Micro Data Analysis of Information System Investment in IT Industiy

本文首先利用信息通信产业的面板数据，运用随机前沿方法对IT资本生产率进行了测算。结果发现，在假设总收入和总增加值被定义为采用柯布-道格拉斯型生产函数的产品的情况下，IT资本的弹性在2000年至2003年期间为正。此外，硬件生产率远远大于软件生产率。其次，利用数据包络分析法对互联网服务提供商（ISP）的效率进行了估计。因此，地方城市的ISP效率相对低于大都市地区和全国各地。采用Wilcoxon秩和检验对两组进行检验，发现两组有效率差异在5%水平上具有显著性。第三，我们分析了生物技术产业，食品加工，化学生产和医疗生产，以及它们公布的财务报告。我们将拥有较小软件资产的组与拥有较大软件资产的组进行了划分。这一小组的资产平均为300万日元，中位数为200万日元。相反，大型集团的软件资产平均为600万日元，中位数为500万日元。此外，封闭软件资产信息的企业的标准差也大于披露软件资产信息的企业。第四，利用财务报告交换市场，估计了日本银行业IT资产的托宾q值。结果，对于被不良贷款贬值的银行贷款，托宾q值的估计值为1.1或1.5。此外，信息系统资产的托宾q值为9.0或9.5，远高于1.0。

At first we started to estimate IT capital productivity with stochastic frontier approach using panel data on information and communication industry. As a result of it, we found out that the elasticity of IT capital was positive during the period from 2000 to 2003 under the assumptions that total revenue and/or total value added were defined as product employing Cobb-Douglas type production function. Moreover, the hardware productivity was much greater than the software productivity. Second, we estimated the efficiency of internet service provider (ISP) with Data Envelopment Analysis. As a result of it, ISP efficiency in local cities was relatively lower than the metropolitan areas and all over the nation. We examined these two groups with Wilcoxon rank-sum test, and found out that the difference of efficiency level between these two groups was significant on five percent levels. Third, we analyzed biotechnology industry ; food processing, chemical production, and medical production, with their published financial reports. We divided the group having smaller software asset with the group having larger software asset. The small group's assets had average of three million yen and median of two million yen. On the contrary, the large group's software asset had average of six million yen and median of five million yen. In addition to it, the standard deviation of the firms which closed software asset information was greater than the firms which disclosed it. Forth, we estimated Tobin's q of IT asset in Japanese banking industry using financial reports to exchange markets. As a result of it, the estimated value of Tobin's q was 1.1 or 1.5 with regard to the banking loans depreciated by nonperforming ones. Moreover, the Tobin's q of information system asset was 9.0 or 9.5 which was much higher than 1.0.