Recent Posts

Friday, October 21, 2022

Runs, Cascades, Squeezes, and Crises

Recent discussions about bank runs in economics has led me to the conclusion that economists want to lump practically all behaviour around default as a “run,” which is not helpful. In order to have a better grasp of the situation, we need to distinguish various possibilities — runs, cascades, squeezes, and good old fashioned financial crises.

Preferred Definitions

I would divide these concepts as follows.

  • A run is what happens when a borrower that has multiple sources of funding has those sources pull their funding. In order for such an entity to be forced to default, a run has to occur, almost by definition. (It is possible that a firm can voluntarily enter into default while funding is still available, since it might be better positioned under bankruptcy protection.)
  • A cascade is a situation where one default (or projected default) weakens the financial profile (or projected profile) of another entity. This can either be within some segment of the financial system, or in one or more sectors of the real economy (typically real estate). Cascades offer a fundamental reason why entire sectors of the credit markets can move together in a coordinated fashion — a “run” on a sector.
  • A squeeze is a situation where one or more entities in the financial markets (including commodity markets) has a large position in a type of asset and other entities are aware of the position. The other entities then push the price of that asset either up or down in an effort to cause mark-to-market losses that forces a liquidation. The liquidation is forced as a result of a liquidity drain caused by the need to post collateral against the market losses, or alternatively, risk limits would be hit. (One of the perverse problems of risk management is that the risk of the position is normally scored as increasing as the position moves against you.) What makes a squeeze interesting is that the price movements do not have to bear any sensible relationship to the fair value determined by “fundamentals” — although there is usually something one can point to in the public news flow that allegedly justifies the movement.
  • A financial crisis is what happens when a cascade and/or squeeze puts the projected financial profile of core banks in a country at risk.

To rephrase a point made above: the distinction between a run and a squeeze is that runs are largely based on the fundamentals of the entity facing the run, while a squeeze is based on the fundamentals of targeted participants of the market instruments are traded in. So long as the uncertainty/disagreement about the fair value of the instruments is not too small relative to the deviation from fair value, the state of the issuer of the instruments in the squeeze is of secondary importance. To the extent that we believe that markets engage in “price discovery” with the price converging to “fair value,” markets during a squeeze are dysfunctional or disorderly.

As an aside, the existence of squeezes explains why fixed income and derivatives markets tend towards greater opacity than things like (cash) equities. Equity investing features a wide disparity of actors, they have large risk tolerances, and leverage used tends to be lower. Although short squeezes occur, their effects tend to be isolated. Conversely, fixed income and derivatives tend to highly concentrated, leverage is widely used, and is therefore more susceptible to squeezes. Opacity allows large players to make large transactions between themselves: both sides have concentrated risk positions (of opposite signs), and until they can lay off the risk, they both have an incentive to keep the existence of the transaction out of trader chatter channels.

The Problem With Run-Centric Analysis

The methodological ideal for neoclassical economists is to have a model where agents trade fixed endowments to optimise utility in an unfettered manner, because allegedly everything works out optimally for society. Credit relations have matched assets/liabilities that pop up out of thin air, and those assets/liabilities vary by agent. It takes a lot of work to force something that vaguely looks like a bank run into the framework.

All this theoretical effort distracts people from the essential triviality of the observations one can make. Pretty much every default of a borrower with multiple sources of finance is the result of a run or the borrower pre-empting an expected run. Defaults happen all the time. Unless you invest in the defaulted entity, not that remarkable an event.

The usual solution to runs — a lender-of-last-resort — does not matter for squeezes, and possibly default cascades in the real economy.

  • In the model of a run, the price of credit assets drops because of “multiple equilibria.” In a squeeze, what matters is the price of some asset(s). They are dropping below “fair value,” (or above fair value if shorts are being squeezed), but the entities that might want to buy cannot do so because they have probably hit risk limits, and liquidity might be secondary. (In the recent gilt market kerfuffle, the problem appears to be more of a margin liquidity issue.) Meanwhile. other knowledgeable market participants are aware of the squeeze, and are either part of it, or they are not going to jump in too early to save the day — finance is not a charity business. The price needs to rebound towards fair value to force the squeezing entities to close their positions (“the hunter becomes the hunted”).

  • In a default cascade in the real economy, a lend-of-last-resort cannot fix negative equity of failed borrowers, it can only contain the damage to others.

Traditional Bank Versus Non-Bank Finance

Loan books at traditional banks do not face daily mark-to-market, nor is there any expectation that they can be sold at the whim of risk managers. As a result, the “traditional banking” divisions of banks are immune to squeezes. However, large banks in the real world rely on non-bank markets for risk management and funding, so they can run into problems in those markets. (Smaller community banks can operate on traditional lines.)

This immunity to squeezes does not come without a cost. The ability to bury losses and “extend and pretend” with problem borrowers can lead to a zombie banking system. Banks clamp down on risk, and use incoming net cash flow to slowly liquidate insolvent clients. Part of the rise of non-bank finance was the desire to get risk off traditional bank balance sheets to avoid that outcome (as well as dealing with pre-existing problems).

The Financial Crisis Of 2008

I am generally unimpressed with economist stories about the Financial Crisis. There are two common failings. One is over-simplifying what happened in favour of some pet theory. The other is digging deeply into some arcane aspect of the crisis (“CDOs!”), where most of the discussion is explaining details.

As an involuntary front seat observer of that crisis, the only way to understand it is from a high level perspective that takes into account the decision-making processes of the larger actors. The details of the various toxic assets matters, but this is because of how they affected the risk management framework. The high level perspective is that 2008 was a mixture of both a cascading credit failures mixed with a squeeze in risk assets, with the squeeze being both for margin calls and risk limits. Everything went wrong, so there is no point in focussing on one set of instruments.

Following the lead of finance academics, leverage (both explicit and implicit) built up throughout the financial system in the run up to 2007. The system started unravelling about them, but this only became obvious to people outside the funding markets when the brain trust in the United States decided that throwing Lehman Brothers under the bus would be a good idea (in 2008).

Since large investors were impaired, they needed to reduce risk in every market they were involved in. This implicated most of the credit complex. Everything got ridiculously cheap versus pre-crisis fair values.

Even if investors wanted to take some risk (and bailing out the squeezed institutions by stabilising prices), they rapidly hit their risk limits. The toxic assets created by deranged financial engineers were burning a hole in balance sheets they were on, and nobody else wanted any more of them.

Lending money to market participants (including bank holding companies) might have alleviated stress due to derivatives collateral needs, but it would not do anything about risk limits. The Fed had to backstop prices in various structured products by getting them off investor balance sheets.

Although the Fed’s purchases caused controversy, there were not a lot of alternatives at that point in time. (If they had not listened to people like Ben “Subprime is Contained” Bernanke, it might have been possible to take less drastic action earlier.) Arguing that the “private sector can take the losses” is exactly the attitude that led to the demise of Lehman — there was no way for credit markets to function while the losses were being meted out. Sure, the financial markets could have left to their own fate, and after six months of zero credit activity, things might have been restored. However, the drop of real economic activity would have been even worse than what actually transpired.

Modelling Difficult (Of Course)

We cannot model runs and squeezes the same way — runs reflect conditions for the borrowing entity, while squeezes reflect the risk/liquidity management practices of market participants. I have no doubt that somebody could come up with a toy model of a market with a squeeze, but such a model does not answer the “so what?” question. Only a neoclassical economist would think that a model that allows for the existence of something that we know happens is a useful applied mathematical model; the model needs to have some predictive powers in the real world.

The issue such a hypothetical model faces is complexity: it needs to simultaneously model the fundamentals of the entity issuing the instruments, as well as the conditions in the market for the instruments. Even if we build such a model (such as an agent-based model), it would have too many behavioural parameters to fit to the data. If we do not fit the data, we are stuck with “squeezes can happen,” which we already knew.

Can We Do Better?

One of the structural problems of our society is that it is too inherently optimistic. Nobody wants to hear that a problem cannot be fixed. Instead, the focus is always on how to fix problems. In the case of financial crises, we see solutions that can be predicted by the ideological sympathies of the person giving the solution. On the left, capitalism/finance/markets are the problem, getting rid of them is the solution. On the free market side, just deregulate and the markets will give optimal outcomes. And the middle ground “serious” camp offers technocratic solutions based on highly technical twiddling around of whatever blew up in the previous crisis.

Capitalism is unstable. That’s the message of my “Recessions: Volume I,” which is just my popularisation of ideas that I largely got from Hyman Minsky. Minsky in turn points to Keynes/Kalecki. Fixed investment adds to business sector profits — which justifies investment. This creates an investment accelerator, leading to borrowing to invest. As Minsky emphasised, previous successes leads to the erosion of any conservative norms in lending, leading to more lending, and thus more profits. Unfortunately, lending standards eventually go too far, and the process reverses rapidly.

Unless fixed investment is socialised, the real side of the private sector will blow itself up using the investment accelerator. The financial sector is at best a glorified casino, and it can blow itself up on its own. However, disturbances that are purely within the financial sector have roughly the same effect as gambling within the real economy — there are winners and losers, but life goes on for outsiders. The problems show up when the real economy credit losses impairs finance — then everything shuts down real quick.

Squeezes happen all the time. In my writing, I try to distinguish between “fair value” and “observed market prices.” Under normal circumstances, we do not have a good reason to believe that the observed market price is radically different than fair value. However, a squeeze is a situation where that is the case. It is difficult for outsiders to identify a squeeze, but the whole point of a squeeze is that the market participants generally know about it — otherwise they would just lean against the silly pricing.

If the central bank wants to direct the economy with interest rate policy, it has little choice but to ensure orderly markets in central government bonds. This is also of interest to the fiscal arm of government. The Bank of England sort of figured this out, but as the small size of the intervention demonstrates, it does not take a whole lot of purchases for a central bank to calm an unruly bond market. The massive bond purchases of the QE era was just the result of Monetarists running amok, and not needed to restore order to government bond markets.

The more serious problems are the ones involving the meltdown of private credit markets. Markets can self-correct, and so the central bank can get away with allowing some turbulence. Periodic default events keeps everyone on their toes. However, if pressures get too large, the choice is either to intervene or let a major component of the real economy collapse. The trend towards tighter risk management limits the ability of the private sector to lean against a crisis: the bureaucratic overhead required to increase risk budgets will not allow timely responses to market conditions. Only the central bank has dual advantage of a risk budget that is best measured as a percentage of GDP and has teams operating in markets on a daily basis (unlike the fiscal arm of government).

Financial chaos does lead to less risk taking in the private sector. However, memories fade, and there is natural selection towards more aggressive lenders during an expansion. Problems will show up somewhere, typically in areas that were not directly involved in previous crises, and thus escape whatever regulations were put in place in response. The best you can hope for is that central bankers can act as responsible grown ups when supervising the financial sector.

Email subscription: Go to 

(c) Brian Romanchuk 2022

No comments:

Post a Comment

Note: Posts may be moderated, and there may be a considerable delay before they appear.

Although I welcome people who disagree with me, please be civil.

Please note that my spam comment filter appears to dislike long "anonymous" posts.

Note: if you want to post comments from Apple devices (iPhone, iPad), you apparently need to turn off "prevent cross-site tracking" in Safari privacy settings. (The reason presumably is that another URL handles comments, and so the user session needs to be preserved when redirected to that site. I don't like this, but this is not enough to make me switch my hosting service.)