The number of studies examining the relation between inflation and structural and long-term factors in Turkey is quite low. In this blog post, we try to shed light on the relation between inflation and demography via the change of weights of child and senior dependents, who have no labor income, and working age population in total population. As the weight of age groups with different economic behavior in the total population changes, so does the total production and spending levels, hence the course of inflation. We test our argument by employing fixed effect panel data models and discuss to what extent it matters for inflation.[1]
Economic theory suggests that the less the consumption levels of individuals fluctuate, the higher the prosperity of individuals becomes. Nevertheless, individuals’ labor income varies according to the age group they are in. For this reason, individuals need to make intertemporal transfers or engage in simultaneous transfers with other individuals to ensure the income- expenditure balance. These transfers may either occur in the private sector’s savings dynamics or be supported by transfers carried out by the public sector. While child dependents’ consumption is met by intra-family transfers (private) and public transfers (like the primary education services), individuals of working age save for their retirement when their current income exceeds their current expenditures. Senior dependents (the group above 65) meet their consumption needs either via the savings they accumulated during their working age and/or transfers from the savings of working population depending on the structure of the social security system applied in the country.
Employment age, retirement age, average life expectancy and wage-pension difference may vary across countries. Accordingly, consumption and income profiles by age cohorts are expected to vary across countries as well. With this perspective, the National Transfer Accounts (NTA) project estimates the consumption and labor income profiles by age for countries based on the household income survey and national income accounts (Chart 1). According to the estimates made for Turkey, labor income exceeds the current consumption of those aged between 29 and 62. Moreover, it is remarkable that the difference between labor income and consumption is more limited in Turkey compared to developed countries.
As Turkey is going through a demographic transition, the age structure of the population is changing too. The decrease in birth rates and the rise in average life expectancy contribute to the change in the age structure of the population as well. These developments bring about a radical change in the weight of population groups with different economic behaviors. For example, while there was approximately one employee matching every dependent (child-senior) in 1964, according to projections, there will be two employees matching one dependent by 2025 (Chart 2).
In this blog post, we analyze how the changes in the age structure of population affect inflation in a sample of developing countries. We limited our analysis to the period 1996-2015 by considering the high inflation experiences of these countries.[2] In the first estimation model, only demographic variables were included, while other models have a richer structure as economic variables were added. In the base model (2), the output gap and the real interest rates are used as explanatory variables. In the next model (3), real exchange rates are used as a control variable, and in the last model (4), openness, labor cost indicator and import unit value index are incorporated as control variables. We present the estimated effects of age cohorts on inflation from our models in Chart 3.
Our findings show that child dependents and those at the early working age are inflationary. The deflationary effects start to be observed in ages between 26 and 30 and continue until 56-60 years. These age cohorts exert downside pressure on inflation as they consume less than what they produce. Elderly dependents are identified as inflationary and according to estimates, the 80 and above age group is the last deflationary group.
We can make a historical accounting as well as projections for the coming period using the estimations we have obtained. The results show that the demographic structure of the countries covered by the analysis had a disinflationary effect in the sample period (Chart 4). Turkey is among the countries benefiting from this effect. We calculated the annual average disinflationary effect for Turkey over the period 1996-2015 as 0.6 percentage points. As observed, demographic developments contributed to the decline in inflation over the estimation period.
As the demographic transition matures, the rise in the share of working age population in total population slows down, and the share even decreases in some countries. In this context, one of our findings is that the disinflationary effect of demography in developing countries will decrease gradually in the upcoming years. However, the figures we calculated for Turkey indicate that this change will not reach a level that may play a decisive role on inflation dynamics in the very near future. Finally, we would like to emphasize that change in the age structure of the population is only one of the many structural determinants of inflation.
[1] This blog post comprises the initial findings of the ongoing study of Kalafatcılar and Özmen (2019). For more details, please refer to Box 3.2 in Inflation Report 2018-IV.
[2] A similar analysis was applied to advanced economies by Juselius and Takats (2015).
References
Fair, R., Dominguez, K. (1991). Effects of the Changing US Age Distribution on Macroeconomic Equations. American Economic Review, 81(5).
Juselius, M., Takats, E. (2015). Can Demography Affect Inflation and Monetary Policy? BIS Working Paper No: 485.
Kalafatcılar, M. K., Özmen, M. U. (2019). , A Glance at the Structural Components of Inflation: Demography, forthcoming study.
United Nations, (2013). National Transfer Accounts Manual: Measuring and Analysing the Generational Economy.