It has been a decade since the declining business start-up rate was first cited as a problem in Japan, and today we see a substantial number of government policies designed to promote entrepreneurships comparable to those in other industrialized countries. The difficulty in securing funding has been cited as one of the biggest obstacles in starting up new businesses. To address this problem, the government set up a new special loan program for business start-ups in January 2002 under which the National Life Finance Corporation, a government financial institution, provides up to ¥7.5 million in unsecured start-up loans for business. Previously, people wishing to set up a joint stock corporation, for instance, had to put up a minimum of ¥10 million in capital and present proof of deposit to show that they had sufficient funds in the bank. The new Law for facilitating the Creation of New Business, effective February 2003, however, allows for the formation of joint stock companies and limited liability companies with a paid-up capital of only ¥1.
The number of business start-ups utilizing such new institutional arrangements has been steadily growing. To quantitatively assess the effectiveness of these policy measures in raising Japan's business start-up rate, it will be necessary to conduct, at some point, policy evaluation based on quantitative analytical methods. In reality, however, it is extremely difficult to accurately capture business start-up rates - which is the first step in conducting such evaluation - because the premises are somewhat ambiguous. How does such absurdity arise and how can we deal with it? In the next section I discuss problems in measuring business start-up rates and future analytical challenges.
How start-up rates are determined at present
Currently, the most frequently used data in determining start-up rates for industry is the "Establishment and Enterprise Census," conducted by the Ministry of Public Management, Home Affairs, Posts and Telecommunications. The survey, a census covering all corporations and other business establishments in Japan, collects information by region and industry on the year of establishment and current status of each enterprise. In addition, corporations are asked to specify their registered year of incorporation. The rate of business start-ups at a particular point in time can thus be determined by dividing the total number of start-ups - newly established corporations plus newly established individual proprietorships - by the total number of corporations and individual proprietorships at the beginning of the period. Business start-up rates used in the government's White Paper on Small and Medium Enterprises in Japan are calculated in this manner.
However, it has been often pointed out that business start-up rates thus calculated have their limitations because of various statistical problems inherent to the Establishment and Enterprise Census.
One of the most frequently cited problems is that the Establishment and Enterprise Census is conducted only once every two or three years. A series of studies on business start-ups both in Japan and abroad has revealed that failure rates for start-up firms are very high. In the case of the Japanese manufacturing sector, for instance, 20 percent to 30 percent of start-up companies cease to exist after one year. Therefore, a census conducted once every two to three years may not accurately capture the number of new start-ups created during the survey period.
Another problem frequently pointed out is the possibility of miscounting arising from the district-by-district collection of data in the Establishment and Enterprise Census. Because the survey is conducted separately in each district, when an individual proprietorship is relocated from one district to another, it may be counted as a business closure in the district from which the business owner has moved, and as a start-up in the district it has moved into, depending on the year of the survey. In other words, when an individual proprietorship moves out and ceases to physically exist in one district, it is counted as a closure. It is not clear to what extent business start-up rates are affected by such relocations because no large-scale data based on follow-up surveys exist. What we know is that survey districts are arranged in such a way that each unit district covered by the survey has 30 establishments, and that there are a total of some 248,000 such districts nationwide. It is therefore reasonable to assume that individual proprietorships relocate from one district to another fairly often.
A business establishment is defined as a place of economic activity. In the case of online businesses, however, their existence is hardly noticeable. Someone operating a popular homepage, for example, could be making a lot of money from banner ads and remain unnoticed, even by his or her family members.
In addition to the points raised above, there are other minor problems. For instance, the definition of the term kaigyo (business start-up) used in full-fledged surveys differs from the one used in simplified surveys. However, what I would like to emphasize here is that the inability to accurately measure business start-up rates is not due to negligence or laziness on the part of government officials.
Business start-ups, by their very nature, are one of the most difficult economic phenomena to capture. Indeed, it is extremely difficult to judge whether activities by a particular person can be defined as business in the sense of being economic activities. For instance, imagine a man working as a painter: At first, no one notices his work; gradually, however, people begin to show interest and his paintings begin to sell. Then, eventually, his paintings become so popular that almost all of his works find buyers and he can make a living solely from selling his works. In this final stage, he is, without doubt, a professional painter. But at what point he actually became a professional is debatable.
Because of the inherent ambiguity in determining when certain businesses actually begin, and the high probability of closure after a short period of time, other countries also have difficulty capturing business start-up trends . As Professor David Storey, a British authority on entrepreneurship, points out:
The most important of these (statistical difficulties) is that for no developed country is there a comprehensive list of all enterprises in the economy... It is the case that the longer an enterprise has been in business, the higher is the probability of it appearing in some official records. Yet if we are interested in births of firms, these firms by definition are new, and also both very small and likely to have a short lifespan. For all these reasons they are the least likely to appear in public records (Storey, 1994).
As such, grasping the trend of business start-ups is one of the most difficult tasks in gathering statistics on economic activity.
Legislators and members of the media often criticize government officials for their failure, despite their vigorous promotion of new businesses, to provide accurate statistical data on business start-ups. However, given the aforementioned characteristics of business start-ups, and considering the cost involved, it is beyond the capability of government statisticians to determine precise trends in the development of new businesses.
Problems in utilizing tax statistics and their solution
This, however, does not mean there is no room for improvement in current methods of collecting and compiling data. One way to do so is to utilize tax statistics. The National Tax Agency publishes the "NTA Annual Statistics Report" every year, compiling statistical information collected through its operations. Included in this report are various useful data such as the number of incorporated companies, the number of legal persons, total capital invested, the rates of profitable and non-profitable corporations respectively against the total number of corporations, the status of wages in the private sector, and so on. These data, collected by taxation office staff throughout the country in the course of their regular duties, are quite reliable.
Data collected for tax administration purposes do not include specific information on business start-ups, but it is possible to grasp the trend of business start-ups through such data. For corporate taxation purposes, newly established corporations (i.e., domestic ordinary corporations, etc.) are required to file a notification of incorporation with the relevant taxation office within two months of the registered date of incorporation. Likewise, individuals who start up a business, or who open, add to, relocate, or close their office, are required to report to the relevant tax office within one month.
If data relating to business start-ups collected during the process of tax administration are incorporated into the NTA Annual Statistics Report, more accurate information on business start-ups will become available.
The advantages of tax statistics are not limited to their quality. By utilizing tax-related data, information on business start-ups can be collected annually or even monthly, along with information on relocations. Business entities that are harder to spot, such as those engaged in small-office home-office (SOHO) work, can also be covered. Thus, most of the problems raised with regard to the Establishment and Enterprise Census can be resolved.
In the United Kingdom, for instance, business formation is tracked by checking the number of companies registered for value-added tax (VAT). In the United States, business start-up rates are calculated based on the "Statistics of U.S. Businesses" published by the U.S. Census Bureau. At the same time, separately compiled data based on tax statistics are utilized for various types of research. In Germany, applications filed with each state for permission to set up a business are used as basis to calculate business startup rates. Thus, each country efficiently utilizes information obtained in the course of implementing its own administrative procedures.
The general perception in Japan is that tax statistics should be used solely for tax administration purposes. But isn't it possible to make an exception, at least with regard to those data related to business startup ratios? This would enable us to develop a more accurate picture of the status of business formation in Japan.
