Tuesday, March 14, 2017

Capital Illogic

More Bank Capital Could Kill the Economy write Tim Congdon and the usually sensible Steve Hanke in today's Wall Street Journal.

I was expecting a quantitative disagreement on plausible channels -- some explicit violation of the Modigliani Miller theorem, some reason that splitting the pizza into 8 slices rather than 4 will help your diet, some argument that relationship lending is inherently tied to short-term funding, and so forth. Instead, we got treated to one of the most illogical conclusions I've seen on the WSJ pages for a long time.

This is a really important argument to revisit, at a really sensitive time.  Right now, the Administration wants to rethink Dodd-Frank. Great. But they could go in two ways: 2) increase capital a lot, and get rid of all the intrusive and stultifying risk regulation and anti-competitive regulation; 2) reduce capital requirements a lot, so the big banks go on an orgy of government-guaranteed borrowing and risky investment. From my last post you can see it going either way. Anti-capital fallacies just pour slops into the second trough.

Their argument:
Here’s what really went wrong in the fall of 2008... Regulators were in such a panic that they hastily increased banks’ capital requirements from 4% to 7%, without thinking through the long-term ramifications.
There are two ways for banks to raise their capital-to-asset ratios: by increasing capital or decreasing assets. Which is most likely during a crisis? Issuing new equity or bonds would be difficult under crisis conditions, so banks will instead shed risky assets. In late 2008 and early 2009 that meant a drop in lending to the private sector and a credit squeeze. As businesses repaid loans, new ones were not issued in their place and the quantity of money in the economy fell. That hit demand, spending and jobs, just as it had in the Great Depression.
What was the fallout? In the five years preceding October 2008, bank lending to the private sector had soared by more than 75%, according to Fed data, from $4.2 trillion to $7.4 trillion. In the five years after, bank lending stagnated, increasing by less than 10%.   
The stock of loans actually went down during the first two years of this period, the only time such a thing has happened on a significant scale since the 1930s. The reversal was most extreme for industrial and commercial loans, which plunged from $1.6 trillion at the end of October 2008 to $1.2 trillion two years later. 
...The blame for this credit crunch falls on the Fed, acting in concert with the Bank for International Settlement 
Let's leave aside quibbling about the facts -- just what capital requirement they are talking about, when it it, and so forth. Leave aside the great argument whether any of this even happened in the crisis -- whether regulatory capital constraints were binding (the banks said no), and whether banks as a whole shed risky assets. Leave aside the "the quantity of money in the economy fell," and ignore the graph below:

Consider the logic of the argument. A rise in capital requirements in Fall 2008, singlehandedly caused a "credit crunch," and lending to "plunge," and bank lending to "stagnate" for five years. Among other influences not held constant here, let us not forget the TARP, which, like it not (I don't) gave the banks a massive shot of... capital.
That brings us to Mr. Kashkari’s proposal to further double capital requirements. What might happen if the Trump administration enacted his plan? Bank stocks would take another dive. They would find it impossible to raise new capital through equity or bond issues, so they would be forced to shed assets. As in 2009 and 2010, banks would refuse applications for new loans. They might go so far as to wriggle out of contracts for existing loans and ask for early repayment.
(Kashkari and a lot of others I might note -- especially the courageous Admati and Hellwig, and on the political side Jeb Hensarling. )

Here you see the deep illogic of the oped: It concludes, from the assertion that banks in the middle of a once per half century financial crisis, cannot raise capital in a week by means other than selling assets (even if that assertion is true), that banks cannot over five years or more, of normal and healthy financial markets, raise adequate capital to bear the brunt of the next crisis without cutting lending.

Banks have many ways to raise capital in functioning markets, and even in relatively dysfunctional markets. Even in 2007, if memory serves me right (there was a Bloomberg.com article on this I can't find right now, I welcome a source), banks raised something like a trillion dollars of new equity, in order to cover losses in their asset positions. When buyers come and take over a bank, that is an equity injection. Lehman itself was poised to be bought in this way, until UK regulators nixed the deal. Banks can reduce dividend payments, which increases the total value of equity, or big payouts to senior employees. If banks aren't lending, in a functional regulatory market (not ours) new banks can IPO and take over their business. In two weeks, in the middle of a crisis? Maybe not. In five years? That does not follow.

Yes, in a city-wide conflagration, "Everybody run down to Home Depot, buy and install sprinklers,  and grab a carful of fire extinguishers" is not going to work. It does not follow that in the five following years a city code that requires sprinkler retrofits and fire extinguishers cannot stop the next fire from happening.

And really -- after all the stress tests, after all the slow, modulated, carefully pre announced capital raises by the Fed and others, do you really think banks would be faced with an overnight, increase capital by tomorrow morning or else? I know regulators can be a bit thick at times, but not that much!
Banks have been required to hold more capital against their risky assets, in the belief that this would make them—and the economy—safer. 
Yes, dear sirs, and that is an entirely correct belief!  Deeply, Congdon and Hanke miss the point of capital: it offers superior returns on average, but takes the losses in a crisis without needing regulators to spot the crisis, prop up markets, inject capital, lend of last resort, bailout and so forth. It's not just there to gum up the works! This is entirely missing in the article. Not enough capital did indeed cause the crisis. If banks had 20 percent capital going in, there would have been no crisis, and no bailouts, because no bank would have gone under.

Anytime a commentator writes "hold" capital you know poor logic is coming. Capital is a source of funds, not a use of funds. Capital is not reserves.


  1. I fully agree that the WSJ article of Congdon and Hanke was rubbish. I actually said as much on my own blog a few hours ago. However, I think there is a point that JC has missed.

    Requiring banks to hold more capital might well raise their costs A BIT (though Google is funded 90% by capital and it seems to be doing OK). However, any deflationary effect of increased capital can easily be countered with bog standard stimulus - monetary and/or fiscal.

    The net effect would be less lending based activity (i.e. less debt) and more non-lending based activity. Given that it is generally held that private debts are excessive at the moment, where's the harm in that?

    Also, the assets and liabilities of UK banks have expanded TEN FOLD relative to GDP since the 1960s, and where has that got the UK? Growth is no better than in the 1960s, that's for sure. As Adair Turner, former head of the UK's Financial Services Authority put it, much of what banks do is "socially useless".

    A contraction in the bank industry, with GDP held constant, would do no harm at all.

  2. Almost all money out there exists purely because bankers created loans. No loan, no deposit, no money. Sure, there is some basic money in coins and notes, but this is peanuts. Most money comes from creating loans.

    If most money comes from creating loans, isn't it then quite normal that banks hold not that many assets? When a loan (and money) is created, bank assets go up and liabilities go up as well, and this money ends up somewhere in the economy. Equity stays constant during this transaction. So by creating the loan, the equity ratio went down in the banking sector.

    So is it right to conclude that, for the sector as a whole, either investors have to massively change their investment allocations (towards risky equity)? Only then the banking sector as a whole can attract more equity. If this doesn't happen, then the only way of increasing capital ratio's is by doing less business, i.e. stop creating as much loans.

    In other words, I'm not really sure it's that easy to just increase capital for the sector as a whole. The money must come from somewhere. So keeping the stock of money constant, investors must change their portfolios towards banking stocks. And if you allow the stock of money to increase, it must follow that loans are created, which by definition decreases capital ratios.

    Does this make any sense? I don't have the feeling I'm missing out on something here...

    1. Actually, this is no longer true. It is how standard money and banking courses are taught, but they are stuck in about 1965. Right now, with $3 trillion of reserves, most money (checking accounts) comes from holding treasuries. The Fed and banks are together one big money market fund, offering deposits which are backed by the Fed's treasury holdings.

    2. If the two St Louis Fed charts linked to below are any guide, then Fed created money as a % of commercial bank created money rose from about 10% in 2008 to about 30% now. But of course there is no hard and fast definition of “money”, plus I might have got something disastrously wrong with my 10% / 30% claim.



    3. John,

      "The Fed and banks are together one big money market fund, offering deposits which are backed by the Fed's treasury holdings."

      ??? - Here is my understanding of quantitative easing (QE):


      "On the other hand, QE can fail to spur demand if banks remain reluctant to lend money to businesses and households."

      And so, I don't know how you arrive at the conclusion that FOMC purchases of Federal Government Debt create deposits. My understanding is that those purchases create loanable funds - hence, if banks are unwilling to lend out those additional funds, QE won't do a whole heck of a lot.

      It is true that deposits can be created when government (or any other entity) borrows and spends money into existence, exactly as Kurt describes.

    4. Ralph,

      You are missing M3, something that the Fed stopped publishing data on in about 2006. See my post below regarding conversion from M3 to M1 in the time frame that John alludes to in his chart above.

    5. "Right now, with $3 trillion of reserves, most money (checking accounts) comes from holding treasuries. The Fed and banks are together one big money market fund, offering deposits which are backed by the Fed's treasury holdings."

      How long has this been true? Since the great recession or longer? Just curious

    6. Frank, You say “I don't know how you arrive at the conclusion that FOMC purchases of Federal Government Debt create deposits.” The answer is as follows, unless I’ve missed something.

      Fed prints $X of new money and buys $X of debt off Y. Y deposits the check at a commercial bank, and the commercial bank gives the check to the Fed and says “credit my account at the Fed, please” (i.e. increase my stock of reserves). Net result: bank deposits and the commercial bank’s reserves rise by $X.

      You then say “if banks are unwilling to lend out those additional funds, QE won't do a whole heck of a lot.” I agree that QE does not give rise to a vast amount of stimulus, but even if the above commercial bank does not lend on the $X, another possibility is that Y spends the money on some investment or other. Indeed that’s main desired effect of QE as I understand it, and that has a stimulatory effect.

    7. Ben Bernanke explained it clearly back in 2008 in congressional testimony. The Fed expanded the money supply and bought financial assets with the newly created money, but at the same time, 'sterilized' (Bernanke's word) it by paying interest on excess reserves.

      That's why the money creation didn't stimulate spending.

    8. Patrick,

      So the Fed bought a bunch of federal debt from banks -
      Banks get money
      Fed gets government bonds

      The Fed then uses the interest (or at least that portion they did not remit back to the Treasury) on that federal debt to pay banks to hold that money as reserves.

      John's statement was:

      "The Fed and banks are together one big money market fund, offering deposits which are backed by the Fed's treasury holdings."

      The money the Fed gives to banks for their bonds are loanable funds (not deposits).

      The status of the interest payments the Fed gives to banks (interest on reserves) is less clear - does it become operating income for the bank or does it remain as a portion of a bank's reserves?


      In your example - who is "Y"? If "Y" is itself a commercial bank, then "Y" would not reasonably deposit the check from the Fed at a competing bank. "Y" would either lend it out or sit on it and collect interest from the Fed for the reserves that it holds.

      If "Y" is a money market or other mutual fund, it is unlikely they would deposit a check from the Fed at a commercial bank collecting no interest. Presumably the interest that a commercial bank collects from the Fed on its reserves are retained by the bank and not passed onto depositors. In any case, the mutual fund would be better served buying Treasuries directly at auction and collecting full interest payments rather than getting interest payments 3rd hand - through the Treasury, through the Fed, and finally from the commercial bank.

    9. 'The money the Fed gives to banks for their bonds are loanable funds (not deposits).'

      Actually, they are deposits. The Fed pays banks for their bonds by creating deposits to their accounts.

    10. I don’t follow you, prof. Cochrane. Banks reduce their equity ratio when they create loans (assets go up, liabilities go up, equity remains). This is a fact. But even when the FED buys the bonds, commercial banks reduce their capital ratios! (again, assets go up, liabilities go up, equity remains). Because QE is not a transaction between the FED and bond holders. No, it’s a transaction in which the commercial bank is an intermediary.

      (Intermezzo: textbooks stuck in de 1965’s talk about money creation as if there is a certain amount of reserves (created by the central bank) in the economy, held by commercial banks. They then lend out money by “multiplying” the reserves. This is indeed how most people view banking, but I don’t. There’s an excellent Bank of England text on this. Google for “Money in the Modern Economy – Bank of England”. It’s great.)

      Anyway, when government bonds are bought, this ALSO creates deposits on checking accounts, no matter who buys them (FED or commercial banks). First, if commercial banks buy them, they gain an asset (the bond) and gain a liability (the deposit which the government can use). These deposits are FRESH money that never existed before. It’s obvious that this decreases equity in the banking system. Second, if the FED buys bonds, the FED does NOT create deposits. Instead, it creates reserves. Reserves are base money, which only central banks can create. It’s a liability for them and they use it to buy assets, like government bonds. Reserves cannot be created without an asset being bought from somebody. Let’s say a mutual fund has a bond that the FED want to buy. The FED then creates reserves, they transfer these reserves to the commercial bank where the mutual fund has its checking account, and that bank in return creates deposits on the checking account of the fund. The bond is transferred to the FED and we have the following:

      FED: +bond (asset) and +reserves (liability)

      Bank: +reserves (asset) and +deposits (liability)

      Fund: -bond (asset) and +cash (asset).

      All balance sheets are still in equilibrium. This is how QE works. At least, this is how it works in Europe and England. I’m not sure if in the US, the FED can buy straight from the Treasury, but I’m guessing not, since only commercial banks can create deposits that you and I and everybody but banks themselves can pay with. Nobody in the real economy cares about reserves. Reserves are only used by the FED and commercial banks. It’s the money of banks and banks alone.

      Nevertheless, this shows that an increase in the amount of deposit money ALWAYS decreases capital ratios. So both QE and loan creation decreases capital ratios. In order for them to go up again in the banking system, there MUST be a transfer from deposits to equity, i.e. money destruction. For example, when I buy a bank stock in an IPO, bank equity goes up (equity+) and deposits go down (liabiaility-). The balance sheet is in equilibrium.

      So it’s simple: in order to increase capital ratios for the banking sector, there has to be a transfer from deposits to banking equity. Me buying a bank stock might change my allocation, but it won’t change the aggregate allocation, since somebody else will have my deposit and I will have his bank stock. So it’ simple: banks need to either retain profits (which is a transfer from liabilities to equity) or banks need to sell new equity (which is also a transfer from liabilities to equity). For the aggregate portfolio, this means the amount of cash must come down (since deposits decrease) and the amount of banking stock must go up (since equity increases). Therefore, it implies a change in investment allocations of investors. Less cash, more equity. I.e. a transfer of risk from banks to the people. Banks become safer but investors hold more risky equity.

    11. Kurt,

      "Let’s say a mutual fund has a bond that the FED want to buy. The FED then creates reserves, they transfer these reserves to the commercial bank where the mutual fund has its checking account, and that bank in return creates deposits on the checking account of the fund."

      Why would a mutual fund ever willingly sell a government bond to the central bank? Commercial banks have their hands forced since they maintain access to the Fed's discount window. Mutual funds are not in the same position.

      Of course MBS purchases change the calculus. A mutual fund may want to offload the risk associated with mortgages and may be a willing seller to the central bank. But even in this case, a mutual fund would likely reinvest the proceeds of the sale rather than deposit the money at a commercial bank.


      "But even when the FED buys the bonds, commercial banks reduce their capital ratios!"

      See below for a description of Tier 1 capital:


      "Tier 1 capital, the more important of the two, consists largely of SHAREHOLDERS' EQUITY and DISCLOSED RESERVES. This is the amount paid up to originally purchase the stock (or shares) of the Bank (not the amount those shares are currently trading for on the stock exchange), retained profits subtracting accumulated losses, and other qualifiable Tier 1 capital securities."

      As long as the commercial bank holds onto those reserves (no lending from them), those reserves are treated as Tier 1 capital.

  3. John,


    Tier 1 capital, the more important of the two, consists largely of shareholders' equity and disclosed reserves. This is the amount paid up to originally purchase the stock (or shares) of the Bank (not the amount those shares are currently trading for on the stock exchange), retained profits subtracting accumulated losses, and other qualifiable Tier 1 capital securities.

    -- Notice, the purchase value of the equity is used, not the market value. A rally in the shares of a bank does not increase the capital adequacy of the bank. Likewise, a market correction in the the value of the outstanding shares of a bank does not decrease the capital adequacy of the bank.

    Tier 2 capital, or supplementary capital, comprises undisclosed reserves, revaluation reserves, general provisions, hybrid instruments and subordinated term debt.


    CAR = (Tier 1 Capital + Tier 2 Capital) / (Risk Weighted Assets)


    The original Basel III rule from 2010 required banks to fund themselves with 4.5% of common equity (up from 2% in Basel II) of risk-weighted assets (RWAs). Since 2015, a minimum Common Equity Tier 1 (CET1) ratio of 4.5% must be maintained at all times by the bank.

    The minimum Tier 1 capital increases from 4% in Basel II to 6%, applicable in 2015, over RWAs. This 6% is composed of 4.5% of CET1, plus an extra 1.5% of Additional Tier 1 (AT1).

    -- I don't know where these guys come up with their 7% Tier 1 CAR number - if they were indeed referring to only Tier 1 Capital (Equity, retained earnings, etc.).

    Basel III also introduced a minimum leverage ratio. This is a non-risk-based leverage ratio and is calculated by dividing Tier 1 capital by the bank's average total consolidated assets (sum of the exposures of all assets and non-balance sheet items). The banks are expected to maintain a leverage ratio in excess of 3% under Basel III.

    In July 2013, the U.S. Federal Reserve announced that the minimum Basel III leverage ratio would be 6% for eight Systemically important financial institution (SIFI) banks and 5% for their insured bank holding companies.

  4. John,

    How much of the rise seen in your M1 graph is correlated with a decrease in M3?

    The federal reserve stopped publishing M3 data in 2006, though there are sites that try to reconstruct it:


    I am not sure of the validity of the data being used, but there is a pronounced contraction in M3 from 2008 to 2010 that corresponds with the large increase in M1.


    1. M0 refers to the currency in circulation, such as coins and cash.

    2. M1 includes M0 plus demand deposits such as checking accounts as well as traveler's checks; M1 includes the currency that is out of circulation but is readily available.

    3. M2 includes all of M1 (and all of M0 as a result) plus savings deposits and certificates of deposit, which are less liquid than checking accounts.

    4. M3 includes all of M2 (and all of M1 and M0 as a result) but also adds the least liquid components of the money supply that are not in circulation, such as repurchase agreements that do not mature for days or weeks.

    1. Frank,

      Having read Hanke in the past, he is almost certainly referring to M3 and other broad money measures.

    2. Actually, Hanke is using a new, broader measure, called Divisia M4. It attempts to 'weight' different types of money;


      'Divisia M4. It is a broad aggregate, including negotiable money-market securities, such as commercial paper, negotiable CDs, and T-bills. M4's components are similar to those of the monetary aggregate once called L, but modernized to be consistent with current market realities.'

  5. I appreciate this post.

    I can now confidently say after 10 years of perusing econ-blogs about macroeconomics and monetary policy, and bank policy, I am confident of nothing.

    Oh, back in the 1970s I got an econ degree from a major university.

    I suspect Cochrane is right, but maybe he is more persuasive.

    But then, maybe requiring banks issue a fat layer of convertible bonds would be enough.

    So when a bank starts going down due to bad loans, it first wipes out shareholders (who have board seats, so in theory wanted the bank to be well-managed), then the convertible bond holders (who assume control of the board and do not want to lose their money, and have extracted conditions to but the bonds in the first place), and then the regular bond holders.

    After all that is burned through, I have no problem with the Fed buying the bank with printed money, taking the bad loans off the books, and re-selling it on the open market.

    The important thing is that the lending infrastructure stay in place and functioning.

    If shareholders, convertible bond holders and then bondholders are wiped out, there is no moral hazard issue.

    It is a bail out of the institution, but not of the investors.

  6. This WSJ piece is really dumb. Nice job of explaining why. And of course Modigliani Miller is an important rebuttal to the argument they did not make even as the 1st comment here tries to. But I'm disappointed that cutting the pizza pie into 4 slices does not help my diet. Funny!

  7. I think Cochrane has it right and tells us why politicians should not be involved in things they know very little or nothing.

    Higher capital requirements makes good sense to me. The bank is safer and therefore requires less regulation. It requires less FDIC insurance, maybe none. More folks want their money there if they are paying attention.

    Example: City National Bank of Florida, where I have banked since the crisis did not have a crisis. It had plenty of capital and was not involved in the subprime market. It made, for the most part, good loans which I suspect required a sane risk analysis. They were able to cover their loses.

    Conclusion: If banks know what they are doing,have their eye on the future and not quick profits and commissions, leave them the hell alone.

  8. 'Higher capital requirements makes good sense to me.'

    How do you know how high they should be?

    'Conclusion: If banks know what they are doing,have their eye on the future and not quick profits and commissions, leave them the hell alone.'

    Which was the problem, in the first place. The kinds of loans that blew up weren't being made by banks when they were left alone, prior to the early 1990s. It was the banking regulators (prodded by noisy 'community organizers') who had the bright idea to change the terms under which home loans were made, not the bankers.

    There's a very good explanation in Calomiris and Haber's 'Fragile By Design' of how it took place.

  9. I wander why no one discusses fractional reserve requirements and the problems it causes. Is it passe?

    1. Reserve requirements are about $50 billion. Reserves are about $3,000 billion. Reserve requirements are interesting economic history.

    2. This is similar to "one week of five years" to increase capital. RR are in big excess now may be not in a few years. Also, in the rest of the world it is a different story altogether

    3. John,

      Reserve requirements in the U. S. are pretty much relegated to the history books. But they are still practiced and enforced in a number of other countries.


      Some countries with relatively high reserve requirements;
      Brazil - 45%
      China - 17%
      Nigeria - 20%

    4. Frank, That Wiki site says the requirement for large banks in the US is 10%.

    5. Ralph,

      Read further:

      "In the United States, a reserve requirement (or liquidity ratio) is a minimum value, set by the Board of Governors of the Federal Reserve System...The deposit liability categories currently subject to reserve requirements are mainly checking accounts. There is no reserve requirement on savings accounts and time deposit accounts owned by individuals."

      Checking accounts comprise a very small portion of demand deposits in the U. S.

  10. Let's hope reserve requirements remain in history, and that talk of the 100% Chicago plan is never revived. Domestic US banks can get by with $3 billion in short-term US debt, financed by a decent equity cushion. If we made them hold much more (as a byproduct of capital requirements), there would be none left for the PBoC.

  11. Capital is a source of funds. Banks issue capital. In 2008, Citibank offered $4.5 billion in common. The hedge fund guys lined up and bought. Fannie,May,2008, offered 82,000,000 shares of common. 2008, GE offered $12B in common, Buffet was allowed to bid for up to $6B. When prices cratered in 2008, our firm bought the market with vols near 80. As for MM, we know prices are set in the market. A home traded at 500,000 can be financed with 20% equity, 80% debt or 100% equity. The price of the house is still 500,000.

  12. leverage adjustment - If financial markets are not perfectly liquid so that greater supply of the asset tends to put downward pressure on its price, then there is the potential for a feedback effect in which weaker balance sheets lead to greater sales of the asset, which depresses the asset’s price and lead to even weaker balance sheets.

  13. John, my recollection is that this is the Friedman and Schwartz story that the increase in required reserves in 1937 led to a contraction in loans and the sharp economic contraction in 1937 and 1938. I am not suggesting that episode resembles this one in any useful way.


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.