Economics Letters 123 (2014) 183–186

Contents lists available at ScienceDirect

Economics Letters

journal homepage: www.elsevier.com/locate/ecolet

Age structure and the current accountGudmundur S. Gudmundsson a, Gylfi Zoega b,c,∗

a Department of Economics and Business, Universitat Pompeu Fabra, Ramon Trias Fargas 25-27, 08005 Barcelona, Spainb Department of Economics, University of Iceland, Saemundargata 2, 101 Reykjavik, Icelandc Birkbeck College, University of London, Malet Street, London WC1E7HX, United Kingdom

h i g h l i g h t s

• The current account balance depends on demographics.• The young and the old reduce savings and the current account balance.• China’s current account would be much smaller were it not for its demographics.• The surplus of Japan would be greater were it not for demographics.• The surplus of Germany would be even greater were it not for demographics.

a r t i c l e i n f o

Article history:Received 15 November 2013Received in revised form28 January 2014Accepted 2 February 2014Available online 12 February 2014

JEL classification:J1E2

Keywords:Current accountAge structureLife-cycle behavior

a b s t r a c t

We adjust current account surpluses and deficits of 57 countries in the period 2005–2009 for differencesin the age structure of their populations and find that these differences can account for a significant partof the variation in the data. Among the large countrieswe find that the adjustment increases the surplusesof Germany and Japanwhile the surpluses of China, Singapore, Hong Kong, Korea, Thailand, Indonesia andMalaysia are significantly diminished.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

In this paper we explore the extent to which observed currentaccount imbalances between countries in recent years can be at-tributed to differences in the population age structure and towhichextent they can be attributed to other factors such as economic pol-icy. Our hypothesis that savings behavior reflects age structure isbased on the life-cycle hypothesis (see Modigliani, 1975). Thus anation largely at work should have a larger current account sur-plus comparedwith anotherwith proportionatelymore young andold people and the resulting imbalances do not necessarily indicateeconomic mismanagement.

∗ Corresponding author at: Department of Economics, University of Iceland,Saemundargata 2, 101 Reykjavik, Iceland. Tel.: +354 525 5239; fax: +354 552 1331.

E-mail address: gz@hi.is (G. Zoega).

http://dx.doi.org/10.1016/j.econlet.2014.02.0010165-1765/© 2014 Elsevier B.V. All rights reserved.

2. Literature

Leff (1969) found a relationship between dependency rates andsavings rates in a cross section of a large sample of countries.Later, Fry and Mason (1982) and Mason (1987, 1988) rekindledinterest in the relationship by finding empirical support for anegative relationship between the youth dependency rate andsavings rates. Taylor and Williamson (1994) explained the capitalflow from Britain to Australia, Canada, the US and Argentina inthe late 19th century by high youth dependency ratios in theNew World. Higgins (1998) found, in a sample of 100 countries,that an increase in both the youth- and old-age dependencyratios is associated with lower savings and larger current accountdeficits. Taylor (1995) used demographic trends to predict that thegrowing elderly dependency ratios in the developed world andthe falling youth dependency ratios in the developing world werelikely to create a pattern of current account surpluses in someof the developing countries in the first part of the 21st century.

184 G.S. Gudmundsson, G. Zoega / Economics Letters 123 (2014) 183–186

Fig. 1. The average age structure in the top 10 surplus and deficit countries.

Lindh (1999) found a positive relationship between the share oftheworking-age population and savings forOECD countries. Bloomand Williamson (1998) attributed East Asia’s economic miracle toits working age population growing at a much faster rate thanits dependent population between 1965 and 1990. Herbertssonand Zoega (1999) detected a relationship between age structureand the current account in a sample of 86 countries. Finally, Laneand Milesi-Ferretti (2012) find that more rapidly aging advancedeconomies run more positive current account balances while ahigh old-age dependency ratio lowers the balance for developingcountries.

3. Age structure and the current account

Our data consist of twelve population variables, measuring theproportion of population in each of twelve age groups, for 57 coun-tries over the period 1980–2009.1 Fig. 1 shows the age structure ofthe working age population in the top ten current account deficitand surplus countries, averaging over the period 2005–2009.

We note that the 35–49 age group is larger in the surplus coun-tries than in the deficit countrieswhile the share of population over50 is higher in the deficit countries.

Fig. 2 shows the correlations between the share of the popu-lation in each of the twelve age groups and the current accountsurplus as a percentage of GDP over the period 1995–2009. Thecorrelation is positive for ages 30–55 but negative for both youngerand older groups. This implies that the larger the share of the popu-lation in the younger and the older age groups and the smaller theshare of the middle-age group, the larger is the current accountdeficit. This is what the life-cycle theory would suggest.

4. Regressions

To further analyze the relationship between age structure andthe current account we regress the current account surplus in per-centages of GDP on three composite population variables, mea-sured in percentages of the total population, in addition to thegrowth rate of output per capita and an interaction term betweenthe youngest age group and growth, following Fry and Mason(1982).2 We use data on all 57 countries from 1980–2009 for theregressions. The three age groups are 0–24, 25–64 and 65+.

1 The countries are: Argentina, Australia, Austria, Belgium, Brazil, Bulgaria,Canada, Chile, China, Colombia, Croatia, Czech Republic, Denmark, Estonia, Finland,France, Germany, Greece, Hong Kong, Hungary, Iceland, India, Indonesia, Ireland,Israel, Italy, Japan, Korea, Rep., Latvia, Lithuania, Luxembourg, Malaysia, Mexico,Morocco Netherlands, New Zealand, Norway, Pakistan, Peru, Philippines, Poland,Portugal, Romania, Russia, Saudi Arabia, Singapore, Slovak Republic, Slovenia SouthAfrica, Spain, Sweden, Switzerland, Thailand, Turkey, Ukraine, the United Kingdomand the United States.2 Economic growth is included because the younger cohorts enjoy higher

permanent incomes and higher consumption when labor productivity growth ispositive.

Fig. 2. The correlation between the current account and population shares(1995–2009).

Table 1The current account and the age structure.

Pooled OLS Fixed effects

(1) (2) (3)

Young −0.550***−0.607***

−0.615***

(0–24) (0.142) (0.168) (0.188)Old (65+) −1.136***

−1.271***−1.306***

(0.335) (0.461) (0.488)Growth per capita −35.292 −14.889

(30.262) (29.232)Growth∗ young 0.610 0.227

(0.613) (0.655)Constant 33.921***

(9.389)Average fixed effect 37.387*** 38.216***

(11.212) (12.367)

R2 0.119 0.451 0.469F-statistic 51.15 21.509 21.425Observations 1518 1578 1518

Standard errors in parentheses.*** Denotes significance at the 1% level.

Two different estimators are used: The within-groups estima-tor, also known as the fixed-effects estimator, and the pooled OLSestimator. When the relationship is estimated with fixed effects,one of the population variables has to be omitted to prevent per-fect multicollinearity. For this reason, the middle group variablewas omitted. For consistency, we also omit the same variable inthe pooled OLS regression. Themodels were estimatedwith robuststandard errors clustered by country.

Column (1) of Table 1 presents the results of estimating therelationshipwith the pooled OLS estimator. The results are roughlyconsistent with the graphical representation in Fig. 2. In particular,the current account surplus is lower the larger is the share of the65+ and the 0–24 age groups. The coefficients of the two growthterms are statistically insignificant. In columns (2) and (3) werepeat the estimation with the fixed-effects estimator, both withand without the growth terms. The fixed effects estimator yieldsvery similar results, which indicates that the coefficient estimatesin column (1) are not merely picking up the effect of omitted time-invariant country effects.3 Because the demographic variables canbe considered predetermined with respect to contemporaneous

3 We also tried a specification with dummy variables for countries exposedto financial crises, oil producing countries, GDP per capita and governmentsurpluses. Their inclusion did not significantly affect the estimated coefficients ofthe demographic variables and the last two turned out to be insignificant.Moreover,we found no structural breaks coinciding with European integration.

G.S. Gudmundsson, G. Zoega / Economics Letters 123 (2014) 183–186 185

Fig. 3. Average adjusted and unadjusted current account surpluses (2005–2009).

economic variables, we rule out endogeneity of the populationstructure.

5. Age-adjusted imbalances

We subtract the age effect from the observed current accountvalues using the estimated coefficients from column (2) in Table 1and then subtract the average value of the fixed effects. This trans-formation changes the position of many countries significantly asshown in Fig. 3 where the horizontal axis has the unadjusted val-ues and the vertical axis has the adjusted current account. Obser-vations above the 45° line indicate countries that have a greateradjusted surplus or smaller adjusted deficits.

There are some notable departures from the 45° line, indicat-ing that demographic differences explain part of the global imbal-ances. Thus the adjusted surpluses of Japan, Germany and Swedenare larger than the actual ones while the surpluses of South Korea,China, Hong Kong and Singapore are significantly smaller.

In Table 2we show the actual current account surplus (% ofGDP)for 2005–2009, the correction – a positive number implies that thecurrent account surplus (deficit) of the country is larger (smaller)once the age distributionhas been taken into account – and the cor-rected surplus.4 Among surplus countries, the surpluses of Japan,Germany, Sweden and Belgium are increased while the surplusesof Singapore, Hong Kong, Korea, China, Thailand and Saudi Arabiaare significantly decreased. The adjustment did not have a big ef-fect on the ranking of the highest deficit countries. Spain’s deficitincreases slightly and the same applies to the United States.

China’s surplus of 7.80% of GDP in 2005–2009 is reduced by 4.4%by the adjustment. In contrast, Germany’s surplus of 6.18% is in-creased by 3.14%. This is explained by the fact that from 1990–2009, the 0–24 group shrunk by 12.9% in China due to its one-childpolicy while it has decreased by only 5.1% in Germany. Reinforcingthe effect on the current account, the share of the 65+ has grownby 2.1% in China and by 5.3% in Germany.5 In Fig. 4 we show the

4 Note that Thailand, Korea, Canada and Indonesia are surplus countries in termsof actual surpluses but deficit countries when it comes to the corrected surplusesin Table 2. Similarly, the deficit countries Italy and France turn out to be surpluscountries after the correction.5 See Table A.1 in the Appendix.

Fig. 4. Actual and age-adjusted current account balances of China, Germany andthe USA.

evolution of the actual and adjusted current account for China, Ger-many and the United States, three countries that have played a bigrole in the generation of the pattern of global imbalances.

6. Conclusions

More than a half of the Chinese surpluses can be explained bythe age structure of the nation, largely a consequence of China’sone-child policy. Similarly, the current account surpluses of Sin-gapore, Hong Kong, Korea, Thailand, Indonesia and Malaysia aremuch diminished when the age structure is taken into account.In contrast, the age-adjusted surpluses of Germany and Japan arehigher than the actual ones. These surpluses can possibly be tracedto policies, such as Germany’s membership of the Euro zone. Bigdeficit countries, such as Spain and the US, remain big deficit coun-tries after the age adjustment.

Acknowledgment

Financial support from the Icelandic Centre for Research isgratefully acknowledged.

186 G.S. Gudmundsson, G. Zoega / Economics Letters 123 (2014) 183–186

Table 2Actual and corrected current account surpluses 2005–2009 (% of GDP).

Surplus countries Deficit countriesActual surpluses Correction Corrected surpluses Actual surpluses Correction Corrected surpluses

Saudi Arabia 22.78 −3.05 Saudi Arabia 19.74 Iceland −19.54 −0.41 Iceland −19.96Singapore 20.36 −7.17 Malaysia 14.88 Bulgaria −17.32 0.73 Bulgaria −16.59Malaysia 16.37 −1.48 Norway 14.81 Latvia −12.42 2.43 Latvia −9.99Norway 14.33 0.47 Singapore 13.19 Greece −11.73 1.80 Romania −9.94Hong Kong 11.61 −5.73 Sweden 11.12 Portugal −10.93 1.21 Greece −9.93Switzerland 10.20 −0.08 Switzerland 10.13 Romania −9.65 −0.29 Portugal −9.72Luxembourg 8.71 −1.24 Germany 9.32 Estonia −9.45 2.41 Slovak rep. −9.50Sweden 8.05 3.06 Japan 7.60 Spain −8.23 −0.31 Spain −8.54China 7.80 −4.45 Luxembourg 7.47 Lithuania −8.15 1.51 New Zealand −7.21Russia 7.36 −1.43 Russia 5.93 New Zealand −7.14 −0.06 Estonia −7.04Netherlands 6.40 −0.69 Hong Kong 5.87 Croatia −6.63 1.48 Turkey −6.95Germany 6.18 3.14 Netherlands 5.71 Slovak rep. −6.29 −3.21 Lithuania −6.64Philippines 3.74 1.33 Philippines 5.07 Hungary −5.95 −0.10 Hungary −6.05Japan 3.69 3.91 Israel 4.83 South Africa −5.39 −0.65 South Africa −6.04Finland 3.23 1.27 Finland 4.50 Australia −5.16 −0.48 Poland −6.03Austria 3.21 0.80 Austria 4.01 Turkey −4.90 −2.05 Ireland −6.00Israel 3.02 1.81 Argentina 3.95 US −4.88 −0.51 Australia −5.64Denmark 2.96 0.64 Denmark 3.60 Pakistan −4.86 3.36 US −5.39Thailand 2.45 −4.09 China 3.35 Poland −4.61 −1.42 Croatia −5.15Argentina 2.36 1.59 Belgium 2.50 Ireland −4.18 −1.82 Czech rep. −4.95Korea 2.01 −5.22 France 1.41 Slovenia −3.43 −1.02 Slovenia −4.45Chile 1.81 −1.65 Italy 0.91 Czech rep. −2.41 −2.54 Colombia −3.42Indonesia 1.38 −2.18 Peru 0.90 UK −2.23 1.89 Korea −3.21Belgium 0.43 2.08 Chile 0.16 Colombia −2.21 −1.22 Brazil −1.84Peru 0.37 0.53 Ukraine −2.16 0.83 India −1.83Canada 0.30 −1.73 Italy −1.71 2.62 Morocco −1.71

India −1.51 −0.32 Thailand −1.64Morocco −1.36 −0.35 Pakistan −1.50France −1.06 2.47 Canada −1.43Mexico −0.81 −0.14 Ukraine −1.33Brazil −0.05 −1.79 Mexico −0.96

Indonesia −0.80UK −0.34

Appendix

Table A.1Changes in population shares and the current account (calculated as the differencebetween the last value and the first).

Young Middle Old CA

1990–2009

China −12.9% 10.8% 2.1% 2.2%Germany −5.1% −0.2% 5.3% 2.7%USA −2.3% 1.9% 0.4% −1.3%

2000–2009

China −4.1% 3.0% 1.0% 3.5%Germany −1.9% −1.9% 3.9% 7.4%USA −1.2% 0.7% 0.5% 1.5%

Data sourcesIMF: World Economic Outlookhttp://www.imf.org/external/pubs/ft/weo/2012/02/weodata/

index.aspx.World Bank:http://data.worldbank.org/data-catalog/health-nutrition-and-

population-statistics.

References

Bloom, D.E., Williamson, Jeffrey G., 1998. Demographic transitions and economicmiracles in emerging Asia. World Bank Econ. Rev. 12 (3), 419–455.

Fry, M., Mason, A., 1982. The variable rate-of-growth effect in the life-cycle model.Econ. Inquiry 20, 232–233.

Herbertsson, T.T., Zoega, G., 1999. Trade surpluses and life-cycle saving behavior.Econom. Lett. 65, 227–237.

Higgins, M., 1998. Demography, national savings, and international capital flows.Internat. Econom. Rev. 39 (2), 343–369.

International Monetary Fund, ‘‘World Economic Outlook (WEO)’’, 2012. http://www.imf.org/external/pubs/ft/weo/2012/02/weodata/index.aspx.

Lane, Philip R.,Milesi-Ferretti, GianMaria, 2012. External adjustment and the globalcrisis. J. Int. Econ. 88, 252–265.

Leff, N.H., 1969. Dependency rates and saving rates. Amer. Econ. Rev. 59 (5),886–896.

Lindh, T., 1999. Age structure and economic policy: the case of saving and growth.Popul. Res. Policy Rev. 18, 261–277.

Mason, A., 1987. National savings rates and population growth: a new model andnew evidence. In: Johnson, D., Lee, R. (Eds.), Population Growth and EconomicDevelopment: Issues and Evidence. Wisconsin University Press, pp. 523–560.

Mason, A., 1988. Saving, economic growth and demographic change. Popul. Dev.Rev. 14, 113–144.

Modigliani, F., 1975. The life-cycle hypothesis of saving twenty years later.In: Parkin, M. (Ed.), Contemporary Issues in Economics. Manchester UniversityPress.

Taylor, A.M., 1995. Debt, dependence and the demographic transition: LatinAmerica into the next century. World Dev. 23, 869–879.

Taylor, A., Williamson, J., 1994. Capital flows to the new world as an intergenera-tional transfer. J. Polit. Econ. 102, 348–369.