Tucker and Bagehot at Hoover

I had the pleasure last week of attending the conference on Central Bank Governance And Oversight Reform at Hoover, organized by John Taylor.

Avoiding the usual academic question of what should the Fed do, and the endless media question will-she-or-won't she raise rates, this conference focused on how central banks should make decisions. Particularly in the context of legislation to constrain the Fed coming from Congress, with financial dirigisme and "macro-prudential" policy an increasing temptation, I found these moments of reflection quite useful.

Some of the issues: Should the Fed follow an "instrument rule," like the Taylor rule? Should it have "goal," like an inflation target, but then wide latitude to do what it takes to attain that goal? What structures should implement such a rule? Implicit in a rule that the Fed should do things, like target inflation and employment, is an implicit rule that it should ignore others, like asset prices, exchange rates and so on. (I think this is much too often overlooked. As financial reform should start by delineating what is not systemic, and hence exempt from regulation, monetary policy rules should start by saying what the Fed should ignore.) Should that limitation be more explicit? What's the right governance structure? Should we keep the regional Feds? How should Fed meetings be conducted? Is "transparency" the enemy of productive debate? How much discretion can an agency have while remaining independent?  And so on.

I was going to post thoughts on he whole conference, but John Taylor just posted an excellent summary, so I'll just point you there.

My job was to discuss Paul Tucker's (ex Deputy Governor of the Bank of England) thoughtful paper, "How Can Central Banks Deliver Credible Commitment and be “Emergency Institutions" Paul's paper starts to think deeply about independent regulatory agencies in general, and monetary and fiscal policy together. My discussion is narrower. I'll pass on the discussion (pdf here) as today's blog post, as it might be interesting to blog readers.

Comments on “How Can Central Banks Deliver Credible Commitment and be “Emergency Institutions” By Paul Tucker

May 21 2015

Let me start by summarizing, and cheering, Paul’s important points.

The standard view says that perhaps monetary policy should follow a rule, but financial-crisis firefighting needs discretion; a big mop to clean up big messes; flexibility to “do what it takes”; “emergency” powers to fight emergencies.

I think Paul is telling us, politely, that this is rubbish. Crisis-response and lender-of-last-resort actions need rules, or “regimes,” even more than monetary policy actions need rules. At a basic level any decision is a mapping from states of the world to actions. “Discretion” just means not talking about it.

More deeply, you need rules to constrain this mapping, to pre-commit yourself ex-ante against actions that you will choose ex-post, and regret. Monetary policy rules guard against “just this once” inflations. Lender of last resort rules guard against “just this once” bailouts and loans.

But you need rules even more, when the system responds to its expectations of your actions. And preventing crises is all about controlling this moral hazard.

To stop runs, our governments guarantee deposits and other loans; they bail out institutions and their creditors; they buy up assets to raise prices, and they lend like crazy. But knowing this, financial institutions take more risk than they would otherwise take, and investors lend without monitoring, making crises worse. Institutions that can borrow at last resort don’t set up backup lines of credit, don’t watch the quality of their collateral, and don’t buy expensive put options and other insurance, making crises worse. Investors who know that the Fed will stop “fire sales,” don’t keep some cash around for “buying opportunities,” making fire sales worse. “Big banks are too complex to go through bankruptcy,” the mantra repeats. But why do people lend to them, without the protections of bankruptcy? Because they know creditors, if not management and equity, will be protected.

“The world is ending. A crisis is no time to worry about moral hazard,” bankers and government officials told us last time, and will tell us again. But the world does not end, and actions taken in this crisis are exactly the cause of moral hazard for the next one.

This isn’t theory. When the Fed and Treasury bailed out Bear Stearns, and especially its creditors, markets learned “Oh, Fed and Treasury won’t let an investment bank broker-dealer go under.” Lehman turned down capital offers, and Reserve Fund losses on Lehman paper were enough to cause it to fail in a run. (This is an update: see below.)

The severe crisis and recession coincident with Lehman’s failure, together with the massive and improvised response — many flavors of Tarp, auto company bailouts, and so on — have arguably created the “rule” in participants’ minds about what will happen next time.

Plans, self-imposed rules, promises, guidance, and tradition are not enough. Given the power, every one of us will bail out. We won’t risk being the captain of the Titanic, and we’ll let the next guy or gal deal with moral hazard. A central banker facing a crisis is like a father holding an ice cream cone, facing a hungry three-year old. Sure, Mom’s rule says dinner always before dessert. We know what’s happening to that ice cream cone.

The central bank and Treasury must not be able to bail out what they should not bail out, to lend where they should not lend, to protect creditors who should lose money. That’s the only way to stop it. More importantly, it’s the only way to persuade the moral hazarders that all the fine words in the boom will not melt quickly in the emergency.

Two central quotes summarize the Tucker view, and I entirely agree.

Prerequisites for any such regime are that its terms should mitigate the inherent problems of adverse selection and moral hazard; be time-consistent; and provide clarity about the amount and nature of ‘fiscal risk’ that the central bank is permitted to take on the state’s behalf.

At a schematic level, a money-credit constitution for today might have five components: inflation targeting plus a reserves requirement that increased with a bank’s leverage plus a liquidity-reinsurance regime plus a resolution regime for bankrupt banks plus constraints on how the central bank is free to pursue its mandate.

***

Now, let me offer a gentle critique.

How are we doing towards the Tucker regime? Not well.

The Dodd-Frank and Basel “regime” has no serious limits at all. Ask yourself, what institutions are not “systemic” and cannot become so designated? What institutions or creditors won’t be bailed out; can’t be bailed out? What are the securities the Fed or Treasury won’t and can’t  buy or lend against? What are the asset prices prices that they won’t and can’t prop up?

Paul points out the difficulties. Yes, “constraints” are good. But just what constraints? We can channel Bagehot, “against good collateral,” to “illiquid but not insolvent” institutions. Except, as Paul reminds us, what’s good collateral, when noone will take anything but Treasuries? How do you tell illiquid from insolvent when prices have tanked and markets are frozen? It’s not so easy.

More deeply, the Bagehot rules are flawed. If it were clear who is illiquid and who is insolvent, there wouldn’t be a crisis. Private lenders would happily support the clearly solvent. And runs happen at institutions that investors fear are insolvent. If you want to stop runs you have to prop up at least the creditors of potentially insolvent institutions. Bagehot’s rules may constrain the central bank; they may be good rules for a prudent investor, they may address moral hazard. But they are not obviously optimal rules to stop crisis or to prevent them from occurring in the first place.

Worse, when we figure all this out, how do we write binding laws or regulations that will effectively constrain bailout-hungry officials?  For example, Paul Volcker proposed a fine clear rule, “thou shalt not finance proprietary trading with deposits.” Which, 600 pages and counting later, is utter mush.

So here we are, 6 years after our crisis — or 82 years after 1932, or 113 years after 1907, or, heck, 300 years after 1720— and as eminent a thinker and practitioner as Paul still needs to invite future thought on what these rules ought to be, let alone just what legal restrictions will actually enforce them and communicate that expectation.

I fear that the next crisis will be upon us long before Paul has figured it out, and a century before he gets the Basel committee, the Fed, ECB, FSOC, Congress, Parliament, SEC, and so on to go along.

***

So, I agree with pretty much all Paul has to say. but I infer the opposite message. If this is what it takes to rescue the house of cards, then we need a different house, one not made of cards. We need to stop crises from happening in the first place.

To its credit, that is the other half of our contemporary policy response: This time, finally, the army of regulators and stress testers will see the crisis coming; with their Talmudic rules and interpretations, and their great discretion, they will stop any “systemically important” financial institution from losing money, despite the moral hazard sirens, and without turning that financial system into something resembling the Italian state telephone company circa 1965.  Good luck with that.

Consider an alternative: Suppose banks had to fund risky lending by issuing equity and long-term debt. Suppose mortgage-backed securities were funded by long-only, floating NAV mutual funds, not overnight repo. Suppose all fixed-value demandable assets had to be backed 100% by our abundant supply of short-term Treasuries. Then we really would not have runs in the first place. And a lot of unemployed regulators.

Why do we not have such a world? Originally, because you can’t do it with the financial, computational, and communications technology of the 1930s or 1960s.  But now we can. More recently, I think, because moral hazard so subsidizes the current fragile system. But now we can change that.

Paul mentioned this possibility, but gave up quickly, conditioning his remarks on a view that society has decided it wants fractional reserve banking. Well, maybe society needs to rethink that decision.

Really, just why is it so vital to save a financial system soaked in run-prone overnight debt? Even if borrowers might have to pay 50 basis points more (which I doubt), is that worth a continual series of crises, 10% or more downsteps in GDP, 10 million losing their jobs in the US alone, a  40% rise in debt to GDP, and the strangling cost of our financial regulations?

***

A last point. Paul unites financial with monetary and fiscal policy. That’s crucial. The last crisis raised US national debt from 60% to over 100% of GDP. The next one will require more. At some point we can’t borrow that much.

But take this thought one step further. The next crisis could well be a sovereign debt crisis, not a repetition of a real estate-induced run. Crises are by definition somewhat unexpected, and come from unexpected sources.

To be concrete, suppose Chinese financial markets blow up, surprise, surprise discovering a lot of insolvent debt. The stress is too much for the IMF and Europe, so Greece goes, followed by Italy Spain and Portugal, half of Latin America and a few American States. Pair that with war in the middle east — Isis explodes a dirty bomb, say — requiring several trillion dollars.

Now Governments are the ones in trouble. They won’t be able to borrow trillions more, bail out banks or lend of last resort.   In a global sovereign debt crisis, even Paul’s regime would turn out to be a superb Maginot line. The current regime wouldn’t be that strong.

A financial system deeply dependent on the government put would be finished.  This is the lesson of Europe. A southern government default would have little consequences if its banks were not so embroiled in government finances.

But a financial system uncoupled from government finances would survive.

***

In sum, I cheer pretty much everything Paul said. But It’s an outline for a plan that will take decades to fill in. And all in the service of keeping the house of overnight debt cards going.

So the lesson I take is that instead, we should finally take seriously the other centuries - old, simple alternative: equity-funded banking, government-provided interest-paying money, mirroring that great 19th century innovation, government-provided banknotes, and a purge of run-prone assets.

 ***

PS:

• Thomas Humphrey writes an interesting  history of Bagehot's rules in the Richmond Fed Review, Averting Financial Crises: Advice from Classical Economists. 
• Renee Haltom has an excellent short article in the same issue, Last-Resort Lending for the 21st Century summarizing current views.
• A spate of news articles came out last summer suggesting Lehman might have been "solvent" after all, here, here, here. Of course "solvent" at ex-post prices selects on one state of the world. Same comment for how much money the government and Fed made on bailout deals. 
• One interesting point came up at the conference (I forget who said this). If the central bank lends against "good collateral," that takes away important assets that rightfully belong to debt-holders, and makes them more likely to run.   

** Update: I originally wrote incorrectly that the Reserve Fund had 40% of its assets in Lehman. A correspondent corrected me and pointed me to McCabe, Holscher, Cipriani, and Martin's BPEA paper whose footnote 27 states

The Primary Fund’s losses were caused largely by its $785 million in holdings of Lehman debt obligations (1.3 percent of the fund’s assets) at the time of Lehman’s bankruptcy. RCMI, the adviser to the fund, announced at about 4 pm on Tuesday, September 16, 2008, that the NAV of the fund’s shares had dropped by 3 percent, to 97 cents, presumably because large redemptions had further eroded the NAV. 

The correspondent adds that money funds can’t have that much  exposure to one counter-party because of limits in rule 2a-7.  1.3 / 3 is about 40%, which must be the number Im remembering -- 40% of the losses, not the assets, came from Lehman. This is even more interesting, because it suggests a run on the fund, rather than large actual losses, was the central problem.  Moral, check your numbers, even ones you think you remember really well.