ABSTRACT
In this article, we investigate Granger’s causality among the exchange rate, Tobin’s marginal q and investment-capital ratio with quarterly data on firms categorized as small and medium-sized enterprises (SMEs) in Japan. We utilize the data of the following industries: chemical, iron and steel, production machinery, electric machinery and equipment, automobile and accessories and all manufacturing. The empirical results we obtain show that the null hypotheses of no Granger’s causality from the exchange rate to other variables, investment and marginal q, are accepted in all industries. The reason is the lower percentage of export by SMEs. Thus, it is considered that the effects of Abenomics on SMEs are limited, and that other kinds of reflation measures for SMEs are needed.
Acknowledgements
The author would like to thank Masaya Ishikawa (Tokyo Keizai University). He gave some comments for the author's other article, but these also helped the author to improve this article. But remaining errors are author's responsibility.
Notes
1 The English version is downloadable from following URL: http://www.chusho.meti.go.jp/pamflet/hakusyo/H26/download/2014hakusho_eng.pdf.
2 The data on investment and Mq are seasonal adjusted by X11.
3 According to the Small and Medium Enterprises Agency, the definition of an SME is that (1) the number of employees is not more than 300 or (2) capital is not more than three thousand million yen in the case of manufacturing. However, there is no category of ‘capital of not more than three thousand million yen,’ and we utilize the manufacturing industry’s data on ‘capital of not more than one thousand million yen.’
4 The English version of this the White Paper is downloadable from following URL: http://www.chusho.meti.go.jp/pamflet/hakusyo/h20/h20/2008hakusho_eng.pdf.
5 By calculating with the rate of retirement of equipment printed in the Quarterly Report on National Accounts of 1980 (Economic Planning Agency of Japan), the vintage of manufacturing industries in 1980 is 8.3 years. Then, we calculate with the rate of retirement of equipment printed in the Quarterly Report on National Accounts of 2010.The data series of 93SNA is available only after 1980. And we thus calculate the vintage with 68SNA, which is available before 1980 and connect to 93SNA at 1980:Q1.