Forces Behind the Nonmetropolitan Population
Growth Rebound of the 1990s

Calvin L. Beale
Economic Research Service
U.S. Department of Agriculture

The current trend of renewed growth in the nonmetro population has been rather well publicized by now, having been reported by major newspapers and magazines. I want today to update it, since we have new data from the Census Bureau, and emphasize what I believe to be its main features, plus a couple of its least reported aspects.

The basic event we are following is one in which 3/4 of the country's 2,300 nonmetropolitan counties have increased in population since 1990, after less than 1/2 had done so during the extended farm crisis and general rural economic recession of the 1980s.

I should note that the basic criteria for defining metropolitan (and by implication nonmetropolitan) have become rather complicated. They are established by the Office of Management and Budget, and not by the Census Bureau as is often commonly thought. In general, a metro area is defined along county lines and must have an urban nucleus of at least 50,000 people. Additional counties are added to the area if they meet certain criteria of worker commuting and metro character. All other counties are nonmetropolitan.

From 1990-96, nonmetro counties had an overall population increase of 6 percent compared with 7 percent in metro areas. Not much of a difference. By contrast, the metro areas saw growth in the `80s at a rate 4 1/2 times that of nonmetro communities.

The key to the recent nonmetro increase is that half of it (about 1 1/2 million) has come from net inmovement of people from metro places. Another 10 percent stems from direct foreign immigration. Only 2/5 of it has been produced by natural increase -- the surplus of births over deaths. The metro areas saw somewhat faster growth, despite their migration losses to the smaller places, because of their much wider margin of natural increase and their disproportionate receipt of immigrants. (If data for just two States - California and New York -- are removed from the results, the metro balance sheet would show domestic inmovement as large as the outmigration implied here).

All broad economic types of nonmetro counties have shared in this rebound --farming, mining and manufacturing counties, plus those dependent on government, services and trade, or having unspecialized economies. But not equally.

Farming and mining areas have had the lowest overall growth (4% and 3%) and are the most likely to have continued to decline. Even these modest levels of increase have required some net inmigration to attain, and where growth has occurred it often has been from sources outside of farming or mining.

Counties that are retirement destinations and/or recreation areas, and counties adjacent to metro areas have the greatest propensity for population growth. The 190 or so nonmetro counties that we have typed as retirement areas increased by 16% from 1990-96. But with their general attractiveness, and the jobs created by retirees, by far the major part of their growth was among persons under age 65.

Counties with 1/3 or more of their land in Federal ownership were also a high growth type, with 14% increase. Some of this is related to recreation or retirement and such growth extended over time adds a dimension of its own to the complex issues surrounding Federal lands and their use.

Regional Aspects

The geography of population change reflects these patterns. As shown on the map, areas with above average population increase are very common in the Mountain West. Much of this territory is still thinly settled, but new growth is rapid enough to be noticeable and the character of many places is changing as a result. Elsewhere, the Upper Great Lakes and Ozarks recreation/retirement districts continue to show above average increases, as do the southern Blue Ridge Mountains counties, northern Florida, and many communities that adjoin thriving metro areas.

Areas that have declined in population sine 1990 are most prevalent in the Great Plains and adjoining parts of the Corn Belt, where continued losses in farm employment have not yet been offset by other job growth. The only significant grouping of declining counties elsewhere is in the lower Mississippi Valley, especially in the Delta. Here as in most of the Plains and Corn Belt farming areas, declines were typically modest and well below those of the 1980s. It is remarkable to note, though, that the Farm Belt has some counties that have declined in every census since 1900 and have continued to do so through 1996. This illustrates how very lengthy the adjustment process can be to continually falling labor requirements in agriculture, unless other sources of employment are developed.

The eastern half of the country is the most likely to have had population growth at low to average levels of less than one percent annually. Such counties often have major dependence on industrial work, even if there is also a farm base, and lack either the widespread amenity attraction of the West or the sparse settlement and farm and ranch dominance of so many of the declining places.

Natural Decrease

One manner in which nonmetro pop change has become widely different from metro is in the rising incidence of natural decrease -- deaths exceeding births. Back in 1950, there were only 2 counties in the whole country with more deaths than births (both semi-abandoned Western mining areas). From 1990-1996 there were 615 in nonmetro areas, and only 28 in metro areas. A phenomenon that in my early days in demography was associated only with France, Austria, and San Francisco has become commonplace in America at the local level. Childbearing among most rural and small town people has declined to metropolitan levels, roughly just adequate for generational replacement or somewhat below it. But in addition, counties that are or have been dependent on agriculture have lost so many young adults over the years that those present now are not numerous enough to bear sufficient children to offset deaths among the larger older population.

Some of the areas with natural decrease are beginning to have inmigration of working age adults now. If this persists, it will, over time, reduce the age distortion that is a major factor in natural decrease. But, the natural decrease counties did not reach their current age profile overnight and most are unlikely to emerge from it quickly. We now have some farming areas whose residents are declining in number now both from further outmigration and from more deaths than births, including a few in this region.

Change in the Growth of the Older Population

The aspect of the current rural demographic trend that has surprised me the most in this decade is the change in the number and proportion of people 65 and over. The national population 65+ continued to grow faster than the rest of the population from 1990-96 (a growth of 9% versus 6.3% for the population under 65). But, the opposite was true in nonmetro America, with the older population growing just 5.5% vs. 6% for the younger group. This is in sharp contrast to the pattern in the 1980s, when the older population had a decade growth of 15% versus less than 1 percent for the under 65 class. Despite the rapid increase of older people in rural retirement areas in the 90s, nationally the current rural growth rebound has occurred only among persons under 65.

Fully 1/3 of all nonmetro counties are estimated to have had declining older populations since 1990, and the number with a diminished proportion of older people is much larger than this.

I believe this trend stems principally from the heavy past depletion during youth of cohorts now reaching 65, as rural young people moved away to the cities in the 1940s or gave up farming in the 50s. The burden of elderly dependency has thus already started to lessen in many rural areas. And this is in advance of the more widespread trend now beginning as the small birth cohorts of the 1930s reach 65. And except in rural areas now having substantial population growth, the rapid national increase in the elderly expected as the baby boom generation reaches its sixties will be somewhat muted in rural communities that are not retirement destinations.

Factors Producing the Rebound in Nonmetro Growth

After this excursion into fertility and age matters, let me return to the general growth of the nonmetro population and the forces behind it. In an effort to get beyond the inferences that can be drawn from the statistics I have been reporting here, I have talked with informed people in a number of counties that (1) have reverted from 1980s loss to 1990s growth, and (2) have migration as the source of most of their increase.

These inquiries have surfaced many instances of employment growth, as one would expect. Many of these relate to manufacturing -- a new candy factory here, meat packing plants here and there, an electric weed cutter factory or its equivalent somewhere else. But we still face the fact that despite the major role of manufacturing in the rural and small town economy, it is not a growth sector nationally, having 800,000 fewer workers than it did in 1990. Local gain may represent some other place's loss or only a partial recovery of closures in the 1980s.

In a number of counties the most conspicuous rural growth industry is prisons - the punishment of largely urban crime. This isn't totally new. It was true of the 1980s, but has been more intense in the 90s. In a sad commentary on the times, I can identify at least 55 nonmetro counties that moved from population decline in the 80s to growth in the 90s either solely or largely as the result of acquiring a prison. More than half of all new prison space is being placed in nonmetro settings, although just a fifth of the total U.S. population is nonmetro. Inmates become counted in the population, but there are also typically about 3 direct new jobs for each 10 inmates. A particularly notable example is occurring in eastern Oregon at Ontario where a modest-sized facility that opened in 1991 is expected to have 3,000 inmates by mid-1998 with 1,100 staff, plus 60-90 teachers. The economic and residential effect of this growth is expected to extend well into nearby Idaho.

But as often as not, local informants have given me explanations for new or renewed growth that relate to residential preferences, exclusive of retirees. Direct surveys on residential preference (such as those done by Glenn Fuguitt of the University of Wisconsin over several decades) have consistently shown a large reservoir of people not living in rural and small town areas who say they would like to do so, whether from positive attraction to such places or discontent with urban conditions.

The growth I'm thinking of takes several forms:

  1. Influx of commuters. Many nonmetro areas that are metro adjacent are being clearly affected by outward sprawl from these urban centers. Some of it is incipient suburbanization. Some of it is more distant in location and still very rural in nature. All told, metro-adjacent counties acquired 2/3 of all the net migration into nonmetro areas, although the degree of change from the 1980s was greater in the nonadjacent group.

    But I've had more reports than I expected of dispersal of workers out from nonmetro employment centers to neighboring counties. This type of dispersal does not contribute to overall nonmetro growth, but does distribute it over a wider area. I think of examples such as Butte Co., S.D. where the gambling boom in Lawrence Co. (Deadwood) has limited space and increased costs enough to impel many new workers to live in other communities (a phenomenon also common to some of the West's other recreation towns); or Gosper Co., Neb., where effects of the huge meatpacking plant in Dawson Co. (Lexington) have persuaded some of that county's residents to move out and commute back; or Lewis Co. Tenn, which attracts some workers from GM's Saturn plant in Maury Co. who want a rural place;


  2. But then there also frequent perceptions of nonmetro growth in the absence of commuting opportunity, or any commensurate local job growth, or linkage to elderly retirees. This is not limited to the West, but seems especially common here. Some of it I would have to call a bit low brow. I recall the county extension agent in Idaho who told me of living in a predominantly Quaker community where newcomers from outside openly wore firearms. Most of it is much more middle class or upscale, but all with a common theme of people said to be wanting to get away from urban areas, and -- if parents -- wishing to get their children into the better school environment that they believe small communities have.
So often the description is of people who have sold an urban home for a large capital gain and have moved to an area where they can buy or build one at cheaper cost, and retain a nest egg. (Presumably such a trend would be accelerated if the capital gains tax exemption on home sales is greatly increased as is commonly suggested to be an action this Congress my take.) Time and again I hear the description of newcomers wanting a place on some acreage away from town. As one extension agent in a remote Pacific Northwest county put it to me, "Everyone who comes in wants 10 acres 5 miles from town. If I were to let it be known in the coffee shop one morning that I had such a place for sale, I would have to rush home to get there before the phone began to ring, and the offers would be as likely to come from New York or abroad as from here."

It has become commonplace in the West to attribute much growth to an influx of urban Californians. Quite aside from local impressions, the outflow of people from that State has been so heavy as to lend credence to this. In just the 6 years, 1990-96, there was a net outflow of 1.9 million domestic migrants from metropolitan California. Not more than 2% of this amount was acquired by nonmetro parts of that State. The rest dispersed to other States, with the numerical potential to affect many communities substantially, both rural and urban.

We normally think of population trends as dependent on economic activity, and in those parts of the country where population is declining, I suspect this relationship is still very valid. But there is no doubt about the fact that a number of nonmetro countries with above average recent population growth seem to have acquired it with little or no employment growth, but rather by newcomers using savings, income from investments, or transfer payments. Nationally, employment as grown at a modestly higher rate than population since 1990, but in a third of nonmetro counties -- many of them in the West -- employment has risen faster than employment.

It is also common to have informants volunteer mention of the inmovement of people who have computer- enabled employment, whether just of themselves or as entrepreneurs with employees. Yet their motivations for moving seem to be largely nonpecuniary, oriented toward quality of life.

Conclusion

In sum, the nonmetro population growth of the 1990s is not as spectacular as that of the 70s when the nonmetro growth rate was higher than the metro one in nearly every year. It has only been so once in the 90s. But the change from the 80s is very real and widespread.

In the short run, I would not be surprised if the nonmetro growth of the 90s has already peaked. July 1994- 95 is the only year in which the nonmetro growth rate exceeded that in metro areas, and the inmovement is estimated by the Census Bureau to have been lower in 1995-96. The attractiveness of nonmetro areas was enhanced in the first part of the decade by the recession that was metro dominated. This led metro employment to grow at a lower rate than nonmetro jobs from 1990 through 94. We know from a couple of data sources that nonmetro outmovement was reduced in this time and metro outmovement rose. But by 1995 the period of higher nonmetro employment growth ended, and since then the pace of metro job growth has revived to a level somewhat above that in nonmetro areas. This is bound to have an effect, but the net flow of migration still remains toward rural and small town areas, even though at a lower pace.

In the longer run, I think we have reached a point where under average economic conditions nonmetro America will continue to receive net migration from metro areas, and where much of this will be driven by quality of life considerations. With just 1/5 of the U.S. population now in nonmetro communities, only conditions of exceptional rural economic stress -- such as those of the 1980s -- can produce enough outmigration to metro areas to offset in absolute terms the movement of a modest rate of inflow from the massive metropolitan population, stimulated mostly by nonpecuniary objectives.