Financial liberalization has led to financial deepening and higher growth in ...
effects of financial liberalization across different groups of countries. We.
FINANCIAL LIBERALIZATION Financial liberalization has led to financial deepening and higher growth in several countries. However, it has also led to a greater incidence of financial crises. Here, we review the empirical evidence on these dual effects of financial liberalization across different groups of countries. We then present a conceptual framework that explains why there is a trade-off between growth and incidence of crisis, and helps account for the crosscountry difference in the effects of financial liberalization. Financial liberalization (FL) refers to the deregulation of domestic financial markets and the liberalization of the capital account. The effects of FL have been a matter of some debate. In one view, it strengthens financial development and contributes to higher long-run growth. In another view, it induces excessive risk-taking, increases macroeconomic volatility and leads to more frequent crises. This article brings together these two opposing views. The data reveals that FL leads to more rapid economic growth in middle-income countries (MICs), but does not have the same effect in low-income countries (LICs). In MICs this process is not smooth, however: it takes place through booms and busts. Indeed, MICs that have experienced occasional financial crises have grown faster, on average, than nonliberalized countries with stable credit conditions. In LICs liberalization does not lead to higher growth because their financial systems are not sufficiently developed so as to permit significant increases in leverage and financial flows. [Insert Figure 1 about here] The contrasting experiences of Thailand and India illustrate these dual effects. Thailand, a liberalized economy, has experienced lending booms and crises, while India, a non-liberalized economy, has followed a slow but safe growth path (see Figure 1). In India GDP per capita grew by only 99 per cent between 1980 and 2001, whereas Thailand’s GDP per capita grew by 148 per cent, despite a major crisis. As will be shown below in a set of data analyses, this trade-off exists more generally across MICs.
-2-
Asymmetric financial opportunities across sectors are key to understanding the effects of FL. In particular, in MICs contract enforceability problems affect the tradable (T) and nontradable (N) sectors differently. Many T-sector firms are able to overcome these problems and gain access to international capital markets, whereas most N-sector firms are financially constrained and depend on domestic banks for their financing. Trade liberalization promotes faster productivity growth in the T-sector, but is of little direct help to the N-sector. By allowing banks to borrow on international capital markets, FL leads to an increase in investment by financially constrained firms, most of which are in the N-sector. However, while FL increases investment, it also increases borrowers’ incentives to take on insolvency risk because there are implicit and explicit bail-out guarantees that cover lenders against systemic defaults. This is why greater leverage and growth is associated with aggregate financial fragility and occasional crises. In the rest of this article we describe the ways in which FL has been measured and the empirical estimates of its effects on growth and crises. We then present a conceptual framework and a review of the policy issues. In a nutshell, any evaluation of FL must weigh its benefits against its costs. Focusing exclusively on the growth effects of liberalization during good times would miss the link between FL and crises. Focusing only on volatility and crises could lead to an excessive cautiousness about the risks of FL. The case for FL requires that its growth and welfare benefits outweigh the costs associated with more frequent financial crises. Measuring financial liberalization There are three classes of FL indices. First, there are de jure indices based on official dates of policy reforms. An example is the index based on the IMF Annual Report on Exchange Arrangements and Exchange Restrictions (Grilli and Milesi-Ferretti, 1995). This class of indices permits a comparison of the periods before and after liberalization. A drawback, however, is that legislated changes take time to translate into liberalization on the ground. Liberalization may even fail to materialize altogether when well-functioning domestic financial markets are absent. Bekaert, Harvey and Lundblad (BHL, 2005) overcome this problem by constructing a de jure indicator of equity market liberalization that records the
-3-
date after which foreign investors are able to invest in domestic securities. A second class of indices uses de facto measures of financial openness, like the capital flows–GDP ratio used by Edison et al. (2004). The drawback is that these measures are contaminated by cyclical fluctuations and thus are imprecise indicators for dating FL. Lastly, de facto indices identify structural breaks in the trend of capital inflows (Tornell, Westermann and Martinez 2003). These indices combine the advantages of the two previous classes as they provide more precise FL dates based on actual, rather than merely legislated, policy reforms. Financial liberalization and growth BHL find that equity market liberalization leads to an increase of one percentage point in average real per-capita GDP growth. Ranciere, Tornell and Westermann (RTW, 2005) find that capital account liberalization leads to a similar gain in growth. To illustrate the link between FL and growth we add liberalization dummies to a standard growth regression: ∆yit = λyi,ini + γXit + φ1TLit + φ2FLit + εjt,
(1)
where ∆yit is the average growth rate of GDP per capita; yi,ini is the initial level of GDP per capita; Xit is a vector of control variables that includes initial human capital, the average population growth rate, and life expectancy; and TLit and FLit are the trade and financial liberalization dummies of TMW, respectively. For each country and each variable, we construct ten-year averages starting with the period 1980–9 and rolling forward to the period 1990–9. Thus each country has up to ten data points in the time-series dimension. The liberalization dummies take values in the interval [0,1], depending on the proportion of liberalized years in a given window. We estimate the panel regressions using generalized least squares. [Insert Table 1 about here]
-4-
The FL dummy enters significantly at the five per cent level in all regressions. Regression 1-1 of Table 1 shows that following FL growth in real GDP per-capita increases by 1.5 percentage points per year, after controlling for the standard variables. Trade liberalization increases growth by 0.8 percent per year (column 1-2). When both liberalization dummies are included (column 1-3), both enter significantly. This suggests that trade and financial liberalization have independent effects and jointly contribute to higher long-run growth. [Insert Figure 2 about here] Figure 2 illustrates the link between FL and growth for individual MICs. For each country, we plot growth residuals before and after FL. Growth residuals are obtained by regressing real per-capita growth on initial income per-capita and population growth. Figure 2 shows clearly that for almost all countries growth has been higher in the financially liberalized period. Several studies find mixed evidence on the link between financial openness and growth. This can be attributed either to the indicators of openness used or to the sample considered. First, some studies include low-income countries that do not have functioning financial markets. In these countries we do not expect the financial deepening mechanism to work. One might also expect the growth effect of FL to be smaller in high-income than in middle-income countries as the latter face more severe borrowing constraints. Hence, sample heterogeneity can create a bias against finding a linear growth effect of FL. Klein (2005) finds that FL contributes to growth among MICs but not among poor or rich countries. Second, some studies test the effect of changes in the capital flows–GDP ratio on growth. However, because this index does not identify a specific liberalization date, it is not appropriate for comparing the behaviour of macroeconomic variables before and after liberalization. Furthermore, these measures tend to exhibit year-to-year fluctuations that do not reflect actual changes in the degree financial openness. Financial liberalization and crises
-5-
FL is typically followed by boom-bust cycles. During the boom, bank credit expands very rapidly and excessive credit risk is undertaken. As a result, the economy becomes financially fragile and prone to crisis. Although the likelihood that a lending boom will crash in a given year is low, many booms do eventually end in a crisis. During such a crisis, new credit falls abruptly and recuperates only gradually. The incidence of crises can be measured by analysing countries’ financial histories and by codifying the occurrence of banking crises, currency crises, and sudden stops in capital inflows. Kaminsky and Reinhart (1999) use such a crisis index in a probit model to test whether banking and currency crises are crises are more likely to occur after FL. RTW (2005) use a more parsimonious indicator of financial fragility: the negative skewness of credit growth. Negative skewness is a de facto indicator that captures the existence of infrequent, sharp and abrupt falls in credit growth. Since credit growth is relatively smooth during boom periods, and crises happen only occasionally, in financially fragile countries the distribution of credit growth rates is characterized by negative outliers in a long enough sample. These outliers correspond to the abrupt falls in credit growth that occur during the crisis or ‘bust’ stage of the boom–bust cycle. The advantages of this skewness measure, relative to other more complex indicators of crises, are that it is objective and comparable across countries. In the literature variance is the typical measure of volatility. Variance, however, is not a good instrument to identify growth-enhancing credit risk because high variance reflects not only the presence of boom–bust cycles but also the presence of high-frequency shocks. Table 2 partitions country-years into two groups: liberalized and non-liberalized. The table shows that, across MICs, the financial deepening induced by FL has not been a smooth process but has been characterized by booms and occasional busts. We can see that FL leads to an increase in the mean of credit growth of four percentage points (from 3.8 per cent to 7.8 per cent) and a fall in the skewness of credit growth from near zero to –1.08, and has only a negligible effect on the variance of credit growth. Notice that, across high-income countries, credit growth exhibits near-zero skewness, and both the mean and the variance are smaller than across MICs. This difference reflects the absence of severe credit market imperfections in high-income countries.
-6-
[Insert Table 2 about here] Growth and crises To close the circle we show that countries with a greater incidence of crises countries have grown faster than those with smooth credit paths. We do so by adding three moments of real credit growth to growth regression (1) ∆yit = λyi,ini + γXit + β1µ∆B,it + β2σ∆B,it + β3S∆B,it + φ1TLit + φ2FLit + εj,t,
(2)
where ∆yit , yi,ini, Xit, TLit, and FLit are defined as in eq. (1), and µ∆B,it, σ∆B,it, and S∆B,it are the mean, standard deviation, and skewness of the real credit growth rate, respectively. We estimate eq. (2) using the same type of overlapping panel data regression as for eq. (1). Columns 1-4 through 1-7 of Table 1 report the estimation results. Consistent with the literature, we find that, after controlling for the standard variables, the mean growth rate of credit has a positive effect on long-run GDP growth, and the variance of credit growth has a negative effect. Both variables enter significantly at the five per cent level in all regressions. The first key point is that the financial deepening that accompanies rapid GDP growth is not smooth but, rather, takes place via booms and busts. Columns 1-4 and 1-5 show that negative skewness – a bumpier growth path – is on average associated with faster GDP growth across countries with functioning financial markets. This estimate is significant at the five per cent level. To interpret the estimate of 0.27 for skewness, consider India, which has near-zero skewness, and Thailand, which has a skewness of about minus 2. A point estimate of 0.27 implies that an increase in the bumpiness index of 2 (from zero to minus 2) increases the average long-run GDP growth rate by 0.54 of a percentage point a year. Is this estimate economically meaningful? To address this question, note that, after controlling for the standard variables, Thailand grows about two percentage points faster per year than India. Thus, about a quarter of this growth differential can be attributed to credit risk taking, as measured by the skewness of credit growth.
-7-
The second key point is that the association between skewness and growth does not imply that crises are good for growth. Crises are costly. They are the price that has to be paid in order to attain faster growth in the presence of credit market imperfections. To see this, consider column 1-6. When the FL dummy is included, bumpiness enters with a negative sign (and is significant at the ten per cent level). In the MIC set, given that there is FL, the lower the incidence of crises the better. We can see the same pattern when we include high-income countries in column 1-7. Clearly, liberalization without fragility is best, but the data suggest that this combination is not available to MICs. Instead, the existence of contract enforceability problems implies that liberalization leads to higher growth because it eases financial constraints but, as a by-product, also induces financial fragility. However, because crises occur relatively rarely, FL has a positive net effect on long-run growth. A unified approach An alternative approach to understand the contrasting effects of FL is to combine the linear growth regression with a crisis probit model. In this way one can decompose the net effect of FL into a direct pro-growth effect and an indirect anti-growth effect, via a higher propensity to crises. Using this approach, RTW (2006) find that the direct effect of FL on growth is 1.2 percentage points and the indirect effect is minus 0.25 percentage points. In order to understand this result, one should keep in mind that even in financially liberalized countries crises are rare events. Therefore, even if crises have large output consequences, their estimated growth effect remains modest. In contrast, since FL is likely to improve access to external finance, it has a first-order impact on growth. Conceptual framework To analyze FL and the subsequent boom–bust cycles, consider an economy with two sectors: non-tradables (N) and tradables (T). Alternatively, one can think of ‘new-economy’ and ‘traditional’ sectors, respectively. The key is that each sector uses as input the other sector’s output.
-8-
This economy is subject to severe contract enforceability problems that generate financing constraints. While T-firms can overcome such constraints and finance themselves in bond and equity markets, most N-firms are financially constrained and bank-dependent. Since N-goods serve as intermediate inputs for both sectors, the N-sector constrains the long-run growth of the T-sector and that of GDP: there is a bottleneck. In such an economy, FL increases GDP growth by increasing the investment of financially constrained firms. However, the easing of financial constraints is associated with the undertaking of insolvency risk because FL not only lifts restrictions that preclude risk taking but also is associated with explicit and implicit systemic bail-out guarantees that cover creditors against systemic crises. It is a stylized fact that, if a critical mass of borrowers is on the brink of bankruptcy, authorities will implement policies to ensure that creditors get repaid (at least in part) and thus avoid an economic meltdown. These bail-out policies may come in the form of an easing of monetary policy in response to a financial crash, the defence of an exchange rate peg in the presence of liabilities denominated in foreign currency, or the recapitalization of the financial sector. Because domestic banks have been the prime beneficiaries of these guarantees, investors use domestic banks to channel resources to firms that cannot pledge international collateral. Thus liberalization results in biased capital inflows. T-firms and large N-firms are the recipients of foreign direct investment (FDI) and portfolio flows, whereas most of the inflows to the N-sector are intermediated through domestic banks, which enjoy bail-out guarantees. Insolvency risk often takes the form of maturity mismatch or risky debt denomination (currency mismatch). Taking on insolvency risk reduces expected debt repayments because authorities will cover part of the debt obligation in the event of a systemic crisis. Thus the guarantee allows financially constrained firms to borrow more than they could otherwise. This increase in borrowing and investment is accompanied by an increase in insolvency risk. When many firms take on insolvency risk, aggregate financial fragility arises together with increased Nsector investment and growth. Faster N-sector growth then helps the T-sector grow faster because N-sector goods are used in T-sector production. Therefore, the T-sector will enjoy
-9-
more abundant and cheaper inputs than otherwise. As a result, as long as a crisis does not occur, growth in a liberalized economy is faster than in a non-liberalized one. Of course, financial fragility implies that a self-fulfilling crisis may occur. And during crises GDP growth falls. Crises must be rare, however, in order to occur in equilibrium – otherwise agents would not find it profitable to take on credit risk in the first place. Thus, average long-run growth is greater along a risky path than along a safe one even if there are large crisis costs. This is why FL leads both to higher long-run growth and to a greater incidence of crises. Schneider and Tornell (2004) and RTW (2003) formalize the intuitive argument we described using a general equilibrium model with rational agents. This discussion of the mechanism through which FL affects the growth of MICs also explains why FL does little to improve the growth of LICs. LICs often do not have functioning financial markets and thus lack the infrastructure that allows the financial system to direct international funds to profitable firms. MICs, by contrast, have enough financial infrastructure to allocate funds reasonably well, even though contract enforceability problems prevent them from doing so as efficiently as high-income countries (HICs). Because of the imperfections in their financial systems, the price of fast growth in MICs is financial fragility. The contrasting experiences of Thailand and India during the period 1980–2001 illustrate this trade-off clearly. As we discussed earlier, Thailand experienced booms and busts while India did not. While Thailand experienced spectacular growth, India’s growth was dismal. Recently, India has opened its economy to both trade and finance. Not surprisingly, India is currently experiencing a lending boom. It will be interesting to analyse the evolution of the Indian economy around 2015. Economic policy Several observers have suggested that partial liberalization is the optimal policy to reap the growth benefits of openness without having to suffer from volatility and crises. They suggest the implementation of trade liberalization but not of FL, or the restriction of capital flows to FDI, the least volatile form of capital flows. These recommendations seem impractical. First, an open trade regime is usually sustained by an open financial regime because exporters and importers need access to international financial markets. Since capital is fungible, it is difficult
- 10 -
to insulate the financial flows associated with trade transactions. The data indicates that trade liberalization has typically been followed by FL. As Figure 3 shows, by 1999 72 per cent of countries that had liberalized trade had also liberalized financial flows, bringing the share of MICs that were financially liberalized to 69 per cent, from 25 per cent in 1980. [Insert Figure 3 about here] Second, FDI does not obviate the need for risky international bank flows. FDI goes mostly to financial institutions and large firms, which are mostly T-firms. Thus, bank flows are practically the only source of external finance for most N-firms (Tornell and Westermann, 2005). Curtailing such risky flows would reduce N-sector investment and generate bottlenecks that would limit long-run growth. Bank flows are hardly to be recommended, but for most firms it might be that or nothing. Clearly, allowing risky capital flows does not mean that anything goes. Appropriate prudential regulation must also be in place. In an environment with asymmetric financial opportunities authorities may be tempted to make direct investment subsidies to constrained sectors. The historical evidence indicates that such centrally planned policies typically fail. We now know that either authorities do not possess the appropriate information or crony capitalism and rampant corruption take over. A second-best policy is to liberalize financial markets and allow banks to be the means through which resources are channelled to financially constrained firms. Here, it is important to make a distinction between ‘systemic’ and ‘unconditional’ bail-out guarantees. The former are granted only if a critical mass of agents default. The latter are granted on an idiosyncratic basis whenever there is an individual default. We have argued that, if authorities can commit to grant only systemic guarantees, and if prudential regulation works efficiently, then FL will induce higher long-run growth in a credit-constrained economy. In contrast, if guarantees are granted on an unconditional basis or there is a lax regulatory framework, the monitoring and disciplinary role of banks in the lending process will be negated. In this case, FL will simply lead to overinvestment and corruption. One should not conclude that in order to enjoy the growth and welfare benefits of FL countries have to be exposed for ever to the risk of crises. The amelioration of contract
- 11 -
enforceability problems, through a better legal system and other institutional reforms, is a fundamental source of higher growth and lower volatility in the long-run. However, it often takes time for these reforms to be achieved. In the meantime, countries with functioning financial markets can be made better off by liberalizing and experiencing a rapid but risky growth path, rather than remaining closed and trapped in a safe but slow growth path. Romain Rancière, Aaron Tornell and Frank Westermann See also banking crises; currency crises; foreign direct investment; international capital flows Bibliography Bekaert, G., Harvey, C. and Lundblad, C. 2005. Does financial liberalization spur growth? Journal of Financial Economics 77, 3–55. Edison, H., Klein, M., Ricci, L. and Sløk, T. 2004. Capital account liberalization and economic performance: survey and synthesis. IMF Staff Papers 51, 111–55. Grilli, V. and Milesi-Ferretti, G. 1995. Economic effects and structural determinants of capital controls. Working Papers No. 95/31. Washington, DC: IMF. Kaminsky, G. and Reinhart, C. 1999. The twin crises: the causes of banking and balance-ofpayments problems. American Economic Review 89, 473–500. Klein, M. 2005. Capital account liberalization, institutional quality and economic growth: theory and evidence. Working Paper No. 11112. Cambridge, MA: NBER. Newey, W. and West, K. 1987. A simple positive semi-definite, heteroskedasticity and autocorrelation consistent covariance matrix. Econometrica 55, 703–8. Rancière, R., Tornell, A. and Westermann, F. 2003. Crises and growth: a re-evaluation. Working Paper No. 10073. Cambridge, MA: NBER Rancière, R., Tornell, A. and Westermann, F. 2005. Systemic crises and growth. Working Paper No. 11076. Cambridge, MA: NBER. Rancière, R., Tornell, A. and Westermann, F. 2006. Decomposing the effects of financial liberalization: crises vs. growth. Journal of Banking and Finance (forthcoming).
- 12 -
Schneider, M. and Tornell, A. 2004. Balance sheet effects, bailout guarantees and financial crises. Review of Economic Studies 71, 883–913. Tornell, A. and Westermann, F. 2005. Boom-Bust Cycles and Financial Liberalization. Cambridge, MA: MIT Press. Tornell, A., Westermann, F. and Martinez, L. 2003. Liberalization, growth and financial crisis: lessons from Mexico and the developing world. Brookings Papers on Economic Activity 2003(2), 1–112. Index terms bail-out guarantees banking crises boom–bust cycles capital account liberalization contract enforceability credit growth currency crises economic growth equity market liberalization financial liberalization financial openness measurement foreign direct investment India insolvency risk international capital flows investment investment subsidies lending booms portfolio flows probit models prudential regulation
- 13 -
Thailand tradable and non-tradable commodities trade liberalization
