Monday, June 12, 2017

Living Trusts for Banking

One of the core problems of financial reform is how to "resolve," AKA bankrupt, a big bank -- how can equity holders be wiped out, and debt holders carve up the remaining assets. Big banks are supposed to craft “living wills,” really living vivisection guides, but that effort is clearly in trouble. This blog post expands on a different idea for bank resolution; let’s call it “living trusts” by a similar analogy to estates.

Here's the idea: Let a bank fund its risky investments 100% by issuing equity. The bank then simply cannot fail — it cannot go bankrupt, it cannot suffer a run.  As I've argued elsewhere, I think this is entirely practical.

But suppose it really is important for some reason to carve up bank liabilities into a small amount of highly leveraged equity and a large amount of run-prone short-term debt. Suppose it really is important for banks to "create money," and to take deposits, and to funnel those into risky, illiquid, and otherwise hard-to-resolve assets. Suppose that equity holders really demand highly leveraged high return high risk bank equity, not super-safe low return low risk bank equity, that the return on equity not its Sharpe ratio is a constant of nature.

OK. For $100 of assets, and $100 of bank equity, let, say, $10 of that equity be traded — enough to establish a liquid market. Then, let $90 of that equity is held by a downstream entity or entities— a fund, special purpose vehicle, holding company or other money bucket. I’ll call it a holding company, and return to legal structures below. The holding company, in turn, issues $10 of holding company equity and $80 of debt.

There you have it — $100 of bank assets are “transformed” into $10 of very safe bank equity, $10 of risky and high return holding-company equity, and $80 of short-term debt.

Now if the bank loses money, the value of the bank equity falls. But the bank is failure-proof and run-proof. Shareholders get mad, may throw out management, may even break up the company. But they cannot run, demand their money now, and force bankruptcy.

The holding company can fail however! Suppose he bank loses $20. The holding company owes $80 of short term debt. Its assets are worth .9 x $80 = $72. It’s insolvent. It fails. Holding-company equity holders are wiped out. Holding-company creditors get the assets, common stock in the original bank, worth $72/$80 = 90 cents on their original dollar.

It need not be that drastic. Its likely the previous short-term debt holders don’t want stock, and would want to sell it in a hurry. Dumping 90 shares on the market might be tough.

The holding company could do a 5-minute recapitalization instead. Holders of the $80 of debt get $60 of debt and $12 of new holding-company equity. The holding company is recapitalized by the flip of a switch.

The key: this resolution/recapitalization can happen in about 5 minutes.

It takes no lawyers, no bankruptcy court, no resolution authority, no FDIC, no deep pocketed buyer, no deal sweeteners and toxic asset subsidies from the Fed, no months in court trying to find rehypothecated securities in foreign branches.  The bank itself keeps humming along. There is no interruption in lending activity, no risk that the ATM machines go dark. There is no run by depositors, brokerage clients, derivatives counterparties. There is no destruction of bank human capital.

The holding company assets—common stock in the bank — are completely transparent. They are liquid, and marked to market instantaneously. They can be handed to creditors at the flip of a switch. There is no tension between "illiquidity" and "insolvency," no "impaired assets." We know what the holding company is worth on a millisecond basis.  There is no "contagion," in which failure of one bank holding company leads people to question another, because everyone knows what the holding company is worth at all times. We don't need any regulators or accountants to flip any switches -- the second the market value of holding company equity falls below a given threshold, the failure or recapitalization happens instantly. Electronically.

As a reminder, here is how banks are organized now. In the event of bankruptcy or a run, the large amount of debt has a direct claim on the bank's assets. To realize that claim, though, they have to go through bankruptcy court or the resolution authority, and then sell actual assets. Assets are illiquid, hard to sell, especially in a crisis. This takes years and a lot of lawyer fees. Meanwhile the bank operations are often frozen, and its ability to serve customers and make loans is impaired, so the economy suffers.

Current regulation includes “living wills” that are supposed to make it easy to tear up a bank quickly, but there is a lot of doubt that will work. The Dodd- Frank "Resolution authority" is supposed to step in as the FDIC does, to quickly force a resolution before too much value is lost. But the idea that a few government officials can do over a weekend what bankruptcy court cannot accomplish in months seems weak, at best.
The living will idea has not been a huge success, with the Fed flunking several banks’ proposals. More deeply, the idea that in the midst of the next crisis — imagine early October 2008 — our government really will step in to a troubled big bank — Citi, say — and force big losses on the other creditors, each screaming their own "systemic importance" seems questionable.

I won't here belabor the options, but there are lots of ways to organize the same basic idea. I originally thought of the holding company as a mutual fund, exchange traded fund, or special purpose vehicle, to emphasize how mechanical the whole thing is. Now, I like the idea of a holding company a bit better, as equity also has voting and control rights. But this post is about finance and not law, and I'm not that good at corporate finance or law anyway. So fill the box any way you'd like.

"Holding company" is attractive though because many banks are already organized around a large holding company which effectively owns shares in multiple banks. So really all we're doing here is cleaning up the relationship between holding company and banks, and who has rights to what claims on either.

The central point: Debt, and especially short term debt, is a liability of the holding company not a liability of the original bank. Its rights are limited to recovering common stock of the bank. As stockholders, its owners can choose to liquidate the company if they choose. But they do not have the right to seize bank assets directly.

Clearly, having gone this far, the holding company could be diversified, and perhaps interleaved -- one holding company holds shares in several banks, and each bank could issue equity to multiple holding companies -- or leveraged bank-stock ETFs, which is what they become. Such diversification would make the debt even safer.

The idea has some parallel with cocos -- convertible bonds. The idea there is to issue bonds that trigger conversion to equity under certain conditions. The trouble is just what are those conditions. You need accountants or regulators to peer into notoriously obscure bank balance sheets and decide to trigger the conversion. With the holding company structure, the market price of bank shares does that for you.

What objections remain? Yes, the deposits and short term debt I describe are not perfectly risk free. Diversified holding companies would help. Capital ratios on holding companies would help. Partial funding with long-term debt, as with current banks, would help. But the current structure is not risk free either (!), absent a government guarantee.  In fact, the risk is a good deal smaller, even at the same capital ratios, because this system pretty much eliminates runs and crises. There is no uncertainty about the holding company's status.

The other possibility is that short term debt, with a bankruptcy court claim on assets, is somehow deeply tied to the bank's investment or loan origination activity. Lots of people make such a claim, perhaps that run-prone debt is needed to discipline bank managers in a way that closely-held equity cannot do. If there is such a tie, however, it is severed by government deposit insurance, guarantees, and resolution, so we're not making progress as things are.

(Some of these thoughts are prompted by "Bank Resolution and the Structure of Global Banks" by  Patrick Bolton Martin Oehmke, which Martin recently presented at Stanford. They study a similar issue of where and how regulators restructure banks.)


  1. I wonder if anyone would actually want to lend money to a holding company structured like that. Especially if it really does get unwound instantly, there's not much room for the government to bail you out even if they wanted to. Would investors demand a higher interest rate and buy the bonds or would they rather just buy the bank equity directly? Would the answer tell us how much those potential bailouts are really worth to people?

    1. The point was to avoid government bailouts! I hope I made that clear, and you're pointing out quite correctly that this doesn't have a chance while the government guarantees and bails out competitors, or that you're pointing out the devilish political economy fact that if we do things this way there will be little cause for people to demand bailouts.

    2. Who, actually, demanded bailouts?
      Was it the bankers, who caused the problem, who demanded bailout?
      I don't think it was ordinary folks demanding it.
      I like the notion of letting over-speculative bankers fall on their noses while ensuring that ordinary miniscule-interest depositors get to keep their savings.

    3. "I hope I made that clear". you did.

      What I was wondering is: if banks are currently structured the way they are in order to attract bailouts, would investors even want bank debt instead of bank equity if they knew for sure that bailouts were off the table?

  2. I must be missing something here.

    "Here's the idea: Let a bank fund its risky investments 100% by issuing equity. The bank then simply cannot fail — it cannot go bankrupt, it cannot suffer a run. As I've argued elsewhere, I think this is entirely practical."

    But, if the bank is funded by 100% equity, there are no depositors to run on the bank to begin with. Also, since there are no depositors, we won't even care if it did in fact default which kind of beats the purpose.

    What am I not seeing?

    1. I think the depositors running into the bank would have to resort to one of 3 strategies:

      1. pay the bank to keep their money for them. This is not far-fetched, as people pay (a little) money to brokers for this service. There would be depositors, but they won't be taking any risk, hence no runs.

      2. Have the bank invest depositors' money purely in "risk-free" government bonds and such. I think this is what professor Cochrane has in mind. Again, "no risk" so no runs. Perhaps more accurately, no government subsidy for risk (hopefully). Depositors aren't trusting their money with the bank but rather with the government to begin with, and the bank doesn't get to pay them a risk-free rate while lending out the money at high interest.

      3. Lend their money to the aforementioned bank holding companies, with the bank acting as an intermediary, but not as a borrower.

      So you can still have "depositors" and "deposits", but the deposited money is not lent to the bank. In scenarios 1 and 2, these depositors won't feel much of a difference IMHO. In scenario 3 they would be accepting much more risk relative to today, and should be aware of that.

      Finally, you ask why should we care if the bank fails in this scenario. We might care a lot less since retail investors would not be directly affected (unless they knowingly took on risk by lending money to a holding company). I think it's a matter of removing the (unbounded) bank failure risk from the taxpayer and shifting it to knowing and consenting, and duly-compensated market participants. These market participants would hopefully account for this risk better than the government and prepare for it better, without the need for intense bank regulation.

    2. Excuse me, I didn't read carefully enough. Professor Cochrane clearly suggests that depositors would be lending money to the holding companies - which is a non-trivial acceptance of additional risk for these depositors, IMHO.

  3. I think this proposal is fairly similar to the current living will structure to be honest. They all suggest having holding co and trading entities, with the holding co failing, while the trading entity goes on. The only thing the current living will is missing is the tradable 100%-equity part, which is an interesting idea.

    My concern with this structure would be that the 10% float on the 100% equity bank might be easily manipulated. Any distortion in a crisis of those shares would cause a substantially inflated effect on the holding shares. You'd get the same issue you get when the derived market is significantly bigger than the underlying market, with strong temptation to manipulate it.

    Also any such proposal would seem to require significantly higher capital ratios than current structures. That might well be enough without having to mess with the structure even more.

    I am also somewhat skeptical you'd be able to run a 100% equity based bank. Are you going to issue shares more or less continuosly like an ETF? The nature of bank assets is quite different. If a bank wants to issue a new mortagage, does it need to sell equity to the holding co automatically? At what price? if you have direct and indirect owners, you'd need some transparent price to value the new shares...

    I understand technology allows new solutions to banking, but the current structure evolved over time. To assume we can just change the entire system without significant risk seems very optimistic. I would much prefer a more gradual approach.

    1. You have some really good points. I'm dubious about "manipulated," but the bank equity will certainly be more volatile than assets held at book value. The current system evolved over time subject to huge regulatory distortions. We are part of evolution. Economic evolution takes human creativity.

    2. May I ask, why shouldn't banks issue shares more or less continuously? The price would be the highest bid price offered at the exchange. The existence of the holding companies should drive this price up, until it reflects the public's preference to lend money to banks.

      Direct and indirect ownership is common in financial markets. Why are novel mechanisms needed? Each actor must make its own assessment of the shares' value and act accordingly. Why do banks need total transparency of asset values when no other company has such transparency?

      I'm probably missing something important, being a mere layman, so I look forward to being enlightened.

  4. In accordance to Basel iii. How does an interest payment to a bank, that is made from a deposit account at the bank, contribute to the bank's capital. Have any of the commenters here have any thoughts on this.

  5. I would be curious how one would value the company differently based on this change. Obviously you could still use dividend yields of each of the respective pieces. However, the levered version is also riskier, so should have a different WACC.

  6. Traditional banking typically involves provision of liquidity, with banks offering things like overdraft and liquidity facilities to business. The ability to do this is one of the things that distinguishes a bank from simply being a structure for pooling risk. How do you see this working? Without access to short-term debt, banks would have to carry vastly increased portfolios of liquid assets, wouldn't they?

    1. I don't see that. Deposits are liabilities of the holding company. Loans are assets of the bank. They can be linked, so overdraft protection, offered by the bank, is linked to the deposit account, all seamless to the depositor.

  7. Who would want to buy the holding company debt when it is obviously not risk free? Surely not in the same quantities as short term bank debt. People want "safe assets," AKA cash substitutes. Bank deposits, money market funds, short term repo contracts are all ways to create "fake" safe assets that the government doesn't provide enough of through T-bills.

    How much of this problem of bank runs could be solved if we implemented your proposal and then massively expanded the government's balance sheet by swapping taxes with more debt?

    1. Why isn't it enough to implement the proposal? Why is it mecessary to additionally create goveenment debt?

      Government debt is not the only cash-equivalent. Cash itself is cash-equivalent.

    2. People want safe, risk-free assets. Except for government debt, these instruments pretty much look like short term deposits prone to runs. Prof. Cochrane's proposal is specifically structured so there are no safe assets anymore generated by the private banking system. That generates liquidity concerns unless new safe assets can replace them.

      I would support letting all market participants access central banking accounts at the Fed (also risk-free), which now can only be done by a small handful of banks. There's no difference between this and issuing a bunch of T-bills.

  8. Sadly you believe the people running the system and creating the rules are honest. Ever see the houses of people once they go to Washington? They have been purchased or selected to create a system to reward those that own it! You don't really believe former Goldman Sachs Millionaires are going to set up a system that doesn't reward the gamblers and reward those that don't own the system. Maybe you could get one of those $1/2 Million 20 minute speaking fees....Oh wait you were never in power and won't be! I remember believing in the integrity of leaders under Reagan...I was a naive college student and he was the last smart honest president.
    Nice idea...never will happen.

  9. If you can re-capitalize at the HoldCo level, what precludes you from doing the same at the top-tier level under the current system. I'm failing to see how dropping debt down in the structure does anything but reallocate the risk to that level, and a systemic collapse would be one where the HoldCo entities are at risk. Seems to me, all you've done is raise the capital requirements in the overall banking structure by assuming an additional $10 of equity at the top-tier entity. But perhaps I'm misunderstanding the re-capitalization mechanism and underestimating the value of spreading risk across many HoldCo entities.

  10. John,

    "After the financial crisis of 2007–08, many U.S. investment banks converted to holding companies. According to the Federal Financial Institutions Examination Council's (FFIEC) website, JPMorgan Chase & Co., Bank of America Corp., Citigroup Inc., Wells Fargo & Co., and Goldman Sachs Groups, Inc. were the five largest bank holding companies in the finance sector, as of 31 December 2013, based on total assets."

    What is the difference between a bank holding company and a bank?

    All that it sounds like you are doing is splitting hairs on definitions. Under your proposal - JP Morgan, Bank of America, Citigroup, Wells Fargo, and Goldman Sachs (all now bank holding companies) could lever up again.

    The more things change, the more they remain the same.

    Also, your diversification plan doesn't make sense from a tax perspective - I (as a holding company) can own 80% or more of a single bank's outstanding shares and receive tax consolidation benefits (including tax free dividends) or I can own 20% stakes in four similar size banks and not receive those benefits.

  11. Maybe it is the jet lag, but I cannot understand the proposal you have made. Your charts a very unclear. I suggest re doing it as nested T accounts. If you want help, contact me off-line.

  12. The fundamental problem is the management of the risk associated with the supply of liquid short debt and deposits that the market demands. Whether this supply is provided by a bank or an upstream bank holding company is irrelevant. For a given supply of what the market wants, there’s no difference between a run on the bank and a run on an upstream the banking holding company. It just upstreams the fundamental problem. The holding company equity sandwich doesn’t change this.

  13. Having this trigger when the market cap falls below a certain value seems like a bad idea. The bank shares could simply be depressed by a general market panic, despite the fact that the banks in question might still be perfectly robust. In fact, a diminishing of share value might exacerbate a sell-off, because the shares hitting a certain value is the same as the shares hitting $0 for the (about-to-be-disowned) shareholders.

    It seems unlikely to me that market has a better understanding of the solvency of a bank than management working together with the government/regulatory agencies, who have inside knowledge about the situation.

  14. In reply I argue that one economic function of banks is to warrant the safety of deposits by bearing a first loss position on bank assets here:

  15. JC says at the start of his article "Suppose it really is important for banks to create money..." Very big "supposition" that. Reason is that I don't see any possible excuse for privately created money and the attendant risks.

    In particular, a central bank and government can perfectly well create and spend enough money into the economy to keep it working at capacity. So why court the risk of bank crises and years of subsequent excess unemployment?

    The actual reason we do not have a ban on privately printed money was pointed out by Milton Friedman in the preface of his book, A Program for Monetary Stability". As he put it, “The vested political interests opposing it are too strong….”.

    1. Spot on, and thanks for emphasizing it. In my view, as I think in Ralph's, there is absolutely no need in a modern economy for banks to "create" money by issuing deposits backed by mortgage loans. Our government can create as much money -- both cash and interest-paying liquid debt -- as necessary, in the form of government debt. This is an old saw, possibly useful as a description of the 19th century when governments issued only coins and not banknotes, but does not at all describe today. The post is, from a Cochrane (-Musgrave?) nirvana point of view, pointless. Pure equity financed banks alone (step 1) with abundant government provided money would do nicely. But, I do like to find halfway solutions, and I think it strengthens the case for nirvana that we don't have to jump all the way right at once.

    2. Ralph / John,

      "In my view, as I think in Ralph's, there is absolutely no need in a modern economy for banks to create money by issuing deposits backed by mortgage loans. Our government can create as much money - both cash and interest-paying liquid debt -- as necessary, in the form of government debt."

      The same argument can be made for any private enterprise.

      For instance, Stanford is a private teaching institution. Why should it be allowed to continue to exist? The government can surely do everything that Stanford does now.

      "As he (Milton Friedman) put it, The vested political interests opposing it are too strong…."

      There is a vested interest in private banking just as there is a vested interest in every other private enterprise (education, health care, manufacturing, farming, etc.).

  16. Frank, You missed the point I made above i.e.: "So why court the risk of bank crises and years of subsequent excess unemployment?"

    In other words banning privately printed money and giving the exclusive right to do that job to the central bank does not simply amount to A doing a job rather than B, with all else equal. It results in the elimination of the catastrophic costs of bank crises.

    It does not however result in the elimination of ALL volatility: that is, human beings will always be herd animals or "lemmings" if you like. Outbreaks of irrational exuberance will always occur. But at least some of the volatility is toned down.

    1. Ralph,

      No I didn't miss your point about excess unemployment.

      First, full employment is not the be all and end all of economic policy. The federal government could pay people to dig holes and fill them back again thus generating full employment.

      That is the impetus behind the Humphrey Hawkins Act:–Hawkins_Full_Employment_Act

      "Explicitly states that the federal government will rely primarily on private enterprise to achieve the stipulated goals - full employment, growth in production, price stability, and balance of trade."

      Your statement:

      "In other words banning privately printed money and giving the exclusive right to do that job to the central bank...results in the elimination of the catastrophic costs of bank crises."

      There are approaching 9 million people employed in the financial sector in the U. S.

      Since your sole concern seems to be unemployment, what do you say to those people that lose their job when private banking is eliminated - here's a shovel, now go dig me some holes?

  17. So... How does repo fit into this scenario? How about derivatives with CSAs?

    I think the central problem with your analysis is that the two boxes you have for "Loans, Assets" and "Debt" are really one big jumbled up box designed very carefully for leverage (mandatory for efficient market making/arbitrage) while still issuing debt that the market will be willing to buy (relatively safe instruments like repo and collateralized derivatives).


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.