Finally, knowledgeable individuals in each community and, whenever possible, executives associated with the "potentially outmigrating" firm were contacted by telephone. These persons were requested to confirm or deny the closing or outmigration of particular plants in their community or with their company. In many cases, the plants provided by the "survival" method were too small to include, or they were still in operation but under a different name. If operation had in fact ceased, the following information was solicited:

(1) The year the company ceased production in the community.

(2) The estimated peak employment of the plant.

(3) The principal products produced by the company in the comniunity.

(4) The ownership characteristics of the concern; i.e., was the facility locally owned, not locally owned but an independent operation, or a branch or subsidiary of another company.

(5) If the plant was a branch or sub

sidiary, who was the parent company and where were its headquarters?

(6) Had the plant, or parent company of the plant, been acquired by another individual or company during its life in the community? If so, when did this acquisition (merger) occur and who was the acquiring party?

(7) The reasons for halting production or moving out of the community.

(8) The names of any management personnel who had remained in the area after the local operation had ceased.

If other sources of information (such as corporate executives, previous employees, or the original owners) were available, they were also contacted and requested to answer the above questions. In total, 197 communities plus numerous companies and employees were contacted. During the period covered by this study (1965-75), 128 plants employing over 20 workers failed or outmigrated from rural Iowa. Of these 128 firms, 75 were branch plants and the remaining 53 were locally owned concerns.

DERIVATION OF LOCATIONAL INSTABILITY RATES

It was previously hypothesized that the probability of failure of a unit firm exceeded that of a multiplant corporation, yet branch plants were still less locationally stable because of their greater propensities for migration. To test these hypotheses, locational instability rates were computed for the sample of closed Iowa manufacturers. These rates are simply the ratio of the number of outmigrating and failing firms in a select classification (SIC, employment size, branch plants, local plants, location, etc.) to the total number of

The most frequently contacted individuals in the communities were persons of leadership level in the local development commissions or chambers of commerce, lowa State Extension personnel, or the mavors or bank presidents of the smaller municipalities.

Because of the high degree of cooperation I received during these interviews, it is my judgment that the 128 firms represent almost all of the rural lowa manufacturers which had employed over twenty workers and outmigrated or ceased production during the last ten years. It is also my opinion that the information concerning each is accurate, though this accuracy is possibly less for those closed several vears ago.

rural Iowa manufacturers in such a category at a particular point in time (in this case 1973-74). Plant closings for the entire study period (1965-75) were included in the computation of these rates to minimize the importance of cyclical variation in migrations and failures. The denominators for the ratios were obtained from the 1973-74 Directory of Iowa Manufacturers.

If branch plants were in fact less locationally stable than unit concerns, their locational instability rates must be significantly greater than those of the indigenous firms. However, this rate differential does not conclusively prove that ownership characteristics were the source of the instability differences. Branch and local plants also differed significantly in the types of products produced (x = 3.02) and in the distribution of their plant sizes (x2= 85.45) (refer to Tables 1 and 2). If a disproportionately large number of rural Iowa branches were producing in "declining" or "footloose" industries, or if large plants were more responsive to changes in locational advantage than smaller companies (possibly because of better access to the capital markets or superior management), the locational instability credited to ownership may have actually resulted from disparities in industrial composition and plant size. The influence of these variables was controlled for by computing locational instability rates within select industrial and employment classes. Had multiplant firms exhibited a greater propensity for plant closings. due to the possibility of achieving scale economies and the relative absence of locational inertia, the locational instability rate of branches should exceed that of indigenous companies regardless of the product type or employment categories utilized. A

higher rate of closings for branches could also result if branch plants were more heavily concentrated in the higher transit cost western region of the state than were unit concerns. However, such a distribution did not occur (see Table 3). The proportion of the state's branch plants in each region (eastern and western Iowa) was almost identical to the proportion of the state's unit firms in those areas (x2 = 1.73). Since no significant difference existed, controlling for the influence of location on instability rates was not necessary.

SUMMARY OF FINDINGS

The statistical findings of this study provide strong support for the previous hypotheses.' The probability of an independent company failing exceeded that of a multiplant firm. Plant closings resulting from branch and local company failures were 19 and 37, respectively. These represented 5.43 percent of the 1973-74 base for multiplant firms and 8.73 percent for local concerns. However, in spite of a lower bankruptcy rate, rural Iowa branch plants were still significantly less locationally stable than indigenous manufacturers (Table 4). The locational instability rate of branches was approximately 50 percent greater than that of local firms, and branch plant stability was not significantly influenced by the location of the multiplant corporations' headquarters (Iowa versus non-lowa).

• The line of demarcation closely followed the route of Interstate Highway 35.

The statistical tests of significance applied throughout this section are tests for differences between proportions.

cational stability of branch plants. The probability of local firm outmigration or failure was not significantly influenced by plant size (x2 = .55). Within each employment category, the locational instability rates of branch plants consistently exceeded those of the indigenous concerns. These results indicate that the differences in branch and unit firm locational instability rates cannot be attributed to a skewed distribution of plant sizes.

The locational stability of Iowa manufacturers was also significantly influenced by the type of products they produced (Table 6). The manufacturers of durable goods exhibited rates significantly larger than those of the producers of nondurable goods, and the locational instability of firms in the electrical and lumber industries was significantly greater than that for all Iowa plants. However, within product groups. branch plants were

more

locationally unstable than indigenous firms. The probability of branch plant failure or outmigration was significantly greater than that of unit firms for all product classifications except nondurable goods. Furthermore, the location

al instability of indigenous firms varied little among product classes. With the exception of the wood products manufacturers, the instability rates of local firms within the selected product groups never deviated more than two and one-half percentage points from the Iowa average for unit concerns. Therefore, the observed differences between the aggregate rates of each group resulted primarily from changes in branch plant instability. To summarize, the locational stability of a region's industry will be altered by the size distribution of that region's plants and its industrial composition. However, that instability will be transmitted to the industrial base through the area's branch plants. The stability rates of a community's manufacturers will be only minimally affected by plant size and product mix if all of the community's industries are locally owned.

CONCLUSION

In recent years, renewed interest has been shown in rural development. Regional, state, and local organizations were formed to undertake the measures necessary to raise area incomes, reduce the rural-to-urban migration, and augment the amenities

"The relatively high instability rates for manufacturers in these industrial groups may have resulted from the following features: (1) The lumber and electrical industries are engaged in the production of durable goods, and are, therefore, more susceptible to fluctuations in aggregate demand. (2) The US. South and Japan had developed comparative advantages in the production of wood products and electrical equipment, respectively Competition from these regions forced lowa manufacturers to seek "lower cost" locations.