Research

Working Papers

Conditional Initialization for Enhanced Inference in Nonlinear State-Space Models

This paper advances inference in nonlinear state-space models by deriving the distribution of initial endogenous states conditional on initial observations, addressing the general unknown nature of the unconditional distribution. Together with the inversion filter, this allows the likelihood to be evaluated deterministically and conditional on the initial observations. Such conditional likelihoods are commonly used in practice, e.g., when estimating autoregressive processes via least squares. Monte Carlo studies demonstrate that employing the conditional distribution of endogenous states can significantly enhance both Frequentist and Bayesian inference concerning states and parameters. The paper illustrates the method’s practical relevance with two applications: first, a Monetary Business Cycle Accounting analysis of the COVID-19 recession and subsequent inflation surge in the US and Eurozone using a global solution; second, an evaluation of uncertainty shocks on the US business cycle during the Great Moderation via a third-order perturbation method. Additionally, the paper discusses limitations of the inversion filter and proposes remedies, particularly for spurious many-to-one policy function mappings that arise with higher-order perturbation solutions.

A “Marginal” Tale of Two Germanies: Accounting for the Systemic Divide [DP](with Vasilij Konysev)

The comparative economic performance between the former socialist German Democratic Republic (GDR) and the capitalist Federal Republic of Germany (FRG) remains inconclusive due to valuation problems. We address these problems by applying wedge-growth accounting to a newly compiled dataset. More precisely, we compare the allocation efficiency using wedges between marginal utility and productivity, as well as Total Factor Productivity (TFP) growth. Wedges in marginal utility account for binding quantity constraints in GDR’s goods and FRG’s labor market. We analyze the resulting unitless wedges and swap them in an equivalent growth model for the two Germanies to quantify their impact on output and economic welfare. The analysis reveals that the consequences of GDR’s rationing were multiple times more drastic than FRG’s unemployment. An observed faster output growth in the GDR stems from excessive labor input—depressing consumption-based welfare by a fourth—rather than from physical capital or TFP. Instead, GDR’s economic activity fell comparatively ten years further behind due to lower TFP growth. Lastly, persistent, substantial net inflows increase GDR’s welfare by 25 %.

To converge or not to converge:
Accounting for the German reunification
[DP] (with Vasilij Konysev)

German reunification in 1990 marked the first sudden integration of a socialist and capitalist economy. Despite East Germany’s (EG) economic catch-up with West Germany (WG), the integration remains unfinished, as indicated by per capita output in EG still being about one-third lower. To study this unfinished regional convergence, we apply wedge-growth accounting using a human capital-augmented, two-sector, two-region model, incorporating labor supply constraints to capture key qualitative differences between EG and WG. Our findings show that sectoral labor and capital wedges are similar within regions and have significantly converged between regions, with EG initially overusing inputs. While productivity in the nontradable goods sector has fully converged, the tradable sector in EG remains less productive than in WG. Counterfactual analysis suggests that this productivity gap, together with persistent net inflows to EG, explains EG’s lower economic activity. However, reducing the inflows would result in significant welfare losses in EG. Furthermore, we account for the reunification event, identifying a substantial productivity catch-up in EG between 1989 and 1991. Our findings offer clear policy insights, highlighting the trade-offs between economic activity and fiscal transfers.

Hone the Neoclassical Lens and Zoom in on Germany’s Fiscal Stimulus Program 2008-2009 (with Johannes Huber) [version: Business cycle accounting for the German fiscal stimulus program during the Great Recession].

Business Cycle Accounting (BCA) by Chari, Kehoe, and McGrattan (2007, Econometrica) completes the “…through the lens of a neoclassical model”-approach. This paper refines and extends the methodology in four primary dimensions, creating a manual. i) the choice of the level of aggregation is critical and thus must be case-dependent. ii) a strict distinction between growth and cycle is beneficial. iii) BCA requires Maximum-Likelihood, even if it is difficult. Given these difficulties, we introduce a procedure that reliably and quickly locates the maximum and enables a detailed evaluation of the likelihood function and robustness checks. iv) it is revealing to discuss the results in the context of economic and political events. To illustrate the necessity and benefits of the refinements, we apply BCA to the Great Recession in Germany. The main driver was efficiency, followed by net exports and distortions in the markets for business investments. Government consumption and durable consumption acted counter-cyclically. We attribute the latter to a high subsidy for new cars or, more generally, for durables.

Refereed publications

Hedging Against Inflation: Housing versus Equity [open access][code files], Empirical Economics 65 (6) (2023), December 2023, pp. 2583 – 2626.

Housing and the Business Cycle Revisited [accepted manuscript][online appendix], Journal of Economic Dynamics and Control 99 (2019), February 2019, pp. 103-115.