Wednesday, 27 June 2012

The money machine

The financial system is short of liquidity.

"What???" you say. "Despite the trillions of dollars that the Fed, the Bank of England, the Bank of Japan and even the ECB have been pumping into it through their various varieties of money creation? How can the financial system POSSIBLY be short of liquidity?"

Trust me, it is. But the conventional banking system isn't short of liquidity. The shadow banking system is. And this is a very serious matter, because the shadow banking system and conventional banking system are critically interlinked. If money stops circulating in the shadow banking system, it can't circulate in the conventional banking system either. The effect is that conventional banks end up awash with cash that they daren't lend out, and the shadow banking network grinds to a halt. Which is what is happening.

So why isn't money circulating in the shadow banking system, and why does this affect the transmission of money in the conventional banking system?

Firstly, let's explain what the shadow banking system is. Essentially it is a network of non-bank financial institutions that together perform functions akin to those of conventional banking. The three principal activities that ensure the flow of funds through the shadow banking system are repo (repurchase agreements), securities lending and rehypothecation. These three also create the links between conventional and shadow banking and are the means by which money flows through the financial system as a whole.

Let's imagine a large bank funding itself in the interbank repo market. A repo consists of two transactions - a spot sale of assets and a future purchase of the same quantity and quality of assets. The combination of these two transactions is effectively the same as a short-term secured loan and is normally recorded as such in the books of the parties concerned, although there have been notorious examples of repos being recorded as sales (Lehman 105, MF Global). The collateral on this secured loan is the assets sold, which have to be of sufficiently good quality to be acceptable to the lender. Note that as these assets are collateral - i.e. do not belong to the holder unless the owner defaults - they do not form part of the asset base of the holder.

Prior to the 2008 financial crisis, the US repo market ran largely on mortgage-backed securities and their derivatives, most of which were assigned ratings of AAA or similar. Those securities are now regarded as toxic, so since then, markets have turned to other forms of "safe" assets - particularly the bonds of highly-rated sovereigns. Regulation on both sides of the Atlantic has encouraged this by requiring large banks to hold substantial quantities of safe liquid assets (usually their own government's debt) as a liquidity buffer. If a large bank doesn't have sufficient acceptable collateral, it is not able to borrow in the repo market.

However, a large bank that is short of collateral does have a saviour at hand - funds. Hedge funds, pension funds, other kinds of investment funds. These funds hold safe long-term assets such as government bonds. But they like to get a short-term return on these assets, so for a fee they will lend them out. So the bank can borrow securities from a fund and pledge them as collateral in a repo transaction. This would typically be done as a reverse repo/repo pair: the fund pledges securities to the bank and the bank re-pledges those securities in the repo market. This practice is called rehypothecation. The amount borrowed is less than the market value of the securities: the difference is called the haircut. Haircuts get bigger each time the securities are rehypothecated. The difference in the haircut on each rehypothecation is equivalent to the spread between borrowing and lending and represents a real return to the lender.

People familiar with conventional banking will recognise this transaction series as credit intermediation and maturity transformation - borrowing short and lending long - which is the traditional business of fractional reserve banking. Both fund and bank have lent out long-dated securities in return for short-term cash funding. And further, credit creation is involved. The bank pays the fund cash it doesn't yet have and settles it when it receives the cash it borrows in the repo market - exactly as conventional banks do when they lend money ahead of obtaining settlement funding. The accounting for this payment creates a deposit in exactly the same way as conventional fractional reserve banking and therefore expands broad money just as normal bank lending does. Much of the expansion of M4 (broad money) prior to the 2008 financial crisis occurred in this way.

So this is "shadow" banking.  It's just like conventional banking except in two critical respects: there is no deposit insurance and usually no lender of last resort (shadow financial institutions typically have no access to central bank funding). And it is largely unregulated. Hence the requirement for safe collateral. Really, it's financial pawnbroking.

And conventional banks depend on it. The deposit base of conventional banks has shrunk over the last few decades as people have invested their savings in pensions, long-term savings schemes linked to insurance, financial assets and property. All of those except the last are providers of funds and collateral to the shadow banking system: the last has until recently been an important source of "risk-free" assets. The shadow banking network is the means by which the funding that previously came mainly from bank deposits finds its way back into conventional banks to enable them to perform their lending function: the mechanism by which this happens is the interbank market, which conventional banks use to fund deposit drawdowns and other outflows. 

So the whole financial system, including conventional banking, depends on there being a plentiful supply of safe assets for use as collateral. But just look at what has been happening to sovereign debt! Sovereigns are losing their AAA ratings: even the mighty USA has been downgraded, although as the US Treasury is still the world's most liquid financial instrument by a country mile the downgrade hasn't made much difference. And around the world, sovereigns are finding ways of removing sovereign debt from circulation, either through fiscal consolidation or by buying it back.  The Fed and the Bank of England have both had two rounds of Quantitative Easing (QE) in which they have removed from circulation large quantities of the safest and most liquid government debt in the world. As a result there is now a serious shortage of AAA-rated assets in the shadow banking system. This is of course partially offset by the large amount of cash held by banks - but as they will only lend to safe prospects or in return for good collateral, the cash is not doing much. Lots of it is parked at central banks for a few basis points in interest.

I am amazed that politicians are arguing about why the yields on UK gilts are at all-time lows. The Government insist it is because their fiscal consolidation programme is increasing market confidence, while the Opposition claim it is because markets don't think the UK will come out of recession any time soon. I don't believe either story. To me, one look at the state of the global money machine is enough to explain why UK gilt yields are below inflation. Demand for these is outstripping supply, so their price is rising (which causes yields to fall). The same is happening to other high-quality sovereign debt.

The financial system is short of liquidity. But not cash. It has lots of that, and the central banks keep providing more and more - while removing from circulation the liquidity that is really needed, namely government debt. It's like a car owner who keeps putting petrol in his car to keep it running when the problem is an oil leak. Eventually his engine will seize up and the car will stop running even though the petrol tank is full to overflowing.

I have some sympathy (well, not much really) for the central banks' predicament. Governments aren't usually too happy for central banks to dish out cash willy-nilly, and they want to know that central bank management are being prudent, because central bank losses do rebound to the State. So central banks generally will only lend cash to banks that can provide decent collateral. But the same governments that don't like central banks to accept risky collateral (thereby keeping safe assets in circulation) also don't want to issue new debt to replace that being removed from circulation. Fiscal consolidation is fashionable and EVERYONE is trying to cut their deficits and run balanced budgets. And those same governments, in the interests of making the financial system safer and protecting small investors, also insist that funds and banks hold increasing quantities of their bonds, while reducing their new debt issuance and encouraging central banks to remove existing bonds from circulation. No wonder UK and US government debt yields are crashing through the floor.

The financial system is at serious risk of seizing up completely due to lack of collateral. There are a number of possible solutions to this. One is to accept that cash is the future and guarantee all cash deposits, including wholesale and interbank deposits. This would remove the need for safe assets as collateral - the repo market could become an unsecured funding market. An interesting side effect of this would be to make it much more difficult for funds to lend out their assets to gain short-term returns, and they might start acting more as custodians.

A second would be for regulators and investors to reconsider what is meant by "safe" assets. A quick look at this chart shows that the Basel assumption that sovereign debt is intrinsically safe is clearly very wrong - just as prior to 2008 the habitual classification of mortgage-backed securities as "safe" turned out to be very wrong. No sane person would regard corporate bonds issued by Apple as riskier than, say, Brazilian sovereign debt - yet that is the assumption. The asset bases of multinational companies are larger and more secure than most countries, and many of them have been in existence for longer, too. The direction of regulation at the moment is to push investors towards government debt. Perhaps that should change, in the interests both of protecting investors and providing more liquidity to the shadow banking system. Goldbugs reading this post will no doubt suggest gold as an alternative safe asset, but there is an even greater scarcity of gold than there is of government debt - and guess who has been buying it? Yup, central banks. Anyone would think they WANT this system to seize up!

The third, and most radical, solution would be to reconsider the nature and purpose of government debt. I've mentioned previously that in financial terms there is no significant difference between government bonds and currency. Government bonds are usually callable term-limited, interest bearing bonds, though they can be perpetual (no maturity date) and they can be zero-coupon (non-interest bearing). All fiat currency (notes and coins) is callable zero-coupon perpetual Government debt. In the shadow banking system, both currency and Government debt are forms of money, and the shadow banking system acts as a sort of money exchange system, intermediating government debt into currency and back again. All that conventional QE does is expand one form of money at the expense of the other, and by so doing it makes the shadow banking system's currency intermediation function  extremely difficult. We should be looking for the balance between the two forms of money that enables the money transmission and intermediation machine to function most efficiently. But to do this, we have to stop thinking of government debt as something that governments need to keep society working, and redefine it as something that banks and markets need to keep the money machine working. 

Therefore I fundamentally disagree with all those who believe that governments should stop issuing debt, whether because they think governments can simply create the money they need or because they think governments should meet spending commitments only from tax receipts. Debt issuance isn't about government spending: MMT theorists are correct that a fiat currency-issuing sovereign can simply create the money it needs and doesn't need to borrow from investors (though fiat currency issuance is of course borrowing at zero interest, as I explained in the paragraph above). No, government debt is a GOOD for which there is a worldwide demand. We should produce as much of it as the world demands. It is one of the most productive assets of our economy.

Now, I can just hear the howls of outrage at the idea that the government should issue MORE debt in order to help the financial system. After all, we pay interest on this debt, don't we? Well, at the moment, no we don't. The amount we are paying in interest on government debt is below inflation - and other trusted sovereigns, notably the US and Germany, actually have negative yields at the moment, which means that investors in those bonds are making a loss. Yields so low that the real rate of return is below zero are in my view a clear indication that there is not enough of the asset in circulation to meet demand. So there is definitely scope for government to issue more debt without incurring real cost.

I propose that the right amount of government debt for a trusted sovereign to issue is the amount that maintains the real rate of return at the same rate as the rate paid by the central bank on cash deposits - currently about 25 basis points, I think. Government spending needs should have no influence on this basic principle. Government then will be indifferent as to whether it meets its spending commitments through debt issuance or currency creation, because the real cost will be equivalent: if it chooses to issue more debt, it will pay interest above inflation; if it chooses to issue currency, it should experience reduction in the purchasing power of the currency equal to the interest cost on the equivalent debt issuance.

I am well aware that this is a controversial proposal. But I have approached it from a purely monetary perspective, without getting into arguments about the morality of government debt and spending. What government should do with the money it receives from the sales of its quality product, obviously, is to spend it on things that improve the quality of its product still further. What exactly those things are is an entirely separate economic and political issue that I do not intend to address in this post.

The financial system is short of liquidity. Let's create some.


  1. Is there anything in this arrangement that allows newly created money to be placed in the hands of people who will be free to spend it? By which I mean "the populace." Greece would be a good example at the moment, but everywhere will do! The lending economy is all ying and no yang.

    1. As I said at the end of the post, exactly what is done with the money received is an economic and political issue that is beyond the scope of this post. Personally I think the money should be spent relieving personal debt - as I've said elsewhere. But there are many other calls on the money.

    2. Which is exactly what the likes of Steve Keen and Michael Hudson have been arguing. That private debts need to be reduced to the capacity to pay.

      I know this post is old, but yields on sovereign debt hasn't really climbed much. In some places at least. So would you say this post is still relevant?

  2. Thank you, very interesting.

  3. Very interesting! Controversial perhaps, but that is a virtue, especially as conventional economic policies around the world continue to fail.

    Keynes was controversial in his time. It took him several years to persuade the UK Treasury of the correctness of its arguments. His ideas were finally accepted, implemented and shown to work.

    Great stuff!

  4. Related matter? Banks Everywhere Are Passing The Trash—Again

  5. The Financial times issued an article this morning in the european edition on the same subject, with less facts, less spirit and later than you did.

    You beat them. Easy...

  6. I agree that “there is no significant difference between government bonds and currency.” I also agree that the return on that debt should be about zero (you say 25 basis points).

    Re what the optimum amount of public debt is, I think Modern Monetary Theory has got this right. MMTers, of which I’m one, tend to call the combination “currency and debt” by the name: “private sector net financial assets”. And the optimum amount of PSNFA is the amount that induces or helps induce the private sector to spend at a rate that brings full employment. (Not that that is the only factor influencing aggregate demand, and hence employment levels.)

    Whether banks have a liquidity problem or not is much less important. Given the profligate lending that was taking place before the crunch, banks may well down size their operations over the next few years. That is, firms may in future rely more on equity and less on bank loans. Or they may borrow less by going for less capital intensive methods of production, or they may rely more on good old consumer demand than on bank loans. That may be difficult and awkward for banks, but I’m not too concerned.

    1. Ralph,

      This is a short-term problem, not a long-term one. The "profligate lending" prior to the 2008 crisis is now a thing of the past and if anything we now don't have enough lending. Yet the collateral shortage threatens to bring what lending there is to a halt. My post was concerned with short-term measures to deal with the shortage of collateral in the financial system.

      Unless you plan either to guarantee unsecured deposits (plan a) so that banks don't need collateral, or to allow the central bank to lend unsecured funds, the need for safe collateral as liquidity in the banking system is not going to go away.

    2. Given that we’re agreed there is little difference between government debt and monetary base, I thought you’d be amused by this article by Milton Friedman which claims that QE would have done wonders for Japan – see section half way down entitled “Increase the money supply”:

      I think he is talking thru his hat.

    3. Particularly ironic, as the Bank of Japan then did QE for the whole of the next decade to no effect.

      Friedman's whole argument is based on the idea that if you give banks money they will automatically lend it out. That is fundamentally wrong, as you know. Bank lending decisions are entirely commercial, driven by the expected return in relation to risk, and have nothing whatsoever to do with the availability of reserves. If banks don't want to lend, they won't. And in Japan they didn't want to, because their balance sheets were already stuffed full of non-performing loans. They were zombies, and all the QE did was keep them alive but unproductive.

      Friedman's ideas have dominated much economic thinking in the last half-century, but as far as I can see they are based on a false premise. He has much to answer for.

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

  8. The decisions taken in case of lending are entirely commercial, and have nothing to do with the availability of reserves.

  9. Nice blog to be honest and would like to share this blog. you can also refer this Press Components Hyderabad