Crisis is the normal state

Strangely, I’ve read a lot more about the 3 year recession/depression in 1893 than the 1975 NYC bankruptcy… are they really comparable events?

The events you have listed are financial crises but not bank panics. As I understand the term, a bank panic is where a crisis of confidence in the ability of one or more banks to repay its depositors results in systemic withdrawals across the banking system. Rumours that banks are not sound thus leads to massive withdrawals, which leads to the collapse of otherwise healthy banks and losses to depositors.

None of the events you listed resulted in systemic runs on banks, so I don’t think it’s fair to call them bank panics in particular, as opposed to financial crises in general.

Along previous lines, I’d dispute most of the crises before 1997, which were basically just big defaults by non-financial-sector debtors (the S&Ls weren’t systemically important). From 1997 onwards, we were seeing rehearsals for the GFC, so I agree with you and disagree with Krugman about those.

The collapse of Bretton Woods was certainly a huge event, but it was a success for financial markets (breaking down government restrictions) not a failure.

@ John Q –

Sure, but “defaults by non-financial-sector debtors” are what happens right before “defaults by financial-sector debtors”. When governments default on their debt, banks holding that debt are put under pressure. When subprime borrowers default on their mortgages it has major ramifications for banks holding those loans. If counterparty risk is high you get crisis.

One definition of “systemic crisis” hinges on whether or not the government felt the need to bail out financial firms to prevent contagion. Bailouts are politically unpopular, so (presumably) governments won’t do it unless they have to. Under that definition several of the pre-’97 events qualify: ’75, ’82, ’85 and onward (747 failed S&Ls and the ’90-’91 recession qualifies as systemically important I think), ’94. Maybe others, but those are the ones I know off the top of my head. In any case, Dimon’s definition — which echoes Kindleberger and probably came from him — holds up much better than Krugman’s.

Bretton Woods was a different animal of course, but we’re not going back there.

the S&Ls weren’t systemically important

Are you kidding me?

I think any crisis that ends with interest rate policy being dictated by the need to keep banks afloat is reasonably described as systemic so the early 90s has to be in there. Even Goldman had to be recapitalised and that was when investment banks weren’t banks and didn’t need much capital.

I’m surprised that John wants to dismiss the S&L crisis so lightly. Even if it seems small potatoes compared to what we’ve just experienced, it is a pretty good prototype. Isn’t the GFC just the ‘seventh wave’ of financial crises? In so far as it is different to the very many other crises it just involved larger, more globalised and faster moving organisations.

The S&L crisis was systemic, for sure. It was a major factor leading to the early 90s recession.

On Gorton, here’s a crosspost of something I posted on Yglesias’s blog:

Look, Gary Gorton is an excellent economist and this book has a lot to recommend it. But he was also the MAIN RISK MODELLING CONSULTANT FOR AIG, responsible for doing their credit default swap risk modelling. He was paid a million dollars a year for that. His picture of the crisis is of a collateral run on assets that really weren’t so bad (specifically, on firms that would have been able to stay healthy given the normal decline in asset values absent the run). There’s definitely truth to that, there was a bank run. But it is also the model that makes his own risk forecasts look the best. He didn’t call what happened the housing market at all, but he can say this is because of the underlying financial instability instead of the fundamental weakness, the overexposure to debt on the part of the American consumer.

Dsquared in 15: Yeah, there’s that, there’s understandable irritation at his role in the crisis. But I welcome all converts, even late ones. It’s that when you listen to him talk about this, he really hammers on his belief that the decline in the value of the underlying assets would not have been enough to trigger massive CDS payments, it was the run on collateral that did the damage. In other words, his credit risk modelling was right, and if there had been a mechanism in place to stop the run (e.g. government backstop on collateral, like deposit insurance) then it would have been proved correct in the long run. It’s the old solvency vs. liquidity argument. (I don’t really believe one can counterpose those two as opposites, but that’s the role they end up playing).

He might be correct, but I wonder about the ways in which his own involvement has affected his views on this. I haven’t read the book, but there are a number of obvious issues with introducing deposit-insurance type mechanisms to back up financial leverage. If we did have a government collateral backstop, should responsible regulators have permitted the credit risks to be taken that were taken?

As there wasn’t a government collateral backstop doesn’t it still amount to an admission that he failed to do his job correctly – he ignored/forgot/didn’t factor in at least one type of risk. Surely the point of risk management is to rate and price ALL risks, not just the ones you think you can most easily model. Of course one of the fundamental problems of the free market in these types of insurance is that the most optimistic risk models gets the majority of the business, so especially in a bonus led corporate culture the boom/crisis/bust cycle seems likely to continue for the foreseeable future.

Daniel, the key is magnitude. How big were these various crises? For example, if the relative sizes were from 1-5 between the late 1930’s and 1980, and then 10’s during the 1980’s and 90’s, and then 100 for the 2007 crash, then Krugman’s point is valid.

Re the crash of 1857
Dec. 8, 1857, Marx to Kugelmann

That the capitalist, who scream so much against the right to employment, now everywhere demand public assisttance from their governments in Hamburg, Berlin, Stockholm, Copenhagen, even England [in the form of the suspension of bills, thus seeking to enforce a “right to profit” at the public’s expense, is beautiful.”

dsquared, I’d be tempted to assume that Krugman misspoke – writing “systemic crises” when he meant “bank panics”. There were lots of financial crises during that period, but how many bank panics?
Second, I don’t think you could really count Black Wednesday as a systemic crisis without weakening the term to the point of uselessness. It was just a stock market crash. Even if it threatened or brought down a few banks, it didn’t IIRC threaten the entire banking system – hence not ‘systemic’.
As for Bunbury’s definition, a lot of financial disturbances lead to monetary policy responses just because they lead to economic downturns, which tend to produce monetary policy responses. But I don’t think that cutting interest rates after the dotcom bubble (say) was aimed specifically at keeping banks afloat rather than just at helping the economy and keeping inflation near target.

But changing the definition to “banking crises” (and I am still claiming the S&L crisis) seems to me like a “Golden Arches Theory Of Geopolitics” move – it changes the assertion from an interesting statement about the world to a pub quiz question. When the banks are guaranteed by an OECD sovereign, fair enough they don’t have panics. But the question is – did this reduce the overall level of financial-sector-caused risk in the system as a whole?

There is reasonable evidence that there was greater macroeconomic stability in the postwar period (or in other words, that the invention of stabilisation policy did work – although this is still controversial and by no means everyone agrees with me on this one). But if what we saw was a situation in which a) most of the time, banks were more stable simply because the macroeconomy was more stable and b) some of the time, stabilisation policy was actually bent around the objective of trying to counteract the effect of financial crises, then I don’t think that supports a view of a more stable financial sector post 1934. If you then add c) (a proposition I don’t necessarily believe myself but it’s at least plausible) that a lot of the stability of the “Great Moderation” era was just the building up of a financial bubble, facilitated by the banking system and contributing to a massive crash in 2007, then I think the underlying Gorton/Krugman hypothesis really does fall down.

I also want to make another objection based on Goodhart’s Law – which is that you can’t just say that “we had a nicely working stable and well-regulated financial system, and then this awful shadow-banking thing came along”. Despite the name, shadow banking wasn’t created as a bane on mankind by the Lord of Mordor – it’s an inevitable and predictable consequence of the regulatory regime. Another application of the POSIWID principle.

But the question is – did this reduce the overall level of financial-sector-caused risk in the system as a whole?

This is a very good question. How would you go about answering it? You’d want some sort of economy-wide VAR measure or something, to measure “overall risk in the system”, and then you’d have to have some way of dividing the risk into “financial sector caused” and “other” (which I suppose would be things like, oh, the movement in oil prices caused by Opec reducing production in 1973, to pick one obvious example).
Any thoughts?

I thought #17 deserved an answer, if only as an acknowledgement of their referencing Tombstone Blues.