Developed countries are ageing fast. Although longer and healthier lives imply huge welfare gains for individuals, the effects on economic growth have been elusive. This column uses a structural model to estimate the contributions of demographic change to the historical and projected future economic growth in the four largest European economies. The model projects that the demographic headwinds for economic growth will be large in the coming decades. Several reforms have been suggested to increase growth. The welfare heterogeneity of these reforms across age, income, and wealth groups provides some insights into the opposition to their implementation.
One of the most striking characteristics of advanced economies has been the secular rise in life expectancy. During the last 50 years, life expectancy at birth in advanced economies has increased by almost ten years, and it is expected to continue increasing according to UN projections (United Nations 2024). This remarkable increase in longevity combined with a decrease in fertility has resulted in shifts in the cohort distribution and an increased average age of the populations of most developed economies. Longevity gains are, in isolation, almost unambiguously good for growth and welfare. Changes to the cohort distributions, which mean smaller young and larger old cohorts, have, in contrast, been associated with lower growth. The sum of these has been seen as a possible macroeconomic problem with adverse implications for growth and public finances (e.g. Baldwin and Teulings 2014, Bloom 2019, Kuhn and Prettner 2023).
In a forthcoming article (Cooley et al. 2024), we provide a structural framework to give quantitative estimates of the contribution of demographic change to economic output. This framework is applied to the impact of changing demographics on aggregate growth in Europe’s G4 economies: France, Germany, Italy, and the UK. Since the early 1990s, these four economies have experienced a reduction in long-run growth rates that has been persistent but not uniform. Compared to the prior two decades, annualised long-run growth over the last 20 years fell by between 0.9 percentage points in France and 2.2 percentage points in Italy. At the same time, these countries have experienced persistent declines in both mortality and fertility rates, resulting in increases in individual life expectancy and average population age.
For all these economies, we disentangle the growth effects of demographic change into an individual decision effect (i.e. the growth effect due to gains in life expectancy) and an aggregating composition effect (i.e. the growth effect from shifts in the age-cohort distribution). At the individual level, increases in life expectancy affect consumption, labour supply, and savings decisions as households must adapt to a longer expected life span. Our estimated individual effect is positive along all margins considered, meaning that gains to life expectancy in isolation contribute positively to growth. The composition effect impacts the aggregation of individual decisions. We find that the composition effect more than offsets the individual-level labour supply effects and amplifies the individual-level savings decisions in all four countries considered.
Late-life labour supply
Put together with demographic data, labour statistics exhibit another striking feature. During the last 50 years, while there have been considerable gains in life expectancy, on average, the pattern of hours worked as a function of age and the age of leaving the labour force have hardly changed. Loosely speaking, longer life spans have almost exclusively gone into longer time in retirement and hardly at all to longer working lives.
We don’t have the data to firmly conclude to what extent these choices are due to regulations and incentives from pensions and social security schemes and to what extent they are due to individual preferences for reducing hours and leaving the labour force at a given age. Since most advanced economies are also highly functional democracies, regulations, pensions, and social security schemes should, over time, reflect aggregate preferences.
An important requirement for any structural model is, therefore, both to allow for higher disutility of late-life labour supply and to calibrate the parameters of this disutility function such that the model respects and accounts for the choices that historically have been made.
Demographic contributions to historical and future growth
Table 1 shows the model’s result for demographic contribution to historical growth. When estimating the contribution of demographic change to historical growth, the calibrated parameters of the model are fixed, and conditional survival probabilities and cohort distributions are the only exogenously changing variables. With the model’s endogenously generated macroeconomic time series, we then perform growth accounting between 1975 and 1995 and between 1995 and 2015. As we see from this table, a cautiously conservatively calibrated structural model predicted an average reduction of annual GDP per capita growth rates of 0.64 percentage points just from demographic change.
Table 1 Estimated demographic contribution to historical growth
Note: The two first columns show the benchmark model’s estimated demographic contribution to average annual GDP growth rates in the two periods 1975-1995 and 1995-2015, respectively. The next column is the percentage-point difference in estimated growth rates between these two periods, and the last column is the model’s estimated change in growth rates relative to the historical slowdown.
Table 2 shows the estimated future demographic contributions to growth. To compute these predictions for future growth, all structural parameters are kept fixed. The only exogenously changed parameters are future survival probabilities and cohort distributions, which are given by demographers’ projections. The reported projected growth rates between 2020 and 2040 are then obtained by growth accounting using the model’s endogenous macroeconomic time series between these two dates. As we see from this table, the demographic headwinds to economic growth will be large. Demographic factors will reduce annual per capita GDP growth by between 0.35% and 1.07% every year.
Table 2 Estimated demographic contribution to future growth, 2020-2040
Note: The two first columns show the benchmark model’s estimated demographic contribution to average annual GDP growth rates in the two periods 1995-2015 and 2020-2030, respectively. The next column is the percentage-point difference in estimated growth rates between these two periods.
Policy focus: Late-life labour supply
The analysis points to the fact that the focus of policy to counter the effects of demographics on economic growth and governments’ fiscal situation should be whether it could be welfare-improving to further stimulate late-life labour supply. In the counter-factual situation where people, on average, had decided to split gains to longevity equally into longer working lives and longer retirement, ageing would have increased growth and improved governments’ fiscal situation.
Sound policies not related to late-life labour supply and savings should be evaluated and, possibly, implemented without taking ageing into account. Policies that increase prime-age labour market participation could potentially increase both economic output and welfare. Likewise, policies that stimulate human capital accumulation and education (Bloom et al. 2024a, 2024b). Other measures to increase TFP growth, such as skill accumulation through college attainment, would potentially increase growth (Conesa et al. 2020). These measures could all be evaluated and implemented irrespective of ageing.
Ageing is primarily caused by longer and, to a large extent, healthier lives. The challenge is how those additional years are spent (e.g. Neumark et al. 2019, Acemoglu et al. 2022a, 2022b).
The political economy of pension reforms
In many countries, labour market reforms are proposed in response to longer life spans and increasing average age. The growth projections presented above support the urgency of analysing reforms that may increase growth and welfare. With this motivation, we analyse the welfare implications of two potential policy reforms: a 10% reduction in benefits and a five-year increase in the eligibility age for retirement.
Even though these reforms increase economic growth and the welfare gains are sizeable for young households, the welfare effects are heterogeneous along the age, earnings, and wealth dimensions. Figure 1 shows the welfare gains and losses for different age and wealth levels. Support for reforms decreases uniformly by age, and if we assume the same propensity to vote for all age groups, these reforms will reduce welfare for about two-thirds of the electorate. These results further highlight the importance of understanding labour supply decisions, life-cycle (dis-)incentives to work, and evaluating pensions and other reforms to increase growth under a welfare measure.
Figure 1 Welfare gains and losses from pension reforms across age and wealth groups
Conclusion
Structural estimates indicate that demographics will cause massive headwinds for economic growth and put further pressure on government budgets. While demographic change has previously contributed positively to per capita economic growth, in the coming decades, it will reduce the growth rate of the European G4 economies by 0.3 to 1 percentage point per year. However, in terms of welfare, the utility gains from longer lifespans almost surely outweigh the adverse effects on per capita growth and fiscal budgets. When evaluating potential reforms to increase growth and improve governments’ fiscal situation, the welfare gains from higher growth should be weighed against the disutility of longer working lives. Further identifying and measuring the fundamental sources of the disutility of late-life labour supply may inform better policy decisions.
Source : VOXeu
