VoxEU Column Development Health Economics Macroeconomic policy

Diseases and development: Does life expectancy increase income growth?

Does rising life expectancy boost economic growth? Existing evidence is mixed, with the relationship appearing to change over time. This column presents recent research showing that living longer may have a negative effect on growth to begin with, but once fertility declines the effect becomes significantly positive. Moreover, higher life expectancy increases the probability of such a switch in fertility behaviour.

Across countries, high life expectancy is associated with high income per capita. But do improvements in life expectancy cause increases in per capita income? The answer to this question has far-reaching implications. Support to implement better health infrastructure may not only affect individual well-being, but also foster economic development.

Contradictory evidence

With rising awareness of econometric problems in the identification of causal effects in recent years, there has been renewed interest among economists in the question of a causal effect of life expectancy on economic development. Part of the literature, such as recent work by Lorentzen et al (2008), has used cross-country variation in geo-climatological conditions to identify the effect of interest and has found evidence for higher life expectancy leading to faster economic growth. Acemoglu and Johnson (2007), on the other hand, propose a new identification strategy that is based on within-country variation. The instrument exploits the so-called ‘Epidemiological Transition’ after World War II that led to exogenous reductions in mortality from several important infectious diseases.1 Their findings document that increasing life expectancy has a small effect on growth in aggregate incomes but a larger effect on population growth, thereby leading to a significant reduction in per capita income growth. These findings have fuelled an intense debate on the causal effect of life expectancy on economic growth, and on the fundamental causes of long-run growth.

Does the effect change over time?

In theory, an increase in life expectancy may have positive or negative effects on per capita income. On the one hand it may increase the productivity of available resources, e.g., by improving health of workers, and it may increase the incentives to undertake long-term investments like (most notably) human capital. On the other hand, higher life expectancy may lead to an increase in population that may depress income per-capita in the presence of Malthusian (congestion) effects.

A further theoretical prediction that has been neglected in the available empirical tests stems from the unified growth literature (see Galor 2005 for a survey) and relates to the presence of non-monotonic population dynamics associated with the ‘demographic transition’.

In virtually all countries and at all times, the onset of reductions in mortality preceded the onset of reductions in fertility. Before the onset of the fertility transition, which leads to substantial reductions in birth rates, population growth indeed rises in response to reductions in mortality. After the onset of the fertility transition, the opposite is true, with population growth decreasing in response to further reductions in mortality. Furthermore, the positive effects of increasing longevity, like the higher acquisition of human capital, are expected to be stronger in association with the reduction in fertility. According to these theories the effect of life expectancy can be negative before the transition but there should be no reasons to expect an unambiguous effect after the onset of the fertility drop. 

The existing empirical studies on the causal effect of life expectancy on economic growth focus on the problem of econometric identification within a linear growth framework, thereby neglecting the possibility of non-monotonic effects that depend on the states of the demographic transition. In recent work (Cervellati and Sunde 2011a) we estimate the causal effect of life expectancy on economic growth while allowing for a potentially non-monotonic effect of life expectancy that may depend on the demographic development. Using the same data and identification strategy of Acemoglu and Johnson (2007) as a benchmark, we extend their estimation framework by explicitly considering whether or not a country had already experienced the onset of the demographic transition before the treatment in 1940.

Figure 1 plots changes in log life expectancy at birth from 1940 to 1980 against changes in log GDP per capita over the same period. The overall relationship is negative, as reflected in the findings of Acemoglu and Johnson.

Figure 1. Changes in life expectancy and GDP per capita, 1940–80

However, distinguishing between countries that had not entered the fertility transition by 1940 (‘pre-transitional’ countries) and countries where fertility had already begun to drop (‘post-transitional countries’), reveals that this negative overall effect masks substantial heterogeneity in the effect across the two sub-samples. While pre-transitional countries still exhibit a modest negative relationship, the relationship is strongly positive for post-transitional countries.

Our results from instrumental variables estimations (exploiting fully and partially interacted models) document that the level of demographic development at which a country experiences an increase in life expectancy is crucial for the subsequent effect on economic growth. The growth effect is negative, although often insignificant, for demographically underdeveloped countries, where increases in life expectancy lead to an acceleration of population growth. For countries past the demographic transition the growth effect of life expectancy is significantly positive. The findings also show that increasing life expectancy may be indirectly beneficial for growth by increasing the likelihood of a country undergoing the fertility transition.

Moreover, our results are not due to use of any specific identification strategy, they emerge with both ordinary least squares and with alternative instruments, including those used by Lorentzen et al (2008). In subsequent work (Cervellati and Sunde 2011b), we document the same patterns by estimating more flexible finite mixture models that allow us to test for the non-monotonic effect of life expectancy in two latent regimes (pre- and post-transitional) without having to rely on any classification of countries. The results of a recent study (Cervellati and Sunde 2012) complement the previous evidence by showing that the positive growth effects of increases in life expectancy in countries that have undergone the demographic transition can be attributed to an acceleration in the acquisition of human capital and a sizable reduction in fertility rates.

Implications for researchers and policymakers

Our empirical findings suggest that the effect of life expectancy on economic growth depends on the stage of the demographic transition. Improving health in countries in which the fertility transition has not yet taken place can have negative effects on economic development, mainly by increasing population size. Improving health in countries that have undergone the fertility transition, however, has positive growth effects. In addition, increases in life expectancy increase the likelihood that the fertility transition occurs. These findings have relevant policy implications. From a technical perspective, the interpretation of the average (treatment) effect of life expectancy may be difficult, and potentially misleading, in the presence of non-monotonic effects. The mixed findings in the literature appear to be driven to a large extent by the different sample composition with respect to countries in different phases of their demographic development rather than, as previously suggested, by different identification strategies.

References

Acemoglu, D, and S Johnson (2007), "Disease and Development: The Effect of Life Expectancy on Economic Growth", Journal of Political Economy,115(6):925-985.

Cervellati, M, and U Sunde (2011a), "Life Expectancy and Economic Growth: The Role of the Demographic Transition”, Journal of Economic Growth, 16(2): 99-133.

Cervellati, M, and U Sunde (2011b), “Disease and Development: The Role of Life Expectancy Reconsidered”, Economics Letters, 113(3):269-272.

Cervellati, M and U Sunde (2012), “Life Expectancy, Education and Fertility: The Demographic Transition as Critical Turning Point”, mimeo.

Galor, O (2005), "From Stagnation to Growth: Unified Growth Theory", in P Aghion and S Durlauf (eds.), Handbook of Economic Growth, Elsevier.

Lorentzen, P, J McMillan, and R Wacziarg (2008), "Death and Development", Journal of Economic Growth, 13(2):81-124.


1 The idea behind the instrument is that the exogenous reduction in mortality had to do with path-breaking medical innovations in the development of treatments and vaccines and the establishment of institutions such as UNICEF or the WHO that granted the worldwide dissemination of these innovations.

20,157 Reads