Working Paper: CEPR ID: DP3060
Authors: Ryan Sullivan; Allan G. Timmermann; Halbert White
Abstract: Data sharing is common practice in forecasting experiments in situations where fresh data samples are difficult or expensive to generate. This means that forecasters often analyze the same data set using a host of different models and sets of explanatory variables. This practice introduces statistical dependencies across forecasting studies that can severely distort statistical inference. Here we examine a new and inexpensive recursive bootstrap procedure that allows forecasters to account explicitly for these dependencies. The procedure allows forecasters to merge empirical evidence and draw inference in the light of previously accumulated results. In an empirical example, we merge results from predictions of daily stock prices based on (1) technical trading rules and (2) calendar rules, demonstrating both the significance of problems arising from data sharing and the simplicity of accounting for data sharing using these new methods.
Keywords: bootstrap; calendar effects; data sharing; forecast evaluation; technical trading
JEL Codes: C10
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| data sharing (D16) | statistical dependencies (C29) |
| statistical dependencies (C29) | statistical inference distortion (C46) |
| data sharing (D16) | statistical inference distortion (C46) |