resources (insurance provision). Yet, what is not clear from the literature is to what extent the
beneficial effects of migration strategy in protecting household members left behind from
economic pitfalls are able to improve the productive-investment capacity in sending rural
households6.
Few empirical studies have shown that earnings of international migrants have a positive
impact on crop productivity and may also serve as a source of capital accumulation in rural
households (Rozelle et al., 1999; and Lucas, 1987). Yet, little evidence exists, in a
simultaneous framework, on the role migration costs (or the shadow value of remittances)
play in shaping the potential impact of migration at origin.
A large body of migration literature has focused on remittances as the main way households
are able to smooth their consumption and overcome liquidity or risk constraints for
investment purposes. Indeed, the contribution of remittances - defined as the money and/or
goods sent home by migrant workers - to the income level of sending households has been
typically considered the key variable to assess the impact of migration on economic
development. It is fairly well known that for many developing countries, remittances are an
important source of income7. Moreover, remittances are now largely recognised as part of an
informal familial arrangement that goes well beyond altruism and entails “exchange motives”
(Lucas and Stark 1985, Cox et al. 1998)8.
However, the linkage between the level of remittances and the development of sending
households is not straightforward for three orders of reasons. Firstly, the sign and dimension
of their economic effects depend on a host of intervening (and often conflicting) variables
such as informational and financial costs of migration and the opportunity cost to move (in
terms of forgone human and physical capital)9. In second place, a critical point in order to
6 In other words it is not clear whether remittances sent back by migrants are able to compensate for the
opportunity cost of allocating a marginal unit of family time to migration, that is the loss of net income from
production. Household may not be able to simultaneously devote time to migrant labour and to investment
activities in home areas. Moreover, it has been argued that human and physical capital embodied in (‘certain
types’ of) migration is likely to complement other family resources in production, strengthening the negative
effect from less family labour (i.e. “brain drain” argument. See Faini, R. 2003). Another argument in this
direction provided by the literature is the one of moral hazard phenomena in sending households: if migrant
work is lucrative enough household members remaining behind may entirely forgo productive activities and live
primarily on remittances receipts. For evidence on this see Gubert, F 2000. On the other hand, though, people
left behind may invest more so as to motivate the migrant to send more remittances (de Janvry et al., 1997).
7 E.g. South Africa gold workers to neighbouring countries; Mexican migrants in the US; unskilled South Asia
(e.g. India, Pakistan, Bangladesh) migrants in the Gulf.
8 According to the “exchange hypothesis” remittances must be seen as repayments for services provided by
parent household such as childcare, education, bequests and inheritance, coinsurance, social standing. The
exchange motive can be further divided into insurance motives (to spread risk across a broader portfolio) and
investment motives (to build up household assets to be inherited later, or to repay previous investments that
allow the migrant to retain the right to inherit).
9 Differently said, the same expected value of remittances may not have the same effect on the probability to
migrate for households at different points in the wealth distribution.