The Fiscal Multiplier

This paper builds a Heterogeneous Agent New Keynesian (HANK) model – combining the standard incomplete-markets model of consumption and saving with New Keynesian price and wage rigidities, capital accumulation, and a government budget constraint specified partly in nominal terms – to quantify the fiscal multiplier for essentially any combination of realistic monetary and fiscal policy. The nominal specification of government debt is deliberate: it lets the model exploit a result (Hagedorn 2016, 2018) guaranteeing a uniquely determined price level even when the nominal interest rate is pegged, avoiding the indeterminacy problem that afflicts representative-agent New Keynesian models at a fixed rate and allowing the authors to compute a well-defined multiplier at the zero lower bound. The authors find the multiplier is highly sensitive to financing: with a pegged nominal rate, it is 1.34 (cumulative 0.55) when spending is deficit-financed but only 0.61 (cumulative 0.43) when contemporaneously tax-financed, with broadly similar values obtained in a simulated liquidity trap; once monetary policy instead follows a Taylor rule, both multipliers fall and largely converge, to 0.66 and 0.54 respectively, because the larger inflationary impact of deficit-financed stimulus triggers a correspondingly larger monetary tightening that offsets much of its extra stimulative power. Decomposing household consumption responses into an intertemporal-substitution channel and a redistribution channel, the paper traces the multiplier’s size to the interaction of market incompleteness with dynamic, forward-looking behavior: unlike in tractable two-agent (TANK) models, in which hand-to-mouth households respond only to current income, households in this model also respond to anticipated future income changes induced by the stimulus, producing materially larger multipliers than TANK models calibrated to the same current-period marginal propensity to consume – a difference the authors attribute to “dynamic anticipation effects arising in the HANK model that are absent in TANK.”

The Geography of job creation and job destruction

This paper asks why unemployment rates differ so persistently across local labor markets, and what role job creation and job destruction play in generating those differences. The authors document a comprehensive set of spatial labor market facts using administrative and survey microdata from Germany, the United States, and the United Kingdom, then build and calibrate a quantitative theoretical framework that accounts for all documented regularities.

Data and scope. For Germany, the authors use administrative data from the German employment office (universe of vacancies and unemployed, 1999–2020) and the IAB social security sample (SIAB, 2% of all workers, 2000–2017) aggregated to 194 commuting zones. For the U.S., they use BLS Local Area Unemployment Statistics (2000–2019) at commuting zones, CPS worker flows at metropolitan areas, and JOLTS vacancy data for the 18 largest MSAs (covering roughly 40% of the U.S. labor force). For the UK, they use Nomis data and Jobcentre Plus vacancy records (2004–2006) for 378 Local Authority Districts.

Iourii Manovskii | Macro Paper Warehouse

The Fiscal Multiplier

The Geography of job creation and job destruction