2009/12/04 by Hiroshi Iyetomi, Yasuhiro Nakayama, Iyetomi, Hiroshi +11
Economics, Econometrics and Finance · #Computational Finance (q-fin.CP) #Economic Growth and Productivity #FOS: Economics and business #General Finance (q-fin.GN) #Global trade and economics #Monetary Policy and Economic Impact
paper · pdf · doi:10.48550/arxiv.0912.0857
openalex publication_date 2009/12/04 · openalex created_date 2025/10/24 · openalex updated_date 2026/07/28
We explore what causes business cycles by analyzing the Japanese industrial\nproduction data. The methods are spectral analysis and factor analysis. Using\nthe random matrix theory, we show that two largest eigenvalues are significant.\nTaking advantage of the information revealed by disaggregated data, we identify\nthe first dominant factor as the aggregate demand, and the second factor as\ninventory adjustment. They cannot be reasonably interpreted as technological\nshocks. We also demonstrate that in terms of two dominant factors, shipments\nlead production by four months. Furthermore, out-of-sample test demonstrates\nthat the model holds up even under the 2008-09 recession. Because a fall of\noutput during 2008-09 was caused by an exogenous drop in exports, it provides\nanother justification for identifying the first dominant factor as the\naggregate demand. All the findings suggest that the major cause of business\ncycles is real demand shocks.\n