Nominal GDP Targeting: An Optimal Rule for the West African Economic and Monetary Union?

This dissertation investigates the prospect of adopting a flexible exchange rate regime with the nominal gross domestic product (GDP) as the intermediate target for the conduct of monetary policy for the members of the West African Economic and Monetary Union (WAEMU). In the first chapter, we begin by assessing the choice of exchange rate regime based on three criteria: exchange rate volatility, exposure to real shocks, and institutional capacities. The current peg with the euro remains the de facto political choice for the WAEMU, but also a reliable option for ensuring monetary policy discipline and allaying risks of currency convertibility. However, we argue that exchange rate flexibility is increasingly becoming more viable due to the overall improvement in the central bank's operational and statistical capacities, the high degree of exposure to terms of trade shocks, and the recent change in the WAEMU's trade patterns away from Europe. Therefore, given the case for exchange rate flexibility, we examine the pros and cons of three types of flexible exchange rate regime: inflation targeting, monetary aggregate targeting, and nominal GDP targeting. The main conclusion is that while all three types of flexible exchange rate regime would provide a firm commitment to a nominal anchor, monetary aggregate targeting would be the most practical option to pursue, because monetary aggregate data are highly observable, prone to almost no revisions, and price stability would be guaranteed by preventing excessive money growth. Alternatively, given the poor growth performance realized in the last decade and increasing vulnerability to supply shocks (for example, falling commodity prices and drastic droughts), we argue that under the right circumstances (for example, when the economy is operating far below potential output), choosing a flexible exchange rate regime that allows monetary authorities to prioritize both employment and price stability such as nominal GDP targeting would be most desirable as compared to adopting a flexible exchange rate regime in which the intermediate target of monetary policy is to keep a monetary aggregate or the inflation rate on target. A commitment to stabilize nominal GDP would force the monetary authorities to prioritize employment in the short run without giving up on achieving price stability in the long run. In the second chapter, we examine the stabilizing properties of using nominal GDP as the intermediate target as compared to using the inflation rate, a monetary aggregate, or the exchange rate. We provide a contribution to Bhandari and Frankel's (2015) model by imposing a cost for data uncertainty into the central bank quadratic loss functions, especially under inflation targeting and nominal GDP targeting. First, we show that no precommitment to a nominal anchor leads to a suboptimal outcome. Then, we derive the optimal conditions to determine whether a precommitment to a nominal GDP target would minimize the variability of real output, inflation, and the nominal exchange rate more efficiently than a precommitment to an inflation target, a monetary aggregate target, or an exchange rate target. We draw three main conclusions from the result of the simulations. First, the exchange rate is the least optimal among the choices of nominal anchors considered for the WAEMU. Second, nominal GDP targeting minimizes the central bank's quadratic loss function in an open economy setting given some parameter values for nominal GDP data uncertainty. Third, monetary aggregate targeting is slightly more robust than inflation targeting and exchange rate targeting. In conclusion, the main policy recommendation is that the central bank should prioritize nominal GDP among the choices of nominal anchors. However, in the event of large nominal GDP data revisions or poor data quality, then the secondary option should be to commit to a monetary aggregate target. In the third chapter, we devise a strategy for implementing nominal GDP targeting for the WAEMU by taking into consideration the existing practical concerns of data revisions and the unavailability of quarterly estimates of aggregate nominal GDP. Given the macroeconomic structure of the WAEMU and the magnitude of the estimated shocks affecting the economy, we demonstrate that a monetary aggregate instrument is superior to an interest rate instrument for the WAEMU. Hence, we propose a money-based rule similar to McCallum's (1987) rule, where the central bank is to announce an annual target for nominal GDP growth and use the monetary base as the operational target. Given the discovery of a robust empirical link between aggregate nominal GDP and aggregate nominal trade (value of exports plus imports) and the availability of trade data on a high frequency with minor revisions, we propose to rely on quarterly estimates of nominal trade as a means to solve the existing data issues coming from large revisions and low frequency data. The proposal to rely on nominal trade as a means to approximate trends in nominal GDP emanates from the fact that nominal trade shocks explain about close to 85 percent of variation in nominal GDP growth every year and trade data can be made available on a quarterly basis. Therefore, relying on quarterly estimates of nominal trade would allow the central bank to detect and respond to nominal GDP fluctuations more accurately and on a timely basis.