An information-based approach to understanding why America let Lehman Brothers collapse but saved everyone afterwards

In addition to his previous comments on the bailouts [25 Aug27 Aug28 Aug], which I highlighted here, Tyler Cowen has added a fourth post [2 Sep]:

I side with Bernanke because an economy can withstand only so much major bank insolvency at once. Lots of major banks were levered up 30-1 or so. Their assets fell in value more than a modest amount and then they were insolvent, sometimes grossly so. (A three percent decline in asset values already puts you into insolvency range.) If AIG had gone into bankruptcy court, some major banks would have been even more insolvent. Or if Frannie securities had been allowed to find their non-bailout values. My guess is that at least 15 out of the top 20 U.S. banks would have been flat-out insolvent if, starting at the time of Bear Stearns, all we had done was loose monetary policy and no other bailouts. Subsequent contagion effects, and the shut down of short-term repo markets, and a run on money market funds, would have made even more financial institutions insolvent. The world as we know it then becomes very dire, both for credit reasons and deflation reasons (yes you can print up currency to keep measured M up and running but the economy still collapses). So we needed not just emergency lending but also resource transfers to banks, basically to put them back into the range of possible solvency.

I really like to see Tyler’s evolving attitudes here.  It lets me know that mere grad students are allowed to not be sure of themselves. 🙂  In any event, let me present my latest thoughts on the bailouts:

Imagine being Bernanke/Paulson two days before Lehman Brothers went down:  you know they’re going to go down if you don’t bail them out and you know that to bail them out creates moral hazard problems (i.e. increases the likelihood of a repeat of the entire mess in another 10 years).  You don’t know how close to the edge everyone else is, nor how large an effect a Lehman collapse will have on everyone else in the short-run (thanks, in no small part, to the fact that all those derivatives were sold over-the-counter), but you’re nevertheless almost certain that Lehman Brothers are not important enough to take down the whole planet.

In that situation, I think of the decision to let Lehman Brothers go down as an experiment to allow estimation of the system’s interconnectedness.  Suppose you’ve got a structural model of the U.S. financial system as a whole, but no empirical basis for calibrating it.  Normally you might estimate the deep parameters from micro models, but when derivatives were exempted from regulation in the 2000 Commodities Futures Modernization Act, in addition to letting firms do what they wanted with derivatives you also gave up having information about what they were doing.  So instead, what you need is a macro shock that you can fully identify so that at least you can pull out the reduced-form parameters.  Letting Lehman go was the perfect opportunity for that shock.

I’m not saying that Bernanke had an actual model that he wanted to calibrate (although if he didn’t, I really hope he has one now), but he will certainly have had a mental model.  I don’t even mean to suggest that this was the reasoning behind letting Lehman go.  That would be one hell of a (semi) natural experiment and a pretty reckless way to gather the information.  Nevertheless, the information gained is tremendously valuable, both in itself and to society as a whole because it is now, at least in part, public information.

To some extent, I feel like the ideal overall response to the crisis from the Fed and Treasury would have been to let everyone fail a little bit, but that isn’t possible — you can’t let an institution become a little bit bankrupt in the same way that you can’t be just a little bit pregnant.  To me, the best real-world alternative was to let one or two institutions die to put the frighteners on everyone and discover the degree of interconnectedness of the system and then save the rest, with the nature and scale of the subsequent bailouts being determined by the reaction to the first couple going down.  I would only really throw criticism at the manner of the saving of the rest (especially the secrecy) and even then I would be hesitant because:

(a) it was all terribly political and at that point the last thing Bernanke needed was a financially-illiterate representative pushing his or her reelection-centred agenda every step of the way (we don’t let people into a hospital emergency room when the doctor isn’t yet sure of what’s wrong with the patient);

(b) perhaps the calibration afforded by the collapse of Lehman Brothers convinced Bernanke-the-physician that short-term secrecy was necessay to “stop the bleeding” (although that doesn’t necessarily imply that long-term secrecy is warranted); and

(c) there was still inherent (i.e. Knightian) uncertainty in what was coming next on a day-to-day basis.

The three best things I’ve read on the US car (auto) bailout …

… are this opinion piece in the FT by Joseph Stiglitz, this brief blog entry by Matthew Yglesis and this blog entry by Robert Cringley.

Stiglitz’s piece makes, to me, a compelling argument for letting the firms go into Chapter 11 bankruptcy, albeit (given the state of the market) with government guarantees for any further financing they may need for restructuring. The following four paragraphs are among the most succinct and clearly written on the US car industry:

 Wall Street’s focus on quarterly returns encouraged the short-sighted behaviour that contributed to their own demise and that of America’s manufacturing, including the automotive industry. Today, they are asking to escape accountability. We should not allow it.
[…]
The US car industry will not be shut down, but it does need to be restructured. That is what Chapter 11 of America’s bankruptcy code is supposed to do. A variant of pre-packaged bankruptcy – where all the terms are set before going before the bankruptcy court – can allow them to produce better and more environmentally sound cars. It can also address legacy retiree obligations. The companies may need additional finance. Given the state of financial markets, the US government may have to provide that at terms that give the taxpayers a full return to compensate them for the risk. Government guarantees can provide assurances, as they did two decades ago when Chrysler faced its crisis.

With financial restructuring, the real assets do not disappear. Equity investors (who failed to fulfil their responsibility of oversight) lose everything; bondholders get converted into equity owners and may lose substantial amounts. Freed of the obligation to pay interest, the carmakers will be in a better position. Taxpayer dollars will go far further. Moral hazard – the undermining of incentives – will be averted: a strong message will be sent.

Some will talk of the pension funds and others that will suffer. Yes, but that is true of every investment that has diminished. The government may need to help some pension funds but it is better to do so directly, than via massive bail-outs hoping that a little of the money trickles down to the “widows and orphans”.

I would perhaps suppliment Professor Stiglitz’s words by proposing that government support to workers laid-off as part of the restructuring could be improved dramatically over the provisions currently available. They should not only include lengthening the duration of unemployment payments and paying for retraining programmes, but also payments to help with relocation if anybody is willing to (voluntarily!) move to find work. An Obama administration might also be reasonably expected to look to Michigan for skilled manual labour in it’s push for infrastructure renewal/expansion.

Yglesis’ brief note observes a vital co-ordination problem when it comes to restructuring what is genuinely a global industry:

One thing here is that as best I can tell none of the five countries — US, Japan, Germany, France, Korea — with substantial auto industries are willing to let their national favorites fail. And yet there seems to be substantial global overcapacity in car manufacturing. If a few of the existing firms are allowed to fail, then the survivors will be in good shape. But if nobody fails, then all the firms worldwide will be left suffering because of overcapacity problems, all potentially drawing bailouts and subsidies indefinitely.

Finally, Cringely’s piece investigates how a successful US car firm ought to be run by imagining that Steve Jobs (of Apple) was running it.  The idea is not his.  Thomas Friedman briefly mentioned in early November that …

… somebody ought to call Steve Jobs, who doesn’t need to be bribed to do innovation, and ask him if he’d like to do national service and run a car company for a year. I’d bet it wouldn’t take him much longer than that to come up with the G.M. iCar.

It was something of a trite comment, and it was picked up by many people in the IT industry who got a little over-excited when imagining the details of what functionality the iCar should have (for example).  In contrast, Cringely looked at the most important thing that somebody trying to emulate Apple might bring to the car industry:  it’s design and manufacturing process:

… embracing these [new technologies] requires the companies do something else that Jobs came to embrace with Apple’s products – stop building most of their own cars.

There are two aspects to this possible outsourcing issue. First is the whole concept of car companies as manufacturing their own products. There is plenty of outsourcing of car components. Most companies don’t make their own brakes, for example. Yamaha makes whole engines for Ford. Entire model lines are bought and rebadged from one maker to another. But nobody does it for everything, yet that’s what Steve Jobs would do.

All the U.S. car companies are closing plants, for example, and all are doing so because of overcapacity. But what would happen if just one of those companies — say Chrysler — decided that two years from now it would no longer actually assemble ANY of its own vehicles? Instead they’d put out an RFQ to every company in the world for 300,000 Chrysler Town & Country minivans as an example. Now THAT would be a dramatic move.

And a good one, frankly, because with a single pen stroke most of the overcapacity would be removed from the U.S. car market. Chrysler would have to shut down all those plants and lay off all those people, true, but doing it all the way all at once would change the nature of the company’s labor agreements such that there wouldn’t be a whimper. When you are eliminating 8 percent of capacity the tussle is over WHICH 8 percent. When you are eliminating ALL capacity, there is no tussle.

So Chrysler reaches out to contract manufacturers in this scenario and you know those manufacturers would fight for the work and probably give Chrysler a heck of a deal. For current models, for example, Chrysler could probably sell the tooling and maybe even the entire assembly plant for a lot more than they’d get from the real estate alone. But that particular advantage, I’d say, would be unique to the first big player to throw in the production towel.

In this scenario, Chrysler becomes a design, marketing, sales, and service organization. What’s wrong with that? They can change products more often and more completely because of their dramatically lower investment in production capital. They can pit their various suppliers against each other more effectively than could a surviving car manufacturer. It’s what Steve would do.

This is brilliant stuff.