Wednesday, July 15, 2015

Miles Looks Back

David Miles, retiring from the Monetary Policy Committee of the Bank of England, gave a fascinating speech on the occasion.  (Pdf with graphs here.) David's voice is particularly interesting since he's a real-world central banker, not an ivory-tower academic who can afford to have radical views. Many central bankers seem to evolve to the view that yes, they can push all the levers and run things just right. Not David.

Looking back: lessons from the global financial crisis
..the simplest, and arguably most effective, policy [to avoid financial crises] may well have low long run costs. That policy is to gradually change the funding structure of banks so that they are much better able to deal with shocks by relying less on debt and more on equity...

There are two fundamental reasons why having financial intermediaries fund their acquisition of assets with significant amounts of equity makes sense. First, it directly addresses the problems of improving incentives and preventing even limited falls in expected asset values triggering big rises in perceived risks of insolvency. Consider why the very large fall in asset values after the bubble burst did not have such devastating effects on the US economy. It was because all that frenzied activity was largely financed by equity and not debt. People who had funded much of the bubble lost money, but this did not trigger a whole series of insolvencies in the financial sector and disrupt the flow of credit to the wider economy.

Second, the long run cost of even rather big increases in the amount of equity funding of financial intermediaries is plausibly quite small. Substantial changes in the use of equity funding have already taken place since the crisis – and on some metrics required capital is as much as ten-times greater than pre 20087. And yet there is little evidence that the overall cost of bank funding has increased substantially. The paths of bank lending rates, both in absolute terms and relative to Bank of England Rates, have tended to fall (charts 2 and 3). And direct measures of the cost of bank funding have been on a steady declining path as capital ratios have risen (chart 4).
But won't the cost of capital rise and thus the cost of loans rise?
Simple finance theory suggests why, starting from very low levels of equity (high debt leverage), the impact of large proportionate changes in the use of equity on the overall cost of funds is likely to be small.

Consider the impact of doubling capital – or halving leverage – using the simplest possible back of the envelope calculation of a bank’s weighted average cost of funds. Suppose we start with leverage of 40 and cut it to 20 (that is with equity initially of 2.5% of total assets rising to 5%). Let’s imagine that the cost of debt financing is 5% and the required return on equity (its cost) at the original level of capital is 15%. First, if we assume that these costs will not change (a pretty big and unrealistic ‘if’ for a dramatic change in leverage), this will lead to total cost of financing increasing from 5.25% (0.975*5%+0.025*15%) to 5.5% (0.95*5%+0.05*15%), a rise of only 0.25pp. 25 basis points is what people used to think of as one typical MPC rate change at its monthly meetings.

And this is an extreme case in which the costs of equity and debt do not change. Theory suggests they should change so as to reflect the shift in riskiness as equity rises – debt becomes safer and equity returns less variable. At the extreme (and if the conditions for the famous Modigliani-Miller (MM) theory hold) there would be no change in the weighted cost of funds.
I hadn't thought of this. Even if MM is completely false for banks, the actual rise in costs of capital is small.
A combination of the limited liability of shareholders and deposit insurance almost certainly makes MM not hold for banks. But many of these factors may mean that while MM does not hold, the private cost of banks using more equity is not a true social cost.
In simpler terms, equity financed banks may face a higher cost of funds, because our governments subsidize debt. That fact does not mean that society as a whole as a higher cost of borrowing through equity-financed banks.

David goes on to an interesting question: Let us compare equity financed banking to the current rage, using monetary policy to identify and prick asset price "bubbles."
Might it then be that a better way to control risk taking and financial fragility is to use ...changes in the general level of interest rates ...

My own view is that skewing monetary policy towards trying to stop financial instability problems is, in general, unlikely to be the right answer. Yet many seem to think that the crash showed that having narrower aims of monetary policy – centred around an inflation target – was somehow proved wrong. I think that view fails to look at the deep reasons for the crash, which to my mind were the existence of excess leverage (too little equity funding) in banks. Excess leverage is not something effectively countered by a general rise in the level of interest rates. Higher interest rates will tend to increase required returns on both debt and equity and so it is not at all clear they encourage the use of relatively more equity. Capital requirements – a macro prudential tool – get to the heart of the problem.
I'm less in love with the "macro prudential" agenda, but in this case I cheer.

David makes another interesting point:
..bankers are right to say: For them raising equity is costly; and imposing a higher capital requirement will reduce aggregate lending.

Both statements are correct. But both miss the point. There may be too much lending in the unregulated state. Equity may look costly to banks but it has an overall beneficial side effect in better aligning the interests of shareholders with those of other claimants on the bank. To put the point another way: there is an inherent tendency in banking markets for there to be excessive risk taking. 
This is a nice point, which had not occurred to me. If the cost of debt financing rises, borrowers may choose equity financing instead. It's not obvious that the total amount of investment declines, or that it declines in a socially inefficient way. There is such a thing as too much debt!

Lessons about Monetary Policy: QE, ZLB and deflation
The global recession led many central banks to lower their policy rates to near zero. With the exception of in Japan, this was pretty much unchartered territory for monetary policymakers...

...the predictions from mainstream theoretical macroeconomic models for what would come next were not comforting... [For example] Eggertsson and Woodford (2003, EW) had analysed what happens at such low levels of policy rates and the likely effectiveness of asset purchases. They suggested that on hitting the lower bound an economy could suffer a deep deflation and recession and that asset purchases were not likely to help much. Their analysis suggested that the effective way to avoid deflation in such circumstances would be to commit to future inflation overshooting the target.

I found these predictions somewhat unrealistic, ...

I also doubt that there is a deflation cliff at the ELB. The evidence for thinking that deflation risks become great at the ELB is actually quite weak. There were no dramatic deflations among OECD economies (except for Ireland, which saw an exceptionally sharp fall in economic activity), and there was no clear difference in the change in inflation rates between countries that were constrained by the ELB and those that were not. Inflation fell in most OECD countries in 2009, but only a few experienced outright deflation.

... Neither actual nor expected inflation displayed the deflation cliff at the effective lower bound.
He's not quite "neo-Fisherian." But clearly the prediction of a deflation "spiral" or "vortex" at the zero bound troubled him at the start -- as it should have -- and no longer sits well.

I disagree mildly on the effectiveness of quantitative easing. David seems to think it worked. And his story for the absence of deflation seems to be in part that QE stopped it. But, he acknowledges Ben Bernanke's famous quote, "the problem with QE is that it works in practice, but it doesn’t work in theory." I'm reluctant to really believe anything works until we have at least a vaguely plausible understanding of how it works. Doctors believed in bleeding for a long time. One can see though how practical experience and academic reserve might differ here.


  1. John,

    My only notations would be:

    1. Banks (like any other business) use debt for capital improvements and other ventures unrelated to their operating business. Looking at a bank's balance sheet or cash flow statement will not tell you which dollar was used to build brick and mortar office in Chicago and which dollar was used to speculate on mortgage backed securities.

    Are banks now required to fund capital improvements with equity, while any other enterprise can fund it with debt? Does this push the stand alone banks into GE Capital type industrial bank arrangements - send jet engines and toasters out the front door and loans out the back?

    2. A bank can manage it's debt / equity ratio in real time fairly easily. A bank cannot manage it's asset / liability ratio in real time easily.

    1. Frank,

      Re 1:

      Does the disparity between treatment of banks vis-a-vis other firms seem unfair to you? I would think that because banks have access to lines of credit that others don't in good times, and a benevolent central bank in bad times, with this state of affairs comes a propensity to leverage unwisely that must be limited. They cannot be seen as similar entities to be treated equally from the perspective of liability profiles.

    2. Anoop,

      Not really unfair, just able to be gamed. If I am the "Bank of Frank" and I don't like issuing equity to do everything I simply buy out "Anoop's Car Dealership", change the name of my company to "Frank's One Stop", and use the car dealership to do my borrowing for me.

      Really, how do you say that this enterprise is a bank while this enterprise is not? You will not be able to tell when a company is performing credit intermediation in real time.

    3. Your car dealership will not be able to issue government insured deposits, access the federal funds market, get emergency lending from the Fed, or too-big-to-fail support. So when it proposes that people lend it money leveraged 30:1, those people will laugh.

    4. John,

      Under your various proposals, those things (eliminate Fed lender of last resort, eliminate deposit insurance, etc.) are gone away for traditional banks as well.

      So the question still remains - what stops traditional banks from "sneaking" back into the borrowing business by combining forces with a traditional non-bank enterprise?

      Your proposal isn't to reduce bank debt to equity ratios (if I read you correctly), your proposal is to eliminate bank borrowing entirely.

      I might lend a bank money on a 1:1 debt to equity ratio, but under your proposal - "Make banks issue equity to be able to lend" - even that is too much bank debt.

    5. Best I can do is in "toward a run free financial system,"

      Your car dealership will not be allowed to offer demand deposits. As, today, even banks are not allowed to issue bearer notes. That's about it for regulation. Not zero, but 19,999 pages less than Dodd Frank

    6. John,

      Thanks for the link, I have read most of it but I don't believe you addressed the possibility of industrial banking companies - for instance J. P. Morgan's U. S. Steel company in the early 1900's.

      Also (and this is my opinion) I don't think the U. S. Government can reasonably expect the banking industry to submit regular bids on U. S. Treasury debt when those banks must fund that lending by selling equity shares.

      The primary reason that the lender of last resort feature exists is so that the private banking industry can borrow from the central bank and lend to the federal government (always at a profit). If banks cannot borrow from the central bank, then it highly likely that U. S. government bond auctions would suffer regular failures.

      You address the fiscal angle at the end of one of your papers, but fail to comment on the primary dealer arrangement with the private banking industry.

  2. One of the social costs of capital requirements is the effect of poorly-chosen risk-weights. When the capital requirement binds, lending decisions are constrained by the weights; sometimes we get housing bubbles, other times reduced business lending. These problems of misallocation have been socially costly in recent years.

    We need a financial innovation that improves solvency of banks during downturns, without also producing the distortions related to the risk-weighting system.

    1. The fact that banks mis-judge risks is not important: EVERY business does that. The big benefit of plenty of equity is that equity holders foot the bill when things go wrong, not taxpayers.

      Next, why does the “solvency of banks” need to be improved “during downturns”? Downturns can be dealt with by the normal stimulatory measures: interest rate cuts, bigger deficits, QE, or whatever. I see no reason to concentrate on one particular form of stimulus (more bank lending) during a downturn. That is, BROAD BASED stimulus (increasing consumer spending and public spending) is the best remedy. And having done that, that will induce a finite increase in bank lending. Nothing encourages a bank to lend to a business like the sight of large numbers of customers coming thru the front door.

  3. As John Cochrane says, Miles makes the point that just because the cost of funding banks rises and the amount of lending falls, that is not necessarily undesirable because that might be a move TOWARDS the optimum amount of lending, not a move AWAY from the optimum. (Passage starting “David makes another interesting point”).

    That raises a big question, namely WHAT IS that optimum? Well the standard answer in economics, with which I agree, is that the optimum number of apples or anything else is produced in a genuine free market: i.e. a market where there are no subsidies.

    Thus the optimum amount of lending will arise where the state openly declares there shall be no form of taxpayer funded assistance for banks whatever: no lender of last resort – nothing. But that ipso facto means that all those who fund banks (including bondholders and depositors) become shareholders: that’s shareholders as in “someone who at worst stands to lose everything”.

    But what about depositors: those who are promised $X back for every $X they deposit? Well under a regime where depositors can “lose everything”, the very notion “depositor” is fraudulent. I.e. any entity that lends and funds that lending via deposits should be prosecuted.

    Ergo the optimum is a system under which entities that lend are funded JUST BY shares. As to those who want to deposit money in a totally safe manner, they can be catered for by the state.

    And that system is called “full reserve” banking: what Milton Friedman and Lawrence Kotlikoff advocate/d. It’s also (as I understand it) what John Cochrane advocates here:

    1. Any proposals for the Greek banking system, Ralph? In a recent blog, Prof. Cochrane insisted that sovereign debt should not be treated as risk-free by anyone.

    2. Anwer, My preferred solution for Greece is Grexit and re-introduction of the Drachma. As to the riskiness of sovereign debt, there are all shades of grey there. The debt, particularly short term debt of responsible countries (e.g. US, UK, Germany etc) is as near 100% safe as you’ll ever get. As to the Greeces, Argentinas, Zimbabwes etc, that’s obviously more risky.

    3. Yes, Greece can leave. But that doesn't give us a working model for banking within a currency union, for those who want to stay. Specifically, I don't see how the Chicago plan could be applied within the monetary union in its current form. Prof. Cochrane suggests that European banks can be made safe, not by using a market for exchange of regional risks, but by having banks that are so large that they span the continent and thereby achieve diversification internally.

    4. Anwer,

      "Prof. Cochrane suggests that European banks can be made safe, not by using a market for exchange of regional risks, but by having banks that are so large that they span the continent and thereby achieve diversification internally."

      One problem with having banks that span the continent is in uniformity of regulation. I don't think John wants to eliminate all governance regarding banks and lending (bankruptcy proceedings for instance). And so banks and individuals will be under the temptation to forum shop - lenders will lean on sovereigns that have tight creditor protections, borrowers will do the opposite.

      This is precisely what happened when commercial banking and investment banking were consolidated - banks were given latitude in choosing which regulators to listen to - SEC Regulators or Federal Reserve regulators.

      Should a bank that spans the European continent listen to regulators in Germany or France? Or should they pick and choose which to listen to and which to ignore?

    5. That's a great point, Frank. I like the idea of relatively small local banks that sell off aggregate regional risks via securities. That requires relatively little regulatory integration, and does not set up games of forum shopping. Of course the securities should be designed better than the toxic assets of the past.

  4. By all means, a financial system should be rock solid. Increase reserve requirements or the equity required.

    That said, a central bank can always print money and make a bank solvent. Moreover, banks tend to go insolvent at the very time a central bank should be printing money.

  5. David Miles is in the same camp as yourself Prof Cochrane, as well as Profs. Admati, Kotlikoff and many others. More equity has been a constant theme of the response to the financial crisis. I disagree with the overall thrust of this argument, as I do not see capital structure as the key solution, instead preferring the failures of risk management, governance and supervision to be addressed. Capital may be a symptom, but it is not the disease.

    However, assuming that more (equity) capital is a desirable policy, we should do well to bear in mind the following points:
    1. Cailbration. Proponents often suggest leverage ratios of 10%, 20%, 50% and even 100%. Since the theories behind the policy lack rigour (in my humble opinion), the actual level of calibration is hardly discussed and there is ambivalence as to whether 10% is better or worse than 100%.
    2. Capital mix. Increased levels of fixed-income sub debt (aka TLAC) offer an increased level of resilience while enabling capital efficiency through risk tranching and matching to investor needs. This is a more intelligent solution than plain old "debt or equity"
    3. Investor needs. As equity requirements rise, the volume of equity supply does not rise in tandem. Many investors do not want to -- or cannot -- take the investment decision on an equity basis; the supply of equity to invest in the banking system is constrained. More equity capital means less financing from banks
    4. Supply of capital to the economy. If a "better" capitalised banking system is indeed smaller, then we should expect the rational implications: fewer financing options eg. mortgages; more risk materialising outside the banking system eg. shadow banking in arguably weaker risk management environments; potentially increased volatility. Not all of these are necessarily bad, of course...
    5. One might be disturbed to see that the EC's Jonathan Hill is now wondering whether "more capital" is having unintended consequences: "To what extent have [the rules] affected the level of capital held by banks? Are they always proportionate to the risks? What impact are they having on lending to smaller businesses, and infrastructure?" ( My concern is that politics is overriding good risk management and we will have more lending that is "directed" in nature, which is not in the collective interest

    Perhaps the proponents of the "all equity" camp could illustrate how they would see the banking industry working, including size/volume and risk location? David Miles' earlier speeches seemed to use the 1850s as a role model, but I fail to see the Victorian system being appropriate to the modern context (eg. then there were no mortgages, no pensions, multiple fortunes won and lost).

    1. Re your claim that the arguments for 100% (or any other percentage) “lack rigor”, what’s wrong with my above argument for 100%, which is thus.

      There is no excuse for any sort of taxpayer funded support for banks or other entities that lend, any more than there is an excuse for taxpayers supporting garages or restaurants. Ergo lender of last resort and ALL OTHER forms of support for banks should be removed.

      That in turn means that all of those funding banks in effect become shareholders: that’s shareholder as in “someone who at worst stands to lose everything”.

  6. Jose Romeu RobazziJuly 16, 2015 at 8:12 AM

    Prof. Cochrane
    I like the discussion on bank's leverage, but I think it is incomplete. Leverage can be harmful depending on the level of riskiness of assets an entity is taking on. Leverage cannot be evaluated alone, in my view, it should be measured also taking into consideration asset/liability term mismatch. if you are highly leveraged but your assets are very little volatile and your liability term structure matches that of the assets, leverage is not as dangerous. For commercial banks, one could say, since most of its liabilities are deposits, there is always a potential large asset/liability term mismatch, and increasing equity would remedy that. I agree.

  7. I'm reluctant to really believe anything works until we have at least a vaguely plausible understanding of how it works. Doctors believed in bleeding for a long time. One can see though how practical experience and academic reserve might differ here."

    I am not sure this is a very helpful approach (or the analogy) to either understanding how the world works or how to implement good policy to make it a better place. I guess it depends what you mean by "theory" - perhaps you can be more specific. Friedman advocated purchases of long term government bonds in the early 2000s - I am sure he had a coherent view on why, perhaps informed by historical knowledge. If you say good policy must conform with neo-classical micro theory (and if this the theory you are talking about), well surely that's nonsense. We are not in the business of theology here. Micro theory, to rub in the irony should not be regarded as gospel, in fact it is as wrong as it is right. I am more sympathetic to the view of "unchartered territory". But actually I think there has been more experience with quantitative easing policies in various countries than a lot of macro-theorists seem to realise and more ground up historical investigation is called for. Once we have the historical facts sorted out we are in a position to (cautiously) construct generalised theories.

    1. "I'm reluctant to really believe anything works until we have at least a vaguely plausible understanding of how it works. Doctors believed in bleeding for a long time." Agreed, an insightful quote. QE is just an "asset swap". The private sector get the reserves or bank deposits and loses the bonds. Private sector financial assets are unchanged. Furthermore, since the reserves or bank deposits yield less than the Treasury securities, the private sector loses interest income. This loss has amounted to about $80B annually in recent years. Hence, while QE's indirect impact (portfolio rebalancing effect, wealth effect) may have been stimulative, the direct impact was mildly deflationary. Hardly what the "hyperinflation crowd" was telling us.


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.