Financial Development, Shocks, and Growth Volatility

This paper argues that studying the effect of financial development and shocks on aggregate growth volatility will not be informative because they affect growth volatility through its different components. Volatility declines either a consequence of a change in the nature of shocks or a change in how the economy reacts to shocks. If two economies differ only in terms of volatility of shocks experienced, the GDP growth spectrum of one economy will lie proportionately below that of another at all frequency ranges so that both business cycle and long-run variances will be lower. Conversely, if change in volatility is due to propagation mechanism such as financial development, a country having developed financial markets will have disproportionately lower variance at the business cycle than at other frequencies relative to that of a country having less developed financial markets. Therefore, the variance at only the business cycle frequency range will be influenced by financial development. The novelty of this paper is that different components of growth volatility are extracted using spectral method. Empirical evidence provides qualified support for both hypotheses. Higher private credit, which is used as proxy of financial development, dampens business cycle volatility but not the long-run volatility. Shocks, as measured by changes in the terms of trade, affect both business cycle and long-run volatility negatively. These results are robust to alternative market-based measure of financial development, and corrections for reverse causality. These results have important implications for growth theory as they shed lights on the factors causing permanent and transitory deviations from the steady state.