WPCJ 2n ZB!HP LaserJet IIIr 3   4 HP3.PRSP F2|x6X@X@ Inquiry paper #Document[1]rM Document StyleI+M IIII_+M F   ׃  Document[2]rM Document StyleI+M IIII_+M *    2pq5ee Document[3]rM Document StyleI+M IIII_+M 0     Document[4]rM Document StyleI+M IIII_+M   . Document[5]rM Document StyleI+M IIII_+M   Document[6]rM Document StyleI+M IIII_+M   2pp}Document[7]rM Document StyleI+M IIII_+M  ` ` ` Document[8]rM Document StyleI+M IIII_+M ` ` ` Technical[1]rM &e+IZ I+M IIII_+M $ Technical[2]rM &e+IZ I+M IIII_+M / 2G LTechnical[3]rM &e+IZ I+M IIII_+M : Technical[4]rM &e+IZ I+M IIII_+M E Technical[5]rM &e+IZ I+M IIII_+M P Technical[6]rM &e+IZ I+M IIII_+M [ 2O y 8   Technical[7]rM &e+IZ I+M IIII_+M f Technical[8]rM &e+IZ I+M IIII_+M q  1rM Document StyleI+M IIII_+M F!"   ׃  2rM Document StyleI+M IIII_+M *#$    2Q q e e 3;rM Document StyleI+M IIII_+M 0% &    4rM Document StyleI+M IIII_+M ' ( . 5.rM Document StyleI+M IIII_+M )* 6rM Document StyleI+M IIII_+M +, 2hpp}c7rM Document StyleI+M IIII_+M -.` ` ` 8CrM Document StyleI+M IIII_+M /0` ` ` 9rM &e+IZ I+M IIII_+M $12 10rM &e+IZ I+M IIII_+M /34 2(-t11crM &e+IZ I+M IIII_+M :56 12rM &e+IZ I+M IIII_+M E78 13hrM &e+IZ I+M IIII_+M P9: 14rM &e+IZ I+M IIII_+M [;< 2ZpqS15rM &e+IZ I+M IIII_+M f=> 16~rM &e+IZ I+M IIII_+M q?@ 1$VrM Document StyleI+M IIII_+M AB` ` ` 2%rM Document StyleI+M IIII_+M C D . 2ee[pS3&$rM Document StyleI+M IIII_+M EF 4'zrM Document StyleI+M IIII_+M GH 5(rM Document StyleI+M IIII_+M *IJ   6)TrM Document StyleI+M IIII_+M KL` ` ` 2%ncBibliogrphyrM Bibliography I+M IIII_+M MN Right Par[1]rM Right-Aligned Paragraph NumbersIIII_+M 8OP@  Right Par[2]rM Right-Aligned Paragraph NumbersIIII_+M AQR@` ` `  ` ` ` 7-!rM Document StyleI+M IIII_+M 0S T    2W^Right Par[3]rM Right-Aligned Paragraph NumbersIIII_+M JUV` ` ` @  ` ` ` Right Par[4]rM Right-Aligned Paragraph NumbersIIII_+M SWX` ` `  @  Right Par[5]rM Right-Aligned Paragraph NumbersIIII_+M \YZ` ` `  @hhh hhh Right Par[6]rM Right-Aligned Paragraph NumbersIIII_+M e[\` ` `  hhh@ hhh 2 ORight Par[7]rM Right-Aligned Paragraph NumbersIIII_+M n]^` ` `  hhh@  Right Par[8]rM Right-Aligned Paragraph NumbersIIII_+M w_`` ` `  hhh@ppp ppp 84J)rM Document StyleI+M IIII_+M Fab   ׃  Doc InitrM Initialize Document StyleM IIII_+M cd    I. 1. A. a.(1)(a) i) a)Document2# !"q"a1DocumentM I:Document Style&e+IZ I+M IIFef   ׃  a2DocumentM I:Document Style&e+IZ I+M II*gh    a3DocumentM I:Document Style&e+IZ I+M II0i j    a4DocumentM I:Document Style&e+IZ I+M IIk l . 2$eP#e#p$p$a5DocumentM I:Document Style&e+IZ I+M II mn a6DocumentM I:Document Style&e+IZ I+M II op a7DocumentM I:Document Style&e+IZ I+M IIqr` ` ` a8DocumentM I:Document Style&e+IZ I+M IIst` ` ` 2',%}&&'Tech InitrM Initialize Technical Style IIII_+M uv 1 .1 .1 .1 .1 .1 .1 .1 Technicala1Technical I<JF2=I&e+IZ I+M II$wx a2Technical I<JF2=I&e+IZ I+M II/yz a3Technical I<JF2=I&e+IZ I+M II:{| 2*'|(%))a4Technical I<JF2=I&e+IZ I+M IIE}~ a5Technical I<JF2=I&e+IZ I+M IIP a6Technical I<JF2=I&e+IZ I+M II[ a7Technical I<JF2=I&e+IZ I+M IIf 2:-*+,,a8Technical I<JF2=I&e+IZ I+M IIq 9=.rM Technical Document Style+M IIII_+M &  . 10.rM Technical Document Style+M IIII_+M &  . 116/rM Technical Document Style+M IIII_+M *    2/l-..E/12/rM Technical Document Style+M IIII_+M '   130rM Technical Document Style+M IIII_+M &   14.1rM Technical Document Style+M IIII_+M 4$     151rM Technical Document Style+M IIII_+M &  . 225/04162rM Technical Document Style+M IIII_+M &  . PleadingrM Header for numbered pleading paperIII_+M     X  y*dddyy*dddy HHH1 HHH2 HHH3 HHH4 HHH5 HHH6 HHH7 HHH8 HHH9 H10 H11 H12 H13 H14 H15 H16 H17 H18 H19 H20 H21 H22 H23 H24 H25 H26 H27 H28   ӋX0Í ÍX0Í Í2d5l5#|b\*hN='3,,d UUUUUUUUUUUUU ,,,888EEEQQQaaaqqq -' k-'VAY#~ #Y PP#  0  k  VAYU  #Y PP#  5  k==WB#Y PP#  10  kooWB#Y PP#  15  kWBG#Y PP#  20  kWBz#Y PP#  25  2 kk  b b     Z Z     R R    2 I I     A A     99  00 2 ((       2    zz  qq  ii2  aa  XX  PP  HH2  ??  77  ..  &&2        !! !!  " " "" ##2 ## ## x$x$ $$ p%p% %% g&g& &&2U@_(#Y PP#  0U@= ] (#Y PP#  2U@<(#Y PP#  4U@(#Y PP#  6U@(#Y PP#  8VAv(#Y PP#  10VAU(#Y PP#  12VA3"#(#Y PP#  14VA&'(#Y PP#  16nYaP# PP#  Figure 2: K/S Basic ModelaL[P# PP#  Panels, 1980sbMIm-#Y PP#  Per Capita GDPiZT7r #Y PP#  Predicted Growth Ratek_b  &Z Z  R  I   A  P90B0B(c"   z I      z r qY qY B i.  a   X  P  H H  ?  7   .  &    !| !q "e "W "W #G #5 #! x$ $ $ p% % kNb  IZ Z  `R  I 6  A  B9  0 0 M (      o Y F 8 8 , $ z  q q ! i& . a8 C C XP _ Po  H H  ?  7   .  &2 E  X  X i !z ! " " " # # # x$ $ $ p% %   k k b  b    Z  Z    R  R    I  I    A  A    9 9 0 0 ( (          z z q q i i a a X X P P H H ? ? 7 7 . . & &       ! !! ! "  "" "# ## ## #x$ x$$ $p% p%% %g& g&&      XC M#\ P P#  LDCsWBM#\ P P#  DCs !'ePUD$#U PP#  Using LDC Medians  p N aLq#c #\ P P#  DC Median Y/N@+ [# PP#^IHP #\ P P#  Median Y/NWB$ P #\ P P#  LDC@+ ~ [# PP#dO U#U PP#  Using DC MedianscDD c iZT- #4f PXP#  Predicted Growth Rate     Toward a Cure for the Myopia and Tunnel Vision of the Population Debate: A Dose of Historical Perspective  by  Allen C. Kelley Robert M. Schmidt* Copyright 1995  @1.0 The Population Debate: Why Does it Continue? Debates about the economic consequences of population growth and size have been too narrow, and they have been too focused on impacts that occur in the short run.The major exception to this statement relates to concerns about the longrun impacts on the environment, and the important, but lamentably nebulous and to date unmeasurable, concept of "sustainable" resource base. Such tunnel vision and myopia have resulted in alarmist assessments. However, when the analysis of the economic impacts is broadened in scope and extended over time, not only are the conclusions altered, but the analysis becomes at once more complex and more reasonable. In short, the pitfalls of misjudging the economic consequences of population growth are attenuated when one gains perspective. This paper has three objectives. Section 2.0 surveys the evolution of thinking over the postwar period about the net overall economic impacts of rapid population growth--the bottom lines, as it were.eWe will focus primarily on American contributions to the population debate.e We advance a surprising conclusion: most economistsUnless otherwise specified, reference to "economists" in this paper relates to the subset of economists who specialize in population matters, sometimes denoted as economicdemographers. * Kelley is James B. Duke Professor of Economics, and Associate Director, Center for Demographic Studies, Duke University; Schmidt is Associate Professor of Economics, University of Richmond. We are grateful for financial support from the Australian International Development Assistance Bureau, and the editorial and manuscriptpreparation assistance of Ms. Gail McKinnis. Draft (2/28/94). To be presented in Canberra, Australia, April 7, 1994. Ī who have specialized in population issues have held a balanced and distinctly non alarmist position on the economic impacts of population growth. While most would consider slower population growth to be beneficial to per capita economic growth, especially in those poor countries with exceptionally high rates of fertility and mortality, there is insufficient empirical evidence to justify the conclusion that the negative impact is particularly large on average.TWhen discussing any issue relating to 100 plus developing countries, exceptions to the rule are the norm. Debaters on both sides of the population issue highlight such exceptions. The generalizations advanced below represent our best judgment about the predominant weight of the evidence for the strong majority of the developing countries, but these must be qualified by the reality that no generalization holds for all countries and across all periods of time. To qualify each statement with such a disclaimer, while rigorous,would be unnecessarily tedious. T To the contrary, there is considerable evidence that some of the negative impacts are small or muted over time. From a policy perspective, while most economists would embrace some forms of population policies, they would place them as secondorder in importance. More specifically, the economic benefits of population policiesThese include programs affecting mortality (e.g., immunization, child and maternal health); fertility (e.g., family planning, infertility); and migration (e.g., resettlement schemes). Quite apart from ethical considerations, economists, both professionally and intellectually, look with great skepticism at any forms of compulsion and coercion of individual decision making unless "externalities" are important.Ķ are likely to be modest by comparison with a host of alternative policies to improve the living conditions in poor countries. This statement applies even to those specific areas (e.g., education, saving, exhaustible resource use, etc.) assessed by population pessimists as being compromised by rapid population growth. Population policies and programs in most of these areas can, at best, be justified as being complementary, albeit with relatively slowacting and modest impacts, to more fundamental policies designed to accomplish specific economic ends. These assessments by economic demographers may come as a surprise since most writers who advance strong pessimism about population matters base their arguments on alleged powerful adverse economic impacts. The key insight to be gained from section 2.0 will be an answer to the question: Why don't economists share the strong pessimism about the economic impacts of population growth so widely embraced by biologists, clergy, ecologists, journalists, and demographers? While the answer to this question will leave plenty of room for disagreement about economists' conclusions, it should clarify the "rules of the debate" and the perspective of this paper. Section 3.0 moves behind the economists' conclusions and takes up the question: What is the assessment by economic demographers about the specific impacts of population growth in areas such as saving and investment, the environment, poverty, education, and health? This section can be brief since these themes are taken up by other authors participating in the AIDAB Inquiry. The goal here is to collect in one place the summary judgments held by most economic demographers as of the mid 1980s in order to provide a benchmark against which we may compare the current assessments provided as a part of the AIDAB Inquiry. We will find that there are two areas where "accepted wisdom" requires some modification. In particular, we find that 1)adverse educational outcomes of rapid population growth are not particularly high; and 2)feeding a doubled population using a sustainable production format will likely be possible, but that this challenge will be increasingly costly and there is some uncertainty about whether new technology packages will be available when needed. A final area where accepted wisdom requires qualification relates to the evidence showing a lack of correlation between population growth and per capita aggregate economic growth in the 1960s and 1970s. In particular, recent research reveals that a negative association emerges for the decade of the 1980s. While we do not consider this type of macroeconomic correlation evidence as especially interpretable or illuminating for any of the three decades, given the exceptional importance of these correlations to the population debate our paper concludes in section 4.0 by examining the new evidence.  2.0 The Population Debate in Perspective: "Why Aren't Most Economists Strongly Pessimistic About the Costs of Rapid Population Growth?",This and the next section draw upon a presentation by Kelley at the Ministerial Seminar on Population and Development in the AsiaPacific Region, at the Australian Academy of Science, Canberra, Australia, 34 November 1993; and Kelley (1994) and Kelley and McGreevey (1994)., @Pessimism about the economic impacts of population has dominated the thinking of population analysts since the original alarmist treatise by the Reverend Thomas Malthus, published over two centuries ago. This "traditionalist" perspective has been challenged from time to time, most recently in the postwar period and peaking in the 1980s with the spread of what we term "revisionism." Revisionists have downgraded the relative importance of population growth as a source of economic growth. Defining revisionism is critical: the key to understanding the "new" assessments is not the direction of the impact (whether positive or negative) of rapid population growth.Most revisionists conclude that many, if not most, Third World countries would benefit from slower population growth. Revisionists include, for example, Julian Simon (optimistic about the net positive impacts of population growth in the longerrun, and likely the most influential person in many dimensions of the population debates in the postwar period); Nobel Laureate Simon Kuznets (somewhat pessimistic about the consequences of rapid population growth in the Third World, but always guarded due to the ambiguity of the evidence); and Joseph J. Spengler (even more pessimistic, but the strongest early advocate for assessing short and longrun perspectives); and many others. Rather, the key is the adoption of a methodological perspective that highlights the intermediate to the longerrun, taking into account both direct and indirect impacts, and feedbacks within economic, political, and social systems. This is in contrast to the "traditionalist" perspective which emphasizes shortrun adverse effects deriving from the apparent "fixity" of resources (i.e., diminishing returns), and which downplays the ability of economic systems to adjust. It may come as a surprise that revisionist thinking is not all that new, although it is associated by some observers with events in the 1980s. Indeed, the term itself may be a misnomer since it is arguably the dominant perspective of American economists working in the population field in recent decades. This interpretation is documented in a paper presented at the Nobel Symposium in Economics in Stockholm, Sweden (Kelley [1994]), which examined in considerable detail the writings of the economists who contributed to two United Nations (1953 and 1973) and two National Academy of Sciences (1971 and 1986) reports on the consequences of population growth. The 1953 United Nations Report was seminal, balanced, and wideranging in scope.The Report was distinctly revisionist, emphasizing a methodology that eschewed the "fixity" of resources in favor of one that treats "constants" as "variables" (p. 181). It noted that population growth and size exerted positive impacts (scale, organization), negative impacts (diminishing returns), and uncertain impacts (technology and social progress). The Report's bottomline assessment was modestly pessimistic, but no quantitative assessment was provided, and the conclusions were guarded and qualified.The Report emphasized that international trade and migration could notably attenuate the impacts of population growth; and that the impacts on saving and investment were, at least theoretically, quite uncertain (pp. 137, 237). This Report was revised in 1973, and the net assessment became somewhat more pessimistic, largely due to an emphasis on shorterrun impacts of population growth. Importantly, the Report's conclusions were strongly qualified by a powerful voice. Simon Kuznets, later to win the Nobel Prize, emphasized that across countries and over time, there did not appear to be a correlation between per capita output growth rates and population growth rates. While he emphasized that such correlations are difficult to interpret, they could not be ignored. Surely if population impacts were strong, they should show up in the macroeconomic data. They did not. The most influential assessment of the consequences of population growth during this period was the 1971 National Academy of Sciences Report. Solidly "traditionalist" in orientation--focusing explicitly on shortrun impacts--the Report's assessment, as presented in the crisply written "Executive Summary," appeared to be strongly pessimistic, indeed alarmist.The Report cites around twentytwo negative impacts of population growth, provides almost no qualifications relating to such effects, and fails to enumerate or assess positive or countervailing impacts. In fact, however, this Summary was at variance with the scholarly papers supporting the Report. Since the Summary was unauthored, not vetted with the participants of the study, and plausibly heavily influenced by the financial sponsor (Agency for International Development), the conclusions of the Summary should be heavily discounted.VFor a detailed description of this study, see Kelley (1994).V Fortunately, the scholarly papers underlying the report did not partake of any such politicallygenerated bias, and represented solid contributions to the scholarly literature. On average these papers were revisionist in methodological orientation, although collectively they could be assessed as conforming to the moderately pessimistic but qualified bottom line assessment of the 1973 United Nations report. The 1986 NAS report is the most revisionist of the major studies. It was written almost entirely by economists, and the executive summary represented a total collaboration by the scholars who compiled the working papers. The report emphasized individual and institutional responses to initial impacts of population change--conservation in response to scarcity, substitution of abundant for scarce factors of production, innovation and adoption of technologies to exploit profitable opportunities, and the like. These responses are considered to be pervasive and important. While the 1986 NAS working group judged the net impact of rapid population growth in the Third World to be negative, their conclusion was highly qualified. It merits quotation since it plausibly represents in tone the predominant assessment of most American economicdemographers over the postwar period: On balance, we reach the qualitative conclusion that slower population growth would be beneficial to economic development of developing countries (emphasis ours). @Examining this carefully worded statement (arduously negotiated to obtain unanimous support by the working group) is instructive because it exemplifies several attributes of modern economic thought on population: 1) there are both important positive and negative impacts of population growth (thus, "on balance"); 2)the actual size of the net impact--even whether it is strong or weak--cannot be determined given existing evidence (thus, "qualitative"); 3) only the direction of the impact from high current growth rates can be discerned (thus, "slower," and not "slow"); and 4) the net impact varies from country to country-- in most cases it will be negative, in some it will be positive, and in others it will have little impact one way or the other (thus "developing countries"). Our historical review of the major summary reports on the consequences of population growth leads to the following conclusions. First, if one ignores the Executive Summary of the 1971 NAS Report as inconsistent with its supporting scholarly papers, all the major assessments have been quite tempered, especially those parts that are contributed by economists. Second, all major studies have come out with a pessimistic assessment; all are quite qualified; none are willing to conclude that the net negative impact of rapid population growth is quantitatively large; and none are alarmist. Finally, the perception of the emergence of revisionism in the 1980s is due mainly to three factors: 1) the increasing influence of economists visavis others in the population debates; 2) the cumulative impact of Julian Simon, whose writings, and especially his influential book The Ultimate Resource, posed a threat to the populationprogramming/familyplanning establishment, and, importantly, stimulated the compiling of several survey articles (including the NAS report) which systematically exposed the results of two decades of empirical research; and 3) the political environment of the period. While politics have since changed, this is the least important of the factors accounting for revisionism. Returning to our question "Why aren't most economists strongly pessimistic about the costs of rapid population growth?," the answer is straightforward. Most economists have recognized and accounted for longerrun positive feedbacks that counter the shortrun negative impacts of population growth; have emphasized the role of institutions (mainly markets, but also governments) in responding in constructive ways to populationinduced scarcities; and see an important role for human capital and technological change as sources of economic growth. Indeed, once technology and human ingenuity in response to any pressures --population or otherwise--are given significant weight, and once one considers technological change as an (the?) important ingredient of longrun economic progress, then the shortrun costs of population growth can be notably counterbalanced by longerrun benefits, and the netassessment picture becomes clouded. One should not, having said this, dismiss the shortrun costs of population growth. They are real and they can be large. They tend to attract the greatest attention of politicians whose perspectives are understandably shortrun: budgets must be proposed annually, elections are often on the immediate horizon, and constituents want benefits now. However, sound policy analysis of population issues requires that the impacts of population growth be assessed over the longer run. While children undeniably impose immediate costs for their rearing, they also grow up and contribute to economic output. To focus mainly or exclusively on the "costs" of a relatively small portion of the human life cycle--childhood--is myopic and partakes of tunnel vision. Such a perspective can lead to inappropriate economic assessments and wrongheaded government policies. The review has focused, however, on bottomline assessments. What about the impacts of population on individual sectors? 3.0 Status Report on Current Knowledge about Specific Economic Consequences of Population Growth: "How Has Recent  ` ` ` Empirical Evidence Altered the "Accepted Wisdom"? @The 1986 NAS Report still provides the most current and useful rendering by American economic demographers on the specific economic impacts of population growth and, as such, offers a useful benchmark for assessing the technical papers supporting the Australian Inquiry Report. In six major areas, the NAS Report provided relatively good news.For a more complete summary and appraisal, see Kelley (1986), Simon (1986), Potter (1986), and Daly (1986). For brevity, this section will not provide detailed bibliographic citations but instead refer the reader to the hundreds of citations supporting the NAS conclusions as found in NAS (1986). For a wideranging consideration of the literature on population consequences, including additional bibliographic references, see Birdsall (1988), Kelley (1988), McNicoll (1984), World Bank (1984), and Srinivasan (1988). For anupdate of citations on specific areas, see the individual papers of the Australian Inquiry.ą  On exhaustible resource use. The concern about population growth resulting in resource exhaustion appears to be misplaced: the relationship between population growth and global resource use is not as strong as had been assumed. Only if we value more highly the relative merits of future generations over the present can a strong case be made that our exhaustible resources are being used improperly over time due to population growth.This conclusion is based on studies of 1) the determinants of resource supply and demand (related most strongly to per capita income); 2) the relative impact of priceinduced versus serendipitous technological change on resource discovery and efficiency of use, and lower costs of extraction; 3) the responsiveness of conservation in the face of resource scarcity; and 4)an assessment of the efficacy of markets and political processes in allocating exhaustible resources over time.  On pollution. Pollution problems are solved largely by government policies correcting market weaknesses; they are not caused significantly by population growth per se, although slower population growth gives governments more time to act.  On savings and investment. The traditional concern about a substantial reduction of saving due to rapid population growth does not appear to be sustained by the data. While some capital shallowing will occur, the quantitative impact of this on economic growth is not substantial.The conclusion regarding saving is based on the inability to obtain reasonably robust empirical results demonstrating the impact of population growth and age structure on saving, deriving mainly (but not exclusively) from international cross section data. The conclusion regarding investment is based on empirical assessments using computable, generalequilibrium growth models.   On education. The concern that population growth will shift resources from productive physical capital formation into such areas as education is not sustained by the data. Educational enrollments have expanded dramatically in the face of population pressures. This expansion has been financed not by diverting investment funds from other areas, but rather by improving the efficiency of resource use, and by reducing expenditures per pupil. While the latter will have some adverse impact on the quality of education, the quantitative importance of this impact is uncertain.  On city growth and urbanization. Problems associated with excessive city growth are to be solved primarily by modifying government policies that encourage people to live in, or move to, cities. Reduction of population growth does not represent a primary means of solving such locational imbalances.  On adjustments to rapid urban growth. There is little evidence that urban unemployment is caused primarily by rapid population growth. Moreover, the deterioration of urban services is not the result of diseconomies in the provision of such services, and slower population growth per se would not significantly improve the quality of the services provided. Rapid population growth exacerbates urban problems but does not represent their primary causes. In three areas, the NAS Report provided relatively bad news.  On renewable resource degradation. The adverse effects of population growth in the form of renewable resource degradation where property rights are difficult to assign or maintain--rain forests, fishing areas, etc.--pose a serious problem.This result, which elevates population pessimism, is also revisionist in orientation since it explicitly highlights the role of feedbacks. In this case, market and political feedbacks needed to attenuate excessive resource use are weak. These feedbacks are likely to remain weak in the intermediate future when substantial, and in some cases irreversible, resource degradation will continue to take place.   On the health and education of children. Large families implied by rapid population growth result in households spending less time and money on the health, nutrition and education of individual children. The quantitative impact of this allocation is difficult to assess, but the qualitative effect is detrimental to the welfare of children.  On offsetting effects. Positive impacts of population in the form of scale economies, and changes in the economic system, resource use, and technology induced by population growth, are not sufficient to offset the negative direct effects of population change. @Given this set of propositions applicable to the mid 1980s, we ask the question: in what areas has recent research altered the assessments of the National Academy Report? There are several. As before, there is good news supporting a more optimistic perspective, and bad news supporting a more pessimistic view. First, some good news supporting optimism.     On education. Based on a sizeable number of new studies using household data, and on recent countrywide experience, it appears that the impacts of large families on educational attainments is fairly small. Moreover, countries continue to be quite successful in advancing education (including modest reductions in class size) even in the face of rapid population growth and, notably, during a decade of languishing economic growth. Finally, based on the judgment of policy analysts, the benefits in the form of tangible educational outcomes (e.g., measurable impacts on achievement scores) of an infusion of new resources into the education sector from any source (including eased financial pressures deriving from reduced numbers of children) could be disappointing in many countries given the notable need to improve the efficiency and effectiveness of education provision. In sum, while the microeconomic and macroeconomic impacts of reduced population growth on educational provision and outcomes will likely be positive in the short run, there is little reason to believe that these benefits will be especially sizeable or that they could not be attained more effectively by alternative means. There is also bad news that supports some pessimism.  On offsetting effects. Recent macroeconomic evidence shows a negative correlation between rapid population growth and per capita output growth during the 1980s, providing some support for a bottomline conclusion that reducing the rate of population growth will enhance the pace of economic activity.yThe 1986 NAS Report's bottomline conclusion was little influenced by the "nocorrelation" results for the 1960s and 1970s, but was based instead on the empirical results of several studies, as well as an assessment that the positive economicgrowth feedbacks in response to the shortrun costs of rapid population growth are quantitatively important.y @This evidence requires close consideration, a task to which we now turn.  4.0The Bottom Line Revisited: What is the Meaning of the New MacroEconomic Correlations of Population Growth, Density and Size with Per Capita Output? @ 4.1 The NonCorrelation  No empirical finding has been more important to conditioning the population debate than the widelyobtained statistical result showing a general lack of correlation between the growth rates of population and per capita output. Figure 1 provides ppp Figure 1 a representative example. Per Capita Income & Population Documented in more than hhhGrowth: 1970811Kelley (1988), p. 1701.1 (annual percent)  X two dozen studies, Reviews of this literature are provided by Blanchet (1988a), Browning (1982), Chesnais (1985), Kelley (1988), Lee (1983), McNicoll (1984), and Stavig (1979). Lee's assessment is most encompassing: "Dozens of studies, starting with Kuznets', have found no association between the population growth rate (n) and percapita income growth rate ( y /y), despite the obvious fact that at least since WWII population growth rates have varied considerably. These studies control for other factors such as trade, aid, and investment in varying degrees. Two recent studies add historical depth to this analysis; even within countries (and thus looking only at disequilibrium), over periods as long as a century, or as short as 25 years, there is no significant association of n and  y /y for either the DCs or LDCs; put differently, one can't reject the hypothesis that the regression coefficient of  Y /Y on n is unity" (Lee 1983, p. 54). For examples of this literature, see Bairoch (1981), Easterlin (1967), Hazledine and Moreland (1977), Isbister (1975), Kuznets (1967), Stockwell (1962, 1966), and Thirlwall (1972). For some methodological assessments, see Blanchet (1988a), Coale (1986), National Research Council (1986), and Simon (1989).   such a (lack of) statistical regularity flies in the face of strongly held beliefs by those who expect rapid population growth to deter the pace of economic progress.rA similar finding is uncovered by Ahlburg (1988) for 17 island economies of the Pacific.r The correlations have therefore become a   point of contention. On the one hand, most analysts agree that simple correlations between population and economic growth are difficult to interpret, plagued as they are by failure to adequately account for reverse causation between demographic and economic change, complicated timing relationships associated with the Demographic Transition, excessive reliance on crossnational data, sensitivity to the selection of countries, complexity of economicdemographic linkages that are poorly modeled, spurious correlation, econometric pitfalls, and data of dubious quality. On the other hand, the virtual absence of a systematic relationship in the face of such strongly held priors has quite literally kept the population debate alive. Ronald Lee's summary evaluation of this literature is instructive: ... these crossnational studies have not provided what we might hope for: a rough and stylized depiction of the consequences of rapid population growth: unless, indeed, the absence of significant results is itself the result. (1983, p. 54) @Ironically, while the simple correlations between population and economic development have constituted a major force in causing a reassessment of the impacts of population growth, the appearance of "new" correlations in several recent studies could well cause the pendulum to swing back toward a more pessimistic (alarmist?) interpretation.These studies include Barlow (1992), Blanchet (1988a), Bloom and Freeman (1988), Brander and Dowrick (1992), and United Nations (1988).ġ These studies appear to reveal a negative association between population and economic growth based on international crosscountry data for the early 1980s. Even though the authors of these studies are generally guarded with respect to the strength or importance of this finding, the intriguing question arises: Has the impact of population growth changed? Could it be that the negative consequences of rapid population growth associated with diminishing returns to capital and the environment are emerging as relatively more important forces than, say, the positive impacts of scale, induced innovation/technical change, and/or attenuating feedbacks? Or, alternatively, is it possible that the 1980s--a period encompassing significant structural adjustments, world recession, wars, and droughts--constitute an "exceptional" decade for untangling these economicdemographic interactions? Are these recent statistical correlations robust, or are they the result of necessarily arbitrary research decisions that are not yet fully assessed--decisions relating to choice of statistical procedures, data sets, time periods, countries included, modeling structures, functional forms, and the like? In short, is a negative population and economicgrowth correlation emerging? If so, why; and so what? While we believe that neither the results of the recent studies, nor those similar to them over the last two decades, should carry very much weight in assessing the net consequences of demographic change on economic growth,Julian Simon has been the most vocal in interpreting the results. He argues: "The most important fact in today's population economics is the lack of observed correlation between the rate of a country's population growth and the rate of its economic development.... The absence of a negative statistical relationship is evidence that within a century, or within even as short a time as a quarter of a century, the positive benefits of additional people at least balance the shortrun costs" (Simon [1992], p. ix; see also Simon [1989]). Most other observers are more cautious. With reference to the literature, Kelley (1988, p. 1701) concludes: "...these statistical correlations provide little prima facie information about the size or nature of the net impact of population growth on economic growth." In a similar vein, the National Research Council (1986, p. 7) observes: "[S]imple correlations between population growth and per capita income, although intriguing, ultimately provide little insight into the causal impact of a policydriven decline in fertility." Caution is also expressed by Blanchet (1988a), who finds a possible reversal of previous findings. He argues: "...it is necessary to reask the question of the relevance of such statistical results, and, in fact, this question can be asked when they are not significant (as was done by Malthusians up to now), as well as when they are (what could be done by antiMalthusians in the future, if the emergence of the negative correlation happened to be confirmed)" (p. 4).  we also believe that such a message will be selectively heard, crowded out in the minds of those who would like to accept or reject the recent (as well as the past) empirical findings. Accordingly, it is important that the recent studies be carefully evaluated in terms of their econometric muster; and, to the extent that they appear to be yielding new findings, that the robustness and meaning of such findings be examined. Timing is important. After all, it was Kuznets' empirical finding showing a lack of association between population and economic growth in the early 1970's that profoundly conditioned the population debate for more than a decade. Moreover, the outpouring of results seemingly confirming Kuznets' finding both sustained and buttressed "revisionism"--a balanced, nonalarmist assessment of the role of population--now arguably the prevalent perspective of American economists specializing in the population field. 4.2 Are New Correlations Emerging? A Review of the  ` ` ` Literature @What can we make of the recent empirical studies of population and economicgrowth connections? We addressed this question in considerable detail in Kelley and Schmidt (1992). Here we summarize the bottomline results of that inquiry, providing sufficient detail to indicate the nature of the argument and evidence. All of the abovementioned studies present findings for simple (what we term "naive") models that correlate per capita output and populationgrowth (or births); and all, save one, present modeling embellishments. We examined the studies at two levels. The first represents an empirical appraisal, including replication, of the sensitivity of the naivemodel results to: alternative estimation procedures (e.g., corrections for heteroscedasticity), samples (differing country selections), country groupings (high versus lowincome, DC versus LDC), data sets (Summers and Heston, World Bank, United Nations), and time periods (five and tenyear intervals). The second represents an appraisal of some modeling embellishments. Taken together, our evaluation provides a reasonably consistent picture. Based on the five studies which cover the period up to 1985, we conclude that:  A negative correlation between output and populationgrowth may have occurred, although there are enough "exceptions"+The results in the UN (1988) study are reasonably robust with respect to estimation procedure and country selection. (They are also strengthened by our own extension of the UN model to 1990.) Bloom and Freeman's (1988) findings do not notably alter their previous "neutralist" position. Blanchet (1988a) makes a convincing case for the exceptional difficulty in interpreting the various naivemodel results. His own new evidence, strongest for agricultural output and the weighted regressions, is more suggestive than conclusive. For example, his model that attempts to control for influences on output growth fails to uncover a statistically significant impact of population growth. Barlow's (1992) highly detailed model is very difficult to interpret, and his results and interpretations are not convincing. Brander and Dowrick's (1992) paper is the most insightful and rigorous of the various studies. Their naivemodel results for population growth are quite robustwith respect to countries selected and estimation procedures (two exceptions of note are found), but they are somewhat fragile with respect to country grouping as between high versus lowincome, and DCs versus LDCs. Since in their study, as in most other studies, country grouping counts (Chow tests indicate that the countries are not drawn from the same sample), isolating those factors that justify one grouping versus another becomes central to the analysis. This is especially the case when the populationgrowthimpact findings are sensitive to this classification.+ to this finding that such a correlation must be considered as tentative at this point.  While the various correlations are to some extent fragile with respect to necessarily arbitrary modeling choices, they are nonetheless no more shaky in this regard than most of the studies that have preceded them. If anything, some of the recent studies appear to be on relatively firmer grounds econometrically, and they are based on better data.  Even though the number of new studies is small, added support to their findings is provided by our own sensitivity/robustness analysis that experimented with variations in the data used, countries selected, time demarcations, and estimation procedures.The studies examine three data sets (Summers and Heston 1993, United Nations 1993, World Bank 1992), five country selections, five country groupings, and eight periods; and they employ five estimation procedures. @As a modest addition to these studies, we extended the naive model to cover the period of the late 1980s.  A reasonably consistent negative correlation shows up in the late 1980s that appears to be more robust than similar correlations recorded in the various recent studies covering the early 1980s.This conclusion is quite robust with respect to estimation procedure (OLS, OLSWhite corrected, weighted least squares), sample (a narrow versus a broad country sampling), and development classification (low versus highincome, LDC versus DC). @Taken together, the above conclusions suggest that a negative correlation may well characterize the 1980s, and that more elaborate modeling is warranted to assess the nature of this relationship. 4.3 The Kelley/Schmidt Basic Model @To this end, we have undertaken an extensive econometric inquiry into these relationships, resulting in a booklength manuscript too dense and broadranging to summarize here.Kelley and Schmidt (1994). Our study examines four major models (and many variants): 1) the K/S Basic Model, similar to equation 2 below; the K/S Augmented Model, which appends population density and size; 3) a Barrotype model that is augmented with population growth, size and density; and 4) the K/S Decomposition Model, that examines separately the impacts of current births, current deaths, and lagged births (each in addition interacted with Y/N). The robustness of the major findings from the various models are then assessed with respect to: 1) alternative output measures (GDP/Ngr versus GDP/Lgr); 2) alternative functional forms in Y/N (linear, logarithmic, quadratic, cubic); 3) alternative data samples (with versus without small countries, DCs versus LDCs, pooled); 4) the impact of specific countries exercising unusual statistical influence; 5) OLS versus Whitecorrected standard errors; 6) crosssection versus panel estimation; and 7) five versus tenyear data intervals. hhh However, apropos to the issue at hand, it would be instructive to present one of the representative findings from this ongoing study.YAppendix A details the sample, data, and estimation procedures.Y A rendering of the empirical formulation using panel data for the Kelley/Schmidt "Basic" Model is presented in equation 1. [1] Y/Ngrit =  + 1(Y/N)it + 2(Y/N)it  2 + 3(Y/N)it  3 + tNgrit + t(Y/N . 0 Ngr)it + i + )t + it where Y/N = gross domestic product per capita, N = total population, gr denotes growth rates, t = time period, and the subscripting of the population coefficients (t and t) indicates that these relationships may vary over time, i represents a country fixed effect, )t represents a period fixed effect (T1 = 1960s, T2 = 1970s, T3 = 1980s), and it represents a classical disturbance term. @It is hypothesized that the rate of per capita output growth varies with 1) the level of economic development and 2) the rate of population growth; and that 3) the impact of population growth itself varies with the level of economic development. The first hypothesis encompasses the general notion that, as countries develop economically, the "returns" from capital and other resources can decline (traditional neoclassical growth theory), but that this can be countered by the positive scale effects associated with knowledge creation (the new growth theory).For the neoclassical paradigm, see Solow (1956), Cass (1965) and Koopmans (1965); for the new growth theory, see Romer (1986).Ę Economic theory admits many alternative patterns on the posited relationships, and thus our empirical specification must be flexible.See Srinivasan (1992). We do not posit signs on the cubic. Linear, logarithmic, and quadratic formations are somewhat more constraining theoretically and, as a result, not as consistent with our somewhat eclectic theoretical priors. Extensive empirical analysis of functional form supports the appropriateness and viability of the cubic in most instances.Ā The second hypothesis is eclectic: population growth can exert a positive or negative net impact due to the many factors outlined above in sections 2.0 and 3.0. The third hypothesis admits the possibility that the impacts of population growth can be different in the LDCs versus the DCs, or for any other factor that is associated with the level of per capita output. In subjecting equation [1] to empirical analysis, we restrict 1, 2, and 3 to remain constant over time while allowing the population coefficients to vary. The panel constitutes 267 observations (3 periods x 89 countries). The results are presented in equation [2]fStandard errors are Whitecorrected for heteroscedasticity; gr = growth; and T1, T3, and T5 = dummy variable for 1960s, 1970s, and 1980s, respectively. No constant is shown since the data were demeaned on a countrybycountry basis to simplify estimation. The quadratic and cubic in Y/N are divided by 10 and 100, respectively. f: [2] Y/Ngr = 3.45** (Y/N) + .0264** (Y/N)  2 - .0074** (Y/N)  3 .78 T1 . 0 Ngr 1.20** T2 . 0 Ngr 2.11** T3 . 0 Ngr + .33** T1 . 0 Y/N . 0 Ngr + .26** T2 . 0 Y/N . 0 Ngr + .23* T3 . 0 Y/N . 0 Ngr + 1.46* T2 + 2.63** T3 Adjusted R2 = .47 * and ** = significant coefficient at 5% and 1% levels, respectively. @The results show that 1) the pace of economic growth varies systematically with the level of economic development--declining for a while and, at later stages, rising somewhat; 2) the direct impact of population growth is statistically insignificant in the 1960s, and negative in the 1970s and 1980s; 3) the size of this negative impact increases systematically over time; 4) the direct negative impacts of population growth are countered by positive impacts that increase with the level of economic development; and 5) the positive population impacts diminish in strength over time. Overall, the results tend to support the "neoclassical" characterization of development whereby growth rates decline and converge over time. In terms of demography, the results are consistent with the "nocorrelation" result for the 1960s, although a negative impact emerges in the 1970s and gains strength in the 1980s. To assess the Figure 2 nature of the specific K/S Basic Model: Panels, 1980s   impacts of population growth, figure 2 compares, for the 1980s, two alternative economicgrowth scenarios corresponding alternatively to high population growth rates prevailing in the LDCs versus low population growth rates prevailing in the DCs. The results are intriguing.   First, for the LDCs the impact of population growth is negative. Second, this negative impact holds for the DCs at relatively lower income levels, but for most DCs population exerts a positive impact. These findings assist to resolve a puzzle in the literature. Everyone party to the population debates--pessimists and optimists alike--advance (often strong) reasons for expecting positive and/or negative impacts of population growth, yet for two decades there has been no obvious "net" impact revealed in the correlations. Is that because these positive and negative impacts are unimportant? Not necessarily. It could be that the impacts are offsetting in ways similar to the pattern displayed in figure 2. The two scenarios depicted in figure 2 are extreme, spanning the entire demographic transition from high to low fertility and mortality rates. It might be more interesting to consider an alternative assessment that applies to the LDCs for the 1980s. One might pose the following counterfactual: What if the rate of population growth had declined from 2.54% in 1980 to 2.34% in 1990an assumed decline of only .02% per annum? The implied Y/Ngr from this population growth rate decline scenario can then be compared with the Y/Ngr based on an historically relevant experience of virtual constancy in the population growth rate. The result is notable. In 1990, the historical constancy scenario yields a Y/Ngr for the median LDC of .52%; in contrast, the population growthrate decline scenario results in a Y/Ngr of .36%. The increase in the annual rate of per capita output growth of .16% represents the impact of lowered population growth. This impact, while not large, is both notable and plausible. Based on an assumed and robust longrun rate of per capita output growth of 2.0% as a norm, a reduced rate of population growth would have raised longrun economic growth by around 8% per annum. This is a conservative estimate based on a counterfactual that assumes a pace of population decline that is considerably lower than that used by the United Nations in its "medium" population growth rate projections. As a result, based on population's estimated impact in the 1980s, a figure of 10% (or .2 growth points) would represent a defendable scenario depicting the impact of population growth on an assumed longrun economic growth rate of Y/Ngr of 2.0%. 4.4 So What? @What can we make of these new results? This can be addressed at two levels: the results for the 1980s, and the implications for the 1990s and beyond. With respect to the 1980s, if one is inclined to place weight in these types of correlations and believe, for example, the nonexistence of a net population impact in the earlier decade(s), then one would conclude that a negative impact prevailed in the 1980s. For the reasons noted above, we are cautious in forming strong conclusions based on such results. Having said this, however, we are prepared to observe that our empirical findings for the 1980s are more believable than those in the earlier literature. Our results are based on higher quality and conceptually better data, and our results have been subjected to stringent sensitivity analysis. They moreover incorporate the plausible hypothesis that the impact of population growth varies systematically with the level of economic development. And we find the results plausible: population impacts are positive in some countries, and negative in others. Finally, the key findings are found not only in the K/S Basic Model, but also 1) in models augmented by population size and density; 2) in models where population impacts are decomposed into births, deaths, and lagged births; and 3) in models that control for intervening economic influences (education, urbanization, political change, saving and investment, and the like). With respect to assessing our results as they pertain to the 1990s and beyond, not much can be said unless we can ascertain why these new results have emerged. Without an answer to this question, we cannot predict whether a negative relationship will persist or increase in strength in the 1990s, whether the correlation will revert to the previous pattern of "no relationship" found for the earlier decade(s), or even whether a positive correlation might emerge for reasons yet unknown. While our own ongoing research has made considerable progress in clarifying these issues, it is not yet sufficiently complete to share. In short, we do not conclude that a "new" result has "emerged" for the 1990s and beyond; we only conclude that for the 1980s, the result is different from that of previous decades. Returning to the AIDAB assessment, however, at a minimum we can conclude that the new findings are fully consistent with, and indeed buttress, the bottom line conclusion of the 1986 NAS report that slower population growth would be beneficial to the economic development of most developing countries. References  #wd6X@@#Dennis A. Ahlburg. 1988. "Is Population Growth a Deterrent to Development in the South Pacific?", Journal of the Australian Population Association, Vol. 5, No. 1, pp. 4657. Bairoch, P. 1981. "Population Growth and LongTerm International Economic Growth." International Population Congress. Manila, Philippines: International Union for the Scientific Study of Population, 3, 141163. Barlow, Robin. 1992. "Demographic Influences on Economic Growth." University of Michigan. Mimeo. Birdsall, Nancy. 1988. "Economic Approaches to Population Growth and Development." In Hollis B. Chenery and T.N. Srinivasan, eds., Handbook of Development Economics. Amsterdam: Elsevier Science Pub. Blanchet, Didier. 1988a. "Estimating the Relationship Between Population Growth and Aggregate Economic Growth in LDC's: Methodological Problems." In Consequences of Rapid Population Growth in Developing Countries. New York: Taylor and Francis. Bloom, David E. and Freeman, Richard B. 1988. "Economic Development and the Timing and Components of Population Growth." Journal of Policy Modeling 10(1) (April) (Special Issue: Modeling Demographic and Economic Dynamics, Dominick Salvatore, guest ed.), 5781. Brander, James A. and Steve Dowrick. 1992. "The Role of Fertility and Population in Economic Growth: New Results from Aggregate CrossNational Data." University of British Columbia. Mimeo. Browning, Mark. 1982. An Econometric Investigation of the Relationship Between Population Growth and Income Growth in Developing Countries, unpublished dissertation, Department of Economics, University of Michigan, USA. Cass, David. 1965. "Optimum Growth in an Aggregative Model of Capital Accumulation." Review of Economic Studies, 12, 110. Chesnais, JeanClaude. 1985. "Progres economique et transition demographique dans les pays pauvres: Trente ans d'experience (19501980)." Population 40(1, JanFev), 1128. Coale, Ansley J. 1986. "Population Trends and Economic Development." In Jane Menken, ed., World Population and U.S. Policy: The Choices Ahead. New York: W. W. Norton for the American Assembly, pp. 96104. Daly, Herman E. 1986. "Review of Population Growth and Economic Development: Policy Questions." Population Development Review 12(3, September): 582585. Easterlin, Richard A. 1967. "Effects of Population Growth on the Economic Development of Developing Countries. The Annals of the American Academy of Political and Social Sciences, 369 (January), 98108. Hazledine, Tim and Moreland, R. Scott. 1977. "Population and Economic Growth: A CrossSection Study." Review of Economics and Statistics. 59(3) (August), 253263. Isbister, John. 1975. "Population Growth and the Productivity of Capital in the Third World." In Leon Tabah, ed., Population Growth and Economic Development in the Third World. Brussels: Ordina. Kelley, Allen C. 1986. Review of Population Growth and Economic Development: Policy Questions. Population and Development Review 12(3, September): pp. 563-568. Kelley, Allen C. 1988. "Economic Consequences of Population Change in the Third World." Journal of Economic Literature. 27 (December), 16851728. Kelley, Allen C. 1994. "Revisionism Revisited: An Essay on the Population Debate in Historical Perspective." In Population, Development and Welfare: The Nobel Jubilee Symposium in Economics, edited by Rolf Ohlsson (Berlin: SpringerVerlag). Kelley, Allen C. and McGreevey, William Paul. 1994. "Population and Economic Development." Forthcoming in Robert M. Cassen, ed., Population and Development. Third World Policy Perspectives series. Washington, DC: Overseas Development Council. Kelley, Allen C. and Schmidt, Robert M. 1992. "Is a Negative Population and EconomicGrowth Correlation Emerging? If So, So What?" Mimeo. Duke University, Durham, NC. Kelley, Allen C. and Schmidt, Robert M. 1994. "Is a Negative Population and EconomicGrowth Correlation Emerging? If So, Why, and So What?" Mimeo, Duke University and University of Richmond. Koopmans, Tjallig C. 1965. "On the Concept of Optimal Economic Growth." In The Economic Approach to Development Planning. Amsterdam: NorthHolland (for Pontificia Acad. Sci.). Kuznets, Simon. 1967. Population and Economic Growth." Proceedings of the American Philosophical Society. 111 (3) (June), 170193. Lee, Ronald D. 1983. "Economic Consequences of Population Size, Structure and Growth." International Union for the Scientific Study of Population, Newsletter No. 17 (JanuaryApril), 4359. McNicoll, G. 1984. "Consequences of Rapid Population Growth: An Overview and Assessment." Population and Development Review. 10 (2) (June), 177240. National Academy of Sciences. 1971. Rapid Population Growth: Consequences and Policy Implications. 2 Volumes. Baltimore: Johns Hopkins University Press for the National Academy of Sciences. National Research Council. 1986. Population Growth and Economic Development: Policy Questions. National Research Council, Working Group on Population Growth and Economic Development, Committee on Population, Commission on Behavioral and Social Sciences and Education. Washington, DC: National Academy Press. Potter, Joseph E. 1986. Review of Population Growth and Economic Development: Policy Questions. Population and Development Review 12(3, September): pp. 578581. Romer, P. M. 1986. "Increasing Returns and LongRun Growth." Journal of Political Economy, 94 (5), 10021037. Simon, Julian. 1981. The Ultimate Resource. Princeton: Princeton University Press. Simon, Julian. 1986. Review of Population Growth and Economic Development: Policy Questions. Population and Development Review 12(3, September): pp. 569577. Simon, Julian. 1989. "On Aggregate Empirical Studies Relating Population Variables to Economic Development." Population and Development Review. 15, (2), 323332. Simon, Julian. 1992. Population and Development in Poor Countries: Selected Essays. Princeton, N.J.: Princeton University Press. Solow, Robert M. 1956. "A Contribution to the Theory of Economic Growth." Quarterly Journal of Economics. 70 (1) (February), 6594. Srinivasan, T. N. 1988. "Population Growth and Economic Development." Journal of Policy Modeling. 10(1, Spring): pp. 728. Srinivasan, T. N. 1992. "Development in the Context of Rapid Population Growth: An Overall Assessment." Yale University, mimeo., pp. 30. Stavig, G.R. 1979. "The Impact of Population Growth on the Economy of Countries." Economic Development and Cultural Change. July, 735750. Stockwell, E.G. 1962. "The Relationship Between Population Growth and Economic Development." American Sociological Review. April, 250252. Stockwell, E.G. 1966. "Some Demographic Correlates of Economic Development." Rural Sociology. 31 (2), 216224. Summers, Robert and Alan Heston. 1993. "Data Update 5.5," computer diskette based on "The Penn World Table (Mark 5): An Expanded Set of International Comparisons, 19501988." Thirlwall, Anthony. 1972. "A Cross Section Study of Population Growth and the Growth of Output and Per Capita Income in a Production Function Framework." Manchester School of Economics and Social Studies. 40 (4, December), 339356. United Nations. 1953. The Determinants and Consequences of Population Trends. Department of Social Affairs, Population Division, Population Studies No. 17. New York: United Nations. United Nations. 1973. The Determinants and Consequences of Population Trends. Department of Economic and Social Affairs, Population Studies No. 50. 2 Volumes. New York: United Nations. United Nations. 1993. World Population Trends and Policies: 1992 Monitoring Report. Department of International Economic and Social Affairs, Population Studies No. 103. ST/ESA/SER.A/103. New York: United Nations. World Bank. 1984. World Development Report, 1984. Washington, DC: The World Bank. World Bank. 1992. "STARS: World Bank Data on Diskette, 199091: User's Guide." International Economics Department, World Bank. #x6X@X@#Appendix A  Data Equation 2 is based on the following data definitions, estimates, and sources. A.1The Country Sample Our country superset consists of the 135 countries designated as predominately market oriented within our primary data source, the Summers and Heston Penn World Table, Mark 5 (Summers and Heston, 1991). We imposed four conditions for including a country in our estimation sample. (1)` ` ` The country must have reasonably complete data for our key economic series. (28 countries were eliminated.) (2)` ` ` The country's GDP must not exhibit an "excessive" resource dependency over the period 197090. "Excessive" is defined as 30% or more of GDP originating from oil, or 50% or more originating from minerals. (6 countries were eliminated.) (3)` ` ` The country must not exhibit "gross" data problems. One country was eliminated because estimates of its population growth rate differed by more than one percentage point between United Nations and World Bank data sources in four of six quintennial periods. Another country was eliminated because remittances from abroad exceeded 30% of its GDP. (2 countries were eliminated.) (4)` ` ` The country's population must have exceeded one million in 1975. (10 countries were eliminated.) In ascending order of 1975 per capita GDP, the sample includes: Ethiopia, Myanmar, Burundi, Malawi, Tanzania, Chad, Zaire, Rwanda, Niger, Togo, Nepal, Central African Republic, India, Uganda, Somalia, Kenya, Haiti, Pakistan, Ghana, Indonesia, Cameroon, Bangladesh, Madagascar, Sudan, Nigeria, Mauritania, Benin, Sierra Leone, Senegal, Zambia, Egypt, Zimbabwe, Honduras, Sri Lanka, Morocco, C=te d'Ivoire, Philippines, Thailand, Paraguay, Papua New Guinea, Bolivia, El Salvador, Tunisia, Dominican Republic, Guatemala, Algeria, Nicaragua, Korean Republic, Colombia, Ecuador, Malaysia, Panama, Turkey, Jamaica, Peru, Chile, Costa Rica, Syria, South Africa, Brazil, Uruguay, Argentina, Portugal, Mexico, Singapore, Greece, Iran, Ireland, Hong Kong, Israel, Spain, Venezuela, Japan, Italy, Austria, UK, Finland, Norway, Belgium, Germany, Netherlands, Denmark, France, New Zealand, Australia, Sweden, Canada, Switzerland, and the United States. ٌA.2The Data Our primary source for economic variables is the Penn World Table as described in Summers and Heston (1991). We employ version 5.5 of the S&H data, an interim release distributed in June 1993. Our primary source for demographic and education variables is the World Bank's (WB) research database. These annual data are updated and extended on an ongoing basis; ours were received in March 1993. This study focuses on the period 196090, subdivided into three decennial growth periods, 196070, 197080, and 198090. Two types of transformations have been performed on the analysis variables. (1) Per capita GDP is presented as the level at the beginning of the period. This "levels" is actually a threeyear (where possible) average centered on the first year of the period. (2) Per capita GDP and population growth rates are calculated as continuous rates between the endpoint years. X` hp x (#%'0*,.8135@8: