We develop metrics based on Shapley values for interpreting time-series forecasting models, including “black-box” models from machine learning. Our metrics are model agnostic, so that they are applicable to any model (linear or nonlinear, parametric or nonparametric). Two of the metrics,...

We develop metrics based on Shapley values for interpreting time-series forecasting models, including "black-box" models from machine learning. Our metrics are model agnostic, so that they are applicable to any model (linear or nonlinear, parametric or nonparametric). Two of the metrics,...

We propose an out-of-sample prediction approach that combines unrestricted mixed-data sampling with machine learning (mixed-frequency machine learning, MFML). We use the MFML approach to generate a sequence of now- and backcasts of weekly unemployment insurance initial claims based on a rich...

We develop metrics based on Shapley values for interpreting time-series forecasting models, including "black-box" models from machine learning. Our metrics are model agnostic, so that they are applicable to any model (linear or nonlinear, parametric or nonparametric). Two of the metrics,...

This paper establishes a methodology to elicit and quantify narratives from survey data using textual analysis. We extract thirteen narratives from daily US stockholder questionnaires conducted during the first-wave COVID-19 period and measure their prevalence over time. Survey-based narratives...

We show that Google search activity on relevant terms is a strong out-of-sample predictor for future employment growth in the US over the period 2004-2019 at both short and long horizons. Starting from an initial search term ''jobs'', we construct a large panel of 172 variables using Google's...

This paper documents the asset pricing implications of the data release process of National Income and Product Accounts (NIPA) consumption expenditure. We show that initial consumption data releases are more suitable for asset pricing than final revised releases. This is because most revisions...

The authors use a dynamic latent factor model to analyze comovements in OECD surpluses. The world factor underlying common fluctuations in budget surpluses across countries explains an average of 28 to 44 percent of the variation in individual country surpluses. The world factor, which can be...

We analyze comovements in four measures of budget surpluses for 18 OECD countries for 1980–2008 with a dynamic latent factor model. The world factor in national budget surpluses declines substantially in the 1980s, rises throughout much of the 1990s to a peak in 2000, before declining again in...