کنترل سرمایه بر روی جریان، نوسانات نرخ ارز و آسیب پذیری های خارجی

We use high frequency data and a new econometric approach to evaluate the effectiveness of controls on capital inflows. We focus on Chile's experience during the 1990s, and investigate whether controls on capital inflows reduced Chile's vulnerability to external shocks. We recognize that changes in the controls will affect the way in which different macro variables relate to each other. In particular, we consider the case where controls co-exist with an exchange rate band aimed at managing the nominal exchange rate. We develop a methodology to deal explicitly with the interaction between these two policies. The main findings may be summarized as follows: (a) a tightening of capital controls on inflows depreciates the exchange rate and (b), we find that a tightening of capital controls increases the unconditional volatility of the exchange rate, but makes it less sensitive to external shocks.

During the last few years the economics profession has made important progress in understanding the determinants of currency crises. This research has helped reshape the way in which monetary and fiscal policies are conducted in emerging and transition nations. Scholars and policy makers, however, continue to disagree on some important aspects of macroeconomic policy. One of the key topics of debate refers to the role of capital controls and the adequate degree of financial integration of emerging markets to the rest of the world.1 According to some authors, limiting the extent of financial integration reduces speculation, and helps countries withstand external shocks and avoid extreme exchange rate fluctuations (Bhagwati, 1998, Krugman, 1999, Stiglitz, 2000, Stiglitz, 2002 and Rodrik, 2006).2 Authors that support restricting capital mobility have mentioned Chile's experience with market-based controls on capital inflows between 1991 and 1998 as an example worthwhile emulating.3 In late 2006 Thailand's economic authorities justified the imposition of controls on short term capital inflows, by referring to Chile's experience during the 1990s.4 In 2007, Colombia imposed hort term capital inflows in an effort to reduce the extent of (nominal) exchange rate appreciation; in rationalizing this policy the authorities also referred to Chile's experience with controls on inflows.5 Authors such as Stiglitz (2002), Eichengreen (2000), Eichengreen and Hausmann (1999), Stallings (2007) and Williamson (2003) have argued that Chile-style controls on inflows have three important effects: (a) they reduce the degree of vulnerability to external shocks; (b) they result in lower exchange rate volatility; and (c) they help avoid the extent of currency appreciation during episodes of capital inflows. According to these authors, controlling short term inflows were one of the keys to Chile's economic success during the 1990s. Calvo and Mendoza (1999), however, have argued that Chile's success during the 1990s was mostly the result of favorable external conditions, including very positive terms of trade. In their view, macroeconomic policies — including the controls on inflows — had little to do with “the notable accomplishments of the Chilean economy.”6 The empirical literature on Chile's controls has tended to support Calvo and Mendoza (1999); most works on the subject have found that Chile's controls had limited macroeconomic effects. De Gregorio et al. (2000), for example, found that during the 1990s controls on inflows altered the composition of capital flows, with short term flows declining and longer term flows increasing. Controls, however, failed to stop currency appreciation or to increase the Central Bank's ability to control monetary aggregates over the medium or long run. Similar results were found by Edwards (1999) and Valdes-Prieto and Soto (1998). Forbes, 2003 and Forbes, 2005 uses firm-level data to investigate whether Chile's controls had microeconomics effects. Her results indicate that by restricting access to external funding, the controls increased the cost of capital to small and mid size firms (see, also, Ulan, 2000). Although the results reported by these early papers are useful, they are subject to some limitations and potential econometric problems. In particular, these works have ignored the fact that controls on capital inflows were only one component of Chile's external macroeconomic policy, and of the authorities' efforts to avoid “excessive” nominal exchange rate fluctuations and, in particular, currency appreciation. A second key element of this policy was a band of varying width that constrained the movement of the nominal exchange rate. Ignoring this exchange rate band can introduce an important bias in the estimation of equations that attempt to assess the effects of the controls on key macroeconomic data, such as the exchange rate (nominal or real). The reason for this is that the controls themselves affected the width and realignment of the band, and the existence of the band affected the behavior of macroeconomic variables such as interest rates and the exchange rate. The purpose of this paper is to develop a new methodology that allows us to evaluate the effects of capital controls on inflows in countries that intervene in the foreign exchange market. In particular, this new approach allows us to investigate whether restricting capital inflows will reduce nominal exchange rate changes and volatility. We do this by using a two-step estimation technique that incorporates the concept of shadow or equilibrium exchange rate developed by Bertola and Caballero (1992). In the first step, we use data on exchange rate fundamentals and on the nature of the foreign exchange rate intervention policy (or, if appropriate, the exchange rate band) to estimate the shadow exchange rate.7 In the second step, we use an augmented GARCH approach to evaluate whether changes in the restrictiveness of capital controls affected the level and volatility of the nominal exchange rate. In the empirical section we use high frequency daily data for Chile for 1991–1998; in some of the estimates, and in order to investigate the robustness of our estimates, we use monthly data. The methodology and results presented in this paper go beyond the historical interpretation of Chile's economic performance, and are useful to evaluate future initiatives aimed at restricting capital mobility in countries that pursue an active exchange rate management policy. This exchange rate intervention policy may take place through an explicit band, as in Chile, or through implicit feedback rules that rely on more implicit intervention thresholds. As pointed out above, both Thailand and Colombia recently imposed controls on inflows as a way to avoid nominal exchange rate appreciation. Our analysis differs from previous work on the subject in, at least, four respects: First, we use high frequency (daily) data to analyze the effects of controls on capital inflows on the nominal exchange rate. Previous work, in contrast, has used relatively low frequency data (monthly or quarterly) to analyze real exchange rate behavior. Second, we explicitly take into account the fact that an active exchange rate policy affects the evaluation of capital controls. All previous papers on the subject that we are aware of ignored this important fact. Indeed, one of the key objectives of introducing capital controls is to allow the monetary authority to exercise some control over exchange rates. As we explain in detail in Section 3, we do this by estimating a shadow exchange rate, which captures the response of the exchange rate to changes in fundamentals in the absence of the exchange rate band. Third, we focus on the effects of the controls on the level and volatility of the nominal exchange rate. In contrast, most previous research deals with the impact of controls on the level of the exchange rate only. And fourth, we use a two-step augmented ARCH and GARCH, while most previous analyses have relied on VARs and/or standard regressions. It is also important to clarify at the outset what our paper doesn't do: we don't provide a complete cost–benefit analysis of Chilean style capital controls. In particular, we don't deal with the potential efficiency (and other) costs of restricting capital mobility. Also, this paper doesn't deal with the effects of capital controls on the probability of a currency crisis, or their effects on interest rates and foreign debt maturities. These are important issues, but they are beyond the scope of the present paper.8 The main findings from our analysis may be summarized as follows. First, a tightening of capital controls results in a depreciation of the domestic currency. This level effect on the nominal exchange rate should have been expected, given that tighter capital controls reduce capital inflows, and cause a deterioration in the balance of payments.9 To return to equilibrium, then, an improvement in the current account is required, and hence a real exchange rate depreciation should take place; this real exchange rate change takes mostly place through changes in the nominal exchange rate. Surprisingly, most of the papers that have studied the Chilean experience have not found significant effects of the controls on the real exchange rate.10 We believe that this is because early studies on the subject ignored the endogenous response of the exchange rate to monetary policy. Second, we find that the “vulnerability” of the nominal exchange rate to external factors decreases with a tightening of the capital controls. More specifically, we find that Chile's controls on capital inflows were effective in (partially) isolating the nominal exchange rate from external shocks to import and export prices and international interest rates. Third, we find that a tightening of capital controls increases the unconditional volatility of the exchange rate. This effect can be explained by the fact that tighter controls are likely to have segmented the Chilean foreign exchange market further. On the other hand, isolating the foreign exchange market contemporaneously means that, in the end, exchange rate volatility is larger in the following periods. Capital controls introduce a tradeoff stabilizing contemporaneous exchange rates (in terms of external shocks), but destabilizing future nominal rates. The rest of the paper is organized as follows: In Section 2 we discuss the functioning of Chile's controls on inflows, and we review the empirical literature on the subject. Section 3 is the core of the paper: we present our model, and we discuss a two-stage strategy for estimating the effects of controls on inflows on the level and volatility of the exchange rate. In this Section we compare the results obtained using a shadow exchange rate and the observed exchange rate. Additionally, we present some robustness tests and we discuss issues for future research. Finally, Section 4 is the conclusions.

An important policy objective of restrictions on capital inflows has been to avoid — or at least control — nominal exchange rate changes, and to allow the central bank to have (some) policy independence. This was, for instance, a stated policy objective of Chile's renowned controls on inflows during most of the 1990s. More recently, Thailand (2006) and Colombia (2007) have put in place restrictions on inflows as a way of slowing down the appreciation of their currencies. When there is an active exchange rate management policy, it is not possible to evaluate the effectiveness of capital controls by analyzing the co-movement between the observed exchange rate and external shocks. In this case, simple estimates are likely to capture the combination of both the controls and the active exchange rate policies. Without a clear model of how exchange rate (or monetary) policy is conducted, this exercise cannot be solved, and it is likely to produce biased results and misleading policy analyses. Furthermore, it is not enough to specify a parsimonious monetary policy reaction function, because one of the purposes of capital controls is to change the stochastic properties of the fundamentals driving the exchange rate (mean, variances, vulnerabilities, and so on). Therefore, the monetary reaction function is also likely to change when controls are imposed, or when the extent of controls changes. In this paper we have attempted to disentangle the role played by capital controls from the management of the exchange rate. In doing so, we specify how monetary policy is conducted — which in the case of Chile between 1991 and 1999 is described by a target zone model based on the contribution by Bertola and Caballero (1992). This is equivalent to estimate a structural model, where the monetary policy reaction function is specified. The contributions of this paper are twofold. First, to estimate a shadow exchange rate that “cleans” the observed exchange rate by the endogenous monetary policy reaction function. This procedure takes into account that the mapping between the two changes through time, and is an explicit function of the stochastic properties of the fundamentals. Second, using the shadow exchange rate estimated in step one we are able to evaluate how the capital controls have affected the exchange rate. As we pointed out in 1 and 2, our results are quite different from those in the previous literature. More specifically, we find that a tightening of the controls on capital inflows is associated with: (a) a depreciation of the nominal exchange rate; (b) an increase in the unconditional variance of the nominal exchange rate; and (c) a reduction in the vulnerability of the nominal exchange rate to external shocks (mainly in the mean equation). Our results are important because using standard techniques it is not possible to evaluate properly if controls have been effective (see the discussion in Section 3). Our results indicate that capital controls on inflows have been — at least in Chile — more effective than what previous studies had suggested, in the sense of helping reduce the impact of external shocks on the nominal exchange rate. This, however, does not mean that capital controls on inflows played a central role in Chile's economic success during the 1990s. Indeed, we are persuaded by Calvo and Mendoza's (1999) comprehensive analysis of Chile's performance in the 1985–1998 period, and by their conclusion that macroeconomic policy played a relatively minor role. It is also important to point out, once again, what we haven't done in this paper: we have not provided a complete cost–benefit analysis of Chilean style capital controls on inflows. In particular, we have not dealt with the potential efficiency (and other) costs of restricting capital mobility. A complete policy evaluation of the controls would consider both the macroeconomic and microeconomic aspects of the policy, including their effects on the probability of crises, interest rates and debt maturities. These are important issues, but they are beyond the scope of the present paper.29