Asymmetric Loss of Macroeconomic Forecasts in South Asia: Evidence from the SPF Survey of India, Indonesia, and Singapore

ABSTRACT

This paper examines asymmetry of the loss function of professional forecasters for output growth, inflation, and exchange rate forecasts, based on the survey of professional forecasts (SPF) data of three South Asian countries (India, Indonesia, and Singapore). Our results provide India’s unbiased output growth forecasts and under-predicted inflation and exchange rate forecasts; Indonesia’s broadly unbiased forecasts; and Singapore’s under-predicted output growth forecasts and unbiased inflation and exchange rate forecasts. Testing the rationality of all three countries’ SPF forecasts, we find that all are rational under an asymmetric loss function but not under a symmetric loss function.

KEYWORDS:

• Macroeconomic forecasting
• asymmetric loss
• professional forecasts
• South Asia

JEL:

• C12
• C53
• E17
• E37

Notes

1. The purposive sampling method reaches a targeted sample based on the characteristics of a population. BI chooses targeted respondents who are concerned with macroeconomic conditions, such as economists, economic researchers, market analysts, economics professors, and banking practitioners. The random sampling method used for the other two countries is unlikely to make a significant difference because their populations are typically well informed and not very heterogeneous.

2. Since the weighting matrix depends on the estimate $a$, the estimation in this paper is iterative, assuming that the weighting matrix in the first round is an identity matrix. The continuously updated estimator of Hansen, Heaton, and Yaron (1996) and the Bartlett kernel are used for possible gains in finite-sample efficiency.

3. A constant is included as an instrument in all models considered in this paper. However, for notational convenience, reference to it is dropped hereafter. For the GMM estimation of exchange rates, the exchange rate changes in actual values and forecasts are used for the stationarity assumption to be met.

4. See Appendix A. in Elliott, Komunjer, and Timmermann (2005) for details.

5. Irrespective of the forecast horizons and countries, these macroeconomic forecasts also efficiently incorporate the information provided by the other macroeconomic forecasts.