Tracking Down the Business Cycle: A Dynamic Factor Model For Germany 1820-1913

We use a Bayesian dynamic factor model to measure Germanys pre WorldWar I economic activity. The procedure makes better use of existing time seriesdata than historical national accounting. To investigate industrializationwe propose to look at comovement between sectors. We find that Germanysindustrial sector developed earlier than stated in the literature, since after the1860s agricultural time series do not comove with the business cycle anymore.Also, the bulk of comovement between 1820 and 1913 can be traced back tofive out of 18 series representing industrial production, investment and demandfor industrial inputs. Our factor is impressingly confirmed by a stockprice index, leading the factor by 1-2 years. We also find evidence for earlymarket integration in the 1820s and 1830s. Our business cycle dating aims toresolve the debate on German business cycle history. Given the often unsatisfactory quality of national accounting data for the 19th century we show theadvantage of dynamic factor models in making efficient use of rare historicaltime series.