Machine learning algorithms can find predictive signals that researchers fail to notice; yet they are notoriously hard-to-interpret. How can we extract theoretical insights from these black boxes? History provides a clue. Facing a similar problem - how to extract theoretical insights from their...

We propose two new procedures for comparing the mean squared prediction error (MSPE) of a benchmark model to the MSPEs of a small set of alternative models that nest the benchmark. Our procedures compare the benchmark to all the alternative models simultaneously rather than sequentially, and do...

We argue that comprehensive out-of-sample (OOS) evaluation using statistical decision theory (SDT) should replace the current practice of K-fold and Common Task Framework validation in machine learning (ML) research. SDT provides a formal framework for performing comprehensive OOS evaluation...

We develop a model to predict consumer default based on deep learning. We show that the model consistently outperforms standard credit scoring models, even though it uses the same data. Our model is interpretable and is able to provide a score to a larger class of borrowers relative to standard...

An employee's annual earnings fall by 10% the year her firm files for bankruptcy and fall by a cumulative present value of 67% over seven years. This effect is more pronounced in thin labor markets and among small firms that are ultimately liquidated. Compensating wage differentials for this...

A serious difficulty in determining the importance of credit constraints in education arises because standard data sources do not provide a direct way of identifying which students are credit constrained. This has forced researchers to adopt a variety of indirect approaches. This paper...

In this paper, we investigate how personal bankruptcy law affects small firms' access to credit. When a firm is unincorporated, its debts are personal liabilities of the firm's owner, so that lending to the firm is legally equivalent to lending to its owner. If the firm fails, the owner has an...

There is little evidence on how the large market for credit score improvement products affects consumers or credit market efficiency. A randomized encouragement design on a standard credit builder loan (CBL) identifies null average effects on whether consumers have a credit score and the score...

Partisan perception affects the actions of professionals in the financial sector. Using a novel dataset linking credit rating analysts to party affiliations from voter records, we show that analysts who are not affiliated with the U.S. president's party downward-adjust corporate credit ratings...

Lending standards are a direct measure of credit conditions. We use the micro data merged from three separate sources to construct this measure and document that an uncertain macroeconomic outlook, rather than banks' balance sheet positions, was an important reason that a majority of banks...