Tracking down the business cycle: A dynamic factor model for Germany 1820–1913

Abstract

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in AbstractWe use a Bayesian dynamic factor model to measure Germany's pre World War I economic activity. The procedure makes better use of existing time series data than historical national accounting. To investigate industrialization we propose to look at comovement between sectors. We find that Germany's industrial sector developed earlier than stated in the literature, since after the 1860s agricultural time series do not comove with the business cycle anymore. Also, the bulk of comovement between 1820 and 1913 can be traced back to five out of 18 series representing industrial production, investment and demand for industrial inputs. Our factor is impressingly confirmed by a stock price index, leading the factor by 1-2 years. We also find evidence for early market integration in the 1820s and 1830s. Our business cycle dating aims to resolve the debate on German business cycle history. Given the often unsatisfactory quality of national accounting data for the 19th century we show the advantage of dynamic factor models in making efficient use of rare historical time series.

Irish GDP between the Famine and the First World War: estimates based on a dynamic factor model

“…Dynamic factor models have been used previously in the estimation of business cycle fluctuations in both contemporary (Stock and Watson 1989) and historical (Sarferaz and Uebele 2009;Ritschl et al 2016) contexts. The basic idea is that a time series is likely to be influenced by one or potentially more common factors as well as an idiosyncratic component.…”

“…This paper proposes a solution to the standstill, which can be applied in other contexts in economic history for which similar conditions prevail. Building on the business cycle literature that identifies the cycle using factors models (Sarferaz and Uebele 2009;Ritschl et al 2016), we develop a two-stage approach to estimate GDP. In the first stage, a dynamic factor model is estimated to identify the common movement in a set of key macroeconomic variables.…”

The U.S. Business Cycle, 1867–2006: A Dynamic Factor Approach

“…To our knowledge, this approach was first employed for the long term by Gerlach and Gerlach-Kristen (2005) for Switzerland between the 1880s and the Great Depression of the 1930s. Sarferaz and Uebele (2009) employ dynamic factor analysis to obtain an index of economic activity for nineteenth-century Germany, comparing it to different chronologies based on reconstructed national accounts.…”

Output growth in German manufacturing, 1907–1936. A reinterpretation of time-series evidence

“…Second, my approach is closely linked to multivariate dynamic factor models that estimate co-movement across various observed series. This research avenue has been explored by Sarferaz and Uebele (2009) and Uebele (2011) in an attempt to track down business-cycle movements in pre-WWI Germany. To the best of my knowledge, this study is the first in the field of economic history to combine the two by using the state-space form in a multivariate setting for estimating common components.…”