Thursday, October 19, 2017

Tyler: Equity financed banking is possible!

Tyler Cowen wrote an extended blog post on bank leverage, regulation and economic growth on Marginal Revolution. Tyler thinks the "liquidity transformation" of banks is essential, and that we will not be able to avoid a highly levered banking system, despite the regulatory bloat this requires, and the occasional financial crisis. As blog readers may know, I disagree.

A few choice quotes from Tyler, though I encourage you to read his entire argument:
I think of the liquidity transformation of banks in terms of...Transforming otherwise somewhat illiquid activities into liquid deposits. That boosts risk-taking capacities, boosts aggregate investment, and makes depositors more liquid in real terms.  
Requiring significantly less bank leverage, at any status quo margin, probably will bring a recession. 
...many economies are stuck with the levels of leverage they have, for better or worse. 
I fear ... that we will have to rely on the LOLR function more and more often. 
I don’t find the idea of 40% capital requirements, combined with an absolute minimum of regulation, absurd on the face of it. But I don’t see how we can get there, even for the future generations.
Depressing words for a libertarian, usually optimistic about markets.

This is a good context to briefly summarize why "narrow", or (my preferred) equity-financed banking is in fact reasonable, and could happen relatively quickly.

Tyler's main concern is that people need a lot of "liquidity" -- think money-like bank accounts -- and that unless banks can issue a lot of deposits, backed by mortgages and similar assets, bad things will happen -- people won't have the "liquidity" they need, and businesses can't get the investment they need.

Here are a few capsule counter arguments.  In particular, they are reasons why the economy of, say 1935 or even 1965 might have required highly levered banks, but we do not.

1) We're awash in government debt.

We've got about $20 trillion of government debt. That could back about $20 trillion of risk free assets. (It would be better still if the Treasury would issue fixed-value floating-rate debt, needing no intermediation at all.) Add agency debt -- backed by mortgage backed securities that are already guaranteed by the Treasury -- and you have another $8 trillion. Checking accounts are about $1.5 trillion and total bank liabilities about $9 trillion.

In the past, we may have needed to create money-like deposits by backing them with bank assets. A happy side to our debt expansion is that government debt -- the present value of the governments' taxing authority -- provides ample assets to back all the money-like deposits you want.

2) Liquidity no longer requires run-prone assets. Floating value assets are now perfectly liquid 

In the past, the only way that a security could be "liquid" is if it promised a fixed payment. You couldn't walk in to a drugstore in 1935, or 1965, and trade an S&P500 index share for a candy bar.  Now you can. (And as soon as it is cleared by blockchain, it will be even faster and cheaper than credit cards.) There is no reason your debit card cannot be linked to an asset whose value floats over time.

This is the key distinction. The problem with short term debt is that it is prone to runs. Financial crises are runs, period.  Short term debt is prone to runs because it promises a fixed amount ($1), any time, first come first serve, and if the institution does not honor the claim it is bankrupt.

Seriously. Imagine that your debit card was linked to an ETF that held long-only, full allocations (not risky tranches) of high grade mortgage backed securities. Its value would float, but not a lot. Bank assets are, curiously immensely safe. So it might go up or down 2% a year. In return you get a higher interest rate than on pure short-term government debt (of which there is $28 trillion under my scheme).  You would hardly notice. Yet the financial system is now immune from runs!

3) Leverage of the banking system need not be leverage in the banking system. 

Suppose even this isn't enough and we still need more risk free assets. OK, let's lever up bank assets. But why should that leverage be in the bank. Let the banks issue 100% equity. Then, let most of that equity be held by a mutual fund, ETF, or bank holding company, and let those issue deposits, long term debt, and a small amount of additional equity. Now I have "transformed" risky assets into riskfree debt via leverage. But the leverage is outside the bank. If the bank loses money, the mutual fund, ETF, or holding company fails... in about 5 minutes. The creditors get traded equity of the bank, which is still at 90% of its initial value. There is no reason bank creditors should dismember a bank, go after complex and illiquid bank assets, stop operation of the bank. If bank assets must be leveraged, put that leverage outside the bank.

And, if you need even more leverage, well, these leveraged ETF can hold other assets too. There is no reason not to leverage up stock, corporate bonds, REITS, mortgage backed securities or other assets if we desperately need to provide a riskfree tranche. We don't see this. Why not? Maybe "riskfree" assets aren't so important after all!

Tyler sort of acknowledges this, but with fear rather than excitement:
But what if a demand deposit is no longer so well-defined? What about money market funds? Repurchase agreements? Derivatives and other synthetic positions? Guaranteeing demand deposits is a weaker and weaker protection for the aggregate, as indeed we learned in 2008. The Ricardo Hausmann position is to extend the governmental guarantees to as many areas as possible, but that makes me deeply nervous. Not only is this fiscally dangerous, I also think it would lead to stifling regulation being applied too broadly. 
A lot of commercial bank leverage can be replaced by leverage from other sources, many less regulated or less “establishment.” Overall, on current and recent margins I prefer to keep leverage in the commercial banking sector, compared to the relevant alternatives....One big problem with attempts to radically restrict bank leverage is that they simply shift leverage into other parts of the economy, possibly in more dangerous forms...
Absolutely. In my view nobody should issue large quantities of run-prone assets -- fixed value, immediate demandability, first come first serve -- unless backed by government debt. However, we should cherish the rise of fintech that allows us to have liquidity without run-prone assets. And don't fear even leverage outside commercial banks without thinking about it. My ETF, whose assets are common stock, and liabilities are say 40% "deposits", 40% long-term debt, and 20% equity, really could be recapitalized in 5 minutes, without any of the adverse consequences of dragging a bank through bankruptcy court.

4) Inadequate funds for investment I'm not quite sure where Tyler gets the view that without lots of unbacked deposits, funds for investment will be scarce -- just how leverage
boosts risk-taking capacities, boosts aggregate investment,...
Requiring significantly less bank leverage, at any status quo margin, probably will bring a recession. 
The equity of 100% equity financed banks would be incredibly safe. 1/10 the volatility of current banks. It would be an attractive asset. The private sector really does not have to hold any more risk or provide any more money to an equity financed banking system. We just slice the pizza differently. If issuing equity is hard, banks can just retain profits for a decade or so.

Or, better, our regulators could leave the banks alone and allow on on-ramp. Start a new "bank" with 50% or more equity? Sure, you're exempt from all regulation.

And, in case you forgot, we live in the era of minuscule interest rates -- negative in parts of the world; and sky high equity valuations. All the macroeconomic prognosticators are still bemoaning a "savings glut." A scarcity of investment capital, needing some sort of fine pizza slicing to make sure just the right person gets the mushroom and the right person gets the pepperoni does not seem the key to growth right now.

Update: Anonymous below asks a good question: "What about payrolls, debiting and crediting exports and foreign transactions, escrows."  And I could add, accounts receivable, trade credit and so forth.

Answer: We need to eliminate large-scale financing by run-prone securities. Not all debt is run prone. It needs to be very short term, demandable, failure to pay instantly leads to bankruptcy, and first come first serve. And it has to be enough of the institution's overall financing that a run can cause failure. An IOU for a bar bill -- pay for my beer, I'll catch up with you next week -- is a fixed-value security, yes. But it is not run prone. You can't demand payment instantly, bankrupt me if I don't pay, I have the right to postpone payment,  it's not first come first serve, and such debts are a tiny fraction of my net worth.

Update: A correspondent writes
[Equity financed banking] Already exists! Albeit not at scale yet. It’s called asset management. See, for example, Alcentra, a UK-based company that lends directly to mid-sized European companies. They are largely “equity financed,” meaning that they sell shares in their funds, mostly to institutional investors. They also offer separate accounts, which you can also think of as “equity financing.” They are not a bank, but an asset manager, taking advantage of reduced lending since the crisis by banks to mid-sized and low credit firms in Europe. They have about 30 billion in AUM. This is a “disintermediation” story no one is talking about, and direct lending by asset managers is on the rise more broadly as well.


  1. John,

    1) "We're awash in government debt.
    We've got about $20 trillion of government debt."

    But think of where that debt is held:
    a. Private Retirement plans
    b. Overseas buyers
    c. Social Security holdings
    d. Central Bank holdings

    None of this government debt is very liquid in terms of transforming it into funds that can be spent / invested elsewhere.

    2. "...In return you get a higher interest rate than on pure short-term government debt (of which there is $28 trillion under my scheme)."

    No you don't. Sheesh. You can't directly compare the interest rate paid on a mortgage (amortized interest) vs. the interest rate received on a government bond (either coupon or accrual interest). With a mortgage, the interest rate is only applied to the amount of principal remaining to be paid on the loan. With a government bond, the interest rate is applied to the full principle amount over the term of a loan.

    For instance, if you have a $100,000, 5%, 30 year mortgage, you will pay about $93, 256.00 in interest over the life of the loan. If you buy $100,000 of 5%, 30 year Treasuries you will receive $150,000 in coupon interest payments over the term of the bond.

    A 5% 30 year mortgage nets about the same interest as a 2.2% 30 year Treasury bond.

    3. "Now I have transformed risky assets into riskfree debt via leverage. But the leverage is outside the bank. If the bank loses money, the mutual fund, ETF, or holding company fails... in about 5 minutes."

    Where do mutual funds, ETF's, holding companies get the money to buy the equity shares sold by banks? Are we eliminating Fed lender of last resort or is the U. S. Fed making short term loans to these other parties so that they can turn around and buy bank equity? How is this an improvement over what we have now?

    1. No, but you can compare the *yield (to maturity)* on a government bond and the "amortized interest" on a mortgage. The entire point of the concept of "bond yield" is that it is the interest rate (when computing yields, we assume it is held fixed over the entire maturity period of the bond) a one period, zero coupon bond has to pay next to the longer maturity bond so that there's no opportunities for arbitrage in the market.

      You're right that if the yield of a Treasury is equal to the "amortized interest" on a mortgage (assuming equal maturity), then the mortgage pays out less dollars in total over its lifetime than the Treasury does. But how is this surprising in any way? If you will get the money further in the future (the Treasury pays the par value of the bond in lump sum at the end of the maturity term, so the payment structure is shifted further into the future), holding yields fixed, the price of the asset falls. To compensate for this and for the two assets to have the same price, the Treasury has to pay a greater amount in total to the investor. This doesn't mean the two assets have a different yield; and this is why we quote YTM instead of something else when talking about the bond market. (Formally, there are many possible short term interest rate paths over time which could give the same price on a Treasury bond; when computing the yield, we choose the one which is flat - this is by no means a canonical choice.)

      Maybe you know all of this, and you decided that the point about the absolute quantity of interest paid being different, without regard to the term structure of those payments, invalidated the statement that "you get a higher interest rate than on pure short-term government debt". (This is not relevant to the main point, but since when is a 30 year Treasury considered to be 'short term'?) In that case, note how Prof. Cochrane says "interest *rate*" instead of "interest". But you can set this all aside - just by understanding the efficient markets hypothesis properly, you can see that investing a fixed amount of money in mortgages or government bonds, and reinvesting the proceeds in the same asset, will (in the absence of credit risk) give you the same return on your investment. In real life, mortgages would yield higher returns due to credit spreads; but there's no world in which mortgages yield more return than short term Treasuries (unless we have a sovereign debt crisis, but one hopes that this will never happen). If you want to argue against this, you have to point out a deep irrationality in bond and MBS markets, that nobody realized that if only they saw your one paragraph "argument", they would be able to get higher returns by dumping all of their money on 30 year Treasuries.

      Oh well, people do believe weird things.

  2. Currency banker do not know time to completion and shouldn't be issuing term debt. The central bankers could buy and sell securitie like they once did. The simplest method is for member banks to buy and sell wads of cash to each other, and the central banks acts as risk sharing market maker. The second option can be automated. We now how to compute the safe rate as an outcome, after the deals are done. Insurance companies can use that. Time to completion is inside information held in the real economy, leave it out of currency banking.

  3. Thanks for highlighting the Cowen article. I left several comments after it. I have couple of quibbles with the above Cochrane article.

    First, re No.1 “capsule counter-argument”, I don’t see what government debt has to do with it. I.e. I don’t see why a stock of government debt is needed in order to create “money-like deposits”.

    Take the extreme case: where there is no government debt and government declares that ALL bank loans must be funded via equity rather than deposits. That would probably (but I don’t think NECESSARILY) cause a decline in the bank industry and a rise in interest rates. In fact Cowen suggests that would cause a recession. But that’s easily countered by having the state print and spend base money into the private sector, until the private sector has whatever stock of base money induces it to spend at a rate that brings full employment. Problem solved and without there being a stock government debt to start with.

    Of course sundry rules in the US and EU prevent straightforward “print and spend”: i.e. government debt must be created as an intermediate product. But that’s a technicality.

    The net effect would be less loan based economic activity and more non-loan based activity (assuming interest rates rise). I.e. debts would decline. Assuming the numerous complaints we hear about excessive debt are justified, that doesn’t sound like a disaster.

    Also, current ultra low interest rates are a mixed blessing: they may exacerbate asset price bubbles and they don’t do any favors for savers or pensioners.

    Second quibble: John Cochrane says “The equity of 100% equity financed banks would be incredibly safe.” Personally I like the 100% equity system advocated by Lawrence Kotlikoff under which savers can choose whatever level of risk they like. That is, banks would set up mutual funds, some specializing in very safe loans, e.g. mortgages where house owners had a minimum 30% equity stake. Others would specialize in NINJA mortgages. So 100% equity banks could be as safe or unsafe as households want them to be.

  4. Valter Buffo, Recce'd, MilanOctober 20, 2017 at 3:43 AM

    Quote: "Imagine that your debit card was linked to an ETF that held long-only, full allocations (not risky tranches) of high grade mortgage backed securities. Its value would float, but not a lot. Bank assets are, curiously immensely safe. So it might go up or down 2% a year. In return you get a higher interest rate than on pure short-term government debt (of which there is $28 trillion under my scheme). You would hardly notice. Yet the financial system is now immune from runs! "
    End of quote. A few remarks could solve a misunderstanding.
    The above sentence reminds me of the Mid-2000 years, when everybody was saying "House prices will never go down. Confront with statics".
    Now we all know: house prices DO go down. And not just by 2%.
    Financial asset prices (ALL assets, even Sovereigns) also go down, more rapidly, a much more that 2%.
    The year 2017 will go down in history books as the "lowest volatility year ever": everything in 2017 financial markets is just "sellable". The illusion of "financial asseet fixed prices" creates the conditions that make a large amount of securities "sellable".
    The absence of volatility is, for reasons that I do not need to illustrate, artificial. It is due to non repeatable events.
    Assets which today are sellable, and in fact sold, to the public, will soon loose this quality. No, they are NOT always liquid (as in: market liquidity). They will show their NON-liquidity once these non repeatable market conditions eclipse, when NOBODY will buy at the margin.
    If someone opposes that Central Banks will guarantee unlimited liquidity with "QE forever", I reply that market liquidity is one thing, and Central State programmed buying of debt is NOT liquidity, and has totally different implications for the banking system and the economy at large.

  5. I think I made the same comment on one of the earlier posts.

    There's nothing all that special here about banks. Nothing really changes if you replace banks with corporations (more broadly). It's really about a way to speed up the bankruptcy process by separating the operating assets (or loans in the case of banks) from the way those assets are financed. There's no reason you couldn't create a structure that offered different levels of leverage for some corporation. One stock would hold 25% equity in the parent and finance the remainder with debt (d/e=3), while another might hold 50% equity (d/e=1). Perhaps the other way to think about it is that you have a 100% equity stock, which represents firm value (as an accounting concept). The different share classes each make different financing decisions (or maybe no financing), the equity component of those share classes would thus be valued differently depending on how it is financed.

    1. You're right that there is nothing in principal different about banks. What's special about banks is that their assets are unbelievably safe compared to those of other corporations. Compare the safety of a portfolio of mortgage backed securities and business loans to the portfolio of, say, Google's various moonshot projects, or even GM's cars. The bank assets are way, way, safer. Which is why it is amazing that our government puts so much effort into regulating the risk of ... the safest assets in the economy? Why? because they are levered to the hilt. They are levered for two reasons -- one, the safety of the assets leads naturally and sensibly to larger leverage. But two, because the government subsidizes leverage massively. As usual subsidize something to encourage it, then regulate it to stop it.

  6. How payments, payrolls, debiting and crediting exports and foreign transactions, escrows and other functions will be served by this system? These essential banking services are forms of debt.

    1. Payments are easy -- electronic transactions of government debt. Remember, we still have $20 trillion or so of completely safe assets to move around. Payrolls, debiting and crediting exports, etc. see the update above.

  7. I've responded to this post here:

    Carolyn Sissoko

    1. This comment has been removed by the author.

    2. John-how would you respond to these arguments?

  8. Why complicate the argument by relying on liquidity and stability of any financial asset other than government currency? What's wrong with dollar-backed full-reserve checking accounts? Depositors get practically nothing (or worse!) in interest from their banks in the developed world, don't they? And yet people are perfectly happy to hold "cash" in their checking accounts.

  9. The references to all Treasuries as "completely safe assets" for backing short-term deposits doesn't make sense to me. There is still meaningful interest rate risk except for short-term Treasuries. Here is a table that includes the year-by-returns for the 10-year Treasury bond, including both interest and price changes:

    In 2013, that annual return was -9.1%. In 2009, it was -11.1%. In 1999, it was -8.3%. So these are three calendar years in the past twenty years where the return on these securities declined much more than 2%. There have also, as you'd expect, been several years when the total return was greater than 10%.

    I can't readily locate similar return figures for high-grade MBS, which are also mentioned in the post, but I assume that total returns have proven to be at least as volatile since duration also increases when interest rates rise, due to fewer refinancings by existing borrowers.

  10. I'm not seeing how the hypothetical ETF idea solves the problem of runs. It strikes me as the same as proposing that a bank holding company wholly own the equity of a bank subsidiary, with all liabilities (including deposits) at the holding company level.

    If the depositors decide to flee, however, their 40% of the total capitalization has to come from somewhere. That could be by finding new deposits, or by raising more long-term debt, or by raising more equity. What if none of those options are available, however? Then the ETF needs to get the money by having the bank sell assets to send funds to the ETF, or it fails to meet depositors demand for their money. And then we're back to the prospect of a bank having to find a way to turn potentially illiquid assets into cash quickly, perhaps implying a distressed fire sale. And there's also still rollover risk on long-term debt, which after all isn't "long-term" forever because it does have various maturity dates.

    There are some interesting ideas here, but I think that this post waves away a lot of very real problems by simply describing structures similar to existing entities in different terms.

  11. Putting the above points together, doesn't the experience of interest-rate risk for early 1980's S&L's show how this model can break down from interest-rate risk alone? (And it would obviously get even riskier than that if we introduce credit rate risk into the equation as well.)

    The old "3-6-3" model of simple banking - pay depositors 3% interest, lend to borrowers at 6% interest, and play golf by 3 pm - breaks down if interest rates increase dramatically and the institution is long fixed-rate loans being funded with short-term deposits. At current interest rates, the institution is insolvent.

    So true "all-equity" banking implies that perhaps everyone suffers double-digit losses on their checking accounts, which isn't the sort of checking account that anyone wants.

    Or, if we get to the point of having to wipe out layers of capital in the hypothetical ETF, including some of the long-term debt, doesn't that imply that we've reached a messy enough legal situation that sorting out how to value the underlying assets of the bank and who gets what value will take some time? We can call that process bankruptcy or an FDIC resolution or whatever we'd like, but we have still have the issue of figuring out how we'll operate the underlying bank as a going concern if it doesn't have market access to new financing while we sort out the mess.


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.