Year: 2013
Type: article
Abstract: Previous articleNext article FreePart I: Exchange RatesCapital Account Policies and the Real Exchange RateOlivier JeanneOlivier JeanneJohns Hopkins University and NBER Search for more articles by this author Full TextPDF Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinked InRedditEmailQR Code SectionsMoreI. IntroductionThere are debates about the extent to which emerging market and developing countries that have accumulated large amounts of foreign exchange reserves in the 2000s are doing so in order to undervalue their currency. However, we do not have a simple model of how a country can achieve persistent real exchange rate distortions through reserve accumulation. The main purpose of this paper is to present such a model and to use it to answer a few questions about undervaluation policies.Real exchange rate undervaluation is often presented, in policy debates, as the result of a monetary operation. For example, it is argued that the People's Bank of China (PBOC) has resisted the appreciation of the renminbi by pegging the nominal exchange rate and accumulating reserves. But pegging the nominal exchange rate is not the same thing as pegging the real exchange rate, and we know that in an environment without nominal frictions monetary policy has virtually no impact on real variables. It is unlikely, furthermore, that nominal frictions alone give monetary policy enough leverage to have a persistent impact on the real exchange rate. Standard estimates suggest that nominal stickiness is not persistent enough to induce large and persistent deviations of the real exchange rate from its flexible-price equilibrium value (Rogoff 1996; Chari, Kehoe, and McGrattan 2002). Thus, in order to achieve persistent real undervaluation, monetary policy must rely on something else than just nominal stickiness.I focus in this paper on the role of imperfect capital mobility. Imperfect capital mobility can be defined, for the purpose of my analysis, as any friction inducing a deviation from Ricardian equivalence in capital flows.1 Imperfect capital mobility could result from "natural causes," such as financial frictions that prevent the private sector from borrowing abroad, or deviations from rational expectations that mitigate or delay the private sector's Ricardian response to reserve accumulation. Imperfect capital mobility could also be policy-induced and result from capital account restrictions that are imposed by the government. The fact that the country that has accumulated the most reserves in the 2000s, China, also imposes tight restrictions on its capital account suggests that the link between the two is worth looking at. Thus, this paper will focus on the question of how the real exchange rate is affected by capital account policies, defined in a broad way as the accumulation of foreign assets and liabilities by the public sector plus all the policies that affect the private sector's access to foreign capital. However, most of my results can be extended to the case where Ricardian equivalence fails because of frictions other than capital account restrictions.In order to simplify and streamline the analysis, I use a model that is entirely real-there is no money and no monetary policy. I consider a small open economy that consumes a tradable good and a nontradable good. The government accumulates foreign assets and imposes controls on capital flows. This combination of policies allows the government to effectively control the level of net foreign assets for the country as a whole. The other properties of the model then follow in a straightforward way. The government controls the current account balance (since it is the change in net foreign assets) and therefore the trade balance. The real exchange rate, then, has to be consistent with the trade balance. Other things equal, accumulating more net foreign assets will depreciate the real exchange rate.I then use the model to look at several questions related to capital account policies and real exchange rates. How can we detect in the data that these policies influence the real exchange rate? Are there limits to the impact of capital account policies on real exchange rates and how are they determined? If capital account policies can lead to real exchange rate undervaluation, how different are they from trade protectionism? What is the welfare cost from resisting to currency appreciation? I also look at the recent experience of China through the lenses of the model.The paper is related to several lines of literature. First, it is related to the literature on global imbalances, the "global savings glut," and the "upstream" flow of capital from relatively poor high-growth countries to relatively rich low-growth countries. The problem in that literature is to explain high saving rates in emerging market economies. One line of explanation is precautionary savings against idiosyncratic risk (see, e.g., Mendoza, Quadrini, and Ríos-Rull 2009; Carroll and Jeanne 2009; or Sandri 2010). Chamon and Prasad (2010) argue that precautionary savings against idiosyncratic risk is the most likely cause of the high saving rate in China. Precautionary savings could also be against aggregate risk, in particular the risk of sudden stop (Durdu, Mendoza, and Terrones 2009; Jeanne and Rancière 2011). Capital outflows from high-growth countries could also result from domestic financial frictions as in Caballero, Farhi, and Gourinchas (2008) or Song, Storesletten, and Zilibotti (2011). A common feature of these contributions is that the saving rate is determined by the behavior of the private sector. Reserve accumulation and capital account policies play no role and it is by hap-penstance that a substantial share of foreign assets ends up being accumulated as reserves.The evidence, however, suggests that the upstream flow in capital is linked to public flows and in particular reserve accumulation (Aguiar and Amador 2011; Gourinchas and Jeanne 2011). My model explains the link between reserve accumulation and net capital flows as more than a coincidence. In equilibrium, reserve accumulation must reduce net capital inflows by reducing saving (keeping investment constant). Another way of looking at the real undervaluation policy in my model, thus, is that the accumulation of foreign assets induces "forced saving" in the domestic economy. The capital controls prevent the domestic private sector from offsetting the public accumulation of foreign assets by borrowing abroad. The model thus provides a simple explanation for the high saving rate in countries such as China.2Second, the paper is related to the literature on exchange rate undervaluation. Dooley, Folkerts-Landau, and Garber (2004) argue that China and several other emerging markets and developing countries have been resisting the appreciation of their currencies in order to promote exports-led growth, a phenomenon that they dub "Bretton Woods II." An empirical literature has studied whether real exchange rate undervaluations increase growth (see, e.g., Rodrik 2008). Policy discussions often take for granted that a country can resist the real appreciation of its currency by accumulating reserves but the literature lacks a clear model of how this comes about. In a related contribution developed independently of this paper, Ghosh and Kim (2012) show the equivalence between capital account restrictions and trade restrictions in a two-period small open economy model.Third, the paper is related to the literature about optimal capital account policies. One recent line of literature has studied the normative case for prudential capital controls aimed at smoothing the boom-bust cycle in capital flows (Bianchi 2011; Korinek 2011; Schmitt-Grohé and Uribe 2012). Another line of literature has studied the case for "mercantilist" real exchange rate undervaluations (Aizenman and Lee 2007; Korinek and Serven 2010). Costinot, Lorenzoni, and Werning (2011) study equilibrium capital account policies in a two-country model. By contrast with that literature, I take capital account policies as given and do not look at the reasons that real exchange rate undervaluation might be desirable from a welfare perspective.Fourth, the paper is a contribution to the literature on the impact of sterilized foreign exchange reserve interventions. The empirical literature until the 2000s was primarily focused on advanced economies and motivated in part by the apparent success of concerted interventions following the 1985 Plaza Accord (see Sarno and Taylor 2001 for a review). The focus of the attention has shifted more recently on how the sterilized accumulation of reserves can help emerging markets and developing countries deal with large capital inflows and resist the appreciation of their currency (see, e.g., Disyatat and Galati 2007; Adler and Tovar 2011). Adler and Tovar (2011) examine whether the impact of sterilized foreign exchange interventions for a panel of 15 economies (mostly in Latin America) covering 2004 to 2010. They find that interventions slow the pace of appreciation, but that (consistent with the model presented here) this effect is stronger for countries with more closed capital accounts. On the theoretical side, the Ricardian irrelevance result for sterilized interventions was stated by Sargent and Smith (1988), and Backus and Kehoe (1989) showed that it holds under more general conditions. However, it is possible to design realistic stochastic environments with incomplete markets in which sterilized interventions have real effects (Kumhof 2010). In the deterministic model presented here, sterilized interventions matter because of a simple policy-induced friction in international capital flows.The paper is structured as follows. Section II motivates the model by looking at the capital account policies of China. Section III presents the model. Sections IV through VII present various properties of the model and Section VIII goes back to the Chinese experience, this time examining it through the lenses of the model.II. Capital Account Policies of ChinaThe capital account is very restricted in China. On the side of inflows, foreign direct investment (FDI) is largely liberalized and even encouraged in some cases through tax incentives, but other inflows are constrained. Inward FDI in manufacturing is almost completely liberalized, with the exception of restrictions in some strategic sectors.3As for financial inflows, the financial assets that foreigners might want to invest in are equity, debt securities, and bank deposits, but their access to these assets is severely restricted. These assets are not scarce. Figure 1 reports the outstanding stocks of the different types of financial assets as shares of GDP. At $3,408 billion at the end of 2011, the Chinese stock market capitalization is significant even relative to that in advanced economies.4 The market for debt securities is less developed. The market for government bonds is not very large (about 17 percent of GDP at the end of 2010) and the local market for corporate bonds remains small and dominated by a handful of large state-owned institutions.Fig. 1. Outstanding stocks of Chinese financial assets: Bank deposits, bonds, and stock market (% of GDP, 2000–2010) Source: People's Bank of China, China Securities Regulatory Commission, Shanghai and Shenzhen Stock Exchanges.View Large ImageDownload PowerPointThe main vehicle for households' and firms' financial saving is bank deposits, which amounted to about 140 percent of GDP on average in the 2000s, and have been increasing over time. Most of those deposits are time and saving deposits that bear an interest rate.The access of foreign investors to Chinese financial assets is severely limited. For equity, two types of shares are traded in the Shanghai and Shengzhen stock markets-"A shares" that can be owned only by domestic investors, and "B shares" that can be purchased by foreigners. The value of B shares has never exceeded 3 percent of total stock market capitalization since 2000. Foreign investors can invest in financial assets other than B shares through the Qualified Foreign Institutional Investor (QFII) program. This program allows about one hundred selected foreign institutional investors to invest in a limited range of Chinese domestic financial assets. The overall quota allocated to this program has remained small and the range of investable assets limited (Lardy and Douglass 2011; Cappiello and Ferrucci 2008).5 Foreign investors cannot otherwise invest in domestic debt securities or hold bank deposits.Capital outflows are restricted too. The restrictions on outbound FDI were relaxed over the past decade as authorities have begun to view it as a valuable way to secure commodities and further integrate China into the global trading system. But Chinese investors cannot, as a rule, hold foreign financial assets. The Qualified Domestic Institutional Investor (QDII) program, introduced in 2006, allows selected domestic financial institutions to invest abroad using a structure similar to that of QFIIS, but the quota allocated to this program has remained small. The state-controlled policy banks do the bulk of China's external lending, often to accompany the FDI of state-owned enterprises.Obviously, the Chinese capital controls are not perfectly tight and there have been leakages. Chinese banks can draw down their overseas claims on international banks and the corporate and household sector can take advantage of the more liberalized current account through leads and lags in trade payments and remittances. However, the large and persistent spread between the onshore yield on the renminbi and its offshore counterpart suggest that China's existing official capital controls on inflows have been binding, especially after 2002 (Ma and McCauley 2008; Cappiello and Ferrucci 2008). Furthermore, the composition of China's capital flows and external assets and liabilities reflects the constraints imposed by the capital account restrictions. As shown in figure 2, which reports the breakdown of Chinese foreign assets and liabilities at the end of 2010, most of the foreign liabilities are accounted for by FDI and most of the foreign assets take the form of foreign exchange reserves.Fig. 2. Composition of Chinese foreign assets and liabilities (%, 2010) Source: State Administration of Foreign Exchange (SAFE).View Large ImageDownload PowerPointIt should be noted that inward FDI is not completely liberalized as it is subject to authorizations from the Chinese authorities. In principle, thus, the Chinese authorities can influence the level of FDI inflows. However, it is unlikely that this influence is used for macroeconomic fine-tuning as most of the decision-making power regarding the screening and approval of FDI is held by local governments. This being said, even if it takes FDI inflows as given, the central government can still influence total net capital inflows through reserve accumulation as long as a change in international reserves is not offset one-for-one by a change in FDI inflows. The Chinese authorities could in this way indirectly control the current account balance-a key feature of the model presented in the next section.III. ModelThe model aims at capturing in the most simple possible way the essential features of the Chinese capital account documented in the previous section. The model is deterministic and in continuous time. It features a small open economy populated by an infinitely-lived representative consumer who consumes a tradable good and a nontradable good. The utility of the representative consumer is given by where ct = c(cTt, cNt) is a function of the consumption of the tradable good, cT, and the consumption of the nontradable good, cN, which is homogeneous of degree 1. I denote by pt the price of the nontradable good in terms of the tradable good, and by qt the price of the tradable good in terms of domestic consumption. I will call qt the real exchange rate (an increase in q is a real depreciation) and by an abuse of language that is common in the literature, I will sometimes call the tradable good "dollar."The domestic consumer receives exogenous flows of nontradable and tradable goods. The budget constraint of the domestic consumer is where at and at* are the consumer's holdings of bonds, respectively, denominated in consumption good and tradable good; yTt and yNt are the country's endowments of the tradable good and nontradable good; and zt is a lump-sum transfer from the government. I will call at and at* the private sector's holdings of "domestic bonds" and "foreign bonds," respectively. The assumption that the output of tradable good and non-tradable good are endowments can be interpreted as the fact that labor is not mobile between the two sectors. I will assume, to simplify the analysis, that the consumer's psychological discount rate is equal to the interest rate, ρ = r*, but this assumption can easily be relaxed.The budget constraint of the government is where bt* and bt are the government's holdings of bonds denominated in dollars and in domestic consumption good, respectively. I call bt* "international reserves." If the government accumulates foreign assets by issuing domestic liabilities, bt is negative and -bt is the government's domestic debt.Government policy consists in the announcement of paths for public assets, (bt*, bt), that satisfy the transversality condition,The impact of government policy crucially depends on the extent of capital mobility between the country and the rest of the world. With perfect capital mobility, government policy has no effect on the domestic economy and the real exchange rate (Ricardian equivalence). This result is well known, but going through the proof will allow us to make some points that are useful for future reference.First, adding the budget constraints for the representative consumer and the government, (2) and (3), using interest parity rt = r* + qt/qt as well as the fact that the consumption of nontradable good is equal to its supply in each period, cNt = yNt, one derives the consolidated budget constraint for the country as a whole,where nt* denotes the country's net foreign assets,Second, using the first-order condition, qt = ∂ct/∂cTt, and again cNt = yNt, the real exchange rate can be written in reduced form as a function of cTt and yNt,The equilibrium under perfect capital mobility is then characterized by the following two conditions,The first equation says that the marginal utility of consuming the tradable good must be constant over time (since the dollar interest rate is equal to the consumer's psychological discount rate). The second equation is the country's intertemporal budget constraint. Together, these conditions pin down the path for the consumption of tradable good, (cTt)t≥0, and through the country's budget constraint (5), the path for the country's total net foreign assets, (nt*)t≥0, but they do not determine the individual components of foreign assets.6 In particular, an open market operation in which the government purchases reserves by issuing domestic debt has no impact on the domestic economy. This is clear if the government makes the transaction with foreign investors, since in this case nothing changes for the domestic private sector. This is also true if the government's debt is not traded internationally and must be sold to the domestic private sector. In this case, the domestic private sector simply finances the purchase of domestic government debt by selling foreign assets (or issuing foreign liabilities) to foreign investors. Government policy is irrelevant if the domestic private sector is connected to the international financial market through the frictionless trade of one asset or liability.7The situation is quite different if the access of the domestic private sector to foreign borrowing and lending is restricted. To simplify, let us consider the extreme case where the government is the only agent in the economy that can enter into financial relationships with the rest of the world Denoting by (a closed capital account).8 Let us assume that domestic government debt can be held only by the domestic private sector (at + bt = 0) and that foreign assets can be held only by the government (at* = 0). This assumption is meant to capture Chinese-style capital account policies in which the access of foreign investors to domestic financial assets and the access of domestic private investors to foreign assets are very limited. Then, the country's net foreign assets are equal to its reserves and its consolidated budget constraint can be written,By setting the path for reserves, , the government completely determines the paths for the consumption of the tradable good, , and for the trade balance, . It also determines the path for the real exchange rate, .This result is, as a matter of accounting, obvious. If the government can determine the country's total net foreign assets, then it can also determine the current account balance (the change in the country's net foreign assets) and the trade balance (the change in the country's net foreign assets minus the return on these assets). In particular, the government can induce "forced saving" in the domestic economy by forcing the private sector to buy domestic debt and by using the proceeds to buy foreign assets. With a closed capital account, the domestic private sector cannot undo this operation by selling assets to-or borrowing from-the rest of the world. Denoting by cT(q, yN) the level of tradable good consumption when the real exchange rate is q, we have the following result.Proposition 1. With a closed capital account, the government can implement any real exchange rate path, (qt)t≥0, satisfying the country's external budget constraintProof. See discussion above.Inequality (8) is binding if the stock of reserves satisfy the transversality condition as an equality,But the left-hand side of (9) could be strictly positive if the government lets the rest of the world play a Ponzi game with domestic reserves (which is equivalent to "throwing away" a fraction of the reserves, as in Korinek and Serven 2010).A realistic application of the model is the case where the government uses capital account policies to resist a real exchange rate appreciation resulting from a takeoff in the tradable good sector. One can capture this case in the model by assuming that the endowment of tradable good, yTt, increases over time. This could lead to a trade deficit if the consumption of tradable good, reflecting the anticipation of higher future tradable income, exceeds the endowment.I will assume that (for a reason outside of the model), the government tries to smooth the trade balance by limiting trade deficits, or even maintaining a surplus. This is possible if the government closes the capital account and accumulates reserves. More formally, I will define an episode of "resistance to real exchange rate appreciation" as follows.First, I denote with tilde the values of the variables in the undistorted equilibrium (with free capital mobility). For example, is the path for the consumption of tradable good when the domestic consumer has unrestricted access to foreign borrowing and lending, and is the path for foreign exchange reserves that is consistent with the undistorted equilibrium (assuming that reserves are the only foreign assets). An episode of resistance to appreciation is when the government depreciates the real exchange relative to the undistorted level by purchasing reserves.Definition 2. There is resistance to real exchange rate appreciation between time 0 and time t if:• The government closes the capital account between time 0 and time t.• The government accumulates more reserves than in the undistorted equilibrium while the capital account is closed: for 0 < s ≤ t.• The initial real exchange rate is depreciated relative to its undistorted value: .The difference is a measure of the initial real exchange rate undervaluation. Note that the resistance to appreciation is assumed to last a finite time t, after which there is free capital mobility and Ricardian equivalence applies. After time t, the economy follows its undistorted path conditional on the init
DOI: 10.1086/669583
Author: Olivier Jeanne
Institution: Johns Hopkins University
Language: English
Subfield: Finance
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W3122849034