Working Paper: CEPR ID: DP8828
Authors: Claudia Foroni; Massimiliano Marcellino; Christian Schumacher
Abstract: Mixed-data sampling (MIDAS) regressions allow to estimate dynamic equations that explain a low-frequency variable by high-frequency variables and their lags. When the difference in sampling frequencies between the regressand and the regressors is large, distributed lag functions are typically employed to model dynamics avoiding parameter proliferation. In macroeconomic applications, however, differences in sampling frequencies are often small. In such a case, it might not be necessary to employ distributed lag functions. In this paper, we discuss the pros and cons of unrestricted lag polynomials in MIDAS regressions. We derive unrestricted MIDAS regressions (U-MIDAS) from linear high-frequency models, discuss identification issues, and show that their parameters can be estimated by OLS. In Monte Carlo experiments, we compare U-MIDAS to MIDAS with functional distributed lags estimated by NLS. We show that U-MIDAS performs better than MIDAS for small differences in sampling frequencies. On the other hand, with large differing sampling frequencies, distributed lag-functions outperform unrestricted polynomials. The good performance of U-MIDAS for small differences in frequency is confirmed in an empirical application on nowcasting Euro area and US GDP using monthly indicators.
Keywords: Distributed Lag Polynomials; Mixed Data Sampling; Nowcasting; Time Aggregation
JEL Codes: C53; E37
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| UMIDAS regressions (C29) | GDP growth (O49) |
| MIDAS regressions (C29) | GDP growth (O49) |
| UMIDAS outperforms MIDAS when frequency mismatch is small (C22) | UMIDAS yields lower MSE (C51) |
| MIDAS outperforms UMIDAS when frequency mismatch is large (C52) | MIDAS performance improves (C22) |