As I say above, following the initial mortality decline all societies are effectively ageing, the ageing is continuous, and at the present time it is hard to identify a natural barrier to this process. In this sense the transition doesn't really seem to have an 'end state', and thus can hardly be called a transition, since the word transition seems to imply something. If there is in fact a transition it is one from a society homeostatically balanced around high mortality to one which is pivoted around low and steadily declining mortality.
Having said this, and in fairness to Lee, what may be meant by ageing is a society with a comparatively high proportion of dependent elderly. On this view the initial mortality decline creates a dependency ratio which is considerably higher than that in the earlier agricultural society. This 'imbalance' takes many years to correct as fertility rates remain high and societies slowly recover the earlier ratios. But equilibrium is not recovered, and dependency ratios once more start to rise, this time amongst the elderly population. So this is what many may mean by ageing societies: societies where elderly dependency ratios rise (and continue to increase) above a certain notional level.
This way of looking at things has a certain validity, but it does beg one very important - indeed possibly critical from a policy perspective - question: just what do we mean by 'old'. The expression, like the terms modern and post-modern is a deceptive one, since it gives the impression of veing carved eternally in time, when in fact it is, of course, an extraordinarily relative one. To give one illustrative example, one populist Turkish politician got himself elected on a promise to introduce male pensions from the age of 43 and female ones from the age of 39 (something which, of course, resulted in the worst pension's crisis in history). He presumeably thought that 43 was 'old' and those who voted him into power evidently agreed. What we consider to be old is a socially defined (and hence relative) concept. It will hold different values at different times, and as life expectancy reaches ever higher limits we can expect our definition to adjust accordingly. This topic however, will have to await a later stage in the argument to receive the elaboration which it deserves. Simply consider this a foretaste.
Whatever the ultimate verdict on the validity of the phases schema, it should be noted that societies which enter the transition later tend to pass through it at an ever increasing rate. This if we take the mortality decline component we can see that gains in life expectancy have occured in the twentieth century in developing countries at rates which are rapidly by historical standards. In India, life expectancy rose from around 24 years in 1920 to 62 years today (a gain of 0.48 years per calendar year over 80 years), while in China, life expectancy rose from 41 in 1950–1955 to 70 in 1995–1999, (a gain of 0.65 years per year over 45 years.(Lee 2003) Fertility transitions since World War II have typically been more rapid than those for the developed countries, with fertility reaching replacement in 20 to 30 years after onset for those countries that have now completed the transition. Fertility transitions in east Asia have been particularly early and rapid, while those in south Asia and Latin America have been slower in starting but now seem to be accelerating (Casterline, 2001, United Nations Population Division, 2003).
Saturday, October 15, 2005
Tuesday, October 11, 2005
The Discovery of Age Structure I
The Discovery of Age Structure
The question of just how demographic change interacts with economic and social development has been debated in the social sciences since time immemorial. In the 18th century the mercantilists held that a large population stimulates economic growth, and this argument has continued to rear its head from time to time across the years, especially in its more refined form that population growth through stimulating demand and investment may stimulate development. In the 19th century Malthus advanced the argument that population growth, by producing decreasing returns in agriculture, leads to lower per capita income. Since this time both these arguments have tended to come and go, with the pendulum swinging now this way, now that.
In general, neo-malthusians (like the Swede Knut Wicksell at the end of the 19th century) have continued to argue that population growth is harmful, while Keynesians have tended to see population growth as a stimulus to investment demand and, thus, to income growth (Perlman 1975). A third, more neutralist view, which has gained considerable influence since the 1970s onwards, argues that population growth rates are not an economically determining factor, one way or the other, and is not a significant variable when it comes to understanding differences in per capita income growth.
More recently this view has been subjected to increasing criticism after having maintained considerable influence all through the 1990s as study after study seemed to reveal little cross-country evidence that would justify thinking there was any significant demographivc effect, either in the form of dividend or in that of penalty. One obvious limitation of much of this population 'neutralist' growth research they largely involved holding virtually every other factor which could conceivably infuence the situation constant in order to test exclusively for correlations between rates of population growth and per capita income growth.
Whether results obtained in this way truly reflect the unimportance of population growth, or, as some have argued, are the summative outcome of the impact of diffent, and mutually offsetting, negative and positive influences of population on economic growth still continues to be debated to this day.
Critics of such studies tend to cite the existence of inadequate control variables or other such model specification errors, or the relatively poor quality of the available data, or the presence of reverse causality, or all of these combined. As a result, and despite a not insignificant quantity of empirical work, the population issue continues to be a relatively open question.
Nonetheless, this earlier body of empirical research surely tended to support what has come to be known as the population neutralist view. This view has been the dominant academic belief in this area since the early 1980s, a state which critics claim has lead to the marginalization of population and reproductive health questions as instruments of economic development within key development agencies like the World Bank (Birdsall 2003, Kelley 2003).
However of late interest in the macroeconomic consequences of population change has renewed itself (and forcefully so). The change in attitudes can be traced to new evidence that comes in three forms.
First, a series of empirical studies based on aggregate level panel data have concluded that demographic factors do in fact have a strong, statistically significant effect on aggregate saving rates (Bloom and others, 2003; Deaton and others, 2000; Kelley and others, 1996; Kinugasa, 2004; Williamson and others, 2001) and on economic growth (Bloom and others, 2001; Bloom and others, 1998; Kelley and others, 1995).
Second, detailed case studies of the East Asian miracle have provided compelling and consistent evidence that what has come to be called the "demographic dividend" was an important element economic success of the region (Bloom and others, 1998; Mason, 2001b; Mason and others, 1999). In a study which really made it impossible to continue to ignore the relevance of demographic changes Bloom and Williamson (1998) using econometric techniques concluded that about one-third of East Asia’s increase in per capita income was due to the demographic dividend. While not everyone has accepted these conclusions at face value (Schultz, 2005), there is little doubt that an increase in the proportion of the population of working age was a significant factor in their growth tramnsition. In a similar vein, Mason (2001a), using growth accounting methods, estimates that the 'demographic dividend' accounted for about a quarter of the Tiger’s economic growth during the 'growth spurt'.
In the third place the combination of having a rapidly increasing proportion of the population over retirement age coupled with an unprecedentedly low level of fertility in the world's second and third largest economies (Japan and Germany) has concentrated attention on the "population topic" as never before.
While it should be obvious, it is perhaps worth noting here that all the above-cited authors argue that even though age-structure variables do have predictive power and can 'explain' a significant portion of economic growth during the development process, the relationship between demographic variables and the economy is not a deterministic one. Population matters, but policy matters to, and whether or not governments put in place an appropriate package of policies makes a big difference between whether the opportunity offered by the demographic dividend is put to productive use or simply frittered. The difference between the experiences of some Latin American countries and some Asian ones is clearly salutory here.
The Coale Hoover Hypothesis
The recent blossoming of interest in the demographic and age-structure related components of growth in fact date-back to the year 1958 and the publication of an influential book by Ansley J. Coale and Edgar Hoover - Population Growth and Economic Development in Low-Income Countries.
The Coale-Hoover hypothesis, as it has come to be called, was based on one simple but powerful intuition: rapid population growth arising from falling infant and child mortality swells the ranks of dependent young, and this single demographic event in and of itself increases consumption at the expense of saving.
Using simulation results from a mathematical model calibrated with Indian data, Coale and Hoover concluded that India's development would be substantially enhanced by lower rates of population growth.
Their analysis rested on two premises. Firstly, that in post initial-mortality-decline 'child-heavy' societies household and aggregate saving is reduced by the generalised presence of large families. And secondly, the existence of such high ratios of dependent children skews aggregate investment away from more self-evidently economically productive activities, since there is a continuous pressure for funds funds to be transferred towards so-called 'unproductive' population-sensitive social expenditures (like health and education).
The key novelty in the Coale Hoover model was to link this 'crowding out' process to the age composition of the high-fertility population, and not simply to its size, density, or growth, per se.
Pioneering in several dimensions, their book:
1) identified several possible theoretical linkages between population and economic growth which were in harmony with the economic-growth paradigms of the time (e.g., an emphasis on physical capital formation);
2) formalized these linkages into a mathematical model that was parameterized and simulated to generate forecasts of alternative fertility scenarios over the intermediate-run;
3) provided a case study of an important country whose prospects were considered by many analysts to be grim. The Coale-Hoover framework was transparent and easy to understand, the assumptions were made explicit and qualified, and the findings were clearly expounded and accessible to a wide readership.
The model identified, and the simulations quantified, three adverse impacts of population growth:
1) capital-shallowing--a reduction in the ratio of capital to labor because there is nothing about population growth per se that increases the rate of saving;
2) age-dependency--an increase in youth-dependency, which raises the requirements for household consumption at the expense of saving, while diminishing the rate of saving;
3) investment diversion--a shift of (mainly government) spending into areas such as health and education at the expense of (assumed-to-be) more productive, growth-oriented investments.
In particular it attracted attention a good deal of attention from economists since its principal focus was on physical capital (as distinct from the Malthusian focus which was on land). At the time physical capital accumulation was considered by many to be 'the' key to economic development.
Their work, which at the time had a substantial impact on U.S. population policy and thinking, did not go unchallenged. In fact over time its impact waned since subsequent empirical research failed to uncover the empirically strong and consistent impacts of population movements on saving in the developing and less developed world which the Coale Hoover thesis was thought to have anticipated.
Nathaniel Leff in an early study using a sample of 74 countries found the log of gross savings rates to be inversely related to the proportion of the population either under 15 or over 64 (Leff, 1969), a finding which appeared to place the youth dependency hypothesis on a solid empirical footing. Subsequent research, however, (Goldberger,1973, Ram, 1982) failed to find confirmation of the dependency hypothesis and researchers even cast doubt on the validity of the empirical methods employed in the initial Leff study.
Contemporaneous theoretical developments also seemed to undermine the foundations of the dependency hypothesis. The principal rival, Tobin’s (1967) life-cycle model, took it as axiomatic that the savings rate should increase with as population growth did. The reason for this is simple, at least in Tobin's original version of the model: faster population growth tilts the age distribution toward young, saving, households and away from older, dissaving ones.
The representative agent version of Robert Solow’s neoclassical growth model also pointed in the same direction, with faster population growth resulting in higher savings rates in response to heightened investment demand (Cass 1965, Phelps 1968, and Solow 1956). However, neither clas of model directly addressed the dynamics of the dividend implicit in the ongoing impact of the demographic transition.
The "age tilt" in Tobin’s steady-state model, while interesting in and of itself, occurs simply because the model takes the decision from the outset to describe a world restricted to active adults and retired dependents. Had the model incorporated the idea of youth dependency a very different tilt-effect would have been produced.
In similar fashion, the standard neoclassical growth models assume exogenously fixed labor participation rates, and, by implication, assume no change endogenously driven in the dependency ratio. Clearly this kind of assumption is strictly speaking 'necessary' if one's objective is the conceptualisation of "steady-state" behaviour, but it precisely fails to capture what one might argue are the key "transitional dynamics" of a continuing process of demographic change, and in so doing it seems to beg the very question it is expected to answer. In effect the whole neoclassical school of models gain in rigour precisely at the price of sacrificing the rich population dynamics which were implicit in Coale and Hoover’s early theorising on the East-Asian demographic transition.
The empirical findings which went against Coale and Hoover are not perhaps entirely surprising. At the household level, the saving impacts they attempted to describe are fundamentally based on a 'life-cycle' conceptualization of behaviour and such a conceptualisation requires a substantial 'forward looking' planning horizon. The behavioural transition which is required also involves a considerable evolution in institutional structure (developed capital markets, reliable pension options etc) in order to make the implementation of such lifetime plans feasible. At the time of Coale and Hoover these conditions simply did not pertain in the vast majority of the then third world countries. For many families living in an agricultural context spending on children represents a form of saving (e.g., parents may expect transfers from their children in old age) and, children, as many studies reveal, can be viewed as a productive asset both in the household and on the farm (Doepke 2004, Doepke and Zilibotti 2005).
New Wave Age Structure Theory
The climate at the end of the 80s, and the evaluation which was made of Coale and Hoover is perhaps well summed-up in the following observation by Angus Deaton (1992): "Although there are some studies that find ... demographic effects, the results are typically not robust, and there is no consensus on the direction of the effect on saving."
As has been said above, in recent years there has been a considerable revival of interest in the Coale-Hoover model, and this despite the known limitations (see eg Williamson 2001), and in fact a revised form of their dependency hypothesis has enjoyed something of a renaissance. The original hypothesis has now evolved into class of explicit economic models that, suitably calibrated, account tolerably well for cross-country savings variations in macro time-series. Almost all recent analyses of macro data confirm Coale-Hoover effects to a greater or lesser extent(Collins 1991; Harrigan 1996; Higgins 1994, 1998a; Kang 1994; Kelley and Schmidt 1995, 1996; Lee, Mason, and Miller 1997; Masson 1990; Taylor 1995; Taylor and Williamson 1994; Webb and Zia 1990; and Williamson 1993).
This succesful renaissance of the Coale Hoover hypothesis is in large measure due to the work of fairly limited number of researchers. What might be termed the 'new wave' of age-structure research possibly begins with a paper by Mason and Fry (1982) Subsequently elaborated further by Mason (1987,1988). These researchers developed what they called a 'variable rate-of-growth effect' model which sought to establish a link between youth dependency ratios and national saving rates. The model relied principally on the insight that given the existence of positive labor productivity growth, younger cohorts enjoy higher permanent incomes and higher consumption than their elders. If consumption is shifted from child-rearing to later, non-childrearing stages of the life cycle, aggregate savings rise with a strength that depends directly on the growth rate of national income. The dependency and lifecyle perspectives are thus unified through the effect that changes in the youth dependency ratio induce in the timing of life-cycle consumption.
A decline in the youth dependency ratio, for example, should cause consumption to be shifted from the childrearing years to later, non-childrearing stages of the lifecycle. As a result, according to their model, the saving rate depends on the product of the youth dependency ratio and the growth rate of national income (the Tobin 'growth-tilt effect'), as well as on the dependency ratio itself (the 'composition effect').
Their findings contained an important qualitative implication for the "classic" dependency model: the demographic "center of gravity" for investment demand should be located earlier in the age distribution than that for savings supply. In particular, investment demand should be more closely related to the share of young (through its connection with labor-force growth), while savings supply should be more closely related to share of mature adults (through its connection with retirement needs). The divergence between these centers of gravity implies that the effects of demographic change on savings and net capital flows will depend on the economy's degree of openness to capital flows.
In an open economy, a population with a heavily child-centred age distribution should exhibit a tendency towards current account deficits: savings are low due to the high youth dependency burden. Later as increasing numbers of young people enter the labour market investment rises in response to higher labour-force growth. Then as the age distribution shifts steadily upwards, the savings supply should increase pushing the current account into surplus.
It is also important to note that the negative coefficients for the elderly need not indicate that they are actually drawing down their stocks of assets. The burden of supporting the elderly (either directly or through transfer payments) might lead to lower saving by younger households. Alternatively, prime-age households with elderly parents might save less in anticipation of bequest receipts (Weil, 1994). The age coefficients are not behavioural parameters which describe the actions of agents belonging to different age groups, but instead capture the relationship between the age distribution and the behaviour of agents of all ages.
The Demographic Dividend
Use of the expression demographic dividend goes back an early Bloom and Williamson paper (1998). As stated above in the paper they use quantitative results from cross-country econometric regressions in an attempt to calculate the contribution made by age structure changes to the 'spurt' in East Asian economic growth. According to their argument the demographic "dividend" leads to opportunities for growth of output per capita for two reasons.
First, there is an impact on total GDP due to a "growth accounting effect": a rising share of the total population in the working-age group increases the ratio of "producers" to "consumers". Obviously this situation contributes positively to the growth of output per capita.
Secondly, they conjecture that age-distributions might also be associated with what they call "behavioral effects", and these in turn exert an influence on the growth of output per capita.
Formally the the accounting effect and the behavioral effects can be decomposed in that growth in per capita income is a function of both growth in the proportion of the population of working age and growth in productivity per worker. The former is what is termed the "compositional effect" and the latter the "behavioural effect".
Simply stated, the demographic dividend occurs when a fall in the birth rate following an initial mortality decline produces changes in the age distribution of a society, and these changes mean that fewer investments are required to meet the needs of the youngest age groups. Thus resources are released which may be used for investment in economic development and for improved family welfare. That is, a falling birth rate makes for a smaller population at the young, dependent, ages and for relatively more people in the adult age groups who comprise the productive labor force. It improves the ratio of productive workers to child dependents in the population. That makes for faster economic growth and fewer burdens on families.
The demographic dividend, however, does not last forever. There is a limited window of opportunity. In time, the age distribution changes again, as the large adult population moves into the older, less-productive age brackets and is followed by the smaller cohorts born during the fertility decline. When this occurs, the dependency ratio rises again, this time involving the need to care for the elderly, rather than the need to take care of the young.
In addition, the dividend is not automatic. While demographic pressures are normally eased when fertility initially falls, some countries will take better advantage of this easing than others. Some countries will act to capitalize upon the released resources and use them effectively, while others will not. However, in time, the window of opportunity closes, those that who have not found the way to take ample advantage of the demographic dividend may well face renewed resource pressures at a time when their ability to respond is weaker than ever.
The demographic dividend is delivered through the operation of several interconnected mechanisms.
Labor Supply
As the demographic transition follows its course the generations of children born during the high fertility years enter adult life and become workers. Women who are now having fewer children than before are released from childrearing responsibilities and are able to take jobs outside of the home; also, as the transition moves forward and years of compulsory education increase youngerwomen tend to be better educated than those in the older cohorts, and are thus more productive once inside the labor force.
Savings
Working-age adults tend to earn more and thus can potentially save more than new-entrants to the labour market. Thus the larger generations who work their way through the labour force as the age pyramid changes favors greater personal and national savings. This ability to save becomes even greater as these "thick cohorts" move into their 40s, especially as in the first instance the generation-span is smaller, and their own children rapidly become wage-earners themselves and hence require less support. Thus personal savings continue to grow and are able to serve as a source of investment funds. Countries thus steadily move from being heavily dependent on external finance, to a position of relative financial self-sufficiency.
Human Capital
Having fewer children normally enhances the health of both mother and child. Female participation in the labor force, in turn, enhances the social status and personal and financial independence of women. Also fewer children normally means fewer and better educated. More investment is allocated to each individual child.
Theory and Modeling
The more general awareness of these issues among economists coincided with the emergence in the 1990s of the empirical "convergence" models of economic growth. Pioneered by Robert Barro, these empirical paradigms postulate the existence of either a universal, or a country-specific, long-run steady state and then proceed to attempt to identify the factors (economic, political, social, institutional, geographic etc) that determine each country's long-run growth rate, and in the shorter-to-intermediate-run the transition to this longer-run state. These models lend themselves to demographic investigation due to the differentiation between short- and long-run impacts.
Three approaches dominate the extensive literature on economic-demographic modeling: simple correlations, production functions, and convergence patterns.
Simple-correlations studies hypothesize that per capita output growth is influenced by various dimensions of demography in the following fashion: Y/Ngr = f(D),
Production Functions
Production-function studies are based on estimating variants of a model:
Y = g(K, L, H, R, T),
Convergence Patterns
Convergence-patterns studies, rooted in neoclassical growth theory, explore the relationships between economic growth and the level of economic development. They focus on the pace at which countries move from their current level of labor productivity to their long-run, or steady-state equilibrium level of labor productivity.
A revealing feature of the convergence-patterns models can be seen if we take a closer look at some of the variables which have been omitted by Barro and those working in this tradition. In the majority of papers authors emphasize variables that determine longrun, or "potential" labor productivity, and downplay variables that bring about the "adjustment" or transition to long-run equilibrium. An example of one such omitted variable would be the investment share. Leaving aside for the moment the issue of endogeneity, investment can be viewed as an adjustment variable.
Indeed investment might be considered to be the first variable one would think to include in a model which focuses on labour productivity. Levine and Renelt (1992) surveyed numerous empirical growth studies to identify a common set of influential variables and found, somewhat unsurprisingly, that investment rates constituted the single most robust variable. Rather than implying that the investment rate is a viable Z variable for predicting long-run capital-to-output ratios, however, the significance of this finding is that it suggests an incomplete set of Z variables. If the convergence hypothesis were correct and the list of Z variables were complete, the investment coefficient would in theory be insignificant (see Bloom, Canning, and Malaney (2001), Higgins and Williamson (1994), and Kelley and Schmidt (1994)).
The gap between current and long-run labor force productivity largely dictates the return to investment. In theory investment should flow to those countries which exhibit the highest returns. So, rather than investment accounting for growth per se, it could be argued that the it is the "structural" features of a country (among which one could incorporate its demographic strucure) which either impede or facilitate investment that ultimately determine growth.
Attempts to incorporate demography into convergence models have been few and normally relaively ad hoc. Demographic variables that qualify are those that affect long run labour productivity, and those that condition the transition to it. The nature of these two types of demographic variables can be illustrated by examining Barro (1997) and Kelley and Schmidt (1994), both of which highlight long-run impacts of demography but exclude a role for transitions, Kelley and Schmidt (1995) which examines both long-run and transition impacts,and Bloom and Williamson (1998), which captures primarily the transition impacts.
Empirical analysis was enrichened in the 1990s with the emergence of a theoretical framework by Robert J. Barro (Barro 1991) that incorporates demography into convergence (or technologygap) models. He and collaborators concluded that high fertility, population growth, and mortality all exert negative impacts on per capita output growth. In 1994 Kelley and Schmidt extended this list to include population density and size, which revealed positive impacts, although a net negative assessment of combined demographic trends represented the bottom line.
Barro (1997) focuses on a single demographic variable, the Total Fertility Rate (TFR). This variable captures both the adverse capital-shallowing impact of more rapid population growth, and the resource costs of raising children versus producing other goods and services. By its very nature the TFR exerts its impacts mainly on long-run labor productivity, to the detriment of the short-run transitions en route to equilibrium. TFR is, after all, a hypothetical construct that represents what the fertility rate "would be" if the current age-specific fertility rates were maintained over a long period of time.
Bloom and Williamson
In the late 1990s there was a further evolution in the convergence modeling of demography by several Harvard economists (e.g., David Bloom, David Canning, Jeffrey Sachs, and Jeffrey Williamson). Building on the Barro setup (albeit with a different choice of core variables), the Harvard framework focused on population impacts that take place due to imbalanced age-structure changes over the Demographic Transition. Their modeling compactly captured these impacts by just two variables: population growth (Ngr) and working-age growth (WAgr). Such a specification neatly follows from an identity that “translates” a traditional neoclassical model formulated in output per worker growth into a comparable model formulated in output per capita growth. While by construction such a translation represents accounting, it nevertheless provides a way of exposing some shorter-period “population impacts” within the usual long-run neoclassical framework. Assessments of such accounting impacts of demography in numerous Harvard empirical papers are shown to be sizeable, especially in East Asia.
Much of this work takes as its starting point an alternative methodology for exposing such dynamic relationships which was advanced by Bloom and Williamson (Bloom and williamson, 1998, BW hereafter). Demography in the BW model follows neatly from a definition that translates the convergence model from one that explains productivity growth into one that explains per capita output growth. In the Bloom and Williamson version of the demographic dividend thesis, the accounting and behavioral effects of the age transition can be decomposed using fairly straightforward econometric techniques which involve the formulation of an identity function, the taking of natural logarithms of output per capita and the differentiation of the resulting expression with respect to time.
Starting with the definition of output per labor hour, it can be shown that the basic model can be transformed into one which describes the growth process in per-capita terms. Now the impacts of working-hour growth and population growth cancel each other out when they change at the same rate, something which certainly occurs in a steady-state growth situation with a static age pyramid. This condition is normally imposed by assumption in most empirical studies. BW note that the 1960s, 1970s, and 1980s were periods of demographic transition for most developing countries. As a result, neither condition held and differential growth rates should be observed. This of course is what in parctice was the case.
In the BW setup, the workforce share in fact has no impact on output growth. BW replace labour force growth with a pure demographic proxy, the growth rate of the working-age population. That is, as if the only determinant of hours worked were the age-distribution of the population, hence the relative growth of the working-age versus full population constitutes the sole impact of demography in their model. Following the BW setup, sometimes the impact of demography will be positive, sometimes negative, and sometimes zero. The model thus highlights the reality that demographic impacts vary during the transition but is silent on the issue of possible demographic impacts on long-run labor productivity; i.e., demography does not affect the Barro Z's.
As a result, the BW model has a narrower interpretation than most renderings in the recent literature, which often admit both short- and long-run impacts of demographic change as a part of their theoretical structure. On the other hand, it has the desirable attribute of clarity in its interpretation.
Second, consider the focus of the BW model on the transitional impacts of demographic change. The postulated relations between population growth and working age growth, respectively, allow a clear interpretation for the role of demography: relatively rapid growth of the working-age population will speed the transition to long-run economic prosperity. However, two countries with the same Z’s will ultimately arrive at the same level of labour productivity growth, irrespective of their demography.
BW acknowledge the possibility that the rates of growth of working age and total population might impact long run labour productivity, but they fail to model this explicitly. Nor do they include other demographic variables among the Z’s. Rather, they limit themselves to noting that long run influences could result in coefficient estimates which deviate from unity.
While Bloom & Williamson (1998) focus on output per capita as the dependent variable, the recent literature argues that an understanding of international differences in output per worker is needed, since only workers contribute to production. From regressions with the growth rate of output per capita as the dependent variable it is rather complicated to assess behavioral effects (see Bloom et al. (2001) for an assessment of behavioral effects from such regressions).
Kelley and Schmidt (2001) for example demonstrated the implication of this point by evaluating the net impact of demographic change on economic growth in eight empirical renderings.
References
Barro, Robert J. and Jong-Wha Lee. 1993. "Losers and Winners in Economic Growth." NBER Working Paper No. 4341. Cambridge, Mass.
Barro, R.J. 1991. “Economic Growth in a Cross Section of Countries.” Quarterly Journal of Economics 106(2):407-444
Birdsall, Nancy, 2003, "New Findings in Economics and Demography: Implications for Policies to Reduce Poverty" in Population Matters: Demographic Change, Economic Growth, and Poverty in the Developing World Edited by Nancy Birdsall, Allen C. Kelley, Steven W. Sinding, Oxford, Oxford University Press
Bloom D, Canning D, Moore M (2004). The effect of improvements in health and longevity on optimal retirement and saving. NBER Working Paper No. 10919. Cambridge, National Bureau of Economic Research.
Bloom D, Canning D, Graham B (2003). Longevity and life-cycle savings. Scandinavian
Journal of Economics, 105:319-338.
Bloom, D. E., D. Canning and Sevilla (2002). The Demographic Dividend: A New Perspective on the Economic Consequences of Population Change. Santa Monica, CA, RAND.
Bloom, D.D., Canning, D. & Malaney, P. (2001). Demographic change and economic growth in Asia, Population and Development Review, 26, supp., 257{290.
Bloom, D. E. and J. G. Williamson (1998). Demographic Transitions and Economic Miracles in Emerging Asia. World Bank Economic Review vol.12, No.3, pp. 419-56.
Cutler, D. M., J. M. Poterba, et al. 1990. "An Aging Society: Opportunity or Challenge?" Brookings Papers on Economic Activity No. 1, pp.1-56.
Collins, Susan. 1991. “Saving Behavior in Ten Developing Countries.” In B. Douglas
Bernheim and John B. Shoven, eds., National Saving and Economic Performance.
Chicago: University of Chicago Press.
Deaton, A. and C. H. Paxson (2000). Growth, demographic structure, and national saving in Taiwan. In Population and Economic Change in East Asia, A Supplement to Population and Development Review. R. Lee and C. Y. C. Chu, eds. New York, Population Council. vol. 26, pp.141-173.
Doepke, M. 2004, “Accounting for Fertility Decline During the Transition to Growth,” Journal of Economic Growth, September 2004, 9 (3), 347–83.
Doepke, M and F. Zilibotti 2005, "The Macroeconomics of Child-Labor Regulation.". American Economic Review, December 2005.
Fry, Maxwell, and Andrew Mason. 1982. “The Variable Rate-of-Growth Effect in the
Life-Cycle Saving Model: Children, Capital Inflows, Interest, and Growth in a New
Specification of the Life-Cycle Model Applied to Seven Asian Developing Countries.”
Economic Inquiry 20(3, July):426–42.
Goldberger, Arthur. 1973. “Dependency Rates and Savings Rates: Further Comment.”
American Economic Review 63(1, March):232–33.
Harrigan, Frank. 1996. “Saving Transitions in Southeast Asia.” Asian Development
Bank, Manila. Processed.
Higgins, Matthew 1998. “Demography, National Savings, and International Capital Flows.” International Economic Review 39(2, May):343–69.
Higgins, Matthew. 1994. “The Demographic Determinants of Savings, Investment, and
International Capital Flows.” Ph.D. diss., Harvard University, Cambridge, Mass.
Higgins M, and Williamson J (1997). Age structure dynamics in Asia and dependence on foreign capital. Population and Development Review, 23:261-293.
Higgins, Matthew, and Jeffrey G. Williamson. 1996. “Asian Demography and Foreign
Capital Dependence.” NBER Working Paper 5560. National Bureau of Economic Research,
Cambridge, Mass.
Kang, Kenneth. 1994. “Why Did Koreans Save So Little and Why Do They Now Save
So Much?” International Economic Journal 8(4, winter):99–111.
Kelley, A.C. 2003, "The Population Debate in Historical Perspective: Revisionism Revised" in Population Matters: Demographic Change, Economic Growth, and Poverty in the Developing World Edited by Nancy Birdsall, Allen C. Kelley, Steven W. Sinding, Oxford, Oxford University Press
Kelley, Allen C., and Robert M. Schmidt. 1995. “Aggregate Population and Economic
Growth Correlations: The Role of the Components of Demographic Change.” Demography
32(4):543–55.
Kelley AC, Schmidt RM (1996). Saving, dependency, and development. Journal of Population Economics, 9:365-386.
Krugman, Paul. 1994. “The Myth of Asia’s Miracle.” Foreign Affairs 73(6, November-
December):62–78.
Lee RD, Mason A, Miller T (2000). Life-cycle saving and the demographic transition: the case of Taiwan. Population and Development Review, 26(Supplement):194–219.
Lee, Ronald, Andrew Mason, et al. (2003). From transfers to individual responsibility: Implications for savings and capital accumulation in Taiwan and the United States. Scandinavian Journal of Economics vol.105, No.3, pp.339-357.
Lee, Ronald, Andrew Mason, and Timothy Miller. 1997. “Saving, Wealth, and the Demographic Transition in East Asia.” Paper presented at the conference on population
and the Asian economic miracle, East-West Center, Honolulu, Hawaii, 7–10 January.
Processed.
Leff NH (1969). Dependency rates and savings rates. American Economic Review, 59: 886-896.
Mason A (1988). Saving, economic growth, and demographic change. Population and Development Review, 14:113-144.
Mason A (1987). National saving rates and population growth: a new model and new evidence. In: Johnson DG, Lee RD, eds. Population growth and economic development: issues and evidence. University of Wisconsin Press, Madison.
Mason, Andrew (1988). Saving, economic growth, and demographic change. Population and
Development Review vol.14, No. 1, pp. 113-44.
Mason, Andrew (2001). Population Change and Economic Development in East Asia: Challenges Met, Opportunities Seized. Stanford, Stanford University Press.
Masson, Paul. 1990. “Long-Term Macroeconomic Effects of Aging Populations.” Finance
and Development 27(2, June):6–9.
Perlman, M. 1975, Some Economic Growth Problems and the Part Population Policy Plays.
Quarterly Journal of Economics, 1975 Volume 89, No2, Pages: 247-56
Persson, Joakim, 2004, The Population Age Distribution, Human Capital, and Economic Growth: The U.S. states 1930-2000, Department of Economics and Statistics, Örebro University, Sweden, Working Paper
Ram, Rati. 1982. “Dependency Rates and Aggregate Savings: A New International Cross-
Section Study.” American Economic Review 72(3, June):537–44.
Schultz, T. Paul, 2005, Demographic Determinants of Savings: Estimating and Interpreting the Aggregate Association in Asia, IZA Discussion Paper No. 1479,
January 2005
Taylor, Alan. 1995. “Debt, Dependence, and the Demographic Transition: Latin America
into the Next Century.” World Development 23(5, May):869–79.
Taylor, Alan, and Jeffrey G. Williamson. 1994. “Capital Flows to the New World as an
Intergenerational Transfer.” Journal of Political Economy 102(2, April):348–69.
Tobin, James. 1967. “Life-Cycle Savings and Balanced Economic Growth.” In William
Fellner, ed., Ten Essays in the Tradition of Irving Fischer. New York: Wiley Press
Tsai IJ, Chu CYC, Chung CF (2000). Demographic transition and household saving in Taiwan. Population and Development Review, 26(Supplement):174-193.
Webb, Steven, and Heidi Zia. 1990. “Lower Birth Rates = Higher Saving in LDCs.”
Finance and Development 27(2, June):12–14.
Williamson, J.G. (2001), Demographic shocks and global factor flows. In J.N. Little and R.K. Triest eds. Seismic Shifts: The Economic Impact of Demographic Change (Boston, Mass.: Federal Reserve Bank of Boston).
Williamson, Jeffrey G. 1993. “Human Capital Deepening, Inequality, and Demographic Events Along the Asia-Pacific Rim.” In Gavin Jones, Naohiro Ogawa, and Jeffrey G. Williamson, eds., Human Resources and Development along the Asia-Pacific Rim. New York: Oxford University Press.
Young A (1994). Lessons from the East Asian NIC’s: a Contrarian view. European Economic Review, 38:964-973. Young A (1995). The tyranny of numbers: confronting the statistical realities of the East Asian growth experience. Quarterly Journal of Economics, 110:641-680.
The question of just how demographic change interacts with economic and social development has been debated in the social sciences since time immemorial. In the 18th century the mercantilists held that a large population stimulates economic growth, and this argument has continued to rear its head from time to time across the years, especially in its more refined form that population growth through stimulating demand and investment may stimulate development. In the 19th century Malthus advanced the argument that population growth, by producing decreasing returns in agriculture, leads to lower per capita income. Since this time both these arguments have tended to come and go, with the pendulum swinging now this way, now that.
In general, neo-malthusians (like the Swede Knut Wicksell at the end of the 19th century) have continued to argue that population growth is harmful, while Keynesians have tended to see population growth as a stimulus to investment demand and, thus, to income growth (Perlman 1975). A third, more neutralist view, which has gained considerable influence since the 1970s onwards, argues that population growth rates are not an economically determining factor, one way or the other, and is not a significant variable when it comes to understanding differences in per capita income growth.
More recently this view has been subjected to increasing criticism after having maintained considerable influence all through the 1990s as study after study seemed to reveal little cross-country evidence that would justify thinking there was any significant demographivc effect, either in the form of dividend or in that of penalty. One obvious limitation of much of this population 'neutralist' growth research they largely involved holding virtually every other factor which could conceivably infuence the situation constant in order to test exclusively for correlations between rates of population growth and per capita income growth.
Whether results obtained in this way truly reflect the unimportance of population growth, or, as some have argued, are the summative outcome of the impact of diffent, and mutually offsetting, negative and positive influences of population on economic growth still continues to be debated to this day.
Critics of such studies tend to cite the existence of inadequate control variables or other such model specification errors, or the relatively poor quality of the available data, or the presence of reverse causality, or all of these combined. As a result, and despite a not insignificant quantity of empirical work, the population issue continues to be a relatively open question.
Nonetheless, this earlier body of empirical research surely tended to support what has come to be known as the population neutralist view. This view has been the dominant academic belief in this area since the early 1980s, a state which critics claim has lead to the marginalization of population and reproductive health questions as instruments of economic development within key development agencies like the World Bank (Birdsall 2003, Kelley 2003).
However of late interest in the macroeconomic consequences of population change has renewed itself (and forcefully so). The change in attitudes can be traced to new evidence that comes in three forms.
First, a series of empirical studies based on aggregate level panel data have concluded that demographic factors do in fact have a strong, statistically significant effect on aggregate saving rates (Bloom and others, 2003; Deaton and others, 2000; Kelley and others, 1996; Kinugasa, 2004; Williamson and others, 2001) and on economic growth (Bloom and others, 2001; Bloom and others, 1998; Kelley and others, 1995).
Second, detailed case studies of the East Asian miracle have provided compelling and consistent evidence that what has come to be called the "demographic dividend" was an important element economic success of the region (Bloom and others, 1998; Mason, 2001b; Mason and others, 1999). In a study which really made it impossible to continue to ignore the relevance of demographic changes Bloom and Williamson (1998) using econometric techniques concluded that about one-third of East Asia’s increase in per capita income was due to the demographic dividend. While not everyone has accepted these conclusions at face value (Schultz, 2005), there is little doubt that an increase in the proportion of the population of working age was a significant factor in their growth tramnsition. In a similar vein, Mason (2001a), using growth accounting methods, estimates that the 'demographic dividend' accounted for about a quarter of the Tiger’s economic growth during the 'growth spurt'.
In the third place the combination of having a rapidly increasing proportion of the population over retirement age coupled with an unprecedentedly low level of fertility in the world's second and third largest economies (Japan and Germany) has concentrated attention on the "population topic" as never before.
While it should be obvious, it is perhaps worth noting here that all the above-cited authors argue that even though age-structure variables do have predictive power and can 'explain' a significant portion of economic growth during the development process, the relationship between demographic variables and the economy is not a deterministic one. Population matters, but policy matters to, and whether or not governments put in place an appropriate package of policies makes a big difference between whether the opportunity offered by the demographic dividend is put to productive use or simply frittered. The difference between the experiences of some Latin American countries and some Asian ones is clearly salutory here.
The Coale Hoover Hypothesis
The recent blossoming of interest in the demographic and age-structure related components of growth in fact date-back to the year 1958 and the publication of an influential book by Ansley J. Coale and Edgar Hoover - Population Growth and Economic Development in Low-Income Countries.
The Coale-Hoover hypothesis, as it has come to be called, was based on one simple but powerful intuition: rapid population growth arising from falling infant and child mortality swells the ranks of dependent young, and this single demographic event in and of itself increases consumption at the expense of saving.
Using simulation results from a mathematical model calibrated with Indian data, Coale and Hoover concluded that India's development would be substantially enhanced by lower rates of population growth.
Their analysis rested on two premises. Firstly, that in post initial-mortality-decline 'child-heavy' societies household and aggregate saving is reduced by the generalised presence of large families. And secondly, the existence of such high ratios of dependent children skews aggregate investment away from more self-evidently economically productive activities, since there is a continuous pressure for funds funds to be transferred towards so-called 'unproductive' population-sensitive social expenditures (like health and education).
The key novelty in the Coale Hoover model was to link this 'crowding out' process to the age composition of the high-fertility population, and not simply to its size, density, or growth, per se.
Pioneering in several dimensions, their book:
1) identified several possible theoretical linkages between population and economic growth which were in harmony with the economic-growth paradigms of the time (e.g., an emphasis on physical capital formation);
2) formalized these linkages into a mathematical model that was parameterized and simulated to generate forecasts of alternative fertility scenarios over the intermediate-run;
3) provided a case study of an important country whose prospects were considered by many analysts to be grim. The Coale-Hoover framework was transparent and easy to understand, the assumptions were made explicit and qualified, and the findings were clearly expounded and accessible to a wide readership.
The model identified, and the simulations quantified, three adverse impacts of population growth:
1) capital-shallowing--a reduction in the ratio of capital to labor because there is nothing about population growth per se that increases the rate of saving;
2) age-dependency--an increase in youth-dependency, which raises the requirements for household consumption at the expense of saving, while diminishing the rate of saving;
3) investment diversion--a shift of (mainly government) spending into areas such as health and education at the expense of (assumed-to-be) more productive, growth-oriented investments.
In particular it attracted attention a good deal of attention from economists since its principal focus was on physical capital (as distinct from the Malthusian focus which was on land). At the time physical capital accumulation was considered by many to be 'the' key to economic development.
Their work, which at the time had a substantial impact on U.S. population policy and thinking, did not go unchallenged. In fact over time its impact waned since subsequent empirical research failed to uncover the empirically strong and consistent impacts of population movements on saving in the developing and less developed world which the Coale Hoover thesis was thought to have anticipated.
Nathaniel Leff in an early study using a sample of 74 countries found the log of gross savings rates to be inversely related to the proportion of the population either under 15 or over 64 (Leff, 1969), a finding which appeared to place the youth dependency hypothesis on a solid empirical footing. Subsequent research, however, (Goldberger,1973, Ram, 1982) failed to find confirmation of the dependency hypothesis and researchers even cast doubt on the validity of the empirical methods employed in the initial Leff study.
Contemporaneous theoretical developments also seemed to undermine the foundations of the dependency hypothesis. The principal rival, Tobin’s (1967) life-cycle model, took it as axiomatic that the savings rate should increase with as population growth did. The reason for this is simple, at least in Tobin's original version of the model: faster population growth tilts the age distribution toward young, saving, households and away from older, dissaving ones.
The representative agent version of Robert Solow’s neoclassical growth model also pointed in the same direction, with faster population growth resulting in higher savings rates in response to heightened investment demand (Cass 1965, Phelps 1968, and Solow 1956). However, neither clas of model directly addressed the dynamics of the dividend implicit in the ongoing impact of the demographic transition.
The "age tilt" in Tobin’s steady-state model, while interesting in and of itself, occurs simply because the model takes the decision from the outset to describe a world restricted to active adults and retired dependents. Had the model incorporated the idea of youth dependency a very different tilt-effect would have been produced.
In similar fashion, the standard neoclassical growth models assume exogenously fixed labor participation rates, and, by implication, assume no change endogenously driven in the dependency ratio. Clearly this kind of assumption is strictly speaking 'necessary' if one's objective is the conceptualisation of "steady-state" behaviour, but it precisely fails to capture what one might argue are the key "transitional dynamics" of a continuing process of demographic change, and in so doing it seems to beg the very question it is expected to answer. In effect the whole neoclassical school of models gain in rigour precisely at the price of sacrificing the rich population dynamics which were implicit in Coale and Hoover’s early theorising on the East-Asian demographic transition.
The empirical findings which went against Coale and Hoover are not perhaps entirely surprising. At the household level, the saving impacts they attempted to describe are fundamentally based on a 'life-cycle' conceptualization of behaviour and such a conceptualisation requires a substantial 'forward looking' planning horizon. The behavioural transition which is required also involves a considerable evolution in institutional structure (developed capital markets, reliable pension options etc) in order to make the implementation of such lifetime plans feasible. At the time of Coale and Hoover these conditions simply did not pertain in the vast majority of the then third world countries. For many families living in an agricultural context spending on children represents a form of saving (e.g., parents may expect transfers from their children in old age) and, children, as many studies reveal, can be viewed as a productive asset both in the household and on the farm (Doepke 2004, Doepke and Zilibotti 2005).
New Wave Age Structure Theory
The climate at the end of the 80s, and the evaluation which was made of Coale and Hoover is perhaps well summed-up in the following observation by Angus Deaton (1992): "Although there are some studies that find ... demographic effects, the results are typically not robust, and there is no consensus on the direction of the effect on saving."
As has been said above, in recent years there has been a considerable revival of interest in the Coale-Hoover model, and this despite the known limitations (see eg Williamson 2001), and in fact a revised form of their dependency hypothesis has enjoyed something of a renaissance. The original hypothesis has now evolved into class of explicit economic models that, suitably calibrated, account tolerably well for cross-country savings variations in macro time-series. Almost all recent analyses of macro data confirm Coale-Hoover effects to a greater or lesser extent(Collins 1991; Harrigan 1996; Higgins 1994, 1998a; Kang 1994; Kelley and Schmidt 1995, 1996; Lee, Mason, and Miller 1997; Masson 1990; Taylor 1995; Taylor and Williamson 1994; Webb and Zia 1990; and Williamson 1993).
This succesful renaissance of the Coale Hoover hypothesis is in large measure due to the work of fairly limited number of researchers. What might be termed the 'new wave' of age-structure research possibly begins with a paper by Mason and Fry (1982) Subsequently elaborated further by Mason (1987,1988). These researchers developed what they called a 'variable rate-of-growth effect' model which sought to establish a link between youth dependency ratios and national saving rates. The model relied principally on the insight that given the existence of positive labor productivity growth, younger cohorts enjoy higher permanent incomes and higher consumption than their elders. If consumption is shifted from child-rearing to later, non-childrearing stages of the life cycle, aggregate savings rise with a strength that depends directly on the growth rate of national income. The dependency and lifecyle perspectives are thus unified through the effect that changes in the youth dependency ratio induce in the timing of life-cycle consumption.
A decline in the youth dependency ratio, for example, should cause consumption to be shifted from the childrearing years to later, non-childrearing stages of the lifecycle. As a result, according to their model, the saving rate depends on the product of the youth dependency ratio and the growth rate of national income (the Tobin 'growth-tilt effect'), as well as on the dependency ratio itself (the 'composition effect').
Their findings contained an important qualitative implication for the "classic" dependency model: the demographic "center of gravity" for investment demand should be located earlier in the age distribution than that for savings supply. In particular, investment demand should be more closely related to the share of young (through its connection with labor-force growth), while savings supply should be more closely related to share of mature adults (through its connection with retirement needs). The divergence between these centers of gravity implies that the effects of demographic change on savings and net capital flows will depend on the economy's degree of openness to capital flows.
In an open economy, a population with a heavily child-centred age distribution should exhibit a tendency towards current account deficits: savings are low due to the high youth dependency burden. Later as increasing numbers of young people enter the labour market investment rises in response to higher labour-force growth. Then as the age distribution shifts steadily upwards, the savings supply should increase pushing the current account into surplus.
It is also important to note that the negative coefficients for the elderly need not indicate that they are actually drawing down their stocks of assets. The burden of supporting the elderly (either directly or through transfer payments) might lead to lower saving by younger households. Alternatively, prime-age households with elderly parents might save less in anticipation of bequest receipts (Weil, 1994). The age coefficients are not behavioural parameters which describe the actions of agents belonging to different age groups, but instead capture the relationship between the age distribution and the behaviour of agents of all ages.
The Demographic Dividend
Use of the expression demographic dividend goes back an early Bloom and Williamson paper (1998). As stated above in the paper they use quantitative results from cross-country econometric regressions in an attempt to calculate the contribution made by age structure changes to the 'spurt' in East Asian economic growth. According to their argument the demographic "dividend" leads to opportunities for growth of output per capita for two reasons.
First, there is an impact on total GDP due to a "growth accounting effect": a rising share of the total population in the working-age group increases the ratio of "producers" to "consumers". Obviously this situation contributes positively to the growth of output per capita.
Secondly, they conjecture that age-distributions might also be associated with what they call "behavioral effects", and these in turn exert an influence on the growth of output per capita.
Formally the the accounting effect and the behavioral effects can be decomposed in that growth in per capita income is a function of both growth in the proportion of the population of working age and growth in productivity per worker. The former is what is termed the "compositional effect" and the latter the "behavioural effect".
Simply stated, the demographic dividend occurs when a fall in the birth rate following an initial mortality decline produces changes in the age distribution of a society, and these changes mean that fewer investments are required to meet the needs of the youngest age groups. Thus resources are released which may be used for investment in economic development and for improved family welfare. That is, a falling birth rate makes for a smaller population at the young, dependent, ages and for relatively more people in the adult age groups who comprise the productive labor force. It improves the ratio of productive workers to child dependents in the population. That makes for faster economic growth and fewer burdens on families.
The demographic dividend, however, does not last forever. There is a limited window of opportunity. In time, the age distribution changes again, as the large adult population moves into the older, less-productive age brackets and is followed by the smaller cohorts born during the fertility decline. When this occurs, the dependency ratio rises again, this time involving the need to care for the elderly, rather than the need to take care of the young.
In addition, the dividend is not automatic. While demographic pressures are normally eased when fertility initially falls, some countries will take better advantage of this easing than others. Some countries will act to capitalize upon the released resources and use them effectively, while others will not. However, in time, the window of opportunity closes, those that who have not found the way to take ample advantage of the demographic dividend may well face renewed resource pressures at a time when their ability to respond is weaker than ever.
The demographic dividend is delivered through the operation of several interconnected mechanisms.
Labor Supply
As the demographic transition follows its course the generations of children born during the high fertility years enter adult life and become workers. Women who are now having fewer children than before are released from childrearing responsibilities and are able to take jobs outside of the home; also, as the transition moves forward and years of compulsory education increase youngerwomen tend to be better educated than those in the older cohorts, and are thus more productive once inside the labor force.
Savings
Working-age adults tend to earn more and thus can potentially save more than new-entrants to the labour market. Thus the larger generations who work their way through the labour force as the age pyramid changes favors greater personal and national savings. This ability to save becomes even greater as these "thick cohorts" move into their 40s, especially as in the first instance the generation-span is smaller, and their own children rapidly become wage-earners themselves and hence require less support. Thus personal savings continue to grow and are able to serve as a source of investment funds. Countries thus steadily move from being heavily dependent on external finance, to a position of relative financial self-sufficiency.
Human Capital
Having fewer children normally enhances the health of both mother and child. Female participation in the labor force, in turn, enhances the social status and personal and financial independence of women. Also fewer children normally means fewer and better educated. More investment is allocated to each individual child.
Theory and Modeling
The more general awareness of these issues among economists coincided with the emergence in the 1990s of the empirical "convergence" models of economic growth. Pioneered by Robert Barro, these empirical paradigms postulate the existence of either a universal, or a country-specific, long-run steady state and then proceed to attempt to identify the factors (economic, political, social, institutional, geographic etc) that determine each country's long-run growth rate, and in the shorter-to-intermediate-run the transition to this longer-run state. These models lend themselves to demographic investigation due to the differentiation between short- and long-run impacts.
Three approaches dominate the extensive literature on economic-demographic modeling: simple correlations, production functions, and convergence patterns.
Simple-correlations studies hypothesize that per capita output growth is influenced by various dimensions of demography in the following fashion: Y/Ngr = f(D),
Production Functions
Production-function studies are based on estimating variants of a model:
Y = g(K, L, H, R, T),
Convergence Patterns
Convergence-patterns studies, rooted in neoclassical growth theory, explore the relationships between economic growth and the level of economic development. They focus on the pace at which countries move from their current level of labor productivity to their long-run, or steady-state equilibrium level of labor productivity.
A revealing feature of the convergence-patterns models can be seen if we take a closer look at some of the variables which have been omitted by Barro and those working in this tradition. In the majority of papers authors emphasize variables that determine longrun, or "potential" labor productivity, and downplay variables that bring about the "adjustment" or transition to long-run equilibrium. An example of one such omitted variable would be the investment share. Leaving aside for the moment the issue of endogeneity, investment can be viewed as an adjustment variable.
Indeed investment might be considered to be the first variable one would think to include in a model which focuses on labour productivity. Levine and Renelt (1992) surveyed numerous empirical growth studies to identify a common set of influential variables and found, somewhat unsurprisingly, that investment rates constituted the single most robust variable. Rather than implying that the investment rate is a viable Z variable for predicting long-run capital-to-output ratios, however, the significance of this finding is that it suggests an incomplete set of Z variables. If the convergence hypothesis were correct and the list of Z variables were complete, the investment coefficient would in theory be insignificant (see Bloom, Canning, and Malaney (2001), Higgins and Williamson (1994), and Kelley and Schmidt (1994)).
The gap between current and long-run labor force productivity largely dictates the return to investment. In theory investment should flow to those countries which exhibit the highest returns. So, rather than investment accounting for growth per se, it could be argued that the it is the "structural" features of a country (among which one could incorporate its demographic strucure) which either impede or facilitate investment that ultimately determine growth.
Attempts to incorporate demography into convergence models have been few and normally relaively ad hoc. Demographic variables that qualify are those that affect long run labour productivity, and those that condition the transition to it. The nature of these two types of demographic variables can be illustrated by examining Barro (1997) and Kelley and Schmidt (1994), both of which highlight long-run impacts of demography but exclude a role for transitions, Kelley and Schmidt (1995) which examines both long-run and transition impacts,and Bloom and Williamson (1998), which captures primarily the transition impacts.
Empirical analysis was enrichened in the 1990s with the emergence of a theoretical framework by Robert J. Barro (Barro 1991) that incorporates demography into convergence (or technologygap) models. He and collaborators concluded that high fertility, population growth, and mortality all exert negative impacts on per capita output growth. In 1994 Kelley and Schmidt extended this list to include population density and size, which revealed positive impacts, although a net negative assessment of combined demographic trends represented the bottom line.
Barro (1997) focuses on a single demographic variable, the Total Fertility Rate (TFR). This variable captures both the adverse capital-shallowing impact of more rapid population growth, and the resource costs of raising children versus producing other goods and services. By its very nature the TFR exerts its impacts mainly on long-run labor productivity, to the detriment of the short-run transitions en route to equilibrium. TFR is, after all, a hypothetical construct that represents what the fertility rate "would be" if the current age-specific fertility rates were maintained over a long period of time.
Bloom and Williamson
In the late 1990s there was a further evolution in the convergence modeling of demography by several Harvard economists (e.g., David Bloom, David Canning, Jeffrey Sachs, and Jeffrey Williamson). Building on the Barro setup (albeit with a different choice of core variables), the Harvard framework focused on population impacts that take place due to imbalanced age-structure changes over the Demographic Transition. Their modeling compactly captured these impacts by just two variables: population growth (Ngr) and working-age growth (WAgr). Such a specification neatly follows from an identity that “translates” a traditional neoclassical model formulated in output per worker growth into a comparable model formulated in output per capita growth. While by construction such a translation represents accounting, it nevertheless provides a way of exposing some shorter-period “population impacts” within the usual long-run neoclassical framework. Assessments of such accounting impacts of demography in numerous Harvard empirical papers are shown to be sizeable, especially in East Asia.
Much of this work takes as its starting point an alternative methodology for exposing such dynamic relationships which was advanced by Bloom and Williamson (Bloom and williamson, 1998, BW hereafter). Demography in the BW model follows neatly from a definition that translates the convergence model from one that explains productivity growth into one that explains per capita output growth. In the Bloom and Williamson version of the demographic dividend thesis, the accounting and behavioral effects of the age transition can be decomposed using fairly straightforward econometric techniques which involve the formulation of an identity function, the taking of natural logarithms of output per capita and the differentiation of the resulting expression with respect to time.
Starting with the definition of output per labor hour, it can be shown that the basic model can be transformed into one which describes the growth process in per-capita terms. Now the impacts of working-hour growth and population growth cancel each other out when they change at the same rate, something which certainly occurs in a steady-state growth situation with a static age pyramid. This condition is normally imposed by assumption in most empirical studies. BW note that the 1960s, 1970s, and 1980s were periods of demographic transition for most developing countries. As a result, neither condition held and differential growth rates should be observed. This of course is what in parctice was the case.
In the BW setup, the workforce share in fact has no impact on output growth. BW replace labour force growth with a pure demographic proxy, the growth rate of the working-age population. That is, as if the only determinant of hours worked were the age-distribution of the population, hence the relative growth of the working-age versus full population constitutes the sole impact of demography in their model. Following the BW setup, sometimes the impact of demography will be positive, sometimes negative, and sometimes zero. The model thus highlights the reality that demographic impacts vary during the transition but is silent on the issue of possible demographic impacts on long-run labor productivity; i.e., demography does not affect the Barro Z's.
As a result, the BW model has a narrower interpretation than most renderings in the recent literature, which often admit both short- and long-run impacts of demographic change as a part of their theoretical structure. On the other hand, it has the desirable attribute of clarity in its interpretation.
Second, consider the focus of the BW model on the transitional impacts of demographic change. The postulated relations between population growth and working age growth, respectively, allow a clear interpretation for the role of demography: relatively rapid growth of the working-age population will speed the transition to long-run economic prosperity. However, two countries with the same Z’s will ultimately arrive at the same level of labour productivity growth, irrespective of their demography.
BW acknowledge the possibility that the rates of growth of working age and total population might impact long run labour productivity, but they fail to model this explicitly. Nor do they include other demographic variables among the Z’s. Rather, they limit themselves to noting that long run influences could result in coefficient estimates which deviate from unity.
While Bloom & Williamson (1998) focus on output per capita as the dependent variable, the recent literature argues that an understanding of international differences in output per worker is needed, since only workers contribute to production. From regressions with the growth rate of output per capita as the dependent variable it is rather complicated to assess behavioral effects (see Bloom et al. (2001) for an assessment of behavioral effects from such regressions).
Kelley and Schmidt (2001) for example demonstrated the implication of this point by evaluating the net impact of demographic change on economic growth in eight empirical renderings.
References
Barro, Robert J. and Jong-Wha Lee. 1993. "Losers and Winners in Economic Growth." NBER Working Paper No. 4341. Cambridge, Mass.
Barro, R.J. 1991. “Economic Growth in a Cross Section of Countries.” Quarterly Journal of Economics 106(2):407-444
Birdsall, Nancy, 2003, "New Findings in Economics and Demography: Implications for Policies to Reduce Poverty" in Population Matters: Demographic Change, Economic Growth, and Poverty in the Developing World Edited by Nancy Birdsall, Allen C. Kelley, Steven W. Sinding, Oxford, Oxford University Press
Bloom D, Canning D, Moore M (2004). The effect of improvements in health and longevity on optimal retirement and saving. NBER Working Paper No. 10919. Cambridge, National Bureau of Economic Research.
Bloom D, Canning D, Graham B (2003). Longevity and life-cycle savings. Scandinavian
Journal of Economics, 105:319-338.
Bloom, D. E., D. Canning and Sevilla (2002). The Demographic Dividend: A New Perspective on the Economic Consequences of Population Change. Santa Monica, CA, RAND.
Bloom, D.D., Canning, D. & Malaney, P. (2001). Demographic change and economic growth in Asia, Population and Development Review, 26, supp., 257{290.
Bloom, D. E. and J. G. Williamson (1998). Demographic Transitions and Economic Miracles in Emerging Asia. World Bank Economic Review vol.12, No.3, pp. 419-56.
Cutler, D. M., J. M. Poterba, et al. 1990. "An Aging Society: Opportunity or Challenge?" Brookings Papers on Economic Activity No. 1, pp.1-56.
Collins, Susan. 1991. “Saving Behavior in Ten Developing Countries.” In B. Douglas
Bernheim and John B. Shoven, eds., National Saving and Economic Performance.
Chicago: University of Chicago Press.
Deaton, A. and C. H. Paxson (2000). Growth, demographic structure, and national saving in Taiwan. In Population and Economic Change in East Asia, A Supplement to Population and Development Review. R. Lee and C. Y. C. Chu, eds. New York, Population Council. vol. 26, pp.141-173.
Doepke, M. 2004, “Accounting for Fertility Decline During the Transition to Growth,” Journal of Economic Growth, September 2004, 9 (3), 347–83.
Doepke, M and F. Zilibotti 2005, "The Macroeconomics of Child-Labor Regulation.". American Economic Review, December 2005.
Fry, Maxwell, and Andrew Mason. 1982. “The Variable Rate-of-Growth Effect in the
Life-Cycle Saving Model: Children, Capital Inflows, Interest, and Growth in a New
Specification of the Life-Cycle Model Applied to Seven Asian Developing Countries.”
Economic Inquiry 20(3, July):426–42.
Goldberger, Arthur. 1973. “Dependency Rates and Savings Rates: Further Comment.”
American Economic Review 63(1, March):232–33.
Harrigan, Frank. 1996. “Saving Transitions in Southeast Asia.” Asian Development
Bank, Manila. Processed.
Higgins, Matthew 1998. “Demography, National Savings, and International Capital Flows.” International Economic Review 39(2, May):343–69.
Higgins, Matthew. 1994. “The Demographic Determinants of Savings, Investment, and
International Capital Flows.” Ph.D. diss., Harvard University, Cambridge, Mass.
Higgins M, and Williamson J (1997). Age structure dynamics in Asia and dependence on foreign capital. Population and Development Review, 23:261-293.
Higgins, Matthew, and Jeffrey G. Williamson. 1996. “Asian Demography and Foreign
Capital Dependence.” NBER Working Paper 5560. National Bureau of Economic Research,
Cambridge, Mass.
Kang, Kenneth. 1994. “Why Did Koreans Save So Little and Why Do They Now Save
So Much?” International Economic Journal 8(4, winter):99–111.
Kelley, A.C. 2003, "The Population Debate in Historical Perspective: Revisionism Revised" in Population Matters: Demographic Change, Economic Growth, and Poverty in the Developing World Edited by Nancy Birdsall, Allen C. Kelley, Steven W. Sinding, Oxford, Oxford University Press
Kelley, Allen C., and Robert M. Schmidt. 1995. “Aggregate Population and Economic
Growth Correlations: The Role of the Components of Demographic Change.” Demography
32(4):543–55.
Kelley AC, Schmidt RM (1996). Saving, dependency, and development. Journal of Population Economics, 9:365-386.
Krugman, Paul. 1994. “The Myth of Asia’s Miracle.” Foreign Affairs 73(6, November-
December):62–78.
Lee RD, Mason A, Miller T (2000). Life-cycle saving and the demographic transition: the case of Taiwan. Population and Development Review, 26(Supplement):194–219.
Lee, Ronald, Andrew Mason, et al. (2003). From transfers to individual responsibility: Implications for savings and capital accumulation in Taiwan and the United States. Scandinavian Journal of Economics vol.105, No.3, pp.339-357.
Lee, Ronald, Andrew Mason, and Timothy Miller. 1997. “Saving, Wealth, and the Demographic Transition in East Asia.” Paper presented at the conference on population
and the Asian economic miracle, East-West Center, Honolulu, Hawaii, 7–10 January.
Processed.
Leff NH (1969). Dependency rates and savings rates. American Economic Review, 59: 886-896.
Mason A (1988). Saving, economic growth, and demographic change. Population and Development Review, 14:113-144.
Mason A (1987). National saving rates and population growth: a new model and new evidence. In: Johnson DG, Lee RD, eds. Population growth and economic development: issues and evidence. University of Wisconsin Press, Madison.
Mason, Andrew (1988). Saving, economic growth, and demographic change. Population and
Development Review vol.14, No. 1, pp. 113-44.
Mason, Andrew (2001). Population Change and Economic Development in East Asia: Challenges Met, Opportunities Seized. Stanford, Stanford University Press.
Masson, Paul. 1990. “Long-Term Macroeconomic Effects of Aging Populations.” Finance
and Development 27(2, June):6–9.
Perlman, M. 1975, Some Economic Growth Problems and the Part Population Policy Plays.
Quarterly Journal of Economics, 1975 Volume 89, No2, Pages: 247-56
Persson, Joakim, 2004, The Population Age Distribution, Human Capital, and Economic Growth: The U.S. states 1930-2000, Department of Economics and Statistics, Örebro University, Sweden, Working Paper
Ram, Rati. 1982. “Dependency Rates and Aggregate Savings: A New International Cross-
Section Study.” American Economic Review 72(3, June):537–44.
Schultz, T. Paul, 2005, Demographic Determinants of Savings: Estimating and Interpreting the Aggregate Association in Asia, IZA Discussion Paper No. 1479,
January 2005
Taylor, Alan. 1995. “Debt, Dependence, and the Demographic Transition: Latin America
into the Next Century.” World Development 23(5, May):869–79.
Taylor, Alan, and Jeffrey G. Williamson. 1994. “Capital Flows to the New World as an
Intergenerational Transfer.” Journal of Political Economy 102(2, April):348–69.
Tobin, James. 1967. “Life-Cycle Savings and Balanced Economic Growth.” In William
Fellner, ed., Ten Essays in the Tradition of Irving Fischer. New York: Wiley Press
Tsai IJ, Chu CYC, Chung CF (2000). Demographic transition and household saving in Taiwan. Population and Development Review, 26(Supplement):174-193.
Webb, Steven, and Heidi Zia. 1990. “Lower Birth Rates = Higher Saving in LDCs.”
Finance and Development 27(2, June):12–14.
Williamson, J.G. (2001), Demographic shocks and global factor flows. In J.N. Little and R.K. Triest eds. Seismic Shifts: The Economic Impact of Demographic Change (Boston, Mass.: Federal Reserve Bank of Boston).
Williamson, Jeffrey G. 1993. “Human Capital Deepening, Inequality, and Demographic Events Along the Asia-Pacific Rim.” In Gavin Jones, Naohiro Ogawa, and Jeffrey G. Williamson, eds., Human Resources and Development along the Asia-Pacific Rim. New York: Oxford University Press.
Young A (1994). Lessons from the East Asian NIC’s: a Contrarian view. European Economic Review, 38:964-973. Young A (1995). The tyranny of numbers: confronting the statistical realities of the East Asian growth experience. Quarterly Journal of Economics, 110:641-680.
The Discovery Of Age Structure II
The East Asian Tigers
The East Asian development experience has often been cited in support of the age transition demographic dividend view, since in the south east Asia "tiger" countries, as the transition progressed a successful export-oriented growth-strategy produced more than enough jobs to absorb the rapidly growing workforce. At the same time, a relatively stable macroeconomic environment – at least until the financial crisis of the late 90s – provided a fertile and attractive investment environment.
Ongoing work by a number of researchers has enticingly suggested that a significant part of the impressive rise in Asian savings rates can be explained by the equally impressive decline in dependency burdens, and that some of the notable difference in savings rates between a previously sluggish South Asia and a booming East Asia can be attributed to differences in the relative dependency burdens. This view also fuels the rather optimistic hypothesis that some of the savings gap between the two regions should reduce as youth dependency rates fall in South Asia over the next three decades.
Debate over the East Asian "miracle" really begins with the work of Alwyn Young, who, in a couple of extraordinarily provocative papers (Young, 1994, 1995), attempted to demonstrate that the rapid economic growth that was only too evident in the Asian Tiger countries was principally attributable to increases in factor inputs - notably labor, capital, and education - rather than to general improvements in total factor productivity. If Young was right, then the key to understanding the rise in income levels in East Asia lay in understanding the driving mechanisms such a growth in inputs.
All of the Asian Tiger economies enjoyed a surge in savings and investment during the high-growth period. The private savings rate in Taiwan, for example, rose from around 5% in the 1950s to well over 20% in the 1980s and 1990s. (It is perhaps worth noting that much of the debate surrounding the savings issue has focused on Taiwan since the data on household savings in Taiwan is fairly comprehensive).
As predicted by the life cycle hypothesis, savings rates in Taiwan were found to vary by age, being highest in Taiwan for households whose heads were in the 50-60 age range. Using this date Bloom and Williamson argued that changing age structure was a plausible explanation for the evident increase in aggregate saving. In support of this view they cited a number of previous studies (both "historic" and more recent ones) that had also found a strong connection between these two variables (Fry and Mason, 1982; Higgins, 1998; Higgins and Williamson, 1997; Kelley and Schmidt, 1996; Leff, 1969; Mason, 1987, 1988).
In terms of prior work, Higgins and Williamson (1996, 1997) had estimated the largest macro impacts. Higgins and Williamson attempted to estimate the effect of changes in population age distribution on changes in, rather than levels of, the savings rate as it deviated around the 1950–92 mean. Thus East Asia’s savings rate in 1990–92 was 8.4 percentage points above its 1950–92 average due to the transition to a much lower dependency burden. Similarly, East Asia’s savings rate in 1970–74 was 5.2 percentage points below its 1950–92 average due, in part, to the size of the dependency burden.
They found the value of the total demographic swing to be a sizeable 13.6%, a swing which would appear to account for almost the entire rise in the savings rate in East Asia over the 20 years in question. The figures for Southeast Asia are similar, but not quite so dramatic. Southeast Asia’s savings rate was 7.9% higher in 1990–92 than the 1950–92 average and the mean was 3.6% between 1970–74. The total demographic swing was 11.5%, a smaller figure than for East Asia, but still apparently sufficient to account for the entire post 1970 rise in the Southeast Asia savings rate. Since the demographic transition had been slower in South Asia, the more modest changes in the savings rate were, more or less, to be anticipated.
Higgins (1998) carried out an econometric investigation of the relations between national age distributions and savings and investment rates for a sample of 100 countries, using both time-series and cross-section data. The results once more point to substantial demographic effects, with increases in both the youth and old-age dependency ratios associated with lower saving rates.
Perhaps more important for the agenda it would set for subsequent work, the results pointed to differential demographic effects on savings supply and investment demand, and thus, to a role for demography in determining the residual between the two: net capital flows or the current account balance (CAB).
In particular, he found that high youth dependency rates demonstrated a strong relation with current account deficits, with nations being found to join the ranks of the capital exporters as they mature. The estimated demographic effect on the CAB has exceeded four percent of GDP over the last three decades in many countries.
This Hiigins paper (which was based on his earlier doctoral thesis: he was one of Jeffrey Williamson's students) is best interpreted as extending the youth-dependency thesis Coale and Hoover (1958) had advanced almost forty years earlier, moving from an exclusive focus on saving rates to incorporating the associated impacts on investment and the CAB.
His finding that maturing nations tend to graduate from reliance on capital imports also has important implications concerning the possible evolution of the global savings pool in the decades which lie ahead. In particular, currently developing nations should be expected to be increasingly able to finance their own investment needs — a prospect which stands in in stark contrast to the pessimistic prognoses of the early 1990s ( For example, Depak Lal, 1991)
Higgin's other major finding was that population dynamics accounted for a substantial share of East Asia’s economic miracle. Population dynamics accounted for something between 1.4 and 1.9% of East Asia’s annual growth in GDP per capita from 1965 to 1990, or as much as one-third of observed economic growth during the period. Using a modified definition of "miracle" (and assuming, as Higgins does, that hypothetical "steady-state" growth in East Asia was a rate of about 2% a year) as everything in excess of steady state growth, or about 4.1% (subtracting the notional 2 from the acieved 6.1 % growth rate). Population dynamics could then account for almost half of this observed difference or miracle. Now accounting for one-third or one-half of the East Asian growth certainly does not explain everything, but it does suggest that population dynamics may have been the single most important determinant of this 'extra' growth.
Within Asia itself, the evidence also suggests that demographic divergence contributed significantly to economic divergence during this same period.
Deaton and Paxson (2000) however throw something of a bucket of cold water on all of this since, using household savings data for Taiwan, they find that changes in age structure account for only a modest increase in the overall savings rate, perhaps 4 percentage points. They argue that the rise in the aggregate savings rate has not been mainly due to changes in the age composition of the population but, rather, to a secular rise in the savings rates of all age groups.
Accepting this finding at face value, the question then arises as to why savings rates should have risen at each age. One possible - and intriguing - explanation, proposed by Lee, Mason, and Miller (2000) is that increased savings rates are due to rising life expectancy and an increasing need to fund retirement income. In support of this idea Tsai, Chu, and Chung (2000) show that the timing of the rise in household savings rates does indeed match the recorded increases in the life expectancy of the population.
With a fixed retirement age we would expect such a savings effect. However, Deaton and Paxson (2000) argue that in a flexible economy, without mandatory retirement, the main effect of a rise in longevity will be on the span of the working life, with no obvious prediction for the rate of saving. Bloom, Canning, and Moore (2004) formalize this argument to show that under reasonable assumptions the optimal response to an improvement in health and a rise in life expectancy is to increase the length of working life, though less than proportionately, with no need to raise saving rates at all (due to the gains from enjoying compound interest over a longer life span).
Schultz (2005) has also argued that the estimated magnitude of the dynamic aggregate relationship identified by Higgins and williamson may be appears smaller than reported, as well as suggesting that the HW effect may be sensitive to the choice of econometric methods used to describe it. referring directly to the Deaton and Paxson finding that savings behavior at the household level does not demonstrate sufficient life cycle variation, he asks whether there might not be alternative explanations for the observed trends in savings? One possibility explanation for the empirical regularities would be the substitution of savings for children, and this in fact could be explained within a household lifetime demand framework.
There may also be a demographic effect as a longer prospective life span can change life-cycle behavior, leading to a longer working life or higher savings for retirement. Recent empirical literature on economic growth does indeed provide compelling evidence that health status, longevity and the age distribution do indeed have strong connections to economic growth (Bloom, Canning, and Graham, 2003; Bloom, Canning, and Moore, 2004, Kinugasa, 2004, Kinugasa and Mason, 2005, Mason 2005).
While in theory a longer life span should be associated with a longer working life, in practice this may not be the case. Bloom, Canning, and Graham (2003) find that, even allowing for age structure effects, longer life expectancy is strongly associated with higher national savings rates across countries, which suggests that there is a savings effect.
While the optimal response with perfect markets may be for workers to have a longer working life as their health improves and they have longer life expectancies, mandatory or conventional retirement ages, coupled with the strong financial incentives to retire that are inherent in many social security systems, seem to result in early retirement and increased needs for saving for old age.
Also old age "dependency" is something of a misnomer. Lee (2000) shows that, in all pre-industrial societies for which he was able to assemble evidence, the flow of transfers is from the middle-aged and old to the young. In developed countries, on the other hand, both the young and the old benefit from government transfers, and the net pattern of transfers is towards the elderly. However, at the level in the United States and elsewhere, elderly households in fact make significant transfers to middle-aged households, undoing to some extent the effects of government policy. This seems to suggest that the dependency burden of the elderly is a function of the institutional welfare systems that are in place rather than an immutable state of affairs.
Beyond East Asia, a number of other studies appear to have leant more support to the demographic dividend idea. Bloom and Canning (2003), for example, look at the case of Egypt. Between 1965 and 1990 Egypt’s working-age population grew at the rate of 2.61% per annum. This rate was about about 30 % faster than the growth rate in the dependent population and is closely comparable to the corresponding rates of population change among slow-growing South Asian countries. By contrast, the working-age population of the East Asian countries grew at approximately ten times the rate of the dependent population during the "miracle" years, as the "baby bulge" cohort entered the prime-age worker group. In East Asia this was also associeted with a drop in fertility beyond the replacement level. Using simple econometric regressions Bloom and Canning estimated that - not unexpectedly - during the early phases of Egypt’s demographic transition age changes contributed a modest, but not insignificant, 0.4 percentage points to Egypt’s economic growth rate during 1965–1990.
New section
Despite having achieved a certain degree of recognition and empirical successes, the variable rate-of-growth effect model is unlikely to provide a complete theoretical framework for understanding the behavior of saving rates and capital flows over whole course of the demographic transition. In particular, the variable rate-of growth effect model describes only one possible relation - that of a hypothetical steady-state relationship between dependency and saving rates - a shortcoming which in part from its life-cycle theory ancestry.
Continuing demographic change over the past half century, however, suggests that we look more to transitional dynamics if we want to more fully understand the observed correlation between dependency and saving rates and capital flows.
In this vein Higgins and Williamson (1996) show that the variable rate-of-growth effect model can be subsumed under the standard textbook neoclassical growth model, suitably modified to incorporate an overlapping generations population (see e.g., Blanchard and Fischer, 1988). The standard growth model is, in essence, simply an open-economy, steady-state version of the latter. The textbook overlapping generations model need only be augmented by adding a third period of life - childhood alogside old age — in order to accommodate a more comprehensive study of dependency effects.
References:
Bloom, D. and D. Canning, 2003. From demographic lift to economic lift-off: the case of Egypt, Applied Population and Policy 2003:1(1) 15–24.
Kinugasa, T. (2004). Life Expectancy, Labor Force, and Saving, Ph.D. Dissertation. University of Hawaii, Manoa.
Kinugasa, T. and A. Mason, 2005, The Effects of Adult Longevity on Saving, mimeo, University of Hawaii, Manoa
LAL, Deepak, 1991, "World Savings and Growth in Developing Countries", Discussion Papers in Economics No. 91-05, University College, London.
Mason, A. 2005, Demographic Transition and Demographic Dividends in Developed and Developing Countries, Paper presented at the United Nations Expert Group Meeting on Social and Economic Implications of Changing Population Age Structures, United Nations Department of Economic and Social affairs, Mexico City, Mexico, September 2005
Schultz, P.T. 2004, Demographic Determinants of Savings: Estimating and Interpreting the Aggregate Association in Asia, Economic Growth Centre, Discussion Paper No. 901, Yale university, New Haven.
The East Asian development experience has often been cited in support of the age transition demographic dividend view, since in the south east Asia "tiger" countries, as the transition progressed a successful export-oriented growth-strategy produced more than enough jobs to absorb the rapidly growing workforce. At the same time, a relatively stable macroeconomic environment – at least until the financial crisis of the late 90s – provided a fertile and attractive investment environment.
Ongoing work by a number of researchers has enticingly suggested that a significant part of the impressive rise in Asian savings rates can be explained by the equally impressive decline in dependency burdens, and that some of the notable difference in savings rates between a previously sluggish South Asia and a booming East Asia can be attributed to differences in the relative dependency burdens. This view also fuels the rather optimistic hypothesis that some of the savings gap between the two regions should reduce as youth dependency rates fall in South Asia over the next three decades.
Debate over the East Asian "miracle" really begins with the work of Alwyn Young, who, in a couple of extraordinarily provocative papers (Young, 1994, 1995), attempted to demonstrate that the rapid economic growth that was only too evident in the Asian Tiger countries was principally attributable to increases in factor inputs - notably labor, capital, and education - rather than to general improvements in total factor productivity. If Young was right, then the key to understanding the rise in income levels in East Asia lay in understanding the driving mechanisms such a growth in inputs.
All of the Asian Tiger economies enjoyed a surge in savings and investment during the high-growth period. The private savings rate in Taiwan, for example, rose from around 5% in the 1950s to well over 20% in the 1980s and 1990s. (It is perhaps worth noting that much of the debate surrounding the savings issue has focused on Taiwan since the data on household savings in Taiwan is fairly comprehensive).
As predicted by the life cycle hypothesis, savings rates in Taiwan were found to vary by age, being highest in Taiwan for households whose heads were in the 50-60 age range. Using this date Bloom and Williamson argued that changing age structure was a plausible explanation for the evident increase in aggregate saving. In support of this view they cited a number of previous studies (both "historic" and more recent ones) that had also found a strong connection between these two variables (Fry and Mason, 1982; Higgins, 1998; Higgins and Williamson, 1997; Kelley and Schmidt, 1996; Leff, 1969; Mason, 1987, 1988).
In terms of prior work, Higgins and Williamson (1996, 1997) had estimated the largest macro impacts. Higgins and Williamson attempted to estimate the effect of changes in population age distribution on changes in, rather than levels of, the savings rate as it deviated around the 1950–92 mean. Thus East Asia’s savings rate in 1990–92 was 8.4 percentage points above its 1950–92 average due to the transition to a much lower dependency burden. Similarly, East Asia’s savings rate in 1970–74 was 5.2 percentage points below its 1950–92 average due, in part, to the size of the dependency burden.
They found the value of the total demographic swing to be a sizeable 13.6%, a swing which would appear to account for almost the entire rise in the savings rate in East Asia over the 20 years in question. The figures for Southeast Asia are similar, but not quite so dramatic. Southeast Asia’s savings rate was 7.9% higher in 1990–92 than the 1950–92 average and the mean was 3.6% between 1970–74. The total demographic swing was 11.5%, a smaller figure than for East Asia, but still apparently sufficient to account for the entire post 1970 rise in the Southeast Asia savings rate. Since the demographic transition had been slower in South Asia, the more modest changes in the savings rate were, more or less, to be anticipated.
Higgins (1998) carried out an econometric investigation of the relations between national age distributions and savings and investment rates for a sample of 100 countries, using both time-series and cross-section data. The results once more point to substantial demographic effects, with increases in both the youth and old-age dependency ratios associated with lower saving rates.
Perhaps more important for the agenda it would set for subsequent work, the results pointed to differential demographic effects on savings supply and investment demand, and thus, to a role for demography in determining the residual between the two: net capital flows or the current account balance (CAB).
In particular, he found that high youth dependency rates demonstrated a strong relation with current account deficits, with nations being found to join the ranks of the capital exporters as they mature. The estimated demographic effect on the CAB has exceeded four percent of GDP over the last three decades in many countries.
This Hiigins paper (which was based on his earlier doctoral thesis: he was one of Jeffrey Williamson's students) is best interpreted as extending the youth-dependency thesis Coale and Hoover (1958) had advanced almost forty years earlier, moving from an exclusive focus on saving rates to incorporating the associated impacts on investment and the CAB.
His finding that maturing nations tend to graduate from reliance on capital imports also has important implications concerning the possible evolution of the global savings pool in the decades which lie ahead. In particular, currently developing nations should be expected to be increasingly able to finance their own investment needs — a prospect which stands in in stark contrast to the pessimistic prognoses of the early 1990s ( For example, Depak Lal, 1991)
Higgin's other major finding was that population dynamics accounted for a substantial share of East Asia’s economic miracle. Population dynamics accounted for something between 1.4 and 1.9% of East Asia’s annual growth in GDP per capita from 1965 to 1990, or as much as one-third of observed economic growth during the period. Using a modified definition of "miracle" (and assuming, as Higgins does, that hypothetical "steady-state" growth in East Asia was a rate of about 2% a year) as everything in excess of steady state growth, or about 4.1% (subtracting the notional 2 from the acieved 6.1 % growth rate). Population dynamics could then account for almost half of this observed difference or miracle. Now accounting for one-third or one-half of the East Asian growth certainly does not explain everything, but it does suggest that population dynamics may have been the single most important determinant of this 'extra' growth.
Within Asia itself, the evidence also suggests that demographic divergence contributed significantly to economic divergence during this same period.
Deaton and Paxson (2000) however throw something of a bucket of cold water on all of this since, using household savings data for Taiwan, they find that changes in age structure account for only a modest increase in the overall savings rate, perhaps 4 percentage points. They argue that the rise in the aggregate savings rate has not been mainly due to changes in the age composition of the population but, rather, to a secular rise in the savings rates of all age groups.
Accepting this finding at face value, the question then arises as to why savings rates should have risen at each age. One possible - and intriguing - explanation, proposed by Lee, Mason, and Miller (2000) is that increased savings rates are due to rising life expectancy and an increasing need to fund retirement income. In support of this idea Tsai, Chu, and Chung (2000) show that the timing of the rise in household savings rates does indeed match the recorded increases in the life expectancy of the population.
With a fixed retirement age we would expect such a savings effect. However, Deaton and Paxson (2000) argue that in a flexible economy, without mandatory retirement, the main effect of a rise in longevity will be on the span of the working life, with no obvious prediction for the rate of saving. Bloom, Canning, and Moore (2004) formalize this argument to show that under reasonable assumptions the optimal response to an improvement in health and a rise in life expectancy is to increase the length of working life, though less than proportionately, with no need to raise saving rates at all (due to the gains from enjoying compound interest over a longer life span).
Schultz (2005) has also argued that the estimated magnitude of the dynamic aggregate relationship identified by Higgins and williamson may be appears smaller than reported, as well as suggesting that the HW effect may be sensitive to the choice of econometric methods used to describe it. referring directly to the Deaton and Paxson finding that savings behavior at the household level does not demonstrate sufficient life cycle variation, he asks whether there might not be alternative explanations for the observed trends in savings? One possibility explanation for the empirical regularities would be the substitution of savings for children, and this in fact could be explained within a household lifetime demand framework.
There may also be a demographic effect as a longer prospective life span can change life-cycle behavior, leading to a longer working life or higher savings for retirement. Recent empirical literature on economic growth does indeed provide compelling evidence that health status, longevity and the age distribution do indeed have strong connections to economic growth (Bloom, Canning, and Graham, 2003; Bloom, Canning, and Moore, 2004, Kinugasa, 2004, Kinugasa and Mason, 2005, Mason 2005).
While in theory a longer life span should be associated with a longer working life, in practice this may not be the case. Bloom, Canning, and Graham (2003) find that, even allowing for age structure effects, longer life expectancy is strongly associated with higher national savings rates across countries, which suggests that there is a savings effect.
While the optimal response with perfect markets may be for workers to have a longer working life as their health improves and they have longer life expectancies, mandatory or conventional retirement ages, coupled with the strong financial incentives to retire that are inherent in many social security systems, seem to result in early retirement and increased needs for saving for old age.
Also old age "dependency" is something of a misnomer. Lee (2000) shows that, in all pre-industrial societies for which he was able to assemble evidence, the flow of transfers is from the middle-aged and old to the young. In developed countries, on the other hand, both the young and the old benefit from government transfers, and the net pattern of transfers is towards the elderly. However, at the level in the United States and elsewhere, elderly households in fact make significant transfers to middle-aged households, undoing to some extent the effects of government policy. This seems to suggest that the dependency burden of the elderly is a function of the institutional welfare systems that are in place rather than an immutable state of affairs.
Beyond East Asia, a number of other studies appear to have leant more support to the demographic dividend idea. Bloom and Canning (2003), for example, look at the case of Egypt. Between 1965 and 1990 Egypt’s working-age population grew at the rate of 2.61% per annum. This rate was about about 30 % faster than the growth rate in the dependent population and is closely comparable to the corresponding rates of population change among slow-growing South Asian countries. By contrast, the working-age population of the East Asian countries grew at approximately ten times the rate of the dependent population during the "miracle" years, as the "baby bulge" cohort entered the prime-age worker group. In East Asia this was also associeted with a drop in fertility beyond the replacement level. Using simple econometric regressions Bloom and Canning estimated that - not unexpectedly - during the early phases of Egypt’s demographic transition age changes contributed a modest, but not insignificant, 0.4 percentage points to Egypt’s economic growth rate during 1965–1990.
New section
Despite having achieved a certain degree of recognition and empirical successes, the variable rate-of-growth effect model is unlikely to provide a complete theoretical framework for understanding the behavior of saving rates and capital flows over whole course of the demographic transition. In particular, the variable rate-of growth effect model describes only one possible relation - that of a hypothetical steady-state relationship between dependency and saving rates - a shortcoming which in part from its life-cycle theory ancestry.
Continuing demographic change over the past half century, however, suggests that we look more to transitional dynamics if we want to more fully understand the observed correlation between dependency and saving rates and capital flows.
In this vein Higgins and Williamson (1996) show that the variable rate-of-growth effect model can be subsumed under the standard textbook neoclassical growth model, suitably modified to incorporate an overlapping generations population (see e.g., Blanchard and Fischer, 1988). The standard growth model is, in essence, simply an open-economy, steady-state version of the latter. The textbook overlapping generations model need only be augmented by adding a third period of life - childhood alogside old age — in order to accommodate a more comprehensive study of dependency effects.
References:
Bloom, D. and D. Canning, 2003. From demographic lift to economic lift-off: the case of Egypt, Applied Population and Policy 2003:1(1) 15–24.
Kinugasa, T. (2004). Life Expectancy, Labor Force, and Saving, Ph.D. Dissertation. University of Hawaii, Manoa.
Kinugasa, T. and A. Mason, 2005, The Effects of Adult Longevity on Saving, mimeo, University of Hawaii, Manoa
LAL, Deepak, 1991, "World Savings and Growth in Developing Countries", Discussion Papers in Economics No. 91-05, University College, London.
Mason, A. 2005, Demographic Transition and Demographic Dividends in Developed and Developing Countries, Paper presented at the United Nations Expert Group Meeting on Social and Economic Implications of Changing Population Age Structures, United Nations Department of Economic and Social affairs, Mexico City, Mexico, September 2005
Schultz, P.T. 2004, Demographic Determinants of Savings: Estimating and Interpreting the Aggregate Association in Asia, Economic Growth Centre, Discussion Paper No. 901, Yale university, New Haven.
Sunday, October 09, 2005
The Mystery of Economic Growth First Draft
Attempts to answer the question "why some countries are so much richer than others" have been as numerous as they have been infructiferous.
Durlauf et al (2005) put it like this:
"Understanding the wealth of nations is one of the oldest and most important research agendas in the entire discipline. At the same time, it is also one of the areas in which genuine progress seems hardest to achieve. The contributions of individual papers can often appear slender. Even when the study of growth is viewed in terms of a collective endeavor, the various papers cannot easily be distilled into a consensus that would meet standards of evidence routinely applied in other fields of economics".
Among the biggest problems facing growth theorists are the comparatively small number of countries to work from, the comparatively limited time duration of the available data, and the wide variety of explanatory models which have been proposed in explanation of the phenomenon. Indeed, in the course of a long and arduous literature history, approximately as many growth determinants have been proposed as there are countries for which data are available.
However despite the not inconsiderable amount of research and energy dedicated by economists to the theme of economic growth and development (Nobel economist Robert Lucas, for example, informs us that not a day goes by without his spending at least some time thinking about the topic), collectively we have yet to discovered how to make poor countries rich.
Whatsmore the road to progress seems fraught with difficulty. If we take it as self-evidently true that technology is one of the primary determinants of a country’s growth and income, and if the know-how relating to modern production technologies is essentially free and there for the taking, then why, oh why is it that so many people still find it so difficult to get on the development track?
Simply posing the question is, of course, to already answer it, since even in the case of something so apparently 'neutral' as technology there are clearly self-reinforcing mechanisms, or 'traps' (or 'bad equilibria') that act either as barriers to adoption or impediments to effective use even when adopted. Development traps arise from a variety of circumstances, and again these have generated an extensive literature with much of the focus being centred on market imperfection and institutional failure.
Another explanation often advanced for the poor growth performance of the LDCs is the existence of poor institutions and bad domestic policy. And of course those who argue this do so not without justification. The LDCs as a group are not exactly short on examples of corruption and bad government. And clearly sound governance and healthy market influences are necessary to reap the benefits of economic growth. But isn't this story perhaps just a little too simple.
In a pathbreaking paper on Geography, Demography and Economic Growth in Africa, David Bloom and Jeffrey Sachs make the following plea:
"Our paper could well be misunderstood. Some will regard it as a new case of "geographic determinism," that Africa is fated to be poor because of its geography. Some will regard it as a distraction from the important truth that geographic difficulties or not, African governments seriously mismanaged economic policy in the past generation. Let us therefore be clear at the outset. We believe emphatically that economic policy matters, and our formal econometric results show that to be true, a point we have also made in related recent studies (especially Sachs and Warner, 1997) We nonetheless focus most of our attention on geography for three reasons. First, there is little to be gained from yet another recitation of the damage of statism, protectionism, and corruption on African economic performance. Amen. Second, most economists are woefully neglectful of the forces of nature in shaping economic performance, in general and in Africa in particular. They treat economies as blank slates, upon which another region's technologies and economic history may be grafted. Our profession's formal models tend to be like that; so do our profession's standard statistical analyses of cross-country growth....Third, and perhaps most importantly, good policies must be tailored to geographical realities.If agricultural productivity is very low in Africa for climatological reasons, perhaps the real lesson is that growth should be led much more by outward-oriented industry and services, rather than yet another attempt to blindly transplant "integrated rural development" strategies from other parts of the world that are not customized to Africa’s unique conditions."
If Bloom and Sachs are accused of geographical determinism, it is easy to imagine that I might here be accused of "demographic determinism", so let me just endorse the above argument: economic and political policy also matters. But what I would also like to stress is the importance of their second point, namely that "most economists are woefully neglectful of the forces of nature in shaping economic performance, in general and in Africa in particular. They treat economies as blank slates, upon which another region's technologies and economic history may be grafted." To ram this home: most economists' appreciation of the workings of evolutionary biology and its implications for demographic processes is rudimentary to say the least. And what outrageously many economists understand about demographic processes in and of themselves their relevance for growth and development would provide material for only the shortest of short papers. The fact is, with some emminent and notable exceptions, demography is simply not seen as an important part of the macro-economic panorama.
In this context I would simply single out two underlying prejudices which I think have had an enormous influence.
In the first place the 'state of nature' argument. Basically most economists seem to work from some variant of Cantillion's notion that: "Men multiply like mice in a barn if they have unlimited means of subsistence", ie that in the pre-modern (Mathusian regime epoch) women produced children at an extraordinarily high rate. This is certainly a counter-factual belief, not only because in most foraging societies fertility falls universally well short of the biological maximum, but also since there is often evidence in history for the idea that as subsistance rises fertility declines.
Secondly the homeostatic principle which normally lies behind modern conceptions of steady state growth. Basically - and this is a point which has been often made - most economists seem to see their discipline as much closer to physics than it is to biology. So the preferred yardstick for measuring the validity of our economic models is far more often the celestial motion of the planets than ever it is the Darwinian evolution of living organisms. So the essential idea is that 'modern' (or if you want market-driven) economies represent a break from an earlier high fertility regime to a new growth regime which has a steady state dynamic, or balanced growth path, and has an internalised homeostatic conditioner which produces a population fertility rate of on or around a 2.1 TFR. Whether by accident or design, the excellent Consise History of Population by Massimo Livvi Bacci is the most widely quoted population reference source in modern growth economics. I say whether by accident or design, because, Livvi-Bacci is, of course, the best know demographic exponent of the homeostatic population view.
To concretise, the story that the competitive neoclassical benchmark model of economic systems tells is the following one: Markets are complete, entry and exit is free, transaction costs are negligible, and technology is convex at an efficient scale relative to the size of the market. As a result, the private and social returns to production and investment are equal and a complete set of 'virtual prices' ensures that all projects with positive net social benefit are undertaken. Diminishing returns to the set of reproducible factor inputs implies that as an economy develops the rate of return on investment declines, whilst where capital is scarce the returns to investment (and of course the level of risk) will be concomitanly higher.
The dynamic implications of this benchmark were intitially summarized by Solow (1956), Cass (1965), and Koopmans (1965). Even for countries which start with different endowments, the main conclusion is that there ought to be an underlying convergence process.
Needless to say the influence on development policy of what might be called the neoclassical 'prejudice' has been enormous. One of the clearest examples is, of course, to be found in the series of structural adjustment programs implemented by the International Monetary Fund. The key components of the Enhanced Structural Adjustment Facility —the centerpiece of the IMF’s strategy to aid poor countries and promote long run growth in the years between 1987 and 1999 — were prudent macroeconomic policies and the liberalization of markets. Growth, it was hoped, would then follow automatically.
Yet the evidence on whether or not non-distortionary policies and diminishing returns to capital will soon carry the poor to opulence is mixed to say the least. In recent years even relatively well governed countries have experienced little or no growth. Mali, for example, although not exactly corruption-free, scores relatively well in comparative measures of governance and real resources (Radlet 2004; Sachs et al. 2004). Yet Mali is still desperately poor. According to a 2001 UNDP report, 70% of the population lives on less than $1 per day. The infant mortality rate is 230 per 1000 births, and household final consumption expenditure is down 5% from 1980. And of course Mali continues to have one of the highest TFR rates on the planet. And Mali is not exactly an isolated case.
So are we perhaps missing something here?
Well lets go back to where we started. Now apart from the existence of model uncertainty, another major set of growth issues revolves around the difficulty of identifying empirically salient determinants of growth when the range of potential factors is large relative to the number of observations.
Individual researchers, seeking to identify the extent of support for particular growth determinants, typically emphasize a single model (or small set of models) and then carry out an inference procedure as if that model had generated the data. Standard inference procedures based on a single model, and which are conditional on the truth of that model, can grossly overstate the precision of inferences about a given phenomenon. In particular such procedures often ignore the uncertainty that surrounds the validity of the model itself.
Given that there are usually other models that have strong claims to consideration, the standard errors which arise can significantly understate the true degree of uncertainty about the parameters, and the choice of model to can ultimately appear somewhat arbitrary. This need for a proper account of model uncertainty has lead naturally toward Bayesian or pseudo-Bayesian approaches to data analysis.
Despite the existence of these widely known problems attempts to develop a unified growth theory along the abovementioned lines have not been notably deterred. In particular we have seen a continuing effort to establish common 'covergent' characteristics to growth across countries. One such approach has centred around the attempt to establish a series of 'stylised facts' about growth.
Stylised Facts
As I have said, most neo-classical growth models start from the idea of steady state growth. One idea closely associated with the steady state one is that there exist a number of 'stylised facts' of growth which are waiting to be uncovered. The tradition of attempting to produce stylised facts in economic growth theory goes back at least to the work of Kuznets and Kaldor in the 1960s.
Kaldor (1961) described what he considered to be a number of stylized facts about economic growth:
1. Per capita output grows over time, and its growth rate does not tend to diminish.
2. Physical capital per worker grows over time.
3. The rate of return to capital is nearly constant.
4. The ratio of physical capital to output is nearly constant.
5. The shares of labor and physical capital in national income are nearly constant.
6. The growth rate of output per worker differs substantially across countries.
When Simon Kuznets delivered his Nobel Address back in 1971 he also singled out six stylised 'characteristics' of modern economic growth:
1/. The high rates of growth of per capita product and of population in the developed countries.
2/. The high rate of rise in productivity, i.e. of output per unit of all inputs in comparison with earlier epochs.
3/. The rate of structural transformation of the economy is also high by historical standards.
4/. Structures which are extremely important in general societal terms - eg levels of urbanization and secularization - have also changed rapidly, leading sociologists to use the term modernization process.
5/. The economically developed countries have the propensity to reach out to the rest of the world (we would now call this globalisation).
6/. Despite the spread of modern economic growth, wide inequality in growth rates and income levels persists.
These early efforts (and in particular those of Kaldor) have not worn especially well with time. As Sala i Martin (2002) points out one important innovation of the new generation of growth literature that took off in the late 80s was that it attempted to tie economic theory much closer to empirical studies of growth. The neoclassical literature of the 1960s seem to link theory and evidence by simply 'mentioning' a list of stylized facts (such the Kaldor 'facts' above) and then attempting to show that the theory proposed was consistent with one, two or perhaps several of
these 'facts'. Indeed as Sala i Martin notes, some of these 'stylised facts' were never really derived from careful empirical analysis, but that did not stop them being quoted and used as if they had been. (Kuznets himself, of course, had carried out significant empirical research on economic growth).
Today’s research, on the other hand, tends to be characterised by a drive to derive more precise econometric specifications through direct recourse to the data. The best examples of this kind of work can be found in the convergence literature. Barro and Sala-i-Martin (1992) use the Ramsey-Cass-Koopmans (Ramsey (1928), Cass (1975) and Koopmans (1965)) growth model to derive an econometric equation that relates growth of GDP per capita to initial levels of GDP. Mankiw, Romer and Weil (1992) derive a similar equation from the Solow-Swan model ((Solow (1956) and Swan (1956)). These researchers derived a relationship where the growth rate of per capita GDP is given by the difference between a constant (beta) and the product of that same beta and the natural log of per capita GDP for country and the steady-state value of per capita GDP for country i plus an error term.
The coefficient is positive in the case that the production function is neoclassical, and is zero if the production function is linear in capital (which was usually the case in the first generation one-sector models of endogenous growth, also known as AK models). The central point is that the modern literature took this equation and used it to 'test' competing the models, namely the endogenous growth ones (the so called AK models) which should predict beta = 0 and the neoclassical models (which should predict beta>0.)
Early findings seemd to indicate that there was no positive association between growth and the initial level of income. The possibility existed however that this finding could be a statistical artifact resulting from the misspecification of the original equation. Basically, the problem is that, if researchers make the assumption that countries converge to the same steady state and the finding is that they don’t, then the original equation is, in turn, misspecified. If the steady state is correlated with the initial level of income, then the error term is correlated with the explanatory variable, so the estimated coefficient is biased towards zero.
One proposed solution to the problem was to continue using cross-country data but, instead of estimating the univariate regression, estimate a multivariate regression where, on top of the initial level of income, the researcher would also hold constant proxies for the steady state. This approach came to be known as conditional convergence. Subsequent research has shown that the idea of conditional convergence describes one of the strongest and most robust empirical regularities to be found in the data. The consequence of all this was the exact opposite of the original conclusion: the neoclassical model was not rejected by the data. The AK model was.
Sala i Martin draws a number of lessons from all this:
(i) There is no simple determinant of growth.
(ii) The initial level of income is the most important and robust variable (so conditional convergence is the most robust empirical fact in the data).
(iii) The size of the government does not appear to matter much. What is important is the 'quality of government'.
(iv) The relation between most measures of human capital and growth is weak. Some
measures of health, however, (such as life expectancy) are robustly correlated with growth.
(v) Institutions (such as free markets, property rights and the rule of law) are important for growth.
(vi) More open economies tend to grow faster.
As a result of this re-evaluation it is now possible to outine a rather different group of stylised facts:
1. Over the second half of the 20th century most countries have grown richer, but at varying rates, and vast income disparities remain. For all but the richest group, growth rates have differed to an unprecedented extent, regardless of the initial level of development. Indeed far from a clear convergence process reality has rather conformed to what Lance Pritchett once famously termed 'divergence bigtime'.
2. Past growth has turned out to be a surprisingly weak predictor of future growth. In the developing world distinct 'winners' and 'losers' have begun to emerge. The strongest performers are located in East and Southeast Asia, which have sustained growth rates at unprecedented levels. The weakest performers are predominantly located in sub-Saharan Africa, where some countries have barely grown at all, or even become poorer. The record in South and Central America is also distinctly mixed. In these regions, output volatility is high, and dramatic output collapses are not uncommon.
3. For many countries, growth rates were lower in 1980-2000 than in 1960-1980, and this growth slowdown has been observed throughout most of the income distribution. Moreover, the dispersion of growth rates has increased. A more optimistic reading would also emphasize the growth take-off that has taken place in China and India, home to two-fifths of the world’s population and a greater proportion of the world’s poor.
But despite all the research two oustanding questions still seem to remain.
Firstly, the issue of convergence: are contemporary differences in aggregate economies transient over sufficiently long time horizons?
And secondly, what are the key growth determinants: which factors are better able to explain observed differences in growth?
Two separate approaches have continued to dominate the extensive literature on economic growth modeling: simple correlations, and production functions.
Simple-correlations
Simple-correlations studies hypothesize that per capita output growth is influenced by various dimensions.
Production Functions
Production-function studies are based on estimating variants of a model along the following lines:
Y = g(K, L, H, R, T),
where output (Y) is produced by the stocks of various factors: physical capital (K), labor (L), human capital (H: education and health), resources (R: land, minerals, and environment), and technology (T).
In the neoclassical model, if each country has access to the same aggregate production function the steady-state is independent of an economy's initial capital and labor stocks and hence initial income. In this model, long-run differences in output reflect differences in the determinants of accumulation, not differences in the technology used to combine inputs to produce output. Mankiw (1995, p. 301), for example, argues that for “understanding international experience, the best assumption may be that all countries have access to the same pool of knowledge, but differ by the degree to which they take advantage of this knowledge by investing in physical and human capital.”
Even if one relaxes the assumption that countries have access to the same production function, convergence in growth rates can still occur so long as each country’s production function is concave in capital per efficiency unit of labor and each country experiences the same rate of labor-augmenting technical change.
Klenow and Rodríguez-Clare (1997a) challenge this “neoclassical revival” with
results suggesting that differences in factor accumulation are, at best, no more important than differences in productivity in explaining the cross-country distribution of output per capita. They find that only about half of the cross-country variation in the 1985 level of output per worker is due to variation in human and physical capital inputs while a mere 10% or so of the variation in growth rates from 1960 to 1985 reflects differences in the growth of these inputs. The differences between the results of Mankiw, Romer, and Weil (1992) and the findings of Klenow and Rodríguez-Clare (1997a) in their reexamination of Mankiw, Romer and Weil have two principal origins. MRW found support for the Solow model’s predictions that, in the long-run steady state, the level of real output per worker by country should be positively correlated with the saving rate and negatively correlated with the rate of labor-force growth. However, their estimates of the textbook Solow model also implied a capital share of factor income of about 0.60, high compared to the conventional value (based on U.S. data) of about one-third.
To address this possible inconsistency, MRW considered an “augmented” version of the Solow model, in which human capital enters as a factor of production in symmetrical fashion with physical capital and raw labor. They found that the augmented Solow model fits the data relatively better and yields an estimated capital share more in line with conventional wisdom. They concluded (abstract, p. 407) that “an augmented Solow model that includes accumulation of human as well as physical capital provides an excellent description of the cross-country data.”
[Bernanke and Gürkaynak model]: Assume that in a given country at time t, output Y depends on inputs of raw labor L and three types of accumulated factors: K , H , and Z . The factors K and H are accumulated through the sacrifice of current output (think of physical capital and human capital, or structures and equipment). The factor Z , which could be an index of technology, or of human capital acquired through learning-by-doing, is assumed to be accumulated as a byproduct of economic activity and does not require the sacrifice of current output.]
Y = K, H (Z, L).
Bernanke and Gürkaynak also use the MRW framework to consider some alternative models of economic growth, such as the Uzawa-Lucas model and the AK model. These models are rejected as literal descriptions of the data.
However, the implications of these models, that country growth rates depend on behavioral variables such as the rate of human capital formation and the saving rate, seem more consistent with the data than the Solow model’s assumption that growth is exogenous. Future research should consider variants of endogenous growth models to see which if any provide a more complete and consistent description of the
cross-country data. We believe that the generalized MRW-type framework we have developed here could prove very helpful in assessing the alternative possibilities.
Indeed both Prescott (1998) and Hall and Jones (1999) confirm the view that differences in inputs are unable to explain observed differences in output and Easterly and Levine (2001, p. 177) state that "the 'residual' (total factor productivity, TFP) rather than factor accumulation accounts for most of the income and growth differences across countries."
Despite these concerns and the differences in the precise estimates found by different researchers, it is clear that cross-country variation in inputs falls short of explaining the observed cross-country variation in output. The result that the TFP
residual, a “measure of our ignorance” computed as the ratio of output to some index of inputs, is an important (perhaps the dominant) source of cross-country differences in long-run economic performance is useful but hardly satisfying and the need for a theory of TFP expressed by Prescott (1998) is well founded. Research such as Acemoglu and Zilibotti (2001) and Caselli and Coleman (2003) are promising contributions to that agenda.
Convergence Patterns
Convergence-patterns studies, rooted in neoclassical growth theory, explore the relationships between economic growth and the *level* of economic development. They focus on the pace at which countries move from their current level of labor productivity (Y/L) to their long-run, or steady-state equilibrium level of labor productivity.
The effect of initial conditions on long-run outcomes arguably represents the primary empirical question that has been explored by growth economists. The claim that the effects of initial conditions eventually disappear is the heuristic basis for what is
known as the convergence hypothesis.
The goal of this literature is to answer two questions concerning per capita income differences across countries (or other economic units, such as regions). First, are the observed cross-country differences in per capita incomes temporary or permanent? Second, if they are permanent, does that permanence reflect structural heterogeneity or the role of initial conditions in determining long-run outcomes?
If the differences in per capita incomes are temporary, unconditional convergence (to a common long-run level) may occur. If the differences are permanent solely because of cross-country structural heterogeneity, only conditional convergence may occur. If initial conditions determine, in part at least, long-run outcomes, and countries with similar initial conditions exhibit similar long-run outcomes, then one can speak of convergence clubs.
In practice, the distinction between initial conditions and structural heterogeneity generally amounts to treating stocks of initial human and physical capital as the former and other variables as the latter. As such, both the Solow variables X and the control variables Z that appear in garden variety cross-country growth regression are usually interpreted as capturing structural heteogeneity. This practice may be criticized if these variables are themselves endogenously determined by initial conditions, a point that will arise below.
Here the rate of labour productivity growth is taken to be proportional to the gap between the logs of the long-run, steady state and the current level of labor productivity. The greater this gap, the greater are the gaps of physical capital, human capital, and technical efficiency from their long run levels. Large gaps allow for significant "catching up" through (physical and human) capital accumulation, and through technology creation and diffusion across, and within countries.
Under restrictive assumptions, this type of model predicts "unconditional convergence" by all countries to the same long-run level of labor productivity. Were the steady state rate of growth of labor productivity to be the same for all countries, then low-income countries would be father away from their state determined income level and hence - following the standard equation - their productivity would grow faster as a result. In fact, however, positive rather than negative correlations have been observed between the level and growth rate of labor productivity. The model has not unsurprisingly been modified as a consequence.
As indicated above, models now tend to hypothesize "conditional convergence", whereby long-run labor productivity differs across countries depending on country-specific characteristics following the following equation:
ln(rate of change of labour productivity) = a + bZ.
The actual specification of the determinants of long-run labor productivity (the Z's) varies notably, but the basic model is the same across scores of empirical studies.
X contains a composite - log (n+g+delta) - where delta denotes the depreciation rate. The variables spanned by log income and X thus represent those growth determinants that are suggested by the Solow growth model whereas Z represents those growth determinants that lie outside Solow’s original theory. The distinction between the Solow variables and Z is important in understanding the empirical literature. While the standard Solow variables usually appear across a wide variety of different empirical studies - a situation which reflects the treatment of the Solow model as a baseline for growth analysis - choices concerning which Z variables to include vary greatly.
As outlined above, statistical analyses of convergence have largely focused on the properties of β in regressions. β convergence, defined as β<0 is easy to evaluate because it relies on the properties of a linear regression coefficient. It is also easy to interpret in the context of the Solow growth model, since the finding is consistent with the dynamics of the model. The economic intuition for this is simple.
If two countries have common steady-state determinants and are converging to a common balanced growth path, the country that begins with a relatively low level of initial income per capita has a lower capital-labor ratio and hence a higher marginal product of capital; a given rate of investment then translates into relatively fast growth for the poorer country.
This is why β convergence is commonly interpreted as evidence against endogenous growth models of the type studied by Romer and Lucas, since a number of these models specifically predict that high initial-income countries will grow faster than low initial-income countries, once differences in saving rates and population growth rates have been accounted for. However, not all endogenous growth models imply an absence of β convergence and therefore caution needs to be exercised in drawing inferences about the nature of the growth process from the results of β convergence tests.
In moving from unconditional to conditional β convergence, complexities arise in terms of the specification of steady-state income. The reason for this is the dependence of the steady-state on Z . Theory is not always a good guide in the choice of elements of Z; differences in formulations of the standard equation have led to a “growth regression industry” as researchers have added one plausibly relevant variable after another to the baseline Solow specification. As a result, there are variants of the equation where convergence appears to occur and there are variants where divergence has a consistent interpretation.
Abramowitz (1986), Baumol (1986), DeLong (1988) among many others view convergence as the process of follower countries 'catching up' to leader countries by adopting their technologies. On the other hand some more recent contributors, starting with Barro (1991) and Mankiw, Romer, and Weil (1992), have emphasised the view that convergence is driven by diminishing returns to factors of production.
One hypothesis is that differences in capital accumulation, productivity, and therefore output per worker are fundamentally related to differences in social infrastructure across countries. By social infrastructure is normally meant the institutions and government policies that determine the economic environment within which individuals accumulate skills, and firms accumulate capital and produce output. A social infrastructure favorable to high levels of output per worker provides an environment that supports productive activities and encourages capital accumulation, skill acquisition, invention, and technology transfer. Such a social infrastructure
gets the prices right so that, in the language of North and Thomas [1973], individuals capture the social returns to their actions as private returns.
Some evidence for this can be found in Hall and Jones (1999) who - in a study running across 127 countries - find a powerful and close association between output per worker and measures of social infrastructure. Countries with long-standing policies which they characterise as favorable to productive activities produce much more output per worker. In one example, they find that the observed difference in social infrastructure between Niger and the United States is more than enough to explain the 35-fold difference in output per worker which is to be found between the two countries.
One influential line of research born from neo-classical growth research and which uses cross-country regressions to find the empirical determinants of the growth rate originates in the work of Robert Barro (1991). The basic Barro equation postulates that the growth rate of per capita GDP for a given country country equals a constant (beta) multiplied by a vector of variables that are thought to reflect determinants of long-term growth (the Zs) plus an error term.
Regressions which have this form are sometimes known as Barro regressions, given Barro’s extensive use of such regressions to study alternative growth determinants starting with Barro (1991). This regression model has been subsequantly become the workhorse of empirical growth research. In modern empirical analyses, the equation has been generalized in a number of dimensions. Some of these extensions reflect the application of the vasic equation to time series and panel data settings, while others have introduced nonlinearities and parameter heterogeneity.
The Economic Zs. In particular work in the Barro tradition has attempted to show growth in output per capita as being positively related to:
(1) A lower initial level of productivity. On this account convergence is posited to be more rapid in countries with higher levels of schooling attainment.
(2) Higher male secondary and tertiary schooling attainment, which facilitates the absorption of new technologies.
(3) Higher life expectancy, a proxy for better health and human capital in general.
(4) Improvement in terms of trade, posited to generate added employment and income.
(5) A lower rate of inflation, leading to better decisions with predictable price
expectations.
(6) A lower government consumption share netted of education and defense spending, which is posited to release resources for more productive private investment.
(7) Stronger democratic institutions at low levels of democracy, which promote market activity by loosening autocratic controls. However, stronger democracies at high
levels can dampen growth by the government exerting an increasingly active role in
redistributing income. Democrcy is thus entered in quadratic form, posited to rise and then fall.
(8) A stronger rule of law which stimulates investment by promoting sanctity of contracts, security of property rights, etc.
One revealing feature of the convergence-patterns models can be gleaned by considering variables omitted by Barro and those who have normally worked in this tradition. Here authors tend to emphasize variables that determine longrun, or “potential” steady state labor productivity, and downplay variables that bring about the “adjustment” or “transition” to long-run equilibrium.
An example of one such omitted variable would be investment shares. Putting aside the problem of endogeneity, investment can be viewed as an adjustment variable. The gap between current and long-run labor force productivity largely dictates the return to investment. Investment will flow to those countries with highest returns. Rather than investment accounting for growth per se, it can be argued that the “structural” features of countries that impede or facilitate investment should be highlighted in the modeling of Z (e.g., measures of the risk of expropriation, restrictive licensing, political conditions, etc.). Such features modify long-run potential labor productivity because they impede or encourage investment.
From Mathus to Solow
Another approach to the growth problem has been the attempt to develop a unified growth theory which accounts for the transition from the earlier Malthus regime to the modern growth one. Typical of that approach is the work of Hansen and Prescott (2002). They presnt the problem in the following way:
"Prior to 1800, living standards in world economies were roughly constant over the very long run: per capita wage income, output, and consumption did not grow.Modern industrial economies, on the other hand, enjoy unprecedented and seemingly endless growth in living standards."
The existing theoretical literature on the transition from stagnation to growth has focused mostly on the role played by endogenous technological progress and/or human capital accumulation rather than the role of land in production.4 For example, human capital accumulation and fertility choices play a central role in Lucas (1998), who builds on work by Becker, Murphy, and Tamura (1990). Depending on the value of a parameter governing the private return to human capital accumulation, Lucas’s model can exhibit either Malthusian or modern features. Hence, a transition from an economy with stable to growing living standards requires an exogenous change in the return to human capital accumulation.
Like Hansen and Prescott, Galor and Weil (2000) and Jones (1999) study models where the transition from Malthusian stagnation to modern growth is a feature of the equilibrium growth path, although their approaches differs by incorporating endogenous technological progress and fertility choice. Living standards are initially constant in these models due to the presence of a fixed factor in production and because population growth is increasing in living standards at this stage of development. In Galor and Weil (2000), growing population, through its assumed effect on the growth rate of skill-biased technological progress, causes the rate of return to human capital accumulation to increase. This ultimately leads to sustained growth in per capita income.
In Jones (1999), increasing returns to accumulatable factors (useable knowledge and labor) causes growth rates of population and technological progress to accelerate over time and, eventually, this permits an escape from Malthusian stagnation.
A transition from Malthus to Solow implies that land has become less important as a
factor of production. Indeed, the value of farmland relative to the value of gross national product (GNP) has declined dramatically in past two centuries. The value of farmland relative to annual GNP in the US fell from 88 percent in 1870 to less than 5 percent in 1990.
Steady State Growth in the USA
All of this leads naturally on to another question, can exponential growth be sustained forever?
How do we understand the exponential increase in per capita income observed over the last 150 years? The growth literature provides a large number of candidate theories to address these questions, with the evident characteristic that such theories are nearly always constructed so as to generate a steady state, or balanced growth path, in the long term.
That is to say that the growth rate of per capita income settles down eventually to a constant. In part, this choice reflects modeling convenience. However, it is also a desirable feature of any model that is going to fit some of the 'stylised facts' of growth.
Barro and Sala-i-Martin (1995) for example argue that one good reason to stick with the simpler framework with a steady state is that the long-term experiences of the United States and some other developed countries indicate that per capita growth rates can be positive and trendless over long periods of time... This empirical phenomenon suggests that a useful theory would predict that per capita growth rates approach constants in the long run; that is, the model would possess a steady state.
Clearly there are many examples of countries that display growth rates that are rising or falling for decades at a time, and thus it is a pertinent question to ask just how 'typical' is the US in this case?
As Kaldor of course recognised, the United States has for the last 125 years exhibited positive growth for long periods with no noticeable trend. However while it seems reasonable that a successful candidate for a generalised growth theory should at least admit the possibility of steady-state growth, it is not clear that this steady state generation need be a model property.
References
Abramowitz, M., (1986), “Catching Up, Forging Ahead and Falling Behind,” Journal of Economic History, 46, 385-406.
Baumol, W., (1986), “Productivity Growth, Convergence, and Welfare: What the Longrun Data Show,” American Economic Review, 76, 5, 1072-85
Barro, R.J. 1991. “Economic Growth in a Cross Section of Countries.” Quarterly Journal of Economics 106(2):407-444.
Barro, Robert J. 1997. "Determinants of Economic Growth: A Cross-Country
Empirical Study." Development Discussion Paper No. 579, Harvard Institute for
International Development, April.
Barro, Robert J. and Jong-Wha Lee. 1993. "Losers and Winners in Economic Growth." NBER Working Paper No. 4341. Cambridge, Mass.
Durlauf, Stephen, Paul Johnson and Jonathan Temple (2005), "Growth Econometrics" in Handbook of Economic Growth, Philippe Aghion and Stephen Durlauf editors, North Holland.
Easterly, W. and R. Levine, (2001), “It’s Not Factor Accumulation: Stylized Facts and Growth Models,” World Bank Economic Review, 15, 177-219.
Hall, R., and C. Jones, (1999), “Why Do Some Countries Produce So Much More Output Per Worker Than Others?,” Quarterly Journal of Economics, 114, 1, 83-116.
Hansen, G., and E. Prescott (2002), “Malthus to Solow”, American Economic Review 92: 1205-1217.
Harberger, A., (1987), “Comment,” in Macroeconomics Annual 1987, S. Fischer, ed., Cambridge: MIT Press.
Henderson, D. and R. Russell, (2004), “Human Capital and Convergence: A Production Frontier Approach,” mimeo, SUNY Binghamton and forthcoming, International Economic Review.
Kaldor, Nicholas, “Capital Accumulation and Economic Growth,” in F.A. Lutz and D.C. Hague, eds., The Theory of Capital, St. Martins Press, 1961.
Klenow, P. and A. Rodriguez-Clare, (1997a), “The Neoclassical Revival in Growth Economics: Has it Gone Too Far?,” in Macroeconomics Annual 1997, B. Bernanke and J. Rotemberg, eds., Cambridge: MIT Press.
Kuznets, S (1971), Simon Kuznets – Prize Lecture, Lecture to the memory of Alfred Nobel, December 11, 1971. Available online at http://nobelprize.org/economics/laureates/1971/kuznets-lecture.html
Levine, Ross and David Renelt. 1992. “A Sensitivity Analysis of Cross-Country Growth Regressions.” American Economic Review 82, 4 (September) 942-963.
Lucas, R., (1988), “On the Mechanics of Economic Development,” Journal of Monetary Economics, 22, 1, 3-42.
Mankiw, N. G., D. Romer, and D. Weil, (1992), “A Contribution to the Empirics of Economic Growth,” Quarterly Journal of Economics, 107, 2, 407-37.
North, Douglass C. and Robert P. Thomas (1973) The Rise of the Western World: A
New Economic History, New York; Cambridge University Press.
Prescott, E., (1998), “Needed: A Theory of Total Factor Productivity,” International Economic Review, 39, 525-551.
Romer, P., (1986), “Increasing Returns and Long-run Growth,” Journal of Political Economy, 94, 5, 1002-1037.
Solow, R., (1956), “A Contribution to the Theory of Economic Growth,” Quarterly Journal of Economics, 70, 1, 65-94.
Swan, T., (1956), “Economic Growth and Capital Accumulation,” Economic Record, 32, 334-361.
Durlauf et al (2005) put it like this:
"Understanding the wealth of nations is one of the oldest and most important research agendas in the entire discipline. At the same time, it is also one of the areas in which genuine progress seems hardest to achieve. The contributions of individual papers can often appear slender. Even when the study of growth is viewed in terms of a collective endeavor, the various papers cannot easily be distilled into a consensus that would meet standards of evidence routinely applied in other fields of economics".
Among the biggest problems facing growth theorists are the comparatively small number of countries to work from, the comparatively limited time duration of the available data, and the wide variety of explanatory models which have been proposed in explanation of the phenomenon. Indeed, in the course of a long and arduous literature history, approximately as many growth determinants have been proposed as there are countries for which data are available.
However despite the not inconsiderable amount of research and energy dedicated by economists to the theme of economic growth and development (Nobel economist Robert Lucas, for example, informs us that not a day goes by without his spending at least some time thinking about the topic), collectively we have yet to discovered how to make poor countries rich.
Whatsmore the road to progress seems fraught with difficulty. If we take it as self-evidently true that technology is one of the primary determinants of a country’s growth and income, and if the know-how relating to modern production technologies is essentially free and there for the taking, then why, oh why is it that so many people still find it so difficult to get on the development track?
Simply posing the question is, of course, to already answer it, since even in the case of something so apparently 'neutral' as technology there are clearly self-reinforcing mechanisms, or 'traps' (or 'bad equilibria') that act either as barriers to adoption or impediments to effective use even when adopted. Development traps arise from a variety of circumstances, and again these have generated an extensive literature with much of the focus being centred on market imperfection and institutional failure.
Another explanation often advanced for the poor growth performance of the LDCs is the existence of poor institutions and bad domestic policy. And of course those who argue this do so not without justification. The LDCs as a group are not exactly short on examples of corruption and bad government. And clearly sound governance and healthy market influences are necessary to reap the benefits of economic growth. But isn't this story perhaps just a little too simple.
In a pathbreaking paper on Geography, Demography and Economic Growth in Africa, David Bloom and Jeffrey Sachs make the following plea:
"Our paper could well be misunderstood. Some will regard it as a new case of "geographic determinism," that Africa is fated to be poor because of its geography. Some will regard it as a distraction from the important truth that geographic difficulties or not, African governments seriously mismanaged economic policy in the past generation. Let us therefore be clear at the outset. We believe emphatically that economic policy matters, and our formal econometric results show that to be true, a point we have also made in related recent studies (especially Sachs and Warner, 1997) We nonetheless focus most of our attention on geography for three reasons. First, there is little to be gained from yet another recitation of the damage of statism, protectionism, and corruption on African economic performance. Amen. Second, most economists are woefully neglectful of the forces of nature in shaping economic performance, in general and in Africa in particular. They treat economies as blank slates, upon which another region's technologies and economic history may be grafted. Our profession's formal models tend to be like that; so do our profession's standard statistical analyses of cross-country growth....Third, and perhaps most importantly, good policies must be tailored to geographical realities.If agricultural productivity is very low in Africa for climatological reasons, perhaps the real lesson is that growth should be led much more by outward-oriented industry and services, rather than yet another attempt to blindly transplant "integrated rural development" strategies from other parts of the world that are not customized to Africa’s unique conditions."
If Bloom and Sachs are accused of geographical determinism, it is easy to imagine that I might here be accused of "demographic determinism", so let me just endorse the above argument: economic and political policy also matters. But what I would also like to stress is the importance of their second point, namely that "most economists are woefully neglectful of the forces of nature in shaping economic performance, in general and in Africa in particular. They treat economies as blank slates, upon which another region's technologies and economic history may be grafted." To ram this home: most economists' appreciation of the workings of evolutionary biology and its implications for demographic processes is rudimentary to say the least. And what outrageously many economists understand about demographic processes in and of themselves their relevance for growth and development would provide material for only the shortest of short papers. The fact is, with some emminent and notable exceptions, demography is simply not seen as an important part of the macro-economic panorama.
In this context I would simply single out two underlying prejudices which I think have had an enormous influence.
In the first place the 'state of nature' argument. Basically most economists seem to work from some variant of Cantillion's notion that: "Men multiply like mice in a barn if they have unlimited means of subsistence", ie that in the pre-modern (Mathusian regime epoch) women produced children at an extraordinarily high rate. This is certainly a counter-factual belief, not only because in most foraging societies fertility falls universally well short of the biological maximum, but also since there is often evidence in history for the idea that as subsistance rises fertility declines.
Secondly the homeostatic principle which normally lies behind modern conceptions of steady state growth. Basically - and this is a point which has been often made - most economists seem to see their discipline as much closer to physics than it is to biology. So the preferred yardstick for measuring the validity of our economic models is far more often the celestial motion of the planets than ever it is the Darwinian evolution of living organisms. So the essential idea is that 'modern' (or if you want market-driven) economies represent a break from an earlier high fertility regime to a new growth regime which has a steady state dynamic, or balanced growth path, and has an internalised homeostatic conditioner which produces a population fertility rate of on or around a 2.1 TFR. Whether by accident or design, the excellent Consise History of Population by Massimo Livvi Bacci is the most widely quoted population reference source in modern growth economics. I say whether by accident or design, because, Livvi-Bacci is, of course, the best know demographic exponent of the homeostatic population view.
To concretise, the story that the competitive neoclassical benchmark model of economic systems tells is the following one: Markets are complete, entry and exit is free, transaction costs are negligible, and technology is convex at an efficient scale relative to the size of the market. As a result, the private and social returns to production and investment are equal and a complete set of 'virtual prices' ensures that all projects with positive net social benefit are undertaken. Diminishing returns to the set of reproducible factor inputs implies that as an economy develops the rate of return on investment declines, whilst where capital is scarce the returns to investment (and of course the level of risk) will be concomitanly higher.
The dynamic implications of this benchmark were intitially summarized by Solow (1956), Cass (1965), and Koopmans (1965). Even for countries which start with different endowments, the main conclusion is that there ought to be an underlying convergence process.
Needless to say the influence on development policy of what might be called the neoclassical 'prejudice' has been enormous. One of the clearest examples is, of course, to be found in the series of structural adjustment programs implemented by the International Monetary Fund. The key components of the Enhanced Structural Adjustment Facility —the centerpiece of the IMF’s strategy to aid poor countries and promote long run growth in the years between 1987 and 1999 — were prudent macroeconomic policies and the liberalization of markets. Growth, it was hoped, would then follow automatically.
Yet the evidence on whether or not non-distortionary policies and diminishing returns to capital will soon carry the poor to opulence is mixed to say the least. In recent years even relatively well governed countries have experienced little or no growth. Mali, for example, although not exactly corruption-free, scores relatively well in comparative measures of governance and real resources (Radlet 2004; Sachs et al. 2004). Yet Mali is still desperately poor. According to a 2001 UNDP report, 70% of the population lives on less than $1 per day. The infant mortality rate is 230 per 1000 births, and household final consumption expenditure is down 5% from 1980. And of course Mali continues to have one of the highest TFR rates on the planet. And Mali is not exactly an isolated case.
So are we perhaps missing something here?
Well lets go back to where we started. Now apart from the existence of model uncertainty, another major set of growth issues revolves around the difficulty of identifying empirically salient determinants of growth when the range of potential factors is large relative to the number of observations.
Individual researchers, seeking to identify the extent of support for particular growth determinants, typically emphasize a single model (or small set of models) and then carry out an inference procedure as if that model had generated the data. Standard inference procedures based on a single model, and which are conditional on the truth of that model, can grossly overstate the precision of inferences about a given phenomenon. In particular such procedures often ignore the uncertainty that surrounds the validity of the model itself.
Given that there are usually other models that have strong claims to consideration, the standard errors which arise can significantly understate the true degree of uncertainty about the parameters, and the choice of model to can ultimately appear somewhat arbitrary. This need for a proper account of model uncertainty has lead naturally toward Bayesian or pseudo-Bayesian approaches to data analysis.
Despite the existence of these widely known problems attempts to develop a unified growth theory along the abovementioned lines have not been notably deterred. In particular we have seen a continuing effort to establish common 'covergent' characteristics to growth across countries. One such approach has centred around the attempt to establish a series of 'stylised facts' about growth.
Stylised Facts
As I have said, most neo-classical growth models start from the idea of steady state growth. One idea closely associated with the steady state one is that there exist a number of 'stylised facts' of growth which are waiting to be uncovered. The tradition of attempting to produce stylised facts in economic growth theory goes back at least to the work of Kuznets and Kaldor in the 1960s.
Kaldor (1961) described what he considered to be a number of stylized facts about economic growth:
1. Per capita output grows over time, and its growth rate does not tend to diminish.
2. Physical capital per worker grows over time.
3. The rate of return to capital is nearly constant.
4. The ratio of physical capital to output is nearly constant.
5. The shares of labor and physical capital in national income are nearly constant.
6. The growth rate of output per worker differs substantially across countries.
When Simon Kuznets delivered his Nobel Address back in 1971 he also singled out six stylised 'characteristics' of modern economic growth:
1/. The high rates of growth of per capita product and of population in the developed countries.
2/. The high rate of rise in productivity, i.e. of output per unit of all inputs in comparison with earlier epochs.
3/. The rate of structural transformation of the economy is also high by historical standards.
4/. Structures which are extremely important in general societal terms - eg levels of urbanization and secularization - have also changed rapidly, leading sociologists to use the term modernization process.
5/. The economically developed countries have the propensity to reach out to the rest of the world (we would now call this globalisation).
6/. Despite the spread of modern economic growth, wide inequality in growth rates and income levels persists.
These early efforts (and in particular those of Kaldor) have not worn especially well with time. As Sala i Martin (2002) points out one important innovation of the new generation of growth literature that took off in the late 80s was that it attempted to tie economic theory much closer to empirical studies of growth. The neoclassical literature of the 1960s seem to link theory and evidence by simply 'mentioning' a list of stylized facts (such the Kaldor 'facts' above) and then attempting to show that the theory proposed was consistent with one, two or perhaps several of
these 'facts'. Indeed as Sala i Martin notes, some of these 'stylised facts' were never really derived from careful empirical analysis, but that did not stop them being quoted and used as if they had been. (Kuznets himself, of course, had carried out significant empirical research on economic growth).
Today’s research, on the other hand, tends to be characterised by a drive to derive more precise econometric specifications through direct recourse to the data. The best examples of this kind of work can be found in the convergence literature. Barro and Sala-i-Martin (1992) use the Ramsey-Cass-Koopmans (Ramsey (1928), Cass (1975) and Koopmans (1965)) growth model to derive an econometric equation that relates growth of GDP per capita to initial levels of GDP. Mankiw, Romer and Weil (1992) derive a similar equation from the Solow-Swan model ((Solow (1956) and Swan (1956)). These researchers derived a relationship where the growth rate of per capita GDP is given by the difference between a constant (beta) and the product of that same beta and the natural log of per capita GDP for country and the steady-state value of per capita GDP for country i plus an error term.
The coefficient is positive in the case that the production function is neoclassical, and is zero if the production function is linear in capital (which was usually the case in the first generation one-sector models of endogenous growth, also known as AK models). The central point is that the modern literature took this equation and used it to 'test' competing the models, namely the endogenous growth ones (the so called AK models) which should predict beta = 0 and the neoclassical models (which should predict beta>0.)
Early findings seemd to indicate that there was no positive association between growth and the initial level of income. The possibility existed however that this finding could be a statistical artifact resulting from the misspecification of the original equation. Basically, the problem is that, if researchers make the assumption that countries converge to the same steady state and the finding is that they don’t, then the original equation is, in turn, misspecified. If the steady state is correlated with the initial level of income, then the error term is correlated with the explanatory variable, so the estimated coefficient is biased towards zero.
One proposed solution to the problem was to continue using cross-country data but, instead of estimating the univariate regression, estimate a multivariate regression where, on top of the initial level of income, the researcher would also hold constant proxies for the steady state. This approach came to be known as conditional convergence. Subsequent research has shown that the idea of conditional convergence describes one of the strongest and most robust empirical regularities to be found in the data. The consequence of all this was the exact opposite of the original conclusion: the neoclassical model was not rejected by the data. The AK model was.
Sala i Martin draws a number of lessons from all this:
(i) There is no simple determinant of growth.
(ii) The initial level of income is the most important and robust variable (so conditional convergence is the most robust empirical fact in the data).
(iii) The size of the government does not appear to matter much. What is important is the 'quality of government'.
(iv) The relation between most measures of human capital and growth is weak. Some
measures of health, however, (such as life expectancy) are robustly correlated with growth.
(v) Institutions (such as free markets, property rights and the rule of law) are important for growth.
(vi) More open economies tend to grow faster.
As a result of this re-evaluation it is now possible to outine a rather different group of stylised facts:
1. Over the second half of the 20th century most countries have grown richer, but at varying rates, and vast income disparities remain. For all but the richest group, growth rates have differed to an unprecedented extent, regardless of the initial level of development. Indeed far from a clear convergence process reality has rather conformed to what Lance Pritchett once famously termed 'divergence bigtime'.
2. Past growth has turned out to be a surprisingly weak predictor of future growth. In the developing world distinct 'winners' and 'losers' have begun to emerge. The strongest performers are located in East and Southeast Asia, which have sustained growth rates at unprecedented levels. The weakest performers are predominantly located in sub-Saharan Africa, where some countries have barely grown at all, or even become poorer. The record in South and Central America is also distinctly mixed. In these regions, output volatility is high, and dramatic output collapses are not uncommon.
3. For many countries, growth rates were lower in 1980-2000 than in 1960-1980, and this growth slowdown has been observed throughout most of the income distribution. Moreover, the dispersion of growth rates has increased. A more optimistic reading would also emphasize the growth take-off that has taken place in China and India, home to two-fifths of the world’s population and a greater proportion of the world’s poor.
But despite all the research two oustanding questions still seem to remain.
Firstly, the issue of convergence: are contemporary differences in aggregate economies transient over sufficiently long time horizons?
And secondly, what are the key growth determinants: which factors are better able to explain observed differences in growth?
Two separate approaches have continued to dominate the extensive literature on economic growth modeling: simple correlations, and production functions.
Simple-correlations
Simple-correlations studies hypothesize that per capita output growth is influenced by various dimensions.
Production Functions
Production-function studies are based on estimating variants of a model along the following lines:
Y = g(K, L, H, R, T),
where output (Y) is produced by the stocks of various factors: physical capital (K), labor (L), human capital (H: education and health), resources (R: land, minerals, and environment), and technology (T).
In the neoclassical model, if each country has access to the same aggregate production function the steady-state is independent of an economy's initial capital and labor stocks and hence initial income. In this model, long-run differences in output reflect differences in the determinants of accumulation, not differences in the technology used to combine inputs to produce output. Mankiw (1995, p. 301), for example, argues that for “understanding international experience, the best assumption may be that all countries have access to the same pool of knowledge, but differ by the degree to which they take advantage of this knowledge by investing in physical and human capital.”
Even if one relaxes the assumption that countries have access to the same production function, convergence in growth rates can still occur so long as each country’s production function is concave in capital per efficiency unit of labor and each country experiences the same rate of labor-augmenting technical change.
Klenow and Rodríguez-Clare (1997a) challenge this “neoclassical revival” with
results suggesting that differences in factor accumulation are, at best, no more important than differences in productivity in explaining the cross-country distribution of output per capita. They find that only about half of the cross-country variation in the 1985 level of output per worker is due to variation in human and physical capital inputs while a mere 10% or so of the variation in growth rates from 1960 to 1985 reflects differences in the growth of these inputs. The differences between the results of Mankiw, Romer, and Weil (1992) and the findings of Klenow and Rodríguez-Clare (1997a) in their reexamination of Mankiw, Romer and Weil have two principal origins. MRW found support for the Solow model’s predictions that, in the long-run steady state, the level of real output per worker by country should be positively correlated with the saving rate and negatively correlated with the rate of labor-force growth. However, their estimates of the textbook Solow model also implied a capital share of factor income of about 0.60, high compared to the conventional value (based on U.S. data) of about one-third.
To address this possible inconsistency, MRW considered an “augmented” version of the Solow model, in which human capital enters as a factor of production in symmetrical fashion with physical capital and raw labor. They found that the augmented Solow model fits the data relatively better and yields an estimated capital share more in line with conventional wisdom. They concluded (abstract, p. 407) that “an augmented Solow model that includes accumulation of human as well as physical capital provides an excellent description of the cross-country data.”
[Bernanke and Gürkaynak model]: Assume that in a given country at time t, output Y depends on inputs of raw labor L and three types of accumulated factors: K , H , and Z . The factors K and H are accumulated through the sacrifice of current output (think of physical capital and human capital, or structures and equipment). The factor Z , which could be an index of technology, or of human capital acquired through learning-by-doing, is assumed to be accumulated as a byproduct of economic activity and does not require the sacrifice of current output.]
Y = K, H (Z, L).
Bernanke and Gürkaynak also use the MRW framework to consider some alternative models of economic growth, such as the Uzawa-Lucas model and the AK model. These models are rejected as literal descriptions of the data.
However, the implications of these models, that country growth rates depend on behavioral variables such as the rate of human capital formation and the saving rate, seem more consistent with the data than the Solow model’s assumption that growth is exogenous. Future research should consider variants of endogenous growth models to see which if any provide a more complete and consistent description of the
cross-country data. We believe that the generalized MRW-type framework we have developed here could prove very helpful in assessing the alternative possibilities.
Indeed both Prescott (1998) and Hall and Jones (1999) confirm the view that differences in inputs are unable to explain observed differences in output and Easterly and Levine (2001, p. 177) state that "the 'residual' (total factor productivity, TFP) rather than factor accumulation accounts for most of the income and growth differences across countries."
Despite these concerns and the differences in the precise estimates found by different researchers, it is clear that cross-country variation in inputs falls short of explaining the observed cross-country variation in output. The result that the TFP
residual, a “measure of our ignorance” computed as the ratio of output to some index of inputs, is an important (perhaps the dominant) source of cross-country differences in long-run economic performance is useful but hardly satisfying and the need for a theory of TFP expressed by Prescott (1998) is well founded. Research such as Acemoglu and Zilibotti (2001) and Caselli and Coleman (2003) are promising contributions to that agenda.
Convergence Patterns
Convergence-patterns studies, rooted in neoclassical growth theory, explore the relationships between economic growth and the *level* of economic development. They focus on the pace at which countries move from their current level of labor productivity (Y/L) to their long-run, or steady-state equilibrium level of labor productivity.
The effect of initial conditions on long-run outcomes arguably represents the primary empirical question that has been explored by growth economists. The claim that the effects of initial conditions eventually disappear is the heuristic basis for what is
known as the convergence hypothesis.
The goal of this literature is to answer two questions concerning per capita income differences across countries (or other economic units, such as regions). First, are the observed cross-country differences in per capita incomes temporary or permanent? Second, if they are permanent, does that permanence reflect structural heterogeneity or the role of initial conditions in determining long-run outcomes?
If the differences in per capita incomes are temporary, unconditional convergence (to a common long-run level) may occur. If the differences are permanent solely because of cross-country structural heterogeneity, only conditional convergence may occur. If initial conditions determine, in part at least, long-run outcomes, and countries with similar initial conditions exhibit similar long-run outcomes, then one can speak of convergence clubs.
In practice, the distinction between initial conditions and structural heterogeneity generally amounts to treating stocks of initial human and physical capital as the former and other variables as the latter. As such, both the Solow variables X and the control variables Z that appear in garden variety cross-country growth regression are usually interpreted as capturing structural heteogeneity. This practice may be criticized if these variables are themselves endogenously determined by initial conditions, a point that will arise below.
Here the rate of labour productivity growth is taken to be proportional to the gap between the logs of the long-run, steady state and the current level of labor productivity. The greater this gap, the greater are the gaps of physical capital, human capital, and technical efficiency from their long run levels. Large gaps allow for significant "catching up" through (physical and human) capital accumulation, and through technology creation and diffusion across, and within countries.
Under restrictive assumptions, this type of model predicts "unconditional convergence" by all countries to the same long-run level of labor productivity. Were the steady state rate of growth of labor productivity to be the same for all countries, then low-income countries would be father away from their state determined income level and hence - following the standard equation - their productivity would grow faster as a result. In fact, however, positive rather than negative correlations have been observed between the level and growth rate of labor productivity. The model has not unsurprisingly been modified as a consequence.
As indicated above, models now tend to hypothesize "conditional convergence", whereby long-run labor productivity differs across countries depending on country-specific characteristics following the following equation:
ln(rate of change of labour productivity) = a + bZ.
The actual specification of the determinants of long-run labor productivity (the Z's) varies notably, but the basic model is the same across scores of empirical studies.
X contains a composite - log (n+g+delta) - where delta denotes the depreciation rate. The variables spanned by log income and X thus represent those growth determinants that are suggested by the Solow growth model whereas Z represents those growth determinants that lie outside Solow’s original theory. The distinction between the Solow variables and Z is important in understanding the empirical literature. While the standard Solow variables usually appear across a wide variety of different empirical studies - a situation which reflects the treatment of the Solow model as a baseline for growth analysis - choices concerning which Z variables to include vary greatly.
As outlined above, statistical analyses of convergence have largely focused on the properties of β in regressions. β convergence, defined as β<0 is easy to evaluate because it relies on the properties of a linear regression coefficient. It is also easy to interpret in the context of the Solow growth model, since the finding is consistent with the dynamics of the model. The economic intuition for this is simple.
If two countries have common steady-state determinants and are converging to a common balanced growth path, the country that begins with a relatively low level of initial income per capita has a lower capital-labor ratio and hence a higher marginal product of capital; a given rate of investment then translates into relatively fast growth for the poorer country.
This is why β convergence is commonly interpreted as evidence against endogenous growth models of the type studied by Romer and Lucas, since a number of these models specifically predict that high initial-income countries will grow faster than low initial-income countries, once differences in saving rates and population growth rates have been accounted for. However, not all endogenous growth models imply an absence of β convergence and therefore caution needs to be exercised in drawing inferences about the nature of the growth process from the results of β convergence tests.
In moving from unconditional to conditional β convergence, complexities arise in terms of the specification of steady-state income. The reason for this is the dependence of the steady-state on Z . Theory is not always a good guide in the choice of elements of Z; differences in formulations of the standard equation have led to a “growth regression industry” as researchers have added one plausibly relevant variable after another to the baseline Solow specification. As a result, there are variants of the equation where convergence appears to occur and there are variants where divergence has a consistent interpretation.
Abramowitz (1986), Baumol (1986), DeLong (1988) among many others view convergence as the process of follower countries 'catching up' to leader countries by adopting their technologies. On the other hand some more recent contributors, starting with Barro (1991) and Mankiw, Romer, and Weil (1992), have emphasised the view that convergence is driven by diminishing returns to factors of production.
One hypothesis is that differences in capital accumulation, productivity, and therefore output per worker are fundamentally related to differences in social infrastructure across countries. By social infrastructure is normally meant the institutions and government policies that determine the economic environment within which individuals accumulate skills, and firms accumulate capital and produce output. A social infrastructure favorable to high levels of output per worker provides an environment that supports productive activities and encourages capital accumulation, skill acquisition, invention, and technology transfer. Such a social infrastructure
gets the prices right so that, in the language of North and Thomas [1973], individuals capture the social returns to their actions as private returns.
Some evidence for this can be found in Hall and Jones (1999) who - in a study running across 127 countries - find a powerful and close association between output per worker and measures of social infrastructure. Countries with long-standing policies which they characterise as favorable to productive activities produce much more output per worker. In one example, they find that the observed difference in social infrastructure between Niger and the United States is more than enough to explain the 35-fold difference in output per worker which is to be found between the two countries.
One influential line of research born from neo-classical growth research and which uses cross-country regressions to find the empirical determinants of the growth rate originates in the work of Robert Barro (1991). The basic Barro equation postulates that the growth rate of per capita GDP for a given country country equals a constant (beta) multiplied by a vector of variables that are thought to reflect determinants of long-term growth (the Zs) plus an error term.
Regressions which have this form are sometimes known as Barro regressions, given Barro’s extensive use of such regressions to study alternative growth determinants starting with Barro (1991). This regression model has been subsequantly become the workhorse of empirical growth research. In modern empirical analyses, the equation has been generalized in a number of dimensions. Some of these extensions reflect the application of the vasic equation to time series and panel data settings, while others have introduced nonlinearities and parameter heterogeneity.
The Economic Zs. In particular work in the Barro tradition has attempted to show growth in output per capita as being positively related to:
(1) A lower initial level of productivity. On this account convergence is posited to be more rapid in countries with higher levels of schooling attainment.
(2) Higher male secondary and tertiary schooling attainment, which facilitates the absorption of new technologies.
(3) Higher life expectancy, a proxy for better health and human capital in general.
(4) Improvement in terms of trade, posited to generate added employment and income.
(5) A lower rate of inflation, leading to better decisions with predictable price
expectations.
(6) A lower government consumption share netted of education and defense spending, which is posited to release resources for more productive private investment.
(7) Stronger democratic institutions at low levels of democracy, which promote market activity by loosening autocratic controls. However, stronger democracies at high
levels can dampen growth by the government exerting an increasingly active role in
redistributing income. Democrcy is thus entered in quadratic form, posited to rise and then fall.
(8) A stronger rule of law which stimulates investment by promoting sanctity of contracts, security of property rights, etc.
One revealing feature of the convergence-patterns models can be gleaned by considering variables omitted by Barro and those who have normally worked in this tradition. Here authors tend to emphasize variables that determine longrun, or “potential” steady state labor productivity, and downplay variables that bring about the “adjustment” or “transition” to long-run equilibrium.
An example of one such omitted variable would be investment shares. Putting aside the problem of endogeneity, investment can be viewed as an adjustment variable. The gap between current and long-run labor force productivity largely dictates the return to investment. Investment will flow to those countries with highest returns. Rather than investment accounting for growth per se, it can be argued that the “structural” features of countries that impede or facilitate investment should be highlighted in the modeling of Z (e.g., measures of the risk of expropriation, restrictive licensing, political conditions, etc.). Such features modify long-run potential labor productivity because they impede or encourage investment.
From Mathus to Solow
Another approach to the growth problem has been the attempt to develop a unified growth theory which accounts for the transition from the earlier Malthus regime to the modern growth one. Typical of that approach is the work of Hansen and Prescott (2002). They presnt the problem in the following way:
"Prior to 1800, living standards in world economies were roughly constant over the very long run: per capita wage income, output, and consumption did not grow.Modern industrial economies, on the other hand, enjoy unprecedented and seemingly endless growth in living standards."
The existing theoretical literature on the transition from stagnation to growth has focused mostly on the role played by endogenous technological progress and/or human capital accumulation rather than the role of land in production.4 For example, human capital accumulation and fertility choices play a central role in Lucas (1998), who builds on work by Becker, Murphy, and Tamura (1990). Depending on the value of a parameter governing the private return to human capital accumulation, Lucas’s model can exhibit either Malthusian or modern features. Hence, a transition from an economy with stable to growing living standards requires an exogenous change in the return to human capital accumulation.
Like Hansen and Prescott, Galor and Weil (2000) and Jones (1999) study models where the transition from Malthusian stagnation to modern growth is a feature of the equilibrium growth path, although their approaches differs by incorporating endogenous technological progress and fertility choice. Living standards are initially constant in these models due to the presence of a fixed factor in production and because population growth is increasing in living standards at this stage of development. In Galor and Weil (2000), growing population, through its assumed effect on the growth rate of skill-biased technological progress, causes the rate of return to human capital accumulation to increase. This ultimately leads to sustained growth in per capita income.
In Jones (1999), increasing returns to accumulatable factors (useable knowledge and labor) causes growth rates of population and technological progress to accelerate over time and, eventually, this permits an escape from Malthusian stagnation.
A transition from Malthus to Solow implies that land has become less important as a
factor of production. Indeed, the value of farmland relative to the value of gross national product (GNP) has declined dramatically in past two centuries. The value of farmland relative to annual GNP in the US fell from 88 percent in 1870 to less than 5 percent in 1990.
Steady State Growth in the USA
All of this leads naturally on to another question, can exponential growth be sustained forever?
How do we understand the exponential increase in per capita income observed over the last 150 years? The growth literature provides a large number of candidate theories to address these questions, with the evident characteristic that such theories are nearly always constructed so as to generate a steady state, or balanced growth path, in the long term.
That is to say that the growth rate of per capita income settles down eventually to a constant. In part, this choice reflects modeling convenience. However, it is also a desirable feature of any model that is going to fit some of the 'stylised facts' of growth.
Barro and Sala-i-Martin (1995) for example argue that one good reason to stick with the simpler framework with a steady state is that the long-term experiences of the United States and some other developed countries indicate that per capita growth rates can be positive and trendless over long periods of time... This empirical phenomenon suggests that a useful theory would predict that per capita growth rates approach constants in the long run; that is, the model would possess a steady state.
Clearly there are many examples of countries that display growth rates that are rising or falling for decades at a time, and thus it is a pertinent question to ask just how 'typical' is the US in this case?
As Kaldor of course recognised, the United States has for the last 125 years exhibited positive growth for long periods with no noticeable trend. However while it seems reasonable that a successful candidate for a generalised growth theory should at least admit the possibility of steady-state growth, it is not clear that this steady state generation need be a model property.
References
Abramowitz, M., (1986), “Catching Up, Forging Ahead and Falling Behind,” Journal of Economic History, 46, 385-406.
Baumol, W., (1986), “Productivity Growth, Convergence, and Welfare: What the Longrun Data Show,” American Economic Review, 76, 5, 1072-85
Barro, R.J. 1991. “Economic Growth in a Cross Section of Countries.” Quarterly Journal of Economics 106(2):407-444.
Barro, Robert J. 1997. "Determinants of Economic Growth: A Cross-Country
Empirical Study." Development Discussion Paper No. 579, Harvard Institute for
International Development, April.
Barro, Robert J. and Jong-Wha Lee. 1993. "Losers and Winners in Economic Growth." NBER Working Paper No. 4341. Cambridge, Mass.
Durlauf, Stephen, Paul Johnson and Jonathan Temple (2005), "Growth Econometrics" in Handbook of Economic Growth, Philippe Aghion and Stephen Durlauf editors, North Holland.
Easterly, W. and R. Levine, (2001), “It’s Not Factor Accumulation: Stylized Facts and Growth Models,” World Bank Economic Review, 15, 177-219.
Hall, R., and C. Jones, (1999), “Why Do Some Countries Produce So Much More Output Per Worker Than Others?,” Quarterly Journal of Economics, 114, 1, 83-116.
Hansen, G., and E. Prescott (2002), “Malthus to Solow”, American Economic Review 92: 1205-1217.
Harberger, A., (1987), “Comment,” in Macroeconomics Annual 1987, S. Fischer, ed., Cambridge: MIT Press.
Henderson, D. and R. Russell, (2004), “Human Capital and Convergence: A Production Frontier Approach,” mimeo, SUNY Binghamton and forthcoming, International Economic Review.
Kaldor, Nicholas, “Capital Accumulation and Economic Growth,” in F.A. Lutz and D.C. Hague, eds., The Theory of Capital, St. Martins Press, 1961.
Klenow, P. and A. Rodriguez-Clare, (1997a), “The Neoclassical Revival in Growth Economics: Has it Gone Too Far?,” in Macroeconomics Annual 1997, B. Bernanke and J. Rotemberg, eds., Cambridge: MIT Press.
Kuznets, S (1971), Simon Kuznets – Prize Lecture, Lecture to the memory of Alfred Nobel, December 11, 1971. Available online at http://nobelprize.org/economics/laureates/1971/kuznets-lecture.html
Levine, Ross and David Renelt. 1992. “A Sensitivity Analysis of Cross-Country Growth Regressions.” American Economic Review 82, 4 (September) 942-963.
Lucas, R., (1988), “On the Mechanics of Economic Development,” Journal of Monetary Economics, 22, 1, 3-42.
Mankiw, N. G., D. Romer, and D. Weil, (1992), “A Contribution to the Empirics of Economic Growth,” Quarterly Journal of Economics, 107, 2, 407-37.
North, Douglass C. and Robert P. Thomas (1973) The Rise of the Western World: A
New Economic History, New York; Cambridge University Press.
Prescott, E., (1998), “Needed: A Theory of Total Factor Productivity,” International Economic Review, 39, 525-551.
Romer, P., (1986), “Increasing Returns and Long-run Growth,” Journal of Political Economy, 94, 5, 1002-1037.
Solow, R., (1956), “A Contribution to the Theory of Economic Growth,” Quarterly Journal of Economics, 70, 1, 65-94.
Swan, T., (1956), “Economic Growth and Capital Accumulation,” Economic Record, 32, 334-361.
Subscribe to:
Posts (Atom)
