Comments on Migration and Economic Mobility in Tanzania

1
Comments on Migration and Economic Mobility in
Tanzania

• David McKenzie

2
Main message

• Migration is development large source of
  consumption growth for people in this area
• non-controversial here, as internal migration,
  but a controversial statement in international
  migration
• (see Lant Pritchetts work).

3
Descriptives convincing, and some good steps
towards identification

• Step 1 control for individual differences in
  levels of consumption (diff-in-diff).
• Thought experiment compare two individuals with
  same initial level of consumption (and same age,
  education, gender, etc.) see how much more
  consumption grows for the person who migrates.
• Concerns those who migrate still would have had
  different consumption growth without migration
  (e.g. high ability people would succeed in
  getting out of poverty at home too).

4
Step 2

• Control for (baseline) household fixed effects.
• Thought experiment is now take two people in the
  same household, which has now split up. Compare
  the consumption growth of the individual who
  moved to the individual who didnt move.
• Issue why did one person leave the household and
  not another? (and is this correlated with
  consumption growth).

5
Step 3

• Try and instrument for migration
• Pull distance to citymale aged 18 to 28
• (possible threats access to trade, local labor
  market, etc. affected by geography)
• Push Rainfall shockmale aged 18 to 28
• (surely this affects consumption, if males
  affected differently by rainfall due to different
  occupations, etc.)
• Social relationship whether head or spouse, age
  rank
• (but doesnt this affect how they would do in
  a non-migrating split household?).

6
Discussion

• Return Migration? How much is there? It is surely
  not surprising that after 10 years, the migrants
  who remain away have succeeded. Did those that
  did less well by migration return?
• Impact on remaining household members? Do the
  remaining individuals in Kagera do better if they
  have a household member move away from Kagera vs
  leave the household and stay in Kagera? (sign of
  effect less clear)
• Consumption at the individual level is per
  capita the right measure? Household composition
  changes with migration at least try per adult
  equivalent.
• Look at BMI take two malnourished individuals,
  does the one who migrated have higher BMI 10
  years later?