Point 1: One of the motivations for this study is the striking low rate of urbanization in Ethiopia relative to the rest of Sub-Saharan Africa. Urbanization rate is 17%<36% in SSA (WDI).
Point 1: Agglomeration economies--Having people concentrated in one area can reduce costs of production (e.g., transport costs), and benefit industry since the locals bring up the demand for products.
Migration can free up resources for household members left behind. Migrants also have access to additional labor markets so it could potentially increase employment and auxiliary income for the household through remittances. Migration used to cope with risk. Remittances back to the household creates the flow of capital from cities to rural areas which can be used towards investment in human and physical capital encouraging growth.
Point 1: Few studies. Studies that are out there focus on one site and/or are very qualitative.
Point 2: We wanted to focus on economic reasons. We allowed only people that had initially left for schooling but now were reported to be working elsewhere. It was important to focus on household head relatives because relatives are more likely to have closer links to the household and send remittances. Also, we learned early on in the piloting that it was rather unlikely that households knew where extended family members or farm hands were located.
Here are the 18 kebeles. The colors represent the proportion of households that have at least one tracked migrant, where the burgundy shaded woredas indicate the villages had over 25 percent of the households with at least one tracked migrant.
--This map shows the travel time to the nearest city of 50,000 people for each site.--No clear effect between migration and travel time/roads. Areas with high migration densities don’t necessarily have low travel time (e.g. Shumsha village in Buginaworeda is located in a woreda which is further than 3 hours from a city of 50,000 or more).--Other factors may also explain decision to migrate.--Travel time may explain where people go.
Destinations are diverse spanning 106 woredas across 10 regions. From this map, you can already tell that much of the migration is rural-to-rural which implies that migration is not causing much urbanization.
We computed migration rates by destination type using four definitions: destination is in a woreda with town of 50,000 inhabitants, 50% woreda is agglomerated, and 20% of woreda is agglomerated, and any % of the woreda is agglomerated. The agglomeration index accounts for travel time of 1 hour or less to a city of 50,000 people and a population density of 150 people per square kilometer (refer to Schmidt and Kedir, 2009 for further details). The fifth panel computes migration rates using the Labor Force Survey 2004-5 for comparison. The rates computed with the definition of urban as 20% of the woreda is agglomerated resembles most calculations from the LFS 2004-5 using its own definition of urban, noting it’s also nationally representative. For the remainder of the analysis, we use the 50,000 inhabitant and 20% agglomerated definitions of urban.
Here is a map of the major routes traveled from ERHS sites in our sample. You will see that there are several routes that consist of local moves (noted by the very short arrows which you cannot see but are noted by the highlighted destination woredas. , and most of the major moves are to other rural places.
Only showing characteristics that are significantly different.Point 1: Heads tend to be more involved outside of agriculture.Point 2: There are ethnic and religious differences. Fewer muslims migrate, e.g.Point 3: Migrant households may have surplus labor, as they have a greater number of males of prime working age and also have less land. Less wealthy.Point 4: Migrant households have greater access to networks outside of the village.
Wage differential: The wages faced at the potential destination and the value of marginal product of labor in the household. These are hard to measure. So, we include household demographic and education level of the household head in the regressions to control for this in the regressions.Migrant networks: Reduce costs associated with migration such as uncertainty of employment and other moving costs. Households with better social networks outside of the village are probably more inclined to migrate.
These results combined with what was found in comparisons of migrant and non-migrant households in the ERHS suggest land scarcity is more of an issue these days than land rights.
We time the ERHS household shocks with the year the migrant moved. We look at the number of migrants that moved when there household reported one of these major shocks, and those that moved the year after the households reported their specific shock. Moves coincide with food price rises and droughts.
Migrant households appear to be more vulnerable to the drought in 2000 (EC), and also more likely to have experienced a death or illness in the family in the last five years.
Drought risk effect becomes significant when you substitute the household self-reported measure for the proportion of households within a neighborhood reporting that was a shock they experienced. That is because, the covariate nature of the shock may be more detrimental to livelihoods and income potential than idiosyncratic shocks. There are mechanisms in place like iddirs to support idiosyncratic shocks but less to support covariate shocks, like insurance mechanisms.
If there are benefits to migration, the increased income through remittances is likely not the channel which produce the se benefits.
Very few differences in the changes across groups. Evidence migrants might be worse off at least they had more months satisfying food needs over time, and ate cereals more infrequently in 2009.
Migrants better off in terms of eating more meat and animal products. They do eat cereals fewer times than their source households.
Migrants eating more of all food categories than they would have in the ERHS household in 2004-5.
Very few significant differences by migration status. When, we see a difference in a few measures and the non-migrants households are always happier compared to the non-migrant household.
Permanent Migration and Remittances in Ethiopia
Permanent Migration and Remittances in Ethiopia<br />June 24, 2010<br />Alan de Brauw<br />Lisa Moorman<br />Valerie Mueller<br />Tassew Woldehanna<br />1<br />
Some urbanization is good<br />Geographic concentration of labor conducive for agglomeration economies and growth<br />More effective to provide services in areas concentrated with people (e.g., sanitation, health, electricity, infrastructure) <br />More cost-effective<br />Easier for citizens to access<br />Migration into urban areas may be too low, and we want to understand why<br />3<br />
Relevant policies<br />National Population Policy (1993)<br /> discourages rural urban migration because excessive it can create urban slams, violence and exacerbate crimes <br />It is only after the 2005, the GoE seems to understand rural-urban migration is key for development<br />E.g. for the time, Tigray Regional state planned to facilitate rural urban migration<br />Change of mind <br />The most urbanized region in Ethiopia <br />4<br />
Registration process inhibits RU migration<br />You need to stay 6 months before you register as urban dweller<br />This means you can get government support <br />You need to have an address, i.e. a house or be a member of a HH who owns a house <br />An ordinary migrant can not be successful in this <br />Social network is very important <br />5<br />
Migration benefits people<br />Releases resources for hh members <br />Creates additional employment opportunities<br />Additional income available to hh (remittances)<br />Risk coping<br />Investments in human and physical capital <br />6<br />
Research Gaps<br />Very few migration studies on Ethiopia<br />Migration behavior not well understood<br />Data to analyze migration is often incomplete <br />Studies that focus on source hh reports exclude migrant destination information<br />Studies that focus on migrants lose information on comparable non-migrants and household members left behind<br />We have a matched migrant sample which allows us to examine the benefits realized by the migrant and the relatives he leaves behind<br />7<br />
Objectives<br />Migration patterns out of 18 kebeles<br />Matching migrant and hh panel to examine<br />Determinants of migration by type<br />Migration benefits<br />Experienced by migrant<br />Experienced by migrant households<br />8<br />
Matched Sample<br />Ethiopian Rural Household Survey (ERHS) <br />Focus on 2004-5 and 2009 rounds (18 kebeles)<br />Tracking survey follows migrants from 2004-5<br />Older than 10 years<br />Moved for employment, schooling (now work), loss of land, resettlement program, and to follow family<br />Relative of household head<br />9<br />
Rural-to-rural migration most common<br />Partially explains why urbanization is low<br />Has implications on determinants of migration<br />Migrants will likely move to local places<br />Where they have connections<br />Cost of move is cheaper<br />Opportunity cost of move to household lower if remain close<br />14<br />
Empirical Model<br />23<br />X: female headship, age, occupation, literacy, ethnicity, support network outside of village, livestock, land, male and female labor endowment<br />Shock: self-reported drought 2000, death or illness in last five years<br />
Migration Probability Results<br />24<br />RR: rural-rural RU: rural-urban<br />1: Defines urban as woreda with >50,000 people<br />2: Defines urban according to 20% agglomeration index<br />
Welfare Implications of Migration<br />Channels of welfare benefits<br />Auxiliary income through remittances<br />Additional resources from migrant’s absence <br />Consumption changes<br />Source households<br />Individual Migrants<br />Changes in subjective well-being<br />Heads of source households<br />25<br />
Remittances<br />Generally remittance rate of internal migrants is lower than rate of international migrants<br />Africa tends to have lowest rates<br />China (2000): 66.4 percent<br />El Salvador (2008): 70.8 percent<br />South Africa (1993): 29.7 percent<br />Ethiopia (2009): 33 percent<br />Conditional on sending remittances, average sent to hh in 2009 was 716 Birr (6.7 % per capita GDP)<br />26<br />
27<br />Household Food Scarcity by Migration Status over Time (ERHS)<br />
Comparing Consumption of Migrants versus Source Households (2009)<br />28<br />
Comparing Consumption Changes of Migrants<br />29<br />
Changes in Subjective Well-being<br />30<br />
Discussion<br />Migration low and predominantly rural-rural<br />agglomeration economies and future growth?<br />provision and expansion of public services to many Ethiopians will be cost-prohibitive<br />Migration is insurance related<br />Little land and shocks increase migration<br />Suggestive evidence that migrants may benefit from moving<br />Policies to reduce barriers of migration<br />Lack of remittances and lack of changes in source households welfare suggest members are ejected from household to relax constraints<br />31<br />