The observed change in population trends in the past decade, if they persist, will have a profound effect on transportation problems. It may well be that a reconsideration of transportation problems will be necessary, since many such problems now with us could disappear or be reduced in importance, while others, not now apparent, may emerge.
This piece is concerned with patterns of change in population which will affect many aspects of life, but are reviewed here primarily in terms of their impact on urban transportation. For the most part it is a review of trends in total population growth, and in the distribution of population by locale. Some of the major points I wish to make at the outset are: (1) There has been a rather long downtrend in population growth, and population forecasts may well remain too high; hence, neo-Malthusian concerns seem grossly overstated. (2) Growth of large population centers has slowed appreciably, so that concern about large urban size also seems overdrawn. (3) High growth areas include Florida; the Pacific southwest, exclusive of very large metropolitan areas; the West, more generally; and middle-sized areas in Texas. If we look at these high growth areas statistically, we find a clear inverse relation between size of place and rate of growth, and an absence of other regional growth differentials for the rest of the south and for the northeast regions. Because the south has a disproportionate number of small metropolitan areas and the northeast a disproportionate number of large areas, the inverse relation of population and growth explains the observed growth differentials for these regions. (4) The population shift from central cities to suburbs is continuing, but there may be something of a historical reversal of migration streams so that there are now net population flows from Metropolitan to nonmetropolitan areas. (5) There is some evidence that suburban sprawl is approaching a limit, and that there will be increased density in many urban areas experiencing growth. (6) The slow-down in total growth, the shift away from large areas, and the shift to relatively uncongested areas with good highway facilities suggests the need for less investment in transportation facilities than has been anticipated. In sum, almost all these changes appear to move in the direction of mitigating urban transportation problems.
Since 1960, the annual increase in U.S. population has shown a rather steady decline, dropping from around three million in 1960 to half that figure in 1970. There was some reversal of the trend in the period 1968-1970, but the decline was reestablished thereafter, as shown in Table 1.
If a trend line is mechanically fitted to the data by least squares, as in Figure 1, zero population growth (ZPG) is reached in fairly short order. A linear fit yields a ZPG date of 1989, while a logarithmic function, with a bit more explained variance, moves the date to 1993. In each case, maximum population is only 226 million, given a January 1974 population of 41.1 million. The mechanical fit might be viewed as a limiting case, since it seems plausible that the downtrend will taper off somewhat. The Census Bureau, recognizing the trend, has made substantial reductions in its population projections, reflecting a sharp decline in birth expectations of young wives during the past five years. Yet there are some indications that the Census Bureau has not gone far enough in recognizing the trend. In previous projections, average number of births per woman upon completion of childbearing were alternatively set at 3.1, 2.8, 2.5 and 2.1, to yield projections respectively labeled the B, C, D and E series. Late in 1972, the Bureau dropped the B series, and added an F series, which assumed 1.8 births per woman. For the year 2000, the midpoint population projection for the initial set of alternatives was 296 million. For the revised series, the midpoint projection now is 275 million. The E series corresponds to eventual ZPG, for 2.1 births per woman is the exact replacement level; if that level held indefinitely, population would stabilize at 320 million in 70 years. But actual experience in 1972-1973, as shown in Table 1, squares best with the F series, which assumes a level of births below the replacement level. For the year 2000, the E series projection is 264 million, while that for the F series is 250 million. In the F series, projected annual growth first increases to 1.8 million and then declines to around 1 million by the year 2000. Peak population is around 270 million. Thus, the F series decline is not nearly as drastic as that obtained by straight-line projections, but nonetheless represents a considerably reduced rate of growth, relative to recent expectations.
Some neo-Malthusians expect a growth turnaround, or even dispute the Census figures; but this appears to be taste for a battle that has already been won. Given the Pill and abortion on request, a marked reversal of trend seems unlikely. In retrospect, the furor over ZPG in the United States seems somewhat surprising in the light of the downtrend of Figure 1. Perhaps both ideology (ZPG enthusiasm) and mores (changing lifestyle) follow technology (improved birth control).
The deceleration in population growth has led to overcapacity in such areas as pediatrics and elementary education, and similar problems are conceivable in transportation. However, the considerable lag between time of birth and time of independent travel should yield enough lead time to adjust transportation planning to changed circumstances. It must be added that the reduced birthrate should add more women to the labor force, and thus increase trip-making for a considerable period. Increased participation of women in jobs outside the home has increased the labor force and the number of journeys to work, and for many families the additional income obtained makes a second car feasible. In fact, John Fisher, in a background paper for a National Academy of Sciences Forum on Energy (cited in New Scientist, February 28, 1974, p. 561), suggests that working women are the major cause of the past decade's increased energy demand.
Concern about population distribution generally focuses on urban growth and the possibility that cities are too "large"; some ZPG enthusiasts shifted their interest to such topics when declining national growth became evident. Here again, trends appear to be preempting the problems. The trends include: (1) little growth or even declining population in the largest metropolitan areas; (2) more generally, an inverse relation between growth and size; (3) a continuation of shifts from the older northeastern and north central urban areas to those in the south and west, which typically also involves a shift to places with better highways; (4) continuation of the shift from central cities to suburbs; but (5) some apparent reversal of internal migration flows from non-metropolitan to metropolitan places.
However, most immigrants from abroad settle in metropolitan areas, countering the internal shift somewhat.
According to recent census data, the population had this estimated percentage distribution in March 1970 and in March 1973:
The increase for the nonmetropolitan percentage contrasts with a previous long-term decline. Thus, between 1960 and 1970, the nonmetropolitan population as a percentage of the total declined from 33 to 31 percent. If the estimated March 1970 and March 1973 percentage distributions are taken as given, and applied to estimated total resident populations for the respective periods, the estimated distributions of number of people, in thousands, is:
The central cities show a decline in total population of 3.3 percent, while the "suburbs" (balance of SMSAs), grew somewhat faster than the remainder of the country (7.7 percent versus 4.2 percent). These percentages contrast with respective percentage growth between 1960 and 1970 of 5.3 percent, 28.2 percent, and 6.5 percent for the three locales. The increase in central city population in the 1960-1970 period may well reflect annexations rather than growth within city limits as defined in 1960. (Cities that annexed no territory showed a small population decline in the period.) Hence, the decline in central city population can well be viewed as a continuation of earlier trends. However, the relative shift to nonmetropolitan location seems substantial; if not a reversal of earlier trends in population growth, it does represent a considerable shift in relative strength of those trends.
Much of the nonmetropolitan growth probably involves a form of urbanization and reflects (1) growth in the "far suburbs" around existing metropolitan centers, but outside the county lines defining SMSAs, (2) development of industrial complexes at intersections of major highways at some distance from metropolitan areas, and (3) expansion of service industry employment in rural counties, possibly a variant of category (2). In all cases, improved highway transport is probably a major underlying cause for the shift.
Some perspective on regional migration flows can be obtained from Table 2, which shows the ratio of in-migrants to out-migrants for recent interstate migration flows—people moving from one state to another—classified by region and locale. Central cities in all regions had net outflows of migrants between March 1970 and March 1973, though the change was most pronounced in the northeast and north central regions, with only limited net losses in the south and west. Suburbs in the first two regions had net outflows, but there were net inflows of migrants in the south and west. For nonmetropolitan areas, however, all regions except the northeast had inflows.
For all locales, the ratio of in-migrants to out-migrants was lowest in the northeast. It was highest for central cities and nonmetropolitan areas in the west, and for the balance of SMSAs in the south. For regional aggregates, the order was south, west, north central, and northeast, for both the ratio and the total net inflow of migrants. It is plausible that the same pattern, albeit in less pronounced forms, holds for total population changes.
Applying some recent census information on individual metropolitan areas, we can pick out patterns of recent population change for those areas. From April 1970 to July 1972, the population of all SMSAs combined grew 2.18 percent compared with nonmetropolitan area growth of 2.27 percent, which squares with the pattern discussed above. It is noteworthy that growth for the fifteen largest SMSAs, with populations over 2 million, averaged less than 1 percent, with small declines registered for five of the group (New York City, Los Angeles, St. Louis, Pittsburg, and Cleveland). For almost all of the others, growth was less than the U.S. average. Only Washington, D.C., grew by a shade more than 3 percent. In contrast, 52 of 235 smaller areas grew by more than 5 percent.
Metropolitan areas in three regions grew at a significantly faster rate than those in the rest of the country. The regions, with 13 SMSAs in each, were: (1) Pacific southwest "desert and shore," consisting of Arizona, Nevada, and California SMSAs below 2 million population, but excluding the California Central Valley SMSAs, which typically had somewhat lower growth rates; (2) all Florida SMSAs, save for Melbourne-Titusville-Cocoa, which lost population because of cutbacks in the space program; and (3) Texas metropolitan areas in the "middle range" of population size, ranging from 150,000 to 2 million in 1970 population. When we have accounted for the high growth rate of Florida and Texas, the remainder of the south shows no significant difference in growth from the rest of the country. The west, however, does exhibit a higher regional growth rate.
Some detailed analyses of the individual area statistics were carried out by calculating the ratio of 1972 to 1970 population for each SMSA and then relating the ratio to population size, both for the three high-growth regions as a group and for the remainder of the country. Six population size classes were set up, and the average value of the 1972 to 1970 ratio is shown in Table 3 for each class. An inverse relation between size and growth rate occurs for both the fast-growing group and the remainder of the country, though this is most evident for the latter set of cases. The significance of the relationship was confirmed by statistical test.
The results can be summarized as follows: In the United States, rapid growth tends to be localized sunny regions, presumably with high amenities; however, with increasing size there is an increasing drag on growth. Both relationships may involve migration of retired persons to pleasant places (the sunny coasts) and/or to places with lower cost of living. Since cost of living is higher in the north and increases with size of place, people on fixed incomes have higher real incomes in the south, and their real incomes increase the smaller the locale in which they live. Assuming that similar migration patterns also hold for labor force participants, it can be inferred that real "net" wages are higher in areas with highest rates of growth. It seems plausible that the level of highway services is a factor in both the regional and the urban size pattern of growth. The larger the urban area, the greater the cost of operating a car (e.g., parking costs, insurance, and accident costs not covered by insurance), the longer the journey to work, and the greater the traffic congestion and pollution. Such costs may help limit the growth of large areas. Further, the newer urban areas of the south and west typically have much better highway networks than those of the northeast and north central regions.
A number of conclusions can be drawn from the examination of these trends. First, the problems of growth, at a national level and in terms of the burgeoning of "too large" urban areas, appear to have been grossly overstated. At the national level, in fact, it is likely that lack of growth will eventually be seen as a problem, giving rise to concern about labor "shortages." At the level of the individual urban area, there is evidence to support the contention that there are eventual diseconomies of scale, in terms of increased congestion and pollution effects. But people can and do vote with their feet, and move to more pleasant—and generally smaller—places.
The shifts in population can be viewed as part of an equilibrating process, alleviating congestion problems in dense areas and making them somewhat worse in places that are presently less dense and congested, but with net gains for the system as a whole.
The slowdown in total growth and in growth of large areas and the shift of population to less congested places, both in terms of urban size and regional distribution, should indicate a considerable slowdown in highway construction "needs." The Department of Transportation has estimated a "need" in 1990 for approximately 18,000 miles of additional freeways and expressways within urbanized areas, compared with about 8,000 miles in 1968. That estimate may well be too high if earlier growth patterns were used in its development. In sum, present population trends suggest a future in which there is a considerable reduction in pressure for solutions to urban transportation problems.
A version of this article appeared in print in the June 1974 issue of Resources magazine.
