Microcredit for Sharecropper
Title:
Seasonal Micro-Credit for Tenant Share-cropping farmers (with Hisaki Kono and Kazushi Takahashi)
Synopsis:
One unique feature of the land holdings in South Asian countries especially in Bangladesh is most of the rural population is either landless or owns a very small plots of land. For example the average operation farm size in Bangladesh in only 1.6 hectors whereas for Philippines and Thailand this number is close to 3.7 hector and in Latin American countries like in Brazil this number is about 60 hector. Moreover, the tenancy arrangement for South Asian countries is also quite different from other continents of the world. In Latin American countries where tenancy in largely of the fixed rent variety, in Asia the tenancy is characterized by a high prevalence of sharecropping, in which the tenant yields to the landlord an agreed-upon share of the crop. In Asian context the fraction of tenant land under share-cropping contract varies from 30% (Thailand) to 50% (India) to 60% (Indonesia) to all the way up to 90% in Bangladesh (Ray 1998).
Neo-classical theory of land contract will argue that share-cropping contracts are inefficient than fixed rent contract due to incentive and agency problem. However the high prevalence of sharecropping contracts existence in developing countries suggests that there might be some plausible interpretation for having such contracts. Since majority of the tenant farmers is poor, landless, credit constraint and the agricultural output is uncertain owing to high dependence on traditional method of farming; a risk adverse tenant farmer has rational reasons to favor share-cropping over fixed rent mainly due to the risk sharing feature of the contract. On the other hand, landlords has also rational reasons to go for such a contract as poor tenant can occasionally fail to achieve the desired level of harvest and could not be able to pay the fixed rent. In such cases of implicit limited liability clauses, the rent either needs to be forgiven or essentially advances as a debt which the farmer may not be able to pay in future and ultimately need to be forgiven. Hence landlords who are mostly absentee would like to prefer sharecropping contracts than fixed rent ones (for detailed theoretical explanation, see Basu 1990)
In share-cropping contracts sometimes input costs are share by the landlords, however, there are many occasions when these costs are solely borne by the tenant which is quite a common practice in case of absentee landlord. In such cases, tenant farmers need to borrow money on credit which is mostly done from informal money lenders, due to the lack of access to formal financial institution under rural setup. These informal money lenders typically charges interest rate of 10 percent a month which is one of the reason why they are infamously known as “loan sharks”. Given sharecropping contract, sub-optimal output and high credit interest charged on loan amount, most tenant farmers incur economic loss even after a successful harvesting season and higher yield.
The development of micro-credit in developing countries especially in Bangladesh could actually solve the issue by providing relatively low interest paying credit access to rural tenant farmers. In fact, the emergence of the micro-credit was motivated to tackle the loan-sharks from charging incredibly high interest rate to poor people who are usually been excluded by formal financial institutions (See “Bankers to the Poor” by Mohammad Yunus). However, a major drawback of the micro-credit framework is the usually rigid loan repayment rule. Nearly all contracts are fixed in their repayment schedules which entail constant equal weekly payments. If a poor farmer access to such flagship “Grameen” style micro-credit, she will suffer two consequences. Firstly, it will be difficult for him to repay on weekly fashion since she will not be able to reap the benefit of her investments in farming before the harvesting period. Secondly, since the loan discipline is typically enforced quite seriously, either she has to invest less than the given credit amount which might hamper her crop yield and profitability or she has to borrow from the loan-sharks to regularize the repayment record of the micro-credit which will force her to fall onto debt trap again. It appears that, the design of the micro-credit suits more the non-farm sector where one can get regular stream of income and could repay the weekly installment with minor difficulties. Hence current design of micro-credit needs serious reconsideration on how to accommodate the tenant share-cropping farmers and how to cater more for their needs to improve their income which will eventually reduce the poverty and debt trap.
Scholars and experts will unanimously agree with the fact that the provision of Micro-credit schemes has improved the access to the credit market for rural people. However, we need to know the possible loopholes and side-effects of such instruments, and ought to find ways to improve them. The results of this project will be of importance to the poor credit takers and micro-finance institutions and the policy makers supporting them. If we find that a better design of repayment schemes catering for tenant farmers yields improved welfare without higher risks on the side of MFIs, those institutions will revise their contracts to the benefit of the poor. If we find that risks are too high, then policy makers can either compensate MFIs for bearing those risks, or we can look for other ways to alleviate the problem and might need to think beyond micro-credit. For example, government and donor agencies might want to focus more on capacity building of diverse, alternative skills among both agricultural professionals or might need to think of a well designed micro-insurance product to help the poor farmers.
One of the recent findings in the behavioral aspect of microfinance is that the rigid disciplinary mechanism of micro-finance could act as a commitment devise (Jonathan et al. AER 2012). For present biased borrowers such disciplinary framework works in favor of MFIs to improve the repayment collection. Moreover, Stuart Rutherford et al., in the book (portfolios of the poor) showed that poor in general have multiple sources of cash demand and without the presence of a commitment device, it is extremely difficult for them to restrict the temptation and save. The problem with agricultural based micro-credit is this worry of the present-biasedness, poor farmers do not require the entire money at the beginning of the production process, hence in the absence of any credible commitment devise, poor farmers may use the money taken for farming in other purpose, and may need to either borrow with an exorbitantly high interest rate from other informal sources or under-invest in the inputs, which may result in lower yield and profit.
Design of the experiment:
We would like to contrast the traditional micro-credit with the proposed micro-credit design targeted for tenant farmers. In various studies, it was found that MFIs target both farm and non-farm households as long as the household fulfills the eligibility criterion. For example, in Shonchoy 2011, near about 63% borrowers are primarily agricultural households. However, the targeting has not been done based on principal source of income, as a result it creates a distortion as farmers have seasonal lump-sum income whereas non-farm sector earns on a regular basis.
We have selected Northwest Bangladesh (Dinajpur) as our survey area. Agriculture is the dominant sector in the Northwest Bangladesh with more than average penetration of poverty. Here most of the tenant sharecroppers are landless and they typically grow 2-3 crops within a year. Our counterpart organization for the research will be Gana Unnayan Kendra which is an established NGO in the Northern Bangladesh with operating record of 30 years. We already have worked with GUK before on other research projects and GUK is quiet capable and experienced in doing such intervention at the village level. For instance, one of the co-investigators of this research project (Dr. Abu Shonchoy) who is originally from Bangladesh has been working on issues of Northern Bangladesh for the last 8 years and has already has active collaboration with GUK.
We would like to employ Randomized Control Trial (RCT) method because of its ability to measure unbiased estimation. Due to the group based liability framework of micro-credit, we need to randomize at the group level rather than at the individual level. However, due to the limited budget to execute this experiment, we will conduct this experiment at an individual level. To Understand whether this shift, group lending to individual lending, works or not, we will first do a pilot phase during the cropping season of Boro (January 2015 to May 2015) and based on our experience at the piloting phase, we will conduct the main Phase (June 2015 to June 2016) covering the two crop harvesting season of Aman and Boro.
The treatments that we want to impose are the following.
C: Control: No Microcredit will be given to this group
T1: Traditional Credit: The credit amount will be given at the beginning of the plantation period. Farmers are liable to pay the monthly installments of equal amount (with interest) from the first month of loan disbursement. The Loan phase will finish after harvesting cycle with farmers paying the entire amount. For example: For 10,000 taka credit and 17% interest rate, each farmer will pay 2340 taka installment per month (considering five months for crop-cycle, could also be six months then the installment amount would be 1950 per month). Once the full payment is done for one season, farmers will be eligible to borrow for the next season.
T2: Crop-Credit: The credit amount will be given at the beginning of the plantation period. Farmers are liable to pay the full amount with the interest at the end of the harvesting period. Using the same example above, farmers will receive 10,000 taka at the beginning of the crop-cycle and will return 11,700 taka at the end of the cycle (at the end of 5 months or 6).
T3: Sequential Crop Credit: One can argue that farmers don’t need the entire amount immediately, as a result, giving the entire amount at one instance might provoke consumption smoothing or multiple use of credit (both on productive and non-productive purpose) especially if the individual is present biased. To check if this is true, in this treatment arm, farmers will get sequential credit, for example, 40% during the first month of the production process, then 15% during next four months until harvesting finishes. Please note that the above mentioned limit of the fund is the maximum cap for the production process and anything that is left in the earlier production stage could be used in the next production stage (for example, if a farmer uses only 10 percent for the first month, he is eligible to use 45% for the next month and so on).
T4: Sequential In Kind Credit: It has been highlighted in the literature that the adoption rates of modern agricultural inputs such as hybrid seed and chemical fertilizers are very low in developing countries. Despite the presence of large-scale public interventions that encourage farmers to use such technologies and boost agricultural productivity, their proliferation has been slow and incomplete; hence, agricultural productivity has been stagnant for several decades. One recent development in this stream of literature is on how to induce farmers to use modern techniques, for example Kremer and Duflo (AER 2012) tried to “Nudge” farmers of Kenya to use chemical fertilizers whereas Matsumoto at al. (2014) tried to incentivized farmers of Uganda to use modern inputs by providing it for free. In this treatment arm, we will try similar intervention through a different channel. In the credit component we will bind farmers to use/buy inputs (high yield variety seeds and modern chemical fertilizers) within the recommended choices given by our counterpart NGO.
Our implementing partner NGO will consult with the agriculture department and extension officers about the appropriate inputs, which with a combination of monetary amount for irrigation, plowing and weeding might be helpful for the farmer to get better yield per decimal to repay the loan on time. The service cost for providing this additional task for our NGO (taking farmers to the market to make sure they purchase the input based on the recommendation and within the credit limit) could be calculated in the total credit provided (say 10%) to the farmers. Many agricultural interventions by developing countries are mainly providing “information” of modern farming inputs. Since most of the farmers in developing countries are uneducated and not aware of such modern methods of agriculture, the adaptation rate of modern inputs are strictly minimum. Moreover, farmers are risk averse (we can test this assumption by using behavior game) and would like to follow the tradition method of farming. This intervention might shed some light on how we can incentivize farmers to use better technology and modern farming methods.