Forecast Combination for Discrete Choice Models: Predicting FOMC Monetary Policy Decisions

Abstract

This paper provides a methodology for combining forecasts based on several discrete choice models. This is achieved primarily by combining one-step-ahead probability forecast associated with each model. The paper applies well-established scoring rules for qualitative response models in the context of forecast combination. Log-scores and quadratic-scores are both used to evaluate the forecasting accuracy of each model and to combine the probability forecasts. In addition to producing point forecasts, the effect of sampling variation is also assessed. This methodology is applied to forecast the US Federal Open Market Committee (FOMC) decisions in changing the federal funds target rate. Several of the economic fundamentals influencing the FOMC decisions are nonstationary over time and are modelled in a similar fashion to Hu and Phillips (2004a, JoE). The empirical results show that combining forecasted probabilities using scores mostly outperforms both equal weight combination and forecasts based on multivariate models. Abstract This paper provides a methodology for combining forecasts based on several discrete choice models. This is achieved primarily by combining one-step-ahead probability forecast associated with each model. The paper applies well-established scoring rules for qualitative response models in the context of forecast combination. Log-scores and quadratic-scores are both used to evaluate the forecasting accuracy of each model and to combine the probability forecasts. In addition to producing point forecasts, the effect of sampling variation is also assessed. This methodology is applied to forecast the US Federal Open Market Committee (FOMC) decisions in changing the federal funds target rate. Several of the economic fundamentals influencing the FOMC decisions are nonstationary over time and are modelled in a similar fashion to Hu and Phillips (2004a, JoE). The empirical results show that combining forecasted probabilities using scores mostly outperforms both equal weight combination and forecasts based on multivariate models.

Research on Forecasting of China's Monetary Policy Based on Random Forest Algorithm.pdf

“…Table 5 shows that Li Haitao uses the non-stationary discrete selection model to predict the direction of China's central bank monetary policy, Hu and Philps use the orderly Probit model to predict the Fed target fund interest rate decision, and Andrey Vasnev [9](2013) uses the combined forecasting model to predict the Australian Reserve Bank's currency. Policy operations and Pauwels L [11](2012) use a combined forecasting method to predict the forecast accuracy of the Fed's monetary policy operations.…”

“…All of the above scholars use a single discrete selection model to predict the trend of the target benchmark interest rate. Because the combined model can not only improve the accuracy of model prediction, but also allow the prediction combination to consider more variables than the single multivariate model, Pauwels [8] presented data and models for applying prediction combinations to Hu and Philps. The prediction results are formed by assigning different weights to each single model.…”

“…The empirical results show that the prediction combination improves the accuracy of model prediction [7]. Ruthira Naraidoo [9], Andrey Vasnev [10] and Pauwels L [11] used a combined prediction method based on the ordered Probit model to study the decisionmaking of monetary policy operations of the Reserve Bank of South Africa, the Reserve Bank of Australia and the Federal Reserve.…”

“…Andrey Vasnev [10] selected unemployment rate, consumer confidence index, commodity price, total export value, M3 annual growth rate, owner's housing loan, and monthly growth rate of S&P 500 index when forecasting the direction of the Australian Reserve Bank's monetary policy. Indicator [11]. Li Haitao selected 40 indicators including CPI, GDP, commodity retail price index, PPI and other real economic activities, currency and capital markets, stock index, foreign exchange market, interest rate and price level when forecasting the operation status of China's monetary policy [7].…”

Research on Forecasting of China's Monetary Policy Based on Random Forest Algorithm.pdf

“…All of the above scholars use a single discrete selection model to predict the trend of the target benchmark interest rate. Because the combined model can not only improve the accuracy of model prediction, but also allow the prediction combination to consider more variables than the single multivariate model, Pauwels [7] presented data and models for applying prediction combinations to Hu and Philps. The prediction results are formed by assigning different weights to each single model.…”