Monetary and Fiscal Policy Performance in Sri Lanka: Empirical Evidence and Optimal Policy

Abstract

This thesis consists of four self-contained essays on policy rules and macroeconomic behaviour of an emerging market economy, in addition to the introductory first chapter. It examines monetary and fiscal policy performance, develops a small open-economy dynamic stochastic general equilibrium (DSGE) model and investigates welfare-maximising optimal monetary rules, with reference to the Sri Lankan economy. The second chapter empirically characterises the monetary policy of Sri Lanka using alternative policy reaction functions. The estimation results suggest that the Taylor rule captures the monetary policy reaction characteristics in Sri Lanka more closely than the McCallum rule. The results further implies that the Central Bank of Sri Lanka (CBSL) follows a contemporaneous or backward-looking rule instead of a forward-looking monetary policy rule. The results indicate that the CBSL responds to inflation fairly strongly while stabilising output aggressively. It is also found that the CBSL smoothes out policy action strongly. This suggests that the CBSL implements policy gradually in small steps in the desired direction, without making immediate sharp changes. The third chapter investigates the fiscal policy performance of Sri Lanka in the period subsequent to the implementation of numerical fiscal targets in 2003. It estimates alternative fiscal policy rules widely used in literature including tax difference rules, primary balance rules and Taylor-type fiscal rules for Sri Lanka. I find that the fiscal authority contemporaneously responds to changes in output gap and government expenditure moderately, by changing the tax rate. Moreover, it implies that the fiscal policy in Sri Lanka is procyclical rather than countercyclical. The fourth chapter develops a New Keynesian (NK) small open-economy (SOE) dynamic stochastic general equilibrium (DSGE) model. This study is different from most of the other available SOE studies, as it includes an explicit fiscal sector in the model. The model is estimated for Sri Lanka using Bayesian techniques. The dynamics of the model in response to different shocks are examined to understand the relative importance of the business cycle drivers. The properties of the estimated rules are analysed to identify the policy reaction behaviour of the monetary and fiscal authorities. The estimation results suggest that over the sample period, the CBSL conducted moderately strong anti-inflationary monetary policy while paying substantial attention to output stabilisation, however, with negligibly small concerns for exchange rate movements. The findings suggest a high degree of interest rate persistence as well. The findings further imply that the fiscal authority of Sri Lanka changes its policy instrument, the tax rate, only moderately in response to changes in debt level, government expenditure or output. The fifth chapter computes welfare-maximising optimal monetary policy rules for Sri Lanka based on a slightly simplified variant of the DSGE model discussed in the fourth chapter. I calculate second-order accurate solutions to the model, which facilitate welfare computation across various policy rules. I determine optimal monetary policy rules such that the welfare associated with them is as close as possible to the Ramsey optimal allocation. The welfare cost of adopting alternative rules, instead of the optimal, is determined to evaluate the relative importance of the different policy rules. There are several key findings. First, the optimal monetary policy rule suggests an aggressive response to inflation and a moderate response to output-gap. Second, the optimal policy advocates a muted response to exchange rate fluctuations. Third, the welfare gains from interest rate smoothing are significant. Finally, the welfare losses associated with the current realised monetary policy rule for Sri Lanka can be mitigated significantly by responding to inflation more strongly. Finally, the sixth chapter summarises the key findings and concludes the thesis.