measures of individual attributes and firm-level performance that were already on the right-hand side of the regressions.¹³ In the regression for the whole sample, outside intensity is significant when put on the right-hand side. This result is obviously influenced by the presence of both the unemployed (with zero inside intensity and varying outside intensity) and the separated and reemployed (with zero inside intensity and outside intensity assumed at the maximum level.) We therefore ran the regression only for people who had not separated.

Both inside and outside intensity are now not significant, but this is not necessarily cause for rejecting our first hypothesis. Given the evidence from the cross-tabulations in Table 7-1, the most reasonable interpretation is that use of outside survival strategies is uniformly high and not affected by the inside intensity of work. This is further evidence of a severe crisis for state-sector employees.

Relevant for the second hypothesis, we find both the female dummy and the dummy for a person being of pensionable age to be significantly negative, indicating that these people work less both inside and outside the firm. Years of tenure (i.e., years of employment in that firm) are positive and significant in both regressions. The dummies for education are not significant, and the only significant age dummy is for those aged 26-35 (negative in the inside intensity regression). Thus it seems that women and pensioners work relatively less inside and do not make this up with more work outside the firm.¹⁴