The cost of capital: a summary of results for the U.S. and Japan in the 1980s.
1 See footnotes and references at end of text.
The cost of capital for the U. S. and Japan was an intensively studied and widely debated topic during the 1980s. This article summarizes the results of much of the pertinent literature that has appeared in recent years. It also calculates a broad range of estimates, within which the real cost of capital for the U. S. and Japan might have fallen in the 1980s, employing a skeletal model under a variety of key assumptions. The conclusion is that the real cost of financial capital was higher for the U. S. than for Japan during much of the past decade.
THE CHRONIC balance of trade deficit that the United States ran with Japan during the 1980s prompted many researchers to search for factors that might underlie the apparent reduced competitiveness of the United States. Other than exchange-rate distortions and lower productivity growth, one often-cited explanation is the possible Japanese advantage in tapping inexpensive investment funds. In fact, for some time now, a widespread perception has prevailed, unchallenged in most business [see, for example, Richman (1989) and Drucker (1990)] if not academic circles, that the real cost of financial capital (RCFC) is substantially higher in the U.S. Indeed, one rough, rule-of-thumb calculation has put the U.S. real cost of financial capital at two to three times that in Japan.(1) This estimate was based on research done by Hatsopoulos (1983).(2)
Since then a great deal of research has been completed on this subject. This article attempts to summarize results of the pertinent literature regarding the range of estimates of the U. S. versus the Japanese RCFC under a wide variety of assumptions.
A review of pertinent literature suggests that there was no clear-cut winner in this matter. Neither country had a lower capital cost over all comparable investment projects and time periods.
A few studies concluded that the Japanese were disadvantaged regarding cost of capital. Ando and Auerbach (1985), for example, found that the market rate of return to equity was much higher in Japan (13.6 percent for the median of their sample of firms, versus 2.2 percent for the U.S. firms) [see Frankel (1990) p. 11]. In the same vein, Kester and Luehrman (1989) argued that extensive cross share-holdings among Japanese firms causes serious underestimation of the cost of equity in Japan and that the difference in real interest rates favored the U. S. as often as Japan between 1976 and 1985. They concluded that their evidence does not support the notion that Japan consistently benefited from lower costs of capital over the decade. Elsewhere, Friend and Tokutsu (1987) estimated that from 1962 to 1984, the Japanese inflation-unadjusted equity cost ranged from 8.2 to 16.1 percent while the corresponding U.S. cost of equity estimates ranged from 9.5 to 13.1. Hodder (1989, 1990) saw, at worst, only a small U.S. disadvantage, which he claimed could be explained away by the hefty "monitoring" cost of Japanese banks. Finally, Baldwin (1986) concluded that U. S. and Japanese costs of capital were very similar.
In addition to the selection of results found in Table 1, costs for specific types of investment also were found. What is generally observable for these results was that the U.S. was always at a disadvantage for the 1980s but that the size of the disadvantage varied for specific projects. For example, McCauley and Zimmer (1989) estimated that the cost of an R&D investment with a ten-year payoff was 16.7 percent in the U.S. and 8.4 percent in Japan during the 1980s, a U.S. disadvantage of 8.3 percentage points. Yet for equipment and machinery investment with a twenty-year life, the U. S. disadvantage was only 2.9 percentage points. (See Table 2 for other comparisons.)
Because no consensus had been reached, we decided to estimate RCFCs using parameters and variables that were already available in the extant literature. Thus, in the remainder of this article we seek to identify commonly shared methods and techniques that provide enough flexibility to enable calculations of a wide range of the RCFC in Japan and in the United States. The purpose of this exercise is purely empirical. We want to satisfy our intuition that the real cost of capital was higher in the U.S. than it was in Japan during the 1980s.
One measure often employed to calculate the financial cost of capital is what Frankel (1990) has termed ". . . [a] traditional measure of the cost of capital [which] is a weighted average of the cost of borrowing and the cost of equity . . ." (p. 5). This measure is expressed as:
RC = wRD + (1 - w)RE, (1) where RC is the real cost of financial capital, RD is the real cost of debt financing and RE is the cost of equity capital. The relative weight attached to debt financing is w; the weights of debt and equity financing must sum, of course, to one. It is also traditional to equate w/1-w to the debt-equity ratio [see Frankel (1990) p. 8], which is important because weights can be calculated from debt/equity ratios previously calculated.
Equation (1) requires one parameter, w, and two variables, RD and RE, to arrive at an estimate of the real financial cost of capital. A precise measure of w is probably unrealistic, so that two sets of weights were employed in the calculations. The first set was derived from French and Poterba (1990). These authors provide estimates of the debt/equity ratios for Japan and the U.S. (see Table 4, p. 36). Using the same procedure, a second set of weights was extracted from McCauley and Zimmer (1990, see Chart 5, p. 13). The French and Poterba debt/equity ratios implied w's of .489 for Japanese debt finance and .408 for U.S. debt finance. Alternatively, the McCauley and Zimmer debt/equity ratios implied w's of .692 and .469 for Japanese and U. S. debt financing respectively.
Debt and Equity Financing
In the present context, debt financing requires calculation of real interest rates. The latter were equated to the long-term nominal government bond yield adjusted for the expected rate of inflation. Basically the idea here is to analyze international credit conditions corresponding to freely traded securities of similar risk and equal maturity in each country. "The most obvious choice is government bonds . . . [because they are] . . . relatively free of most risks, except, of course, for those which arise from uncertainty concerning inflation" [see Bernheim and Shoven (1986) p. 6]. Given the choice of long-term government bonds as the comparative debt instrument, inflation uncertainty had to be incorporated into the analysis.
Calculating inflation uncertainty requires an assumption regarding market expectations because it includes the expected rate of inflation over the holding period of the bond. Because the aim of this article is to calculate a wide range of point estimates of the RCFC, two market expectation assumptions were considered. The first is adaptive expectations. The estimate for adaptive expectations was taken from Frankel (1990). These real interest rates were calculated using fifteen years of declining weights on inflation.
A second approach to inflation uncertainty assumes that investors form expectations rationally, incorporating all relevant and publicly available information into the decision-making process. This hypothesis implies that the current inflation rate is, on average, equal to the ex ante expectation of the inflation rate for the relevant forecast period. Finally, the real debt costs were also calculated using the Bernheim and Shoven "Benchmark" real interest rates for the 1980s. [See Bernheim and Shoven (1986), Table 4, p. 16]. In total, this provided us with three alternative real cost of debt variables.
Estimating the cost of equity variable also presents a challenge to the researcher. For example, Hodder (1990) has noted that . . . "[t]he inaccuracy in cost of equity estimation is relatively well-known . . . [a]s a consequence, one frequently tries different estimation approaches to see if the resulting estimates are reasonably consistent with each other." (p. 3.) In fact, it is not the purpose of this article to survey or to summarize or to resolve these issues [see Frankel (1990) for this]. Rather, the intent is to infer a wide range of single-point estimates from the existing literature.
This is a macroeconomic application. In essence, this article compares the relative national real interest rate cost of debt, which would include only a small default risk premium, and the real cost of equity capital, which would also include risk.
Three main techniques and seven methods in total were employed to calculate an equity cost of capital.(3) The three main techniques include: (1) the dividend/share price ratio plus a measure of expected real capital gains accruing to shareholders; (2) an earnings/share price ratio also adjusted for a measure of expected real capital gains accruing to shareholders [see Frankel (1989) p. 7]; and (3) the unadjusted earnings/share price ratio [see Shoven (1990), or McCauley and Zimmer (1989)]. Techniques (2) and (3) were also adjusted (on the Japanese side) for cross-holdings and depreciation accounting using the French and Poterba estimates [see French and Poterba (1990), Table 4, p. 36]. Because there were three methods of calculating the equity cost of capital for each of (2) and (3) and one method for (1), a total of seven possible calculations results. Therefore, these calculations in conjunction with the two sets of weights and the three real interest rate assumptions provide forty-two possible calculations for the real cost of financial capital.
A wide range of national estimates for the cost of capital over the 1980 to 1988 period were calculated. The central tendency for the U. S. RCFC was 7.5 percent and for Japan it was 4.2 percent. The average U.S. disadvantage is, thus, about 3.3 percentage points for the period. The U.S. RCFC ranged from a low of 5.0 percent to a high of 9.7 percent. The calculation for the Japanese RCFC ranged from a low of 1.9 percent to a high of 7.3 percent. As a result, the U.S. disadvantage ranged from a low of 1.3 percentage points to a high of 4.8 percentage points. The overriding conclusion, in all cases, was that the RCFC was lower for the Japanese economy than it was for the U.S. economy, on average, during the 1980-88 period.
A summary, in matrix form, of the U.S. real cost of financial capital disadvantage is presented in Table 3. The rows, RE1 to RE 7, represent the seven alternative real equity cost assumptions noted above, while the columns RD1 to RD3 refer to each of the real interest rate assumptions. The two sets of weights are applied to each pair of debt/equity ratio assumptions resulting in a 7 x 6 matrix of calculated values. The average of each combination of assumptions is presented in the last column and row. The central tendency, or the overall average, is presented in the lower right hand corner of the matrix.
Referring to the previous discussion of the ranges of the general results, we can now see that the largest difference in the RCFC occurs when: (1) the earnings/share price ratio is not adjusted for anticipated earnings growth; (2) no French and Poterba (1990) adjustment is made to the Japanese share price/earnings ratio; (3) the rational expectations assumption is used to generate expected inflation; and (4) the French and Poterba weights are employed. Alternatively, the narrowest difference can be found when we used: (1) the dividend/share price plus expected growth of earnings for the cost of equity; (2) the benchmark real interest rates for the cost of debt; and (3) the McCauley and Zimmer weights.
The results presented in the previous section do not consider the impact of tax structures. They represent the before-tax real cost of financial capital. But the after-tax cost of capital is what matters for investment decisions. In fact, researchers have made various adjustments for domestic tax structures in order to compare net-of-tax capital costs and returns. Table 4 presents the findings from a selection of studies. While these authors had different intents, they all demonstrate that the U.S. was at a disadvantage vis-a-vis the Japanese in terms of the tax burden on the real cost of financial capital.
A great deal of caution must be taken when blending national tax rates into comparisons of real financial cost of capital comparisons. For example, tax and regulatory differences can often be exploited by offshore borrowing or security issuance. The optimal combination may be firm-specific when it has foreign subsidies [see Hodder (1990) p. 12]. Yet this will only tend to mitigate any difference rather than to equalize the net effective after-tax costs. And, in any event, not all firms are multinational. Arguably, most firms must be content to raise financial capital on domestic markets, and so we consider it appropriate to consider the implications of aggregate average (and marginal) tax structures on the real cost of financial capital. In this regard, Bernheim and Shoven (1986) calculated what they termed "tax wedges" using post-1986 codes in the U.S. and Japan. The "tax wedge" is the difference between the financial cost of capital to the firm and the return to the investor. In effect, they found that, for any amount of financial commitment by a potential investor, the U.S. cost of the funds would have to be 1.6 times the Japanese cost of funds if the investor's returns were to be equalized after adjusting for taxes in the respective countries. (See Table 5.)
Elsewhere, Bernheim and Shoven (1986) developed a cost-of-capital model that extends the traditional approach, represented by equation (1), to include time discounting, a risk premium and ". . . a moderately detailed description of the personal and corporate income tax system." [Shoven (1990) p. 14]. Shoven (1990) estimated a multiplicative tax term, (1-A)/(1-c), where A is the present value of the tax savings due to depreciation deductions and c is the corporate tax rate for two types of physical investment. The first was an automobile with a seven-year useful life span. The second was a plant with a 31.5-year useful lifespan. These tax multipliers are also presented in Table 5. Again, for both types of physical investments, the U.S. was shown to be at a disadvantage versus the Japanese. For plants, the investment in the U. S. is about 6 percent higher. But it is noteworthy that taxes increase the RCFC substantially in both the U.S. and Japan. The tax structure adjustment is much less severe for the automobile investment; nevertheless, the U.S. is at a disadvantage, albeit only about 2.5 percent.
Next, we consider the difference between marginal effective corporate tax rates. Two comparisons are shown in Table 4. The first set has the tax structure in place for much of the sample period of this article. It shows that the marginal effective tax rate was substantially lower in Japan than in the U.S. The Japanese Diet did not approve a tax reform package until December 1988 in spite of considerable pressure at home and abroad. In fact, Shoven (1989) has estimated that Japanese tax reform raised the effective tax rate on Japanese investments very sharply from 5 percent in 1980 to 32 percent in 1988. The U.S. also was putting in place its own tax increase on corporate investment at this time. It has been estimated that the Tax Reform Act of 1986 increased the marginal effective corporate tax rate from 29 percent to 41 percent. After these tax changes, the U.S. maintained about a 7 percent effective tax rate disadvantage. However, the gap had been narrowed sharply from the 23 percent disadvantage with which it had to contend for most of the 1980s.
In spite of these words of caution, perhaps it would be instructive to go one step further and contemplate how these national tax rates would alter the relative RCFC comparisons presented in Tables 1-4. International comparisons are always dangerous, but they will continue regardless of the inherent difficulties of interpretation. Therefore, assume that our firms, Japan Inc. and U.S. Inc., are the Shoven (1989) firms making the automotive and plant investments represented by the parameters on his Table 4. The results of these hypothetical investments are presented in Table 5.
The U. S. disadvantage increases in all cases. The national tax multipliers were applied to the extreme cases, again in keeping with the intent to provide more than one, single-number comparison. The largest increase in the U.S. disadvantage is 2.05 percentage points for the 31.5-year plant in the maximum disadvantage case. But the largest percentage increase in the U. S. disadvantage occurs at the other extreme. The minimum U. S. disadvantage for the 31.5-year plant jumps 63.9 percent because of the higher U. S. relative national tax rate structure. The automobile investment cost is raised less, but in both extremes the U.S. is put at an incremental disadvantage as a result of the tax rate structure.
The general conclusion of this article is that Japanese firms, in aggregate, on average and in real terms enjoyed a substantial advantage over U.S. firms when it came to raising financial capital during the 1980s. The range of the U.S. disadvantage was a multiple of between 1.3 and 2.75 times the Japanese RCFC. It was centered at 1.75 times the Japanese RCFC. Additionally, after reforms in both countries, relative national tax structures still appear to favor Japanese over U.S. corporations and investors.
So the real cost of financial capital in Japan during the 1980s appears to have been lower over time and for comparable investments. Arguments that the difference in the RCFC can be explained away by risk differences [Baldwin (1987)] and/or "monitoring" costs [Hodder (1989, 1990)] are, at best, inconclusive. Alternatively, the average estimate of the U.S. disadvantage presented in this article is narrow compared to some others. For example, Ando and Auerbach (1987) estimated a six-percentage-point gap favoring Japan. But even here, Hodder (1989) claimed that "monitoring" costs of only 3.1 percent would close this gap. Also, once financial capital has been raised and spent to install "long-lived" physical capital, industries must live with the results for many years. Thus, even though calculations for 1990 may show the U.S. disadvantage in the cost of financial capital to have been eliminated, cost-of-capital parity must persist for years to come before the competitiveness disadvantage is also eliminated.
1 Drucker (1990) also reports a range of 2-3, but his estimates of the levels of U.S. and Japanese costs are different from ours.
2 These calculations excluded the cost of equity on the Japanese side, which we believe is important in explaining our estimates.
3 The expected rate of inflation in period t + 1 is equal to the current rate of inflation plus a "white noise" term, which itself has an expected value of zero. Over a long period time, say a decade, it is probably not unrealistic to set the white noise term to zero and equate the actual real debt cost to the nominal debt cost deflated by the corresponding actual inflation rate.
4 Share price to earnings ratios have also been used as a measure of the equity cost of capital. This is simply inverse of the earnings/share price ratio.
Ando, Albert, and Alan Auerbach, 1985, "The Cost of Capital in The U.S. and Japan: A Comparison," NBER Working Paper no. 1762, October.
Ando, Albert and Auerbach, 1988, "The Cost of Capital in the U.S. and Japan: A Comparison," journal of the Japanese and International Economics, Vol. 2, pp. 124-158.
Baldwin, Charles, 1986, "The Capital Factor: Competing for Capital in a Global Environment," in ed. M. Porter, Competition in Global Industries, Harvard Business School Press, Boston, MA, pp. 1985-223.
Bernheim B. Douglas, and John Shoven, 1986, "Taxation and the Cost of Capital: An International Comparison" paper presented at the American Council for Capital Formation Conference, September.
Drucker, Peter, 1990, "Japan's Not-So-Secret-Weapon," Wall Street Journal, January 9, p. A14.
Frankel, Jefferey, 1990, "Japanese Finance: A Survey," forthcoming in ed. Paul Krugman, The U. S. and Japan: Trade and Investment, Chicago: University of Chicago Press.
French, Kenneth, and James Poterba, 1990, "Are Japanese Stock Prices Too High?" Mimeo, University of Chicago, Revised February.
Friend, I and I. Tokutsu, 1987, "The Cost of Capital to Corporations in Japan and the U.S.A.," journal of Banking and Finance, Vol. 11, pp. 313-327.
Hatsopoulos, George, 1983, "High Cost of Capital: Handicap of American Industry," Study sponsored by the American Business Conference and Thermo Electron Corp., April.
Hatsopoulos, George and Stephen Brooks, 1986, "The Gap in the Cost of Capital: Causes, Effects and Remedies," pp. 221-280, in eds. Ralph Landau and Dale Johnson, Technology and Economic Policy, Ballinger Publishing Co., Cambridge, MA.
Hatsopoulos, George and Stephen Brooks, 1987, "The Cost of Capital in the United States and Japan," International Conference on the Cost of Capital, Kennedy School of Government, Harvard University.
Hatsopoulos, George, 1990, "Technology and the Cost of Equity Capital," Paper presented at the National Academy of Engineering Symposium, Washington D.C., April.
Hodder, James, 1988, "Capital Structure and Cost of Capital in the U.S. and Japan," Mimeo, Stanford University, July.
Hodder, James, 1990, "Is the Cost of Capital Lower in Japan?" forthcoming in the Journal of Japanese and International Economists.
Jorgenson, Dale, 1963, "Capital Theory and Investment Behavior," American Economic Review, 53 May, pp. 247-259.
Kester, W. and T. Luehrman, 1989, "Real Interest Rates and the Cost of Capital: A Comparison of the United States and Japan," Japan and the World Economy, Vol. 1, pp. 279-301.
Langford, R., 1989, "The Cost of Capital in the U. S. and Japan: Comparisons and Implications," Marakon Associates Commentary, Fall, pp. 11-20.
McCauley, Robert and Steven Zimmer, 1989, "Explaining International Differences in the Cost of Capital," Federal Reserve Bank of New York Quarterly Review, Vol. 13 (2), Summer, pp. 7-28.
Mishkin, Frederick, 1984, "Are Real Interest Rates Equal Across Countries? An Empirical Investigation of International Parity Conditions," Journal of Finance, Vol. 39, pp. 1345-1358.
Richman, L., 1989, "How Capital Costs Cripple America," Fortune, August 14, pp. 50-54.
Shoven, John, 1989, "The Japanese Tax Reform and the Effective Rate of Tax on Corporate Investments," in ed. L. Summers, Tax Policy and the Economy, M.I.T. Press; Cambridge, MA.
Shoven, John, 1990, "Consumption vs. Income Taxes for Deficit Reduction and Tax Restructuring," paper prepared by the American Council for Capital Formation Conference on "Public Policy Options to Increase U.S. National Saving," Washington D.C., October.
Table 1 Cost of Capital Funds: Japan vs. U.S. Selected Studies: Average Annual Interest Cost U.S. Disadvantage/ Japan U.S. Advantage (Percent) (Percent) (Pts.) Ando and Auerbach (1985) (Return to Equity) 13.5 2.2 (11.3) Ando and Auerbach (1985) (Cost of Equity) Small Sample 8.2 10.5 2.3 Pre-tax Large Sample 6.4 12.4 6.0 Small Sample 3.8 5.0 1.2 After-tax Large Sample 2.4 5.5 3.1 Ando and Auerbach (1987) (Cost of Equity) Pre-tax 6.5 12.3 5.5 After-tax 2.5 5.6 2.4 Friend and Tikutason (1987) (Cost of Equity) 8.2-16.1 9.5-13.1 1.3-(3.0) Hatsopoulos (1981) (Japan: Cost of Debt) Pre-tax 2.5 8.9 6.4 After-tax (1.8) 6.4 8.2 Hatsopoulos and Brook (1985) (Cost of Funds) After-tax 1.5 6.0 4.5 Hatsopoulos (1990) (Cost of Equity) After-tax 2.5 7.0 4.5 Table 2 Cost of Capital: Japan vs. U.S., 1980 to 1988 Selected Projects: Average Annual Interest Cost U.S. Japan U. S. Disadvantage Equipment and 8.2% 11.1% 2.9 Pts. Machinery(1) (20 Year Life) Plant, 5.4% 10.7% 5.3 Pts. (40 Year Life) R&D(1) 8.4% 16.7% 8.3 Pts. (10 Year Payoff) Expensed Item(1) 35.6% 39.6% 4.0 Pts. (3 Year Life) Land 6.3% 12.5% 6.2 Pts. Autos(2) (7 year life, 1988 Cost) Equity Financed 4.1% 10.4% 6.3 Pts. Debt Financed 2.5% 8.3% 5.8 Pts. Plants(2) (30 year life, 1988 Cost) Equity Financed 5.1% 12.6% 7.5 Pts. Debt Financed 3.1% 9.7% 6.6 Pts.
(1) McCauley and Zimmer (1989). Adapted from Table 2 p. 16.
(2) Shoven (1990)
Real Cost of Capital Funds: U.S. Disadvantage vs. Japan, 1980 to 1988 Macro Comparisons - Various Assumptions
Real Interest Rates French and Poterba McCauley and Zimmer Weights Weights Equity Assumptions RD1 RD2 RD3 RD1 RD2 RD3 Average RE1 2.30 2.42 1.72 2.29 2.30 1.33 2.06 RE2 4.58 4.70 4.00 4.18 4.20 3.22 4.14 RE3 3.75 3.87 3.17 3.69 3,70 2,72 3.48 RE4 2,93 3.05 2.37 3.18 3.20 2.22 2.83 RE5 4.71 4.83 4.14 3.92 3.94 2.96 4.08 RE6 3.88 4.00 3.31 3.42 3.44 2.48 3.42 RE7 3.05 3.17 2.49 2.92 2,94 1.96 2.76 Average 3.60 3.72 3.03 3.37 3.38 2.41 3.25
Various Tax Adjustments: Japan vs. The U.S., 1980 to 1989
U.S. Japan U.S. Disadvantage "Tax Wedges"' Investment Cost 1.530 2.450 1.600 over Investor Return Multipliers:(2) Average net of interest cost deduction to Corporate tax payment. Plants (31.5 Yrs.) 1.309 1.384 1.057 Autos (7 Yrs.) 1.052 1.077 1.024 Marginal Effective Corporate Tax Rate(3) 1.320 (1989) 1.410 (Post TRA) 1.068 Including debt 1.050 (1980) 1.290 (Pre TRA) 1.229 write off.
(1) Bernheim and Shoven (1986)
(2) Shoven (1988)
(3) Frankel (1990) from Shoven (1986B)
Impact of Relative Tax Structure on Investments
Pretax Cost of Capital U.S. Japan U.S. Disadvantage (Percent) (Percent) (Percent Pts.) (Percent) Minimum 5.82% 4.49% 1.33 Pts. 22.8% Disadvantage Maximum Disadvantage Tax Multipliers 35 Year Plant 1.384 3.301 7 Year Auto 1.077 1.052 Hypothetical Post-tax Cost of Capital U.S. Japan U S. Disadvantage (Percent) (Percent) (Percent Pts.) (Percent) Minimum Disadvantage 30 Yea, Plant 8.06% 5.88% +0.85 Pts. +63.9% 7 Year Auto 6.27 4.72 +0.21 +15.8 Maximum Disadvantage 35 Year Plant 10.39 3.51 +2.05 +42.4 7 Year Auto 8.09 2.82 +0.44 9.1
|Printer friendly Cite/link Email Feedback|
|Author:||Lippens, Robert E.|
|Date:||Apr 1, 1991|
|Previous Article:||Cost of capital for industry and banks.|
|Next Article:||The cost of capital in Japan: update.|