Large financial institutions have failed with much higher frequency than is generally perceived, says Andrew Kuritzkes, a partner at Oliver Wyman and head of the management consulting firm’s public policy practice in North America. “What is surprising is that we’re surprised by how often large banks fail,” Kuritzkes says. His research shows the actual failure rate is an order of magnitude higher than the default rates implied by credit ratings. What’s more, such failures are unavoidable over the long term, he says. In this interview with Knowledge at Wharton, Kuritzkes suggests new rules of the game that would greatly improve the financial system’s ability to absorb the inevitable, if individually unpredictable, shocks of big failures. Kuritzkes was the keynote speaker at the recent 12th annual Financial Risk Roundtable organized by the Wharton Financial Institutions Center and Oliver Wyman Institute.

An edited transcript follows.

Knowledge at Wharton: Today, we’re speaking about risk management with Andy Kuritzkes, a partner with Oliver Wyman who currently heads the firm’s public policy practice in North America. Thanks for joining us, Andy.

Andrew Kuritzkes: Thank you.

Knowledge at Wharton: You’ve been quoted as saying that large financial institutions have failed at a much higher frequency than generally perceived. Can you put that in perspective? That’s a pretty strong statement.

Kuritzkes: Yes. What I’ve looked at is the empirical record of large financial institutions, which I define as global top 100 financial firms — what their actual failure rate was over the last 20 years. There’s a notion we had before the crisis that large financial firms — the Lehman Brothers, the Bear Stearns, the AIGs — were actually very safe firms, that they had very high credit ratings — typically a single A or better — and that these firms should fail at a very, very low rate. But when you actually look at the empirical record, it turns out if you go back over 20 years there were 26 failures of firms that would have been in the list of global top 100 financial firms.

Knowledge at Wharton: What were some of the big ones? I think I can remember Continental Illinois was probably one. Was that before [the particular time period you mentioned]?

Kuritzkes: That was before that time, yes. My analysis shows that over the 20-year period, the default rate or the failure rate for the global top 100 works out to be 1.3%, which is at least an order of magnitude higher than the default rate implied by credit ratings of the global top 100. The median credit rating of the top 100 firms in 2007 was single-A-plus. That’s consistent with the default rate of about four basis points. So … the historical record seems to show that these firms actually fail by… more than 20 times the implied failure rate of the credit ratings.

Knowledge at Wharton: Worldwide?

Kuritzkes: This is worldwide.

Knowledge at Wharton: How many of those top 100 would have been in the U.S., for example?

Kuritzkes: In the U.S. there are eight firms that failed in 2008. That’s a large part of the sample. And so the eight firms that failed, by my definition, in 2008 were Bear Stearns, Freddie Mac, Fannie Mae, Lehman Brothers, WaMu, Wachovia, AIG and Merrill Lynch. And my definition of failure is a firm that went into insolvency, was placed in receivership or was forced into a government-assisted merger. Those eight fell into different categories. Fannie, Freddie and AIG were placed in receivership. Lehman actually went insolvent. The others were really forced into government-assisted mergers.

Knowledge at Wharton: What are the implications of your analysis that the failure rate is higher than most of us thought?

Kuritzkes: The fact that the failure rate is so much higher than implied by credit ratings … which are used by risk managers internally in most risk modeling, says to me that, first, there’s a general misperception of what the true risk is of large financial institutions. And there’s other evidence out there that we could have looked at to correct this misperception — there’s evidence of credit default swap pricing, which shows default rates that are much, much higher than credit ratings. There’s simply the sustainability of very high returns for some financial firms.You have to ask yourself: “For the firms that are earning more than 20% returns on equity for prolonged periods — is that consistent with a firm having a very, very low default rate — a default rate equivalent to a double-A or single-A rated institution?” In general we think that there’s a trade-off between risk and return, and [we wonder how] a very high return firm [can also be] a very low risk firm.

Knowledge at Wharton: Can a firm be both?

Kuritzkes: I think at some point you kind of run out of runway to do that. And it becomes increasingly difficult to sustain very high returns on equity, especially in a business model that depends so heavily on leverage. It’s easy to leverage up returns and show high returns on equity, but I think you sustain that at the cost of solvency protection. In other words, you sustain that by running increased risk, especially to your debt holders. And that’s sort of what, I think, happened in 2008. So those are some of the takeaways from the mismatch between the empirical failure rate and credit ratings. But I think the more significant implication is the one I draw for public policy. And I think the public policy implication is that this isn’t the first banking crisis we’ve lived through. In fact, over the last 20 years we’ve gone through several in different parts of the world and I think the reality of our financial system is that it’s going to be very difficult to reduce the failure rate of large financial firms systematically for a prolonged period of time. We’ve gone through too many banking crises in the post war period to think that this time around we’re going to get it right and we’re going to put in place fixes that are going to basically legislate large bank failure or large financial firm failure out of existence.

Knowledge at Wharton: Not a comforting thought. So what do we do about that?

Kuritzkes: Well, I’m sort of fatalistic. I’m sure we’ll get it right coming out of the crisis for the short-term. And I’m sure we’ll put in place fixes so that next time around it’s not going to be subprime. It will be something else. It won’t be structured credit. It will be a new set of risks that emerge that are probably unknown now and probably unpredictable ex ante. We’ll know about it after the fact and then we’ll think, gee, this is blindingly obvious. How did we miss it? But we’re not going to know about it beforehand. That’s sort of the nature of a banking crisis I think.

Knowledge at Wharton: So when should we plan on the next one?

Kuritzkes: If I knew the answer to that….

Knowledge at Wharton: Is it every 20 years? When do we start getting nervous?

Kuritzkes: It seems to happen a lot more frequently than every 20 years. Within the U.S., it’s been about a 20-year period because the last major banking crisis was in the late 1980s, early 1990s. But then after the U.S. experience you had Scandinavia in the early ’90s. You had other parts of continental Europe that had big banks fail in the early ’90s period. Then you had Japan in the later ’90s. You had the Asian banking crisis, the Russian crisis. You had the Tequila crisis. There are lots of other crises there, which resulted in domestic banking failures even if in some cases they didn’t hit the global top 100 list. So I think the crises recur a lot more frequently than even once in every 20 years. But I think that the policy takeaway is — if you are fatalistic, as I am — we’re going to have bank failures in the future for some new and unknown causes then the premium should be on trying to make the financial system more survivable or better able to survive the failure of banks. That’s where policy needs to be focused: On trying to make the system more resilient. And I think the big failing of policy in the crisis was that it didn’t constrain systemic risk. This is the first crisis where we’ve really had the huge systemic [effects on the broader economy], which then required the degree of government intervention that we’ve had. And so in the future I think we need to take a series of steps to constrain systemic and improve the system’s survivability given that banks are going to get into trouble.

Knowledge at Wharton: Did this happen the way it did this time because the institutions got so large? And how do you inoculate the system?

Kuritzkes: Well, I think size is a part of it. An interesting question that I can’t fully explain is: Why did the institutions become so big, both in terms of absolute scale and in scope of activity? Certainly deregulation played a part, but it’s unclear that there are significant economies of scale that would have driven firms to be as large as they’ve become. But setting that aside, I think that the fixes really should focus first on what really failed systemically in the crisis.And to me, the first thing that failed was we had no resolution regime for non-bank financial institutions that got into trouble. So Bear Stearns, Lehman….

Knowledge at Wharton: No regulatory, or resolution regime?

Kuritzkes: In some cases, no regulatory authority either. That’s right. But to me, in the trifecta of Lehman, Bear Stearns and AIG, we had no adequate mechanism for taking these firms and placing them in receivership and allowing some more orderly unwind of them without — in the case of Bear and AIG — the Fed stepping in and effectively throwing them a lifeline. In Bear’s case to support the merger with J.P. Morgan, and in AIG’s case just effectively taking over the firm through a conservatorship type process.

Knowledge at Wharton: Is this semantics when we talk about receivership versus temporary nationalization? These become politically radioactive terms….

Kuritzkes: Receivership to me is not synonymous with temporary nationalization. What receivership would do is you would take over the firm by wiping out the shareholders and installing new management to oversee the firm as the receiver tries to either wind down the firm’s assets — work out the firm’s assets and maximize the recovery value — or sell off the firm in whole or in pieces. And then use the proceeds to pay off liability holders.Nationalization implies to me the government owning the firm as a going enterprise and that’s not the point of receivership. Receivership is what the FDIC does. So Indy Mac was placed under receivership.

Knowledge at Wharton: So that was an unfortunate term probably — temporary nationalization when they probably really meant receivership.

Kuritzkes: Maybe they did. Maybe that’s a better way of looking at it. But receivership also has this connotation that you only do that for firms that have really failed. And I think when people were talking about should we temporarily nationalize Citi or Bank of America, nobody wanted to put the label out there because they didn’t want to send a message that the two largest banks in the U.S. had failed.

Knowledge at Wharton: Right. And it’s just too complicated to talk about the difference between Chapter 11 and Chapter 7 and all of that.

Kuritzkes: It sounds like you know more about that than I do.

Knowledge at Wharton: Well, in other words, semantics matter.

Kuritzkes: Right. And so, in any event, I think a resolution regime for non-bank financial institutions has to be at the top of the list. We have to find a way of working out these firms and ultimately imposing costs on not only the shareholders who were wiped out — in Lehman and largely in Bear and AIG’s case, but really also on the debt holders who are an important source of potential discipline on the firm. And the notion that all of AIG’s creditors are going to be made whole on a 100 cents on the dollar removes a tremendous incentive for market discipline.

Knowledge at Wharton: So you’re saying they should get what’s called a haircut?

Kuritzkes: Yes. That’s right.

Knowledge at Wharton: Now that’s not a typical attitude in the business world in general. So I’m interested in how you came to this view and how widely do you think it is felt outside of the financial world that has its own interest at heart.

Kuritzkes: That if you’re the creditor of a firm that goes bust that you should not be paid out 100 cents on the dollar? I think … there’s a lot of widespread experience there. I mean there’s a pain that’s been shared across many other industries and … I think for the business world at large this is not a new concept. I think what’s new is the notion that every creditor of a firm that’s deemed to be too big or too interconnected or too complicated to fail gets full recovery.

Knowledge at Wharton: And, putting aside the merits of what’s fair or not or what makes the most business sense, one thing that’s worrisome is that you create a moral hazard.

Kuritzkes: Absolutely.

Knowledge at Wharton: You’re kind of setting us up for another fall, which could be worse if you don’t handle this right.

Kuritzkes: That’s exactly right. That’s why I think the starting point has to be to find a way to let firms fail. Failure is a good thing if the costs of excessive risk taking or lax governance or even bad luck are internalized among a firm’s shareholders and debt holders. That’s actually what we want failure to do. That’s how the market is supposed to operate. So I think resolution regime is the first place that we have to start. And [one of] the two other things I think we need to do to improve the survivability of the financial system given the inevitability of bank failure [is] reducing firm interconnectedness. In many ways systemic risk is directly proportional to the degree of counter-party risk in the system. So I think we need to find mechanisms to reduce the absolute amount of counter-party risk from interconnectedness and then, finally….

Knowledge at Wharton: What people sometimes call contagion? Is that right?

Kuritzkes: That’s right. And then finally I think we also need to introduce counter-cyclical capital measures so that we raise capital levels during boom years, lower them during the bust years, which would reduce the connection between bank losses and the real economy.

Knowledge at Wharton: Well, that’s interesting. I want to read to you a quote and get your reaction to it. This is from Philipp Hildebrand of the Swiss National Bank and he recently said that looking at risk-based capital measures, the two large Swiss banks were among the best-capitalized large international banks in the world. Looking at simple leverage, however, these institutions were among the worst capitalized banks. So can you comment on that?

Kuritzkes: Yes. That’s one of the perverse consequences of Basel Capital regulation is that Basel translates the asset base of a firm into risk-weighted assets based on a series of complex calculations that are run off of a firm’s internal risk models.

Knowledge at Wharton: Do you mean they combined all of these mortgage loans and some were somehow under grade A, but when it came out the end of the pipeline everything was rated AAA?

Kuritzkes: In a sense that’s sort of what happened. UBS for example invested in lots of CDOs that were AAA rated that got very low risk weights under this Basel scheme. So when they compute regulatory capital, they show capital divided by risk-weighted assets. Risk-weighted assets were low. Capital ratio was high. And that’s the basis on which the Swiss banks had some of the highest capital ratios in the world. But when you looked at the total size of the balance sheet, they had a tremendous amount of leverage because you’re not risk-weighting these different asset buckets now. You’re just adding them up in nominal terms. And in nominal terms the leverage was very, very high. And I think what we learned in this crisis was that absolute leverage matters, not just risk-weighted assets.

Knowledge at Wharton: If I’m not mistaken, absolute leverage mattered to the point that the estimated losses at one of those big Swiss banks was greater than the GDP of the country. That’s real leverage.

Kuritzkes: That is real leverage.

Knowledge at Wharton: How does the country bail you out when you’re essentially bigger than they are?

Kuritzkes: Well, it’s too big to bail.

Knowledge at Wharton: Too big to bail?

Kuritzkes: Instead of too big to fail.

Knowledge at Wharton: Do you have other solutions? Because I think you’re coming up with some really interesting ones.

Kuritzkes: Well, I think those are the three main ones that I see for dealing with systemic risk, making the system more survivable. I think a part of trying to reduce counter-party risk and contagion effects is also to try to come up with measures of how interconnected or systemic a firm is. And there are new efforts under way to devise metrics that would do that — that would in effect measure a firm’s contribution to overall system wide risk.

Knowledge at Wharton: And you’re referring to worldwide systemic risk?

Kuritzkes: Well, this would probably be done on a national basis.

Knowledge at Wharton: Right. But then there’s that contagion that crosses borders.

Kuritzkes: You can take it up a level. That’s certainly true. But if you just stick with the national problem now — so we’re talking about say the U.S. systemic risk –I think if you had a measure or even proxies for that, which were less than perfect, one of the things that should be considered is imposing a systemic risk charge. We know that systemically important firms have cost the taxpayers now hundreds of billions — total exposure at risk is in the trillions of dollars. I’m sure the losses won’t ever get to the government but it’s certainly very sizable. I’m troubled by the fact that very large systemically important firms don’t pay any insurance premium in effect for the implicit government guarantee — the bailout that happens when they run into trouble.

Knowledge at Wharton: So when someone puts money in a regular bank and you’re covered by the FDIC up to, I think, $100,000, aren’t the banks paying a premium on that? You’re saying the non-bank banks don’t. Is that right?

Kuritzkes: You’re right that banks do pay deposit insurance premiums to the FDIC on their insured deposit base. But we’ve now moved to a world where it’s not just insured deposits that are guaranteed by the government. It’s in effect almost the entire liability structure. The Treasury announced that the 19 largest U.S. bank holding companies are each now too big to fail. That means that it’s not just their insured deposits that are protected, it’s the whole liabilities of those 19 banks.

Knowledge at Wharton: Including the creditors that aren’t getting the haircuts?

Kuritzkes: That’s right. Including the creditors who aren’t getting the haircuts and including bank holding companies of convenience, as I call them, such as American Express and GMAC and MetLife, which converted to bank holding company status (in the case of GMAC and American Express, once the TARP was announced). But these really aren’t institutions we think of as being banks. They don’t really have deposit bases. We’re now saying, at least for the duration of the crisis, that we’re going to guarantee all of their liabilities; and, again, there’s no haircut for the creditors. So that’s a huge step beyond just deposit insurance. And no one is paying a premium on that.

Knowledge at Wharton: Well thank you very much for speaking with us.

Kuritzkes: My pleasure. I appreciate it.