with J. Mark Ramseyer (Mitsubishi Professor of Japanese Legal Studies, Harvard Law School)
interviewed by MIYAJIMA Hideaki
Miyajima: These days there are many issues concerning corporate governance. In this interview, I'd like to focus on two of them. The first issue is ownership structure and the other is managerial compensation. Let me begin with ownership structure. Since 1997, the ownership structure of Japanese firms has been changing dramatically. So how can we understand the causes and results of this ownership structure change?
My first question is what are your views on the cross-shareholding issue? Is cross-shareholding a unique feature of the Japanese firm system, or not?
Ramseyer: What I would say is that I really don't think that it exists, and that it ever did. What Professor Yoshiro Miwa and I did was look at cross-shareholding arrangements in 1965 and 1975. We found that among the industrial firms, there was virtually no cross-shareholding. So I start with that, and it's hard to believe that between 1975 and 1995 it would have developed. What you do find is that banks buy stock in firms. But I don't see any reason to think that there is anything "cross" about it. Certainly if it is, it's unlikely that the amount that a firm holds in a bank is anywhere close to what the bank holds in the firm. A more plausible explanation to me would be that banks in Japan are allowed to own stock. In America, under Glass-Steagall, banks couldn't own stock, and so they didn't. In Japan, banks can, and they have, of course, lots of funds. So if you're looking for a place to park your money, large industrial firms are a good place to park your money. And so my hypothesis would be that what gets, mistakenly understood to be cross-shareholding, is simply a case of the banks holding a diversified portfolio of investments that include stock in listed firms.
My guess is that sometimes industrial firms also hold diversified portfolios, and it would be sensible to include banks as part of that portfolio. So in many cases where a bank holds an interest in a firm, the firm probably holds an interest in a bank. But I don't see any reason to think that that's part of a cross-shareholding arrangement.
Miyajima: This is a unique understanding. So how can we explain the dramatic decrease of holding shares by banks to client firms since 1997?
Ramseyer: Well, I haven't studied it, so this is just a guess, but I would think probably the banks needed cash. And stock portfolios are among the easiest to liquidate, so they sold some of their stock.
Miyajima: I wrote a paper concerning how cross-shareholding unwound since 1997. And the main incentive of the banking sector to dissolve cross-shareholding is, as you said, the necessity of money for solving, or coping with, non-performing loans. What I observed in this research is that the banking sector has a great need to liquidate, or sell, its stock. So then what companies are sold to the market? It seems natural that highly performing firms are much easier to sell, so that first the good companies are sold by the banks.
Ramseyer: Why would that be? I mean, these are listed companies, so they have a market price. And whether the market price is high relative to performance, or low, I'd think it's just as easy to sell either.
Miyajima: Right, but suppose that a bank, for example Mitsui-Sumitomo held Sony and other Sumitomo group firms. And the Sony stock is sold first, or the bank tries to keep the Sumitomo-related firms, even if this company's stock price is low. So the point is that if the banking sector prefers to sell stock in the market, it is likely to sell the stock with a high stock return.
Ramseyer: I don't understand that. You need cash. Stock has a market value and if you need $200, you can sell one share that sells for $200 or two shares that sell for $100. But in any case you're getting the same amount of cash.
Miyajima: So what you're saying is the process of a bank selling the stocks will transpire evenly throughout the bank's portfolio.
Ramseyer: I would assume they have some sort of target for how they want to balance the portfolio, and they would sell in a way that gives them the balance that they want.
Miyajima: Among academics, conventional wisdom on cross-shareholding is that it has a cost and a benefit. The benefit is this system, or this scheme, made it possible for managers to implement long-term management. And the cost is the insider control and shielding the pressure of the capital market. But, I think you probably don't buy this kind of opinion.
Ramseyer: That's right. The incentive that managers have is to keep stock prices high, because otherwise they get into trouble. The stock price is based on the present value of expected cash flows, net cash flows, to the firm. If it's the present value of expected net cash flows, then that's neither short-term nor long-term. That's everything that's coming in. So there's no reason to think Japanese managers were particularly longer term than Americans, unless they're using a lower discount rate. So, to the extent that interest rates are lower in Japan, I suppose, perhaps, that means they take a longer-term perspective. But that's just a function of the discount rate.
If insider control somehow results in managers not maximizing the net cash flow, I don't know why a bank would want that. I would think a bank would not want that. A bank would want this firm to be maximizing its cash flows, both because that increases the value of the stock, which the bank holds as part of its portfolio, and because if you're increasing your cash flows, you're more likely to pay back your loan. And the bank wants its loan paid back.
Miyajima: That's right. But some observers say that there are two lines of thought about bank ownership. One interpretation is that the bank is a dedicated monitor of other shareholders, and mainly the bank uses its ownership stake for realizing the value maximization of client firms. The other idea is that equity ownership of the bank to the client firms is mainly motivated by maximizing its loans to those firms. And in order to maximize their clients' firms, and when the firm is not facing a bankruptcy situation, the bank tries to maximize, to put it differently, lender money, as much as possible, beyond a value maximization position. But you do not buy this idea. Why?
Ramseyer: Why would a firm want to borrow excessive amounts of money? Japanese firms have relationships with lots of banks, and I assume they play them off against each other, and I don't see why they would have any interest in borrowing excessively. It's true that banks in Japan, I would guess, monitor their debtor firms. But banks everywhere monitor their debtor firms because they want to be paid back. And I've never seen any evidence that there is anything different about what happens in Japan than happens anywhere else. You know, part of the story that's sometimes told is that there's a particular bank, which is a main bank, which is acting as the delegated monitor on behalf of all other banks. And Professor Yoshiro Miwa and I tried to find evidence of this, and we didn't see any evidence.
Miyajima: Professor Miwa and you also haven't found any evidence that bank is delegated as monitor by other shareholders.
Ramseyer: Not by anybody. That's right.
Miyajima: The last question concerning ownership structure is the recently increasing threat of hostile takeover in Japan. There is some movement that Japanese firms again began to stabilize their shareholding. How do you see this phenomenon? It is a minor phenomenon, or a big change, or some symptom of a new type of corporate behavior? Do you have any opinion on that?
Ramseyer: You know, I've never understood what stable shareholding was. I mean, you look at the ownership of the TSE listed firms, and most of them start at 4%, 3%, and then they go down from there. But then these different newsletters will say, "Stable shareholding on average is 35%." I don't know where they get these numbers. It's always been a puzzle when you read these newsletters. And they don't tell you where they get these figures. But just looking at the ownership structure, I've never seen anything stable there. So my guess is that there never was stable shareholding, and there isn't now. Maybe people go to friends and say, we'll buy your stock if you buy ours. I suppose that's part of what this is claiming. But I have not seen evidence of that.
Miyajima: What happened since 1990, and especially 1997, is a declining share held by the financial sector, and an increasing share held by foreign investors. Right now, on average, TSE companies' ownership structure or share held by foreign investors is approximately 28%, which is almost a 20% increase from 10 or 15 years ago. This seems to be a big change. Does this or will this change have any type of substantial effect?
Ramseyer: I guess I don't know why a foreign investor would have any different goals in investing than a Japanese investor. So I don't see why that's a relevant number. The decline in the financial sector would be potentially relevant to the extent that financial banks are heavily regulated. I have not studied this, but it's possible that the heavy regulation of the banks means that they have to behave slightly differently as a shareholder than an unregulated firm. But if it's just some stock held by Japanese industrial firms, some Japanese commercial firms, some Japanese individuals, and some foreign firms, among those I don't know why they would be investing with different goals.
Miyajima: To sum up your understanding, is all investors (institutional, individual, Japanese, and foreign) seem to behave almost the same, so that a decrease in banks' shareholding ratio and an increase in foreign investors' ratio may not have any significant effect?
Ramseyer: That would be my guess.
Miyajima: Let's move to the second question, about compensation. Firstly, could you summarize your current opinion on the compensation issue following the Enron situation in the United States? In my understanding, before Enron, the public opinion, or the popular opinion among economists and law academics was that high-powered incentives or stock options were a good device for allying the interests of the CEO and shareholders. But after Enron this opinion changed and now U.S. CEOs are widely seen as overcompensated. This is my very basic understanding.
Ramseyer: Well, there are puzzles. Let me go back. What Professor Minoru Nakazato and Professor Eric Rasmussen and I find are that Japanese compensation levels are substantially lower than American levels. They're not quite as low as they initially appear. The reason is that compensation, both in the U.S. but also in Japan, is very much a function of the size of the firm. And Japanese firms are smaller than American firms. So if you hold firm size constant, Japanese presidents seem to be making about a third of American compensation. So it's less, but it's a third, it's not one thirtieth, the way it often appears in the popular press, and that's because Japanese firms are smaller.
A second thing to remember, I think, is that if you compare Japanese compensation levels with other countries, they're pretty close to Europe; slightly lower. Among the European countries, Japanese compensation puts it at the lower end of the range within Europe, but it's in the range. We say it's in the ballpark of European compensation levels. And what this means, I think, is that the puzzle is not explaining the low level of Japanese compensation, it is explaining why American compensation levels are so high. American executive compensation started to increase radically in the early- to mid-1980s. And that's consistent with the period when stock options began to be used heavily. I don't know whether that's good or bad. It's just a fact that executive compensation tended to rise very much. And as you point out, one thing which investors want is for executives to have an incentive to maximize firm value. And a good way to get its executives to maximize firm value is to give them an interest in the firm. So if they have a large interest in the firm, they obviously have an incentive to do what investors want them to do to some extent. Now, I think among academics, for the last 15, 20 years, that's been probably the dominant position, at least among economists. You're also correct that the public didn't pay a whole lot of attention to it.
But there's another strand that goes back to the 1930s, that's more common among law professors perhaps; and that's the Bearle-Means position. What Bearle and Means argued in the 1930s is that American listed firms were owned by a very large number of investors, each of whom had a very small interest in the firm. And because each person had a small interest in the firm, nobody had an incentive to monitor what was going on, and as a result managers were behaving selfishly. In terms of the possible range of selfish behavior, one plausible phenomenon would be paying themselves too much. In the last 10 years, the number of law professors and economists who have been arguing that managers are paying themselves too much, and that this occurs because they're not monitored if shares are held by a very large number of people with small interests, has probably increased in the academic community. My colleagues, Professors Lucian Bebchuk and Jesse Fried, are both probably very well known in the law and economics community as people who make that claim, that managers are out of control, and because they're out of control they pay themselves too much. I don't know if that's the dominant view or not, but that is a powerful position that one hears.
Does Enron have something to do with this? I would guess that among economists, it probably doesn't have anything to do with this. Among law professors it might. Among the public it might. You know, I teach corporate law, and after Enron my mother was very upset. She said, "Why aren't you telling your students that this is awful? Why aren't you teaching your students not to steal from the firm?" My answer to my mother was that you can't teach these things in law school. This is the sort of thing that you learn in kindergarten. And if you don't learn not to steal in kindergarten, it's too late to teach somebody once they're 25 that this is just basic theft, it's immorality. Most professors who work in this field would similarly say that there's always a risk that there will be theft. And Enron is just a good example of sophisticated theft. And that's bad.
On the other hand, as you know, the optimal level of crime is not zero. If police are behaving as they should, there will always be some crime, because it would be too expensive to prevent it all. And once in a while that may mean that you have disasters like Enron. But the fact that you get an occasional disaster once in a while, in itself, doesn't mean there's anything wrong that needs to be fixed. There may be, and the puzzle for a scholar after something like Enron is to try to figure out whether there's something wrong with the system. Is this the result of a malfunctioning governance structure or regulatory structure? Or is it instead simply the random crime that occurs in any system, just like there will always be murders in any community?
Miyajima: That's a very interesting explanation. So you have already mentioned about your paper with Professors Nakazato and Rasmussen, and you will present two papers at a RIETI project workshop. Could you summarize the bottom line of your paper or your fact-finding based on the new dataset?
Ramseyer: Well, we find that compensation levels are lower. Holding firm size constant, they are about one-third the level of American firms. They are sensitive to firm size with the same elasticity; about 0.3 as I recall. And that that's pretty close to what it would be in the U.S.
Miyajima: How did you measure firm size?
Ramseyer: We find that it works best if we use stock market capitalization. We also try book value of assets; and it works, but not as well.
Miyajima: Okay, I understand. What else are you finding?
Ramseyer: There doesn't seem to be much sensitivity to performance. As I recall, Professor Rasmussen didn't think that one found it in the U.S. either.
Miyajima: Really? Our understanding on the characteristic of Japanese compensation is, first, low level of compensation compared to the U.S., and second, elasticity of compensation, to corporate performance might be lower than the U.S. But it is not the case, according to your empirical.
Ramseyer: Professor Rasmussen didn't think it was. Part of the complication that comes up in our data set is that we're not measuring compensation, we're measuring income. And income will include investment income. And this has advantages and disadvantages. The advantage is that if you are an investor at a firm, you want your executive to have the right incentives. That doesn't include just salary, it also means you want to know his return from his investments in the firm. And we're able to capture both. In the U.S., executive compensation studies use Securities and Exchange Commission (SEC) data, and SEC includes only salary and compensation. It doesn't include the person's investment income. So we capture something that the studies using SEC data don't capture.
On the other hand, we may miss something that they do have. They're able to get very clean studies of how much compensation the firm is paying, and we don't have that. What we do is we try to divide our CEOs into those who are most likely and those who are least likely to have other investment income. We use things like whether they work at a family firm, what kind of stock ownership they have in the firm, whether they have been on the tax data many years in a row, so that they're likely to have saved up income. We partition the data set up into those two groups, and we do find that those executives who are likely to have substantial outside investment income, have income which is sensitive to firm performance. But others don't. And I think it comes from the fact that they simply have dividend income and it's a function of their shareholding. So they do have an incentive to increase firm performance. But it doesn't seem to come through salaries. It comes instead through stock ownership. What that means is that executives seem to vary in terms of their sensitivity to firm performance. Those who have substantial stockholdings obviously do have an incentive to increase performance. Those who are just, in Japan you'd call them salarymen, the company man, rise through the ranks and have less of an incentive.
Miyajima: So let me confirm. The point of your paper is that although a strict comparison is not possible due to a data difference, in your estimation the elasticity of compensation in relation to performance among two countries seems to be not so big a difference, or almost the same.
Ramseyer: That's what it seems to be.
Miyajima: And the second point is that even in Japanese companies, compensation is sensitive to performance.
Ramseyer: It is sensitive to performance in cases where there is an executive who has a substantial amount of stock; so that salaries are not sensitive to performance, but in many cases the executive's income will be.
Miyajima: Then why does the manager own a certain stock? A possible reason is that (s)he is the owner of that company or related to the founder. This is the owner-manager type of CEO. And the other possibility is the result of the stock option.
Ramseyer: You know, we tried very hard to get stock option data. If you do somewhat of a literature survey, you will find a working paper that claims to have data on who owns stock options and how many an executive holds. We could not locate this data.
Miyajima: Then your finding is based on managerial ownership.
Ramseyer: We looked really hard for this, and what we found is not how many options each president has, but whether a president works for a company that has a stock option plan. And there it looks as though options are complements, not substitutes for salary. So that places in Japan that pay high salaries seem also to be giving stock options. But the amount of qualified options that are possible to give is limited under Japanese law.
Miyajima: That's right. Stock options are not explicitly introduced in your estimation, but introducing stock options in Japanese firms is limited, so there is not so big a problem. So the point is manager ownership is one determinant for elasticity of compensation in relation to performance.
Ramseyer: Basically what we have in our regressions is a dummy variable for stock options, of zero one, and that's because that's all I could find.
Miyajima: Compensation is generally understood as an alternative device of monitoring a large shareholder or through the board of directors. And you introduce ownership-related variables such as outside investors. Do you find this kind of a conjuncture where compensation is an alternative to, say, monitoring by the board of directors?
Ramseyer: Well, I think what we find is that, with one exception, the various governance variables, we have a bunch of governance variables: Do you work for a family firm? Do you have a U.S.-style committee structure? Are there outside directors? Those governance variables are all insignificant. As you know, the fact that they're insignificant doesn't tell us why. There are two obvious explanations. One is; they don't matter. The second, though, is that the firms have adopted firm-specifically sensible governance structures. And if the firms have the right governance structures then they will be endogenous in these regressions and you won't get significant results.
And so I can't tell which of those explanations is what's happening. We do get significant results on one governance variable, but they don't make sense, so I don't want to put much stock in it.
Miyajima: In your explanation, then, the real puzzle in comparing compensation in Japanese firms and U.S. firms is not why Japanese firms' compensation is low, but why U.S. compensation is so high.
Ramseyer: That's right. Let me stress, I think this study has implications for corporate governance debates. We don't try to address those in this article. We see it as; just let's try to figure out what's happening, because I think before one can get in to the corporate governance debates, one needs to know what real firms are paying real presidents and why. And all we're trying to do is sort of lay out the basic facts. I think what you've said is quite right in the sense that the results we get, I think most scholars specializing in Japanese corporate practice will say, "that sounds about right." The levels that are being paid in Japan are what most people thought, I think, firms were paying. And the determinants too, I think most people would say too, "yeah, that sounds about right." But the reason that I think our paper matters is because there haven't been studies that systematically look at compensation levels in Japan, because the data's not there.
Miyajima: So as the last point on the compensation issue, I would like to ask your opinion on the design of Japanese compensation schemes. For example, is there no need to change the current compensation scheme or should we try to introduce a U.S. type of high-powered incentive system? Japanese firms seem to already be overcompensating or overpaying, so it seems we need to reduce the compensation level. Do you have an opinion on the design of Japanese companies' compensation schemes?
Ramseyer: As you know, I used to teach at the University of Chicago, and I'm a believer in markets, and this is a competitive market for CEOs. So my instinct tells me whatever is happening is what's reasonably sensible. It doesn't mean there aren't ways to improve compensation practices. There are always ways to improve the way firms are operating. But running a firm is a really difficult and complicated matter. I think professors like myself just don't know enough to do it. And I would guess that the right compensation scheme varies from firm to firm. We show that Japanese compensation patterns are not as significantly lower than American ones as we thought, but they're still lower. And American compensation levels are significantly higher than Japanese and European levels. And they're both competitive markets, both similar markets. They're all advanced economies, making the same kinds of products and living in the same kinds of cities, and the societies are very close. So I don't understand why there's this difference, and it's a real puzzle.
*Interview conducted on July 25, 2007, transcript was compiled by MIYAJIMA Hideaki and YAOITA Shumpei
