Sunday, February 23, 2014

The Financial Crisis in Retrospect

The release of FOMC transcripts from 2008 is an important event. These are key documents for anyone interested in understanding the role of policy in the financial crisis. Jim Bullard has written a short piece on the events in the latter half of 2008. Bullard's view seems to be that the Lehman failure was widely anticipated, and perhaps beyond policymakers' control, though of course the resulting systemic events appear to have come as a surprise.

Jeff Lacker gave what I think is an extremely important speech this past Friday on the crisis, the monetary policy response, and the role of economic theory during the crisis and in the aftermath. The speech is both enlightening and thought-provoking. Lacker has a unique perspective on the financial crisis. He has been President of the Richmond Fed for 10 years and, before assuming a management role at the Fed, worked on research problems in banking, financial contracts, and incentives. The Bank of America is in the Richmond Fed district, so Lacker would have had an inside track on what was going on in one of the largest financial institutions in the country, and thus in much of the rest of the financial system.

Some of the interesting points in Lacker's speech are the following:

1. Economists had plenty of tools and models available to apply to the problems of the financial crisis. But they may not have applied them very well. Lacker traces the roots of the crisis to August 2007, when interest rate spreads began to rise in the market for asset-backed commercial paper. This brought on a response from the Fed in the form of a change in discount window policy. Lacker questions whether that was the appropriate policy response, and uses this as an example of poor application of the available economic theory to the problem at hand. What Lacker mentions in his speech relates to a particular FOMC discussion, which can be found in the 2007 FOMC transcripts. In the transcripts, there is a very long discussion of the policy, and Bernanke defends it by citing chapter 6 in a book by Allen and Gale. My interpretation is that the policy decision had been made - the discount rate would be lowered - and Lacker was questioning the policy. Bernanke's response was to find a piece of academic work that supported the policy move, but that piece of work had questionable relevance to the problem at hand.

2. Too-big-to-fail is at the heart of the financial crisis. The U.S. financial system is fragile. Unless the features of financial disruption in 2008-09 were somehow only a symptom of some large underlying shock - a view which at this point seems a very long stretch - this seems clear. But where does that fragility come from? One view is that the fragility is inherent to financial systems. This view is framed in some versions of the Diamond-Dybvig model, in which the maturity mismatch and illiquidity inherent in banking imply that bank panics are possible. An alternative view is that the fragility is induced. It is well-known, for example, that deposit insurance induces risk-taking by banks that can be curbed through regulation - e.g. capital requirements. For small banks in the U.S., regulation is simple, because small banks are simple. Further, the failure of small bank is no big deal. In many cases, the FDIC can step in on a Friday, resolve the failed bank over the weekend, and have it open on Monday under new management.

Big banks and other financial institutions are another matter altogether. These large financial intermediaries are complicated and hard to regulate. Indeed, large financial institutions like Bear Stearns and Lehman Brothers could take on some of the characteristics of banks - e.g. maturity mismatch - while falling outside the regulatory umbrella that covers banks. As well, the failure and resolution of a large financial institution is a potential nightmare. The possibility of systemic disruption (which we saw in the financial crisis with the Lehman failure and the subsequent problems at AIG) from a large financial institution is what leads to the too-big-to-fail moral hazard problem. That is, a too-big-to-fail financial institution understands that it is too-big-to-fail, and therefore takes on too much risk, relative to what is socially optimal. This high level of risk could be reflected, for example, in a high-aggregate-risk asset portfolio, or in a maturity mismatch between assets an liabilities.

The too-big-to-fail problem is not only a long-run problem, but could manifest itself within a crisis period, as it may have in 2007-08. For example, it appears to have been well-known to regulators, and to Lehman Brothers itself, that Lehman was on shaky ground, perhaps a year or more before it failed. When Bear Stearns failed in March 2008, the Fed intervened in key ways, and helped engineer a sale to JP Morgan Chase. This could have created the expectation in other large financial institutions that their incipient failure would be met with the same type of intervention. Thus, it is possible that Lehman Brothers could have taken corrective action, including increasing capital and reducing dependence on short-term funding, to ward off failure, if it had correctly anticipated that a bailout would not occur.

Lacker states:
Lehman Brothers filed for bankruptcy on September 15, 2008. One day later, the insurance company American International Group (AIG) received a large credit extension from the New York Fed. At that point, regulators had resolved six large failing financial firms in five different ways. Some positions in the capital structure were rescued in one firm's resolution but not rescued in another's. Each had been handled on an ad hoc basis, without comment on how similar cases might be handled in the future. Market conditions following Lehman and AIG cried out for a general policy statement providing guidance on future interventions.
So this points to a policy failure. There was no explicit policy on how to deal with large financial insititutions, and those large institutions had to be confused about what would happen in the event of looming trouble. The most uncharitable view of Fed policy would be that, rather than saving the U.S. economy from the Great Depression, in the fall of 2008 Ben Bernanke was only cleaning up his own mess.

One might be tempted to think that too-big-to-fail is just an unsolvable commitment problem that we just have to live with. But, as Lacker points out
The thesis that financial market instability is inherent, rather than induced by poor policies, must also confront the fact that instances of instability are quite unevenly distributed across different countries and different regulatory regimes, as exemplified by the contrasting experiences of the United States and Canada. Over the past 180 years, the United States has experienced 14 major banking crises, compared with just two mild illiquidity episodes in Canada over the same period. If financial fragility were an inherent feature of financial markets, financial panics would be ubiquitous, but that's not what we see.
So, somehow the Canadians managed to solve the too-big-to-fail problem. While I'm quite willing to entertain the possibility that Canadians are superior human beings whose financial acumen is beyond the comprehension of Americans, I think we should entertain the possibility that there are things going on in the Great White North that we can easily replicate.

3. Economic researchers need to think more carefully about their policy advice. There is no hint in Lacker's speech that pre-financial-crisis economic theory was somehow lacking. Obviously we have learned a lot from the financial crisis, and that episode raised questions and moved some researchers in different directions. But the economics that Lacker brings to bear in understanding the events of the financial crisis was well in place long ago. That economic theory included information economics, the theory of incentives and insurance, the theory of financial contacts, and banking theory. While some people want to tell us that the financial crisis demonstrated that economics is a failure, I think we can safely argue that any failure was in the inability of regulators and policymakers to apply what was on the shelf.

But we can do better, of course. Lacker is critical of the tendency for economists to produce "possibility theorems," i.e. models which tell us that such-and-such can happen under such-and-such conditions, without worrying too much about how to apply such knowledge in practice. In my experience, contact between policymakers and researchers can cure that problem easily. And that was exactly what the conference at Arizona State University (organized by Ed Prescott), where Lacker gave his speech, was all about.

4. Central bank crisis intervention is not primarily about the overnight nominal interest rate and Taylor rules. This was not one of Lacker's bullet points, but this idea is implicit in his speech. Lacker talked for an hour on monetary policy in the financial crisis without mentioning the overnight interest rate. During a financial crisis, a decision about whether the fed funds rate target should be 2% or 1.5% is second order relative to who is borrowing at the discount window, how much, and at what rate; what Fed officials are saying to what CEO of which large financial institution; or what assets the Fed is buying and putting on its balance sheet.

38 comments:

  1. Two points:
    1."When Bear Stearns failed in March 2008, the Fed intervened in key ways, and helped engineer a sale to JP Morgan Chase. This could have created the expectation in other large financial institutions that their incipient failure would be met with the same type of intervention."
    The intervention at Bear Stearns bailed out the bond holders but the stock holders took a pretty big hit. They got $10 a share for a stock that had seen $133 within the previous year. (The original deal was for $2 a share, and Paulsen in his book regrets that they got a penny more). I'm pretty sure the management was turfed. It's hard to see how the management and shareholders of Lehmann could view this as an excuse to delay action...

    2. "So, somehow the Canadians managed to solve the too-big-to-fail problem".
    This probably has a lot to do with the industry structure of Canadian banks. It was still possible for them to make lots of money doing traditional banking business. (And with all those safe profits at stake, the option value of "putting" the bank to the tax payer was a lot less). But, because of competition, the traditional banking business model just wouldn't work for the large US banks.

    ReplyDelete
    Replies
    1. 1. Yes. It's clear that these financial institutions don't want to fail, as in that event the shareholders and management lose. The argument would have to be that there was the expectation on the part of Lehman that it would not fail because there would be a large intervention to prevent that (different from Bear Stearns - note AIG).

      2. "This probably has a lot to do with the industry structure of Canadian banks." Exactly. the industry is dominated by the 5 largest banks. They are as too-big-to-fail as Bank of America. But they do not fail, and are not bailed out either.

      You may have the idea of "charter value," that the banks are disciplined to through the threat of losing their charters and their monopoly profits. But the whole idea of too-big-to-fail is that the government cannot commit. They can threaten to revoke the charter, but they won't do it.

      Delete
    2. "The argument would have to be that there was the expectation on the part of Lehman that it would not fail because there would be a large intervention to prevent that (different from Bear Stearns - note AIG)." What should we note about AIG in this context? On 1/2/2008, its stock price was $912 per share. On 9/17/2008, the day after the U.S. government received a 79.9% equity stake in the firm (in return for the Fed's agreement to lend up to $85 billion), its stock price was $33.87 per share. Even more so than at Bear, AIG's stockholders took a massive hit.

      Delete
    3. What's the price of AIG stock now? What happened to its management after fall 2008?

      Delete
    4. AIG's stock price is currently at $49.45; the BS stockholders got paid out in shares of JPM stock, which was priced at $34.84 at the time of the BS deal and is now at $58.03. When the AIG deal went through, its head, R.B. Willumstad, was canned and replaced by E.M. Liddy.

      Delete
    5. It's hard to see how AIG stockholders benefited from its TBTF status; they got slammed. The major beneficiaries were AIG's CDS counterparties and FP division traders (kept on and paid -a lot- to unwind their positions).

      Delete
  2. Was TBTF also a problem in the crisis that proceeded the Great Depression?

    ReplyDelete
    Replies
    1. I think not. If you read Lacker's speech, he gives you the history of TBTF. It appears to be a post-1970 phenomenon.

      Delete
    2. Right. Seems that way to me too. So it doesn't seem obvious to me that TBTF was a necessary condition for the 2008 crisis. Plenty of bank-like institutions operated without deposit insurance in the early 2000s; why would these institutions not be just as vulnerable to a wave of bank runs as small, uninsured banks in the pre-Depression period?

      Delete
    3. I should note that my prior is that TBTF was a big problem in the 2008 crisis; I'm just being devil's advocate here.

      Delete
    4. On runs: Lacker talks about this. Gary Gorton thinks that you can think of a "run on repo" in the same way as a Diamond-Dybvig run on the demand deposits of a commercial bank. That's not clear. For shadow banks the assets are liquid, and there is nothing like the "sequential service" idea that you need to make DD work. You need to argue a fire sale phenomenon - some inefficiency that the central bank can exploit.

      Delete
    5. I was also thinking of the Gorton idea.

      Delete
    6. "You need to argue a fire sale phenomenon - some inefficiency that the central bank can exploit."

      I could be mistaken, but I thought that was Gorton's argument concerning mark to market being a bad idea in a crisis:

      "The panic was rooted in the fear of losses, the location and extent of which can't be determined. But there was also a virulent knock-on effect which is a significant force in its own right: liquidity dried up. With no liquidity and no market prices, the accounting practice of marking-to-market became highly problematic and resulted in massive write-downs based on fire-sale prices and estimates."

      Slapped by the Invisible Hand, page 128

      Delete
    7. I guess the question is whether you can write that down formally and have it make sense. I've seen plenty of models that claim to have fire sales in them, but I'm not sure I like any of them much.

      Delete
    8. Cochrane is a fan of the "bank run" theory, but not sure what model he has in mind...

      http://johnhcochrane.blogspot.com/2013/06/stopping-bank-crises-before-they-start.html

      Delete
    9. The run model is Diamond-Dybvig, and DD essentially captures conventional wisdom (based on US experience) about what "bank run" means. That's why people liked it so much - even people who don't actually understand how the model works.

      Now the shadow-bank-run story for the financial crisis has become conventional wisdom, and that's what Cochrane is echoing here.

      Delete
    10. "For shadow banks the assets are liquid, and there is nothing like the 'sequential service' idea that you need to make DD work."

      You lost me here. Why aren't shadow bank withdrawals characterized by sequential service? If a securities firm has a small cash reserve but lots of repo liabilities, then you can imagine self-fulfilling redemptions by repo creditors, as in Diamond-Dybvig.

      True, securities firm asset portfolios are liquid relative to bank asset portfolios, but there are degrees of liquidity; it isn't a binary thing.

      Delete
  3. The Savings and Loan Crisis, which cost the public far more in terms of inflation adjusted bailout dollars seems to refute your TBTF argument.

    There is not any practical free market (small bank) or regulatory solution to the problem as far as I can see. Canada is not comparable - both on scale and on the static, oligarchic nature of banking that Canadians seem willing to tolerate. The obvious solution - more public banking - is, on the other hand, easy.

    ReplyDelete
    Replies
    1. The Savings and Loan Crisis came about because of the standard moral hazard problem. The S&Ls were deregulated in 1980, and there insufficient subsequent regulation of their risk-taking, and that blew up later in the 1980s. This was not about TBTF, but the root of the problem is the same.

      Sure, the U.S. is not Canada, but what you learn from Canada is that having fewer regulators helps, combining investment banking and commercial banking doesn't hurt, and high capital requirements and close scrutiny are important.

      What do you mean by "public banking?"

      Delete
    2. The problem is that banking is inherently unstable and any large scale economy is vulnerable to cascading banking failures. I don't think this has anything to do with the scale of the banks - we have had banking panics since the 1300s.

      The moral hazard is intrinsic to any system of regulation and government bank insurance/central bank lending, no? Under Clinton and particularly Bush, the US system was very "lightly" managed, but it seems to me that there is a paradox inherent to any banking system: you cannot avoid really catastrophic banking failure without moral hazards at some level.

      I think much of the problem of bank regulation could be reduced by offering e.g. a postal bank alternative to the market. The public bank would invest entirely in government debt. This would serve to damp out network effects of private banking failures and so the public could tolerate bank panics without bailouts to a much greater degree.

      Delete
    3. ""we have had banking panics since the 1300s."

      Not in Canada. Banking panics have been non-existent there. So it's not inherent.

      "The public bank would invest entirely in government debt."

      You could carry this one step further, and give that job to the Fed. That is, everyone holds a reserve account with the Fed, which looks just like a checking account, there are Fed ATMs, etc.

      Your postal bank falls under the category of "narrow banking," and narrow banking proposals have been around for a long time -e.g. it's in Friedman's "Program for Monetary Stability." The arguments against narrow banking are two:

      (i) It's easy to get around the regulations. Given the regulatory structure, financial intermediaries find ways to look marginally different from what is prohibited, though they are essentially carrying on some of the important functions of a commercial bank.

      2. Narrow banking means you forego some socially useful financial intermediation. Bank deposits are backed by loan portfolios that are financing investment; narrow bank liabilities are backed by government debt.

      Delete
    4. Canada is too small and static to be a model for the US - and, in fact, it is the model of TBTF.

      Yes, I like the Fed idea - postal makes it easy to have brick&mortar sites too.

      Your objection (1) seems directed at proposals for private banks that are regulated to be narrow rather than a public bank run by civil servants.
      Objection (2) - you could still have private, and even less regulated investment/commercial banking, but the public would not be so concerned about cascading failures in that system because clearing/savings would be outside of it.

      http://krebscycle.tumblr.com/post/37844600416/a-modest-proposal-for-free-market-bank-reform

      Delete
    5. 1. Too small and static? That's a country of 35 million people which is almost as rich in per capita terms as the US. How big do you want it to be? I have no idea what you mean by "static."

      2. Private banks can be narrow now if they want. So why don't they do it?

      3. Question: Which is more efficient, USPS or Fed Ex?

      Delete
  4. Stephen,
    I think the change in TBTF institutions was the move from traditional to shadow banking post 2000. This is roughly the period when repo's exploded in size, and overnight repos grew to about 40% of outstandings. The question is, why did large financial institutions begin to hold such large volumes of securities? This is not a question of securitization or innovation. That is, banks and investment banks had been engaged in securitization for a decade without placing those securities on their own balance sheets. Something happened circa 2000/2001 that encouraged them to book these instruments for a profit. I argue that it was Fed forward guidance. This guidance created quasi-arbitrage profits from the duration carry trades: funding short and lending long. The main asset held was a long-duration, low credit risk asset -- a AAA bond. This is because the Fed signal had information content for the duration trade but not the credit risk trade. As Gorton documents, the shadow banking boom was a AAA phenomenon.

    ReplyDelete
    Replies
    1. Why can't we explain these developments by invoking technological change? How could Fed behavior generate a permanent profit opportunity for this kind of activity?

      Delete
  5. The business of shadow banking involved credit scoring and pooling and tranching. These were two technologies that had been around for at least a decade. Credit default swaps were also not new.

    Think of the information contained in forward guidance: "your funding costs will remain known for the foreseeable future (extended period), after which they will rise at a predictable rate (measured pace)." What is the likely, indeed even the desired, effect of this information on speculator strategies? How is this affected or amplified by herding dynamics? Does the profit opportunity have to appear to be "permanent" to have this impact?

    The important thing, I think, is to ask the question: in a few short years financial system strategies converged towards extreme balance sheet growth via shadow banking. What caused this? If someone thinks financial innovation was at work, I'd like to hear the argument and evidence.

    ReplyDelete
    Replies
    1. The big change in Fed policy was post-Volcker, and this didn't happen until post-2000. Do you think that the Fed can keep real rates low for long periods of time? Seems hard to explain all that growth as being due to monetary policy.

      Delete
    2. Regarding financial innovation, there has been much discussion about the role of the Gaussian copula function in pricing derivatives. See also here:
      http://www.wired.com/techbiz/it/magazine/17-03/wp_quant?currentPage=all

      Delete
  6. From a forward guidance perspective, the big change occurred with Greenspan circa 1998 and, again, in 2001.

    Monetary policy operates by sending signals to financial intermediaries. A broad class of these intermediaries is speculators: i.e. they speculate, for instance, on the yield spread. Real rates are immaterial to these strategies. Further, the informational value of forward guidance is low for predicting the real rate.

    Shadow banking was a form of speculation on the yield spread. There was no asymmetric information about a borrower's credit quality, for instance. All shadow bankers had was information about the volatility of short term interest rates. This information was, in fact, the "innovation". It had never been provided before. So, yes, there was new technology, only it was a central banking, not a banking, one.

    By the way, I'm not saying, "all that growth" was due to monetary policy. I am saying 1) unprecedented growth occurred; 2) unprecedented policy occurred; 3) that policy had information value that can be linked to that growth.

    ReplyDelete
    Replies
    1. I'm not sure why you are focused on explicit forward guidance. I think post-Volcker there was an implicit commitment that people understood.

      Delete
  7. Pre-'98, give an indication of the inter-meeting bias. This later shifted to the "balance of risks" statement. These cannot compare to extended period and measured pace in terms of its impact on speculator strategies.

    From a speculator's standpoint, the relevant information was the signal about future rate volatility, one that had never been sent before. This is surely not controversial.

    ReplyDelete
  8. In any case, I'd like to see competing explanations for the shadow banking explosion: the "high aggregate risk asset portfolio" you write about. This was, in fact, not a "high-risk" portfolio in the credit risk sense. AAA-rated securities were thought to be practically risk-less. The only (ex ante) risk these portfolios carried was duration rise. Why such a high speculative demand for duration risk?

    ReplyDelete
  9. For the ASU conference, just wondering what Prescott presented on "Mechanism Design Theory and the Theory of Value for Fractional Reserve Banking".

    ReplyDelete
    Replies
    1. Nothing new. He thought narrow banking was a good idea.

      Delete
  10. The solutions are simple: increase the capital ratios (Hellwig has a great book out on the topic), separate commercial and investment banking

    ReplyDelete
    Replies
    1. "...separate commercial and investment banking"

      So why does it work in Canada when they don't do that?

      Delete
    2. No idea, you are the expect on Canada. I guess not having a firewall between commercial and investment banks matters less when there is more competition in banking.

      Delete
  11. I simply want to tell you that I am new to weblog and definitely liked this blog site.

    ReplyDelete