LeoF.Schnoren/an/an/an/a

The Separation of Home and Work: A Problem for Human Ecology¹

I.

… Seen in historical perspective, the separation of place of work from place of residence is a relatively recent phenomenon and has been closely associated with the course of industrialization. Liepmann has suggested that these recurrent daily movements between home and work supplement migration and enhance the stability of community structure by contributing to the flexibility of industrial-economic organization. This contribution is most important in effecting adjustments to the changes that occur with the expansion and decline of particular industries, the short-distance relocation of factories, and seasonal fluctuations.…

The separation of home and work, however, is not without its dysfunctional features. Some attention has been directed toward the possibility of severe physiological and psychological strain upon individual employees who must travel long distances to work. In addition, there have been numerous discussions of the problems of cities themselves, increasingly threatened with a drastic shrinkage of their tax bases. The problems of financing municipal services may be expected to multiply with a continuation of the trend toward decentralization. In addition to the costs of daily movement to the family budgets of modern workers, the costs of elaborate transportation systems to the municipality must be considered. In particular, the initial capital costs of underground and overhead systems in the largest cities are enormous. Added to these, however, are operating expenses, many of which elude exact calculation. Still another increasing cost to the city is that represented by the loss of revenue arising out of traffic congestion, for component business units. A significant proportion of this congestion is brought about by the work trips of persons finding employment within the local area.

The traffic problem has persuaded planners and other interested officials to participate in such efforts as the federal program of origin-and-destination traffic studies. These surveys, jointly supported by federal, state, and municipal funds, have been carried out in more than fifty cities, and represent a valuable new source of urban data. With the inception of such studies a large body of by-product material has become available for analysis by social scientists, and a fund of research knowledge is being rapidly accumulated.

II.

In the course of two recent studies of the residential locations of industrial employees, it has been asserted that their distribution is the consequence of the operation of an underlying "principle of least effort."² An application of this hypothesis to account for the residential distribution of industrial workers was first attempted by Carroll, whose principal argument is "that employees of industrial plants seek to minimize the distance between home and work, and that the aggregate choices of large numbers of employees will tend to produce the observed pattern."³ More recently, the staff of the Industrial Areas Study, University of North Carolina, has subscribed to this explanatory device.

The weight of Carroll’s argument rests upon the observation of a gradient pattern of worker residences by distance from the workplace. "The central thesis of this paper," he says, "is that industrial workers will seek to minimize distance from home to work. This generalization was based on data showing that the number of employees resident in each successive mile zone from the plant site beyond the first few miles diminished as distance was increased."⁴ Some attention is given by Carroll to factors other than the possible motives of the industrial employees studied. Some interesting hypotheses pertaining to the possible influence upon residential distribution patterns of such variables as type of industry, wage level, and size of city are presented. In the main, however, these are conceived only as limiting conditions to the operation of the "fundamental principle" of least effort. In Carroll’s words,

It will be sufficient to indicate that, while many factors are involved in the selection of homes and places of work, the persistence of the desire to minimize the distance separating workplace from home acting through each individual worker may be the single element which can create pattern out of the aggregate choices of large numbers of workers. It is, of course, obvious that these choices are differentially limited for each individual worker so that only in large aggregates can patterns begin to appear.⁵

The cause of the observed gradient distribution is thus to be found in a single dominating desire experienced by individual workers. It might be argued, however, that should an individual have at his disposal time and money in quantities sufficient to relieve him, to some extent, from the ordinary restrictions imposed by transport costs, he might locate his residence almost anywhere, and for any of a variety of motives. The latter might include, in fact, a desire to maximize the distance between home and work. The least-effort hypothesis appears to confuse motivation with its external limiting conditions.

Even if the foregoing consideration is omitted, however, the least-effort hypothesis remains subject to serious question on logical grounds. If the tendency to minimize effort is assumed to be constant throughout the population, it appears that the hypothesis offers a plausible explanation of the concentration of residences near work sites but fails to account for the equally obvious scatter away from those sites. This assumption of a constant desire to minimize effort meets still another difficulty if an explanation of change over time is attempted. Given this constant, the antecedent factors responsible for any change must be sought in the external conditions which limit the "basic desire" to minimize effort, for this desire is not conceived as a variable. Thus an explanation, in these terms, of the decentralization movement would appear to require an assumption to the effect that the desire to minimize effort has been on the wane in recent years.

These observations suggest that the factors considered by Carroll as comprising only limitations upon the operation of the least-effort principle may be those worthy of more serious study in their own right. Toward this end, certain findings from a recent study of Flint, Michigan will be presented here.⁶

III.

A. The Distance Between Home and Work. The fact that the distribution of worker residences assumes a gradient pattern with respect to distance might have been anticipated on the basis of Carroll’s research. The upper panel of Table 1 shows that, in the case of all six plants, the great majority of workers live within

TABLE 1 PER CENT DISTRIBUTION AND RATIO OF WORKERS TO RESIDENT POPULATION OF EMPLOYEES OF PRINCIPAL INDUSTRIAL INSTALLATIONS IN FLINT, MICHIGAN, BY DISTANCE, 1950

six miles of their place of employment. The computation of ratios of workers to resident population, however, shows that each plant draws workers from each of the five distance zones in accordance with the number of persons it employs. The gradient pattern of these ratios in strict accordance with size of plant employment could not be readily deduced from prior suggestions to the effect that the average distance between home and work varies directly with the size of plant employment.

The ratio for the largest plant (A) is seen to decline steadily with distance, reaching a plateau at approximately 18 miles and beyond. The ratios for the remaining plants, however, begin to level off at the second zone, with the decline remaining in strict accordance with the size of plant employment. It is interesting to note that the only one of the smaller plants exerting any pulling power over the area in the last distance zone (plant E) is the newest of the major industrial sites in the Flint area. The tendency for newer and more rapidly expanding industrial installations to draw workers from a wider area has been noted by other investigators.

Thus the size of the plant employment and the length of time that it has been located at a given site appear to be variables of more than passing interest to one who would explain the residential distribution of industrial employees. Even if viewed as conditions to the operation of some more basic tendency, their importance should not be overlooked.

B. The Distribution of Workers by Workshift. The fact that the origin-and-destination data used here contain information on the time of arrival at work allows an examination of the spatial distribution of employees working on different shifts. Table 2 summarizes the results of this study. As may be seen, the proportions of workers on the first (day) shift decline regularly with distance, while the proportions of those employed on the two remaining shifts increase as distance increases.⁷ Again, the least-effort hypothesis suggests nothing in the way of an explanation for it assumes that the desire to minimize effort is a constant, that is, an attribute of all workers. Certain other possibilities, given no recognition in the development of the least-effort hypothesis, might be considered here.

TABLE 2 PER CENT DISTRIBUTION OF EMPLOYEES OF PRINCIPAL INDUSTRIAL INSTALLATIONS IN FLINT, MICHIGAN, BY SHIFT, BY DISTANCE, 1950

First of all, it should be remembered that more recently employed workers, those with the least accumulated seniority, are most often assigned to the afternoon and evening shifts. These are the employees hired periodically in response to fluctuations in demand, and represent "marginal workers," a concept widely used in labor force analysis. The relationship found here between workshift and distance suggests the hypothesis that the "marginal labor force" may also be physically marginal to a given industrial community.⁸

From the standpoint of the individual worker, residential location some distance from a center might be advantageous in two ways. For one thing, alternative sources of employment in nearby cities are more accessible. Secondly, a mode of adjustment to fluctuation in labor demand is made possible—a way of life promising more security than can be gained through industrial employment alone. Here we refer to the pattern of part-time agriculture discussed in detail by Firey.⁹

Firey identifies a trend toward what can be described as an urbanization of the originally rural population, and a ruralization of the urbanites, participating in the outward drift from the city. It is his judgment that Genesee County (of which Flint is the center) is one of those counties in Michigan within which the farm population has most fully taken to urban wage employment, while at the same time its decentralizing urban population has begun the practice of extensive gardening and part-time farming in the area surrounding the city. "Thus gardening or parttime farming," he concludes, "has in a certain sense become a way of life for a large proportion of the people in Genesee County. This is particularly true of the zone which immediately surrounds Flint and the radial bands which extend along the paved highways leading outward from the city … (for) within this star-like area part-time farming or gardening is the predominant pattern."¹⁰ Whether or not for the same reasons, this pattern has long been established in the continental countries, where great numbers of workers alternate between agriculture and industrial employment during the course of the year. Its emergence in this country is not surprising, since our urban communities have tended to become increasingly market-oriented in supply, demand, and employment opportunities.

Proportionately greater difficulty in securing off-season employment is encountered where the units of production are highly specialized or where the occupational specialization of the workers is great. An increase in the size of the production unit may also be expected to result in further difficulties should a temporary shutdown become necessary, since great numbers of workers are released at one time and in one place. The automobile industry, which forms the basis of Flint’s economy, serves as an excellent example of this situation. The industry as a whole employs thousands and is, moreover, highly concentrated geographically. "The results," according to one observer, "are aggravated by the sensitive interlinkages among units, which necessitate that the closing of one unit be followed by the closing of others. Thus a cessation of activities in the automobile industry throws such large numbers out of work that it is impossible for the community to absorb them in other types of employment.…"¹¹

Perhaps it is this difficulty of finding other full-time employment in nearby industries that encourages the widespread part-time farming by shop workers observed by Firey. Given a location on one of the major arteries leading to the city, a factory worker is within relatively easy access of industrial employment, yet has ample land on which to raise garden crops in sufficient quantity to supplement purchased foodstuffs. Such a practice would be encouraged if he should be employed on either the late afternoon or evening workshift in the plant, for he would then be able to utilize the daylight hours in work on the land. It is even conceivable that great numbers of these workers prefer work on the later shifts, especially during the planting and harvesting seasons when the daylight hours can be used to greater advantage. It should also be remembered that the factories represent a significant source of extra income to persons whose principal occupation is farming and who maintain their rural residence. At any rate, the observed relationship between distance from workplace and the time of work may have as one consequence the stimulation of the pattern of part-time agriculturism discussed here, and it is within this broader community context that the findings might be interpreted.

C. Ride Sharing and the Ability to Pay Costs of Transportation. Still another matter we might consider is the ability of workers to pay the costs of transportation to and from work. Table 3 shows that as distance increases up to approximately

TABLE 3 PER CENT DISTRIBUTION OF CARS TRAVELING TO PRINCIPAL INDUSTRIAL INSTALLATIONS ON WORK TRIPS IN FLINT, MICHIGAN, BY NUMBER OF PERSONS IN CAR, BY DISTANCE, 195030 miles, the proportion of cars in which only one person is traveling to work declines, while the proportions in which there are two, three, and four or more persons increases. The mean number of passengers per car also increases with distance up to this "breaking point."¹² One possible explanation for the reversal of the observed tendency in the last distance zone may be found in the widely scattered distribution, at this extreme distance, of those who must regularly travel to the city for employment. Since the area surrounding the central city increases as the square of the radial distance from it, workers are presumably more scattered at this extreme distance, and thus find it more difficult to make ride-sharing arrangements with others.

When considered in an a priori manner, ride-sharing might be expected to increase with distance, for such a practice is an effective method of distributing the high costs of automobile transportation.¹³ The significance of this practice may be realized when it is remembered that the ability to pay transport costs to and from centers of activity appears to be one of the most critical selective factors in the centrifugal shift of our decentralizing urban population. It might be suggested here that many family units which otherwise could not participate in the decentralization movement may be able to do so by virtue of such arrangements as ride-sharing. Such a minimization of transport costs, together with the added security obtained by part-time agricultural activities, may account for the presence in these peripheral areas of large numbers of families whose general economic status would otherwise not permit such location. These are the persons whose scattered residential distribution remains inexplicable when the least-effort hypothesis is utilized.

IV.

We have indicated in the foregoing sections certain apparent limitations upon the use of the least-effort hypothesis in this problem area. These limitations, for the most part, appear to be a consequence of the form in which that hypothesis is stated. The postulation of a constant attribute as a fundamental causal factor meets resistance when variation is encountered. The remaining space will be devoted to the consideration of an alternative hypothesis of an intentionally different form.

From one theoretical point of view, the daily journey to work may serve as one of the most easily perceived data in the observation of community organization. As treated by one student of human ecology, the regular ebb and flow of community activity is viewed as itself expressive of community structure. According to Hawley,

Recurrent movements, as the name indicates, comprise all those movements that are routine and repetitive. They might also be called functional, for it is by this type of movement that the functioning of the community is carried on.… Each [of these movements] is an integral part in an established organization and is therefore essential to the maintenance of that organization. Recurrent movements involve no break with the past, no disruption of an established order. They are the means by which an existing equilibrium is maintained.¹⁴

The increasing spatial differentiation of the modern community, of which the separation of home and work is one aspect, might also be considered as reflecting an increasing functional differentiation. Such an interpretation assumes, of course, that space presents at least one measurable dimension of community structure. This assumption is, in fact, given formal expression by Hawley when he suggests that "the distribution of the elements of [the physical] structure [of the city] form a pattern of land uses which presumably is expressive of the interdependence among the various activities comprised by the city."¹⁵

In this spatial pattern, the functional units occupying different sites may be thought of as possessing different locational requirements. One requirement common to all units, of course, is space itself, or room in which to operate. But units may differ in the amount of space required. Because space is limited, particularly at the center of an area, those requiring the greatest amounts of space might be expected to locate away from that center. At the same time, any location involves costs to the occupant by virtue of his occupancy. Costs being highest at the center, the units least able to maintain occupancy of sites at the center may be expected to be found at or near the periphery. This cost, however, is not the only one exacted from a unit.

One other key characteristic of any site is the degree of accessibility to other units it may have. Units may have quite specifically defined needs for accessibility to other units, and this may be taken as another locational requirement. Just as with space, moreover, this need is fulfilled only at a cost to the unit involved. The cost, in this case, is experienced as the cost of transportation, for either the movement of the members of the unit from the site to another place or for the movement of goods and services to the site occupied.

Two assumptions regarding these costs underlie this discussion. The first is that rent, or the cost of occupancy of a site, declines with distance from an activity center. The most frequently observed decline is at a somewhat greater than proportional rate. Secondly, transport costs are assumed to increase with distance, at an approximately proportional rate, although significantly modified by the method of transport utilized. If it is then assumed that costs of location represent the sum of these costs, the following hypothesis suggests itself: The maximum distance from significant centers of activity at which a unit tends to locate is fixed at that point beyond which further savings in rent are insufficient to cover the added costs of transportation to these centers.

It is perhaps, in the interaction of these two broadly conceived "cost" factors that an explanation of the residential distributions of employees may be found. Units obviously differ in their ability to pay both of these costs, and the family unit is no exception. First of all, the work site may be taken to represent one of the most significant "centers of activity" for the family unit of the employee. Exchanges between the family unit and the production unit at the work site—in the form of the physical movement of the worker—are frequent, so that a certain degree of access is an important requirement. Within this broad range, then, it might be expected that the cost-paying ability of the family unit becomes significant. The area in which any given unit may be able to afford location may also be fairly broad. In the aggregate, however, the area open to families of a given cost-paying ability may be distributed about the relevant centers of activity in such a manner that the gradient pattern of residential distribution becomes readily observable. Since the ability to pay costs of transport and occupancy are conceived as variables in this formulation, no difficulty is encountered in the fact of scatter. Indeed, such a gradient distribution would be expected to follow.

It is in the light of this hypothesis that the characteristics of the population participating in the outward drift from the city may be reasonably interpreted. Available census data indicate that the ability to pay transport costs may well be a selective factor in residential decentralization, for peripheral areas of metropolitan districts are found to be occupied by families of higher than average socioeconomic status.¹⁶ In addition to costs of transport, of course, other financial considerations are involved in peripheral location. The entire range of family purchases may be expected to be somewhat more costly in the outlying areas, at least until sufficient densities of population make possible greater economies in the provision of goods and services. These remarks are not intended to imply that the economic aspect of location is the only one that can be identified, or that these costs are somehow the only factors operative. Nor are space and time granted some kind of deterministic role with reference to the location of units of the community. Emphasis is placed upon spatial and temporal relations in this approach for the simple reason that the patterns and processes in which we are interested occur in a space-time context.

In addition to static descriptions of community structure, ecological theory attempts to provide some information upon the processes of change in that structure. The general approach sketched here might also prove useful in an attack upon the problem of change in community organization. The hypothesis suggested above might be utilized in accounting for changes observed in the patterns of population distribution. That important changes have occurred is well known. Although changes in costs of occupancy should not be overlooked, the long-range trend toward residential decentralization can be viewed, in this context, as a consequence of a long-range decline in transportation costs. As such, this interpretation represents a formal statement of the frequent impressionistic observation to the effect that the automobile has "released" population from the immediate confines of the city. At any rate, some attention should be directed toward the development of hypotheses suitable to the description of change, in addition to those which offer only plausible accounts for observations relating to a given point in time, as in the case of the least-effort hypothesis.

With respect to the distribution of other centers of activity significant to the functioning of the family, it has been found that many are located at a lesser average distance than that between home and workplace.¹⁷ The location of these units could also be approached through the use of hypotheses of this general order. Retail shopping centers, for example, might tend to locate at a relatively low average distance from their supporting populations (made up of family units, in the main) by virtue of the necessity for frequent exchange with those populations. The latter need may well be one of the key locational requirements of such units, and the concomitant costs could be treated from the point of view outlined here as those deriving from the necessity for a high degree of accessibility to other units. The study of the location of industrial activities and other work sites might also be approached in this manner. Treatment of the location of all units comprising the community would be necessary in a complete description of communal land uses.

In any event, the approach outlined here—although far from entirely satis-factory—might be productive of more general information than would one in which a single observed relationship is given priority out of proportion to its apparent significance. This discussion is intended primarily to indicate that another approach of apparently equal research feasibility is possible. The fact that the least-effort hypothesis has happened to dominate the little research already carried out in this problem area should not deter the presentation of alternative modes of explanation.

¹ From , 1954, 32:336–343. By permission.

² This hypothesis is given its most detailed elaboration in George K. Zipf, Human Behavior and the Principle of Least Effort (Cambridge: Addison-Wesley Press, 1949). A summary exposition may be found in "The Hypothesis of the ’Minimum Equation’ as a Unifying Social Principle," American Sociological Review, 12 (1947), pp. 627–650.

³ J. Douglas Carroll, Jr., "Home-Work Relationships of Industrial Employees" (unpublished doctoral dissertation, Harvard University, 1950), p. 21. Certain of Carroll’s research findings are summarized in "Some Aspects of Home-Work Relations of Industrial Workers," Land Economics, 25 (1949), pp. 414–422, and in "The Relations of Homes to Work Places and the Spatial Pattern of Cities," Social Forces, 30 (1952), pp. 271–282.

⁴ "Home-Work Relationships of Industrial Employees," p. 130.

⁵Ibid., p. 24.

⁶ The data presented below were gathered in the origin-and-destination traffic study carried out in Flint. The author is indebted to the Michigan State Highway Department for the use of these materials. The time of the traffic survey, the summer of 1950, permitted some use of these data in combination with statistics from the decennial census. Flint, a single-industry city, had a population just in excess of 163,000 in 1950. The study reported here was confined to the 66,000 employees of the six Flint plants of the General Motors Corporation: (A) Buick Motor, (B) Chevrolet Motor, (C) Fisher Body, (D) A. C. Sparkplug—Dort Highway, (E) Chevrolet Assembly, and (F) A. C. Sparkplug—Industrial Avenue. Reference will be made to these plants by these letter designations.

⁷ This same relationship between distance and workshift was found for each of the six individual plants studied, although data for the latter are not presented in this report.

⁸ Estimates of manufacturing employment in Flint have been made for 1950, which has been identified as an extremely stable year when the magnitude of fluctuation in employment from month to month was at a minimum. These figures approximate the employment of the six General Motors plants under study. Manufacturing employment rose from 63,300 in May to a high of 67,400 in September, and then fell off to 66,400 by November. The difference of 4,100 between the high (September) and low (May) figures provides an estimate of the size of Flint’s marginal labor force during this period. (These estimates were abstracted from the Monthly Estimate of the Labor Force prepared by the Flint office of the Michigan Unemployment Compensation Commission, and appear in the appropriate issues of the Labor Market Letter published by that office.)

⁹ Walter Firey, Social Aspects to Land-Use Planning in the Country-City Fringe: The Case of Flint, Michigan (East Lansing: Michigan State Agricultural Experiment Station, Special Bulletin 339, 1946). Although Firey’s study was also limited to Flint, a similar pattern has been found in many other areas. See, for example, Nathan L. Whetten and R. F. Field, Studies of Suburbanization in Connecticut, 2, Norwich: An Industrial Part-time Farming Area (Storrs: Connecticut State Agricultural Experiment Station Bulletin, 226, 1938).

¹⁰Op. cit., pp. 16–17. Statistics revealing the number and size of farm units within the county offer a measure of corroboratory evidence. As farms have increased in number there has occurred a concomitant decrease in their size.

¹¹ Amos H. Hawley, Human Ecology (New York: The Ronald Press, 1950), p. 312.

¹² Although not shown in Table 3, the same relationship between distance and the number of persons in the car was found for each of the six individual plants studied.

¹³ It has been estimated that each mile added to the daily journey to work adds an additional $25.00 to the annual cost of work transportation alone. Richard Dewey, "Peripheral Expansion in Milwaukee County," American Journal of Sociology, 54 (1948), p. 121.

¹⁴Op. cit., pp. 326–327. The material upon which the following discussion is based was drawn from a seminar at the University of Michigan conducted by Hawley.

¹⁵Ibid., p. 382.

¹⁶ Outlying areas of metropolitan districts have been found to contain significantly greater proportions of persons with the following characteristics: one or more years of college education; self-employed; females not in the labor force; professionals and proprietors, managers and officials. Census of Housing data provide interesting supplementary evidence. Among other differences, the outlying areas have been found to contain higher proportions of one-family, owner-occupied homes, of larger and more recent construction.

¹⁷ Trips to 20 other activities have been found to be shorter than work trips. See Donald L. Foley, "The Use of Local Facilities in a Metropolis," American Journal of Sociology, 56 (1950), pp. 238–246. See also Carroll, Home-Work Relationships of Industrial Employees, pp. 71–74 and 86 where it is noted that "trips to work are short, but shopping, school, and church trips are shorter."