Labor contract negotiations in the airline industry: airline labor negotiations take 1.3 years, on average, to conclude, and about half go into Federal mediation; much of the variance in the duration of negotiations can be attributed to which particular airlines and unions are bargaining, not to economic conditions

Monthly Labor Review, July, 2003 by Andrew von Nordenflycht, Thomas A. Kochan

Results. Summary statistics on the economic variables are presented in table 8. The method of ordinary least squares was used to regress the duration of negotiations (duration2) on the variables just described, as well as on dummy variables for the year, airline, occupation, and union. A number of the variables had no significant coefficients, either alone or in various combinations. Carrier size and some measures of growth yielded some significant results, but neither the profitability measures nor the change-in-profitability measures generated any significant coefficients. (8) The results presented use only those measures which had significant coefficients in some specifications. Table 9 gives the results of the ordinary least-squares regressions. Columns 1 through 5 do not include carrier dummies, whereas columns 6 through 12 do.

The coefficient on the time trend is positive and significant in every model, except when year dummies are included. This is true even when carrier dummies are included (columns 6, 7, 9, 10, and 11), suggesting that the duration of negotiations has increased significantly over time, even after controlling for changes in the composition of the sample. However, adding year dummies to control for idiosyncratic year effects renders the trend insignificant.

The coefficient on carrier size (that is, revenue) is positive and highly significant in columns 1 through 5, which take into account cross-airline size variation. The revenue coefficient is insignificant, however, in each model with carrier fixed effects, implying that the duration of negotiations for a given carrier does not increase significantly as the carrier grows larger. Nonetheless, in the cross section, in which size differences among airlines can be quite large, larger airlines do take longer to negotiate contracts than smaller airlines take. The coefficient on carrier size, when it is significant, is approximately 0.7, implying that a $1 billion size differential between carriers entails about a 22-day differential in the duration of negotiations. The difference between the smallest and largest revenue values is $19.3 billion, which translates into a maximum 14-month difference in the duration of negotiations.

There is some evidence that higher carrier growth rates lead to longer negotiations, but these results are not robust, because they depend strongly on the measurement and model chosen. Both a carrier's 1-year growth rate and the industry's 3-year average growth rate have positive coefficients in all models. However, the significance of those coefficients is not consistent. For example, a carrier's 1-year growth rate has a significant positive coefficient in several of the models with carrier fixed effects, implying that, as carriers register higher growth rates, the duration of their contract negotiations gets longer. However, the coefficient is not significant in the cross section (without the carrier fixed effects). The same is true for the 3-year average of industry growth. Furthermore, neither the carrier-level 3-year average growth nor the industry-level 1-year growth has significant coefficients in any model. (These results are not shown in table 9). Hence, there is some evidence that higher growth may lead to a longer duration of negotiations, but the finding is not robust. Adding fixed effects for unions (columns 11 and 12) enhances the precision of the estimated coefficients of the other independent variables, but does not alter any of the basic patterns.