wealth for Bangladeshis households and the main asset (beyond labour) that allow people to
invest in widening opportunities26. Moreover, given that land market in Bangladesh is very thin
or even missing, it is reasonable to treat land ownership as exogenous27 28. Cattle owning, on
the other hand, is a form of saving or liquid asset, whose role is to cope with risk29.
Table 6 reports estimated coefficients and marginal effects for two logit model specifications
of the migration decision rule.
Interestingly, the land asset variable appears significant and negative in the first specification,
but significant and positive when squared (second specification). The same holds for cattle
endowment. This is to say that the negative effect of land and cattle endowment on migration
marginally increases as assets increases.
Being land a typical proxy for household wealth in Bangladesh, this result seems to suggest
that at a lower level of wealth, a small increase in assets tends to discourage households to
participate in the migration process, but the marginal propensity to migrate of better-off
wealthy households increases in assets endowment. In other words, the non-linear U-shaped
relationship between asset holding and migration seems to capture the role of fixed moving
costs that present a barrier to migration for households liquidity or risk-constrained.
26 Poor households typically own less land than the non-poor, and are highly represented among the near-
landless (i.e. those owning less than 0.05 acres) (WB 2003). Moreover, financial assets in the form of micro
credit have become more available to the rural poor in the recent decades, thanks to successful innovations
adopted by various non-governmental organizations in Bangladesh. However, despite tremendous progress in
this area, land is still considered the main form of collateral by micro credit financial institutions and formal-
sector lenders. See note 13 above.
27 Land is actually a great issue in Bangladesh because of its scarcity and because it is taken as a collateral in
credit programmes. On the exogeneity of land ownership in Bangladesh see the debate between Morduch and
Pitt (Pitt, 1988, Morduch, 1998, Pitt, 1999).
28 The exogeneity of land ownership challenges the potential inverse relationship between (past) migration and
(today’s) wealth.
29 It is commonly recognised that dependence on agriculture makes rural households more vulnerable to weather
and price fluctuations and causes income fluctuations. Given the limited access to formal credit markets and
social insurance, savings is an important tool for coping with income risk. Given the limited access to financial
institutions, the most available form of savings for rural households in developing countries is livestock (the
other forms of saving in kind are grain stock and land, but both are not easily realisable due to the tiny land
market and to the lack of stocking utilities, especially for smallholder), which serve as the major form of wealth
and as an insurance substitute, yielding a positive expected return and providing risk-diversification benefits
(Dercon, 1996). The existing literature on the risk-coping role of livestock generally find that the latter perform a
function as a liquid asset , which enables rural households to direct more inputs into high-return activities. Thus,
assuming credit constraints under which few farmers have access to credit, livestock play the same role NELM
attributes to the migration process, that is liquidity and risk alleviation at a household level, although livestock
may be a relatively cheaper investment than migration. In this sense owing livestock and participating in the
migration process might be viewed as substituted in their economic effect in enabling farm households to
overcome production constraints.
18