Search

We demonstrate the usefulness of payment systems data and machine learning models for macroeconomic predictions and provide a set of econometric tools to overcome associated challenges.

We analyze monetary policy responses to AI in a two-sector New Keynesian model, distinguishing augmentation and automation. Both reduce labor demand, requiring accommodation that creates inflation trade-offs. Automation worsens them. Aggregate inflation depends on AI’s form and breadth, making policy stabilization more complex and aggregate data potentially misleading.

Can monetary policy influence the neutral real interest rate (r-star)? Using a new statistical model, we show that interest rate hikes tend to lower r-star and long-run growth, but that monetary policy explains only a small share of the long-run decline in r-star.

Central banks in many advanced economies enjoy a high degree of independence, which protects monetary policy decisions from political influence. But how should independent central banks react if pressured by fiscal policy-makers? We examine whether a central bank should design a monetary policy framework that prescribes acting conditionally on how fiscal policy behaves.

I propose a novel method to identify and estimate the macroeconomic effects of forward guidance and large-scale asset purchases (LSAP) for each FOMC announcement. I find that LSAP is more important than forward guidance in influencing output and inflation. LSAP puts upward pressure on short-term yields, so it should always be used in conjunction with forward guidance.

We investigate the unintended consequences of the Home Affordable Refinance Program (HARP). Originally designed to help borrowers refinance after the 2008–09 global financial crisis, HARP inadvertently strengthened the market power of incumbent lenders by creating a cost advantage for them. Despite a 2013 policy rectifying this cost advantage, we still find significant welfare losses for borrowers.

Nonresponse is a considerable challenge in the Retailer Survey on the Cost of Payment Methods conducted by the Bank of Canada in 2015. There are two types of nonresponse in this survey: unit nonresponse, in which a business does not reply to the entire survey, and item nonresponse, in which a business does not respond to particular questions within the survey.

Is obscuring prices always bad for consumers? The answer depends on the market structure and on the negotiating power between manufacturers and retailers.

Should managers be paid in stock options if they provide stock-market participants with information about the firm? This paper studies how firm owners trade off the benefit of stock-price incentives and better-informed market participants against the cost of potential stock-price manipulation.

We introduce behavioral learning equilibria (BLE) into DSGE models with boundedly rational agents using simple but optimal first order autoregressive forecasting rules. The Smets-Wouters DSGE model with BLE is estimated and fits well with inflation survey expectations. As a policy application, we show that learning requires a lower degree of interest rate smoothing.