Wednesday, October 2, 2013

Financial Reform in 12 Minutes

I was on a panel yesterday at a fascinating conference, "The US Financial System–Five Years after the Crisis."  The conference was run simultaneously at Hooover and Brookings, with a live video link. Which went dead exactly as I was proclaiming the wonders of modern technology, but otherwise worked remarkably well. Alas, the conference was run under "Chatham house rules" so I can't tell you all the remarkable things that the Very Important People said. I can pass on surprise that there was so little disagreement between the Brookings and Hoover sides. I was expecting a spirited defense of Dodd-Frank from the East. Instead, they piled on it, if anything more eloquently that the West, with only one very thoughtful but still lukewarm defense.  I hope the presentations will eventually be made public, as they were uniformly interesting -- even the ones I disagreed with I thought were wrong in very interesting and thoughtful ways.

I was given 12 minutes to comment on the state of financial reform, is too big to fail over, are we ready for the next crisis. My answer follows. Yes, faithful readers will recognize something of a moving average, which will continue both to move and to average.

Have fun

Financial Reform in 12 Minutes

Is too big to fail over? No. Are we ready for the next crisis? Absolutely not. 

The Dodd-Frank act is “more cowbell” – a dramatic expansion of the same regulatory structure that failed before. Unless fixed, it will fail again in even more spectacular fashion.

The basic structure of Dodd-Frank is a huge expansion of regulation – largely “discretionary, judgmental, and micro-managing” as nicely put by an earlier panelist – to try to prevent any large and “systemically important” institution from losing money again, and a “resolution authority” in place of bankruptcy court should it fail anyway.

The premise of “resolution authority” is that large financial institutions are too complex to go through bankruptcy.  So instead, the mess will be dumped into the lap of appointed officials who will figure out over a weekend who gets how many billions of dollars. If it is so complex that bankruptcy can’t fixed, to write down ahead of time who gets what, how in the world are these poor folks going to figure it out on the spot?  This idea is a triumph of discretion over rules.

They won’t. Politically powerful creditors will scream that they too are too “systemic” to lose money, much as Goldman Sachs threatened if it should not get its collateral from AIG.  “Resolution” only matters with a crisis in the background as in fall 2008, amid general panic. Of course these large and well-connected creditors will be bailed out.

Seeing this in advance, and all the rules up for grabs, other creditors will run at the first hint that resolution might be coming. Seeing that run develop, the FSOC will be forced to bail out ahead of time.

The pretense that regulators can and will spot trouble brewing and stop banks from taking risks, as they abundantly failed to do in 2008, and again when faced with European sovereign defaults, is a triumph of hope over experience.

The “macroprudential” idea that the Fed can spot “bubbles” forming, and can and will stabilize asset prices by artfully controlling interest rates, intervening in many markets, and controlling the details of financial flows, so that nobody loses any money in the first place, is a triumph of pipe dreaming. (An earlier panelist eloquently called it “profoundly misguided.”)

Like all previous crises, the next crisis will not conveniently repeat the last one, with a real estate boom and bust provoking a run in shadow banking. It will erupt from an unexpected quarter. What out there today looks today like subprime mortgages did in 2004? Sovereign debt is a good possibility.  Europeans have learned that buying the bonds of high-yield “do what it takes” countries makes money, as Americans learned for subprime mortgages. A realization that sclerotic growth is settling in as the new normal, upending long-term budget projections, could be the trigger.  Detroit could be LTCM, California, Illinois, and Greece could be Bear Stearns, and Italy could be Lehman Brothers.

We used lots of sovereign debt to bail out of the last crisis and stimulate in its wake. Issuing more sovereign debt remains our and Europe’s fire extinguisher. We are utterly unprepared for a crisis of sovereign debt itself.

There is an alternative.

The crisis was a run. The tech stock bust did not cause a crisis, because tech stocks were stocks. When stock prices fall, it’s too late to run. The housing bust led to a crisis because houses were funded, in the end, by overnight debt, which ran.

Institutions are not systemically dangerous. Run-prone assets are dangerous.

So, why not just ban run-prone assets? We could require that all run-prone fixed-value liabilities, including deposits, overnight debt, and money-market shares, must be backed 100% by short-term treasuries; ideally in separate or at least ring-fenced institutions.  Mortgage-backed securities can be held, without government guarantee, via long-only, floating-value mutual funds in your and my 401(k) accounts, by pension funds and by endowments. Banks, and everyone else, must then finance risky investments primarily by equity, with perhaps some long-term debt. Equity will no longer be just a “cushion” but a main source of funds.

Why not? The bank answer is “the Modigliani-Miller theorem fails for banks, so borrowing will be more expensive.” The MM theorem does indeed fail – because the government subsidizes and guarantees debt! Sure, banks want to maximize the value of those subsidies, and greater equity dilutes them.  In addition, even if equity-financed banks charged 20 basis points more for loans, in equilibrium, remember that we lost almost 10% of GDP and 10 million jobs 5 years ago, and they have not yet returned. That’s a big price to pay.  If we want to subsidize borrowing, we can do it transparently, on budget, rather than by subsidizing or even tolerating run-prone debt.

The other answer is, people need a large supply of fixed-value “liquid” assets to make transactions. It is said that we need banks to “transform” liquidity and maturity, even if imperfectly.

That may have been true in 1938. In 2013 financial, transactions, and communication technology have  changed everything. Instant communications – technical change – and index funds – financial engineering -- mean you no longer need fixed-value, first-come-first-serve, run-prone assets to have perfect liquidity. You could bump your iphone and pay for coffee by selling an S&P index fund share, and the store buys a share in a floating NAV mortgage-backed security fund, all in milliseconds. In fact, most current transactions are simply netted by banks, with nothing exchanged. In the 1930s we could not look up the value of the stock index to make the transaction; we needed to offer a fixed-value claim. Now we can.  2013 index funds and ETFs carry none of the asymmetric-information illiquidity that 1938 individual stocks suffer.  And, as a minor benefit of our fiscal profligacy, $18 trillion of Treasuries can back every imaginable genuine economic need for fixed-value run-prone debt.

How do we get there? Much – much – higher capital requirements are a good first step. But they choke on two practical problems: First, what’s the denominator? Risk weights can be gamed, and prescribing a ratio of capital to total assets incents banks to find clever ways to take on more risk at the same asset value. It’s not hard to buy beta. Second, what’s the minimum? 20%? 50%? 100%? The right answer is “the more the better,” and “so big that it doesn’t matter,” but that’s hardly satisfying.

I think a simple tax is the answer – though since “tax” is a dirty word, let’s call it a “systemic externality fee” – on debt, and especially on short-term debt or any other contract where the investor has the right to demand payment, and fail the firm if not received.  Every dollar of such funding will cost, say, a 10 cent fee. Payments due later generate smaller fees. I think we’ll see a lot less run-prone debt, fast. (We could at least stop subsidizing debt!)

Then, we won’t have to argue about risk weights and precise capital ratios, we won’t have to intensively regulate bank assets, we won’t tempt regulatory arbitrage, we won’t ask the Fed to decide whether houses in Palo Alto are a “bubble,” we will not hear the periodic call “we must recapitalize the banks” (at taxpayer expense), and, most of all, we can escape the chokehold on competition and innovation posed by our current expanding regulatory mess, together with the capture, cronyism, and politicization to which it is swiftly leading.

We need a financial system that can absorb booms and busts without creating a run or a crisis, rather than dreaming that regulators can produce a world without booms and busts. We need to regulate financial institutions’ liabilities, not micro-manage their assets, and especially not try to manage the price of every asset in which they might invest. We must escape this crazy system in which our government subsidizes debt, guarantees debt, increases the demand for debt by regulating it as a safe asset, and then tries to regulate financial firms away from issuing that debt.

We need to insulate the financial system from looming government financial trouble rather than more deeply intertwine them.

Update: John Taylor posts the conference program; Russ Roberts asked a great question 


  1. I like these suggestions, but I had to respond. (as a quant trader, I may be biased).

    Many of the problems of 2008 go away if you could not lever assets that were not being traded and had a sufficient level of liquidity. So much of the issue involved the various structuring desks taking higher risk tranches of deals that no one would buy and marking them at a profit. The quants on the desk get paid, the bank loses true book value and increases its leverage on every deal. Its fine if the bank structurers think they can make money on a deal, but they should not get to mark it as worth more than the buy side thinks its worth.

    Also, people rail against VaR. VaR models are good enough as first order approximations on liquid assets. (No one would be trading the E-mini levered 30 to 1.)

  2. "How do we get there? Much – much – higher capital requirements are a good first step. But they choke on two practical problems: First, what’s the denominator?"

    The call is for much higher equity capital requirements. Maintaining a stable debt to equity ratio has problems with the numerator. You can either used when issued pricing for the debt of a bank or the market price of the debt. The denominator (equity valuation) is always market priced.

    "Second, what’s the minimum? 20%? 50%? 100%? The right answer is the more the better, and so big that it doesn’t matter, but that’s hardly satisfying."

    The "right" answer is that the debt to equity ratio that banks maintain should be correlated with the yield curve. A steeper yield curve should allow for a higher debt to equity ratio, a shallower yield curve should require a lower debt to equity ratio.

    That is the problem with "one size fits all" capital requirements. They need to be flexible within the economic environment.

  3. "Flexible within the economic environment" is great for the omniscient benevolent planner and academic who likes to derive optimal policies. Alas, it's an opening to shoot from the hip macro prudential dirigisme in practice. Look how much trouble our Fed is having deciding whether the "environment' is good enough for a little tapering -- to say nothing of proclaiming that Citi had better go raise some capital. One size fits all may not fit all the time, but you have the advantage of not needing to haggle.

    1. John,

      The way you get rid of haggling is through preset rules based upon the prevailing economic conditions. The example I gave was the status of the yield curve. Imagine something like this:

      If yield spread (30 year to 1 year) >= 4%, then debt to equity must be no higher than 10 to 1

      If yield spread (30 year to 1 year) < 4%, >= 3%, then debt to equity must be no higher than 8 to 1

      If yield spread (30 year to 1 year) < 3%, >= 2%, then debt to equity must be no higher than 6 to 1

      If yield spread (30 year to 1 year) < 2%, >= 1%, then debt to equity must be no higher than 4 to 1

      If yield spread (30 year to 1 year) < 1%, then debt to equity must be no higher than 2 to 1

      Debt to equity ratios must be adjusted based upon prevailing economic condition (yield spread). I am not saying that yield spread would be the ideal economic indicator, but it does provide flexibility without haggling.

  4. Interesting stuff (though what happens if Treasuries start carrying default risk?).

    One question: Is there solid empirical work on the question of whether equity price crashes affect the macroeconomy less than waves of debt defaults? It certainly seemed like the tech crash was much less harmful to the macroeconomy than the housing crash (or Japan's commercial property crash).

    1. On default. Good point, which I left out for length. Greece obviously cant do it. For now, let's just say the US will likely inflate rather than default, so short term treasuries are at least immune from bankruptcy. We're about to find out if that assumption is true. There is a whole additional essay coming one of these days on how to insulate money from default.

    2. Hmm...I used to think that, but now I realize that the Fed doesn't coordinate with Congress, and that Congress (or some faction of Congress with veto power) may believe for some reason that default is desirable. But anyway that's a bit off topic...

    3. If Treasuries start to carry default risk then perhaps the GDP-linked securities advocated by Robert Shiller are as well or better-suited for holding in bank accounts. If the masses opt for such holdings, it would also reduce the problems associated with sovereign defaults. I don't see why Prof. Cochrane incorporates Shiller's ideas of national self-insurance into his own proposals to improve financial stability.

    4. Also regarding default: I believe Prof. Cochrane's post from May 4 is helpful - he proposes floating-rate perpetual treasury securities as part of a run-free financial system.

  5. What if the investment we need will not be made by businesses, regardless of how financed (debt or equity), as Christensen argues?

  6. Over at Cafe Hayek, Russell Roberts has some comments on the same conference.

    Read it. I found his thoughts kind of interesting.

  7. This comment has been removed by a blog administrator.

  8. My only issue here is that sufficiently short-term debt is effectively run-able, even if it's not technically withdrawable. For example if you buy overnight commercial paper and just keep rolling it over into new CP every day, it's effectively like having money in the bank. If you lose faith in the company, you can just not roll over that CP, which is analogous to pulling your money out of the bank. Of course the most famous example of this happening was when the Reserve Primary Fund "broke the buck" and there was effectively a run on it.

    So run-ability is not really binary, more of a spectrum, where 30-year bonds are obviously not run-able, and bank deposits clearly are. But then banks do serve the useful functions of maturity and liquidity transformation.

  9. If the U.S. government funded itself with overnight borrowing, then the short end of the yield curve would flatten out, and there wouldn't be any profit in maturity transformation. Is that basically what you're suggesting?

  10. Dr. Cochrane,

    I think we're agreed that there's a limit to how much the shareholders, and creditors, of financial institutions can or should be shielded from their own bad calls. With all due respect, trying to abolish financial crises by abolishing fractional reserve banking (what is what your prohibitive tax on bank deposits would amount to) is frankly about as foolish as trying to abolish bad harvests by abolishing industrial farming and making everyone grow his own food.

    I don't even believe it's even possible in a globalized financial market. A prohibitive tax on demand deposits in any single country would simply drive bigger and more sophisticated depositors offshore. Smaller and less sophisticated depositors would either have to learn to live with not knowing how much was in their S&P-500 linked chequing account minute by minute or start demanding their employers pay them in cash and keeping the rent money in a safe.

    If there's a straightforward proof that such a financial system would really yield welfare improvements over the one we have now, I'd very much like to see it.

    One reason I'm skeptical is that banks do offer run-free deposits. They're called safe deposit boxes. Most people don't use them to store money for transactions purposes. They use theoretically "run-prone" demand deposits because they trust the bank to be reasonably well managed enough that they can call their deposit at any time and the bank will be able to pay out---that is, they consider the risk of a run in practice small enough ignore. At any rate, that risk is a small price to pay for knowing how much money they actually have for transactions minute by minute. Households value not having to worry about getting a nasty surprise at the supermarket checkout counter, and firms value not having to worry that the payment for a big shipment might be refused because the S&P 500 did badly that day.

    (Households also value not having to go to a seedy check-cashing outfit and paying fees greater than their expected loss in the event of a run by several orders of magnitude, and not having large amounts of cash in their house.)

    To put it another way, people use banks as a way to insulate themselves from risk, believing banks are better at managing risk than they are, and are willing to pay banks for the privilege (in the form of accepting a lower return on deposits than they would receive on equity). Occasionally banks will make bad calls. ("Risk-free" government bonds need not be all that risk-free, for example!) Surely they will do so no more often than individuals with less information about financial markets would if left to their own devices.

    Yes, there's always a risk a bank won't be able to pay all its depositors. At well-managed banks, it's usually small enough for people to ignore, or at least not big enough to justify any bot the most risk-averse person keeping money in a safe.

    So to explain people's insistence on keeping money in demand deposits at all and still argue that abolishing demand deposits is welfare-improving, your model would presumably have to assume that people consistently underestimate the possibility of a run. Relaxing the rational expectations assumption takes one into very deep water very fast. It's hard to imagine that such---ahem---irrational people would only be irrational along this dimension, or that the welfare improvements to be gained from financial paternalism would end there.

  11. Have the Kovacevich (WellsFargo) observations been born out? What would have happened if we'd sold-off CITI to the highest bidder, let the mortgage holders take their lumps, and provided liquidity to only sound institutions? A few more attempts to fly-without-wings on wall-street perhaps, but both an object and moral lesson to those who believe models are real. And can't bring themselves to admit that the future is largely unknowable. Save when arbitraging a government guarantee.


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.