Consider the following panel model to examine the effect of retirement on consumption expenditure, consit, of individual i over years t=1,…,3:
Consider the following panel model to examine the effect of retirement on consumption expenditure, consit, of individual i over years t=1,…,3:
(B1) log(consit) = β0 + β1retiredit + β2ageit + β3marriedit + β4healthit + δ1Yr2t + δ2Yr3t + ai + uit
Where:
- retiredit is a dummy variable equal to 1 if individual i is retired on year t and 0 otherwise
- ageit is the individual's age in years
- marriedit is an indicator variable for whether the individual is married (1) or not (0) in year t
- healthit is an indicator variable equal to 1 if the individual is in 'good health' and 0 otherwise
- Yr2 is a dummy variable equal to 1 in year t=2 and 0 otherwise
- Yr3 is a dummy variable equal to 1 in year t=3 and 0 otherwise
We obtain the following results when we estimate Model (B1) using RE and FE methods:
| Variable | Random Effects | Fixed Effects |
| retired |
-0.072*** (0.024) |
-0.054** (0.025) |
| age |
-0.008*** (0.003) |
- |
| married |
0.031*** (0.010) |
0.028 (0.022) |
| health |
0.050*** (0.021) |
0.052*** (0.022) |
| Yr2 |
0.111*** (0.022) |
0.113*** (0.022) |
| Yr3 |
0.164*** (0.022) |
0.167*** (0.022) |
| Constant |
1.048*** (0.215) |
1.054*** (0.255) |
|
R-sq Number of Obs |
0.207 3124 |
0.184 3124 |
Note: Clustered standard errors in () bellow the coefficient estimates.
***p-value <0.01, **p-value<0.05, *p-value<0.10
[i] Interpret the estimated coefficient on the Yr3 dummy variable in the RE model.
[ii] Notice that the age coefficient is not estimated when we use FE. Explain why this is the case.
[iii] For Model (b1), which estimator do you prefer, RE or FE? Explain your reasoning.
Step by step
Solved in 2 steps
