land area and cropping pattern. The basis for the profiles is simplified farm budget data2, and the
 models are static, not behavioral; i.e., they do not allow farmers to change production and
 consumption patterns in response to price changes. Only five crops are included, and maize is
 the only staple food crop. In reality, cropping patterns would be more complex. In spite of the
 simplification, the profiles allow us to identify short-term losers and gainers.

 Table 2.1 shows characteristics of the five households, the relative importance of these
 household types in the rural economy, their cropping patterns and how they fared over the
 five-year period from 1990/91 to 1994/95.

 Overall, families who have access to purchased inputs and hence can plant tobacco and '
 hybrid maize have done well with liberalization. Their income has increased, both in current and '
 in real terms, to such an extent that they can acquire enough savings to withstand a drought year.
 These families often have above average farm size. The household which grows cotton is an
 intermediate case. Until 1994/95, nominal income was stagnant. In the last year, however, the
 cotton market was opened up to competition and prices rose substantially. This household would
 normally be food secure, but, until last year, it would not have sufficient reserves for a bad year.
 However, if cotton prices stay high, this situation may change.

 .Households which are maize deficient and have no resources to invest in improved maize
 varieties have lost out due to rising market prices, although, as they have very little marketed
 output, the size of this loss is limited. These households tend to have below average size. One
 type of household which is not represented in the models above is the ery or family who has
 virtually no interaction with the market. There are indications that the poorest rural households
S )) in Malawi, whose number is difficult to estimate, exist in a world of their own production and
 bartL They perform ganyu largely for maize, and they barter any spare vegetables or pulses for
 food. Market liberalization has almost certainly had very little impact on them.

 If these findings are representative, they raise important issues about how(j-oruira 4 at
 households can be given some ability to participate in markets on better terms than they are doing V. Pa 7
 now. Unless they can develop the ability to buy inputsr aise cash crops, and otherwise improve
 their productivity, they are fated to a life of increased marginalisation if they continue to depend
 on agriculture for their major income source.

 The main hope for these households to break away from poverty and become food secure
 is through employment or other income generating activities. The principal form of employment
 in the rural area is ganyu. Other income generating activities will primarily come through off-
 farm activities such as petty trading, food processing and catering. Poor households' ability to
 find employment or engage in inconrenerating activities depends on two factors: general growth



 2 Details on methodology and data sources are given in Annex C.