Household Geographic Mobility and the Impact on Macro Market Segments

ABSTRACT - After more than a quarter of a century (19471975) in which one-fifth of the population consistently changed addresses each year, the annual mobility rate declined to 17.3% only to rise again in 1984-85 to 20%. Because of the numerous macro and micro implications, other academic disciplines, in addition to marketing, have studied the mobility phenomenon in recent years. In the present study, three emergent societal groups long distance mobiles, short distance mobiles, and renovators - are studied in the contemporary societal context. The relevance of household geographic mobility as a contemporary segmentation base is evaluated by rigorously defining, and testing the purchase behavior of resulting segments. The applications of these findings for mobility theories are also addressed.

Citation:

John Gottko and Paul Sauer (1989) ,"Household Geographic Mobility and the Impact on Macro Market Segments", in NA - Advances in Consumer Research Volume 16, eds. Thomas K. Srull, Provo, UT : Association for Consumer Research, Pages: 85-92.

Advances in Consumer Research Volume 16, 1989      Pages 85-92

HOUSEHOLD GEOGRAPHIC MOBILITY AND THE IMPACT ON MACRO MARKET SEGMENTS

John Gottko, Oregon State University

Paul Sauer, SUNYAB

ABSTRACT -

After more than a quarter of a century (19471975) in which one-fifth of the population consistently changed addresses each year, the annual mobility rate declined to 17.3% only to rise again in 1984-85 to 20%. Because of the numerous macro and micro implications, other academic disciplines, in addition to marketing, have studied the mobility phenomenon in recent years. In the present study, three emergent societal groups long distance mobiles, short distance mobiles, and renovators - are studied in the contemporary societal context. The relevance of household geographic mobility as a contemporary segmentation base is evaluated by rigorously defining, and testing the purchase behavior of resulting segments. The applications of these findings for mobility theories are also addressed.

Introduction

The rate of annual geographic mobility in the U.S. is intimately intertwined with such societal trends as the increased number of single adult households, two-person cohabitant households, and group households. The latter consist of several persons of the same or opposite sex sharing expenses by living together. Such non-family households are growing more rapidly in number than are the conventional family households (Kotler, 1980). Interestingly, the phenomenon of contemporary geographic mobility is multidimensioned in that increases are perhaps occurring in the rate of geographic mobility within the nontraditional household sector at the same time that the traditional household sector had been experiencing a decreased rate of growth. Yet, the demand for housing and for related products and services, might be experiencing double jeopardy: not only is the mobility rate down among traditional households, but the net impact of the increased rate of mobility among the nontraditional sector might be to decrease the number of new household formations.

The mobility and non-mobility decisions of both traditional and non-traditional households are emerging, and increasingly being shaped by societal and other "environmental" developments. At the same time, these decisions are reshaping the nature of this environment, as well as the demand for products and services.

Both the academic and practitioner sides of the discipline of marketing, are affected at the macro as well as the micro levels. At the macro level, the role of marketing, and the marketing concept, in the greater U.S. society is affected. For example, Hirschman (1983) has reported that "the marketing concept, as a normative framework" is not always applicable "because of the personal values and social norms" that characterize a particular process. The selection of market segments and the design of marketing mixes are impacted at the micro level as well.

The three objectives of this research are:

1. To reevaluate the existence of household geographic mobility segments in a framework provided by theoretical contributions of earlier geographic mobility researchers.

2. To use these evaluations to test hypotheses stemming from these theories.

3. To examine the relevance of household geographic mobility as a segmentation base for demand in the contemporary society.

Rationale for Mobility Segments: Resulting Product Decisions

Over twenty years ago, Andreasen (1966) proposed his "tentative theory of geographic mobility.' This was based on the proposition that "long-distance" geographic mobiles, as opposed to "short distance" or local movers, possessed certain demographic and lifestyle characteristics which facilitated both the opportunity and willingness to undertake more complex changes in their places of residence. Andreasen tested the theory in a study of Philadelphia (PA) households which had recently moved to that city. A principal finding was that long distance mobiles might be regarded as a market segment because they exhibited a distinct demographic and life-style profile (See Table 2). He presented data on the nature of actual or planned purchases or major assets (Included were: major appliances; minor appliances; furniture; automobiles; clothing; drapes; slipcovers; and other.) after the long distance mobiles' arrival in Philadelphia. Based on the percents of sampled long-distance mobiles who indicated either actual or purchase intentions-- percent ranged from 15.5 of the sample indicating actual or planned purchases of minor appliances to 83.8 for drapes and slipcovers-- Andreasen concluded that these mobiles would be important target customers for many such products. While Andreasen contended that these purchases would exceed those of either local movers, or non-movers, he did not support this with data. In all of the categories of major products, or assets, studied by Andreasen, a greater proportion of the sample reported actual, as opposed to planned or purchase intentions, so that by the time of the survey, purchases had already occurred. This provided a much stronger behavioral statement of demand than purchase intentions which may or may not occur.

Rationale For Mobility Segments: Resulting Retail Patronage Decisions

Soon after the Andreasen Philadelphia study, Bell (1967) conducted research on a sample of long-distance mobiles in an urban Michigan location. To test Andreasen's contention that long-distance mobiles would experience considerable store switching, Bell investigated the nature of patronage changes in ten retail categories (The ten were: food store; beauty parlors; dry cleaners; financial institutions; insurance firms; furniture stores; appliance stores; women's clothing stores; men's clothing; and medical services.), Bell found the length of time it took to rebuild the retail sources of supply could vary from almost immediate for supermarkets and banks to nine weeks to select dentists and sources of supply for women's clothing. Similarly, the number of potential suppliers visited and screened, ranged from one in medical service to four in selecting a furniture store (Bell, 1967). Like Andreasen, Bell concluded that long distance mobiles constituted a unique market segment.

However, though Andreasen and Bell contended that the long distance mobiles represent a viable segment, they did not empirically demonstrate that these mobiles differ in their marketplace responses, from shorter distance mobiles or non-mobiles.

Household Geographic Mobility as a Segmentation Base

A requirement of segmentation is that there be congruity or homogeneity within segments, but a lack of congruity or heterogeneity across segments (Kotler, 1980). The two existing studies did not show that the long-distance mobiles, as a "segment," respond differently (with their product and retail patronage marketplace decisions) than do those who are not long-distance mobiles. While other researchers have suggested possible homogeneity within the local movers segment, marketplace responses again were not tested (Quigley and Weinberg, 1917). Thus, geographic mobility as a segmentation base has not been tested by marketplace behavior for neither homogeneity within, nor heterogeneity across, resulting segments.

The mobility segmentation base embodies both practitioner and theoretical considerations. Some of the practical implications of the earlier marketing mobility studies have been noted above. While both Andreasen and Bell focused on the inter-urban mobiles as a market segment, Quigley and Weinberg (1977) reviewed 19701975 Census data and suggested the possibility of a homogeneous segment consisting of intra-urban mobiles. However, some research challenges the notion that significant differences in marketplace responses result from mobility status. For example, Dickson (1982) performed a role-playing experiment which contrasted the patronage preferences of long distance mobiles vs. all other consumers of major appliances. Dickson found that mobility status did not affect selection of chain stores vs. local stores.

The study of geographic mobility has received research attention over the years from academic fields other than marketing and while the marketing studies have focused on the long-distance mobiles, the other fields have tended to concentrate on the local, or intraurban, mobiles. A theoretical rationale for possible divergent marketplace behavior between the two groups of mobiles is provided by Roseman, a researcher from the field of geography.

The migration process is formally distinguished by Roseman according to a human movement concept utilizing partial and total displacement.

Partial displacement is defined as a move in which the family stays in the same general area keeping some "location activities" (movement for work, shopping, and entertainment), but changing its center of gravity (the home). Total displacement occurs when a household moves to a new general area which results in a complete change both in its "center of gravity" and its "location activities" (Roseman, 1971).

Roseman suggests that the inter-urban, or longer-distance moving, households would experience more marked and extreme changes in their marketplace behavior, resulting from their "total displacement" than would the "partially displaced" in a-urban, or shorter distance moving, households. Moreover, it appears implicit that the "partial displacement" might involve marketplace changes which exceed any changes experienced by non-mobile households.

Characteristics of Mobility Segments

Andreasen (1966) and Bell (1967,1969) have demonstrated that there are unique demographic, and lifestyle, characteristics associated with the inter-urban mobile households. However, about three-fourths of all Household moves in the U.S. are within, rather than across counties. Furthermore, the U.S. Census (Current Population Reports, 1984), Quigley and Weinberg (1977), Gottko (1983) and others have shown that intraurban moving households also possess identifiable demographics and life-style characteristics.

Another consideration in the identification of possible segments is a temporal one. The rate of household geographic mobility in the U.S. is undergoing change. After more than a quarter of a century (1947-1975) in which one-fifth of the population consistently changed addresses each year, the annual mobility rate fell to 17.7% in the 1976 Census Bureau study and continued to decline to 17.2% between March, 1980 and March, 1981 and 16.6% between March, 1982 and March, 1983 before rising to 20.2% between March, 1984 and March, 1985, according to Census officials (Current Population Reports, 1983,1987). The major reason cited for this fluctuation in mobility is the cost of money as reflected in the rate of interest on mortgage and home improvement loans (Wall Street Journal, 1987). While the mobility groups have differing demand effects, income is still a major factor in segmentation.

A related trend is the one involving remodeling, converting housing space to other uses, and adding on, all as alternatives to moving up to better accommodations. This is because interest rates have made it more difficult for the average wage earner to qualify for a home. As a result, people are looking more and more toward staying in their current neighborhood, and house, and improving rather than moving. This is evidenced by $80 billion spent by American households in 1985 to maintain and remodel their homes. This represented a 15 percent increase over 1984. [According to a 1985 Consumer Preference Survey of 920 homeowners conducted by the National Association of Home Builders, the most common remodeling project is kitchen remodeling with an average price of $12,400. The second most frequent remodeling project is finishing a basement-recreation room at an average price per job of $11,000 (Corvallis Gazette Times(A), 1986)].

EXHIBIT 1

MACRO AND MICRO SEGMENTATION USING MOBILITY SEGMENTS

Thus, it is not likely that the inter-urban and intra-urban mobile households have remained static over time, with respect to either consumer characteristics, or marketplace responses, nor is it likely that in the future they will remain constituted as they presently are. Furthermore, theoretical contributions, such as Andreasen's (1966) "tentative theory of geographic mobility," which emphasized upward career mobility, might also become outdated. The societal developments cited above might signal a shift in attitudes and this shift might be impacting inter-urban mobility decisions. Perhaps those who are opting not to undertake long distance moves are instead undertaking remodeling, converting space to other uses, and adding on to existing houses as a means of "trading up." For example, according to Merrill Lynch Relocation Management, Inc., the number of relocated employees per company declined from 188 in 1984 to 183 in 1985 (Gazette Times(B), 1986). Perhaps others are choosing to forego short distance moves and also seek renovations as a means of increasing household satisfaction (Carson, 1982).

While much less study has been directed at renovators, compared to the mobiles, there are some theoretical underpinnings. These underpinnings are involved with financially or economically motivated renovation decisions on the part of homeowners (Smith, Tschappat and Racster, 1981). Consequently, a realistic segmentation study, employing household geographic mobility as the base, must examine not only the potential existence of such traditional groupings as inter-urban, and intra-urban mobiles as segments but also the potential among the emergent group of non-mobile renovators.

Models of Demand

The value of segmentation to marketers lies in its implications for marketplace behavior, or demand. Household geographic mobility, like social class, might be termed macro, in contrast to, micro segmentation variables. That is, just as social class is a composite of a number of variables (income, occupation, education, etc.), so is household geographic mobility (income, occupation, life-style, etc.). On the other hand, individual variables such as income and occupation are micro segmentation variables. Exhibit 1 illustrates the linkage of both macro and micro segmentation variables to the demand for products. Alternatively, the linkages between micro and macro segmentation variables and resulting marketplace demand can also be demonstrated by Exhibit 2.

METHODOLOGY

Sample Design

Respondent households were randomly selected from the publication Daily Reporter, published in Columbus, Ohio on weekdays. This publication prints the names and addresses of those who have recently engaged in Franklin County, Ohio house transactions as well as those who have recently been issued building permits.

Franklin County, and Columbus, Ohio are both frequently regarded as representative "middle America" locales. Indeed. the Wall Street Journal recently indicated that Columbus is a very highly regarded test market city for mass market products. Additionally, Franklin County provides for the replication of an urban environment which was the focus of both the Andreasen and J.E. Bell inter-urban mobility studies. Homeowning respondents who recently acquired a Franklin County building permit with a declared value of 51,000 to $10,000 were selected from the Daily Reporter. For tax reasons, recipients of building permits are suspected of understating the actual value on these permits. On the other hand, building permits are also issued to builders -of new houses as well as to those constructing or expanding/improving commercial properties. With both these considerations in mind, Franklin County real estate officials suggested the building permit dollar range of $1,000 to $10,000 of declared value as a pragmatic guideline in the selection of Franklin County households who have engaged in remodeling activity, in part, as an alternative to moving.

EXHIBIT 2

MICRO AND MACRO SEGMENTATION VARIABLES AND RESULTING DEMAND

Homeowning respondent households who had recently moved were also selected from the Daily Reporter. Commercial property mortgage and deed transactions were eliminated, as were those listings which involved financial transactions, rather than actual household movements. Daily Reporter mortgage and deed transactions do not include pre-move addresses of the households. This made it impossible to distinguish inter-urban from intra-urban mobiles in the Daily Reporter listings. Thus, long-distance vs. short distance mobiles were self-identified.

Renovating and mobile Franklin County homeowning households were selected from Daily Reporter listings which appeared between mid-April and mid-May, as well as between mid-July and mid-August, 1982. This corresponds with Andreasen's and Bell's sampling patterns and also coincides with the traditionally heavier rates of national mobility during the second and third calendar quarters. This time frame provided a potential census of 803 respondent homeowning mobile households and 307 homeowning renovating households for a mail survey. As the survey occurred early in 1983, about four to nine months had elapsed from the time of the Daily Reporter listings. This allowed time for respondents to settle and react to their decisions, and at the same time the interval is thought to be short enough to minimize diminished consumer recall.

The mail survey included a personalized cover letter carrying the signature of the researcher. It also provided the names, addresses, and telephone numbers of researcher's dissertation committee members including both an Associate Dean and the Chairman of the Finance Department and Director of the Center for Real Estate Education and Research, (CREER) from the College of Administrative Science, at The Ohio State University.

The pre-tested questionnaire was preceded by questions ensuring that respondent households did live and own the address where the questionnaire was sent, and the move or remodeling, had occurred in the previous nine months. A self-addressed return envelope was provided, as was the promise of a household gift to respondents returning completed questionnaires. Telephone calls were made to the respondents by female undergraduates to announce the ma lout; reminder telephone calls were also utilized as boosters.

Utilizing the durations and guidelines presented above, a total of three hundred (300) usable questionnaires were received from the census of Franklin County mobiles and renovators. This was based on a mailout of 1110 questionnaires, 307 to renovators and 803 to mobiles. Returns were distributed as follows:

Data Analysis and Results

A key element affecting the measurability of inter-urban and intra-urban moving households are the bounds applied. The current study divided the two groups of mobiles on the basis of whether the move was made within the studied county (in ra-urban) or from outside the studied county (inter-urban). The U.S. Bureau of the Census (1983) groups household moves as to: same county; different county; same states; different states; same region; different region and abroad. The census mobility data is also categorized by various demographic characteristics. The availability of this census data argues for utilizing such politically defined geographic units. The availability of country records, such as the ones used for this study (see Method above), further supports this rationale.

"Within county" seems a logical definition for the local, or short distance, or intra-urban mobile households. If these mobiles can be shown to have homogeneous demand patterns, the number of such households can easily be measured. It might be necessary to aggregate mobiles from contiguous counties within a Standard Metropolitan Statistical Area (SMSA) to determine the boundaries of demand.

TABLE 1

SURVEY RETURN RATES

Households which moved into the studied county, from outside this county, defines the long-distance, or inter-urban mobiles in this study. As in the discussion of intra-urban mobiles, it is noted that segments could be defined to distinguish between households moving between adjacent or contiguous counties and those moving from non-adjacent counties within the same state. Aggregation would continue as long as market response remained homogeneous, and the increments did not significantly affect the criterion. In so doing, the mobility/nonmobility base might conceivably contain a number of unique long distance mobile segments, ranging from cross-county to cross-country.

In this study, only homeowning mobiles are studied along the mobility/nonmobility segmentation base. In part, this is to reduce the high level of transients among homeowners in major university neighborhoods, such as that of Columbus, Ohio and the Ohio State University. However, this was also done as it was thought that homeowners make for more opportune market targets. Further, it also facilitates the inclusion of nonmobile, renovating households, as virtually all of this group would be expected to be homeowners.

It may be desirable to subdivide mobiles and renovators in alternate ways and test for the possible existence of several, rather than just one, viable segment with each group. The noteworthy demographic, and life-sty e, differences between the categories of mobiles are summarized in Table 2.

It is pointed that these results act as a test of Andreasen's conclusions regarding the demographics and life-styles of long distance mobiles (see p. 3). Table 2 demonstrates that long distance mobiles are: 1. relatively young, and in the stages of family life cycle (Characteristics A and B); 2. well educated (Characteristic C); 4. have above average incomes (B); 5. in higher social classes (B, C); 7. upwardly mobile (F); 8. motivated to move for economic reasons (H).

It is also noted that the current study extends Andreasen's work by providing additional descriptor variables for long distance mobiles (Characteristics D, E, G, I, J, K).

Demand Characteristics of Nlobility Segments

A key segmentation criteria is homogeneity or congruity. Two measures are used as a means of determining whether the three studied groups can be said to be homogeneous segments. The two measures of absence or presence of homogeneity of marketplace response are post-move, and post-renovation (within one year), combined planned and actual purchases in eight (8) product categories; and (2) switches in retail patronage patterns in fifteen (15) retail categories (see Table 3). These responses were chosen as they represented many of the products, and retail categories studied by Andreasen (1966) and Bell (1967).

The volume of combined actual and purchase intentions was greatest for short distance mobiles, then long distance mobiles, and least for renovators.

The retail patronage change magnitudes were greater for long-distance mobiles, than for short distance mobiles, and least for renovators. Table 3 presents the results.

DISCUSSION

Testing of Existing Theories:

Andreasen's tentative theory of geographic mobility suggests that demographics and life-styles of long-distance mobiles differ from those of short-distance mobiles, and others. This theory would also seem to hold that such marketplace responses as actual and planned purchases, and retail patronage shifts, would be greater for the long-distance mobiles than for the others. The present study tested this theory, in a contemporary marketing environment. It appears that the demographics, and life-styles, of long-distance mobiles do differ from those of the short-distance mobiles, and others, in this case, renovators (see Table 2). The long-distance mobiles also differ from both groups as to retail patronage changes. However, these mobiles do not differ from renovators with respect to actual and planned purchases. Also, the volume of actual and planned purchases of short-distance mobiles is actually greater than for the long-distance mobiles (see Table 3).

TABLE 2

DEMOGRAPHICS AND LIFE STYLE CHARACTERISTICS OF RENOVATORS, LONG DISTANCE MOBILES, AND SHORT DISTANCE MOBILES CURRENT STUDY

TABLE 3

HOMOGENEITY/CONGRUITY OF MARKETPLACE RESPONSES

Roseman's displacement theory would lead to the hypothesis that the most actual and planned purchases would result for long-distance mobiles, the next most for short-distance mobiles, and the least for renovators. This would also be the order for the hypothesis regarding retail patronage change. In the current study, the latter hypothesis (retail patronage) is born out while the former (purchases) is not, just as for the tentative theory of geographic mobility. Thus, actual and purchase intentions are perhaps not a matter of "displacement" any more than can they be explained by the tentative theory of geographic mobility.

New Findings:

Table 3 demonstrated that the pairwise tests between the two groups of mobiles held up for both of the marketplace responses. This is the clearest finding: long distance and short distance mobiles differ not only in their demographics, but also in their marketplace responses. However, the magnitude of combined actual and planned purchases was actually less for the long-distance mobiles, than the short distance mobiles. This is the opposite of what Andreasen expected. The retail patronage shifts were in the order that would have been predicted by Andreasen and Bell.

Future Research Directions:

While descriptors or the demographics and lifestyles of consumer groups (or potential segments), may differ, this may not produce significantly divergent marketplace behavior. While Andreasen and Bell assumed such a linkage between descriptors and response, their work represented an early segmentation investigation. The body of knowledge has grown considerably since then. Segmentation studies which test the linkage between descriptors and marketplace responses are more likely to be welcomed by practitioners.

The current study also underscores the need for the monitoring of segments as they are not static over time (e.g. the recent emergence of "Yuppies"). Too, the study demonstrates that segments are not a monolith. This is, the likelihood is that a "true segment" (measurable, accessible, profitable, congruous) is not unique in all marketplace situations. Thus, marketers must determine when it is appropriate, and when it is not appropriate, to combine or aggregate consumer groups. The present study illustrates that in some marketplace settings, as many as three segments exist along the mobility/nonmobility homeowning base, but there were sometimes less than three. Correspondingly, it is reasonable to expect that there could also be situations in which more than three segments exist along this base, depending on how aggregated. This needs to be explained.

Limitations and Conclusions:

The two marketplace responses studied are not sufficient to suggest that results can be generalized across all marketplace responses, nor even across all product and retail categories. Indeed, results would likely differ if other product. or retail types were tested.

Also, more far reaching managerial implications would have resulted if more traditional (e.g. convenience, shopping, specialty) marketing categories were tested. For example, it might have been more meaningful if such alter late categories of products as convenience, shopping, and specialty had been tested rather than the diverse product types represented in both the Andreasen and this study.

BIBLIOGRAPHY

Andreasen, R. R. (1966). "Geographic Mobility and Market Segmentation," Journal of Marketing Research, Vol. m, November, 1966, 341-348.

Bell, J. E., Jr. (1967). An Analysis of the Decision Process Utilized by Long Distance Mobile Families in Selecting New Sources of Supply for Goods and Services, Ph.D. dissertation, Michigan State University.

Bell, J. E., Jr. (1969). "Mobiles - A Neglected Market Segment," Journal of Marketing, Vol. 33, April, 3744.

Carson, R. N. (1982). "Do-it-yourself Dream Built by Board," Columbus Dispatch, January 17, J1.

Churchill, G. A. (1979). "A Paradigm for Developing Better Measures of Marketing Construct," Journal of -- Marketing Research, Vol. XVI, 64-73.

Corvallis Gazette Times (A) (1986). "$80 Billion Spent on Remodeling," May 8.

Corvallis Gazette Times (B) (1986). "Employees Move Around More in '85," May 11.

Current Population Reports (1984). U.S. Department of Commerce Bureau of the Census, Ser P-20, No. 317, January 1.

Current Population Reports (1987). U.S. Department of Commerce Bureau of the Census, Ser P-20, No. 318, January 1.

Dickson, P.R. (1982). An Interactionist Study of Buyer Behavior, Ph.d. dissertation, University of Florida.

Gotzko, John J., Jr. (1983). Geographic Mobility in an Urban Environment: Impact of Life-style Economic and Corporate/Organization Policy Variables, Ph.D. dissertation, The Ohio State University.

Kotler, P. (1980). Marketing Management Analysis Planning and Control, Fifth Edition, Prentice-Hall, Inc.

Quigley, J. M., and Weinberg, D. H. (1977). 'IntraUrban Residential Mobility: A Review and Synthesis," International Regional Science Review, Fall, 441-66.

Roseman, C. C. (1971). "Migration as a Spatial and Temporal Process," Migration, 589-598.

Smith, H. C., Tachappat, C. J., and Racster, R. L. (1981). Real Estate and Urban Development, Third Edition, Irwin.

Wall Street Journal (1987). "Record Number of Moves Posted in U.S. for 84-85."

----------------------------------------