«Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.8, No5, November 2012 563 Why Did the Debt Maturity of the Japanese ...»
Table 3 shows the estimation results of equation (1) for two groups of Japanese firms, manufacturing and non-manufacturing industries. First, if we look at the regressions for the total debt growth, we find quite remarkable differences between the manufacturing and non-manufacturing industries. The estimated coefficient of ROA is positive and significant at 10% for the manufacturing industries, but is negative and significant at 1% for the non-manufacturing industries. Thus, for non-manufacturing industries, it is clear that the firms performing better have been reducing their debt more aggressively.
When looking at the coefficients of capital investment, they are positive for both industries at 1% level, but the size of the estimated value of the non-manufacturing industry was about nine times larger than that of manufacturing industry. The constant term in the non-manufacturing regression is also much larger than in the manufacturing regression. Overall, explanatory power of the regression as a whole is much higher for non-manufacturing with R2 of 61 %.
Let us move on to the estimation results for the long-term borrowings in panel (2). The main focus of this paper is the relationship between the firm’s debt growth and its business performance. As the explanatory power of the total debt growth regression for manufacturing is rather low compared with the regression for nonmanufacturing, it will be safe to say that influences of factors other than ROA and capital spending are much more important in manufacturing industries. This will explain why adding the total debt growth as an explanatory variable greatly increases the performance of the regression for the long term borrowings. Also, the coefficient of ROA becomes negative and statistically significant when total debt is added to the explanatory variables. This is consistent with the estimation results in Table 3 for which the all industries are pooled.
In contrast to that, in non-manufacturing industry sample, the coefficient of the ROA in the total debt growth regression is statistically significant in a large negative value and the performance of the regression as a whole is also high. Thus, while ROA is not statistically significant in the regression including total debt growth as an explanatory variable, it is Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.8, No5, November 2012 577 conceivable that ROA reduces long-term bank borrowings by reducing total debt.
As turning to the estimation results of the short-term borrowing regressions in Panel (3), for the manufacturing industry sample, the coefficients of ROA are positive and statistically significant regardless total debt was included or not as an explanatory variable. Hence, positive correlation between the increase in short-term debt and the corporate performance is quite clear.
The discussions on the estimation results reported in Table 3 so far are consistent with the empirical results for the overall industry sample in Table 2. They are also consistent with the signaling story of the debt maturity choice. However, in the estimation results for short-term borrowings of non-manufacturing industries, ROA has a negative effect at 5% in the estimation without total debt growth. When total debt is added, the estimated coefficient of ROA becomes positive, but it is not statistically significant. Overall, whether the correlation between ROA and the short-term borrowings is significantly negative or not is not clear. However, there is no doubt that the relationship between short-term debt and corporate performance in manufacturing industry is quite different from that of for all industries and manufacturing industries.
Finally, let us focus on the estimates of the constant term. Again all estimated coefficients in the long-term borrowing regressions have positive sign and statistically significant, except for the regression for non-manufacturing with total debt as an explanatory variable.. As a contrast, the estimated constant terms are negative and statistically significant in most of the cases for short-term borrowings. Even for the only exception, the regression for the non-manufacturing industries that does not include total debt, the absolute value of the parameter is less than the estimation with long-term borrowings of the same specification. Therefore, it is safe to argue that a shift to long-term borrowings from a short-term borrowings has been induced by some factors other than the corporate performance measured by ROA.
Though these empirical results are not fully satisfactory, the following statement can be made. First of all, it has been confirmed that the industries that had better performance measured by ROA had decreased more debt among non-manufacturing industries at least for the sample of Japanese firms since mid-1980’s examined in this study. This means that in the non-manufacturing sector in recent years, if the high quality borrower firms wanted to differentiate themselves by sending a signal to the lenders, they should reduce their debt more aggressively, because it would be more difficult for the lower quality company to do so.
The remaining discussions admittedly involve some speculations if the reduction of debts were working as an important signaling device, it is not necessary for the firm to send an additional signal by choosing the maturity length of the debt. If this is true, it is quite a sensible strategy for the firm to cut the short-term debt to increase the bargaining power against financial institutions, by reducing the frequency of refinance or “debt roll over”.
However, the estimation results on the growth rate of short-term and long-term borrowings shown in Table 2 and 3 suggest that their growth rates are almost proportional to the growth T Iwaisako / Public Policy Review rates of total debt. However, the firms with higher ROA are apt to use short-term debt for fundraising or less aggressive in reducing short-term borrowings. After all, if a company cannot distinguish itself from lower quality rivals by the ability to reduce total debt, it will turn to the debt maturity choice as an additional signal. This argument will be consistent with the signaling model of debt maturity choice suggested by Flannery=Diamond.
On the other hand, the difference in estimated constant terms of short-term and long-term borrowing regressions suggest that the factors other than corporate performance has been a major influence for the Japanese firms in increasing their average debt maturity. For this regard, the analysis in this section leaves a large open question and it should be examined in more details in the future.
Since mid 1970’s, the Japanese firms had reduced the proportion of debts and borrowings in their capital structures. Trend of increasing debt maturity, measured by the ratio of fixed debt to total debt, parallels with this movement. In this paper, first, I carefully reconfirmed such increasing trend of debt maturity of the Japanese non-financial corporations. Then, focusing on bank borrowings from financial institutions rather than the debt amount as a whole, I carefully examined the mechanism behind increasing debt maturity from the viewpoint of the signaling model of the financial intermediation.
First, I examined the possibility that the low interest rate in recent years can explain the increase of average debt maturity. By looking at the data on the spreads between deposit and loan interest rates and also by looking at the long-short spread, it is concluded that this explanation is mostly incompetent.
Next, using FSSCI data, empirical analysis is carried out to examine if the story that both total amount of debt and the debt maturity choice serve as signaling devices of borrower firms, an extension of the signaling model of the choice of debt maturity such as Flannery and Diamond. At least for the non-manufacturing industry sample, it has been confirmed that there is a clear correlation between corporate performance and the reduction of corporate borrowings. On the other hand, when the growth rates of the total debt are at the similar level, the industries that have better performance measured by ROA are more prone to increase or less aggressive to reduce the short-term borrowings. Since the shorter debt maturity is more costly for the firms of poor performance, it is argued that the maturity length plays the role of signal as the theoretical model of Flannery=Diamond suggest.
The result of the empirical analyses in this paper is consistent with the static model of the debt maturity choice by Flannery=Diamond, when we look at the relationship between the growth rate of the short/long-term debt and ROA alone. However, when the growth rate of total debt is taken into account, significant parts of the variations in the growth rates of short-term Policy Research Institute, Ministry of Finance, Japan, Public Policy Review, Vol.8, No5, November 2012 579 and long-term debts remain unexplained in our regression results. Therefore, although we believe this paper has made some important contributions, the reason why Japanese industries actively increased its debt maturity should be a subject of more comprehensive future research.
In addition, this paper used a panel data of industries. In future studies, the panel data of the individual firms should be examined.
T Iwaisako / Public Policy Review
Cai, J., Y. Cheunga, and V. K. Goyal (1999) “Bank Monitoring and Maturity Structure of Japanese Corporate Debt Issues”, Pacific-Basin Finance Journal 7:3-4, 229-250.
Dawkins, R. (1989) The Selfish Gene, 2nd ed., Oxford: Oxford University Press.
Diamond, D. W. (1991) “Debt Maturity Structure and Liquidity Risk”, Quarterly Journal of Economics 106:3, 709-737.
----- (1993) “Seniority and Maturity of Debt Contracts”, Journal of Financial Economics 33:3, 341-368.
Flannery, M.J. (1986) “Asymmetric Information and Risky Debt Maturity Choice”, Journal of Finance 41:1, 19-37.
Fukuda, S. and S. Kei, (1996), “Asymmetric Information, Agency Problems, and Bank Loan Maturity in Japan (in Japanese, Jyo-hou no hi-taisyou-sei to chouki shikin-no yu-shi hiritsu)”, Economic Research (Keizai Kenkyu) 47:3, pp.204-216.
Hoshi, T. and Anil K. K. (2001) Corporate Financing and Governance in Japan: The Road to the Future, MIT Press.
Iwaisako, T. (2010), “Aggregate Corporate Saving and Recent Changes in Capital Structures of Japanese Firms (in Japanese, Macro-no kigyou-chochiku to nihon-kigyou no shikin chotatsu-no doukou)”, Economic Research (Keizai Kenkyu) 61:1, pp.18-32.
Iwaisako, T., (2012), Financial Behaviors of Households and of Firms in Japanese Economy – Micro Structural Change and Macroeconomic Implications (in Japanese, Kakei-kigyou-no kinyu-koudou to nihon-keizai), Nikkei Publishing Inc.
Miwa, Y. (2010), “A Study of Financing Behavior of Japanese Firms with Firm-Level Data from Corporate Enterprise Quarterly Statistics – 1994-2009 (in Japanese, Hou-jin kigyou toukei kihou kohyou wo mocha-ita nihon-kigyo no shikin-choutatsu-koudou no kenkyu 1994-2009)”, CARF Working Paper J-063.
Uchida, H. (2010), Economic Analysis of Financial Functions and Banking Business (in Japanese, Kinyu-kinou to ginkou-gyo no keizai-bunseki), Nikkei Publishing Inc.
Watanabe, H. and Uesugi I., (2008), Empirical Analysis of Mid-Small Firms’ Finance –
Reexamination of Prevailing View (in Japanese, Kensyo chyusho-kigyou-kinyu: