The Formal Divide: Customary Rights and the Allocation of Credit to Agriculture in Tanzania

ABSTRACT

It is generally held that one mechanism to enable inclusive growth in Tanzania is enabling farmers to access credit to raise productivity and incomes. The formalisation of property rights in Tanzania is being undertaken by a multiplicity of actors at great expense to donors, individuals and the government. While there have been a variety of different justifications for allocating Certificates of Customary Rights of Occupancy (CCROs) to farmers in Tanzania, perhaps the most prominent argument is that it will enable farmers to finally overcome the divide between ‘informal’ customary rights and the formal banking sector. CCROs would provide the collateral that would induce banks to lend money to small-scale farmers. As part of a six-year investigation in Manyara, Mbeya and Dodoma regions, our research team evaluated the impact of formalisation on farmers’ access to credit. The paper will present the results while pointing to the continuing institutional and market imperfections that perpetuate the formal divide.

Acknowledgements

The earlier version of this paper was prepared for (though not presented) at REPOA’s 18th Annual Research Workshop (3–4 April 2013) The Quest For Inclusive Growth and presented at a seminar at REPOA in July 2013.

We would like to acknowledge generous support from the National Science Foundation, Grant No. 0921757, ‘Transformations in Property Rights and Poverty in Rural Tanzania,’ and the Royal Danish Embassy, Dar es Salaam, Tanzania. And also thank the following individuals: Kate Owens, Lydia Nyeme, Menan Jangu,Vedast Makota, Chambi Chachage, John Ulumara, Mathew Bukhi, Thabit Jacob, Faida Zacharia, Elineema Ezekiel, Salome Swai, Shymala Nagaraj, Emmanuel Sulle and two anonymous referees.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. ‘If the income is to be raised for enabling the vast majority of the population to enjoy a more decent standard of living, there seems to be immediately no alternative to increasing the productivity of agriculture. It will be thoroughly unrealistic to expect the farmers, the vast majority of whom are on or below the poverty line to have enough savings of their own to finance the capital investments. They have, therefore, to be enabled to make these investments by giving them access to the necessary credit facilities’ (Bank of Tanzania, 1979 qtd. in Haidari, Amani, Msambichaka, Hedlund, & Lundahl., 1987).

2. Wanyonge, meaning ‘the weak’ or ‘the oppressed’, features in the full Swahili name of this programme but not in the English version or the Swahili acronym ‘because Mkapa wanted it that way. He wanted to place local emphasis on the wanyonge’ (Interview with MKURABITA Director of Finance and Administration, Dar es Salaam, 7 July 2010).

3. This was their stated objective on their website through 2010. It has been adjusted now slightly and states ‘empower owners of the formalised asset to use them in accessing financial capital (credit) and other socio-economic benefits found in the modernizing formal economy of the country’. The reasons for the adjustment arose out of the slow recognition that CCROs were not being directly used as collateral. As we will see, however, this has not stopped MKURABITA telling villagers that they could take their certificates, use it for collateral and go to the bank for loans. See http://www.mkurabita.go.tz/index.php?&chooselang=1. Information on the collateral issue is from interviews at MKURABITA in May 2012.

4. In an interview at MKURABITA in June 2010 a senior director told us that they focused on the rural areas because they saw the disadvantages that rural areas have regarding growth requirements. Bankers [were] not ready to offer them loans. A key challenge was ‘on getting bank institutions… to honor what is owned in the rural areas’.

5. The document justifying the new DFID/G8 land formalisation programme argues: ‘Titling may also ease credit constraints to investment by allowing households to borrow against land titles’ (Locke, Henley, & Nshala, 2014)

6. See the Online Appendix for a discussion of our methodology.

7. Fernqvist (2012) interviewed 72 members in four villages involved with the first phase of the World Bank titling project. 69 per cent of the people responding indicated that they remember being promised loans by the group representing the WB/Ministry of Lands project during the village assembly which was convened as part of the titling awareness campaign. Other benefits like security and the reduction of conflict were mentioned but were remembered with less frequency by interviewees.

8. An estimated 43.8 per cent of the population in Tanzania lived under the international poverty line in 2011/12. These people are likely to be too poor to save anything (World Bank, 2015).

9. Interview with villagers in an Mbozi district village, May 2012.

10. Interview with representatives of the village council, May 2010.

11. It is important to emphasise that people typically do not self-select into the group of CCRO holders since virtually all titling is initiated exogenously by external agencies. Therefore, we are confident that the variable for having a CCRO is not endogenous. See the Online Appendix for details about theoretical discussion and robustness tests investigating this further.

12. The independent variables used in this regression are not endogenous with the indicator for having a loan. As discussed in footnote 11, self-selection into the group of CCRO holders is not a relevant problem. Our interviews showed that selection into the titling process is exogenously determined by NGOs and government departments. This is the primary evidence against endogeneity, since endogeneity cannot be ‘proved’ but rather must be carefully argued on a theoretical basis (in absence of treatment randomisation). In addition, all correlations between all independent variables (including the indicator for having a CCRO) and the residuals of the regression are between 0.1 and −0.1, except for one correlation, which is 0.11. These correlations are too weak to be of concern. This suggests that endogeneity is not an issue, since a variable is endogenous if it is correlated with the error term (and residuals are estimations of the error in a model), and our variables are not correlated with the residuals.

13. To be clear, this is not a model of loan supply or loan demand. Because we do not have data on interest rates for the different types of credit used in our sample areas, we cannot model supply or demand curves for loans. However, because credit rationing is evident in our descriptive statistics (among interviewees who did not have a loan or a CCRO – about 76% of the sample – roughly 40% said they would use a CCRO to obtain a loan), this model could be considered a loan supply equation if interest rate data were available. It is clear from our interviews that there is a need for credit among rural farmers, and that financial institutions only provide a limited amount to select groups.

14. In the following regressions a logit regression model is used, as is standard when working with a binary dependent variable. We ran several robustness tests to be sure of the quality of our model. We also identified any outliers that may have influenced the results and re-ran the regression with those observations omitted. The results were not substantively different. We also tested for multicollinearity and found no tolerances below 0.1, which is the standard benchmark for a tolerance worthy of investigation. Other robustness tests were ran, including testing for theoretical issues regarding where CCROs were distributed, and the results did not differ substantively or significantly from our original model.

15. Plans are afoot in Tanzania to launch an agricultural bank to specifically serve farmers, which will complement the newly launched womens’ bank. It remains to be seen whether these new financial institutions will address the problems of credit access for their rural constituencies.

16. The Bank of Tanzania (BOT) established two guarantee schemes for commercial bank loans in 2003–04 with about 80 per cent loans in agriculture. One focused on export guarantees with a total capital base of 55 billion shilling. The minimum loan is 500 million Shs with no maximum limit. A second fund is for small and medium-size enterprises and has a capital base, which is much smaller at Shs 6 billion. They finance requests of between 5 million and 500 million shillings. They supported many types of activities including SACCOs, warehouse receipt credit systems, cooperative societies, and so forth. They were being phased out by the central bank monitoring board arguing that central banks should not be encouraging risky loans when their main function is to support prudential regulation and financial stability in banks. One of the officers admitted the real pressure was probably from the IMF and World Bank and lamented. ‘The Bank of Tanzania needs to be more proactive [not less] in giving credit to poor people in villages’ (Interviews Credit Guarantee Officers, BOT, February 2013).

17. Interview, CRDB, Vwawa, May 2012.