Table 6: Parameter estimates for selected variables
Variable |
Model I: 1992-2005 |
Model II: 1999-2005 |
Firm-specific factors | ||
a) Size category | ||
Less than 49 employees |
0.15* (25.25) |
0.12*(18.67) |
50-99 employees |
0.11*(19.49) |
0.09*(14.77) |
100-249 employees |
0.08*(14.34) |
0.07*(11.30) |
250-499 employees |
0.06*(9.50) |
0.04*(7.16) |
500-999 employees |
0.04*(5.73) |
0.03*(4.60) |
More than 1000 employees |
— |
— |
Production share in industry |
-0.07* (-2.2) |
-0.11*(-3.26) |
Number of owners working in the firm |
0.01*(12.79) |
0.01*(10.23) |
R&D intensity |
-0.14*(-6.9) | |
b) Outsourcing activities | ||
Quota of material inputs |
0.01*(22.06) |
0.01*(18.39) |
Quota of external contract work |
0.04*(16.82) |
0.04*(13.57) |
Quota of external services |
-0.02*(-3.3) |
-0.05*(-6.36) |
Quota of temporarily employed labor |
0.03(1.35) | |
Quota rents and leases |
1E-07 (0.10) | |
Sample selection control | ||
Number of years observed |
-0.005*(-2.56) |
-0.002*(-4.87) |
Number of observations |
38,641 |
24,339 |
Notes: It is not possible to present all estimates, since ANCOVA gives an estimate for every category of a
nominal variable, resulting in 256 estimates for each industry etc. Estimates for all categories are available
upon request; statistical significance at the 1 percent level is indicated (*).
t-values in parentheses.
Fourth, the location effect is captured by including 440 dummy variables for the German
districts (Kreise). It is worth noting that with this approach we not only capture differences
in the performance of the firms located in the eastern or western part of Germany (e.g., Funke
and Rahn, 2002), but also assess the efficiency of firms at a much smaller geographical scale.
The results for firm location suggest that regional factors play a fairly important role. The
explanatory power of location in terms of partial R2 is 3.12 percent for the 1992-2005 period and
2.77 percent for the 1999-2005 period (Table 5). Thus, these finding are grounds for accepting
Hypothesis 2. The location variable refers to the firm’s headquarters, not to the location of
branch plants, which may be located in other regions. However, since more than 90 percent of
the firms in the Cost Structure Census are single-establishment firms, the effect of branch plants
located in other regions is not expected to be large or important.
Furthermore, firm size is the only firm-specific determinant that explains a large part of pro-
ductive efficiency (Table 5). Other factors, such as the share of R&D expenditure, the firm’s
legal form, and indicators for the degree of outsourcing are not important. The parameter es-
timates (Table 6) show a negative effect of R&D on productive efficiency. This confirms the
15