Are daily financial data useful for forecasting GDP? Evidence from Mexico

Abstract

This article evaluates the use of financial data sampled at high frequencies to improve short-term forecasts of quarterly GDP for Mexico. In particular, the mixed data sampling (MIDAS) regression model is employed to incorporate both quarterly and daily frequencies while remaining parsimonious. To preserve parsimony, factor analysis and forecast combination techniques are used to summarize the information contained in a dataset containing 392 daily financial series. Our findings suggest that the MIDAS model that incorporates daily financial data lead to improvements for quarterly forecasts of GDP growth over traditional models that either rely only on quarterly macroeconomic data or average daily financial data. Furthermore, we explore the ability of the MIDAS model to provide forecast updates for GDP growth (nowcasting).

A hybrid MIDAS approach for forecasting hotel demand using large panels of search data

“…Since the original study conducted by Bates and Granger (1969), forecast combinations have attracted extensive attention. This method combines the prediction results of individual models by using predetermined weighting schemes, thereby improving the forecasting performance over that offered by individual models (Gomez-Zamudio and Ibarra, 2017). Timmermann (2006) and Kim and Swanson (2014) have discussed this topic.…”

“…This approach has been applied to a large number of empirical studies. For example, to forecast Mexico’s GDP, Gomez Zamudio and Ibarra (2017) used factor analysis to extract a few common factors from 392 daily financial series and combined the forecast results of the single-factor MIDAS models with forecast combinations. The results show that this approach significantly improves prediction accuracy relative to the single-factor MIDAS model.…”

“…A large number of prior studies have shown that forecast combinations can improve prediction performance by using information from all models rather than relying on a single specific model (Aiolfi et al, 2010;Stock and Watson, 2004). Following the recent work by Andreou et al (2013) andGomez Zamudio andIbarra (2017), and so on, this study introduces dynamic factor model and forecast combinations as two complementary approaches to deal with large-scale data that cannot be handled in one go.…”

Forecasting recovery from COVID-19 using financial data: An application to Vietnam

“…Their findings point to improvements in quarterly forecasts of U.S. real GDP growth beyond quarterly macroeconomic factors. Applying a similar approach to Mexico, Gómez-Zamudio and Ibarra (2017) find that financial data can improve quarterly forecasts of GDP growth over traditional models that rely on quarterly macroeconomic data. Mitchell (2020) combines daily and weekly financial data with monthly macroeconomic indicators in a mixed frequency probit (MFP) regression to accurately forecast and nowcast U.S. and Canadian recessions.…”