Density-dependent dynamics in regulated industries: founding rates of banks and life insurance companies

Administrative Science Quarterly,  March, 1991  by James Ranger-Moore,  Jane Banaszak-Holl,  Michael T. Hannan

• E-mail
• Print
• Link

For each subpopulation, adding the effect of density improves the fit significantly (columns 2 and 5). In addition, both first-order and second-order effects are significant individually for each subpopulation. Most important, each of these effects has the predicted sign: both first-order effects are positive (and less than unity) and both second-order effects are negative. Thus the effect of own density is nonmonotonic for each subpopulation.

Figure 5 (right vertical scale) shows that the founding rate of commercial banks reaches a maximum when its density reaches 54, at which point the rate is 7.0 times higher than at zero density. Note that this is close to the result for the whole population. The founding rate of savings banks reaches a maximum when density equals 22. At this density, the founding rate for savings banks is 10.3 times higher than the rate at zero density (Figure 5, left vertical scale).

There is no evidence here of links between these subpopulations. Adding the cross-effect of the density ([N.sub.j]) scarecely increases log-likelihoods, as can be seen by comparing the log-likelihoods in columns 2 and 3 and columns 5 and 6 in Table 2. Despite the plausible claims that interlocking directorates between commercial banks and savings banks would have created links between the density of commercials and the founding rate of savings banks, it appears that the efforts by state regulators to maintain institutional barriers between the subpopulations trimphed in practice, at least insofar as founding rates are concerned.

Mutual and Stock Life Insurance Companies

Table 3 shows that the period effects are quite different for mutual and stock life insurance companies. Initial regulation (period 2) raised the rate substantially and significantly for stock companies but the other two period effects are smaller and insignificant. By contrast, initial regulation had essentially no effect on foundings in the mutual population, but more extensive regulation in the third period depressed it greatly, and the crisis following the Armstrong Investigation (period 4) depressed it still more. These negative effects are deflections in the overall strongly positive time trend in the mutual founding rate. Economic conditions affected both populations similarly, with depressions lowering founding rates by 30-40 percent. Wartime periods do not appear to have affected either subpopulation. And each subpopulation has a nonmonotonic pattern of effects of foundings in the previous year.

When we add effects of own density, the patterns of signs of effects of own density agree with the theory (columns 2 and 5). For mutuals, the pioneer subpopulation, both first-order and second-order effects of own density are significant. While the signs of the two effects of own density for stock companies agree with the theory, only the second-order effect is significant (column 5).

For mutuals, the turning point in the density-dependent founding process occurs at a density of 53 firms, according to the estimates in column 3 of Table 3, while the empirical population reaches a maximum of 96 firms. As Figure 6 (left vertical scale) shows, the rate at the turning point is 36 times greater than at zero density.