Ever since the seminal paper of Coase (1937), economists have known that transactions costs can hinder the efficiency of exchange. If transaction costs are present, some mutually beneficent transactions may not take place.
In developing countries, poor institutions mean that many such transactions are left on the table since transaction costs are too high. For example, property rights are often vaguely defined and contracts hard to enforce legally. Well-functioning institutions support well-functioning markets through low transaction costs. In these contexts, there is increasing evidence that households instead rely on their social networks. One example is how social ties play an important role in informal insurance schemes (Fafchamps and Lund 2003, Mazzocco and Saini 2012). In fact, the importance of social ties shows up in very diverse contexts where they can help to decrease transaction costs; other examples from the economics literature are in the selection of an international trading partner (Granovetter 1995, Topa 2001) and in labour markets where seeking and getting a job is affected by social networks (Rauch 2001, Chaney 2014).
There are good reasons to think that social networks can also reduce barriers to the exchange of production factors. Social connections can increase trust between individuals and important information can be exchanged. Social ties can also reduce the risk of violation of agreements, since the violator risks losing not only the contract but also the social connection. These are some of the ways in which social ties are thought to lower transaction costs. However, the extent to which social ties can offset the negative impacts of high transaction costs for exchange of production factors is an open research question. Earlier papers provide indirect evidence that this may be the case (Sadoulet et al. 1997, Holden and Ghebru 2005, Macours et al. 2010).
In a recent paper (Beck et al. 2015), we add to this ongoing discussion by directly testing the impact of social ties on efficiency of factor exchange. We use a unique network dataset from 51 rural villages in Gambia, the Gambia Networks data 2009 (Jaimovich 2015). We use this data to answer the following questions:
Do transfers of land between rural households in Gambia improve efficiency?
And do social networks help or hinder efficiency-enhancing transfers?
We consider three types of social networks: family ties, ethnic groups, and geographical neighbourhoods.
Efficiency in rural Gambia
Our paper focuses on the allocation of production factors among small-scale agricultural farmers in Gambia. Agriculture in Gambia is characterised by low levels of mechanisation, and the most important inputs in production are land and labour. We focus most of our efforts on understanding transfers of land, but we also consider labour transfers. Land is typically lent or rented out on an annual basis at the beginning of the agricultural year. In the villages of our study, land endowments are highly unequal. However, we observe a dense system of land transfers within the villages. We are concerned with how transfers of land change the allocation of land from the initial allocation of endowments to the post-transfer allocation of land used among farmers.
Figure 1 shows the family network in a single village. It is apparent that the network of family ties is quite dense. However, while the households at the centre of the network have connections to many households, the periphery of Figure 1 contains some more marginalised households with none or only a few links.
Figure 2 shows the network of land transfers in the same village. In this village, two households are senders of most of the land transfers that take place in the village. If one overlays Figure 2 onto Figure 1, it would become apparent that quite a few – but not all – land transfers take place where there is an existing family connection, as one would expect if family connections lower transaction costs and increase the potential for efficient exchange.
Figure 1. The family network in a single village
Note: Each dot represents a household, and each link represents a family connection.
Figure 2. The land network in a single village
Note: Each dot represents a household, and each arrow represents a land transfer to the household which the arrow is pointing to.
We are concerned with the notion of allocative efficiency. Allocative efficiency is achieved if production factors (namely, land and labour) are distributed such that overall production cannot be increased by shifting factors of production across households.
We begin by deriving efficient allocations of factors across farmers for different types of production functions as well as with and without heterogeneity in the production functions across households. Some configurations of production functions are rejected trivially by the data, so we do not consider them further. An important and empirically relevant subset of configurations provides testable predictions regarding the direction in which land should flow, if land exchanges are efficiency enhancing.
We further exploit the network-level data available to investigate the efficiency properties of the pre-existing social networks. The crux of our empirical method is a quantification of the degree to which land exchanges – and the subset of land exchanges which overlaps with the social networks – improve allocative efficiency. This quantification allows us to estimate the impact of social networks on the likelihood of efficiency-enhancing transfers taking place.
Networks matter – but efficiency is not achieved
Our results support the hypothesis that land transfers improve allocative efficiency. However, we reject that efficiency is achieved.
Contrary to what we expected, we do not find that efficiency-enhancing land transfers are more likely to happen within ethnic groups and between family-related households on average.
In fact, we find some evidence of the opposite result. However, when we investigate further, we find that this result is driven by land transfers from a small number of households who have large land holdings.
Once we mitigate the impact of these outliers on the results, we find that land does flow between family-connected and ethnically related households such that a more efficient allocation is achieved, compared to what is achieved between unconnected households.
We also investigate the properties of transfers of the other important production factor, namely labour. There is no evidence that labour transfers improve efficiency – either between or within social networks. It is likely that transfers of labour serve other purposes, such as to deal with the consequences of unanticipated shocks during the agricultural cycle.
Can social ties offset the negative impact of a limited or non-existing institutional framework? Our answer is a tentative yes.
Interestingly, we find that some of the largest landowners do not perform efficiency-enhancing transfers, but once these are controlled for, the average transfer is efficiency-enhancing. While this can be interpreted as discouraging evidence as to the potential of social networks to offset the limits imposed on transfers by the institutional environment, it is not all bad. One interpretation of our results is that while smaller landowners do use social networks to offset the negative impacts of transaction costs, the largest landowners are not inhibited by the costs of transferring land outside the social networks. In this sense, it is encouraging that the socially marginalised households of the village – those who are not kin-related to or shares ethnicity with the large landowners – are not marginalised when it comes to receiving land.
Our analysis does not attribute any motives as to why households transfer land and labour in the ways that they do. Our method is only concerned with the properties of the factor exchange which occurs in these villages. This column is not concerned with the mechanisms driving the exchanges of land. However, there is some evidence that for the rural farmers in Gambia, factor allocations are at least partly motivated by altruistic behaviour (Beck and Bjerge 2015).
Beck, U and B Bjerge (2015), “Insuring the Poor: Inter-household Land Transfers and the Importance of Land Abundance and Ethnicity in The Gambia”, Working paper.
Beck, U, B Bjerge, and M Fafchamps (2015), “Social Ties and the Efficiency of Factor Transfers”, CEPR working paper no. 10957.
Coase, R H.(1937), “The nature of the firm”, Economica, 4(16):386–405.
Fafchamps, M and S Lund (2003), “Risk-sharing networks in rural Philippines”, Journal of Development Economics, 71:261–287.
Granovetter, M (1995), Getting a job: A study of contacts and careers, University of Chicago Press.
Holden, S and H Ghebru (2005), “Kinship transaction costs and land rental market participation”, Department of Economics and Resource Management, Norwegian University of Life Sciences.
Jaimovich, D (2015), “Missing links, missing markets: Evidence of the transformation process in the economic networks of Gambian villages”, World Development, 66:645–664.
Macours, K, A dJanvry, and E Sadoulet (2010), “Insecurity of property rights and social matching in the tenancy market”, European Economic Review, 54(7):880–899.
Mazzocco, M and S Saini (2012), “Testing efficiency risk sharing with heterogeneous risk preferences”, American Economic Review, 1:428–468.
Rauch, J E (2001), “Business and social networks in international trade”, Journal of Economic Literature, 39(4):pp. 1177–1203.
Sadoulet, E, A de Janvry, and S Fukui (1997), “The meaning of kinship in sharecropping contracts”, American Journal of Agricultural Economics, 79(2):394–407.
Topa, G (2001), “Social interactions, local spillovers and unemployment”, The Review of Economic Studies, 68(2):pp. 261–295.