Recently, survey data analysis has emerged as a new method for testing hypotheses and for clarifying the relative importance of different factors in corporate finance decisions. This paper investigates the advantages and drawbacks of survey data analysis, methodology of survey data analysis such as questionnaire design, and analytical methods for survey data, in comparison with traditional large sample analysis. We show that survey data analysis does not replace traditional large sample

more »

... but complements it and derive a new theory based on new realistic assumptions which are consistent with the results of survey data analysis. Introduction In the area of corporate finance, there are many competing theoretical models and hypotheses based on these models. In the area of payout policy alone, hypotheses include free cash flow, the pecking order, the lifecycle, signaling, catering, etc. One of the reasons for this proliferation is that none of the models have great explanatory power regarding firms' financial decisions. Consequently, there are many puzzles and unexplained facts. Recently survey data analysis has been used to test hypotheses in corporate finance. Three papers by Graham and Harvey(2001), Brav et.al. (2005), and Graham-Harvey-Rajgopal(2005) -established a method for testing hypotheses in corporate finance. They are very impressive papers and each won honors. 1 Before them, there were several papers using survey data analysis but they were small surveys and/or had simple questions which did not permit the testing of hypotheses or the elucidation of important factors. The three papers by Graham and Harvey are different from past survey researches in that they are more rigorous in testing hypotheses and explaining managers' motives in financial decisions. Recently we conducted two surveys for Japanese firms. One is Serita-Hanaeda( 2007 ), a survey on perceptions of stock price and financial actions particularly on stock splits and on changing the minimum unit of one trade(kukuri-naoshi). The other is Hanaeda-Serita(2007), a survey on the payout policies of Japanese firms comparable to the U.S. firms by Brav et al.(2005) . From the two surveys we learned much about survey data analysis.