Journal Issue

Research Summaries: Population Aging and International Capital Flows

International Monetary Fund. Research Dept.
Published Date:
September 2005
  • ShareShare
Show Summary Details

Robin Brooks

The population of the world is aging. The United Nations projects that the global old age dependency ratio—the number of people 65 and older relative to those between the ages of 15 and 64—will rise from the current 11 percent to 25 percent by 2050. How will this trend affect saving and investment rates around the world? What are the implications for international capital flows? To answer these important questions, this article reviews recent IMF research on how the global population shift will affect international capital markets.

Empirical work suggests that there is a systematic association between demographics and saving and investment. Brooks (1998) uses data for industrial countries to establish that saving rates are negatively related to youth and old age dependency. Chinn and Prasad (2003) use cross-section and panel regressions to investigate the medium-term determinants of current accounts around the world using data for 18 industrial and 71 developing countries. They find that youth and old age dependency are negatively associated with current account balances.

The IMF (2004) confirms these results, based on the empirical relationship between demographics and saving and investment for 115 countries from 1960 to 2000, and explores how projected population trends may affect current account positions going forward. It finds that in advanced countries, the aging of populations will result in deteriorating current account balances. For Japan, the effect could be around 2.5 percent of GDP by 2050. For Europe, it would be smaller, at less than 0.5 percent of GDP over the same period. The major exception is the United States, where it is predicted that demographic effects will boost the current account by more than 1 percent of GDP. Elsewhere, demographic change could contribute to an improvement in Africa (close to 3 percent) and the Middle East (around 0.5 percent), but a deterioration in central and eastern Europe (around 1 percent) and emerging Asia (less than 0.5 percent). Heller and Symansky (1997), surveying the existing literature, forecast how the aging of populations will affect the “Asian tiger” countries. These countries will contribute to world saving as their populations move into prime saving years. Beyond 2025, however, they will drag down the global saving rate, as their populations begin to dissave in retirement.

The empirical analysis discussed above, however, is constrained by the fact that an important equilibrium condition—that saving must equal investment at the global level—is difficult to impose. In addition, historical correlations may not reflect causality. To address these issues, Brooks (2003) simulates the effects of population change on external balances in an overlapping generations model with eight regions. The model explicitly incorporates the population age structure and assumes that households accumulate wealth as they work to finance consumption in retirement. Capital is assumed to be perfectly mobile across borders. The simulations suggest that the European Union and North America will run large current account surpluses around 2010, exporting capital to Latin America and Africa where population growth will be faster. Around 2030, dissaving by baby boomers switches the current accounts in the EU and North America to deficit. Both regions will in effect be repatriating foreign assets from regions such as Latin America and the rest of the world.

The main result of the simulations—that current accounts in industrial countries will increasingly move into surplus before switching into deficit as the baby boomer generation retires—is remarkably robust. Faruqee (2002a)—using a multiregion model that has a very different specification for population dynamics that extends Faruqee (2002b)—shows that the current account of an industrial country like Japan will gradually swing into deficit as an aging population consumes the net foreign assets it accumulated earlier. Studies that focus on individual countries, typically treating them as small and open economies with the world interest rate exogenous, also provide supportive evidence. Cheng (2003) builds an overlapping generations model for China and finds that low fertility rates will imply future capital outflows if capital mobility is high. If capital is less mobile, low fertility today will lower the domestic return to capital and raise the domestic return to labor. Brooks (2004) builds an overlapping generations model for a small and open economy to assess how much of Singapore’s present current account surplus can be explained via demographic factors. He finds that up to half of Singapore’s current account surplus could be due to demographic factors alone and that the surplus will decline in the years ahead as the population ages.

Of course, the model-based literature is subject to several caveats. First, it typically assumes perfect capital mobility and perfect foresight, ignoring the presence of capital account restrictions and political risk in developing countries. As a result, the magnitude of flows to and from developing countries is likely overstated. Another common assumption is that labor is not mobile, which again tends to overstate the role of capital flows.

Second, the magnitude of current account swings depends significantly on how governments respond to the aging of their populations. Schimmelpfennig (2000) argues that tax financing, which is equivalent to prefunding, is the best path for pension reform in a small and open economy with a weak current account position. This is because debt financing leaves the current account unchanged, while tax financing improves the current account. Faruqee and Mühleisen (2001) argue that public investment cuts, measures to broaden the base for income taxes, some increase in the consumption tax, and reductions in social security benefits are the best way to promote orderly fiscal and current account adjustment in Japan. In addition, the multiregion model in IMF (2004) shows that a reduction in the replacement rate of pay-as-you-go pension systems in Europe would boost the European current account substantially relative to the baseline scenario, as it forces households to save more for retirement, which spills over into increased net foreign asset accumulation.

Third, and perhaps most importantly, these models generally fail to explain the current constellation of external balances. Notably, they tend to suggest that the US current account should be moving into surplus, when in fact there is a widening deficit. This is in part due to omitted factors, such as monetary and fiscal policies, but it also reflects the deeper problem, which is that, at best, these frameworks provide an incomplete description of actual saving behavior.


Visiting Scholars, April–June, 2005

Adeola Adenikinju; University of Ibadan, Nigeria; 2/22/05–4/1/05

Marco Aiolfi; University of California, San Diego; 6/15/05–6/24/05

Joshua Aizenman; University of California, Santa Cruz; 3/28/05–4/1/05, 4/25/05–4/29/05

Lloyd Amaghionyeodiwe; University of Ibadan, Nigeria; 2/28/05–4/8/05

Michael Atingi-Ego; Central Bank of Uganda; 3/21/05–4/29/05

Jushan Bai; New York University; 4/18/05–4/22/05

Gian Luca Benigno; London School of Economics; 4/4/05–4/15/05

Sami Bennaceur; IHEC, Carthage, Tunisia; 2/28/05–4/1/05

Fernando Broner; Universitat Pompeu Fabra, Spain; 3/21/05–4/1/05

Anetta Caplanova; University of Economics, Bratislava, Slovakia; 3/14/05–4/15/05

James Cassing; University of Pittsburgh; 4/11/05–4/13/05, 4/25/05–4/29/05

Luis Cespedes; Banco Central de Chile; 3/21/05–4/1/05

Oluwatoyin Chete; Nigerian Institute of Social and Economic Research; 5/2/05–6/10/05

Ehsan Choudhri; Carleton University, Canada; 4/4/05–4/8/05

Jeffrey Chwieroth; Syracuse University; 5/16/05–7/15/05

Mario J. Crucini; Vanderbilt University; 4/12/05–4/14/05

Betty Daniel; University of Albany; 4/27/05–4/29/05

Allan Drazen; University of Maryland; 4/1/05–4/29/05

Todd Gormley; Massachusetts Institute of Technology; 4/4/05–4/8/05

Pierre-Olivier Gourinchas; Princeton University; 6/6/05–6/17/05

Stephen Haber; Stanford University; 6/27/05–7/1/05

Franz Hamann; Bank of Colombia; 6/8/05–6/21/05

Jiandong Ju; University of Oklahoma; 4/25/05–4/29/05

Michel Juillard; CEPREMAP, France; 4/18/05–4/25/05

Arvind Krishnamurthy; Northwestern University; 5/18/05–5/20/05

Kenneth Kuttner; Oberlin College; 2/25/05–4/29/05, 6/2/05–6/3/05, 6/8/05–6/10/05, 6/13/05–7/29/05

Philip Lane; Trinity College Dublin; 4/21/05–4/29/05

Chia-Hui Lu; Academia Sinica; 4/15/05–7/15/05

Jane Mariara; University of Nairobi, Kenya; 3/21/05–4/29/05

Enrique Mendoza; University of Maryland; 4/11/05–4/22/05, 6/6/05–6/6/05, 6/9/05–6/10/05

Enrico Minelli; CORE, Belgium; 4/5/05–4/6/05

Alessandro Missale; Università di Milano; 3/30/05–4/8/05

Peter J. Neary; University College, Dublin; 4/25/05–4/29/05

Christopher Otrok; University of Virginia; 4/14/05–4/29/05, 5/11/05–5/13/05, 5/23/05–5/27/05, 6/7/05–6/8/05, 6/16/05–6/17/05

Sandra Poncet; University of Clermont-Ferrant and University of Paris XIII; 5/9/05–5/13/05

Jay Shambaugh; Dartmouth College; 4/4/05–4/8/05

Jeremy Stein; Harvard University; 6/15/05–6/17/05

Phillip Swagel; American Enterprise Institute; 3/21/05–4/29/05

Allan Timmermann; University of California, San Diego; 6/15/05–6/21/05

Eric van Wincoop; University of Virginia; 5/9/05–5/13/05, 6/6/05–6/10/05

Thierry Verdier; Delta, France; 4/4/05–4/15/05, 4/25/05–4/26/05

Randall Wright; University of Pennsylvania; 4/21/05–4/22/05, 4/28/05–4/28/05

Pierre Yared; Massachusetts Institute of Technology; 6/7/05–6/10/05

Other Resources Citing This Publication