Labour force ageing and productivity growth

ABSTRACT

This paper reassesses the empirical relationship between ageing and total factor productivity (TFP) growth at the country level. We contribute to the literature by (i) extending the sample coverage to 2014; (ii) adding a new identification strategy to instrument labour force ageing both in advanced (AEs) and emerging market economies (EMEs); and (iii) refining the empirical test with a focus on the age structure of employed workers. Our econometric evidence indicates that ageing has played a significant role in slowing down TFP growth both for AEs and EMEs over the recent decades, and may remain a drag on productivity growth in the years to come.

KEYWORDS:

• Ageing
• total factor productivity
• economic growth
• employed workers

JEL CLASSIFICATION:

• D24
• O40
• J14
• J11
• C33

Acknowledgments

I would like to thank Gustavo Adler, Shekhar Aiyar, Oya Celasun, Romain Duval, Christian Ebeke, John Fernald, Davide Furceri, Bertrand Gruss, Zsoka Koczan, Ksenia Koloskova, Gian Maria Milesi-Ferretti, Weicheng Lian, Malhar Nabar, Maurice Obstfeld, Pierre-Daniel Sarte, Xiaobo Shao, and Petia Topalova for helpful comments and suggestions. Sung Eun Jung provided excellent research assistance. The opinions expressed herein are those of the author and do not necessarily reflect those of the International Monetary Fund.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ See related work by Jaimovich and Siu (2009), and Wong (2007).

² Compared to Aiyar, Ebeke, and Shao (2016), we (i) analyze the effects of changes in (instead of the level of) older worker labour force participation and in old workers employed on productivity growth; and (ii) extend their analysis to EMEs.

³ The raw dataset initially covers 202 countries between 1960 and 2014. See also Adler et al. (2017).

⁴ They are obtained by dividing the share of young (0–14 years old) and old (65+ yearsold) population to the active population.

⁵ The results (available upon request) are qualitatively robust to excluding time fixed-effects or using 15-year averages (as in Acemoglu and Restrepo 2017) since 1955, particularly for the OECD sample. We further differ methodologically from those authors by (i) focusing on TFP rather than GDP per capita; (ii) using a more restrictive definition of ageing based on the share of employed workers; (iii) allowing for dynamic effects on this relationship by using five-year panel (with time fixed-effects) – instead of cross-section; and (iv) instrumenting the latter using lagged demographic characteristics.

⁶ The solid dots are constructed with the estimates from columns (4) and (8) of Table 2 and with the difference in the average share of older labour force in the 1990s (13.8% of the total labour force) minus the average share of older labour force in the 2000s (10.6% of the total labour force).