Research

We propose a new tool to filter non-linear dynamic models that does not require us to fully specify the model and can be implemented without solving the model. If two conditions are satisfied, we can use a flexible statistical model and a known measurement equation to back out a hidden state. The first condition is that the state is sufficiently volatile or persistent to be recoverable. The second condition requires the possibly non-linear measurement to be sufficiently smooth and to map uniquely to the state absent measurement error. We illustrate the method through various simulation studies and an empirical application to a small open economy model with an occasionally binding constraint.

Paper.

We show that, in a canonical model of policy formation, campaign contributions, and electoral competition, despite donor polarization, candidate agendas converge. If purely office-motivated candidates move away from the centrist agenda, they increase their opponents' contributions more than their own. An extension that introduces a candidate preference for absolute levels of campaign contributions generates divergence of political agendas in equilibrium. We provide a micro-foundation for this extension based on the name recognition of political candidates, and show that it is consistent with a range of empirical evidence.

Paper. Philadelphia Fed Working paper.

Identification in VARs has traditionally mainly relied on second moments. Some researchers have considered using higher moments as well, but there are concerns about the strength of the identification obtained in this way. In this paper, we propose refining existing identification schemes by augmenting sign restrictions with a requirement that rules out shocks whose higher moments significantly depart from independence. This approach does not assume that higher moments help with identification; it is robust to weak identification. In simulations we show that it controls coverage well, in contrast to approaches that assume that the higher moments deliver point-identification. However, it requires large sample sizes and/or considerable non-normality to reduce the width of confidence intervals by much. We consider some empirical applications. We find that it can reject many possible rotations. The resulting confidence sets for impulse responses may be non-convex, corresponding to disjoint parts of the space of rotation matrices. We show that in this case, augmenting sign and magnitude restrictions with an independence requirement can yield bigger gains.

Paper (latest). Earlier NBER working paper and Philadelphia Fed Working paper.

Identification in VARs has traditionally mainly relied on second moments. Some researchers have considered using higher moments as well, but there are concerns about the strength of the identification obtained in this way. In this paper, we propose refining existing identification schemes by augmenting sign restrictions with a requirement that rules out shocks whose higher moments significantly depart from independence. This approach does not assume that higher moments help with identification; it is robust to weak identification. In simulations we show that it controls coverage well, in contrast to approaches that assume that the higher moments deliver point-identification. However, it requires large sample sizes and/or considerable non-normality to reduce the width of confidence intervals by much. We consider some empirical applications. We find that it can reject many possible rotations. The resulting confidence sets for impulse responses may be non-convex, corresponding to disjoint parts of the space of rotation matrices. We show that in this case, augmenting sign and magnitude restrictions with an independence requirement can yield bigger gains.

Paper (latest). Earlier NBER working paper and Philadelphia Fed Working paper.

 Partisanship of state governors affects the efficacy of U.S. federal fiscal policy. Using close election data, we find partisan differences in the marginal propensity to spend federal intergovernmental transfers: Republican governors spend less than Democratic governors. Correspondingly, Republican-led states have lower debt, (delayed) lower taxes, and initially lower economic activity. A New Keynesian model of partisan states in a monetary union implies sizable aggregate effects: The intergovernmental transfer impact multiplier rises by 0.58 if Republican governors spend like Democratic governors, but due to delayed tax cuts, the long-run multiplier is higher with more Republican governors, generating an intertemporal policy trade-off.

Paper (updated: January 2023), NBER working paper (older version), VoxEU column.

We argue that social and political risk causes significant aggregate fluctuations by changing bargaining power. To that end, we document significant changes in the capital share after large political events, such as political realignments, modifications in collective bargaining rules, or the end of dictatorships, in a sample of developed and emerging economies. These policy changes are associated with significant fluctuations in output. Using a Bayesian proxy-VAR estimated with U.S. data, we show how distribution shocks cause movements in output and unemployment. To quantify the importance of these political shocks for the U.S. as a whole, we extend an otherwise standard neoclassical growth model. We model political shocks as exogenous changes in the bargaining power of workers in a labor market with search and matching. We calibrate the model to the U.S. corporate non-.financial business sector and we back out the evolution of the bargaining power of workers over time using a new methodological approach, the partial .filter. We show how the estimated shocks agree with the historical narrative evidence. We document that bargaining shocks account for 28% of aggregate fluctuations and have a welfare cost of 2.4% in consumption units.

Last working paper version, NBER working paper (older version) (supersedes "Political Distribution Risk and Aggregate Fluctuations")

We propose ranking restrictions on impulse-responses to sharpen set identification with sign restrictions in Bayesian vector autoregressions (VARs) and develop tools for inference with tight restrictions. Ranking restrictions come from micro data on heterogeneous industries in VARs, bounds on elasticities, or restrictions on dynamics. We characterize analytically when ranking restrictions sharpen the identified set in small VARs, including for variables not subject to ranking restrictions. Our tools for inference are of independent interest: (1) A simple algorithm testing whether the identified set is nonempty, (2) a fully Bayesian algorithm that directly samples from the identified set, and (3) a prior-robust algorithm to sample the posterior bounds of the identified set. Using our tools, we analyze productivity news shocks. Both heterogeneity and slope restrictions, but not sign restrictions alone, imply that news shocks raise output temporarily, but significantly. With ranking restrictions the output contribution of news shocks drops by 20-40pp.

Paper, FRB Philadelphia working paper (older "Identification Through Heterogeneity" version), code.

Structural DSGE models are used for analyzing both policy and the sources of business cycles.  Conclusions based on full structural models are, however, potentially affected by misspecification.  A competing method is to use partially identified SVARs based on narrative shocks.  This paper asks whether both approaches agree.  Specifically, I use narrative data in a DSGE-SVAR that partially identify policy shocks in the VAR and assess the fit of the DSGE model relative to this narrative benchmark.  In developing this narrative DSGE-SVAR, I develop a tractable Bayesian approach to proxy VARs and show that such an approach is valid for models with a certain class of Taylor rules.  Estimating a DSGE-SVAR based on a standard DSGE model with fiscal rules and narrative data, I find that the DSGE model identification is at odds with the narrative information as measured by the marginal likelihood.  I trace this discrepancy to differences in impulse responses, identified historical shocks, and policy rules.  The results indicate monetary accommodation of fiscal shocks. 

Last working paper version, Appendix, FRB Philadelphia working paper (2016 version), Code.

There are substantial differences in startup activity across U.S. local labor markets. We study the causes and consequences of these differences. Startup productivity shocks are found to drive much of these cross-city differences in startup activity. Examples of such shocks include breakthroughs in biotech that spurred startup formation in San Diego and Philadelphia. Overall, these shocks explain half of the forecast error variance of startup job creation, accounting for 40% of population growth and long-run changes in employment. Shocks to barriers to firm entry have economy-wide effects similar to those of startup productivity shocks but operate largely through the number of startups, rather than their size. We use a novel spatial panel VAR, identifying shocks using shift-share external instruments.

Draft paper, FRB Philadelphia working paper (2019 version), code.

New businesses are important for job creation and have contributed more than proportionally to the recent economic fluctuations. Given the risk exposure of entrepreneurs, this paper asks whether changing risk can explain the dynamics of new businesses. It makes two contributions. First, it provides a tractable, quantitative framework for analyzing business creation when entrants are exposed to idiosyncratic risk. Second, it provides conditions under which data on the size distribution of new businesses and their exit rates identifies entrepreneurial risk. According to the structural estimates, fluctuations in such risk explain around 40% of the employment fluctuations at new U.S. businesses.

Last working paper version, web appendix, code. 

Using a multi-industry real business cycle model, we empirically investigate the microeconomic origins of aggregate tail risks. Our model, estimated using industry-level data from 1972 to 2016, indicates that industry-specific shocks account for most of the third and fourth moments of GDP growth. 

FRB Philadelphia working paper.

We quantify the fiscal multipliers in response to the American Recovery and Reinvestment Act (ARRA) of 2009. We extend the benchmark Smets-Wouters (Smets and Wouters, 2007) New Keynesian model, allowing for credit-constrained households, the zero lower bound, government capital and distortionary taxation. The posterior yields modestly positive short-run multipliers around 0.52 and modestly negative long-run multipliers around -0.42. The multiplier is sensitive to the fraction of transfers given to credit-constrained households, the duration of the zero lower bound and the capital. The stimulus results in negative welfare effects for unconstrained agents. The constrained agents gain, if they discount the future substantially.

FRB Philadelphia working paper, Column on voxeu.org