Revisiting the Link Between Finance and Macroeconomic Volatility
Author: Ms. Era Dabla-Norris and Mr. Narapong Srivisal1
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

Contributor Notes

Author’s E-Mail Address: edablanorris@imf.org; etihuu@gmail.com
Publication Date:
30 Jan 2013
eISBN:
9781475543988
ISBN:
9781475543988
Language:
English
Keywords:
WP; yes; external shock; consumption volatility; GDP; terms of trade volatility; investment growth; open economy; investment decision; GDP volatility
Download PDF (1.4 MB)

This paper examines the impact of financial depth on macroeconomic volatility using a dynamic panel analysis for 110 advanced and developing countries. We find that financial depth plays a significant role in dampening the volatility of output, consumption, and investment growth, but only up to a certain point. At very high levels, such as those observed in many advanced economies, financial depth amplifies consumption and investment volatility. We also find strong evidence that deeper financial systems serve as shock absorbers, mitigating the negative effects of real external shocks on macroeconomic volatility. This smoothing effect is particularly pronounced for consumption volatility in environments of high exposure - when trade and financial openness are high - suggesting significant gains from further financial deepening in developing countries.

Abstract

This paper examines the impact of financial depth on macroeconomic volatility using a dynamic panel analysis for 110 advanced and developing countries. We find that financial depth plays a significant role in dampening the volatility of output, consumption, and investment growth, but only up to a certain point. At very high levels, such as those observed in many advanced economies, financial depth amplifies consumption and investment volatility. We also find strong evidence that deeper financial systems serve as shock absorbers, mitigating the negative effects of real external shocks on macroeconomic volatility. This smoothing effect is particularly pronounced for consumption volatility in environments of high exposure - when trade and financial openness are high - suggesting significant gains from further financial deepening in developing countries.

I. Introduction

The global financial crisis has re-ignited the policy debate on the role of finance in propagating and dampening macroeconomic fluctuations. Shallow financial systems in low-income and emerging market countries imply insufficient instruments for households, enterprises, and governments to diversify risk, manage the volatility of their income streams, and insure against unexpected events. Given their higher vulnerability to external shocks, and inherently greater volatility, these countries seem to have the most to gain from financial deepening. On the other hand, larger financial systems may also indicate higher leverage on the part of economic agents, which implies more risk and lower stability. Indeed, it is argued that the excessive size of financial systems in many advanced economies was a causal factor behind the global crisis (Smaghi, 2010).

This paper aims to shed light on these issues. Do deeper financial systems dampen output, consumption and investment volatility? Are there non-linearities in this relationship, with financial depth amplifying macroeconomic volatility at very high levels? Does financial depth reduce the effect of real external shocks on volatility, and what role do structural characteristics related to trade and financial openness play in this regard? Answers to these questions have important welfare implications, especially in light of the fact that macroeconomic volatility is not only a major source of business cycle uncertainty but also a determinant of growth performance.1

Economic theory and evidence at the industry, firm, and household levels suggest a number of channels through which financial depth can affect volatility. Yet, in contrast to the finance-growth literature (Levine et al., 2005), there is little robust empirical evidence at the aggregate level of a causal link between finance and volatility. A number of empirical studies have attempted to examine whether financial depth reduces macroeconomic volatility using a variety of approaches. The results, however, appear to be sensitive to the measures of financial development considered, the sets of controls, aggregation periods, country samples, and the estimation techniques employed.

Our paper contributes to this literature by rigorously examining the link between financial depth and volatility using a dynamic panel analysis for 110 advanced and developing countries over the years 1974-2008. We find strong support of a beneficial role of financial depth (proxied by the level of private credit extended by banks and other institutions as a share of GDP) in dampening the volatility of output, consumption, and investment across countries, but only up to a certain point. At very high levels (over 100 percent of GDP), well above those observed in most developing countries, we find that financial depth magnifies consumption and investment volatility.

We find strong evidence that deeper financial systems serve as shock absorbers, mitigating the negative effects of real external shocks on macroeconomic volatility. When country-specific external shocks are interacted with its level of financial depth, the results prove to be robust and highly significant. Moreover, this effect is more pronounced in smoothing consumption and investment volatility, which tends to be relatively high in developing countries. We also find that the volatility dampening role of finance is related to the degree of openness of an economy. In particular, financial depth helps reduce the impact of terms of trade, external demand, and commodity price shocks on consumption volatility particularly in environments of high exposure—that is, when trade and financial openness are high.

The results appear to hold intact against a variety of standard robustness tests, including allowing for alternative control variables, countries’ income levels, and considering different time periods. Alternative measures of financial depth and shocks are also considered and to address the problem of endogeneity of financial depth, we use techniques within the GMM methodology. All these tests contribute to making us confident that the empirical results are indeed robust and capture the causality from financial depth to macroeconomic volatility.

The remainder of the paper is organized as follows. Section II provides a review of the literature. Section III describes the data and methodology. Section IV discusses the main findings, while Section V presents a battery of robustness checks. Section VI concludes.

II. Theoretical and Empirical Evidence

The theoretical literature outlines various mechanisms through which financial depth can affect macroeconomic volatility, which motivates our analysis. Financial frictions and the underlying agency and informational asymmetries can play a role in propagating real sector shocks through the credit channel (i.e., conditions that limit the availability of funds or increase the cost of funds needed to make production and investment decisions). In particular, shocks to the net worth of non-financial borrowers in the presence of credit market imperfections limit the country’s ability to reallocate resources, amplifying macroeconomic fluctuations and contributing to their persistence (Bernanke and Gertler, 1990; Kiyotaki and Moore, 1997; Greenwald and Stiglitz, 1991). While these papers do not explicitly model the level of financial development as a source of instability, an implication is that financial institutions and intermediaries that mitigate or even overcome these frictions can dampen aggregate volatility.

A related literature shows that deeper financial systems relax borrowing constraints and promotes risk sharing, thereby enhancing the economy’s ability to absorb shocks. Deeper financial systems can dampen volatility by alleviating firms’ cash constraints, particularly in economies with tight international financial constraints (Caballero and Krishnamurty, 2001). Aghion et al. (1999) show that in the presence of financial market imperfections and unequal access to investment opportunities, economies with poorly developed financial systems tend to be more volatile, as both the demand for and supply conditions for credit tend to be more cyclical. Deeper financial systems may also facilitate greater diversification, reducing risk and dampening fluctuations (Acemoglu and Zilibotti, 1997).

At the same time, financial development can lead to more risk-taking by entrepreneurs and banks or facilitate over-leverage, both of which could potentially drive up volatility (Shliefer and Vishny, 2010; Wagner, 2010). Several recent papers show that while financial intermediaries and institutions can mitigate frictions, propagation and amplification mechanisms within the financial sector and from the financial sector to the real economy can exacerbate volatility (see Brunnermeier et al., 2012; and Quadrini, 2011, for surveys of recent models).

A vast body of empirical literature has examined the relationship between financial depth and volatility. One strand of literature to which our paper is related finds that deeper financial systems stabilize intra-sectoral output and induce an inter-sectoral reallocation of output away from sectors with a large contribution to aggregate volatility. Braun and Larrain (2005) and Raddatz (2006) use data on sectoral value added in a large cross-sections of countries, and find that financial development lowers output volatility, more so in financially vulnerable industries. As long as industrial shares and the correlations of sectoral output remain constant, these results imply a reduction in overall volatility. Manganelli and Popov (2012) find that financial development affects the speed with which an economy’s actual industrial composition converges to its optimally diversified benchmark level.2

Our paper is more closely related to cross-country empirical studies that examine the link between finance and volatility. Using panel data for 60 countries, Easterly et al. (2000) find a U-shaped relationship between volatility and financial sector depth. Their point estimates suggest that output volatility starts increasing when credit to the private sector reaches 100 percent of GDP. Using a similar methodology, but different controls and aggregation periods, Denizer et al. (2002) generally support a negative correlation between financial depth and growth, consumption, and investment volatility. However, they do not find private sector credit as a fraction of GDP to be a significant determinant of macroeconomic volatility. Moreover, Acemoglu, et al. (2003) and Beck et al. (2006) do not find a robust relation between financial intermediary development and growth volatility.

Finally, our paper is related to a large literature that examines the structural determinants of macroeconomic volatility and the role of financial development in mediating the impact. Kose et al. (2003) find that trade and financial openness is positively correlated with output volatility suggesting that more open economies are more vulnerable to external shocks. Rodrik (1998) provides evidence that higher income and consumption volatility is strongly associated with exposure to external risk, proxied by the interaction of overall trade openness and terms of trade volatility. Popov (2011) finds that countries with deeper domestic credit markets benefit more from financial liberalization in terms of lower tail risk. Beck et al. (2006), find weak evidence that financial intermediaries dampen the effect of terms of trade shocks on output volatility, while Loayza and Raddatz (2007) find that domestic financial depth plays a role in stabilizing more open economies.3

III. Methodology and Data

Our approach in this paper relies on a dynamic panel regression framework. While this has some limitations, it enables us to provide a broad-brush characterization of the effects of financial depth on volatility at the macroeconomic level. One potential problem we have to contend with is that of reverse causality—the possibility that macroeconomic volatility itself hinders financial development, and the related problem of endogeneity—macroeconomic volatility and financial depth could both be responding to some other unobserved forces.

We tackle the endogeneity issue, in the presence of unobserved country fixed effects, by using the system GMM dynamic panel data estimator developed in Arellano and Bond (1991), Arellano and Bover (1995) and Blundell and Bond (1998) and we compute robust two-step standard errors by following the methodology proposed by Windmeijer (2004). This approach addresses the issues of joint endogeneity of all explanatory variables in a dynamic formulation and of potential biases induced by country-specific effects.

We estimate the following baseline regression:

Vi,t=αVi,t−1+γ1FDi,t+βXi,t+ut+μi+εi,t(1)

where V is a measure of macroeconomic volatility at time t for country I; FD is a measure of financial depth; X is a set of other control variables; ut is the time-specific effect (for each non-overlapping five-year period); μi is the country-specific effect; and εi,t is the error term.

As is now standard in the literature, a panel dataset is constructed by transforming the time series data into five-year averages. Our panel comprises data over the period 1974–2008 for 110 countries (34 high-income, and 76 middle- and low-income countries). Although the potential instability associated with a large financial sector is central to our argument, we exclude the recent crisis from the sample period in order to draw more general conclusions. We use data from the IMF’s International Financial Statistics and World Economic Outlook, and supplement this with data from various other sources, including databases maintained by the World Bank. The panel of country and time-period observations is unbalanced. Appendix I presents the list of countries included in the sample.

To evaluate the empirical predictions advanced by the variety of theoretical models on the relationship between financial depth and macroeconomic volatility, we would ideally like to construct measures of the ability of financial systems to ameliorate information asymmetries, ease risk management, and facilitate resource mobilization. Financial depth, however, is a multidimensional concept and, as such, difficult to quantify, particularly for a broad cross-section of countries over the past four decades. Following King and Levine (1993), we measure financial depth by the aggregate private credit provided by deposit money banks and other financial institutions as a share of GDP. To assess the robustness of our results, we also consider three alternative measures, including total liquid liabilities, depository banks’ assets, and total deposits in financial institutions, all measured as a share of GDP.

Following Denizer et al. (2002), we distinguish between overall macroeconomic volatility (volatility in real per-capita output growth) and sectoral volatility (volatility of real per-capita final and private consumption and total investment growth). Such a sectoral disaggregation can reveal whether financial depth has different influences on the household and business sectors. Our benchmark measure of volatility is the standard deviation of the per-capita growth rate of the relevant variable. To assess the robustness of our results, we also consider the deviation of the per-capita growth rate of output, consumption, and investment from its Hodrick-Prescott filter trend.

To study the responsiveness of macroeconomic volatility to real external shocks, we consider two measures of real external shocks, including shocks to the country-specific terms of trade and external demand. External demand shocks are measured as the standard deviation of partner country growth. As a robustness test, we also consider country-specific commodity price shocks, using the approach outlined in Aghion et al. (2010).4 Financial crisis is a dummy variable indicating the frequency of a banking crisis within each five-year interval, taken from Laeven and Valencia (2010).

As controls we include a number of variables drawn from papers that have examined various aspects of volatility. The set includes beginning-of-period real GDP per capita to control for economic size, as smaller economies may concentrate more on particular industries, resulting in higher macroeconomic volatility. We also include the standard deviation of real exchange rates changes to control for macroeconomic stability, as exchange rate changes can affect production and consumption decisions. Other control variables, measured as averages within each 5-year period, include the ratio of government balance to GDP, the inflation rate, and growth rates of real per-capita GDP, consumption or investment. Structural variables to capture exposure to shocks include trade openness, as measured by the ratio of exports plus imports to GDP, and financial openness, measured by gross inflows as in Kose et al. (2009). Finally, we include an index of the type of political regime (Polity index) which captures the institutionalized qualities of governing authority, and may have a bearing on economic stability. Definition and sources for all variables are given in Appendix II.

Tables 1-2 present the summary statistics. Simple bivariate regressions show a negative correlation between financial depth and volatility. Figure 1 plots private credit as a share of GDP against real per capita GDP, consumption, and investment growth volatilities over the horizon covered in dataset. The linear fitted lines and 95-percent confidence interval band are also displayed in the figures, suggesting that countries with higher levels of financial system depth experience less volatility.

Table 1.

Summary Statistics

(Five-year panel)

article image
Table 2.

Correlation Matrix

(Five-year panel correlations between variables reported in Table 1)

article image
Figure 1.
Figure 1.

Financial Depth and Macroeconomic Volatility, 1974-2008

(with linear fitted lines and 95-percent confidence interval bands)

Citation: IMF Working Papers 2013, 029; 10.5089/9781475543988.001.A001

Figure 2.
Figure 2.

Marginal Effect of Private Credit on Final Consumption Volatility

Citation: IMF Working Papers 2013, 029; 10.5089/9781475543988.001.A001

Note: This figure shows the projected effect using the coefficients on both the ratio of private credit to GDP and its square in the final consumption volatility regression.

IV. Regression Results

This section presents regression analysis of the relationship between financial depth and macroeconomic volatility. We start with simple cross-section regressions and then move on to dynamic panel regressions to exploit the time series dimension of the data as well. We then examine the presence of non-linearity in the financial depth-volatility nexus. Finally, we examine the role of financial depth in dampening the effect of external shocks on macroeconomic volatility, taking into account the extent of trade and financial openness of the economy.

A. Baseline Results

We begin with simple reduced-form cross-section regression to more formally characterize the correlation between financial depth and macroeconomic volatility across different time periods. Columns 1-4 of Table 3 show the basic cross-country regression for the full sample. The coefficients on private credit are negative and statistically significant in all regressions, suggesting that countries with deeper financial systems experience less macroeconomic volatility. The control variables generally enter with the expected signs. Trade openness is associated with higher variability of consumption but not investment or GDP, while financial openness is associated with greater GDP volatility. Exchange rate volatility is associated with higher GDP volatility. There is also evidence that countries with stronger fiscal positions experience lower private and total consumption volatility. Finally, stronger political institutions are associated with reduced macroeconomic volatility.

Table 3.

Financial Depth and Volatility: Cross-Section Regressions (OLS)

article image
Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1

The empirical results are broadly consistent across different country groups. Columns 5-8 of Table 3 report the regression results for the sample of low- and middle-income countries. We still find a positive and statistically significant correlation between private credit and output and consumption volatility, with coefficient estimates that are larger for consumption volatility than those obtained for the full sample. It is well documented that consumption volatility tends to be substantially higher in developing countries (Loayaza et al., 2007). In line with this, we find that the beneficial effect of financial depth in smoothing consumption and investment volatility is more significant in these countries.

Financial depth has changed markedly over time across countries. To exploit the time series variation in the data, we now move on to using dynamic panel regressions based on nonoverlapping five-year averaged data for each country. Table 4 reports the GMM regression results for the full sample (Columns 1-4), and separately for a sample of developing countries (Columns 5-8).5 The coefficient on private credit is significantly negative in all regressions, implying that greater financial depth is associated with lower macroeconomic volatility across different country groups. The estimates are also economically significant: a one-standard deviation increase in private credit to GDP (37.4 percent of GDP in the sample) reduces the volatility of growth, consumption, and investment by over 15 percent. As in the simple OLS regressions, the point estimate of the coefficient on private credit is larger in the sub-sample of middle-and low-income countries, suggesting that greater financial depth has a more pronounced effect in smoothing consumption volatility in these countries.

Table 4.

Financial Depth and Volatility: Panel Regressions (GMM)

article image
Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1

Non-linear relationship

We have so far considered a linear relationship between financial depth and volatility. To investigate the presence of a non-monotonic relationship between the two, as in Easterly et al. (2000), we introduce a quadratic term to the baseline regressions reported in Table 4. Specifically, we estimate the following regression:

Vi,t=αVi,t−1+γ1FDi,t+γ2FDi,t2+βXi,t+ut+μi+εi,t(2)

From the above equation the marginal impact of financial depth on volatility is given by γ1 + 2γ2 FDi,t. Therefore, the threshold at which the impact of financial development on volatility changes from negative to positive (i.e. magnifies volatility) is (− γ1/2γ2)

As shown in Table 5, both the linear and the quadratic terms are statistically significant in almost all regressions. In general, the results suggest a U-shaped effect of private credit in smoothing volatility. Deeper financial systems reduce the volatility of macroeconomic aggregates, but only up to a certain threshold.6 The coefficient estimates indicate that this threshold is approximately 132 percent of GDP for final consumption, 119 for private consumption, and above 106 percent of GDP for investment (Columns 2-4). Above these levels, consumption and investment volatility increases with the level of financial depth.7 This suggests that consumption and production smoothing possibilities provided by the existence of a deep financial system might reduce volatility, but up to a limit. As the financial system becomes larger relative to GDP, the increase in risk and leverage become relatively more important, and act to reduce stability.

Table 5.

Non-linear Relationship Between Financial Depth and Volatility: Panel Regressions (GMM)

article image
Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1

Figure 3 plots the marginal effect of credit to the private sector on final consumption volatility based on the estimates of Column 2 in Table 5. It shows that the positive effect of financial depth in reducing consumption volatility is no longer statistically significant when credit to the private sector reaches 105 percent of GDP (over 18 percent of the observations in the regression of Column 2 are above this threshold). In 2008, the last year included in the dataset, there were 20 countries above this level, most of which are advanced countries significantly affected by the global financial crisis: the United States, Ireland, the United Kingdom, Spain, Portugal, and Cyprus.

B. Shocks

It is well documented that external shocks, such as fluctuations in the terms of trade, commodity prices, and external demand, play an important role in fueling macroeconomic volatility. In this section we first augment the baseline specification to examine whether financial depth reduces volatility, controlling for real and financial shocks. Second, we examine the role of financial depth in mitigating or amplifying the impact of external shocks on macroeconomic volatility. Finally, we consider whether this impact depends on the structural characteristics of an economy (e.g., degree of trade and financial openness).

Controlling for shocks

Table 6 extends the analysis in Table 4 and studies whether the association between financial depth and volatility is robust to controlling for measures of external shocks (measured as standard deviation of the terms-of-trade and external demand) and financial shocks (as proxied by banking crises). All regressions include the full set of control variables and control one-by-one, respectively for banking crises (Columns 1, 5, 9, and 13), fluctuations in the terms-of-trade (Columns 2, 6, 10, and 14) and external demand (Columns 3, 7, 11, and 15). In line with the literature, fluctuations in the terms-of-trade increase the overall volatility of output, consumption, and investment. However, external demand (partner growth) volatility is only associated with higher output volatility. The coefficient on external demand is insignificant for consumption and investment volatility. Finally, the results suggest that banking crises are an important determinant of output and investment volatility.

Table 6.

Financial Depth, Shocks, and Macroeconomic Volatility: Panel-Regressions (GMM)

article image
Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1

In almost all regressions, the impact of financial depth in reducing macroeconomic volatility remains negative and statistically significant. Columns 4, 8, 12, and 16 in Table 6 show that the association between financial depth and macroeconomic volatility also remains significant when all the shock variables are included simultaneously. All these results taken together suggest that financial depth helps mitigate macroeconomic volatility even after controlling for external shocks, thus providing support for the role of financial systems in fostering risk diversification and reducing information asymmetries within an economy.

Interacting financial depth and exogenous shocks

We now turn to the interaction between exogenous shocks and financial depth to examine whether finance plays an amplifying or mitigating role through these shocks. The following form of equations is estimated:

Vi,t=αVi,t−1+γ1FDi,t+γ2Shocki,t+γ3FDi,t·Shocki,t+βXi,t+ut+μi+εi,t(3)

where Shock is a measure of real external shocks and the other terms are defined as earlier.

Table 7 illustrates the impact of domestic financial depth in reducing the impact of terms-of-trade fluctuations across different measures of economic volatility (Columns 1-4). The interaction term is negative and statistically significant, suggesting that the more financially developed an economy, the less it is adversely affected by terms of trade volatility. We also find some evidence on the role of financial depth in reducing the impact of external demand shocks. The regressions indicate that the interaction term of partner country volatility and financial depth is negative and significant for private consumption volatility (Columns 5-8), although the effects are less significant for output and investment volatility.

Table 7.

Interacting Financial Depth and External Shocks: Panel-Regressions (GMM)

article image
Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1