SME Financing in Japan: Characteristics and problems, predominance of low-interest rate loans
Senior Fellow, RIETI
Associate Professor, Keio University
Following the global financial crisis, although overall business conditions for Japanese small and medium-sized enterprises (SMEs) became extremely difficult, relationships between SMEs and their creditor banks have been relatively stable. Yet, it has long been said that Japanese banks are not making sufficient efforts to improve the efficiency of financing to SMEs. In this article, we focus on this longstanding problem with SME financing in Japan. We also discuss what factors make it difficult for Japanese banks to cultivate new customers (borrowers) with medium risk, a market segment targeted by the now-defunct Incubator Bank of Japan.
♦ ♦ ♦
The current recession has been extremely tough and Japanese companies have been in a tight cash-flow position due to declining sales. Even so, banks have not hardened their lending attitude as much as one might have expected. The degree of credit tightening has been fairly limited, particularly in comparison with situations in the United States and Europe, where banks rapidly tightened credit standards for loans as shown by several indicators.
A questionnaire survey conducted in February 2009 by the Research Institute of Economy, Trade and Industry (RIETI), in which we asked companies about changes in the lending attitude of their creditor banks, found that only slightly more than 10% of respondents experienced a drastic deterioration in the lending attitude of at least one of their creditor banks following the crisis, mostly in cases where they had short-term transactions with megabanks. Indeed, the outstanding balance of SME loans turned upward at the end of 2008 or in the first half of 2009 at credit unions (i.e., shinkin banks and cooperatives) and regional banks (i.e., first- and second-tier regional banks). Thus, apart from a limited number of exceptional cases, it is unlikely that the credit squeeze was caused by factors attributable to banks, such as the accumulation of bad loans and losses on securities holdings.
In the latest financial crisis, Japanese banks suffered relatively small losses compared to their U.S. and European counterparts. This, with the help of government liquidity support measures, enabled Japanese banks to function, to some extent, as a steady supplier of funds to companies. However, whether or not they have done this efficiently is another story.
From 2000 onward, banks and the government have taken various initiatives to facilitate greater efficiency in the supply of funds to SMEs while accelerating the process of bad loan disposals. Perhaps the most important element of these initiatives is the promotion of relationship banking. When making decisions on lending to SMEs that do not provide reliable financial information, hard-to-quantify information (soft information) such as the quality of management and viability of their business is very important. Banks can obtain such soft information only through their close, long-term relationships with SMEs. Lending practices that take advantage of such relationships are referred to as relationship banking.
♦ ♦ ♦
Do banks actually collect soft information on SMEs and effectively use it in their decision-making for lending and other undertakings?
Before answering this question, let us be clear about one characteristic of Japanese banks: they maintain longer relationships with their corporate borrowers and are better placed to collect soft information than their counterparts in other countries. The question is whether or not Japanese banks actually lend on the basis of soft information and, if they do, whether such lending practices represent widespread behavior of banks in general. At present, the answer to this question appears to be no. Japanese banks set most of their loans at low interest rates, although they could fully utilize soft information on individual borrowers and adjust these rates greatly according to a careful assessment of credit risk.
Throughout the 2000s, Japanese banks' average contractual interest rates on their loans (outstanding balance-based) remained low and were concentrated on the lower end. During the same period, the percentage of the amount financed at relatively high interest rates of 3% to 5% continued to diminish, from 14% at the end of 2000 to 7% at the end of 2009. Moreover, the percentage of funds loaned at 5% or higher interest rates, which accounted for a mere 1.6% at the end of 2000, dropped even lower to 1.0% by the end of 2009.
Although the above figures include loans to large firms and general consumers, and not just SMEs, interest rates of 5% or more are similar to those applied to unsecured loans by the Incubator Bank of Japan and other banks that newly entered the lending market for SMEs in the 2000s.
It is not clear whether those banks' target market segment (defined as medium-risk borrowers) represents only a small fraction of the whole or whether more established banks have offered low interest rates to relatively high-risk companies. In either case, as the percentage of companies that borrow without security has increased by almost 10% in the last few years, as well as the present competitive environment surrounding financial institutions and the low profitability of companies in general, it must have been very difficult to find borrowers who would accept interest rates above 5%, even without security.
The predominance of low-interest rate loans with the reduced percentage of the amount financed at higher interest rates suggests that soft information retained by banks is not used when making credit decisions or setting interest rates. This is because, as indicated in a 2009 study by Professor Hans Degryse of Tilburg University of the Netherlands and others, interest rates tend to diverge when banks make credit decisions or set interest rates for SMEs without having sufficient management information (hard information), since most of such decisions are left to the discretion of banks' loan managers.
What is necessary for Japanese banks to start lending based on the active use of hard-to-quantify, non-financial soft information? To answer this question, an empirical analysis using loan contract data in other countries yields useful insight.
The 2009 study on an Argentine bank by Associate Professor Jose Liberti of Tilburg University and others indicated that the "distance" between a bank's lending personnel working directly with clients and head office influences the degree of use of soft information in credit decision-making. In other words, a flat structure within a bank, that is, close proximity between the field and the headquarters, encourages the use of soft information. As this distance increases, there are fewer incentives for collecting soft information in the field, resulting in inaccuracies in client information and leading the bank to rely more on hard, financial information in making credit decisions. The study concludes that delegating to staff in the field effectively promotes the use of soft information.
In applying this finding to Japanese banks, it is necessary to design a mechanism to effectively exploit the advantages associated with delegation to the filed (branches) and, at the same time, minimize the disadvantages. At present, most of the lending decisions of Japanese banks are made by the loan application examination division in their head office. A questionnaire survey conducted by RIETI of mainly credit unions found that about three quarters of them make decisions on over 60% of their loans to SMEs (in terms of the amount financed) at their head office. Probably for this reason, they usually use soft information in drawing up consensus-forming circulars concerning lending proposals and revising internal credit ratings, but rarely in setting interest rates. Greater delegation to branches is thus expected to encourage lending based on the effective use of soft information.
However, in banks where the performance of lending personnel is mainly evaluated by the number of new loan contracts and other quantitative results, further delegation could cause an expansionary race among personnel to win more contracts and financing while ignoring soft information. To minimize this risk, it is necessary to devise measures such as considering profitability when evaluating the performance of lending personnel and considering problems with borrowers that occur soon after a change of personnel when evaluating preceding personnel.
♦ ♦ ♦
To realize both efficient financing and improved loan application screening systems of banks, it is necessary to carefully examine the efficacy of measures that the government applies to the lending market. Following the financial crisis, a number of governmental measures including the Emergency Credit Guarantee Program have facilitated an abundant supply of funds, while little attention has been paid to the question of whether these funds are channeled into areas that are sure to generate long-term high profitability.
It is important to assess whether these emergency measures have actually achieved an efficient flow of funds from the standpoint of medium- and long-range institutional design. A 2010 study by Uesugi and others on the efficacy of a special guarantee provided 10 years ago on a scale similar to that of the present emergency measure, found that financing has improved for firms that took advantage of the system but ex-post performance has not always improved, except for firms with a high ratio of net worth to total capital.
Governmental involvement greatly affects the behavior of banks that lend at their own risk. It is important to thoroughly discuss and clarify who will take charge of the efficient supply of funds, and how, when the current emergency measures introduced by the government expire.
* Translated by RIETI.
September 14, 2010 Nihon Keizai Shimbun
February 7, 2011
Article(s) by this author
April 20, 2018［Newspapers & Magazines］
August 23, 2017［Newspapers & Magazines］
July 3, 2017［Column］
Rethinking the Government's Policy for Small and Medium-sized Enterprises: Make a clear distinction between emergency measures and support measures in normal times
August 10, 2015［Newspapers & Magazines］
July 3, 2015［Column］