On the Fund Raising by Japanese Firms Using Commercial Paper Issues

这项为期三年的研究计划分析了：(1)商业票据（OP）问题的决定因素；(2) OP事项公告的影响；(3)与事件研究相关的计量经济学问题。OP问题的决定因素。1989年对日本公司行为的分析发现，利用商业票据筹集资金的决定取决于公司规模；破产的概率；主要银行关系的存在；还有现金流。考虑到发布资格标准和忽略这些标准的CP问题分析结果的比较表明，资格标准对某些公司具有约束力。OP问题公告的影响。对《日本经济新闻》1987年至1997年期间的调查显示，有23家公司宣布将发行新的cp。利用这23家公司的数据，进行了一项事件研究，以确定新OP计划的公告对发布公告的公司价值的影响。发现这些公告与任何统计上显著的异常收益无关；观察到的产生的回报与：fin-n是否有一家主要银行；公司债券的评级；或者是宣布的时间。事件研究的计量经济学问题。在包括上述事件研究在内的事件研究中，通常使用两步估计方法，但这会导致对感兴趣的参数的估计效率低下。对于所有解释变量都是可观测的情况，提出了一种有效且计算简单的估计这些参数的方法。当解释变量包含生成的回归量时，对其带来的计量经济学问题进行了解释。

This three-year research program has analysed : (1) the determinants of commercial paper (OP) issues ; (2) the impact of announcements of OP issues ; and (3) econometric problems associated with events studies.1. Determinants of OP Issues. An analysis of the behavior of Japanese firms in 1989 found that the decision to raise funds using commercial paper depended on firm size ; the probability of bankruptcy ; the presence of a main bank relationship ; and cash flows. A comparison of the results obtained by analysing CP issues taking account of issuing eligibility criterion and ignoring those criterion suggests that the eligibility criterion were binding for some firms.2. impact of the Announcement of OP issues. A search of the Nihon Keizai Shinbun over the period 1987 to 1997 revealed 23 firms that made announcements that they would be making new issues of CPs. Using the data on these 23 firms, an event study was conducted to determine the impact of announcements of new OP programs on the value of the firm making the announcement. It was found that these announcements are not associated with any statistically significant abnormal returns ; and that the observed returns generated are not related to : whether the fin-n has a main bank ; the rating of the firm's bonds ; or the timing of the announcement.3. Econometric Problems of Event Studies. In event studies including the one described above, it is typical to use a two step estimation method but this leads inefficient estimates of the parameters of interest. An efficient and computationally simple method for estimating these parameters is proposed for the case when all the explanatory variables are observable. When the explanatory variables include generated regressors, the econometric problems they bring about are explained.