Kurt G. Lunsford

We study how congressional testimony about monetary policy by the Chair of the Board of Governors of the Federal Reserve System affects interest rates and stock prices. First, we study testimony associated with the Federal Reserve’s Monetary Policy Reports (MPRs) to Congress. Testimony for a particular MPR is usually given on two days, one day for each chamber of Congress. We separately study the first day and second day of MPR testimony. We also study testimonies not associated with MPRs but that are still related to monetary policy. We find that first-day MPR testimonies cause the largest movements in interest rates and generate negative co-movement between interest rates and stock prices. Testimonies not associated with MPRs have similar but weaker effects. Second-day MPR testimonies cause the smallest movements in interest rates and generate no co-movement between interest rates and stock prices.

I show that business cycles can generate most of the low-frequency movements in the unemployment rate. First, I provide evidence that the unemployment rate is stationary, while its flows have unit roots. Then, I model the log unemployment rate as the error correction term of log labor flows in a vector error correction model (VECM) with intercepts that change over the business cycle. Feeding historical expansions and recessions into the VECM generates large low-frequency movements in the unemployment rate. Frequent recessions from the late 1960s to the early 1980s interrupt labor market recoveries and ratchet the unemployment rate upward. Long expansions in the 1980s and 1990s undo this upward ratcheting. Finally, the VECM predicts that the unemployment rate will be near 3.6 percent after a 10-year expansion and that lower unemployment rates are possible with longer expansions.

We study the use of a zero mean first difference model to forecast the level of a scalar time series that is stationary in levels. Let bias be the average value of a series of forecast errors. Then the bias of forecasts from a misspecified ARMA model for the first difference of the series will tend to be smaller in magnitude than the bias of forecasts from a correctly specified model for the level of the series. Formally, let P be the number of forecasts. Then the bias from the first difference model has expectation zero and a variance that is O(1/P-squared), while the variance of the bias from the levels model is generally O(1/P). With a driftless random walk as our first difference model, we confirm this theoretical result with simulations and empirical work: random walk bias is generally one-tenth to one-half that of an appropriately specified model fit to levels.

We collect data from Worker Adjustment and Retraining Notification (WARN) Act notices and establish their usefulness as an indicator of aggregate job loss. The number of workers affected by WARN notices ("WARN layoffs") leads state-level initial unemployment insurance claims, and changes in the unemployment rate and private employment. WARN layoffs move closely with aggregate layoffs from Mass Layoff Statistics and the Job Openings and Labor Turnover Survey, but are timelier and cover a longer sample. In a vector autoregression, changes in WARN layoffs lead unemployment rate changes and job separations. Finally, they improve pseudo real-time forecasts of the unemployment rate.

We collect rich establishment-level data about advance layoff notices filed under the Worker Adjustment and Retraining Notification (WARN) Act since January 1990. We present in-sample evidence that the number of workers affected by WARN notices leads state-level initial unemployment insurance claims, changes in the unemployment rate, and changes in private employment. The effects are strongest at the one and two-month horizons. After aggregating state-level information to a national-level “WARN factor” using a dynamic factor model, we find that the factor substantially improves out-of-sample forecasts of changes of manufacturing employment in real time.

Proxy structural vector autoregressions identify structural shocks in vector autoregressions with external variables that are correlated with the structural shocks of interest but uncorrelated with all other structural shocks. We provide asymptotic theory for this identification approach under mild α-mixing conditions that cover a large class of uncorrelated, but possibly dependent innovation processes, including conditional heteroskedasticity. We prove consistency of a residual-based moving block bootstrap for inference on statistics such as impulse response functions and forecast error variance decompositions. Wild bootstraps are proven to be generally invalid for these statistics and their coverage rates can be badly and persistently mis-sized.

This paper studies the effects of Federal Open Market Committee (FOMC) forward guidance language. I estimate two policy surprises at FOMC meetings: a change in the current federal funds rate and an orthogonal change in the expected path of the federal funds rate. From February 2000 to June 2003, the FOMC only gave forward guidance about risks to the economic outlook, and a surprise increase in the expected federal funds rate path had expansionary effects. This is consistent with models of central bank information effects, where a positive economic outlook causes private agents to revise up their expectations for the economy. From August 2003 to May 2006, the FOMC also gave forward guidance about policy inclinations, and a surprise increase in the federal funds rate path had contractionary effects. These results are consistent with standard macroeconomic models of forward guidance. Overall, the effects of forward guidance depend on the FOMC’s choice to use one or both of the economic-outlook and policy-inclination aspects of forward guidance.

We study long-run correlations between safe real interest rates in the United States and over 20 variables that have been hypothesized to influence real rates. The list of variables is motivated by the familiar intertermporal IS equation, by models of aggregate savings and investment, and by reduced form studies. We use annual data, mostly from 1890 to 2016. We find that safe real interest rates are correlated as expected with demographic measures. For example, the long-run correlation with labor force hours growth is positive, which is consistent with overlapping generations models. For another example, the long-run correlation with the proportion of 40- to 64-year-olds in the population is negative. This is consistent with standard theory where middle-aged workers are high-savers who drive down real interest rates. In contrast to standard theory, we do not find productivity to be positively correlated with real rates. Most other variables have a mixed relationship with the real rate, with long-run correlations that are statistically or economically large in some samples and by some measures but not in others.

Proxy structural vector autoregressions (SVARs) identify structural shocks in vector autoregressions (VARs) with external proxy variables that are correlated with the structural shocks of interest but uncorrelated with other structural shocks. We provide asymptotic theory for proxy SVARs when the VAR innovations and proxy variables are jointly α-mixing. We also prove the asymptotic validity of a residual-based moving block bootstrap (MBB) for inference on statistics that depend jointly on estimators for the VAR coefficients and for covariances of the VAR innovations and proxy variables. These statistics include structural impulse response functions (IRFs). Conversely, wild bootstraps are invalid, even when innovations and proxy variables are either independent and identically distributed or martingale difference sequences, and simulations show that their coverage rates for IRFs can be badly mis-sized. Using the MBB to re-estimate confidence intervals for the IRFs in Mertens and Ravn (2013), we show that inferences cannot be made about the effects of tax changes on output, labor, or investment.

Using a factor-augmented vector autoregression (FAVAR), this paper shows that residential investment contributes substantially to GDP following monetary policy shocks. Further, it shows that the number of new housing units built, not changes in the sizes of existing or new housing units, drives residential investment fluctuations. Motivated by these results, this paper develops a dynamic stochastic general equilibrium (DSGE) model where houses are built in discrete units and traded through searching and matching. The search frictions transmit shocks to housing construction, making them central to producing fluctuations in residential investment. The interest rate spread between mortgages and risk-free bonds also transmits monetary policy to the housing market. Following monetary shocks, the DSGE model matches the FAVAR's positive co-movement between nondurable consumption and residential construction spending. In addition, the FAVAR shows that the mortgage spread falls following an expansionary monetary shock, providing empirical support for the DSGE model's monetary transmission mechanism.

This paper develops a simple estimator to identify structural shocks in vector autoregressions (VARs) by using a proxy variable that is correlated with the structural shock of interest but uncorrelated with other structural shocks. When the proxy variable is weak, modeled as local to zero, the estimator is inconsistent and converges to a distribution. This limiting distribution is characterized, and the estimator is shown to be asymptotically biased when the proxy variable is weak. The F statistic from the projection of the proxy variable onto the VAR errors can be used to test for a weak proxy variable, and the critical values for different VAR dimensions, levels of asymptotic bias, and levels of statistical significance are provided. An important feature of this F statistic is that its asymptotic distribution does not depend on parameters that need to be estimated.

We study how one form of Federal Reserve communication, the congressional testimony by the Fed Chair, affects interest rates on 2-year and 10-year Treasury Notes.

The United States has two measures of economic output: gross domestic product (GDP) and gross domestic income (GDI). While these are conceptually equivalent, their initial estimates differ because these initial estimates are computed from different and incomplete data sources. I study the difference, or “statistical discrepancy,” between GDP and GDI in percent and document three features. First, its size does not materially shrink on average as more data become available. Second, the size of the initial discrepancy in absolute value does not predict the size of the discrepancy in absolute value after revisions. Third, the initial discrepancy has some predictive information about revisions to lagged GDP growth but no predictive information about revisions to lagged GDI growth.

The underemployment rate, the percent of employed people who are working part-time but prefer to be working full-time, moves closely with the unemployment rate, rising during recessions and falling during expansions. Following the Great Recession, the underemployment rate had stayed persistently elevated when compared to the unemployment rate, that is, until the COVID-19 recession. Since then, it has been consistent with its pre-2008 levels. We find that changes in relative industry size account for essentially none of the underemployment rate increase after the Great Recession nor the underemployment rate decrease after the COVID-19 recession. Based on this finding, we do not expect the underemployment rate to revert to its pre-COVID-19 levels if industry composition reverts to its pre-COVID-19 structure.

The unemployment rate in the United States falls slowly in expansions, and it may not reach its previous low point before the next recession begins. Based on this feature, I document that the frequent recessions prior to 1983 are associated with an upward trend in the unemployment rate. In contrast, the long expansions beginning in 1983 are associated with a downward trend. I then estimate a two-variable vector autoregression (VAR) that includes the unemployment rate and a recession indicator. Long-horizon forecasts from this VAR conditioned on no future recessions project that the unemployment rate will go to 3.6 percent after a long period with no recessions.

We use advance layoff notices filed under the Worker Adjustment and Retraining Notification (WARN) Act as an indicator of current and imminent labor market conditions. We have constructed a database of establishment-level notices starting in 1990 by scraping state government websites, contacting state officials, and retrieving historical data. We find evidence that these notices, aggregated to the national level, lead other prominent labor market indicators, such as initial unemployment insurance claims, the change in the unemployment rate, and changes in private employment. The lead relationship seems strongest at one month with economically meaningful magnitudes. Most recently, WARN data suggest a slight increase in labor market slack.

Measuring economic growth is complicated by seasonality, the regular fluctuation in economic activity that depends on the season of the year. The BEA uses statistical techniques to remove seasonality from its estimates of GDP, but some research has indicated that seasonality remains. As a result, the BEA began a three-phase plan in 2015 to improve its seasonal-adjustment techniques, and in July 2018, it completed phase 3. Our analysis indicates that even after these latest improvements by the BEA, residual seasonality in GDP growth remains. On average, this residual seasonality makes GDP growth appear to be slower in the first quarter of the year and more rapid in the second quarter of the year. Rapid second-quarter growth is particularly noticeable in recent years. As a result, business economists and policymakers may want to take seasonality into account when using GDP to assess the health of the economy.

An inverted Treasury yield curve—a yield curve where short-term Treasury interest rates are higher than long-term Treasury interest rates—is a good predictor of recessions. Because of this, economists and policymakers often assess the risk of a yield curve inversion when the yield curve is flattening. I study the forecastability of yield curve inversions. Professional forecasters did not predict the beginning of the yield curve inversions prior to the 1990–1991, 2001, and 2008–2009 recessions. In all three cases, professional forecasters failed to predict the magnitude of the rise in short-term interest rates. Prior to the 2008–2009 recession, forecasters also overpredicted long-term interest rates.

Despite the unemployment rate’s return to low levels, inflation-adjusted or “real” interest rates have remained negative. One popular explanation for persistently negative real interest rates is that long-run productivity growth has slowed. I study the long-run relationship between real interest rates and productivity growth from 1914 to 2016 and find a negative correlation between these two variables. Hence, low productivity growth has been historically associated with high real interest rates. Since World War II, the correlation between these variables has been near zero. This suggests that slow long-run productivity growth is not driving real interest rates to be persistently negative.

Measuring economic growth is complicated by seasonality, the regular fluctuation in economic activity that depends on the season of the year. The Bureau of Economic Analysis uses statistical techniques to remove seasonality from its estimates of GDP, and, in 2015, it took steps to improve the seasonal adjustment of data back to 2012. I show that residual seasonality in GDP growth remains even after these adjustments, has been a longer-term phenomenon, and is particularly noticeable in the 1990s. The size of this residual seasonality is economically meaningful and has the ability to change the interpretation of recent economic activity.