Elsevier

Journal of Monetary Economics

Volume 117, January 2021, Pages 473-488
Journal of Monetary Economics

Foreign exchange reserves as a tool for capital account management

https://doi.org/10.1016/j.jmoneco.2020.02.006Get rights and content

Highlights

  • In theoretical papers variable capital taxes are a tool for managing capital flows.

  • Empirical work shows we rarely see these variable capital controls in practice.

  • This variable tax can be equivalent to sterilized foreign exchange intervention.

  • In many countries a variable tax is a reasonable proxy for sterilized intervention.

Abstract

Recent theoretical papers argue that countries can insulate themselves from volatile world capital flows by using a variable tax on foreign capital as an instrument of monetary policy. But empirical papers argue that we rarely observe these cyclical capital flow taxes used in practice. We construct a small open economy model where the central bank engages in sterilized foreign exchange intervention. When private agents freely trade foreign bonds, sterilized intervention has no effect. But we prove that when frictions prevent the free trade in foreign bonds, optimal sterilized foreign exchange intervention is equivalent to an optimal tax on foreign capital. The model is then calibrated to match the levels of capital account restrictions that we observe in the data. For levels of capital account openness that we observe in many emerging market economies, a variable tax on capital flows is a close approximation for sterilized foreign exchange intervention.

Introduction

The Global Financial Crisis and subsequent recovery has led to renewed focus on the potential merits of capital controls as a tool for capital account management. Rey (2015) argues that the presence of a Global Financial Cycle that drives capital inflows to emerging market economies means that the “trilemma” of international finance is actually more of a “dilemma”, and that “independent monetary policies are possible if and only if the capital account is managed.”1 In 2012 the IMF argued that active capital controls might be a useful policy instrument to manage the macroeconomic and financial risks associated with large swings in capital inflows and outflows (International Monetary Fund, 2012). This has led to a renewed interest in capital controls as a tool for active capital account management. Jeanne et al. (2012) argue that capital controls “properly designed ... might even be a regular instrument of economic policy” (p. 95). Farhi and Werning (2012) introduce capital account management measures as a variable tax, τt, on the price of foreign bonds.2 In this model, and others that model capital controls in a similar way, the variable tax affects fluctuations in net capital outflows. This is a policy instrument that the policy maker can adjust in real time to achieve some policy objective like welfare maximization. Theory suggests that these variable capital controls should be cyclical and should vary with net capital flow surges and stops.

But while these variable capital taxes are intuitive and easy to implement in a model, the empirical evidence that policy makers make use of these variable capital controls is lacking. Eichengreen and Rose (2014) and Fernandez et al. (2015) find that in practice capital controls are acyclical and do not respond to booms and busts in output, the current account, or the real exchange rate.

Klein (2012) and Forbes et al. (2015) argue that temporary, episodic type capital controls, what Klein refers to as capital control “gates”, tend to be ineffective. Moreover, Klein argues that permanent capital controls, what he refers to as capital control “walls”, can be effective at deflecting capital inflows (although they may come at a high cost in terms of low growth, so while they are effective, whether they are beneficial is another story).3

We begin this paper with two simple empirical observations. First, emerging markets maintain significant capital controls, while the advanced economies have basically eliminated capital controls. These capital controls are very persistent and unchanging in emerging markets, i.e. Klein’s “wall-type” capital controls. Second, central bank foreign exchange interventions are frequently observed in emerging market economies but not in advanced economies. While central bank foreign exchange reserve accumulation tends to be small and uncorrelated with other types of capital flows in most advanced economies, reserve accumulation is a large and volatile component of the balance of payments in most emerging markets, and it is highly correlated with capital inflows from abroad.4 In simple cross-country regressions we show that countries that have greater capital controls tend to be more active in using foreign exchange intervention.

We then construct a small open economy dynamic stochastic general equilibrium (DSGE) model where the central bank can engage in sterilized foreign exchange intervention that explains these results. In the model we introduce central bank balance sheet into a model of a small open economy. The central bank potentially has two instruments, the size of its balance sheet and the composition of its balance sheet. Both are set optimally to maximize household welfare. The first instrument determines the money supply and thus the interest rate, and the second instrument determines the central bank’s stock of foreign bonds, which can (potentially) be adjusted to smooth fluctuations in net capital flows.

The effectiveness of sterilized foreign exchange intervention depends on frictions or adjustment costs which determine how easily private agents can buy or sell foreign bonds, as in Gabaix and Maggiori (2015). We show that when private agents can freely buy and sell foreign bonds, sterilized foreign exchange intervention has no effect, and optimal policy with two instruments, both the size and composition of its balance sheet, is equivalent to optimal policy where the central bank only has one instrument, the size of the balance sheet. This finding echoes that in Obstfeld (1981) and Backus and Kehoe (1989). In our contribution we prove that when private agents cannot freely buy and sell foreign bonds, optimal sterilized foreign exchange intervention is equivalent to the optimally chosen variable capital control tax from Farhi and Werning (2012). We provide analytical proofs of the two sets of equivalence results under the extreme cases of a perfectly open or a perfectly closed capital account.

We then build a fully specified model for numerical simulation of the equivalence results. The model is similar to the one in Chang et al. (2015).5 To calibrate the numerical model we map the observed levels of capital controls in the data into the adjustment cost parameter in the model that determines how easily private agents can buy or sell foreign bonds. As in Gabaix and Maggiori (2015), capital controls increase the effectiveness of foreign exchange intervention.6 We show that under levels of capital controls that are observed in many emerging market economies, sterilized foreign exchange intervention delivers results similar to those under an optimally chosen variable capital control tax. Or to put it another way, under the levels of capital controls that are observed in many emerging markets, a model with an optimally chosen variable capital control tax is a reasonable approximation to a more complex model with a central bank balance sheet and sterilized foreign exchange intervention.

The remainder of this paper is organized as follows. In Section 2 we present some simple descriptive statistics describing capital controls and central bank reserve accumulation. The model where the central bank can engage in sterilized foreign exchange intervention is presented in Section 3. The proof of the equivalence between sterilized foreign exchange intervention and a variable capital tax is presented in Section 4. The results from a numerical model are presented in Section 5. Finally Section 6 concludes.

Section snippets

Descriptive Statistics on Reserve Accumulation

The index of capital account restrictions from Chinn and Ito (2006) is presented in Fig. 1. This figure shows the simple average of the Chinn-Ito index, KAOpen, updated through 2015, for a sample of 23 advanced countries and a sample of 66 emerging market and developing countries. We re-normalize the original Chinn-Ito index to a 0–1 scale, where 0 represents no capital account restrictions and 1 represents a closed capital account.

The figure shows that on average, advanced economies have fewer

Model

In the model there are two countries, home and foreign. The home country is of size n and the foreign country is of size 1n, and we consider the case of the small open economy where n → 0. The home economy is populated by a representative household and a continuum of firms. There is a central bank which sets the domestic money supply, and under different policy scenarios may set a tax rate on foreign borrowing/lending or may alter their stock of foreign exchange reserves.

In this section, we

Proofs of two equivalence results

The Ramsey policy maker sets policy to maximize household welfare (1) subject to structural constraints that include the economy-wide budget constraint (11), the UIP condition (5), a set of first order conditions from households’ utility maximization, a set of first order conditions from firms’ profit-maximization and cost-minimization problems, and market clearing conditions.

We present here analytical proofs of our two equivalence results. First we show that, when households can freely buy and

Numerical results

The model to calculate the numerical results is a standard open-economy New Keynesian model. Details about the specific functional forms used in the numerical model, the household’s utility function and firm side production functions, are presented in the Online appendix. In addition, in the Online appendix we provide a full list of all equations in the numerical model and prove the equivalence results with a generalized Ramsey solution.

Here we discuss calibration of this numerical model and a

Conclusion

In one extreme, when private agents face restrictions that prevent them from buying and selling foreign bonds, equilibrium under optimal sterilized foreign exchange intervention is equivalent to equilibrium when agents can buy and sell foreign bonds subject to a tax τt that is chosen optimally by the central bank. In the other extreme, when private agents can freely buy and sell foreign bonds, sterilized foreign exchange intervention has no effect at all.

This explains the descriptive statistics

References (45)

  • K. Forbes et al.

    Capital flow waves: surges, stops, flight, and retrenchment

    J. Int. Econ.

    (2012)
  • A.R. Ghosh et al.

    Two targets, two instruments: monetary and exchange rate policies in emerging market economies

    J. Int. Money Finance

    (2016)
  • A.R. Ghosh et al.

    Surges

    J. Int. Econ.

    (2014)
  • P.A. Neumeyer et al.

    Business cycles in emerging economies: the role of interest rates

    J. Monet. Econ.

    (2005)
  • N. Prasad

    Sterilized interventions and capital controls

    J. Int. Money Finance

    (2018)
  • S. Schmitt-Grohe et al.

    Closing small open economy models

    J. Int. Econ.

    (2003)
  • A. Abiad et al.

    A new database of financial reforms

    IMF Staff Papers

    (2010)
  • K. Aoki et al.

    Monetary and financial policies in emerging markets

    mimeo

    (2016)
  • C. Baba et al.

    Effectiveness of capital controls in selected emerging markets in the 2000s

    IMF Working Papers WP/11/281

    (2011)
  • M. Bussere et al.

    For a few dollars more: reserves and growth in times of crises

    J. Int. Money Finance

    (2015)
  • P. Cavallino

    Capital flows and foreign exchange intervention

    Am. Econ. J.

    (2019)
  • A. Costinot et al.

    A theory of capital controls as dynamic terms-of-trade manipulation

    J. Polit. Econ.

    (2014)
  • Cited by (0)

    We thank the editor, Yuriy Gorodnichenko, the associate editor, Vivian Yue, and an anonymous referee. We have also benefited from conversations with Marvin Goodfriend, Zheng Liu, Enrique Mendoza, Andrew Rose, Mark Spiegel, Martin Uribe, Yi Wen, and seminar participants at the Federal Reserve Banks of Dallas and Boston and Osaka University, the China Meeting of the Econometric Society, and the Shanghai Macro Workshop for many helpful comments and suggestions. Fujiwara acknowledges the financial support from JSPS KAKENHI Grant-in-Aid for Scientific Research (A) Grant Number 15H01939 and 18H03638. The views presented here are those of the authors and do not necessarily represent the views of the Federal Reserve Bank of Dallas or the Federal Reserve System.

    View full text