Inflation forecasting, relative price variability and skewness

Abstract

This article presents out-of-sample inflation forecasting results based on relative price variability and skewness. It is demonstrated that forecasts on long horizons of 1.5–2 years are significantly improved if the forecast equation is augmented with skewness.

Notes

¹ Vining and Elwertowski (1976) and Parks (1978) contain empirical evidence on the (noncausal) relationship between inflation and price variability. Kücük and Tuğer (2004) survey the models discussed in this paragraph.

² The wording in this paragraph follows Barnett and Serletis (1990) closely. They provide further insights about the stochastic interpretation of the Divisia index.

³ The recorded expenditure flow is zero on one or more occasions in 9 of these 108 series. They have, subsequently, been deleted from our data set.

⁴ Using autoregressive inflation models, we initially compared the direct forecasting method with the iterative method. We found that the direct method yielded, on average, 23% lower forecast errors. This finding supports our choice of forecasting method.

⁵ All other results are available from the authors upon request.