0 5 26 a. World Bank (2014). b. A tube well is a type of well with a wide tube or pipe a few inches long that provides water from an aquifer. FIGURE 3: : Median physical assets and financial net worth of the Smallholder Diaries samples (US$) $20,000 $18,000 $17,661 $16,000 $14,000 $12,000 $10,000 $8,000 $6,000 $4,000 $2,000 $440 $1,139 $0 $3 $0 –$2,000 –$410 Mozambique Tanzania Pakistan Physical assets owned (median) Financial net worth (median)

SAMPLE DEMOGRAPHICS AND AGRICULTURAL CHARACTERISTICS | 23 MOZAMBIQUE SAMPLE ¦f the three samples in the Smallholder ‰iaries, the one in –o—ambiue includes the most vulnerable population, where respondents identiﬁed a clear “hunger season” between late ‰ecember and –arch €n the ﬁeld site of ªapale district in northern –o—ambiue, the Smallholder ‰iaries sample included žŸ households across three villages, with an average family si—e of six members (see Œable Ÿ) Œhe household head had an average of ﬁve years of schooling and …• percent of households were female-headed Œhe median value of all physical assets owned ‘’ was ©S±““’ (see igure Ÿ) Œhe main receptacles of physical asset wealth were vehicles, such as motorbies, and land (see ˆnnex …) †rop production was notably diverse among the –o—ambiue sample, but the use of agricultural inputs was limited Sample households grew mai—e, cas- sava, peanuts, beans, rice, and a range of vegetables, and they consumed a sig- niﬁcant proportion of their production at home ƒivestoc holdings were mostly limited to poultry, and occasionally goats and pigs Some families channeled water from the river to their plots via informal irrigation schemes, but most depended entirely on rainfall Households in the –o—ambiue ‰iaries tended to exhibit some matriarchal tendencies, reﬂecting traits of the –aua culture, the largest ethnic group in –o—ambiue ˜otably, women can inherit land, and women and men often come to a marriage each with their own plots €n many cases, each is responsible for farming their own small area €n most households in the –o—ambiue sam- ple, the male household head made most decisions regarding agriculture and his wife followed his lead ™omen generally cultivated food crops and men cash crops (though these were very limited in the sample) €n addition to any agricul- tural activities, women were solely responsible for cooing, obtaining and pre- paring food, and raising the children ™omen also organi—ed funerals and other cultural events in the community, and the few respondents that participated in accumulating savings and credit associations (ˆS†ˆs) were women Some men in the sample did not disclose their entire income to their wives, in some cases due to polygamous marriages or other relationships TANZANIA SAMPLE Œhe Smallholder ‰iaries sample in –beya, Œan—ania, was comprised of ¡¢ households, with an average family si—e of ﬁve members ™omen headed about a third of sample households (‘ž percent) and levels of education were modest ‘“ percent of household heads had no formal schooling and ¢ž percent had only completed up to primary school Œhe median asset portfolio of house- holds in the sample reached ©S±…,…Ÿž, and emphasi—ed the primary residence and farmland, with notable holdings of other property and livestoc (see ig- ure Ÿ and ˆnnex …) Œhe sample in Œan—ania was located in two villages, each in a diﬀerent dis- trict ¦ne-third of families came from the ªu„ewa district, where rice is the main cash crop, and two-thirds were drawn from the –beya ªural district, where potato is the dominant cash crop Œhe village in ªu„ewa district sits at a lower elevation and has access to irrigation, unlie the village in –beya ªural

Financial Diaries with Smallholder Families - Page 35

24 | SMALLHOLDER DIARIES Tanzania. Photo by Erin Scronce. district, which is located eight ilometers down a gravel road branching oﬀ the paved road that leads to –alawi amilies typically sold potatoes to independent agents who visit the village Œhe two villages have diﬀerent agricultural cycles, so data related to harvests are presented by village in this report €n the Œan—ania sample, men are typically the household head and the ﬁnal decision-maers ™omen were generally in charge of household activities (eg, cooing, cleaning, caring for children) and contributed to household decision- maing ™omen and men both shared farming responsibilities, but it was more common for women to wor on other farms than men, and for less money ¦ver the course of the Smallholder ‰iaries, women from “¢ diﬀerent households wored as casual laborers on neighboring farms on ¡’ž occasions, earning an average of ©S±‘¢• each time –en from ‘“ households, in contrast, wored as ‘… casual laborers on farms ‘¢¡ times and earned an average of ©S±”Ÿž each time PAKISTAN SAMPLE Œhe research team selected two villages about ‘” ilometers apart, both growing rice as the main cash crop €n the ﬁrst, semi-perennial irrigation canals provided water about six months out of the year, while in the second, smallholders did not have access to the canal and instead depended on tube well irrigation (see Œable Ÿ) Some families in the ﬁrst village used tube well irrigation as well to supple- ment water from the canal Œhe use of chemical pesticides and fertili—ers was universal among the aistan sample and almost all households were connected to the national power grid, though electricity may be available for only a few hours each day

SAMPLE DEMOGRAPHICS AND AGRICULTURAL CHARACTERISTICS | 25 Œhe aistan sample universally grew wheat, and fodder to feed livestoc ˆ smaller percentage of households (…‘ percent) also grew vegetables, such as green chilies and ora –ost families cultivated rice (ž¢ percent) and about a third (Ÿ• percent) grew cotton –ost families had buﬀalos (¡ž percent) and more than half also raised goats (”” percent) Œhe median physical asset value in the aistan sample was signiﬁcant ©S±…•,¢¢… However, ﬁnancial net worth was negative -©S±“…’ at the median (the credit-heavy portfolio in the aistan sample is discussed more in Section ”) armland was the ma„or household asset, followed by the primary residence and livestoc, which was signiﬁcant and formed an important part of household livelihood strategies (see igure Ÿ and ˆnnex …) Œhe family si—e was the largest across the three samples, with an average of seven members per household Sixty-one percent of the household heads had no formal education, and ‘¡ percent completed primary school –en were gen- erally the household head and women were in charge of household chores €n the aistan Smallholder ‰iaries sample, men headed all but one of the ž“ households €n the aistan sample, most husbands and wives tended to mae farming decisions together, but the husband had the ﬁnal say €n some cases, the household head was the husband, and the wife operated her own side business Œhe husband made decisions regarding farming, but the wife operated her busi- ness independently Œhe women in the aistan Smallholder ‰iaries sample generally played a ma„or role in all aspects of agricultural production, from cultivation to harvest, and were primarily responsible for looing after livestoc, which could be uite numerous ™omen conducted the ma„ority of harvesting and cleaning related to cotton picing, rice planting, and wheat cutting, and generally mae up the hired labor force for these tass in the region –ost women in both vil- lages also engaged in casual labor on farms doing similar tass, though earn- ings from temporary agricultural labor were very low at most, women earn ©S±‘ to ©S±Ÿ per day ATTITUDES ABOUT AGRICULTURE IN THE SMALLHOLDER DIARIES SAMPLE orn and raised in agriculture, most adults in the Smallholder Diaries sample belieed that they would continue to farm, and in many cases would continue to supplement agriculture with other income sources ˜early all adults in the sample were born into agricultural households—ž• percent in each of –o—am- biue and Œan—ania, and …’’ percent in aistan—and began woring in agriculture at a young age ˆmong the Smallholder ‰iaries participants in –o—ambiue, their two most important aspirations were building a sturdy house out of uality materials (¢¢ percent) and improving their farm (”• percent) providing an education for their children was a relatively low priority Œhe more commercially active smallhold- ers in the Œan—ania and aistan samples viewed expanding or improving their farm as a top life goal (•“ percent and ¢ž percent, Photo by Mwangi Kiribi.

26 | SMALLHOLDER DIARIES respectively) –any households in both countries also wanted to purchase their own farm euipment (”” percent and Ÿ’ percent in aistan and Œan—ania, respectively), educate their children (“” percent and ““ percent in aistan and Œan—ania, respectively), and support their children in achieving their future goals (“” percent and …¢ percent in aistan and Œan—ania, respectively), which speas to the multiple ob„ectives and trade-oﬀs facing smallholder households ew respondents from the –o—ambiue sample could envision a life outside of agriculture in the next few years (” percent) or were attracted to the possi- bility of migrating to the city (Ÿ” percent) Œhe largely noncommercial small- holders in the –o—ambiue sample voiced concern with meeting their most basic needs—food and shelter—and considered the main advantage of woring in agriculture the guarantee of providing at least a minimum level of food for their family Œhe thought of not being able to mae this contribution to their survival worried them Œhe Smallholder ‰iaries sample in –o—ambiue may have been less attracted to migration out of agriculture due to limited alterna- tives to earn income, modest levels of education, and few social connections outside the village ªespondents in the Œan—ania and aistan samples also had limited desire to leave agriculture ¦nly • percent and ž percent, respectively, envisioned them- selves woring primarily outside of agriculture within the next ﬁve to …’ years, and ‘Ÿ percent and Ÿ¡ percent, respectively, would move to urban areas if ﬁnan- cial constraints were not an issue amilies in the Œan—ania and aistan Small- holder ‰iaries shared a range of reasons why they stayed in agriculture and remained in rural areas Œhe sample of smallholders in Œan—ania felt that people in urban areas had access to everything, such as health care, education, and very important social services, but they viewed life in the city as full of “bad inﬂu- ences” €n the rural areas, it was also “easier to get food” and “everything is cheaper” €n addition, rural areas were “all € now” for many respondents €n aistan, one smallholder in the sample explained that nonfarming „obs were not guaranteed “™hether you have a „ob or not, whether you have a business or not, you can mae do with the land” ˆnother explained that he doesn’t have the sills reuired for other „obs “™e don’t have other options € have been doing this wor my whole life € am an expert in farming” ƒooing to the future, not everyone wanted their children to follow their footsteps into agriculture, despite their own desire to eep woring in agricul- tural production Œhe ma„ority of the –o—ambiue sample wanted all or some of their children to continue farming (¡” percent), while a smaller proportion of the Œan—ania and aistan samples did so (‘“ percent and ”’ percent, respec- tively) (see igure “) amilies that wanted their children to pursue other activ- ities emphasi—ed that nonfarming „obs could provide a steadier or higher income source, which also reﬂected that households were relying on multiple sources of income and not only on their agricultural production

SAMPLE DEMOGRAPHICS AND AGRICULTURAL CHARACTERISTICS | 27 FIGURE 4: “If you have children, do you want them to continue farming” MAY 2015 100% 80% 60% 57% 61% 49% 40% 36% 28% 20% 15% 18% 14% 6% 0% Mozambique Tanzania Pakistan Yes, all of them Yes, but only some of them No Sample size per country includes only households with children: Mozambique n=61; Tanzania n=62; Pakistan n=80.

Smallholder Diaries interview, Mozambique. Photo by Erin Scronce.

1. INCOME SOURCES AND THE ROLE OF IN-KIND CONSUMPTION Agricultural production was just one source of cash income for Smallholder Diaries households. But their numerous nonagricultural production activities provided the majority of household net cash income at the median. When in-kind consumption of crops produced at home was factored into total household income, agricultural production income did increase markedly and provided approximately half of total household income at the median. Casual labor was also an important source of income for smallholder families, and in many cases it, too, involved tasks related to agriculture. ™hen the Smallholder ‰iaries began, all respondents in the sample indicated that agriculture was their primary source of income, cash or in-ind, though they were balancing a range of income sources in and out of the agricultural sector Œhey made it a priority to earn money from a range of sources due to the irregularity of their agricultural income and relatively more stable nonfarm income £ut they still self-identiﬁed as agricultural households and indicated that this was their most signiﬁcant source of income Œhe data, however, showed a more nuanced picture Smallholder families in the sample had numerous sources of cash income, tending to fall into three categories: (i) agricultural production, (ii) casual labor, and (iii) other oﬀ-farm, nonagricultural sources such as managing a ‘‘ small business, receiving remittances, or engaging in wage labor ˆt the median, households had a total of eight income sources in –o—ambiue, …… in Œan—ania, and nine in aistan‘Ÿ Sample households in –o—ambiue and Œan- —ania had far more sources of nonfarming income, independent of their own agricultural production €n the Œan—ania sample, for example, at the median, households had two sources of income in agricultural production and nine related to other, oﬀ-farm activities (see igure ” and †ase …) Families in the Smallholder Diaries also earned the majority of their house- hold net cash income from their numerous nonagricultural production actiities ˆmong the sample families, the median proportion of household net cash income (ie, revenue less any associated expenses, such as stoc purchases for side businesses) from nonagricultural production sources was žŸ percent in ‘“ –o—ambiue, •“ percent in Œan—ania, and ”¡ percent in aistan (see igure ”) ut focusing only on cash income underplays the importance of agriculture to smallholder households Œhe relative importance of agricultural produc- tion income increases maredly when household consumption of crops is | 29

30 | SMALLHOLDER DIARIES ‘” included in the analysis €n the sample in Œan—ania and aistan, the median proportion of household income derived from crop production increased from ‘¢ percent to “¢ percent and “‘ percent to ”Ÿ percent, respectively, when also considering in-ind consumption £ut the change was most dramatic in –o—am- biue, „umping from • percent to “ž percent ƒooing more closely at the sample in –o—ambiue, the žŸ percent ﬁgure for total household net cash income from nonagricultural production income sources may seem high, especially among a sample that self-identiﬁes as farm- ing households £ut consider that the sample in –o—ambiue was comprised largely of noncommercial smallholders they generated little surplus to sell (ie, revenue) and purchased very little fertili—er or pesticide (ie, spending on agri- cultural inputs), and thus reported very low incomes related to their agricul- tural production Œhey instead relied largely on other income sources for cash (see £ox ‘) £ut these households were heavily involved in agriculture, produc- ing a wide variety of crops, consuming them at home, and trading it with friends and family, outside the cash economy FIGURE 5: Smallholder Diaries: Household income from agricultural and nonagricultural production JUNE 2014–JULY 2015 Pakistan Income sources 4 5 Proportion of 42% 58% income Income sources 5 5 w/in-kind Proportion of income 53% 47% w/in-kind Tanzania Income sources 2 9 Proportion of 26% 74% income Income sources 3 9 w/in-kind Proportion of income 46% 54% w/in-kind Mozambique Income sources 2 6 Proportion of 7% 93% income Income sources 6 6 w/in-kind Proportion of income 49% 51% w/in-kind Agricultural production income Nonagricultural production income (1) Median number of household income sources (2) Median proportion of total household net cash income (3) Median number of household income sources, factoring in in-kind consumption (4) Median proportion of total household income factoring in in-kind consumption Note: Each crop or livestock byproduct (e.g., milk, eggs) that was sold at least once in the Smallholder Diaries is considered a distinct source of agricultural production income. When tracking in-kind consumption, the Smallholder Diaries recorded only activities or transactions related to crops, not livestock byproducts.

INCOME SOURCES AND THE ROLE OF IN-KIND CONSUMPTION | 31 BOX 2 a ACTIVE IN AGRICULTURE, BUT EARNING CASH OFF THE FARM: THE NTHIAS (MOZAMBIQUE) José and Thalia Nthia, a family in the Smallholder Diaries market where José sold groceries and airtime provided their sample in Mozambique, lived with their two daughters, main source of income, though this ﬂuctuated over the year aged four and one, in a brick house on a very small plot of (see Figure B2-1). land. José, 21, had studied until the ninth grade and Thalia, The Nthias faced a major income shock in April. Thieves 19, did not ﬁnish primary school. They used half of the area broke into the area where José had stored his stock of gro- around their house to plant crops and raise livestock, includ- ceries and stole everything. To make up for the cash short- ing goats. The Nthias also cultivated another larger plot of fall, he sold one of his goats (see the red line in Figure B2- land, which they considered their main farm, on a hectare 1). José then identiﬁed another source of income: buying close to a neighboring village. Thalia maintained the family peanuts from his neighbors to sell in Nampula, the regional farm, while managing all household tasks and caring for the children. In this area of Mozambique, it was typical for hus- urban center, at a higher price. bands and wives to each maintain their own macshamba, After the theft, Thalia also looked for additional work to or farming plot, and for women to do much of the daily supplement their income. She started weeding on a neigh- farm labor. bor’s farm, but this was short-lived. She was injured and, as José and Thalia were active in agricultural production, she recovered, Thalia struggled to keep up with the work but they typically did not sell any of their livestock or the ﬁve at the neighbor’s on top of the demands of their own farm crops they grew. The household viewed their crops and live- and family. Finally the Nthias decided that Thalia would stop stock primarily as sources of food for their own consumption, working for their neighbor, ending this source of income not monetary income. They earned their cash income from from casual farm labor, and focus on the needs of their fam- work away from their own farm. A small stall in the village ily and its agricultural activities. FIGURE B2-1: The Nthias various income sources over time (US$) JULY 2014–JUNE 2015 $250 $200 $150 $100 $50 $0 Jul Aug Sep Oct Nov Dec Jan Feb MarApr May Jun Sell airtime Sell groceries Work on neighboring farms Buy and sell peanuts Sell animals a. The names of all respondents have been changed to protect their privacy. ”asual labor was an important source of income for smallholder families, ‘¢ and much of it was related to agriculture €n all three samples, casual labor was an important source of cash income (see igure ¢) ˆmong the dominant sources of net cash income, Ÿ’ percent, …ž percent, and …“ percent of total net cash income came from casual labor in –o—ambiue, Œan—ania, and aistan, respectively ˆ signiﬁcant proportion of opportunities for casual labor was related to agriculture, including tass such as planting and harvesting crops €n terms of its relative value, casual labor income from wor related to agriculture

32 | SMALLHOLDER DIARIES was ‘” percent in –o—ambiue, ”‘ percent in Œan—ania, and Ÿ¢ percent in ai- stan of total income from casual labor „ signiﬁcant proportion of households in all three samples gae and receied resources among family and friends, in both cash and in-ƒind Œhe Small- holder ‰iaries traced both cash or in-ind contributions to income from friends or family (ie, “resources received”), as well as the payments that respondents made to family and friends outside the household in their wider social networ (ie, “resources given”) (see igure •) ˆmong the Smallholder ‰iaries sample in –o—ambiue, •” percent of households received resources, ¡Ÿ percent of which was in cash •“ percent of households also gave resources to family and friends, “’ percent of which was in-ind (see igure •) €n the Œan- —ania sample, žŸ percent received and ¡ž percent gave resources (see £ox Ÿ), FIGURE 6: Mean proportion of income from various sources Percentage of Value at the Sample Level 100% 6% 90% 17% 10% 13% 80% 8% 7% 11% 8% 70% 9% 9% 60% 15% 50% 21% 14% 19% 40% 30% 30% 49% 20% 39% 10% 15% 0% Mozambique Tanzania Pakistan Agricultural income (net) Resources received Casual labor Regular employment Self-employment (net) Other income Proportions of total resources received and resources given in cash FIGURE 7: and in-kind (percentage) Pakistan 76% 24% Tanzania 68% 32% Mozambique 83% 17% Pakistan 29% 71% Tanzania 80% 20% Mozambique 60% 40% 0% 10% 20% 30% 40% 50% 60%870%90% 0% 100% Resources received (cash) Resources received (in-kind) Resources given (cash) Resources given (in-kind)

INCOME SOURCES AND THE ROLE OF IN-KIND CONSUMPTION | 33 BOX 3 THE MANY ROLES OF KINDI IN HOUSEHOLD CASH FLOWS: RACHEL (TANZANIA) In the village focused on potato production in Tanzania, the local beer made from maize, kindi, played a signiﬁcant role in household cash ﬂows in the Smallholder Diaries sample. Households bought and sold it, and also gave and received it as donations or gifts. At the aggregate, kindi accounted for 31 percent of the total value of in-kind resources received over the course of the Smallholder Diaries in Tanzania. The experience of Rachel’s household over the year demonstrates the different ways that kindi was used. Rachel, 61, lived with her six grandchildren, ranging in age from six to 18. She grew pota- toes and maize, and considered herself a farmer. She struggled with pain in her legs and could not work on her farm or walk for a long time, which led her to minimize her investment in farm- ing. As a result, her yields suffered and she earned less from her agricultural production than her neighbors. Informal savings and credit groups were concerned that she could not repay, and she was typically unable to borrow from them. Thus, to make extra money and generate “quick cash,” Rachel relied on selling kindi. Sales of kindi made up 95 percent of her total income in April, and approximately 30 percent of her income in January, February, and May (see Figure B3-1). A signiﬁcant portion of Rachel’s monthly income came from resources received in-kind (i.e., monetary or in-kind contributions provided to respondents through their social networks), including kindi. In No- vember and June, for example, resources received in-kind were her only source of income; in March, resources received in cash and in-kind made up her total income. Rachel also bought kindi for herself and sometimes received it as a gift. Over the course of the year, Rachel’s spending on kindi for her own consumption varied, falling in periods when she received it as a gift, such as March and April. FIGURE B3-1: Total net income by type: Rachel (TANZANIA) JULY 2014–JUNE 2015 (PERCENTAGE) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% JulAug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Other income sources Kindi sales as a percentage of total monthly incomes Resources received in kind as a percentage of total monthly income while in aistan, “ž percent received resources and Ÿ¢ percent oﬀered them, more than •… percent of which was in-ind €n the Œan—ania sample, most of the total amount of both resources received and given circulated within the com- munity (•¢ percent, or •Ÿ percent by volume), while the aistan sample received and sent the most long-distance remittances (¡” percent, or ”¢ percent by volume)

34 | SMALLHOLDER DIARIES CASE 1 JUGGLING MULTIPLE INCOME SOURCES TO REPAY DEBTS: BERTHA (TANZANIA) £ertha too part in the Smallholder ‰iaries in Œan—ania She lived close to the main road in her village Her husband died, leaving her to raise their ﬁve chil- dren, ranging in age from four to ‘’ years old, by herself £ertha was nown as a good neighbor and always helped those who came to her with a problem She wored hard to provide for her family, leaving her house early in the morn- ing and returning late at night She focused on agricultural production and relied primarily on selling potatoes, eggs, and sometimes mai—e She also had a number of other income sources £ertha collected and sold grasses and timber in the neighborhood, provided casual labor on other farms, and from time to time she received money from relatives outside her household (ie, “resources received”) (see igure ¡) €n addition to providing for her family, £ertha was also motivated to tae on a number of „obs because she had several outstanding debts with diﬀerent groups Œhese came about in ‘’…“ when her sister-in-law needed an operation during childbirth and her brother borrowed a large amount (©S±Ÿ“Ÿ) to pay for it ©nfortunately, soon after taing this loan, £ertha’s brother fell ill and stopped maing payments on it Œhe lender wanted to sue him and sei—e his assets to recover the debt, but £ertha stepped in to help her brother She started borrow- ing from four informal groups, two local shops, and four other family members and friends to help repay the loan FIGURE 8: Income from multiple sources: Bertha (TANZANIA) (US$) $70 $60 $50 $40 $30 $20 $10 $- JulAug SepOct Nov Dec Jan Feb Mar Apr May Jun Casual labor Agricultural income Self-employment Resources received Rental income Broker income

INCOME SOURCES AND THE ROLE OF IN-KIND CONSUMPTION | 35 Œhrough „ust one informal group, £ertha was able to cover the interest pay- ments over the course of the year £ut only after she sold her potatoes could she repay the principal in ebruary ™hen £ertha could not cover the interest pay- ments for the loans from her regular income sources, she too on additional casual wor to earn money Her income barely covered her debts plus her household expenses (see igure ž) €n ebruary, when she earned the most from selling potatoes, all of her income was dedicated to loan repayments She felt obligated to engage in numerous types of wor, without regard to the „ob or its location, to cover her household expenses, her children’s basic needs, and her debts to the groups Some of £ertha’s income from selling potatoes was never actually received €n ebruary, she sold potatoes to an agent and earned a lump sum of ©S±”• €n ebruary she sold an additional ©S±¡¢ of potatoes to another agent, with the agreement that he would pay her later, after his buyer paid him ©nfortunately as of Žune, the second agent had never returned with her payment, and £ertha had given up on this payment Since £ertha earned ±¡’ in agricultural income in total that year, this means that she lost more than half of her potential annual agricultural income in a single bad transaction Her experience highlights the vulnerability of commercial smallholders in loose value chains FIGURE 9: Income barely covers debts and household expenses: Bertha (TANZANIA) (US$) $180 $160 $140 $120 $100 $80 $60 $40 $20 $0 Jan Feb Mar Apr May Jun July Aug Sep Oct Nov Dec Agricultural income earned Other income earned Debt paid Household expenses

Farmers working with Wasil Foundation, Pakistan. Photo by Ayesha Vellani.

2. PATTERNS OF AGRICULTURAL PRODUCTION AND SALES Income from agricultural production in the Smallholder Diaries sample showed three distinct patterns of harvest, sales, and consumption: (1) two major annu- al harvests but very limited or nonexistent sales; (2) one major annual harvest, with varied sales into loose value chains; and (3) two major annual harvests, with immediate sale into tight value chains. If and how smallholder households con- sume or sell their crops and the nature of their connections to value chains, for example, have important implications for the roles that ﬁnancial tools can play in their lives and how they are tailored to household circumstances. MOZAMBIQUE SMALLHOLDER DIARIES: TWO MAJOR ANNUAL HARVESTS, BUT VERY LIMITED OR NONEXISTENT SALES Smallholder families in the oambi ue sample were “net consumers” of their agricultural production (ie, they consumed more of their agricultural output than they sold) Households harvested continually between ˆpril and September (see igure …’, green line), with the two ma„or harvests in –ay¤Žune (cassava) and ˆugust¤September (mai—e, vegetables, other grains) amilies maintained a relatively smooth level of consumption throughout the year (orange line) Œhey generally stored their harvests in bags and traditional bam- boo silos in the house Sales of crops and livestoc were FIGURE 10: Mozambique Smallholder Diaries: Monthly average value of limited among Smallholder ‰iaries fam- crops (self-reported), JULY 2014–JUNE 2015 (US$) ilies in –o—ambiue ƒargely noncom- $50 mercial smallholder households, many $45 families in the sample did not sell any- $40 thing over the entire year of data collec- $35 tion (see dotted red line) Œhey were $30 able to consume smoothly what they $25 produced, but they were unable to do $20 much else with their harvest Œhey did $15 not sell their crops for the cash needed to $10 $5 buy other foods, diversify their diet, or $0 meet other household needs July Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Harvest Consumed Sold/traded Lost/given Green “Harvest” line: Self-reported value of harvested crops. Dotted red “Sold/Traded” line: Self-reported value of crops that were sold or traded. Orange “Consumed” line: Self-reported value of crop production that was consumed at home. Blue “Lost/Given” line: Self-reported value of any crop amounts that were lost (due to weather, pests, theft, etc.), given away, or traded. Value of crops stored not included. | 37

38 | SMALLHOLDER DIARIES BOX 3 TANZANIA SMALLHOLDER DIARIES: ONE MAJOR ANNUAL HARVEST, VARIED SALES INTO LOOSE VALUE CHAINS Smallholder households in the anania sample were “net sellers” in loose alue chains (ie, they generally sold more of their agricultural output than they consumed) Œhese households typically had one ma„or harvest of a cash crop (eg, rice, potatoes) each year, which would be sold for cash and, to some degree, consumed at home (see igure …… for the sample producing potatoes) ¦verall, sales of their agricultural production were fairly lumpy (ie, they occurred in distinct periods over the year, not continuously) and reﬂected the timing of the main harvest ˆlmost all of these respondents sold directly to village-level agents and¤or larger aggregating buyers further along the value chain, while those with ‘• the means also transported their production directly to maret (see £ox “) FIGURE 11: Tanzania Smallholder Diaries: Monthly average value of crops (self-reported): Potato production village, JULY 2014–JUNE 2015 (US$) $200 Potato harvest $150 $100 $50 $0 July Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Harvest Consumed Sold/traded Lost/given Green “Harvest” line: Self-reported value of harvested crops. Dotted red “Sold/Traded” line: Self-reported value of crops that were sold or traded. Orange “Consumed” line: Self-reported value of crop production that was consumed at home. Blue “Lost/Given” line: Self-reported value of any crop amounts that were lost (due to weather, pests, theft, etc.), given away, or traded. Value of crops stored not included. BOX 4 HARVEST AND CONSUMPTION OVER THE YEAR: THE LUHENDES (TANZANIA) The experience of Mathias and Zaituni FIGURE B4-1: Aggregate harvest and consumption of agricultural Luhende offers a good representation production for the Luhendes (self-reported) (TANZANIA), JULY 2014–JUNE 2015 (KG) of the sample in the village focused 400 on potato production. The Luhendes 350 sold most of their potatoes in January 300 soon after the harvest (see Figure B4-1, dotted red line). They consumed some 250 of their production during the year 200 (orange line); their maize was destroyed 150 by livestock in September (blue line). 100 50 0 July Aug Sep Oct Nov Dec JanFeb Mar Apr MayJun Harvest Consumed Sold/tradedLost/given

PATTERNS OF AGRICULTURAL PRODUCTION AND SALES | 39 PAKISTAN SMALLHOLDER DIARIES: TWO MAJOR ANNUAL HARVESTS, IMMEDIATE SALE INTO TIGHT VALUE CHAINS Smallholder households in the ‚aƒistan sample were net sellers in tight alue chains, consuming an even smaller proportion of their production than in ‘¡ the Œan—ania sample ˆfter each ma„or harvest in ¦ctober (cotton and rice) and in –ay (wheat), they were usually obligated to sell their output immedi- ately bac to middlemen to repay debts for the costly agricultural inputs they had ﬁnanced (see igure …‘, dotted red line, and £ox ”) Œhis resulted in large spies of sales from agricultural production over the year Œhe sample in aistan sold into a complex value chain ™oring directly with the smallholders are the beoparis, local villagers who buy crops directly from farmers and do not provide agricultural credit or inputs Œhey oﬀer farmers a lower price than atcha (informal) arthis, the next step in this value chain, but they could pay farmers more uicly ˜evertheless, the ma„ority of the sample sold directly to arthis, and in many cases to a speciﬁc arthi family that had been woring with their family for generations Œhe arthi also held a good portion of their cash savings and provided pesticides and fertili—er on credit –oving up the value chain, atcha arthis then sell to the pua (formal) arthis, who in turn sell to factories and mills or broers Tanzania. Photo by Erin Scronce.

40 | SMALLHOLDER DIARIES FIGURE 12: Pakistan Smallholder Diaries: Average value of crops (self-reported), JULY 2014–JUNE 2015 (US$) $1,200 $1,000 Wheat Rice/cotton harvest $800 harvest $600 $400 $200 $0 July Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Harvest Consumed Sold/traded Lost/given Green “Harvest” line: Self-reported value of harvested crops. Dotted red “Sold/Traded” line: Self-reported value of crops that were sold or traded. Self-reported value of crop production that was consumed at home. Orange “Consumed” line: Blue “Lost/Given” line: Self-reported value of any crop amounts that were lost (due to weather, pests, theft, etc.), given away, or traded. Value of crops stored not included. BOX 5 WAITING TO SELL UNTIL PRICES INCREASED: SAMIR AND ZAINAB Samir and Zainab, participants in the Smallholder Diaries in In September, Samir earned some cash from his work as a Pakistan, grew rice, wheat, and fodder and sold milk. Zainab tailor and that helped improve the household’s ﬁnancial situ- also generated income for the household from casual labor ation. Finally, in November, Samir harvested the rice, but the (see Figure B5-1). In June 2014, right before the research prices in the market were very low. If he had sold the rice at began, heavy rains destroyed the rice they had planted. The that time, then he would have suffered a ﬁnancial loss on his family replanted rice, spending money on additional agricul- investment. Instead he stored some of the rice to sell later in tural inputs; July and August were difﬁcult for the family. In December after the price increased. this period, and in the winter months, when the household After their harvests were cut and sold, Samir continued was under intense seasonal pressure to invest in their agri- working as a tailor and Zainab did casual labor. Samir said cultural production, Samir and Zainab spent much of their the family needed to do this additional work. “When we money on agricultural inputs. Agriculture was such a high don’t have crops, we make ends meet with credit [from local priority and related expenses were so high that they could stores]. There is no other source where we can get additional not afford to buy a school uniform for one of their sons and income. That’s why we have to rely on credit.” he was sent home from school for a few days. FIGURE B5-1: Sources of household income: Samir and Zainab (PAKISTAN) (US$) $700 $600 $500 $400 $300 $200 $100 $0 –$100 –$200 Jul-14 Aug-14 Sep-14 Oct-14 Nov-14 Dec-14 Jan-15 Feb-15 Mar-15 Apr-15May-15Jun-15 –$300 Net agricultural production income Net income from tailoring (revenue less expenses) Casual labor

PATTERNS OF AGRICULTURAL PRODUCTION AND SALES | 41 CASE 2 THE BENEFITS AND BURDENS OF A TIGHT VALUE CHAIN: THE MALIKS (PAKISTAN) Saif –ali, a participant in the Smallholder ‰iaries whatever they had on agricultural production Saif in aistan, earned agricultural production income spent a total of ©S±Ÿ’“ on fertili—er, pesticides, and from selling mil from his cows and growing wheat seed in this period ¦verall the family earned nega- and rice, which he sold to three arthis (see igure tive income (ie, spending on farming inputs …Ÿ) ™henever he needed funds, he could borrow exceeded revenue from farm sales) during those from any of the arthis, but he ept multiple rela- months, due to these signiﬁcant expenditures on fer- tionships going with arthis in the event that one of tili—er and pesticides Œo mae ends meet, the family them would not allow him to borrow the amount he relied on credit and sale of mil needed ests destroyed much of their rice crop in ¦cto- Saif also looed for oﬀ-farm wor to supplement ber, but ˜ovember was the best month for the fam- his income from agricultural production “Œhe ily ˜ovember is an important time to harvest in the money we mae from our crops alone is not enough,” area, and his sons were able to wor on other farms he explained “Œhat’s why we wor on other farms to help mae up for the loss in ¦ctober ¦ur crops do suﬀer a little bit, since we aren’t able to Žanuary was again diﬃcult as the family had no give them our full time” ™hat other sources of cash Saif had been earning money almost every income he could ﬁnd were unstable Saif was the month by cutting trees for the forest department main breadwinner, and the family worried about £ut at the end of ‰ecember, the forest department their economic condition told him it did not need him anymore and that it Œhe most diﬃcult periods for the –alis were in would call him again only if wor was available ˆll Žuly and ˆugust and the winter months, the stretches household expenses were purchased on credit between harvests ‰uring these periods, they spent FIGURE 13: Main financial transactions of the Maliks (PAKISTAN), JULY 2014–JUNE 2015 (US$) November December $700 The family experiences a Saif loses his job good financial month. Saif at the forestry April $600 was able to sell his rice crop department. The family’s income and his sons were able to picks up again with $500 work on others’ farms the wheat harvest. during the harvest. January $400 The family has no cash in July/August January. All of their $300 He spends US$304 on expenses are on credit. farming inputs $200 $100 $0 JulAug SepOct NovDec JanFeb Mar Apr MayJun –$100 –$200 –$300 Net income Household expenditures

Pakistan. Photo by Erin Scronce.

3. RELATIONSHIPS BETWEEN INCOME VOLATILITY AND AGRICULTURAL PRODUCTION Agricultural production income was markedly more volatile than other sources of income across the Smallholder Diaries sample, and overall household income volatility depended on the balance between the two. Expenditures ﬂuctuated considerably as well, due to major household shocks or events, or regular, sig- niﬁcant expenses such as school fees. Despite income diversiﬁcation outside of agricultural production and smoothing strategies, the agricultural cycle still ex- erted a strong inﬂuence over the ﬁnancial lives of smallholders and most ﬁnancial hardships were related to agriculture. Smallholder households in the sample knew from past experience what ﬁnancial problems to anticipate, but they lacked the ﬁnancial tools to mitigate them or meet their needs. „gricultural production income was marƒedly more olatile than other sources of income in all three samples, and a household’s oerall income olatility depended on the balance between agricultural production and nonagricultural production income (see igure …“) Œhe wide range of income sources outside of crop and livestoc production did dampen the eﬀects of the agricultural cycle on sample households, but only to a point €t was not elimi- nated entirely ªespondents in all three countries struggled the most, with their ﬁnances and nutrition, in the months between harvests €n the –o—ambiue sample, with very limited crop sales and reliance on other sources of cash income, families experienced less severe ﬂuctuations in overall income, but their paucity of ﬁnancial tools presented other challenges or the sample in aistan, month-to-month agricultural production income was highly volatile (“”¡ percent relative to average income) €t swung from lows well below —ero, when there were ma„or expenditures on inputs, to signiﬁcant highs, when immediately after harvest they sold their output (see igure …”) (Since these expenses are related to agriculture income, they are embedded in the green income line) Œhe involvement of households in other sources of income was relatively more stable, peaing in –arch when farming activity ui- ets down €mportantly, though these other sources of income do not generate periods of large negative income (ie, deﬁcit periods where expenses related to agricultural production or running a business exceeded revenue), total income volatility (red line) still tracs agricultural production income volatility (see igure …” and £ox ¢) †‡penses were smoother than income, but still ﬂuctuated to some e‡tent with income in all three samples ˆt the median, expenditures per month var- ied ¢¡ percent in –o—ambiue, ¡’ percent in Œan—ania, and ”ž percent in ai- stan among Smallholder ‰iaries samples £ut this volatility of expenditures | 43

Financial Diaries with Smallholder Families - Page 58

44 | SMALLHOLDER DIARIES FIGURE 14: Volatility of income: Median standard deviation of monthly income relative to average monthly income, a JULY 2014–JUNE 2015 500% 458% 400% 300% 281% 222% 213% 200% 120% 100% 89% 84% 101% 76% 0% b Tanzania Pakistan Mozambique Agricultural production volatility Nonagricultural production volatility Total income volatility a. Relative standard deviation of income = (Standard deviation of monthly income * 100)/ Average monthly income. Standard deviation of monthly income represents the amount by which a household’s income deviates from the average monthly income of that household. b. In the Mozambique sample, agricultural production income and nonagricultural production income have an inverse relationship, so much so that the medians of the two cancel each other out a bit. Thus total median income volatility is lower than the two parts that compose it. FIGURE 15: Pakistan Smallholder Diaries: Value of agricultural and nonagricultural production income at the sample level, JULY 2014–MAY 2015 (US$) $50,000 $40,000 $30,000 $20,000 $10,000 $0 July Aug Sep Oct Nov Dec Jan Feb Mar Apr May –$10,000 –$20,000 Agricultural net income Nonagricultural net income Total net income should not necessarily always be considered negative €n certain months people have ma„or expenses such as school fees or may choose to spend more for expen- sive items Œhat said, data on spending across the samples showed that basic expendi- tures, such as food, public transportation, clothing, and education, accounted for the largest share of household budgets at the median (see igure …¢) House- hold budgets are tight, leaving little room to pay for other lump-sum or unex- pected needs €n its extreme form, the inability to smooth expenses can translate into the inability to cover some basic needs ¦ver the course of the year, notable

RELATIONSHIPS BETWEEN INCOME VOLATILITY AND AGRICULTURAL PRODUCTION | 45 BOX 3BOX 6 SEASONALITY OF AGRICULTURAL PRODUCTION FOR THE MALIKS (PAKISTAN) The pattern of agricultural production that Adil Malik followed was typical of the sample of Smallholder Diaries households in Pakistan. Adil grew fodder, rice, and wheat and sold buffalo milk. His sister did casual work, and his mother did embroidery. Adil also took care of his two sisters, one divorced and one widowed, and their three children. In June 2014, before data collection for the Smallholder Diaries began, part of the Malik’s wheat crop was destroyed in a hail storm. From September to October, the fam- ily did well, thanks to the sale of the cotton crop (see Figure B6-1). The family’s second major annual harvest, wheat, occurred in May. Most of the wheat was sold, but they also consumed some and gave some away. FIGURE B6-1: The Malik’s aggregate harvest and consumption over time (self-reported), JULY 2014–JUNE 2015 (KG) 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0 July Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Harvest Consumed Sold/traded Lost/given Green “Harvest” line: Self-reported value of harvested crops. Dotted red “Sold/Traded” line: Self-reported value of crops that were sold or traded. Orange “Consumed” line: Self-reported value of crop production that was consumed at home. Blue “Lost/Given” line: Self-reported value of any crop amounts that were lost (due to weather, pests, theft, etc.), given away, or traded. Value of crops stored not included. percentages of households in each sample sacriﬁced visiting a doctor or buying medicine when needed (see igure …•) Despite income diersiﬁcation and smoothing strategies, the agricultural cycle still e‡erted a strong inﬂuence oer the ﬁnancial lies of smallholders Smallholders were ased, “€n the past …‘ months, during which months did your family struggle the most with money?” to identify the typical “hardship months” ˜otably, most of the reasons cited for ﬁnancial hardships were tied to the agricultural cycle, such as “waiting to harvest crops,” “expenses related to agriculture,” and “already sold all crops” (see igure …¡) ªespondents identiﬁed similar reasons when ased why they had cut down on food consumption or worried about food during certain months Farmers in the Smallholder Diaries ƒnew from past e‡perience what money problems to anticipate, but they still lacƒed the tools to smooth consump- tion €n all three samples, household operational expenses were smoother than | 45

46 | SMALLHOLDER DIARIES FIGURE 16: Median share of consumption expenditures on various household needs (Percentage) Food 46% 39% 36% Clothing 17% 7% 11% Transportation 12% 7% 9% 3% Medical Services 6% 7% Drinks/Cigarettes 4% 6% 3% Education 1% 10% 5% Home Fixes 11% 7% Other 17% 14% 22% 0 10% 20% 30% 40% 50% Mozambique Tanzania Pakistan FIGURE 17: Share of households forced to forgo medical care during the study JULY 2014–MAY 2015 (Percentage) 80% 70% 60% 15% 50% 4% 11% 5% 40% 5% 6% 6% 30% 5% 5% 7% 5% 2% 2% 0% 9% 20% 38% 5% 5% 7% 2% 38%37% 5% 10% 28% 2% 3% 24% 23% 15% 19% 14% 10% 12% 0% JulAug Sep Oct Nov Dec Jan Feb Mar Apr May Mozambique Tanzania Pakistan income, but expenditures still ﬂuctuated to a degree with income Œhe samples in both –o—ambiue and Œan—ania most strongly exhibited a “spend-as-you- ‘ž go” expenditure pattern (see igures …ž and ‘’) €n –o—ambiue, income and expenditures dropped to the lowest during the lean hunger season ˆnd in the Œan—ania sample in the village focused on rice production, income and expenses dipped to their lowest point between ebruary and ˆpril, the months that respondents had earlier identiﬁed as their most diﬃcult, when crop stocs and income opportunities related to farming typically dwindle (see £ox •) Œhe aistan sample experienced the largest swings in income across the three countries in the Smallholder ‰iaries, but these households maintained relatively smoother levels of household expenditures Œhey had access to more

RELATIONSHIPS BETWEEN INCOME VOLATILITY AND AGRICULTURAL PRODUCTION | 47 FIGURE 18: “[In the months when the family struggled most with money] what happened to cause this difficulty?” (Multiple responses allowed; percentage of households) 100% 90% 80% 70% 30% 2% 60% 50% 44% 12% 41% 23% 40% 7% 2% 30% 5% 57% 18% 11% 4% 20% 41% 51% 1% 30% 15% 10% 28% 23% 13% 26% 20% 1% 17% 5% 4% 6% 10% 9% 0% 1% Waiting to Expenses Already sold Harvests Price of Extensive Price of No casual Health No specific Repayment Expenses Theft harvest crops related to all crops were lower crops rose crop loss crops work or emergency reason of debt related to agriculture than at time to dropped at income life events expected buy food selling time activities Mozambique Tanzania Pakistan FIGURE 19: Mozambique Smallholder Diaries: Net income and household a expenditures all sample level, JULY 2014–JUNE 2015 (US$) $4,500 $4,000 $3,500 $3,000 Hardship months/ $2,500 hunger season $2,000 $1,500 $1,000 $500 $0 JulAug Sep OctNov DecJan FebMar Apr MayJun Household expenditures Net-income a. The green income line refers to net income. For agricultural production, and small businesses in particular, income refers to revenue less related expenditures. The red expenses line refers to operational expenses of the household separate from income or financial transactions pending on groceries, clothes, education, transportation). (e.g., s credit options and were thus better able to tide themselves over during hard- ship months ˜evertheless, these families were not immune to the ebbs and ﬂows of the agricultural cycle €ncome from agricultural production was nega- tive in Žuly and ˆugust (see igure ‘…) when the sample in aistan spent a si—eable amount on agricultural inputs (see igure ‘‘) Smallholder households in the sample identiﬁed Žune, Žuly, and ˆugust, as well as the winter months between the two ma„or harvests as particularly diﬃcult periods in the year

48 | SMALLHOLDER DIARIES BOX 7 SPENDING AS YOU GO: THE LUHENDES (TANZANIA) Mathias and Zaituni Luhende participated in Smallholder Di- when available cash was low (see Figure B7-1). In July, for aries in Tanzania. They lived with ﬁve of their six children and example, Mathias received an advance of US$23.40 for had a range of income sources. The Luhendes grew pota- casual farming work, which he used on home expenses and toes and worked on neighboring farms; Mathias sometimes kindi, and was given US$1.10 by their son, which they con- earned money from guarding the forest near the village, and tributed to their savings group. At this time, Zaituni cared for both he and Zaituni also sold timber. Their hardest months her sick mother and did some casual work to pay for their were September and April, when income and savings from household expenses, and Mathias borrowed from an unspec- their potato harvest had run out. All their opportunities to iﬁed source to pay for two plots of land. earn income from casual labor were working on other farms, Income and expenses during September were low, as the and this was unavailable in the off-harvest period between family had anticipated. Their 14-year-old son started work- February and May. ing to contribute to the household budget. Mathias contin- Mathias and Zaituni lived in the village oriented to potato ued spending on kindi with money that he borrowed from production and sold potatoes to agents as their main source a friend. of income. They would harvest their potatoes only after they By October, both Mathias and Zaituni mostly spent time found an agent and agreed on a price. Mathias and Zaituni doing casual work in the village. At the beginning of the had no way to properly store harvested potatoes, and the po- month, they managed to save some money at home, but by tatoes would likely rot without proper storage while they tried the end of the month, Mathias had spent some of the sav- to ﬁnd a buyer. They usually consumed the maize they grew, ings on kindi. The couple continued to depend on their son’s but could sell it in an emergency if they needed cash quickly. income to cover household expenses and interest payments The only agricultural inputs that Mathias and Zaituni were on the loans from their informal savings groups. able to afford during the data collection for the Smallholder In January, Mathias and Zaituni harvested some of their Diaries were the potato seeds, which they could buy using potatoes for sale and consumption and sold two sacks. The their earnings from casual labor on other farms. Mathias and household did not earn any income in February. Zaituni har- Zaituni had other sources of income, but they mentioned vested the potatoes on her farm by March and used the rev- throughout the year that they felt it was unaffordable to buy enues for food, kindi, and home expenses. The couple had other inputs. saved some money at home by the end of the month, but Mathias and Zaituni exhibited spend-as-you-go behavior: by the end of April, the family used up all of their savings for their expenses peaked when income was high and dwindled home expenses, food, and kindi. FIGURE B7-1: Major financial transactions, income, and expenses for the Luhendes (TANZANIA) JULY 2014–JUNE 2015 (US$) Mathias received US$23.40 They also received in advance for casual farming US$1.1 from their son. Zaituni earned US$1.70 for work. He spent it on home They contributed all of if casual farming work. Mathias expenses and kindi. to their ROSCA. spent US$0.57 on groceries; $120 Adam spent US$0.86 on Mathias earned US$5.16 for casual kindi in the next 24 hours. farming work. Zaituni spent US$3.43 $100 on groceries; Mathias spent US$0.57 on kindi in the next 24 hours. $80 $60 $40 $20 $0 JulAug Sep Oct Nov Dec Jan FebMar Apr MayJun Income (US$) Expenses (US$)

RELATIONSHIPS BETWEEN INCOME VOLATILITY AND AGRICULTURAL PRODUCTION | 49 FIGURE 20: Tanzania Smallholder Diaries, rice production village: Net income and household expenditures a all sample level, JULY 2014–JUNE 2015 (US$) $3,000 $2,500 $2,000 $1,500 Hardship $1,000 months $500 $0 JulAug SepOct Nov Dec JanFeb Mar Apr May Net income Household expenses a. The green income line refers to net income. For agricultural production, and small businesses in particular, income refers to revenue less related expenditures. The red expenses line refers to operational expenses of the household separate from income or financial transactions (e.g., spending on groceries, clothes, education, transportation). a FIGURE 21: Pakistan Smallholder Diaries: Net income and expenditures all sample level, JULY 2014–MAY 2015 (US$) $50,000 $40,000 $30,000 Hardship Hardship $20,000 months months 10,000 $0 JulAug Sep Oct NovDec JanFeb Mar Apr May –$10,000 –$20,000 Net income Household expenditures a. The green income line refers to net income. For agricultural production and small businesses in particular, income refers to revenue less related expenditures. The red expenses line refers to operational expenses of the household separate from income or financial transactions (e.g., spending on groceries, clothes, education, transportation, etc.). FIGURE 22: Pakistan Smallholder Diaries: Spending on household expenditures and agricultural expenses, JULY 2014–MAY 2015 (US$) $45,000 $40,000 $35,000 $30,000 $25,000 $20,000 $15,000 $10,000 $5,000 $0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Total household expenditures Total agricultural expenditures

50 | SMALLHOLDER DIARIES CASE 3 COPING WITH INCOME VOLATILITY AND ANTICIPATING THE HUNGER SEASON: ALINA (MOZAMBIQUE) Smallholder ‰iaries households were able to reduce the volatility of their income, to a degree, through oﬀ-farm wor, but their struggles still tended to mirror the agricultural cycle €n the –o—ambiue sample, for example, house- holds earned little cash income from agricultural production and confronted most hardships in the months between the harvests, the period called the “hun- ger season” £ut the hunger season was about more than „ust hunger, and was lined to a number of other environmental and nutritional factors that caused hardship €nformation was collected about hunger in the household and forgone medi- cal care Œhe results indicated that households did not experience notable increases in the incidence of household hunger between Žanuary and –arch ‘’…”, but that there was a mared increase in reported health problems and forgone medical attention (see igure ‘Ÿ) €n addition to being nown as the hunger season, Žanuary, ebruary, and –arch are also the rainiest months of the year and present a range of challenges that drive a vicious cycle (see igure ‘“) ƒess variety in the diet results in phys- ical weaness and susceptibility to sicness €n these months respondents reported suﬀering from bad digestion and constipation from eating cassava without much ﬁber or protein Some households still had something to eat in the hunger season, but the variety of their diets was lacing €nclement weather in the hunger season also means that there is less farm wor to do †onstruction pro„ects and other forms of manual labor tend to slow down, and there are fewer opportunities for casual wor €n addition, the roads often ﬂood Œransportation in and out of villages taes much longer, and some- times the buses do not run at all Œhis limits access to marets and health clinics FIGURE 23: Mozambique Smallholder Diaries: Households that reported forgoing medical expenses (Percentage) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% JulAug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Needed a doctor or medicine, but went without Experienced hunger

RELATIONSHIPS BETWEEN INCOME VOLATILITY AND AGRICULTURAL PRODUCTION | 51 and reduces the local availability of medicine, fuel, and other products Standing water also results in an increase of the mosuito populations and resulting vec- tor-borne diseases ƒocal media in –o—ambiue reported that malaria was up by ‘‘ percent in ˜ampula in ebruary ‘’…” due to rains and standing water €n addition, one of the three villages participating in the Smallholder ‰iaries suf- fered a cholera outbrea after the heavy rains and ﬂooding ƒie many in the –o—ambiue sample, ˆlina, ”…, experienced more hardships during the hunger season ƒiving alone since her divorce, ˆlina sold cabanga, locally made beer, lie many single women ˆlina was energetic and managed seven crops woring alone on her small plot (’… hectare) She raised poultry for eggs and meat and hired additional worers only to help her with planting in ˆugust ˆlina also received an old age pension from the government, though pay- ments were inconsistent Œhe pension was meant to provide ©S±¢¢’ every month, but in two months ˆlina received only ©S±”ž’ without explanation ˆlina stored the crops she harvested in September and ¦ctober in a simple bamboo cistern and bags in her house, and then consumed her stored crops over the lean months (igure ‘”) “€ eat well because € live alone, and € don’t drin or smoe,” she said “€ consider that € am eating well when € can eat what is appe- ti—ing, what my heart desires £ecause of this, whatever money € mae, € use to buy foods € lie to eat, lie ﬁsh, meat, rice, cassava ﬂour, and other foods” Œhough these income sources made ˆlina better oﬀ than many other small- holders, she still suﬀered from hunger in the ﬁrst three months of the year Œhe whole community was short of money between Žanuary and –arch, and sales of cabanga dropped during this period (see igure ‘¢) ˆlina survived only on her pension during these months €n September, all seven of her chicens and three of her ducs fell prey to disease and died ™hen ˆlina needed money for oil, soap, and other necessities during the hunger season, she sold cassava and peanuts in Žanuary and mai—e in ebruary to generate cash ˆlina went to sleep without eating for many nights in Žanuary, three nights in ebruary, two nights in –arch, and three nights in ˆpril FIGURE 24: The vicious cycle of the hunger season in Mozambique Run low on Limited stored food; possibility for lack of variety casual work in diet Poor nutrition Rainy season from sickness and breeds disease: hunger further malaria, cholera limits availability to do other labor

Financial Diaries with Smallholder Families - Page 66

52 | SMALLHOLDER DIARIES CASE 3, continued when she did not have money to buy anything to accompany cassava or mai—e ˆlina new what to expect from past hunger seasons, but she did not have access to relevant ﬁnancial tools (eg, savings and credit groups, credit at stores or from agents, storing value on a mobile phone) that could help her anticipate these challenges or cope when circumstances grew diﬃcult FIGURE 25: Alina consuming and selling crops to survive the hunger season, JULY 2014–JUNE 2015 (KG) 160 140 120 100 80 KILOGRAMS60 40 20 0 July Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Harvest Consumption Sale FIGURE 26: Mozambique Smallholder Diaries: Alina’s monthly income and expenses, JULY 2014–JUNE 2015 (US$) $80 $70 $60 $50 $40 $30 $20 $10 $0 JuJulAl ug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Income Expenses

CASE 3, continued 4. RISK MITIGATION AND COPING STRATEGIES Smallholder households were coping with a range of risks, and the ways they attempted to mitigate these risks and how they coped with shocks reﬂect the breadth of their ﬁnancial portfolio and market relationships. Less commercialized smallholders experienced more production-related shocks, while more commer- cialized households faced more market-related ones. Smallholder Diaries house- holds also faced the health and employment shocks common to all families. These not only depleted their ﬁnancial, emotional, and human resources, but also dis- rupted their agricultural activities. Each sample employed a range of strategies to mitigate risk, but most commonly after an agricultural shock there was no spe- ciﬁc response, which could signal a lack of tools with which to cope. Over the 12 months of data collection, Smallholder Diaries households reported no use of insurance of any kind. ˆll proﬁles of smallholder farming families need a myriad ris mitigation solu- Ÿ’ tions, particularly given the covariant riss of agriculture Œo explore the riss smallholders face and their coping mechanisms once shocs occur, the Small- holder ‰iaries included a module of uestions focused on the riss, agricultural and otherwise, faced by smallholder households Œhe module covered attitudes toward ris, actual shocs experienced in the past ﬁve years, perception of riss, and ris management practices SHOCKS €n addition to shocs common to all families—such as sicness, accidents, death of a family member, „ob loss, and business failure—smallholder households are also susceptible to riss related to agriculture, such as production riss from pests and ma„or weather events such as ﬂooding, hail, and drought, and maret riss such as input and harvest price ﬂuctuations Smallholder households may employ a range of ris mitigation strategies, but they cannot always cope when shocs or disasters hit, especially given their limited ﬁnancial portfolios Smallholders faced common health and employment shocƒs that not only depleted household ﬁnancial, emotional, and human resources, but also dis- rupted agricultural actiities (see igure ‘•) €f a household member was sic during harvest, for example, a household may need to hire labor, cutting into their already thin net income and¤or harvest Some respondents in the Œan—ania sample reported woring through illness and in„ury so as not to miss ﬁeldwor, though this aggravated their condition ™hile illness was common in all three | 53

54 | SMALLHOLDER DIARIES FIGURE 27: Households that experienced selected health and employment shocks at least once over past five years, JULY 2015 (Percentage) 100% 91% 90% 85% 80% 70% 70% 66% 62% 60% 58% 50% 42% 40% 35% 30% 27% 25% 25% 24% 23% 21% 20% 17% 20% 14% 15% 14% 14% 15% 11% 8% 10% 5% 0% Household member Household member Household member Someone in Lost a permanent Animals Animals died Was unable temporarily had chronic had an illnes the household job or own were stolen to find work unable to work disease died business failed off farm due to accident Mozambique Tanzania Pakistan FIGURE 28: Households that experienced selected agricultural shocks at least once in the past five years, JULY 2015 (Percentage) 100% 96% 99% 90% 80% 72% 70% 61% 61% 60% 51% 50% 43% 38% 40% 41% 40% 36% 32% 30% 26% 23% 26% 21% 18% 20% 10% 8% 7% 10% 5% 2% 0% Lost land Crops destroyed Crops destroyed Crops destroyed Crops stolen Price of own Price of inputs Could not Could not rent by weather by pests in by pests in crops decreased increased rent enough enough land the field storage significantly significantly land due to due to price increase availability Mozambique Tanzania Pakistan samples, in the –o—ambiue sample more than two-thirds of households (•’ percent) experienced the death of a household member in the past ﬁve years ‰ess commercialied smallholders e‡perienced more production-related shocƒs, while more commercialied households faced more marƒet-related ones (see igure ‘¡) ™eather-related shocs dealt ma„or hardships to families across the sample ¢… percent, Ÿ¢ percent, and •‘ percent of the samples in –o—ambiue, Œan—ania, and aistan, respectively, had a signiﬁcant proportion (‘” percent or more) of a crop destroyed by weather in the past ﬁve years ests also presented a ma„or challenge œven among the sample in aistan, where

RISK MITIGATION AND COPING STRATEGIES | 55 the use of pesticides was universal, Ÿ¡ percent of smallholder households reported signiﬁcant in-ﬁeld crop loss due to pests ˆmong the largely noncommercial –o—ambiue sample, where the use of inputs and sale of outputs was minimal, production riss were paramount ˆbout one-third of these households had experienced signiﬁcant in-ﬁeld crop loss due to pests (Ÿ‘ percent) ˆfter the harvest households needed to store crops for many months to cover their own food consumption, and close to two-thirds of the sam- ple (¢… percent) had lost crops in storage due to contamination from pests or the more commercial smallholders in the aistan sample, maret riss presented a greater concern than production riss §irtually every household had been aﬀected by increases in the price of inputs and decreases in crop sales £ut crop loss in the aistan sample was much lower than in –o—ambiue, largely for two reasons irst, a ma„or proportion of the harvest in aistan was immediately handed to middlemen Second, the aistani respondents used more secure means of storage, including tin and aluminum containers, than the Œan—anian and –o—ambican respondents, who typically stored crops in bags inside their house he aluation and incidence of crops lost due to weather and other disasters among Smallholder Diaries households in oambi ue and ‚aƒistan were remarƒably large ˆttaching an estimated value to an agricultural shoc, such as how much crops that had been destroyed would have sold for, provides an indication of their perceived value across the sample and the relative impact of Ÿ… the loss on the household Œhe median estimated cost of crops destroyed by weather in the –o—ambiue sample was ©S±ž‘, which represents a very large shoc for families with such low overall incomes €n aistan, respondents estimated the cost of in-ﬁeld crop loss due to pests at ©S±”’’ at the median ˆmong the sample in Œan—ania, smallholder households considered the inabil- ity to rent suﬃcient land as one of their costliest and most common negative events “¡ percent reported this as an issue that carried a median cost of ©S±“¢” ‰rops in the maret price of crops was also named by Ÿ¢ percent of the sample in Œan—ania, with a median cost of ©S±‘”¡ RISK MANAGEMENT †ach sample employed a range of strategies to mitigate risƒ, but most com- monly after an agricultural shocƒ there was no speciﬁc response, possibly signaling a lacƒ of tools with which to cope Œhe diﬀerences among the sam- ples revealed varying degrees of access to ﬁnancial tools and safety nets, as well as their degree of maret engagement ™hen their crops were destroyed by weather, for example, many Œan—anian households in the Smallholder ‰iaries sample did nothing, which could reﬂect an apparent lac of perceived fallbac options and¤or a lower impact of weather-related shocs (see igure ‘ž) ™hen the sample in aistan faced the same situation, some smallholder households borrowed money (Ÿ¡ percent), and about one-third also had no speciﬁc coping response (Ÿ“ percent) Œheir range of active responses indicates both access to a more robust ﬁnancial portfolio, the central role of agricultural production in their household livelihood strategy, and, in most cases, a contractual obligation to fulﬁll ™hen households rely on agricultural production for income and to

56 | SMALLHOLDER DIARIES FIGURE 29: Households that used these coping mechanisms when crops were destroyed by weather (Percentage; multiple answers allowed) 80% 72% 70% 60% 50% 40% 37% 38% 34% 35% 30% 28% 20% 10% 9% 10% 9% 6% 6% 4% 7% 5% 3% 3% 3% 3% 2% 0% Nothing specialBorrowed money Used savings Reduced Worked more Sold crops Sold other assets Received help consumption from friends and expenses and family Mozambique Tanzania Pakistan clear si—eable debts with the arthi for costly production inputs, crop failure has a dramatic impact on the family ƒooing more closely at a range of ris management tools and their preva- lence in the Smallholder ‰iaries, it is helpful to distinguish between (i) tradi- tional ris-management strategies that farmers use without recourse to outside services (ie, beyond the extended family) and (ii) ﬁnancial ris management tools that involve an explicit demand for and use of savings, credit, and insur- Ÿ‘ ance services TRADITIONAL RISK MANAGEMENT STRATEGIES ‰eveloped over generations, these mechanisms can reduce ris and smooth consumption, but they have their limitations Œhey may also entail income loss and discourage on-farm investments and the adoption of innovative technolo- gies (Sees, Ha—ell, and –iranda …žžž) †rop diversiﬁcation provides an exam- ple of this trade-oﬀ • ”rop and liestocƒ management and diersiﬁcation ƒooing for counter- cyclical net ﬂows, farmers planting crops with a deﬁned growing¤harvest cycle will also eep cattle, goats, pigs, and chicens as a way to smooth their sources of income and food €n addition, staggered planting dates, especially in irrigated plots, mean similarly staggered harvest dates, thus mitigating riss such as unexpected drought (or ﬂoods or pests, as the case may be) in “ The number of crops I grow mid-season depends on the amount of †rop diversiﬁcation is perhaps the most dominant strategy of agricultural money and capital I have, and ris prevention (Ha—ell, omareda, and §aldés …ž¡¢) †rop diversiﬁcation, also depends on my health which could mean planting several crops instead of a single crop or planting the same crop in plots with diﬀerent rainfall patterns, is intended to lower and the availability of labor.” the possibility that all crops will be aﬀected by speciﬁc pests or price ﬂuctua- —Smallholder respondent tions, but its main shortcoming is that the yields and proﬁts from a diversi- in Tanzania ﬁed crop portfolio are typically lower that those obtained from speciali—ing

RISK MITIGATION AND COPING STRATEGIES | 57 in a single crop or „ust a few ‰iversiﬁcation can also have the added disad- vantage that production volumes from each crop are relatively small, limit- ing their ability to grow the minimal “critical mass” of produce reuired to access some marets –ost smallholder households in the sample grew a variety of crops Œhe Smallholder ‰iaries samples in Œan—ania and aistan grew an average of Ÿ¢ and “Ÿ crops, respectively €n –o—ambiue, households grew an average of ¢Ÿ crops most had relatively small yields and were mainly intended for con- sumption at home €n each sample, smallholders calibrated the mix of crops that would give an optimal return, but there was only so much they can suee—e out of the land until they started to see diminishing returns –any families wanted to, and in some cases attempted to, grow additional crops but did not have enough land or resources, or would not be able to grow the new crop without a ﬁnancial loss (See †ase “ on the ˜amuacas in –o—am- biue detailing the importance they placed on crop diversiﬁcation and their unfulﬁlled interest in expanding and diversifying yet further without a sup- portive ﬁnancial mechanism) €ntercropping—growing two crops on the same plot (eg, mai—e and beans) to limit transmission of pests or disease, reduce erosion, shade new growth, and control weeds—was also widely practiced throughout the sam- ple Œhe proportion of farmers practicing intercropping was the largest in –o—ambiue (¡• percent), in part to ensure as much diversity in consump- tion as possible (see igure Ÿ’) ¦n the more general point about access to agricultural training and infor- “[Intercropping] is important, mation about agricultural techniues, responses diﬀered greatly across the because when it is harvest time samples ˆpproximately •” percent of the sample in aistan reported that you can harvest at one time they have access to agricultural information, compared to only …¢ percent in all crops and it is easy –o—ambiue and …Ÿ percent in Œan—ania (see igure Ÿ’) to clean the plot for two €n terms of strategies focused on improving agricultural production, almost or more crops.” every smallholder household in the aistan sample used insecticide and fer- tili—er, while in –o—ambiue the use of either was very limited (…‘ percent —Respondent in the Mozambique and ” percent, respectively) (see igure Ÿ’) ˆmong the Œan—anian families, Smallholder Diaries FIGURE 30: Risk mitigation/production maximizing strategies in use among Smallholder Diaries samples (PERCENTAGE) 100% 100% 99% 94% 90% 87% 84% 80% 78% 79% 75% 70% 60% 61% 60% 50% 46% 40% 33% 30% 23% 20% 16% 12% 13% 10% 5% 0% 2% Use fertilizer Use insecticide Use machinery Intercropping Have access Have access to to replace labor to info about newest agricultural prices training and information Mozambique Tanzania Pakistan

58 | SMALLHOLDER DIARIES use of fertili—er (•¡ percent) was as prevalent as insecticide (‘Ÿ percent) Sim- ilarly, labor-saving mechani—ation was prevalent in the aistan sample (ž“ percent), but nearly absent in the –o—ambiue one (‘ percent) • –ﬀsetting price and yield ariations with surpluses †ontingent upon their ability to store produce and the perishability of the product, as well as their overall cash-ﬂow ﬂuctuations, smallholder farmers may be able to manage price and yield variations in an attempt to reduce income variability Storage capacity, a ma„or enabling factor for this strategy, is usually a constraint ©sing registered warehouses (which may oﬀer access to credit) involves transport ŸŸ costs, warehouse fees, and the need to meet uality standards Œhe Smallholders ‰iaries sample in aistan and Œan—ania reported ma„or concerns about potential decreases in maret prices for their crops ˆccess to information about prices could lower this ris, and a large proportion of the sample in aistan (¡“ percent) but less than half in Œan—ania (“¢ percent) reported access to price information (see igure Ÿ’) €nterestingly, both sam- ples of farmers in aistan and Œan—ania reported that their main sources of information about crop prices were agricultural agents, other people in the village, and people they tal with in the city or town Œan—anian respondents were particularly concerned that buyers could tae advantage of them, while respondents in the aistan sample were focused on maing sure they would earn a proﬁt, given the high expenses related to their agricultural activities ˆbout two-thirds of the smallholders in the –o—ambiue sample (¢” per- cent) reported that they learned about crop prices at the maret when they went to buy and sell ˆ few of these households mentioned other sources as well, such as friends and family ‡iven the distance, the poor roads, and the cost of transportation, this eﬀectively meant that they often have to tae whatever price is available at the maret that day, since they could not aﬀord to return home with their crops and then come bac another time • Šncome source diersiﬁcation and nonfarm employment ˆs noted in Sec- tion …, oﬀ-farm agricultural and nonagricultural employment were impor- tant sources of revenue for smallholder households outside their own agricultural activities Some sources ran “increased labor maret participa- “ My farming is small, thus my tion” as the most important ris-ad„ustment strategy after crop diversiﬁca- tion (™aler and Žodha …ž¡¢) –oreover, wage labor in particular conveys a income is relatively low. Crop steady source of income that few agricultural activities oﬀer, with the possi- prices help us not fall into ble exception of dairy production and bacyard poultry farming (ie, egg pro- total darkness of loss.” duction) ˆn important caveat, however, is that the eﬀectiveness of accessing oﬀ-farm employment opportunities to oﬀset ﬂuctuations in agricultural pro- —Respondent in the Tanzania duction income depends to a large extent on the covariance between agricul- Smallholder Diaries tural and nonfarm revenues ˆgricultural shocs that aﬀect an entire region will mean that only access to employment in a diﬀerent region (or country) would be an eﬀective counterbalance • —ecourse to family and friends Œhe use of cash and in-ind contributions from family and friends (ie, “resources received”) to cope with emergencies and meet lumpy expenses (ie, they occurred in distinct periods over the year, not continuously) is documented in the Smallholder ‰iaries (see igure ‘ž) ™hile not a formal ris-prevention or ris-mitigation mechanism, it is con- sidered a general ris-coping tool and plays an important role in smallholder families and low-income households more generally

RISK MITIGATION AND COPING STRATEGIES | 59 • Sale of assets, mainly liestocƒ ˆfter nonfarm employment, using livestoc as uasi-liuid assets that can be converted into liuid assets to compensate for crop losses or cope with unexpected shocs such as medical emergencies is an important ris management tool Stocpiling basic assets, therefore, is the corresponding ris-prevention method, and temporary surpluses are usually “invested” in assets that can be easily liuidated, such as small live- stoc (eg, goats, pigs, chicens) Œhe limitations of this strategy stem from the mortality¤loss rates associated with livestoc, the price disadvantage associated with emergency sales, and asset indivisibility (ie, the need to sell the entire pig even if handling the emergency would call for only half of the pig’s value) ¦ther uasi-liuid assets suitable for stocpiling are building materials (eg, brics, gravel), ﬁrewood, and manure ¦ver the period of data collection in the Smallholder ‰iaries, sales of any ind of livestoc to gener- ate cash or cope with shocs were reported by ”” percent of sample house- holds in Œan—ania, and •‘ percent of sample households in aistan, but not at all among the sample in –o—ambiue FINANCIAL RISK MANAGEMENT TOOLS ©sing ﬁnancial tools such as savings, credit, and insurance products in con„unc- tion with traditional methods of ris management is arguably a more compre- hensive and preferred overall strategy for those farmers with access to those services Œhe literature on agricultural insurance begins with an assessment of farmer demand for insurance ˆ critical uestion in this assessment is whether ris-management methods used by farmers adeuately protect household con- sumption stability and maintain farm productive capacity €f the answer is yes, then there is limited scope for public policies such as crop insurance to help Ÿ“ farmers ad„ust to ris (™aler and Žodha …ž¡¢) • Saings and insurance oor households hesitate to commit their limited cash-ﬂow surpluses to insurance premium payments for the coverage of rel- atively low-impact or unliely riss (ie, with low expected losses) Œhey pre- fer to eep those funds in liuid or uasi-liuid assets that have multiple uses Œhe low uptae of nonlife insurance in low-income countries is usually Ÿ” attributed to this preference Œhe Smallholder ‰iaries reported no use of insurance, of any ind, across the samples in –o—ambiue, Œan—ania, and aistan over the entire period of data collection • ”rop and liestocƒ insurance ˆ separate topic with its own abundant litera- ture, crop insurance, including index-based insurance, has a mixed record in low-income countries Œhis is in no small measure due to the tendency of governments to intervene in the presence of systemic shocs, such as droughts, ﬂoods, or pests, and totally or partially relieve farmers of all obliga- tions (interest and principal) with regard to their creditors Œhe incentives for farmers to purchase insurance are therefore minimal, and more successful eﬀorts to extend insurance rely on bundling the insurance with the purchase of seeds or fertili—er Œhe case for ﬁnancial institutions to purchase index- based insurance has been convincingly made by –iranda (‘’’ž), since their nonperforming loans would drastically and almost immediately increase in Ÿ¢ the event of systemic weather-based shocs

60 | SMALLHOLDER DIARIES Œo understand their overall priorities in responding to ris, participants in the Smallholder ‰iaries were ased to ran their three most important agricultural ris mitigation strategies ˆmong the –o—ambiue sample, buying or leasing machinery (“… percent) and using fertili—er (Ÿž percent) were named the two most important strategies €nterestingly, though they were deemed important, virtually no one in the sample used either approach, which could suggest that they were already convinced about their eﬃcacy but laced the ﬁnancial tools to deploy these strategies CASE 4 DIVERSIFYING CROPS, DIET, AND INCOME, BUT CONSTRAINED FROM DOING MORE: THE NAMUACAS (MOZAMBIQUE) ˆs part of the Smallholder ‰iaries sample in –o—am- œven among this extensive variety of crops, none of biue, the ˜amuacas live with their three daughters the …‘ could be harvested during the hunger season in and two sons on a very small plot of land (’…’ hectare) ‰ecember and Žanuary ™hen food stores ran low, the ˆlberto, ”’, and Œeresa, “’, both started studying but ˜amuacas were tempted to harvest and eat crops, par- did not complete primary school Œhey considered ticularly cassava, from the ﬁeld before they were ripe educating their children a priority and ensured that ¦ther families in the Smallholder ‰iaries sample in their school-age children remained in school –o—ambiue considered this, too, but eating unripe ™oring with three small parcels of land, ˆlberto cassava can be dangerous €t contains a substance that, and Œeresa grew …‘ crops peanuts, sweet potatoes, rice, when consumed, can trigger the production of cyanide cassava, mai—e, beans, sugar cane, bananas, papaya, and be poisonous €n addition, families also made a orange, tomatoes, and mangoes Œhe ˜amuacas con- trade-oﬀ between planting bitter and sweet varieties sumed most of their agricultural production (see igure of cassava Œhe bitter variety contains more cyanide Ÿ…), and sold only four of their crops bananas, mangoes, when unripe and should not be harvested early, yet it is tomatoes, and rice Œhey did consume some of these more resistant to pests sweet cassava is less dangerous four crops, but they were grown primarily for sale Œhe and allows food to stretch across more months of the most important harvests for the ˜amuacas were cassava year, but it is more susceptible to pests Œhe ˜amuacas in ˆugust, peanuts in –ay, rice in Žune, and bananas in valued the resistance to pests and chose to plant the ˆugust¤September and ebruary (see igure Ÿ‘) bitter cassava, which taes more time to mature ƒie other smallholder families in the –o—ambiue sample, the ˜amuacas felt FIGURE 31: In-kind consumption: The Namuacas (MOZAMBIQUE) (KG) that it was important to eat diﬀerent things Cassava 98 over the year Œhey euated good nutrition Peanuts 67 with consuming a diversity of foods, which Maize 66 motivated their extensive crop diversiﬁca- Sweet Potato 47 Rice 40 tion “™e eat well,” ˆlberto explained, Beans 38 “ˆnd € than ‡od that € harvest enough Orange 20 food to feed my children € thin my family Banana 20 has good nutrition because we manage to Papaya 15 eat ﬁsh, cassava ﬂour, sorghum, rice, and € Tomato 13 Sugar Cane 10 also have fruit trees Œhese are foods that Mango 0 neighboring households can’t always get” 020406080 100

RISK MITIGATION AND COPING STRATEGIES | 61 €n the Œan—ania sample, the use of fertili—er was named as an essential pro- duction input by over two-thirds of sample households (¢¡ percent) Having sav- ings (“‘ percent) and strong networs of friends and family (“… percent) were also seen as important –any farmers in the village with access to irrigation also named the use of irrigation or having access to a water reservoir Œhe aistan sample considered three approaches essential having assets that can be sold in times of need (¡… percent), using fertili—er to drive higher yields (¢¢ percent), and maintaining close ties with friends and family who can help (”¢ percent) ¦n a few occasions, the ˜amuacas sipped meals by growing other vegetables Œhey needed the seeds during the Smallholder ‰iaries because they did not for these additional crops, but did not want to spend have caril, a vegetable or meat sauce that accompanies money on buying them Since they did not have access rice, and xima, creamed cassava †aril plays an impor- to a ﬁnancial tool that could support the acuisition of tant role for ﬂavor and nutrition as well as digestion, as these additional seeds (eg, layaway or savings plan, cassava alone is hard to digest amilies often prefer to input credit from a retailer or buyer), their plans to go hungry rather than to eat cassava without caril further diversify and expand their agricultural activi- Œhe ˜amuacas wanted to ﬁll this lull in what they ties were on hold could harvest and consume in ‰ecember and Žanuary FIGURE 32: Crops harvested by the Namuaca family by month (KG) 350 300 250 200 150 100 50 0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Rice Cassava Sugar cane Orange Peanuts Maize Banana Tomato Sweet potato Beans Papaya Mango

Mozambique. Photo by Erin Scronce.

5. HOUSEHOLD FINANCIAL PORTFOLIOS Smallholder Diaries households can be grouped into three loose proﬁles by the breadth of their ﬁnancial portfolios: narrow, moderate, and broad. Most Small- holder Diaries households had access to only a thin scattering of informal ﬁnancial tools. The degree to which sample households could sustain their consumption levels and cope with shocks during lean periods between harvests depended heavily on the range of tools in their ﬁnancial portfolio. Each ﬁnancial mechanism had its own limitations, and across the sample all ﬁnancial portfolios were insuf- ﬁcient to meet the varied needs of smallholder households. SMALLHOLDER DIARIES SAMPLE IN MOZAMBIQUE— NARROW FINANCIAL PORTFOLIOS Smallholder households in the oambi ue sample used only three ﬁnan- cial instruments at the median ˆnd overall they were woring with only two Ÿ• types of ﬁnancial instruments at the median Œheir very narrow ﬁnancial port- folio was mostly limited to savings at home (see Œable “) –nly a fraction of the oambi ue sample was engaged in informal saings and credit groups, and all participants in „S”„s were women ˆmong Small- TABLE 4: Mozambique Smallholder Diaries: Use levels and other data for common ﬁnancial devices (N=93 households) HOUSEHOLD LEVEL MEAN STATISTICS AT DEVICE LEVEL TOP 5 SAVINGS PERCENTAGE OF # OF TRANSACTIONS AVERAGE AVERAGE INSTRUMENTS HOUSEHOLDS LATEST BALANCE PER INSTRUMENT DEPOSIT WITHDRAWAL Savings in the house 87 $17.44 9 $10.90 $12.88 Lending to friends and family 41 $10.49 2 $33.77 $18.23 Credit given 18 $1.92 3 $3.17 $4.64 ROSCA 12 $15.89 7 $10.37 $51.91 ASCA 9 $17.38 6 $6.18 $41.05 TOP 5 CREDIT PERCENTAGE OF # OF TRANSACTIONS AVERAGE AVERAGE INSTRUMENTS HOUSEHOLDS LATEST BALANCE PER INSTRUMENT DEPOSIT WITHDRAWAL Borrowing from friends and family 59 $3.41 2 $10.80 $8.62 Credit at a store 22 $2.57 2 $3.89 $3.28 Borrowing from informal group 5 $6.92 2 $27.69 $6.49 Pawn 5 $6.58 1 $1.83 $1.73 Act as money guard 5 $1.33 1 $17.60 $0.58 | 63

64 | SMALLHOLDER DIARIES holder ‰iaries households in –o—ambiue, only …‘ percent used a ª¦S†ˆ, ž percent an ˆS†ˆ, and ” percent a money guard to save Smallholder families in the oambi ue sample often relied on casual labor as a ƒind of automatic teller machine („) to generate cash during lean times ™ith limited savings and credit options, sample households looed to wages from casual labor to get through the hunger season †asual labor ﬁlled a gap, but was not a perfect solution Œhe timing of this income did not always match the timing of the needs and it was often insuﬃcient to carry families through this diﬃcult period (see £ox ¡) „ fraction of households in the oambi ue sample had heard of mobile money products (ŽŒ percent), and use was none‡istent ©se of mobile money was at least in part impaired by low mobile phone ownership and capability Œhose who did not use it but who had heard of it considered it a useful product BOX 8 CASUAL LABOR AS AN ATM AMONG THE WAPERIWAS (MOZAMBIQUE) Among the sample in Mozambique, the Waperiwa family construction projects in January, at the height of the hunger used casual labor as a kind of ATM. A family of six, they relied season. heavily on casual labor to supplement their subsistence farm- By relying on income from casual labor, the Waperiwas ing. Issa and his wife Fabiana lived with their four teenage did not go hungry during the past lean season, but they children, and also often cared for Fabiana’s mother. The Small- did feel the burden of the “sick season.” Family members holder Diaries found only two ﬁnancial mechanisms at work needed medical care in September, December, and March, in this household: Issa saved at home and borrowed from his but went without. In all cases, instead of going to the doctor, brother, just once at the very beginning of the research. they bought medicine in the village. The family produced cassava, peanut, sweet potato, The one ﬁnancial instrument the Waperiwa household beans, and maize. The harvests in July to September to used, saving at home, was not sufﬁcient to carry the family some extent carried the family through the hunger season. through the hunger season. Casual labor ﬁlled a gap, but it The household sold only small quantities of sweet potato in was an imperfect, insufﬁcient solution, and they were lucky to July 2014 and cassava in June 2015 because they preferred ﬁnd as much casual work as they did. Some households re- to store as much food as possible for their own consumption. ported struggling to ﬁnd employment during the lean months. Although Issa had amassed a signiﬁ- cant amount of savings by October in FIGURE B8-1: Waperiwa household’s monthly savings in the house a secret saving place at home, this sav- balance and casual labor earnings (US$) ings was not enough to carry the fam- ily through the months of the hunger $100 season (see Figure B8-1). To continue to $90 have enough to eat during the hunger $80 season, the Waperiwa family spent all $70 of their savings by January, and did not contribute to their savings at all during $60 this time. Although the family may have $50 been able to borrow from their social $40 network in times of need during these $30 months, everyone in the village tend- $20 ed to have limited ﬁnancial resources. Households in the Mozambique sample $10 had virtually no access to credit during $0 the lean months and instead increased July Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun their work on other farms and small Savings in house Casual labor

HOUSEHOLD FINANCIAL PORTFOLIOS | 65 (hypothetically) ¦ne farmer in the –o—ambiue sample explained that though he was not using it, he lied the idea of the service because “it is a faster way to send and receive money €t is a pity that we don’t have an agent near €f we had € would have tried it € heard that you don’t need a ban account” ‘orƒing with such a narrow portfolio of ﬁnancial tools, the oambi ue sample generally e‡hibited spend-as-you-go behaior with uneen and ola- tile spending on e‡penditures Œhe average family in the –o—ambiue sample actually drew on their savings least during the hunger season from Žanuary to –arch because it had run out by this point Œo cope, they reduced their house- hold consumption, and expenditure levels dropped to the lowest in this period SMALLHOLDER DIARIES SAMPLE IN TANZANIA— MODERATE FINANCIAL PORTFOLIOS Smallholder households in the anania sample used ŒŽ diﬀerent ﬁnancial tools, and oerall si‡ types of ﬁnancial instruments, at the median (see Œable ”) Œhey relied most heavily on current income and short-term savings for both their agricultural and nonagricultural expenses ˆwareness of mobile money (ž¡ percent) was almost universal, but only …ž percent reported using it during the Smallholder ‰iaries to receive or send money (either from their own mobile Ÿ¡ money account or someone else’s) –nly the illage that focused on potato production had access to —–S”„s and „S”„s, but een then loans from informal saings groups were gener- ally not used for agricultural e‡penses Œhose who borrowed had a number of small loans with informal groups and in their social networ ™ith few oppor- tunities to borrow, most relied heavily on short-term savings (including stored crops) and earnings from casual labor to mae investments in fertili—er and other value-adding inputs †learly, savings in the house is too liuid an instru- TABLE 5: Tanzania Smallholder Diaries: Use levels and other key data for common ﬁnancial devices HOUSEHOLD LEVEL MEAN STATISTICS AT DEVICE LEVEL TOP 5 SAVINGS PERCENTAGE OF # OF TRANSACTIONS AVERAGE AVERAGE INSTRUMENTS HOUSEHOLDS LATEST BALANCE PER INSTRUMENT DEPOSIT WITHDRAWAL Savings in the house 100 $18.25 101 $10.86 $3.10 Lending to friends and family 48 $10.52 2 $15.91 $15.84 ASCA 53 $5.83 3 $1.01 $34.47 ROSCA 33 $2.14 28 $1.09 $32.23 Credit given 28 $5.20 6 $2.81 $3.82 TOP 5 CREDIT PERCENTAGE OF # OF TRANSACTIONS AVERAGE AVERAGE INSTRUMENTS HOUSEHOLDS LATEST BALANCE PER INSTRUMENT DEPOSIT WITHDRAWAL Borrowing from friends and family 77 $4.53 2 $11.20 $11.56 Borrowing from informal group 67 $10.74 6 $7.83 $15.74 Credit at a store 60 $2.00 4 $7.31 $4.47 Agent credit 14 $12.44 1 $71.87 $41.35 Act as money guard 13 $0.09 3 $18.58 $18.67

66 | SMALLHOLDER DIARIES ment to be used for such large investments, and without leverage, smallholders had limited options to invest in the productivity of their agricultural activities –any of the Œan—anian respondents planned strategically to invest in their farms Some families borrowed from middlemen and repaid in-ind with their own crops but for many, however, the amounts oﬀered by agents were very low compared with their overall farming expenses ¦ther households saved revenue from sales of their agricultural production for months to purchase inputs and pay worers when needed (see †ase ” on the experience of the £itungwas in Œan—ania, who used this approach) „ majority of smallholder households in the anania sample (˜™ percent) reported that ƒeeping cash in the house was their most important way to sae Smallholder families also used stored crops as a ƒind of “term deposit,” waiting for their production to gain “interest” with price increases oer timeš ‘… percent considered crop storage their most important form of savingsŸž €n many cases, families delayed the sale of rice, and when necessary, mai—e, until funds were needed or there was an emergency †ash that was obtained from selling the crops was then stored at home he anania sample also used casual labor liƒe an „ to generate small amounts of cash for e‡penses such as agricultural inputs amilies wored at odd „obs until they earned enough cash to buy what they needed ¦ne small- holder household in the rice-producing village said they generally do not go to bed hungry when they were low on food, they could “„ust wor in someone’s ﬁelds” and earn money to buy food €n the same vein, respondent households in the potato-producing village sometimes engaged in casual labor when a large expense came up that their other income sources (particularly agricultural pro- duction) could not cover ‰avid, the head of one household, for example, engaged in construction wor to repay debts to his informal savings group He had expected to earn ©S±¡’ from the „ob but, as this was casual wor, he was not woring under a formal contract and in the end received only ©S±“¢ from his employer he anania sample also e‡hibited a spend-as-you-go approach and suf- fered most during nonagricultural production periods ˆmong the Small- holder ‰iaries sample in villages focused on rice cultivation in Œan—ania, families identiﬁed ebruary, –arch, and ˆpril as the hardest months of the year, when money, crops, and opportunities for casual labor related to the previous harvest dwindled amilies were unable to eﬃciently draw on their ﬁnancial tools when they needed them most (see igure Ÿ“) saving withdrawal activity (yellow line) dipped signiﬁcantly and borrowing was almost nonexistent over these three months, mirroring corresponding dips in income SMALLHOLDER DIARIES SAMPLE IN PAKISTAN— BROAD FINANCIAL PORTFOLIOS ‘orƒing with the broadest, most robust ﬁnancial portfolio of the sample, Smallholder Diaries households in ‚aƒistan used almost Œ’ diﬀerent ﬁnan- cial tools from si‡ types of ﬁnancial instruments at the median (see Œable ¢) Œhe average smallholder family’s income ﬂuctuated dramatically with the farm-

HOUSEHOLD FINANCIAL PORTFOLIOS | 67 FIGURE 34: Tanzania Smallholder Diaries, rice production village: Income, borrowing, withdrawals from savings devices, and balances of saving devices at the sample level, JULY 2014–JUNE 2015 (US$) $3,500 $3,000 $2,500 $2,000 $1,500 $1,000 $500 $0 JulAug Sep Oct NovDec JanFeb Mar Apr MayJun Income BorrowingSavings withdrawals Savings deposits FIGURE 35: Pakistan Smallholder Diaries: Average income, borrowing, withdrawals from savings devices, balances of saving devices (US$), JULY 2014–MAY 2015 $50,000 $40,000 $30,000 $20,000 $10,000 $0 July AugOSep ct Nov Dec Jan Feb Mar Apr May –$10,000 –$20,000 Income (Mean) Expenses (Mean) Borrowing (Mean) TABLE 6: Pakistan Smallholder Diaries: Use levels and other key data for common ﬁnancial devices HOUSEHOLD LEVEL MEAN STATISTICS AT DEVICE LEVEL TOP 5 SAVINGS PERCENTAGE OF # OF TRANSACTIONS AVERAGE AVERAGE INSTRUMENTS HOUSEHOLDS LATEST BALANCE PER INSTRUMENT DEPOSIT WITHDRAWAL Savings in the house 85 $72.74 9 $146.65 $82.78 Lending to friends and family 61 $45.32 2 $135.36 $108.62 Money guard 56 $122.68 4 $583.83 $280.89 Credit given 23 $175.95 20 $21.15 $51.92 ASCA 19 $37.55 1 $80.60 $130.14 TOP 5 CREDIT PERCENTAGE OF # OF TRANSACTIONS AVERAGE AVERAGE INSTRUMENTS HOUSEHOLDS LATEST BALANCE PER INSTRUMENT DEPOSIT WITHDRAWAL Borrowing from friends and family 99 $74.08 2 $97.14 $80.67 Agent credit 97 $106.98 9 $199.19 $49.43 Credit at a store 94 $22.87 14 $36.98 $5.99 Joint liability loan 27 $122.16 2 $259.05 $252.80 Business loan 16 $1,244.98 2 $262.38 $939.74

68 | SMALLHOLDER DIARIES ing cycle, but the aistan sample was able to leverage credit (see orange line, igure Ÿ“) so that expenses could at times exceed income (eg, Žanuary) ˆlmost all of the smallholders in the aistan sample had heard of mobile money (¡‘ percent), but none used it during the study ›ouseholds used arious forms of credit to get through the months when spending on agricultural inputs was high and reenue from farming was low, borrowing from family and friends (žž percent), arthis (ž• percent), and local stores (ž“ percent) to get through the summer and winter months ¦ver the year of data collection, the typical household in the aistan sample obtained goods on credit from an individual store on …“ occasions Œhe sample also relied heavily on credit to buy expensive farming inputs, which, given the scale of agri- culture in the region, was probably the only option ut een this relatiely broader, more dierse ﬁnancial portfolio did not meet all needs of the Smallholder Diaries families in ‚aƒistan Spending on agricul- tural inputs was very high, and at times basic household needs were sacriﬁced to aﬀord investments in their agricultural activities ˆlmost one-ﬁfth (…ž per- cent) of households in the Smallholder ‰iaries sample in aistan reported that buying food was their most challenging expense „rthis were embedded in the agricultural and ﬁnancial lies of the sample in ‚aƒistan ˆrthis oﬀered inputs (eg, fertili—er, pesticides) on credit to be repaid after the harvest and then bought crops from farmers at harvest (see £ox ž) ˆrthis also acted as money guards for smallholder households, holding onto their cash savings, which was often the income from their agricultural produc- tion Smallholder households in the aistan sample appreciated that their ﬁnancing arrangements with arthis had some ﬂexibility, in that they could defer repayment in the event of crop failure ™oring with an arthi also reduced some riss, such as ma„or income shocs from crop failure, not ﬁnding buyers when needed, and the probability that the price might go down “After we cut the cr op, we ut smallholders generally found the interest rates high and were re uired sell right away. We have to to repay immediately after harest, which meant smallholders could not return the credit to the arthi. wait to sell their agricultural production until prices were more faorable Because of this helplessness, ˆrthis generally charged smallholders about ‘… percent interest on the amount I have to sell. And sometimes borrowed, citing the ﬂexibility they oﬀered and the ris they were taing on, and expected payment after six months, right after the harvest Some arthis the selling price is low, and were so eager to obtain repayment that some families reported feeling “both- we suffer a loss [ﬁnancially].” ered” by the arthi to repay Some families felt forced to sell their crops at an —Smallholder participant inopportune time, when supply was high and prices low, and saw this as exert- in Pakistan ing a large, negative impact on their income (see £ox …’) ©pon repaying their debt after harvest, many families reported immediately borrowing again to ﬁnance inputs for the next season, creating a cycle of debt Œhere was also some concern about the ledgers that arthis ept on the saved amounts, and the uality and transparency of their record eeping

HOUSEHOLD FINANCIAL PORTFOLIOS | 69 BOX 9 THE FINANCIAL LIFE OF ALI, AN ARTHI (PAKISTAN) Ali has been an arthi for ﬁve years and serves one of the vil- ton), he paid the farmers, usually within a week of the sale. lages with participants in the Smallholder Diaries in Pakistan. His borrowing needs were covered by banks. After his father, who had been a businessman in the oil in- Ali also bought pesticide and fertilizer from various dustry, passed away, Ali used his inheritance to start working shops, stored them at certain stores, and sold them to as an arthi, following the advice of a friend. Ali had about farmers at a 25–30 percent proﬁt. He held about 5–7 mil- 80 ongoing clients in 12 villages, most of whom farmed less lion rupees (about US$49,000–69,000) in savings for farm- than 4.8 hectares of land. ers. He kept a record of transactions in a register, and only Working in competition with other informal arthis, Ali some farmers kept track themselves and cross-checked spent millions of rupees each year buying wheat, rice, with them. Most farmers borrowed for agricultural needs, cotton, and mustard seed. He did not store the crops he and sometimes for expenses such as marriages and funer- bought, but sold right way to pukka (formal) arthis and mill als. He lent only to regular clients and rarely had a problem brokers. The price he obtained for the crops was based on with repayment. When crops failed, farmers could repay bidding, and inﬂuenced by the quality of the crop and the next season. He said he has generally good relations with international market. Prices changed daily, and after deduct- farmers, but that some did get upset if he denied them ing a fee (1.5 percent for staple crops such as wheat and credit. Overall, Ali considered being an arthi a good busi- maize, and 2.5 percent for cash crops such as rice and cot- ness with a “handsome pay off.” BOX 10 PERCEPTIONS OF A MIDDLEMAN: FARID AND SEEMA (PAKISTAN) The experiences of Farid and Seema shed light on how the arthi when they needed them. This had a negative ef- arthis, as well as other service providers, are typically viewed fect on the output of their wheat crop. Similarly, they needed by the respondents in the Smallholder Diaries in Pakistan. money for medicine during the same month, but the arthi Farid, 43, and Seema, 40, live with their four young children did not lend it to them on time. and grow mainly rice and wheat on their 0.8 hectares of land. Farid did not like working with the arthi very much, but The family had trouble meeting all of their expenses and of- felt majboor [helpless or obligated] since he had no other ten borrowed to cover daily expenses, which they cited as way to obtain fertilizer. He thought banks charged too much their most difﬁcult expense to cover, but their ﬁnancial port- interest and that microﬁnance institutions would charge ex- folio consisted only of informal ﬁnancial mechanisms. tra interest that would be difﬁcult to pay back if he could Farid and Seema dealt with one arthi whom they bor- not repay on time. Farid did appreciate that he could repay rowed from and stored money with as well. The family oc- the arthi after the harvest. “With the middleman, we don’t casionally borrowed for nonagricultural expenses from the have a ﬁxed time to repay our debt. Whenever the crop is arthi, but it had difﬁculty obtaining funds from the arthi in a ready, we sell and repay,” he explained. He also preferred to timely manner. In December 2014, for example, Farid and save with the middleman because he felt that money kept at Seema needed diesel to operate the tube well that would home was easily spent. irrigate the wheat, but they were unable to obtain funds from

70 | SMALLHOLDER DIARIES CASE 5 WORKING WITH A MODERATE FINANCIAL PORTFOLIO AND STRUGGLING TO MAKE AGRICULTURAL INVESTMENTS: THE BITUNGWAS (TANZANIA) ™illiam and ˆnitha £itungwa, participants in the ™illiam also borrowed from family and friends and Smallholder ‰iaries sample in Œan—ania, were married saved with an ˆS†ˆ Œhough he had borrowed from and lived with three children, one adopted and two of ﬁve informal savings groups, the £itungwas did not ˆnitha’s from before their marriage Œhey owned more lie them and felt that the groups did not beneﬁt them than two hectares of land and relied primarily on their Œhey both believed that when they had a loan they income from growing potatoes, wheat, mai—e, beans, must wor harder, but only because they wanted to and peas and selling mil Œhey planted potatoes twice repay the loans to the groups, not for some greater a year and, because of a suitable climate, considered beneﬁt to their household €t was social pressure that them “easy to care for” and good for generating “uic led ™illiam to „oin the groups nonetheless ƒocal inter- income” Œhey grew wheat once a year and harvested est groups wanted everyone in the village to be part of in ˆpril for extra income Œhe £itungwas earned addi- these savings groups, and ™illiam was forced to „oin tional income from selling rat poison and received one operating under the political party of which he is a resources from family and friends outside the house- member hold ™illiam also wored as a researcher with a local ˆgricultural production was their main source of university income, and ™illiam and ˆnitha “tae it seriously,” ™illiam and ˆnitha used a number of ﬁnancial which was evident from the amount they spent on mechanisms Œhey both ept money aside at home as a inputs for potatoes and wheat, what they considered form of savings and too informal credit from shops their cash crops (see igure Ÿ¢) Œhe £itungwas used FIGURE 36: Agricultural expenses (labor, seeds, fertilizer) compared to restaurant, food, and grocery expenses: The Bitungwas (TANZANIA) JULY 2014–JUNE 2015 (US$) $160 $140 $120 $100 US$$80 $60 $40 $20 $0 Jul Aug Sep Oct Nov Dec Jan Feb MarApr May Jun Agricultural expenses Food from restaurants, take aways, delivery, street food Groceries/food to be eaten at home (including baby food)

HOUSEHOLD FINANCIAL PORTFOLIOS | 71 their savings in an attempt to meet their farming fertili—er and pesticide and pay for laborers in their expenses and saved revenue from potato sales for up to ﬁelds when needed ™hile they were actively with- three months before withdrawing the money to pur- drawing from and depositing into their savings at chase inputs, but they were unable to meet all of their home between Žanuary and ˆpril, the ©S±…•… that they farming investment needs with the ﬁnancial tools at earned from potato sales in Žanuary clearly corre- hand –ost of the time they bought inputs using money sponds to the ©S±…•‘ they later spent on labor and earned from selling their cash crops, or they used agricultural inputs in ebruary, –arch, and ˆpril when money that they had received from their children living planting potatoes and wheat €n fact, ™illiam and in town Sometimes they sold assets to cover input ˆnitha earned only ©S±Ÿ”‘’ in ebruary but had expenses as well, since they believed that they would be saved enough from selling potatoes to cover their able to purchase more assets after harvest and after expenses, which was not an uncommon approach to they had sold their cash crops or the £itungwas in the meeting agricultural expenses in their village £ut period of the Smallholder ‰iaries, this was in fact true despite the £itungwa’s careful planning to aﬀord (see igure ŸŸ) inputs, they still insisted that they did not have enough ™illiam and ˆnitha did not always need to buy “€ have enough plots, but € don’t have enough capital to inputs when they sold their cash crops, but they pur- invest in all plots € have,” said ™illiam “€f € could have posely set aside money from selling cash crops to buy capital, € would € have gotten more out of my farm” FIGURE 33: Sales from potatoes between January and April 2015: The Bitungwas (TANZANIA) January 8 January 16 February 13 March 3 March 29 April 4 April 11 Sold potatoes Sold potatoes Withdrew Withdrew Withdrew Withdrew Withdrew Withdrew worth worth savings at savings at savings at savings at savings at savings at US$106 US$65 home worth home worth home worth home worth home worth home worth US$6 US$29 US$11 US$11 US$69 US$29 Deposited Deposited Paid for Paid for Paid for Paid for ag Paid for ag Paid for savings at savings at labor on labor on labor on inputs inputs labor on home worth home worth farm worth farm worth farm worth worth worth farm worth US$106 US$65 US$6 US$29 US$28 US$11 US$60 US$29 US$171 They save their revenues (US$171) from US$172 potato sales in January before withdrawing the money in January to purchase inputs in February to April.

Tanzania. Photo by Erin Scronce.

6. ACCESS TO MOBILE PHONES AND USE OF DIGITAL FINANCIAL TOOLS Use of digital ﬁnancial tools in the Smallholder Diaries sample was very limited and only among households in the Tanzania sample. The Smallholder Diaries also point to a crucial gap between basic access to a phone, which itself remains a bar- rier, and the ability to use it to conduct ﬁnancial transactions. Overall, the degree to which smallholder households might engage with mobile ﬁnancial services var- ies across the proﬁles observed in the Smallholder Diaries, particularly in terms of their starting point. he use of mobile phones to access ﬁnancial serices has been of particular interest in smallholder ﬁnance, because of the potential of digital ﬁnancial serices (DFS) to oercome some of the ƒey constraints to proiding rural “’ ﬁnance at scale inancial inclusion in rural areas has been constrained by its inherent lower population density, as well as relatively lower incomes, and therefore higher transaction costs, outside urban centers Œhe possibility of providing a wide array of ﬁnancial services—including easy-access bridge loans, savings mechanisms that target various goals, and insurance for the most devas- tating riss, as well as nonﬁnancial services, such as weather and price informa- tion and linages to value chains—all seem to come within closer reach when using new applications of technology and mobile phones £ut ﬁrst things ﬁrst access to and capability with mobile phones drive the viability of ‰S ost of the most economically actie members in the anania and ‚aƒi- stan samples owned a mobile phone (˜œ percent and žŸ percent, respec- tiely), but less than half in oambi ue did (¡˜ percent) (see Œable •) –ost other household members did not have a mobile phone €n addition to the main respondent, someone else in the household also had a mobile phone in only Ÿ¡ percent, “Ÿ percent, and ‘… percent of the samples in –o—ambiue, Œan—ania, and aistan, respectively –ore broadly, the great ma„ority of respondents in Œan—ania and aistan had used a phone—including both their own or a bor- rowed one—in the past year (•• percent and •Ÿ percent, respectively), but the proportion in the –o—ambiue sample was much lower ¦nly ”” percent of the sample smallholder households in –o—ambiue had used a phone at all in the past year here is a crucial gap between basic access to a phone, which itself remains a barrier, and the ability to use it to conduct ﬁnancial transactions, which entails basic literacy in the operating language S–S functionality is also important for mobile ﬁnancial services for those unable to send or receive S–S, | 73

Financial Diaries with Smallholder Families - Page 91

74 | SMALLHOLDER DIARIES performing ©SS‰- or text-based ﬁnancial transactions on a mobile phone is liely to pose signiﬁcant challenges ¦f the respondents with access to a mobile phone, ¢¡ percent in Œan—ania could use S–S functionality (but not access the internet), compared with only ‘“ percent in aistan and ‘” percent in –o—am- biue (see Œable •) Œhus ‰S remain important tools to explore and expand, but must be carefully targeted to each customer proﬁle Some types of ‰S (eg, mobile ﬁnancial services such as ‘ transfers) may be more appropriate for mobile-literate smallholders who are more capable with their phone, while other digital solutions will need to be explored to reach those with low levels of mobile literacy (eg, ¦Œ†, agent-facilitated transactions) “se of digital ﬁnancial tools in the Smallholder Diaries was ery limited (and only in the anania sample), despite arying leels of awareness of and aspiration to use this ﬁnancial tool across the three countries ‡eneral awareness of mobile money (deﬁned in this study as a transfer of funds using a mobile wallet) as a ﬁnancial tool ranged greatly across Smallholder ‰iaries respondents, from a low of ‘… percent in the –o—ambiue sample to near com- plete awareness among the sample in Œan—ania (see Œable •), which is expected given the strength of the Œan—anian digital infrastructure Œhis was reﬂected in the perception of mobile money within the sample as a ﬁnancial tool relevant to their needs ™hen ased what ﬁnancial mechanisms they might use to send or receive money, “mobile money” was selected as one answer by almost three- uarters of the sample in Œan—ania and more than half of the sample in aistan «et the perceived relevance or aspiration to use mobile money—deﬁned for these purposes as a transfer or transaction using a mobile wallet—did not materiali—e over the year of data collection in the Smallholder ‰iaries £ased on the actual transactions of sample households, only …ž percent of Smallholder ‰iaries fami- a TABLE 7: Mobile phones and mobile money among Smallholder Diaries households (percent), NOVEMBER 2014 MOZAMBIQUE TANZANIA PAKISTAN Access to mobile phones and use of mobile money Had a mobile phone 45 56 70 Had a SIM card 48 56 65 Other household members had a mobile phone 38 43 21 Other household members had a SIM 57 35 22 Had used a phone, even a borrowed one, in the past year 55 77 73 Had heard of mobile money 21 98 82 Selected “Mobile money” as one response when asked “What would you likely use to send or receive money?” (multiple answers allowed) 0 74 57 Had used mobile money (for transfers and transactions on a mobile wallet based on actual cash ﬂows from June 2014 to June 2015) 0 19 0 Self-reported capability with mobile phones “I cannot initiate or receive a call, or send or receive an SMS.” 0 3 1 “I can only receive calls.” 45 9 7 “I can only dial and initiate calls.” 2 0 19 “I can dial and initiate calls and receive calls.” 27 15 37 “I can dial and initiate calls, receive calls, and send and receive SMS.” 25 68 24 “I can dial and initiate a call, receive calls, send and receive SMS, and access the internet.” 0 1 7 a. Responses are from a Smallholder Diaries module administered to the most economically active member of the sample household.

ACCESS TO MOBILE PHONES AND USE OF DIGITAL FINANCIAL TOOLS | 75 lies in Œan—ania used mobile money in the study period ˜o smallholder house- holds in either the –o—ambiue or aistan sample used mobile money at all ”onsidering the potential of DFS to sere smallholder households in areas that traditional bricƒ-and-mortar FS‚s hae failed to reach, DFS remain important tools to e‡plore and e‡pand for purposes of ﬁnancial inclusion, and they must be carefully targeted to each customer proﬁle ‰igital savings and credit products could provide more compelling use cases than payments, as many households in the sample laced access to even informal ﬁnancial ser- vices, such as savings groups ˆdditionally, some smallholders may be well- served by digital payment services that facilitate transactions such as bill payments and school fee payments (ie, person-to-business ¬‘£® and person- to-government ¬‘‡® payments), though smallholders may prefer ¦Œ† pay- ment methods over self-initiated mobile transactions from their own wallet

76 | SMALLHOLDER DIARIES CASE 6 RELYING ON CREDIT IN MONTHS BETWEEN HARVEST: THE BHATTIS (PAKISTAN) Œhe experience of ˆbdul and ªania £hatti, one of the Œhe £hattis grew two main crops, rice in ˜ovem- Smallholder ‰iaries households in aistan, demon- ber and wheat in –ay, as well as a variety of vegetables strates how families in this sample used their relatively and fodder to feed their three buﬀalos and one cow broad ﬁnancial portfolio to access goods and cash when Œhey relied heavily on revenues from selling mil they needed it most ™hile the family had savings both Œhe £hattis incurred signiﬁcant spending on agricul- at home and with its arthi, and also lent money to its tural inputs throughout the year and spent substantial family and friends, it was the liability side of its portfo- amounts on their agricultural production each month lio—informal credit at the store, agent (arthi) credit, relative to what they spent on food €n ˜ovember and occasionally borrowing from friends and family— alone, their spending on agriculture reached ©S±””ž that carried them through tough times Œhe family’s (see igure Ÿ•) views of the arthi were also typical across the sample inancially, ˜ovember was the best month for the ˆbdul, ”•, and ªania, “¡, lived with their son and £hattis Œhey sold their rice and, as a result, were able daughter on a small plot of land (…‘ hectares) Œheir to spend the most £ut the family had the hardest time older son and his wife lived in their house, too, but ept budgeting in Žuly, ‰ecember, and Žanuary, and faced their expenditures separate Œhough ˆbdul and ªania unexpected hardships throughout the year €n Žuly, all were not educated, their ……-year-old daughter was the vegetables that ˆbdul had been growing were enrolled in school Œheir younger son, …Ÿ, preferred to destroyed, costing him about ©S±”’‘ Œhen in Sep- wor rather than go to school and was training in a tember, bad weather completely destroyed the fodder worshop crop as well FIGURE 37: Total spending on food-related items and agricultural inputs: The Bhattis (PAKISTAN), JULY 2014–JUNE 2015 (US$) $700 $600 $500 $400 $300 $200 $100 $0 Jul Aug Sep Oct Nov Dec Jan Feb MarApr May Jun Seeds, fertilizer and agricultural equipment Groceries and cooked food

ACCESS TO MOBILE PHONES AND USE OF DIGITAL FINANCIAL TOOLS | 77 €n the hard winter months, the £hattis spent more igure Ÿ¡) ˆbdul had a line of credit open with „ust than they earned by relying on store credit at seven dif- one arthi, and he lied some aspects of borrowing ferent shops, including a butchery, a vegetable store, from him “™henever € need fertili—er, ¬the arthi® clothing shops, and a irana shop, which sells small gives it to me even if my arthi has a shortage when goods such as soap €n Žanuary, the household experi- € need fertili—er, he’ll get it from someone else and enced a particularly diﬃcult period, and they could bring it to me ˆfter the harvest comes, € pay bac He not even purchase shoes, despite the harsh winter doesn’t demand money from me every day that’s why weather Œheir crops ran out, as did their savings, and he is the best for me” Œhat said, ˆbdul had to repay ˆbdul and ªania’s purchases on credit peaed in Žanu- the arthi with ‘… percent interest on the amount bor- ary €n –ay, after they harvested the wheat, they paid rowed after six months, a rate common across the much of this debt bac ™ithout access to these many Smallholder ‰iaries sample in aistan, which he felt lines of credit, the £hattis would have experienced was too high “ˆfter six months when € need to repay even more hardships during the year him, € have to pay a lot of interest €f the government Œhe £hattis borrowed money to purchase inputs for could give us fertili—er without interest, that would be their rice from the arthi from Žuly to September (see best for us ¦ur ¬future® income would be saved” FIGURE 38: Credit amount and frequency of instrument use: The Bhattis (PAKISTAN) JULY 2014–MAY 2015 (US$) $140 $120 $100 $80 $60 $40 $20 $0 JulAug SepOct Nov Dec Jan Feb Mar Apr May Agent credit Borrowing from friends and family Informal store credit

Tanzania. Photo by Hailey Tucker.

7. IMPLICATIONS FOR FINANCIAL SOLUTIONS: TRANSLATING THE EVIDENCE INTO FINANCIAL TOOLS TAILORED TO EACH SMALLHOLDER HOUSEHOLD PROFILE How FSPs sustainably engage with smallholder households, particularly when it comes to DFS, depends heavily on the household proﬁle they serve. Smallholder Diaries households can be placed into three loose proﬁles based on observed dif- ferences in (i) their patterns of agricultural production and sales and (ii) the breadth of their household ﬁnancial portfolios. Each proﬁle requires its own tailored ﬁnan- cial solutions, and the degree to which smallholder households might engage with mobile ﬁnancial services varies, particularly in terms of their starting point. Œranslating this evidence into ﬁnancial solutions, the Smallholder ‰iaries observed ey diﬀerences across the sample in (i) their patterns of agricultural pro- duction and sales (see Section ‘) and (ii) the breadth of their household ﬁnancial portfolios (see Section ”) £ringing these two variables together, the Smallholder ‰iaries households can be placed into three loose proﬁles, which correspond with the location of the sample (see Œable ¡) (i) noncommercial smallholder house- holds (–o—ambiue) (ii) those in loose value chains (Œan—ania) and (iii) those in tight value chains (aistan) (see igure Ÿž) œach proﬁle has a distinct set of chal- lenges and features that service providers will need to fully understand and address to successfully meet their demands for ﬁnancial tools TABLE 8: Financial Portfolio Breadth and Sales and Consumption of Agricultural Production in the Smallholder Diaries MOZAMBIQUE TANZANIA PAKISTAN Financial portfolio breadth (median) Number of ﬁnancial instruments 3 12 18 Number of savings instruments 2 5 3 Number of credit instruments 1 6 14 Number of households using credit (%) 67 92 100 Main method of borrowing money (self-reported) Friends and family Friends and family Friends and family, agents Main method of funding inputs (self-reported) Savings, current income, Savings, current Credit in-kind income Sales and consumption of agricultural production Percentage of harvests consumed (self-reported) 68 18 12.5 Buyers N/A Village agents; Agricultural agents; agricultural agents; retail village agents consumers at market Crop storage Bags in the house, Bags in the Tin and aluminum traditional bamboo silos house containers Main method of storing value (self-reported) Money at home Money at home Livestock; money at home; money guard (arthi) | 79

Financial Diaries with Smallholder Families - Page 98

80 | SMALLHOLDER DIARIES FIGURE 39: Three smallholder profiles based on degree of agricultural commercialization and breadth of financial portfolio Rapid Tight value chains monetization Smallholders in loose value chains Narrow Broad portfolio portfolio/ credit availability Noncommercial smallholders Non- monetization/ self-consumption SOLUTIONS FOR NONCOMMERCIAL SMALLHOLDER HOUSEHOLDS: STRETCHING SMALL AMOUNTS OF INCOME, IMPROVING CROP STORAGE AND OTHER MEANS OF SAVING Financial tools to store and stretch the small amounts of income earned by noncommercial smallholder households on an infre uent basis, particularly from casual labor, could be useful to these families ˜oncommercial small- holder households receive income infreuently and are not able mae savings deposits on a structured schedule, but ﬁnancial mechanisms such as ˆS†ˆs that allow families to deposit intermittently when they do have a bit of surplus can help households build up small reserves Œhis pool of stored funds could also be useful as a type of household-generated safety net during the hunger season Œhat said, any additional savings derived from their very limited incomes would still liely be insuﬃcient to insulate the household from shocs or meet all consumption needs ”lose to two-thirds of the Smallholder Diaries sample in oambi ue lost crops in storage, which points to a clear opportunity to improe both their agricultural and ﬁnancial lies Smallholders in the –o—ambiue sample stored their agricultural production in burlap sacs in the house or simple bamboo cis- terns that pests could easily penetrate ™ith these rudimentary crop storage sys- tems, households were unable to store crops for many months Securing agricultural output in a more resilient form of storage (eg, reinforced plastic bags, small metal grain silos) could be an improvement for noncommercial “… smallholders ˆ range of ﬁnancial tools and service providers could support the uptae of improved storage methods, including targeted layaway products from retailers and commitment savings plans and tailored credit products from Ss Œhese could be facilitated by mobile phones or use of standard transactions

IMPLICATIONS FOR FINANCIAL SOLUTIONS | 81 his proﬁle is liƒely the most challenging to sere ia mobile ﬁnancial ser- ices, and successful approaches will need to oercome the barriers of lim- ited access to and capability with mobile phones Œhe experience and understanding of mobile phones in the –o—ambiue sample was the lowest in the Smallholder ‰iaries roughly half (”” percent) had used a phone in the past year and many new only how to receive calls with their phone (“” percent) ƒow levels of literacy, the lac of access to electricity to charge phones, and limited infrastructure presented additional challenges, and no one in the –o—ambiue sample used mobile money during the Smallholder ‰iaries €nput retailers, oﬀ-taers, and Ss would need to address these constraints and then develop tailored products that oﬀer the option or provide incentives to use mobile ﬁnancial services (eg, to receive payment, to contribute to a layaway account for inputs or an improved form of crop storage), perhaps in the near term through an agent (though this poses its own challenges) and, with time, independently Šmproed agronomic practices and better agricultural risƒ management would also be important in this proﬁle, and oﬀ-taƒers interested in reach- ing noncommercial smallholders and embedding them in their supply chain would need to bundle this agronomic support with ﬁnancial tools ƒoans from oﬀ-taers to ﬁnance agricultural production may be relevant to some noncommercial smallholder households, but probably only when com- bined with a fairly extensive range of support that enables some maretable surplus, including high-uality seeds and inputs, agronomic information, weather information, and access to marets ˜oncommercial smallholders present the most challenging producers to lin to value chains, and it will be important to explore ways to push beyond the estimated • percent of small- holders currently embedded in tight value chains (†hristen and ˆnderson ‘’…Ÿ), looing for opportunities to broaden this approach to include addi- tional crops and livestoc, oﬀer ﬁnancial services beyond input credit, and reach lower-income, less silled producers •onﬁnancial interentions are also critical to improe the well-being of noncommercial households €mprovements in ﬁnancial inclusion are impor- tant for noncommercial smallholder households, but only one of many rele- “‘ vant and needed interventions Œhe experiences of the Smallholder ‰iaries sample in –o—ambiue emphasi—ed the connections among health and nutri- tion, food security, and irregular incomes €nterventions to improve health- care, provide wider safety nets, increase agricultural production, and better preserve the harvest are important for improvements in the overall well-being of noncommercial households ™hen someone in the family needs medical attention, for example, they cannot pay for transportation to a health clinic or hospital without cash on hand, which may prevent them from seeing care §ouchers that allowed bus and taxi drivers to receive payment from ˜‡¦s or the government for transportation to medical care could ease this constraint

82 | SMALLHOLDER DIARIES SOLUTIONS FOR SMALLHOLDER HOUSEHOLDS IN LOOSE VALUE CHAINS: MOVING SAVINGS FROM THE MATTRESS INTO A FINANCIAL TOOL, LEVERAGING DFS –erwhelmingly, families in the Smallholder Diaries sample in anania ƒept their saings in-ƒind or under the mattress, presenting a clear opportu- nity for FS‚s to oﬀer more aenues to store money Œhese relatively higher- income (or somewhat less poor) households could in some cases more successfully postpone crop sales until maret prices increased after the initial harvest period, and some also purposefully set aside their stored crops to “save” for a targeted lump sum of money or more generally in case of emergency Œai- lored savings instruments could be designed around this seasonal pattern as well as the agricultural and household needs of smallholder families, but Ss must recogni—e that these ﬁnancial savings tools would be competing with the nonﬁnancial crops in storage as their own form of savings, each with its own rate of return and attributes †rops can be consumed or traded in times of need and the gains on crop sales due to changes in maret price can be substantial, but prices are unpredictable and crops can rot or be destroyed Œhese savings tools could also be positioned as compatible, and smallholder households may have an interest in using both ™arehouse receipts may also be useful for both storage and ﬁnancial purposes Serice proiders need to taƒe into account the full range of income sources when ealuating risƒ ›ousehold cash ﬂows and their olatility—from income sources both related to and independent of agricultural produc- tion—should be considered when calculating the risƒ and terms of any loan †ommercial smallholders are in a stronger position to tae on debt than non- commercial smallholders, but a deeper understanding of the dynamics of household cash ﬂows could provide comfort to service providers and borrowers alie by demonstrating that certain forms of agricultural lending may not be as risy as previously perceived, or that they are at least mitigated by other less volatile sources of income outside of agricultural production ¦n the other hand, credit is not always the answer ƒenders may consider some loans too risy, and some consumers may wish to avoid debt in any circumstances, in which case some near-term investments and purchases can be achieved through other ﬁnancial mechanisms, such as layaway products, leasing, commitment savings, and in-ind payments Šn countries with a robust mobile ﬁnancial serices infrastructure, such as anania, access to credit and payments could be facilitated ia digital chan- nels –any smallholder households in the Smallholder ‰iaries sample in Œan- —ania purchased fertili—er, and access to a mobile-based oﬀering that provides this ind of small-scale ﬁnance could build on this interest and be applied to other goods and assets Savings-based and layaway ﬁnancial mechanisms— where goods are saved for in small increments over time and secured when the total amount has been reached—that are delivered via digital channels could also facilitate these purchases Œhey would also provide an alternative for those who prefer to avoid taing on debt or who are not creditworthy €n addition, Smallholder ‰iaries households used cash to mae payments to formal entities such as government oﬃces, schools, and hospitals, which suggests that ‘£ and

IMPLICATIONS FOR FINANCIAL SOLUTIONS | 83 ‘‡ payments could present other use cases for formal ﬁnancial services or mobile payments ”loser connections to buyers and aggregators in the alue chain could also beneﬁt this proﬁle, and in countries such as anania that hae a robust dig- ital infrastructure, these relationships and serices could be enabled ia digital channels Such services could facilitate the creation of purchase agree- ments or formal contracts, for example, against which smallholders could bor- row for fertili—er, an oft-cited need among the Œan—ania sample households FS‚s that oﬀer any type of insurance would need to demonstrate its alue ery early on to build trust, counter scepticism, and distance their product from other negatie e‡periences ˆgricultural insurance may be worth explor- ing, but insurance and insurers had a wea reputation among the sample Households are wary of any insurance due to reports of poor service when using the national health insurance card Stories abound among the Œan—ania ‰iaries sample where someone has gone to a public hospital with their national health insurance card, only to wait for hours to be served, while someone else who pays for services in cash is immediately assisted SOLUTIONS FOR SMALLHOLDERS IN TIGHT VALUE CHAINS: ENGAGING MIDDLEMEN, CHANNELING SAVINGS Smallholders in tight alue chains need ﬁnancial tools that facilitate their relationships with middlemen and could beneﬁt from mechanisms that reduce their dependence on this one FS‚ Œhe households in the aistan Smallholder ‰iaries had by far the widest, most complex portfolio of ﬁnancial tools, but a signiﬁcant portion of their income sources and ﬁnancial tools ﬂowed through an individual middleman Œhere is clearly space for more options on the supply side, but to compete, other service providers would need to match the ﬂexibility and proximity of arthis, or oﬀer better terms and service €n the event of crop failure, for example, the arthi will liely agree to be repaid in the following year, while most Ss would not Smallholder families may have issues with the middleman, particularly the timing of repayment immediately upon harvest, but they do return to them, at least in absence of a compelling alternative Saings is widespread and the opportunity for FS‚s to oﬀer ﬁnancial tools to harness this is great ™hen smallholder households saved from ¦ctober to ˜ovember and ˆpril to –ay, for example, when agricultural income was high, the largest share of these savings went to arthis or to savings in the house, and of course neither earned any interest Šn addition to proiding ﬁnancial serices alternaties to the middlemen, greater transparency could strengthen and better balance the arthi–small- holder relationship Smallholders could beneﬁt from a short-term loan to repay arthi debt and obtain inputs elsewhere, which would also allow breathing room to tae advantage of price increases as time passes after harvest œven when savings are ept with the arthi, digital ﬁnance and tools may help farmers

84 | SMALLHOLDER DIARIES Pakistan. Photo by Erin Scronce. eep better trac of the large amounts of money they have saved Œhe written ledger of savings that arthis eep for farmers, for example, could be converted to a new record eeping system using mobile technology he middlemen themseles may beneﬁt from a wider range of ﬁnancial tools to cope with li uidity constraints, transaction costs, and risƒ ˆddress- ing their pain points may allow greater ﬂexibility on the products and terms that they oﬀer the smallholder households with whom they wor œasing ﬁnancial pressures at the top of the value chain may beneﬁt farmers at the bottom, though middlemen would still hold the balance of power in negotiating with smallhold- ers Œhe eventual impact on smallholder households of improving the ﬁnancial mechanisms available to middlemen should be further explored AREAS FOR FURTHER RESEARCH Œhe Smallholder ‰iaries are designed to spotlight some ey observations in income, expenses, ris management, and ﬁnancial portfolios across three dis- tinct proﬁles of smallholder households in –o—ambiue, Œan—ania, and ai- stan and drive advances in their ﬁnancial inclusion ™hile the speciﬁc context of each sample is important, these ﬁndings also have implications for the small- holder household sector worldwide Œhe sample of smallholder households in each study country has characteristics shared with broader types of smallholder segments identiﬁed in countries around the world, which presents the opportu-

IMPLICATIONS FOR FINANCIAL SOLUTIONS | 85 nity to discuss the ﬁnancial tools these common segments demand regardless of their location ˆs in any research eﬀort, the results spar as many uestions as they may answer and the wor to better understand smallholder households and improve their agricultural and ﬁnancial lives continues ¯ey areas for further exploration include how value chain ﬁnance approaches could reach lower-income, less silled smallholders, embed additional crops and livestoc in tight value chains, and bundle ﬁnancial services beyond input credit in their production support pacages Œhe relationship among partici- pants is also an important aspect of value chains ©nderstanding how small- holders and purchasers beneﬁt from these connections, ﬁnancially and otherwise, and where vulnerabilities remain will be important to their success- ful and euitable expansion €n addition, as Ss develop and tailor a range of ﬁnancial tools to speciﬁc proﬁles of smallholder households, it would be useful to understand how household ﬁnancial portfolios change over time, including what ﬁnancial mechanisms are replaced, how transaction costs change, when informal tools continue to complement formal ﬁnancial tools, and where gaps persist †onsidering the level of ﬁnancial exclusion among smallholder house- holds, additional research on how Ss can most eﬀectively develop products that rapidly and eﬀectively reach scale is also ey orthcoming data and analysis from the national smallholder surveys will help further contextuali—e the results of the Smallholder ‰iaries and contribute to building the evidence base on the ﬁnancial and agricultural lives of small- holder households Œhe segmentation of the smallholder sector in each country will inform Ss developing ﬁnancial tools tailored to each segment and design- ing the business cases to deliver them sustainably €t will also identify client groups whose primary needs are in basic health, education, and infrastructure, and where collaboration with government and ˜‡¦ partners would be even more important Œogether, the data and results from the Smallholder ‰iaries and the national surveys and segmentations should help Ss, government partners, ˜‡¦s, and other staeholders better understand smallholder house- holds and mae strides in advancing their ﬁnancial inclusion

ANNEXES

ANNEX 1 AVERAGE VALUE OF ASSETS PER HOUSEHOLD Asset values were calculated using self-reported, approximate values from respondents themselves. FIGURE A1-1: Smallholder Diaries sample in Mozambique: Average value of assets per household (US$) Vehicles $778 Primary residence $606 Farmland $217 Household items/materials $60 Furniture $60 Electronics $46 Livestock $23 Other $20 Agricultural assets/production $20 Rental property $15 Appliances $13 $0 $200 $400 $600 $800 $1,000 FIGURE A1-2: Smallholder Diaries sample in Tanzania: Average value of assets per household (US$) Primary residence $1,014 Farmland $691 Other property $208 Livestock $190 Furniture $124 Household items/materials $99 Transportation $35 Electronics $32 Agricultural assets/production $25 Appliances $8 Other $7 Rental property $4 $0 $200 $400 $600 $800 $1,000 $1,200 FIGURE A1-3: Smallholder Diaries sample in Pakistan: Average value of assets per household (US$) Farmland $17,510 Primary residence $5,232 Livestock $2,702 Agricultural assets/production $526 Household items/materials $355 Vehicles $223 Furniture $205 Electronics $82 Appliances $71 Other $52 $0 $5,000 $10,000 $15,000 $20,000 88 |

Financial Diaries with Smallholder Families - Page 106

ANNEX 2 DEFINITIONS OF INCOME SOURCES IN THE SMALLHOLDER DIARIES Agricultural production income is money earned from the production and sale of ag- ricultural goods, such as crops, livestock from an established agricultural business, and livestock byproducts (milk and eggs). Casual labor income includes irregular income from short-term employment, such as work on construction sites or helping with the harvest on other people’s farms. People employed in casual labor are not making management decisions or investments, which distinguishes it from self-employment income. Nonemployment income includes grants and other institutionally provided support from charities, hospitals, and government. Regular/waged employment is salaried income received that has been at least tacitly agreed to be earned on a regular basis. Rental income is usually linked to the lease of land or property. Resources received is monetary or in-kind contributions or remittances provided to re- spondents through their social networks. Self-employment income relates to sole-proprietor microbusinesses in which the owner manages an enterprise and invests money in inputs, stock, and tools; the activity may be formal or informal, and the work may be full-time, part-time, or occasional. Self-employ- ment income does not come from agricultural production, as this would be considered agricultural production income. | 89

ANNEX 3 DEFINITIONS OF FINANCIAL TOOLS IN THE SMALLHOLDER DIARIES Accumulating Savings and Credit Associations (ASCAs): Relatively more complex infor- mal savings groups (compared to ROSCAs, see below) that allow members to build up savings over time, lend the group fund to one another, and accumulate interest. A share- out typically occurs once a year when members divide the savings and earned interest among the group. Agricultural middleman credit or agent credit is a loan from an input supplier, usually with the understanding that repayment will be in cash or in-kind after that crop has been harvested. In Pakistan, these middlemen are known as arthis. Farmers sell produce to arthis and obtain fertilizer and pesticides on credit. They can also go to arthis to ﬁnance other major expenditures, such as weddings or emergencies. Borrowing from friends and family includes informal borrowing from their social network. Borrowing from an informal group includes borrowing from ASCAs and other community- led savings and credit groups. Business loans are loans in which someone borrows money for a business and is individu- ally liable for repaying the loan. Usually these loans are from banks, cooperatives, credit unions, or microﬁnance institutions if the borrower does not belong to a group. Checking accounts are current accounts with a formal commercial bank. Credit at a store is an arrangement whereby the shopkeeper lets a household member take goods now and pay later. A household member may buy a sack of ﬂour from a shop- keeper on credit, for example, and promise to pay for it on his or her next visit to the shop. Credit given occurs when the respondent runs any type of small business and lets clients take items on credit and pay later. Hire purchase is when an individual purchases something from a shop but does not pay the full amount upfront. The good is taken ﬁrst, usually upon payment of a deposit or an installment, and then the buyer continues paying installments over time until the good is paid off. Joint liability loan is when someone belongs to a group and the group has an account with a microﬁnance institution or bank that lends them money. The person may take a portion of that as a personal loan, but all the members of the group are guarantors. If the person stops paying, then all the group members are responsible for covering her debt. Layaways are ﬁnancial tools in which a person pays in installments for a good, and ac- quires it only once all payments are made. Lending to friends and family is when members of the household or social network pro- vide others with ﬁnancial services. 90 |

ANNEX 3 | 91 Loan from employer is a loan from your employer (not a wage advance). The repayments may be deducted from pay slips over multiple periods or the borrower may have to repay separately. Money guard is a person who holds money and keeps it safe for someone else. Moneylender is a private individual who lends money and charges interest. Interest rates tend to be high and repayment times tend to be strict, but people borrow from money- lenders during emergencies or when they prefer to keep the reason for the loan private from friends and family. Pawning occurs when a person brings something of value (e.g., gold, phones, appliances) to a pawn shop or an individual who then gives her money for the item. If the borrower repays on time, then they can have the item back. If not, they forfeit the item. Rotating Savings and Credit Associations (ROSCAs) are informal savings groups in which members generally combine their savings together at regular, recurring meetings and take turns giving the entire pot to one member. Savings in the house typically includes cash stored in a safe, readily accessible place. Note that strategies such as storing gold and raising livestock are not classiﬁed as savings at home, but rather household (physical) assets. Supplier credit refers to informal credit given by a small business owner’s regular suppli- ers. The supplier may allow the business owner to pay off the loan after a week, a month, or even longer, depending on the arrangement. Wage advance is when an employer pays an employee’s wages early. The amount the employee owes is either deducted from their salary, repaid in one lump sum, or over a few paychecks.

ANNEX 4 USE OF FINANCIAL TOOLS FOR SAVINGS AND CREDIT AMONG SMALLHOLDER DIARIES HOUSEHOLDS The averages calculated in the tables below refer to instruments in use. Mean withdrawals and deposits refer to average transaction size. Latest balance refers to self-reported balances at the end of the study, when available. TABLE A4-1: Percentage of Households with Financial Tools for Savingsa TRANSACTIONS % HOUSEHOLDS LATEST BALANCE INSTRUMENTS WITHDRAWAL DEPOSIT SAVINGS INSTRUMENTS USING (MEAN) # PER (MEAN) (MEAN) (MEAN) Mozambique Savings in the house 87 $17.44 9 $12.88 $10.90 Lending to friends and family 41 $10.49 2 $18.23 $33.77 Credit given 18 $1.92 3 $4.64 $3.17 ROSCA 12 $15.89 7 $51.91 $10.37 ASCA 9 $17.38 6 $41.05 $6.18 Checking account 8 $246.99 7 $69.14 $106.91 Using money guard 5 $178.86 2 $34.62 $37.50 Layaway 1 $0 5 N/A $14.42 Tanzania Savings in the house 100 $18.25 101 $3.10 $10.86 Lending to friends and family 48 $10.52 2 $15.84 $15.91 ASCA 53 $5.83 3 $34.47 $1.01 ROSCA 33 $2.14 28 $32.23 $1.09 Credit given 29 $5.20 6 $3.82 $2.81 Using money guard 21 $26.54 3 $54.15 $125.48 Layaway 17 $13.87 3 $50.20 $25.80 Checking account 5 $128.18 28 $164.43 $267.27 Pakistan Savings in the house 85 $72.74 9 $82.78 $146.65 Lending to friends and family 61 $45.32 2 $108.62 $135.36 Using money guard 56 $122.68 4 $280.89 $583.83 Credit given 23 $175.95 20 $51.92 $21.15 ASCA 19 $37.55 1 $130.14 $80.60 Checking account 16 $158.00 2 $98.39 $107.83 Layaway 7 $285.28 2 $195.77 $15.06 ROSCA 4 $87.85 7 $126.50 $9.04 a. Certain instruments were uncommon (e.g., some were used by only one household) and in such cases the mean refers to very few observations. 92 |

ANNEX 4 | 93 TABLE A4-2: Percentage of Households with Financial Tools for Credit # TRANSACTIONS DEPOSIT % HOUSEHOLDS LATEST BALANCE INSTRUMENTS BORROWING REPAYMENT CREDIT INSTRUMENTS USING (MEAN) PER (MEAN) (MEAN) (MEAN) Mozambique Borrowing from friends and family 59 $3.41 2 $8.62 $10.80 Credit at a store 22 $2.57 2 $3.28 $3.89 Act as money guard 5 $1.33 1 $0.58 $17.60 Borrowing from informal group 5 $6.92 2 $6.49 $27.69 Pawning assets 5 $6.58 1 $1.73 $1.83 Wage advance 2 $0.43 1 $0 $28.85 Business loan 1 $956.60 3 $1,009.68 $26.54 Supplier credit 1 $0 1 N/A $12.98 Tanzania Borrowing from friends and family 77 $4.53 2 $11.56 $11.20 Borrowing from informal group 67 $10.74 6 $15.74 $7.83 Credit at a store 62 $2.00 4 $4.47 $7.31 Agent credit 14 $12.44 1 $41.35 $71.87 Act as money guard 13 $0.09 3 $18.67 $18.58 Supplier credit 8 $7.95 2 $13.78 $5.55 Moneylender 6 $43.42 2 $125.77 $15.51 Hire purchase 3 $0 2 $8.70 $0.00 Business loan 1 $12.92 2 $26.36 $13.18 Tafu airtime credit 1 $0 4 $0.39 $0.39 Pakistan Borrowing from friends and family 99 $74.08 2 $80.67 $97.14 Agent credit 97 $106.98 9 $49.43 $199.19 Credit at a store 94 $22.87 14 $5.99 $36.98 Joint liability loan 27 $122.16 2 $252.80 $259.05 Business loan 16 $1,244.98 2 $939.74 $262.38 Hire purchase 14 $90.88 4 $176.53 $56.57 Loan from employer 7 $88.35 3 $64.76 $57.92 Pawning assets 6 $71.71 2 $351.39 $100.40 Moneylender 5 $140.56 1 $200.79 $150.59 Wage advance 5 $97.05 4 $82.83 $50.20 Supplier credit 3 $29.72 6 $26.40 $31.43 Act as money guard 2 $107.93 22 $4.17 $28.79

REFERENCES AgriFin. 2015. Agricultural Value Chain Finance. A Bankers’ Guide. Washington, D.C.: World Bank. Armendariz, Beatriz, and Jonathan Morduch. 2010. The Economics of Microﬁnance. Cambridge, Mass.: MIT Press. Bankable Frontier Associates. 2011. “Initial In Focus Report Back to FSP.” Unpublished presentation for Bill & Melinda Gates Foundation. Christen, Robert P., and Jamie Anderson. 2013. Segmentation of Smallholder House- holds: Meeting the Range of Financial Needs in Agricultural Families. Focus Note 85. Washington, D.C.: CGAP. Collier, Benjamin, and Jerry Skees. 2014. Increasing the Resilience of Financial Intermedi- aries through Portfolio-Level Insurance against Natural Disasters. ResearchGate. Collins, D., J. Morduch, S. Rutherford, and O. Ruthven. 2009. Portfolios of the Poor: How the World’s Poor Live on $2 a Day. Princeton, N.J.: Princeton University Press. Conning, Jonathan H., and Christopher R. Udry. 2005. “Rural Financial Markets in Developing Countries.” New Haven, Conn.: Yale University Economic Growth Center Discussion Paper No. 914, June. Cuevas, Carlos E. and Jamie Anderson. 2016. “Understanding Demand, Driving Innova- tion: Smallholder Households and Financial Services.” Working Paper. Washington, D.C.: CGAP. Dalberg Global Development Advisors. 2012. Catalyzing Smallholder Agricultural Finance. Demirguc-Kunt, A., L. Klapper, D. Singer, and P. Van Oudheusden. 2015. The Global Findex Database 2014: Measuring Financial Inclusion around the World. Policy Research Working Paper 7255. Washington, D.C.: World Bank. http://www-wds.worldbank.org/ external/default/WDSContentServer/WDSP/IB/2015/10/19/090224b08315413c/2_0/ Rendered/PDF/The0Global0Fin0ion0around0the0world.pdf GIZ (Deutsche Gesellshaft fur Internationale Zusammenarbeit). 2011. Agricultural Finance—Trends, Issues and Challenges. Eschborn: GIZ. Grossman, Jeremiah, and Michael Tarazi. 2014. “Serving Smallholder Farmers: Recent Developments in Digital Finance.” Focus Note 94. Washington, D.C.: CGAP, June. Hazell, P., C. Pomareda, and A. Valdés, eds. 1986. Crop Insurance for Agricultural Development. Issues and Experience. Baltimore: Johns Hopkins University Press. IFAD (International Fund for Agricultural Development). 2010. Rural Poverty Report 2011. Rome: IFAD. Kenya Financial Sector Deepening Trust. 2014. “Kenya Financial Diaries: Shilingi Kwa Shilingi—The Financial Lives of the Poor.” http://www.fsdkenya.org/new/our-work/ ﬁnancial-diaries.html | 95

96 | SMALLHOLDER DIARIES Lowder, S. K., J. Skoet, and S. Singh. 2014. What Do We Really Know about the Number and Distribution of Farms and Family Farms in the World? ESA Working Paper No. 14- 02. Rome: United Nations Food and Agriculture Organization. Mahul, Olivier, and Jerry Skees. 2007. Managing Agricultural Risk at the Country Level: The Case of Index-Based Livestock Insurance in Mongolia. Policy Research Working Paper 4325. Washington, D.C.: World Bank. Mattern, Max, and Michael Tarazi. 2015. “Designing Digital Financial Services for Small- holder Families: Lessons from Zimbabwe, Senegal, Rwanda, and Cambodia.” Perspec- tives 1. Washington, D.C.: CGAP. Miller, Calvin, and Linda M. Jones. 2010. Agricultural Value Chain Finance: Tools and Lessons. Rome: Practical Action Publishing and the Food and Agriculture Organization of the United Nations. Miranda, Mario J. 2009. “Agricultural Insurance in Developing Countries: Practitioners’ Perspectives. Summary of Findings.” Unpublished. Bill & Melinda Gates Foundation. Miranda, Mario J., and Katie Farrin. 2012. “Index Insurance for Developing Countries.” Applied Economic Perspectives and Policy: 1–37. Pénicaud, Claire, and Arunjay Katakam. 2013. “State of the Industry 2013. Mobile Financial Services for the Unbanked.” GSMA. Rutherford, Stuart. 2000. The Poor and Their Money. Delhi: Oxford University Press. Skees, J., P. Hazell, and M. Miranda. 1999. New Approaches to Crop Yield Insurance in Developing Countries. Washington, D.C.: International Food Policy Research Institute. Udry, Christopher. 1994. “Risk and Insurance in a Rural Credit Market: An Empirical Inves- tigation in Northern Nigeria.” Review of Economic Studies 61 (3): 495–526. Walker, Thomas S., and N. S. Jodha. 1986. “How Small Farm Households Adapt to Risk.” In P. Hazell, C. Pomareda, and A. Valdés, eds. Crop Insurance for Agricultural Development. Issues and Experience. Baltimore: Johns Hopkins University Press. World Bank. 2014a. “World Development Indicators: GDP per capita.” Accessed Sep- tember 2015. ———. 2014b. Global Findex Database. www.econ.worldbank.org Wyman, Oliver. 2007. “Sizing and Segmenting Financial Needs of the World Poor.” Unpublished. Bill & Melinda Gates Foundation.

NOTES 1. See Cuevas and Anderson (2016) for a discussion of smallholders in poverty statistics and their role in food security and ﬁnancial inclusion. 2. The several income sources identiﬁed by the Smallholder Diaries are deﬁned in Annex 2. Note that households may have multiple types of the same income category (e.g., income from cultivation of four crops, wages from casual labor on two different jobs). Each individual income source is counted and tracked separately. 3. The ﬁnancial tools (or instruments) identiﬁed by the Smallholder Diaries are deﬁned in Annex 3. Note that there may be multiple, distinct uses of each type of ﬁnancial tool in Smallholder Diaries households (e.g., participation in two savings and credit groups, informal credit at three different stores). Each individual ﬁnancial tool is counted and tracked separately. 4. ROSCAs are informal savings groups in which members generally combine their savings together at regular, recurring meetings and take turns giving the entire pot to one member. ASCAs are somewhat more complex informal savings groups. They allow members to build up savings over time, lend the group fund to one another, and accumulate interest. A share-out typically occurs once a year when members divide the savings and earned interest among the group. 5. In a layaway purchase agreement, a retailer holds merchandise secured by a deposit until it is paid in full by the customer, usually through a series of payments over time. 6. See Lowder, Skoet, and Singh (2014), Christen and Anderson (2013), Dalberg (2012), and Wyman (2007) for background on the population estimates related to smallholder farmers and households. 7. Data on the number of smallholder households worldwide are fraught with caveats and nuance. See Cuevas and Anderson (2016) for a discussion of smallholders in poverty statistics and their role in food security and ﬁnancial inclusion. 8. See, for example, Conning and Udry (2007), GIZ (2011), and Udry (1994). 9. Not all smallholder households are located in rural areas. Data collected by CGAP in nationally representative surveys of smallholder households in Mozambique and Uganda found an estimated 15 percent of smallholder families in peri-urban and urban areas. 10. The World Bank Group Global Findex database, the world’s most comprehensive database on ﬁnancial inclusion, provides in-depth data on how individuals save, borrow, make payments, and manage risks. Collected in partnership with the Gallup World Poll and funded by the Bill & Melinda Gates Foundation, Global Findex is based on interviews with about 150,000 adults in over 140 countries. See http://www.worldbank.org/en/ programs/globalﬁndex. 11. The Findex 2014 database does not report ﬁndings for an urban/rural split due to inconsistencies in the deﬁnitions of urban and rural across countries, although it includes estimates for account penetration in rural populations. The database effectively does have these estimates; Table 1 was generated using those estimates, and extrapolating the urban account penetration using rural population shares from IFAD Poverty Report (2010). 12. CGAP retained the services of Bankable Frontier Associates (BFA) to manage the Smallholder Diaries. For in-country data collection, BFA worked with International Capital Corporation in Mozambique, Digital Divide Data in Tanzania, and RCons in Pakistan. 13. CGAP retained the services of InterMedia to manage the national surveys of smallholder households, and it worked with Ipsos in Mozambique and Uganda for in-country data collection. National surveys and segmentations of the smallholder sector are also underway in Tanzania, Côte d’Ivoire, and Bangladesh, and results and data from all ﬁve countries will be published in 2016. | 97

98 | SMALLHOLDER DIARIES 14. Key sources on the diaries methods and ﬁndings are Collins et al. (2009) and Kenya Financial Sector Deepening Trust (2014). Other references indicated as appropriate. 15. All data were collected between April 2014 and July 2015, including the initial question- naires and additional qualitative modules. Data collection on household cash ﬂows started in June 2014 and ended in June 2015. The module on risk was administered in July 2015. 16. All data and questionnaires are available at http://www.cgap.org/topics/ﬁnancial-innova- tion-smallholder-families. 17. “In-kind” refers to goods, services, and transactions not involving money (i.e., payments in-kind, barter transactions). 18. The Financial Diaries methodology was developed by David Hulme of the University of Manchester and Stuart Rutherford of SafeSave. 19. Smallholders Diaries households were selected purposefully and not randomly, and the samples are not statistically representative of smallholder farmers in these areas or in the three countries. There are additional reasons the Diaries methodology does not use a random sample. The study is able to include only households that are willing to commit to a year-long study and, to minimize attrition issues, are likely to stay in the community. Households initially oversampled to include 286 households and the study ended with 273. Households left the study by leaving the study villages, seasonal migration, and occasionally by the prompting of the research team due to concerns about the house- hold’s willingness to be forthcoming about important sources of income. The research ﬁrms provided small cash gifts at surprise times throughout the study to thank respon- dents for their participation. The value of these cash gifts were a very small share of income for most households. The gifts were tracked as income and expenditures enabled by these extra inﬂows were also tracked. 20. Figures for assets were self-reported. 21. The exchange rates used in this paper, calculated as the average exchange rate during the period of the study, are as follows unless otherwise indicated: Mozambique Metical to the U.S. Dollar: 34.66; Tanzania Shilling to the U.S. Dollar: 1934.72; Pakistan Rupee to the U.S. Dollar: 99.60. 22. In the terminology of the Smallholder Diaries, wage labor generates salaried income that has been at least tacitly agreed to be earned on a regular basis. This is distinct from casual labor, which is irregular income from short-term employment, such as work on construc- tion sites or helping with the harvest on other people’s farms. 23. The several income sources identiﬁed by the Smallholder Diaries are deﬁned in Annex 2. Note that there may be multiple, distinct income streams from each type of income source in Smallholder Diaries households (e.g., income from cultivation of four crops, wages from casual labor on two different jobs). Each individual income source is counted and tracked separately. 24. Some income sources even go negative in a given month, since the Diaries charted net income from the source. In agriculture production and self-employment, some expenses may be incurred before revenue, resulting in negative net income for a given month. While agricultural revenues in Pakistan can be high, agricultural income is calculated by subtracting farming expenses, which can be substantial, from gross revenue. 25. This includes any crops consumed, traded, or given away for any reason. 26. In the Smallholder Diaries a distinction was made between self-generated agricultural production income and income from casual labor related to agriculture (e.g., working on a neighbor’s farm). Someone providing casual labor is not the main decision-maker or investor in those crops or livestock and their work starts and stops at the will of someone outside the household. 27. For a detailed exploration of value chain ﬁnance, see AgriFin 2015 (forthcoming) and Miller and Jones (2010). 28. Commercial smallholders in tight value chains have the capacity to generate reliable, high- quality outputs that are sold on a contract basis through relatively highly organized value chains. See Christen and Anderson (2013). 29. The village in the Tanzania sample focused on potato production had the same pattern, with expenses dipping to their lowest during hardship months. 30. Covariant risk arises when many households in one area are adversely affected by a single phenomenon such as a natural disaster, epidemic, unexpected change in world prices, macroeconomic crisis, or civil conﬂict. Individual risks, in contrast, randomly affect individual households. 31. Most Tanzanian respondents evaluated the cost of crops destroyed in the ﬁeld as zero, since they were unable to put a monetary value on the loss. 32. Strategies to manage risk are fairly well documented. See for example, Skees, Hazell, and Miranda (1999) on crop insurance; Miranda and Farrin (2012) on index-based insurance in low-income countries; Mahul and Skees (2012) on index-based livestock insurance; and Hazell, Pomareda, and Valdés (1986). 98 |

NOTES | 99 33. An interesting experience evolving in Kenya is the e-warehouse system that records onsite storage and allows its use as collateral against short-term loans (See http://www. grameenfoundation.org/what-we-do/ﬁnancial-services/agricultural-ﬁnance),. 34. The other question posited by Walker and Jodha refers to the effect of risk-management methods on static and dynamic social efﬁciency, which (if detrimental), would justify public policy for the sake of social welfare. 35. See for example Armendariz and Morduch (2010). 36. See Miranda (2009) and Collier and Skees (2014). 37. Note that each tool is distinguished by both its ﬁnancial function and its source. For example, each account at a ﬁnancial institution is a separate device. Each ROSCA is a different device and if it has separate functions—such as merry-go-round, accumulation, lending, and welfare—then each of those functions would be registered separately. Each source of informal borrowing, including each individual moneylender and each individual lender among friends and family, is also tracked separately. See Annex 3 for deﬁnitions of the various types of ﬁnancial tools observed in the Smallholder Diaries. 38. Note that there may be multiple, distinct uses of each type of ﬁnancial tool in Smallholder Diaries households (e.g., participation in two savings and credit groups, informal credit at three different stores). Each individual ﬁnancial tool is counted and tracked separately. 39. When considering a household’s ﬁnancial portfolio, both informal and formal ﬁnancial tools, as well as physical assets, including crops and livestock, mobile phones and radios bought and sold with a clear intention of ﬁnancial management, are important. Also see Kenya Financial Sector Deepening Trust (2014). 40. For additional detail and examples of smallholder households and digital ﬁnance, see also Mattern and Tarazi (2015) and Grossman and Tarazi (2014). 41. See, for example, the Purdue Improved Crop Storage triple-layer storage bag (http:// www.entm.purdue.edu/PICS3/index.php) and the POSTCOSECHA metal silo developed by the International Maize and Wheat Improvement Center (http://www.cimmyt.org/en/ projects/effective-grain-storage-project/about-the-project). 42. See Cuevas and Anderson (2016) for a discussion of variables associated with rural poverty and smallholder well-being.

Consultative GrConsultative Group to Assist the Pooroup to Assist the Poor 1818 H Str1818 H Street, N.Weet, N.W., MSN IS7-700., MSN IS7-700 WWashington, DC 20433 USAashington, DC 20433 USA Phone: +1 202 473 9594Phone: +1 202 473 9594 [email protected]@worldbank.orgg

Financial Diaries with Smallholder Families - Page 118

Financial Diaries with Smallholder Families - Page 121

Financial Diaries with Smallholder Families - Page 122

Financial Diaries with Smallholder Families - Page 123

Financial Diaries with Smallholder Families - Page 124

Financial Diaries with Smallholder Families - Page 125