Thursday, October 20, 2016

TBTF isn't the problem. TSTS is.

TSTS = Too Small to Save

The problem with Too Big to Fail, as generally described, is that we have let certain financial firms become so powerful that we were forced to save them in the crisis.  But, the crisis was a liquidity crisis.  The collapse of the housing market and the related collapse of the subprime mortgage market were products of contractionary monetary and credit policies.  This shouldn't be a controversial statement.  Policy makers at the Fed and the Treasury, plus just about everyone who either supports or questions Fed policies during the boom and bust, have explicitly stated that home prices had to collapse, that the Fed could not provide support for nominal economic activity in 2006 and 2007 because that would be bailing out irresponsible investors and lenders.  It is a matter of public consensus that policies in 2006 and 2007 could have lent support to housing markets and we chose not to, in order to impose discipline on the market.

I suspect that few readers will regard that last sentence as false.  Most will regard it as an obvious statement of wisdom.  At this point, just typing it sort of makes my blood boil.  Maybe a few of you join me on that.

Below the fold is an extended excerpt from Ben Bernanke's "The Courage to Act", about events in early 2008 (pg. 202-205):

Auctions (KE: for auction-rate securities) almost never failed.  If there weren't enough buyers, as happened on occasion, the big investment and commercial banks who sponsored the auctions usually stepped in as back-up bidders.  Except in mid-February 2008, when they refused to buy.  Many of the sponsoring institutions were wary of adding auction-rate securities to balance sheets already stuffed with hard-to-sell complex debt instruments.  On February 14, an astonishing 80 percent of the auctions failed for lack of investor interest.  Issuers with good credit records suddenly faced steep interest penalties through no fault of their own.  The Port Authority of New York and New Jersey, for instance, saw its interest rate nearly quintuple from 4.2 percent to 20 percent.

Elsewhere in the market, financial dominoes had continued to fall.  On February 11, the venerable insurance giant AIG disclosed in an SEC filing that its auditors had forced it to take a $5 billion write-down on its holdings of derivatives tied to subprime mortgages.  (Derivatives are financial instruments whose value depends on the value of some underlying asset, such as a stock or a bond.)  Three days later, the massive Swiss banking firm UBS reported an $11.3 billion loss for the fourth quarter of 2007.  It attributed $2 billion of the loss to a write-down of its exposure to Alt-A mortgages.  UBS's write-down of its Alt-A mortgage securities forced lenders with similar securities to do the same.  Given the level of investor distrust and the vagaries of generally accepted accounting principles, the valuations of the most pessimistic firms and investors seemed to be determining asset prices industrywide.

Two hedge funds with assets totaling more than $3 billion, managed by the London-based Peloton Partners and run by former Goldman Sachs traders, failed on February 28.  On March 3, Santa Fe, New Mexico-based Thornburg Mortgage, with $36 billion in assets, was missing margin calls - demands from nervous creditors for additional collateral in the form of cash or securities.  Thornburg specialized in making adjustable-rate jumbo mortgages (mortgages above the $417,000 limit on loans purchased by Fannie and Freddie) to borrowers with strong credit.  But, it also had purchased securities backed by now-plummeting Alt-A mortgages.  On March 6, an investment fund sponsored by the Carlyle Group, a private equity firm whose partners moved in Washington's inner circles, also failed to meet margin calls.  The fund's $22 billion portfolio consisted almost entirely of mortgage-backed securities issued and guaranteed by Fannie and Freddie.  The holdings were considered very safe because investors assumed Fannie and Freddie had the implicit backing of the federal government.  But the Carlyle fund had paid for its securities by borrowing more than $30 for every $1 in capital invested in the fund.  It could absorb only very small losses.  By Monday, March 10, it had unloaded nearly $6 billion in assets - yet another fire sale.

Peloton, the Carlyle fund, and Thornburg had something in common: Lenders in the repo market were reluctant to accept their assets as collateral - assets they had routinely accepted in the past.  Until the previous summer, repos had always been considered a safe and reliable form of funding - so reliable that a company like Thornburg felt comfortable using them to finance holdings of long-term assets, like mortgages.  Because longer-term interest rates are usually higher than short-term rates, this strategy was usually profitable.  It's effectively what a traditional bank does when it accepts deposits than can be withdrawn at any time and makes loans that won't be paid off for months or years.

But Thornburg wasn't a bank, and, of course, its borrowings were not government-insured.  When concerns about Thornburg's assets surfaced, nervous repo lenders began to pull back.  In what was becoming an increasingly common scenario, some repo lenders shortened the term of their loans and demanded more collateral per dollar lent.  Others wouldn't lend at all.  With no means to finance its holdings of mortgages, even its high-quality jumbo mortgages, Thornburg found itself in serious trouble - much like a bank suffering a run in the era before deposit insurance...

...I called Thornburg.  I was sympathetic.  He and his company were caught up in a panic not of their own making.  But in my heart I knew that use of our emergency authority could only be justified when it served the broad public interest.  Whether the firm was in some sense deserving or not was irrelevant.  Lending to Thornburg would overturn a six-decade practice of avoiding 13(3) loans - a practice rooted in the recognition of the moral hazard of protecting nonbank firms from the consequences of the risks they took, as well as the understanding that Congress had intended the authority to be used only in the most dire circumstances.  The failure of this firm was unlikely to have a broad economic impact, and so we believed a 13(3) loan was not justified.  We would not lend to Thornburg, and it would fail.
Congress had added Section 13(3) to the Federal Reserve Act in 1932, motivated by the evaporation of credit that followed the collapse of thousands of banks in the early 1930s.  Section 13(3) gave the Federal Reserve the ability to lend to essentially any private borrower.  At least five members of the Board needed to certify that unusual and exigent circumstances prevailed in credit markets.  The lending Reserve Bank also had to obtain evidence that other sources of credit were not available to the borrower.  And importantly, 13(3) loans, as with standard discount window loans, must be "secured to the satisfaction" of the lending Reserve Bank.  In other words, the borrower's collateral had to be sound enough that the Federal Reserve could reasonably expect full repayment.  This last requirement protected taxpayers, as any losses on 13(3) loans would reduce the profits the Fed paid each year to the Treasury and thus add to the budget deficit.  But, the requirement also limited the interventions available to the Fed.  Invoking 13(3) would not allow us to put capital into a financial institution (by purchasing its stock, for example) or to guarantee its assets against loss.

The Fed used its 13(3) authority during the Depression, but only sparingly.  From 1932 to 1936, it made 123 such loans, mostly very small.  The largest, $300,000, was made to a typewriter manufacturer; another, for $250,000, was extended to a vegetable grower.  As the economy and credit markets improved in the latter part of the 1930s, the Fed stopped making 13(3) loans.
Within the month, 13(3) would be invoked in order to extend credit on a broader range of securities to primary dealers that traded with the Federal Reserve, and would also be invoked in the arrangement to facilitate the purchase of Bear Stearns.

The mental disconnects here are large.  What is described here is clearly a liquidity crisis.  The problem these firms were facing was that they existed in an economy that was short of cash.  Whether that shortage can be attributed to a sharp rise in demand or a relative decline in supply is not particularly meaningful.  The economy needed cash.

Considering this, what was holding the Fed back?  Maybe the sharp drop in the Fed Funds Rate from 5.25% to 2.25% in just 7 months seemed like accommodation.  But, the Fed was just chasing the collapsing neutral rate down.  They weren't buying treasuries.  In fact, they had already been making emergency loans and were selling treasuries in order to prevent cash from remaining in the economy.

It's strange that the "Too Big To Fail" problem is a core part of criticisms of the Fed, yet here Bernanke essentially states "Too Big To Fail" as a sort of principle.  It's not so much a matter of saving the large firms because they have some power over the Fed or the economy.  It's that, on principle, the Fed refuses to help firms that aren't "TBTF".  But why not?  In the passage above, written by Bernanke himself, he explicitly points out that the original 13(3) authority was used exclusively to support systemically insignificant firms.

In the end, 13(3) authority is somewhat useless.  The reason that the Fed wouldn't invoke 13(3) earlier in the crisis is because the only reason there was a crisis was because they were creating a crisis, and they didn't know it.  There is not a conceivable context where the Fed would utilize 13(3) in a productive way, because in order to do that, they would have to understand that the firms they would be supporting are suffering from a systemic liquidity crisis.  If they understood that, they would invoke more conventional forms of support, and 13(3) would be unnecessary.

You might argue that firms like Carlyle were overleveraged, so it was their recklessness that got them in trouble.  But, in every conceivable liquidity crisis, it will be the most leveraged and most financially vulnerable firms that will need support.  If this is your measure for when nominal stability can be an acceptable policy, then you will inevitably support purposeful instability.  That is actually what we have been doing since 2006.


  1. On the flip side, if the Fed were to react strongly to every liquidity crisis won't you constantly be bailing out those using leverage in the most dangerous areas of the economy? If the Fed had been super-aggressive in the 2001 recession it would have had to bail out leveraged owners of tech stocks, no?

    I personally think "liquidity crisis" have a purpose - transfer capital to those who understand the economy better and invest accordingly.

    1. Tech stocks didn't fall because there was a liquidity crisis. The excerpt from Bernanke in the post is a description of a liquidity panic. The decline in tech stocks wasn't remotely similar. Equities in tech firms weren't near-cash securities, and nobody ever expected them to be treated that way.

    2. "I personally think "liquidity crisis" have a purpose - transfer capital to those who understand the economy better and invest accordingly."

      You are the problem. There are plenty of idiosyncratic ways to take on risk and to fail. Creating new systemic ones, or cheering them on, is dumb.

    3. So you were in favor of bailing out Long Term Capital Management, which was 100% a liquidity crisis (after all, the theoretical values of all their spread trades could be proven to be positive in a normal state of the world)?

    4. There are reasonable arguments for and against intervening in the LTCM case, although in either case, the use of the term "bail out" in these cases where equity values are basically wiped out is part of the mania. It would be like referring to bankruptcy as a "bailout".

      But there is no reasonable argument against providing nominal support where across asset classes and among thousands of firms and banks, there are obvious dislocations in securities that are generally treated as near-cash.

    5. To me there is a direct correlation between Greenspan's strategy of reducing economic volatility and the increased leverage in the economy. My theory is that private actors target a level of wealth volatility. If you dampen the volatility of the economy and assets they will add leverage to retail total volatility at the desired level.

      So you cannot remove volatility simply shift it. My view is that supporting those who are most leveraged you are implicitly reinforcing the behaviors who ask "how can i best game the system" before asking "what investment makes the most long-term sense." Do that for a few decades and my hunch is you end up with very low productivity growth..oh, wait, we have that.

  2. Yes, many useful people and enterprises are leveraged, like homeowners, land developers, some business start-ups (in a sense), and some financial outfits.

    The Fed can always pull the rug out from such people, and then someone can always cry, "Yes the weak and over-leveraged are getting deservedly whacked!|

    I do wonder about a Carlyle, or earlier a Long-Term Capital Management, and their ability to leverage 30 or even 100 to one, and then have macroeconomic effects when they fail. This may be a rare case where a very simple rule is needed. Perhaps mandated but privately provided industry insurance. Like private mortgage insurance, or something to that effect.

    It does us no good if a LTCM tanks and takes down a few nations and millions of people with it.

    Side note: I recently blogged about the trade deficit and property zoning. Unfortunately,. Marcus Nunes put my commentary behind a paywall.

    As Scott Sumner posits, the US trades tee-shorts for homes. But not mobile homes, rather homes on US soil. So the trade deficit is not a deficit, says Sumner.

    Maybe so, but given the reality of property zoning, we then see wealthy foreigners eclipsing the middle class in various housing markets. We now see Australia and Canada erecting barriers against foreign ownership of housing that would make Don Trump blush.

    This is a bit if a replay of the immigration scene. If the supply of housing stock is limited, then calling for open borders has less appeal for renters and others.

    Unless you are a property owner already.

    So the middle class sees competition for limited rental housing from immigrants, and competition for limited housing for purchase from foreigners.

    But if the middle-class complains, they are labelled Neanderthals etc.

    Yes, I have simplified, but….

    1. On the leverage bit, it seems that removing the tax benefits of debt funding would be the easiest way to try to bring down leverage. I think the best way to do that would be to eliminate corporate taxation, which would have the added benefit of making many of the tax advantages of homeownership go away, so it seems to me that would also be a progressive change in the tax code.
      That's probably not ever going to happen, but I'm not sure that it does any good to do some complicated second best policy when we can't manage to do the simple policy. In the meantime, the complaints about providing nominal support for firms that become destabilized during liquidity events seem more motivated by aesthetics than by reason. Our fear of moral hazard from 2006 to the present has created much more damage than moral hazard itself ever did. That sounds crazy, though, to anyone who thinks the housing bubble was an irrational demand event.

  3. So in every economic cycle I should find the most liquidity-sensitive asset and buy it on leverage..heads I win, tails i get bailed out. Then as i become fabulously wealthy, I should seed others to do exactly the same thing. Sounds a lot like what banks have been doing for the last 20 years.

    1. There are so many ways to fix that problem other than purposefully introducing nominal volatility. You're basically applying the broken window fallacy to risk. The defining characteristic of developed economies is the minimization of systemic risk. You're essentially demanding that we act more like a third world country so that unleveraged at-risk investments will command higher incomes in exchange for political uncertainty.

    2. In a perfect world I would agree. In a world full of humans I do not. Capital will always want to chase returns. The asset class being chased will often result in something resembling a bubble. That asset class will often become systemically important and thus always threaten a "liquidity crisis." Reinforcing such behavior seems like a bad idea.

    3. Risk premiums are predictably negatively correlated with risk free interest rates. Times, such as the late 1960s and late 1990s, where there had been long periods of stability and risk premiums were low, were associated with high real risk free interest rates.

      Real risk free rates were very high when the internet bubble happened. You are empirically wrong about this. The price of near-cash securities goes down when expectations of stability are high. This is such a basic and universal part of how capital market participants behave, I can't imagine how so many people could be so wrong.
      It is true that risky assets might have higher prices in a stabilized regime, but those aren't the assets the central bank is trying to stabilize. 2001 and 2007 are very different episodes, the latter episode was clearly more damaging, and risk premiums were not particularly low then. Investors were piling into AAA securities precisely because they were risk averse. Using the fact that financial intermediaries were falling over each other trying to meet demand for low risk securities as a symbol of risk taking is a terrible error.

  4. The Fed did not want to save subprime and allowed the CP to disintegrate. That is irrefutable. However, the Fed allowed lending that did not reflect risk. Anyone could get a loan. Still, there was sound subprime lending in many states.

    So, subprime was seen as something expendable. However, the Fed did not count on all the loans being placed back onto the balance sheets of the banks, causing HELOCs, which drove much of the economy, to shrink. They thought subprime was contained, but it shrunk the balance sheets of the banks. Some would say that was, especially with mark to market, a solvency issue. But either way, liquidity or solvency at stake, and the Fed squeezed the banks and created a major crisis. They got wages to come down, and houses back for the elite at bargain rates.

  5. Back in 2007 and 2008, I hadn't heard of the NGDPLT concept or even NGDP growth rate targeting. My thoughts in late 07 and early 08 were that the Fed probably couldn't avoid a recession while controlling inflation. With hindsight, I still don't know. But at least if they had just been targeting 5% NGDP growth, they would have had a shot at keeping things under control (no matter what combination of RGDP and inflation we ended up with). So with that background, I'd still prefer the Fed to resort to 13(3) lending only if other means of implementing monetary policy were not working to keep NGDP growth on track. If NGDP growth is on target, then failures are fine. There will always be, from time to time, institutions, big and small, that overreach and at times some will go BK. And that's OK with me as long as the Fed isn't the problem because the Fed let NGDP growth fall below 0%, or even much below its targeted growth path. The problem with Bernanke's approach in 2008 was that the Fed sterilized every single dollar it lent out to specific institutions, so (a) it wasn't doing anything to push up NGDP growth and (b) even worse, it was actually making the liquidity crisis even worse for the remaining institutions each time the Fed bailed out one of them. With respect to LTCM, it wasn't really a bailout. What LTCM got was what everyone should be able to get - an expedited chapter 11 if all the lenders are on board. It's a problem that you have to be TBTF to get one, but the solution is to make it easier for everyone else to get one too.

    1. Great comment. Part of the problem is the sloppy use of the word "bail out".