CHAPTER 23 WHAT WILL FUTURE POPULATION GROWTH BE? Chapter 23: Table of Contents: Voodoo Forecasting Do Current Trends Call for Coercive Policies? Who Will Support Whom? The Dependency Burdens Decreasing Population Density in the Future? Conclusion When looking at the demographic facts with an eye to population policy, one wants to know what the future holds - how great the "pressures" of population size and growth will be. Hence governmental agencies and academic researchers make forecasts. The history of demographic predictions calls for humility, however, and teaches caution rather than fear-motivated over-reactive policies. Voodoo Forecasting About 1660, with tongue in cheek William Petty extrapolated population trends to show that by the year 1842 the population of London would about equal the population of England. In the 1930s most Western countries worried about a decline in population growth. The most extensive investigation of the "problem" was undertaken in Sweden by some of the world's best- regarded social scientists. The dotted lines in figure 23-1 show how the future looked to them then. But all their dotted-line hypotheses about the future - intended to bracket all of the conceivable possibilities - turned out to be far below the actual course of population growth, as shown by the solid line; that is, the future turned out far better, from the point of view of those scientists, than any of them guessed it might. It may well be that we are now at an analogous point with respect to the world as a whole, except that now population growth is popularly thought to be too fast rather than too slow. FIGURE 23-1. The Population of Sweden According to Four Hypotheses Made in 1935, and the Actual Population in 1979 (old 11-10) FIGURE 23-2. U.S. Population Forecasts Made 1931-43, and the Actual Population (old 11-11) The Swedes were not alone in making inaccurate "pessimistic" forecasts. A U.S. Presidential Research Committee of eminent scientists reported to Herbert Hoover in 1933 that "we shall probably attain a population between 145 and 150 million during the present century". Figure 23-2 shows a variety of forecasts made in the 1930s and 1940s by America's greatest demographic experts. For the year 2000, none of them forecast a population even as large as 200 million people in fact, the U.S. reached 200 million people sometime around the year 1969, and is far beyond that already. A good many of the forecasters actually predicted a decline in population before the year 2000, which we now know is impossible unless there is a holocaust. The forecasts just mentioned erred by extrapolating the recent past. Another blunder is fitting the data to the wrong ideal mathematical form, among which the logistic curve has been the most dangerously seductive; it assumes that population will level off. Figure 23-3 shows how Raymond Pearl, one of the greatest mathematical demographers, went wrong. And the graph is coming to us with the approval of Walker Shewhart, the greatest of quality- control statisticians. Figure 23-3. Shewhart The accuracy of long-run forecasts has not benefited from the huge errors of the 1930s. Figure 23-4 shows how forecasters went wrong in 1965; they extrapolated a large increase for 1990, but a decline actually occurred. Figure 23-4 [from Economist and Tom Poleman] Even as late as the 1970s there were astonishing flip-flops in world population forecasts for the turn of the century - then only three decades away. As of 1969, the U.S. Department of State Bulletin forecast 7.5 billion people for the year 2000, citing the United Nations. By 1974, the figure commonly quoted was 7.2 billion. By 1976, Raphael Salas, the executive director of the UN Fund for Population Activities (UNFPA) was forecasting "nearly 7 billion." Soon Salas was all the way down to "at least 5.8 billion." And as early as 1977, Lester Brown and his Worldwatch Institute (which has a close relationship to the UN) dropped the estimate again, forecasting 5.4 billion people for the year 2000 - which was surpassed around 1990 (if the data are sound). This variation in forecasts must be astonishing to laymen - to wit, that a United Nations-State Department forecast for a date then only twenty-three years away, when a majority of the people who would then be living were already living, could be later revised by 2 billion people, a change of more than a third of the total forecast. By the time of this writing in 1992, the latest UN "medium" forecast for the year 2000 was 6.3 million. Does this example of forecasting "science" suggest that we should give credence to long-run population predictions? And consider this: In 1972, the President's Commission on Population Growth forecast for the United States that "even if the family size drops gradually - to the two-child average - there will be no year in the next two decades in which the absolute number of births will be less than in 1970." How did it turn out? In 1971 - the year before that forecast by the august President's Commission was transmitted to the President and then published - the absolute number of births (not just the birthrate) was already less than in 1970. By 1975, the absolute number of births was barely higher than in 1920, and the number of white births was actually lower than in most years between 1914 and 1924 (see figure 23-5). Figure 23-5 (old A-15) This episode shows once again how flimsy are the demographic forecasts upon which arguments about growth policy are based. In this case the Commission did not even backcast correctly, let alone forecast well. Then in 1989, the U.S. Census Bureau forecast that U.S. population would peak at 302 million in 2038 and would thenceforward decline. But just five years later in 1992, the Census Bureau forecast 383 million in 2050 with no peaking in sight; one gives the lie to the other by about 50 million people - a sixth of the 1989 forecast. The science of demographic forecasting clearly has not yet reached perfection. With a track record this poor, one wonders why official agencies should make any such forecasts at all, especially when they are based on little more than guesses concerning such matters as future immigration policy and people's decisions about how many children to have. We have no experience to rely on for estimating how people will act under the conditions that will prevail in the future. In short, this history of population forecasts should make us think twice - or thrice - before crediting doomsday forecasts of population growth. (Actually, demographic forecasting is very easy - even movie actresses can do it. Meryl Streep noted that "men were given more than twice as many roles as women in 1989 movies", and then she forecast that "by the year 2000 we will have 13 percent of roles...and in 20 years we will be eliminated from the movies.") Some trusting persons respond to data on past failed forecasts by saying that this shows we need better research, and "the government should do something about it." But "official" forecasts made by government agencies - always at very large public expense, of course - have been shown to be no better than the run of private forecasts. And official forecasts have the grave disadvantage that people are more likely to believe them than private forecasts - and then take legal action against the government if the forecasts are wrong, as the timber industry did in the 1980s when an officially predicted tree shortage never occurred. Not only are methods of population forecasting unreliable, but the base of data used for many countries, and for the world as a whole, is still shockingly weak. In 1992 it was revealed that the populations of several developing countries were much lower than anyone had imagined they could be. Nigeria's population, for example, was said by standard sources to be 122.5 million in 1991. But when the results of the 1991 Census were released - quite likely the best census ever done, by far - the total turned out to be 88.5 million people. Amazing. Do Current Trends Call for Coercive Policies? What population size or rate of growth does the long-run future of the world hold in store? No one can be sure. One frequently hears that zero population growth (ZPG) "obviously" is the only viable state of affairs in the long run. But why? Why shouldn't population get smaller instead of staying level if it already is too large? What is sacred about the present population size, or the size that will be attained if it levels off soon? As one writer put it, the concept of ZPG is simply "a careless example of round number preference." As to whether a larger stationary population, or a larger and still growing population, is plausible or desirable in the long run - the whole of Part II of this book addresses that topic. The mere recitation of a number like "five and a half billion people" can evoke the reaction that the number is very large, or too large, or unsustainable. But no one says that the numbers of trees or birds or bacteria on our planet are too large, or calculates the number of doublings until their space would run out. Why the difference between birds and people? Contemporary data show us that the rate of population growth can go down as well as up. In many poor countries, including the largest of them, fertility has been falling - and very rapidly (see figure 23-6). Many of the countries with the fastest- falling birthrates are small islands, which seem especially quick to respond to new conditions and currents of thought; perhaps it is because they have good communication networks due to their high population density. Hong Kong's crude birth rate fell from 38 to 30 in just the six years from 1960 to 1965. But huge China is no island, and it supports a quarter of all humanity; yet fertility there, too, has apparently dropped sharply in the past decade or two (though the coercion involved makes this case different from other countries). Figure 23-6 [like old **Appendix, figure A-18] The supposed inevitability of population leveling off seems to be an implicit assumption of a scenario like Malthus's two postulated growth rates - exponential for humans, arithmetic for food. But this scenario lacks supporting evidence. Humans and wheat are both biological species, and the growth of each is constrained by various forces. There is no a priori reason why the two species should follow different growth patterns. The recent drops in fertility imply that countries that are now poor and have high fertility rates will sooner or later follow the pattern of the richer countries, whose mortality rates fell years ago and whose fertility rates then likewise fell. This pattern may be seen in figure 23-7, which shows the well-known "demographic transition" as it actually occurred in Sweden. Figure 23-7 (old 11-8) In the more-developed countries, fertility now is low by anyone's standard. In figure 23-8 we see that the birthrate is far below replacement - that is, below zero population growth - for many of the largest countries in Europe. What if those trends continue? Forecasts of population size require assumptions about the fertility of future couples, and about the fertility of present couples who have begun but not yet finished bearing children. Such assumptions have proven wildly wrong in the past, as we have seen. Yet it is interesting to imagine the implication of assuming that childbearing patterns like those practiced at present will continue - population decline in the major Western countries. Already there is a felt "labor shortage" in the Western countries, though it may well be that this is a result of short- term ups and downs of the birth rate rather than a result of the long-term rate. And it might be that because of the long-term growth in income, there might be a felt shortage even if population were to continue growing at a substantial rate in rich countries. This squares with the paradoxical idea expressed in chapter 00 that human beings are the only resource whose shortage is increasing over the decades. Figure 23-8 (like old 11-9) Even if total population continues to grow, there may be declines in some large geographic areas. This is discussed below in the section on population density. As to the long-run future - again, no one knows what will happen. We can expect that people's incomes will rise indefinitely. But how much of that income will people spend on additional children? And what other activities will compete with bringing up children for parents' interest and time? These factors influence population growth, and no one knows well how they will operate. My guess - based on no solid evidence - is that fertility in the rich countries will again begin to be large enough to produce population growth. We can say with confidence, however, that an extrapolation of the last few centuries' world population growth rapidly toward infinity and doom, as seen in the silly figure 22-1, has no warrant in the facts. WHO WILL SUPPORT WHOM? THE DEPENDENCY BURDENS A key issue in the economics of population is the proportion of persons of working age relative to the proportions of child dependents and of old-age dependents. A fast-growing population contains a large proportion of children, who are an economic burden until they are old enough to earn their keep (just like capital investments while they are being constructed). On the other hand, a population that is growing very slowly or not at all contains a large proportion of persons too old to work but who require that others support them. The interrelationships among births, deaths, population size, and age distributions have been changing over the years: 1. Figure 23-9 shows that the total number of persons under 20 has risen little in France in over two hundred years, and now is little higher than in 1881. But the total number of persons 65 and over has multiplied by a factor more than six. Fig 23-9[Chesnais] 2. The total number of deaths each year is no greater than it was in 1950, though the world's population has doubled. Consider the large differences in the child-dependency burdens from one country to another: In 1985, fully 46 percent of the population of Bangladesh (typical of many countries in Asia and Africa) was younger than 15 years old, compared with only 17 percent in Switzerland (typical of many countries in Europe); that is, in Bangladesh there were 91 children for each 100 working-age persons 15-64, while in Switzerland there were 27 children for each 100 persons 15-64. The age pyramids for the two countries are shown in figure 23-10. Clearly, the economic effect of such dependency differences is not trivial. FIGURE 23-10. [Age Distributions in Bangladesh and Switz] Many draw from these data the conclusion that the standard of living will be higher if the birthrate is lower. For the immediate future this is undeniable; the proposition can be demonstrated by the simplest kind of arithmetic: If income per capita is our measure of economic well-being, then we need only divide gross national product by population to calculate income per person. Adding a non-producing baby to the population immediately reduces the calculated income per capita; it is that simple. The implications of this simple arithmetic are less simple, however. Another baby means there is less of everything to go around - for the time being. But the squeezes in schooling, feeding, and housing (the extent of which will be discussed later) bring forth additional efforts on the part of individuals and institutions to mitigate the squeezes. Also very important in assessing the impact of an additional child is the issue of who assumes which part of the burden - the parents or the public. (Later we'll talk more about this, too.) The story of dependency burdens does not end here, however. A modern low-mortality society with a low birthrate supports few children. But each person in the labor force also has a great many old people to support. Consider these variations. a) In 1900 in the U.S., 4 percent of the population was over 65. But perhaps 13 percent of the population will be over 65 in 2000, and 17 percent in the year 2020. b) The proportions of the male population within the prime labor-force ages of 15 to 64 were 70 percent in Sweden in 1940, and 52 percent in Bangladesh in 1985 (that is, there were about twice as many workers in Sweden then as in Bangladesh now to support each dependent). But counting the numbers of people in different age groups does not tell the whole story. The cost of supporting a retired person is much greater in the U.S. than the cost of supporting a child. Least important in a developed society is the difference in food consumption, unlike the situation in a subsistence- agricultural society. Consider that old people need much more expensive health care than do children. And old people may travel for twelve months a year in trailers on public roads, whereas children cannot. Except for schooling, old people consume much more than do children in almost every category of expensive goods and services. This pattern of old-age dependency causes perturbations. The U.S. Social Security system has already been in severe funding trouble, and financing the payments is a serious economic and political problem for the federal government. In the future, the burden of Social Security payments will take up an even larger proportion of a U.S. worker's pay, and of the production of the economy as a whole. Even more fundamentally, the deficit in the U.S. government budget, which troubled so many people so much in the 1980s and beyond, may reasonably be seen as a demographically-caused phenomenon. If people were still dying at the 1900 mortality rate, there would now be no budget deficit, because benefits to the aged would be enormously lower. Or, if couples had continued having children at the fertility rate of the baby boom, there would be more working taxpayers now, and the deficit would not exist or at least would be much lower than it is. (At present, increased immigration is the only feasible solution which would keep the deficit from growing while avoiding the pain of decreased benefits or increased taxes.) Are you now, or will you in the future be, a working person? If so, a lower birthrate means that you have fewer people who will look to you now for support. But the same reduction means that there will be fewer people to support you when you get older, and you will then be a relatively greater burden on others. This two-edged sword involves one of the major themes of this book. The short-run effect of a given demographic factor is often the opposite of its effect in the long run. Deciding which demographic pattern you prefer - faster or slower population growth - requires that you put relative values upon the long-run and short-run effects. And this, of course, requires that you decide who is to contribute and who is to receive. DECREASING POPULATION DENSITY IN THE FUTURE? If one divides the number of square miles on the earth's surface by population size, an increasing population obviously implies a greater density of persons per surface unit. But this simple calculation is seldom (if ever) a meaningful or appropriate calculation. What matters are the densities that people experience. In the United States, a large proportion of rural counties have lost population in recent decades, and in 1986 the entire state of Iowa had fewer people under the age of five than at any time since before 1900; the number declined by fully a third between 1960 and the 1980s. Rural people move to cities. And the people in Iowa worry about it because the trend feeds on itself; the fewer people there are, the poorer the economic opportunities for those who remain (just the opposite of Malthusian theory), and the greater the incentive for them to leave, too. The continued concentration of population could mean that the absolute number of persons living on (say) 75 percent of the area of the Earth might be declining even as total world population increases, due to the reduction in the total number of persons needed for agriculture as countries grow wealthy. (This is likely to continue indefinitely: see chapter 6.) Consider this: Between 1950 and 1985 (and surely continuing beyond 1985), the population density on 97 percent of the surface of the United States has been decreasing. Here is the calculation that shows this to be so: The total population outside of metropolitan areas fell from about 66 million people in 1950 (44.9 percent of total population) to about 59 million people in 1985 (23.4 percent of total population). That is, aside from the (roughly) three percent of the area of the United States incorporated within metropolitan areas, the total population has fallen (see figure 23-11). Those who choose to live in rural areas are experiencing a decline in the density of neighbors. And there is little reason to expect this trend to reverse for a long time, if ever. The total urban area may creep up a bit - say to six percent in a century - but in the rest of the country, density may well decline indefinitely, except for vacation uses. The trend is the same for the rest of the developed world, too (see figure 23-11). FIGURE 23-11 Yes, population density within any given metropolitan area will probably increase. But individuals can choose to live in cities or their surrounding suburbs of whatever density they prefer, and the range of individual choices at any density level will increase in the future. Even more important than population density per unit of surface area is the amount of living and work space available to an individual. And this is continually increasing, in considerable part due to multi-story buildings, made possible by increasing income as well as greater technology, as discussed in chapter 6. Of course this line of thinking does not comfort those who call themselves wilderness lovers, and simply want no human beings there - except themselves. "Rivers like the Hulahula and the Kongakut have lost some of their wilderness character", said George Heim of Alaska River Adventures, a group that organizes trips. "You're not getting what you go up there for if you end up sharing the river with 20 other people and jockeying for campsites.'" One wonders what is the "optimal" number of visitors that authorities should allow - 100? 10? 1? 0? And, should Alaska River Adventures be given a monopoly on the trade? A cute device used to show that a larger population is a bad thing - based just on population numbers alone - is used by Zero Population Growth in their ratings of the quality of life in American cities. The ratings simply assume that a larger population is bad. And hence it is not surprising that the smaller cities that have no growth garner such headlines as "Albany, N. Y. Ties for First Place in Measure of Urban `Livability.'" And such publicity leads the mayor of Phoenix to be "aggressive about growth-control measures". One wonders how the journalistic staff of such an august newspaper as the Wall Street Journal can fall for such simple-minded stuff - unless they desperately wish to believe in it, which leads into the discussion of values in chapters ??. CONCLUSION The doomsday scenario of "standing room only" suggests a juggernaut inexorably bearing down upon the world, subject to no control. But such forecasts have a very poor track record. The data suggest that not only starvation and disease act to control population size; rather, a variety of economic, political, and social forces are also important. To understand these forces is the task of the following chapter, which will show how such an understanding can help us predict what the future holds in store if we adopt one or another policy toward population growth. Even when we have the correct data, the gross numbers can sometimes be misleading. Some of the implications of the long- run trends - an increasing "labor scarcity" in the world, and at least for some time a decline in population density in most areas in the developed world, for example - will surely seem strange if stated out of context. page# \ultres\ tchar23 February 17, 1994