State Formation, Ideological Competition, and the Ecology of Israeli Workers' Cooperatives, 1920-1992 - Statistical Data Included

Administrative Science Quarterly,  March, 2000  by Paul Ingram,  Tal Simons

• E-mail
• Print
• Link

APPENDIX: Baseline Model

The baseline model includes a large number of control variables that are not used to test our hypotheses but are necessary to a well-specified model of coop failure. We present the descriptions of the control variables, discussion of theories that lead to expectations about them, and results together. Table A.1 shows the results for the control variables.

Coop age. There has been an active debate regarding the age dependence of organizational failure. The most recent findings indicate that the risk of failure increases with age and that the cause appears to be the increasing obsolescence of organizations over time (Barron, West, and Hannan, 1994; Ranger-Moore, 1997). There are no indications of the significance of the age-range coefficients in table A.1 because age dependence in a piecewise-exponential model is indicated by differences in the magnitude of coefficients for adjacent age ranges, not by the significance of the coefficients. Chi-squared tests on model 3 indicate the following ordering of the magnitudes of the age coefficients: (0-2 years) [less than] (2-5 years) [less than] (5-10 years) = (10-20 years) = (20-30 years) [less than] (30-40 years) [less than] (40+ years). The resulting pattern, of a failure rate that increases with age but is flat in the middle of the age distribution, has been found in previous studies of organizational failure (Barro n, West, and Hannan, 1994; Ingram and Inman, 1996).

Other coops absorbed. The risk of failure of an organization is typically found to decrease with its size (Baum, 1996), although for coops, many of the advantages of worker-ownership are diffused with the size of the organization (Perry, 1998). Our data do not include the number of participants, total assets, the volume of business, or other common measures of organizational size. We were able to construct a variable that is a count of the number of other coops that have been absorbed into the focal coop (an absorption takes place when two coops merge, with one persisting and the other disappearing). Because internal growth is expensive for current members, when coops grow, it is often through absorptions of other coops. That was the pattern of growth for the largest Israeli coops, the bus coops Dan and Egged (five and eight absorptions, respectively). The number of absorptions can be expected to represent slack resources and whatever other positive characteristics allow one organization to absorb another. Th e absorptions coefficient is negative and bordering on significance (p [approx] .065, one-tailed test).