FOOD security is widely assumed to be a rural problem that will be resolved through technical innovation among smallholders (in the guise of a "green revolution"). What seems to be forgotten in this romantic view of the African rural household is that its food security is not simply, or even mainly, a function of what it does or does not produce itself.

R ural households buy some or most of their food with cash from family members who have moved within the country or across borders to earn income.

Another assumption is that food security in urban areas is about promoting urban agriculture. This may be well intentioned but it derives from the misleading idea that increased food production is the key to urban food security. The main cause of food insecurity in African cities is the lack of access to food and that means the absence of a regular and reliable income with which to buy it.

A third problematic assumption is that the rural and urban areas are separate spheres with a deep divide between them. This is at odds with the observable web of connections that bind them together.

There is regular circulation of people, goods and money between town and countryside. This means it is impossible to fully explain the state of food security of urban households without reference to their rural counterparts, and vice versa.

With migration within and to S outhern Africa having changed dramatically, in recent decades, most poor households in cities either consist entirely of migrants or a mix of migrants and locals. Rapid urbanisation, increased circulation and growing cross-border migration have all meant that the number of migrants in cities has grown exponentially.

The global food security agenda pays little if any attention to migration and urbanisation. The reasons are hard to understand since the connections between migration and food security seem obvious. They cannot be properly understood and addressed independently.

The African Food Security Urban Network conducted a survey in 11 cities in nine Southern African Development Community countries in 2008 and looked at whether migrants are more food insecure than longer-term residents of the poorer areas of Southern African cities.

Across the sample, unemployment rates were high, with nearly half of migrant and non-migrant households receiving no regular wages. This also suggests migrants do not find it any harder to find jobs than permanent residents in the city. Migrant households do, however, find it easier to derive income from casual work while local households were more involved in running informal and formal businesses. Very few households in either category earn any income from the sale of homegrown agricultural produce.

The differences between migrant and non-migrant households are relatively significant. While there are many poor and food-insecure households in both camps, there are more food-secure households in the non-migrant group. One of the reasons may be that migrants remit a portion of their income to households that are in even greater need of the cash.

Research shows that cash remittances are a crucial source of income in Southern Africa, with 74% of all migrant-sending households receiving remittances.

The vast majority of households (93%) buy food and groceries with this income. No other expenditure category comes close, although a significant minority pay for transportation, clothing, education and medical expenses. A mere 15% spend income on agricultural inputs.

The implications of the region's new mobility regime for food security need further exploration and analysis. To what degree is heightened mobility related to problems of food insecurity? Food security shocks and chronic food insecurity can certainly be major motives for migration for income-generating opportunities. War and conflict have led to the displacement of millions, and the collapse of Zimbabwe's economy pushed hundreds of thousands of people out of the country.

Africa faces an increasingly urban future and the feeding of African cities is a growing development challenge. Urbanisation and migration therefore urgently need to be "mainstreamed" into the food security agenda. Without such an effort, food security policies will be undermined by a failure to understand and manage the crucial reciprocal relationship between migration and food security.

. Crush is the CIGI chair in global migration and development at the Balsillie School of International Affairs and honorary professor at the University of Cape Town.