Table 4: Wage Dynamics Within Firmsa
Comparative Advantage
|
Specificationb 1 |
Low Rank L |
Middle Rank M |
Upper Rank U |
EXecutive |
|
Rank Dummies |
dL |
dM |
dU |
dEX |
|
Skill*Ranks |
- |
0.208*** |
0.499*** (0.044) |
0.695*** (0.088) |
|
Unmeasured |
cL |
cM |
cU |
cEX |
|
1 |
0.932*** (0.080) |
1.261*** (0.131) |
1.467*** | |
|
Measured |
βL |
βM |
βU |
βEX |
|
0.813*** |
1.154*** |
1.087*** |
1.159*** | |
|
Ratio |
(0.220) (0.008) |
(0.194) |
(0.233) |
(0.328) |
|
Testsc for Equality |
Joint |
M=L |
U=M |
EX=U |
|
of Slopes cj |
9.36 (0.02) |
0.58 (0.44) |
5.11 (0.02) |
0.89 (0.34) |
|
of Slopes βj of Ratio b0 = 1 Overidentification Test |
9.71 (0.02) |
9.69 (0.00) |
0.27 (0.60) |
0.11 (0.73) |
Comparative Advantage and Learning
|
Specificationb 2 |
Low Rank L |
Middle Rank M |
Upper Rank U |
EXecutive Rank EX |
|
Ranks |
dL |
dM |
dU |
dEX |
|
- |
0.065 |
0.008 |
0.11 | |
|
(0.083) |
(0.233) |
(0.17) | ||
|
Skill*Ranks | ||||
|
Unobserved |
cL |
cM |
cU |
cEX |
|
1 |
1.119*** |
1.556*** |
1.793*** | |
|
(0.254) |
(0.384) |
(0.468) | ||
|
Measured |
βL |
βM |
βU |
βEX |
|
0.908** |
0.716** |
0.890*** |
0.770 | |
|
(0.468) |
(0.310) |
(0.313) |
(0.526) | |
|
Ratio |
b0 | |||
|
b0 |
1.017*** | |||
|
(0.006) | ||||
|
Testsc for Equality |
Joint |
M=L |
U=M |
EX=U |
|
of Slopes cj |
3.03 (0.38) |
0.22 (0.63) |
1.22 (0.27) |
0.57 (0.44) |
|
of Slopes βj |
0.62 (0.89) |
0.37 (0.54) |
0.17 (0.68) |
0.08 (0.78) |
|
of Ratio b0 = 1 |
2.00 (0.15) | |||
|
Overidentification Test |
56.16 (0.25) |
a-Dependent variable is wage in level in thousands of marks. Also included are dummies
for the type of contract, large firm size, public sector, occupations, industries and years.
b-Estimation of Ω using the residuals from NLIV with Ω = I in a first step. Number of
observations is 11159 in the comparative advantage case and 9891 in the learning case.
The instruments for previous wage in the comparative advantage case include all the rank
interactions between t and t-1 as well skills interacted with rank. In the learning case,
current rank affiliation is instrumented using rank affiliation in t-2.
32