July 11, 2007
The eurozone is missing the point
By Paul De Grauwe
Since the creation of the eurozone in 1999, inflation has been very low (on average 2 per cent a year) and moved very little. The European Central Bank deserves much of the credit for this success.
The puzzling thing, however, is that the ECB claims to monitor closely the development of the money stock (in particular of M3, a broad measure of money) and says this monitoring exercise contributes to its success.
That cannot be true. Since the start of the eurozone, average yearly money growth (M3) has been close to 7 per cent while inflation was only 2 per cent, a huge gap. In addition, money growth has been subject to wild fluctuations. From less than 4 per cent in early 2001, it shot up to 8 per cent during much of 2001-03. Then it declined again to less than 5 per cent. In the past few months it has exceeded 10 per cent. During this whole period, inflation has been moving at a steady pace of 2 per cent a year, or a few decimal points lower or higher.
So there has been a great disconnection between inflation and money growth in the eurozone. Inflation has barely budged from its 2 per cent horizontal path while, according to traditional thinking, the “excessive liquidity creation” observed throughout the period should have led to much higher inflation. It did not.
The recent double-digit increases in M3 again prompted warnings that this will inevitably lead to inflation. Maybe. But up to now the facts show that money growth has been irrelevant to explaining inflation in the eurozone.
The remainder of this column can be read here (FT.com subscription required). Discussion from our guest economists is free.
Kris Vanderstede (guest): I do not agree with Professor De Grauwe’s explanation on the way hedge funds create liquidity.
Hedge funds buy assets form a small capital base, using to a large extent borrowed money. The capital base comes from “real money” investors, such as insurance companies and high net worth individuals. The borrowed money comes from banks. They lend money to the hedge funds against the assets of the hedge funds. The perceived quality of these assets determines how much money they are willing to lend. If the quality decreases, the banks lend less money and the liquidity in the system decreases. But in the end, it is still the banks that create liquidity.
Ironically, the assets that the hedge funds buy are in many cases assets that have been securitised by the banks in the first place. So one may wonder what the difference is between a bank holding assets on its balance sheet ; and a bank selling assets to a hedge fund, and then lending money to that same hedge fund to hold those same assets.
This raises the question what happens with liquidity if there is no longer a market price for the assets, because no-one wants to bid for them. This is the ultimate liquidity crunch. Banks call their money back, because they no longer want the assets as collateral ; but hedge funds cannot sell them, because they are offered-only. Not so different from a run on the bank, it seems to me.
What is happening now in the sub-prime sphere starts to feel like an illustration of this mechanism.
Kris Vanderstede is asset manager at Fortis Insurance Belgium
Posted by: Kris Vanderstede | July 11th, 2007 at 7:32 pm | Report this commentMartin Wolf: I think I agree with Mr Vanderstede. Ultimately, liquidity is created by banks, because their liabilities alone are also means of payment. On the assumption that the capital ratio of a hedge fund is lower than that required of banks by their regulators against holdings of risky assets, then hedge funds are essentially devices to allow banks to bypass the regulatory regime and increase leverage in the financial system. But the difference from a run on the bank in the crisis situation Mr Vanderstede describes is that nobody should care very much if hedge funds go bust, provided the banks do not do so, too. Of course, if the losses of hedge funds turn out to be big enough, banks might go bust, too.
Posted by: Martin Wolf | July 12th, 2007 at 12:29 pm | Report this commentPaul De Grauwe: These are very interesting and challenging comments. Let me try to respond. First, I want to eliminate a possible misunderstanding.
My argument is that hedge funds create liquidity because they borrow short and lend long. That does not mean (as pointed out by Martin Wolf) that they create means of payments, like demand deposits. Banks do that. That’s why I also wrote in my article that the liquidity created by hedge funds does not lead to higher prices in the grocery stores.
But there can be no doubt that hedge funds create liquidity in the sense that they transform assets, i.e. they transform short-term liabilities into longer term assets. As a result, the rest of the economy holds claims against the hedge funds that are more liquid than their debts to the hedge funds. The liquidity creation by hedge funds is in fact another way of saying that hedge funds are in the business of credit creation.
I think both Martin Wolf and Kris Vanderstede would have felt more comfortable if I had written that hedge funds create credit.
This leads me to answer Kris Vanderstede. The fact that the hedge funds borrow from banks (instead of directly from the public) is not relevant here. To the extent that their borrowing from the banks has a shorter maturity than their investments, one can conclude that hedge funds have created liquidity and expanded credit.
In this process, banks may or may not have increased liqudity. It will depend on whether their lending to hedge funds has a longer maturity than their liabilities.
One of my conclusions was that banks have been less in the business of liquidity creation lately and hedge funds more. This can also be seen from the fact that the growth rate of M1 (measuring traditional means of payments) has declined sharply during the last few years in the Eurozone. It now stands at about 6 per cent (remember the growth rate of M3 is now above 10 per cent).
Finally, Martin Wolf argues that “hedge funds are essentially devices to allow banks to bypass the regulatory regime and increase leverage in the financial system”. That may be true, but it also means that banks transfer part of their traditional activities of transforming assets (credit creation) to hedge funds. As a result, it is quite inadeqaute to only monitor the assets and liabilities of banks. Hedge funds must be brought into the monitoring process in order to have a correct view of liquidity and credit creation in the system.
Posted by: Paul De Grauwe | July 13th, 2007 at 8:59 am | Report this commentPatrick Minford: Paul’s column is a breath of fresh air after all the pontification we have had recently in FT Letters and from Martin’s domestic columns over the dangers of excess liquidity being signalled by broad money. What Paul talks about in the eurozone is also true of the UK. Financial innovation has indeed caused great instability in the relationships between various measures of ‘inside money’ and aggregate demand. As for measuring ‘liquidity’, in the style of the Radcliffe Committee, the prospects seem as
Paul implies even worse.
Where I part company from Paul is on ‘asset inflation’. Critical to his view is his espousal of behavioural finance as an explanation of asset price movements. But these movements can be well explained within rational agent models which include a rich enough account of uncertainty about the future of productivity change and other fundamentals. I am not sure that behavioural finance models can explain these movements without an unacceptably large menu of ad hoc assumptions that may have to vary from case to case.
In terms of policy for central banks to intervene in supposed bubbles is a dangerous practice as the collapse of the US economy in 1929 and of the
Posted by: Patrick Minford | July 15th, 2007 at 5:11 pm | Report this commentJapanese economy in 1989 illustrate, both of them after Bank policies targeting bubbles. We need much better models of bubbles from Paul and
others working on them before policy could be based on them.
Those economists who feel twinges of unease of varying degrees of intensity when observing the rapid growth of some broad monetary aggregate (M2 in the US, M3 in Euroland, M4 in the UK, M16 and AK47 elsewhere) are the victims of two fallacies: (1) the fallacy of misplaced concreteness and (2) the nominalist fallacy.
According to Wikipedia - a fount of highly doubtful authority - the fallacy of misplaced concreteness, as described by the philosopher Alfred North Whitehead, involves thinking something is a ‘concrete’ reality when in fact it is an abstract belief, opinion or concept about the way things are. The nominalist fallacy is the belief that naming something defines its essence, and explains it.
Every economist grows up learning and at times teaching models in which something labelled ‘M’ has a unique role in the process determining the general price level. Even the most New-Keynesian follower of fashion believes in the long-run neutrality of ‘M’.
Mapping that ‘M’ into some real-world counterpart is awkward. In fiat money economies, the monetary base - currency plus commercial bank deposits with the central bank - is probably the closest ‘real world’ counterpart.
The vast majority of analytical macroeconomic models ignore financial intermediation completely. The only financial assets are base money, government non-monetary debt and equity claims on real reproducible capital. There is also a complete contingent markets literature, which grafts base money onto an Arrow-Debreu complete markets model of a barter economy. This hybrid, although popular, is absolutely useless as a starting point for the analysis of monetary policy in the real world. While it has an exhaustively rich financial asset menu, there are no financial intermediaries and there is no financial intermediation in any meaningful sense.
Where banks are considered, they tend to be a poorly motivated constraint on financial intermediation rather than a means for overcoming informational, monitoring, commitment and enforcement constraints. The same holds a-fortiori for non-bank financial intermediaries.
No macroeconomic model I have ever seen creates a special role for the sight-liabilities of deposit-taking institutions (and their close substitutes) that make up most of M2, M3 and M4 in the real world. Even more emphatically, there is no macroeconomic model that makes the case that that the real world M2, M3 and M4 play a role analogous to ‘M’ in the baby models we play with.
My argument is quite different from the proposition that, if the monetary authorities use the short risk-free nominal interest rate as the instrument of monetary policy rather than the quantity of ‘M’, and if as a result of this, ‘M’ becomes endogenous, then all equilibrium prices and quantities that can be determined in the model can be determined without reference to the behaviour of ‘M’. This proposition goes through only if the economy is ‘block recursive’ in ‘M’, but this is the case for most of the models that litter the literature. When the interest rate is the monetary instrument, ‘M’ is all tail and no dog. However, in such models, the behaviour of ‘M’ over time itself may still be determinate (it will be so unless we are in a world with perfect price flexibility, in which case only the real value of ‘M’ is determinate). Ruling out this uninteresting special case, there would be a systematic relationship between the general price level and the endogeous stock of ‘M’. If in such a world prices had been rising for a long time by 2 percent per annum, real GDP by 3 percent per annum and ‘M’ by 14 percent per annum, we would all be scratching our heads.
The world we appear to be living in is much more complicated than that. Not only is ‘M’ endogenous, it is unobservable, and is represented empirically neither by the broad monetary aggregates M2, M3 and M4 nor necessarily by the real-world monetary base or M0.
We are clearly moving towards a world where the currency component of M0 will be displaced by private cash-on-a-chip and other networked or anonymous payment mechanisms(except for use by the criminal community and by foreign demand from underdeveloped countries with high inflation). The bankers’ balances with the central bank component of M0 could easily be replaced by overdraft facilities or contingent credit lines with the central bank. Functionally M0 will survive, but its conventionally measured quantity could easily vanish.
The non-M0 component of M2, M3, M4 etc. is part of portfolio demand, boosted by expanding and deepening intermediation. It has no close connection with any transactions demand, that is, with imminent spending on currently produced goods and services. It is not at all surprising that it has grown in line with household financial wealth rather than with GDP.
Clearly, the growth of credit to the ‘ultimate spending units’ - households and non-financial enterprises - must be tracked carefully, but even there, we must be aware of the fact that households are growing their net financial wealth -to-income ratios and that gross assets and liabilities may be growing even faster than net financial wealth.
I am a monetarist, through and through. Unfortunately, I don’t have a clue as to what the real-world counterpart of the textbook ‘M’ is. It seems pretty clear though, that whatever it is, it’s not M2, M3 or M4. The ECB, to the extent that it pays special attention to a subset of the liabilities of certain deposit-taking institutions, is barking up the wrong tree.
To the extent that the problem is caused by the nominalist fallacy, part of the solution is to stop referring to M2, M3, M4 etc as ‘money’ or ‘monetary aggregates’. Let’s call them L2, L3 and L4 instead, and stop worrying about them.
Posted by: Willem H. Buiter | July 16th, 2007 at 12:31 am | Report this commentTwo more short points. Paul first calls issuing short-term liabilities to fund the acquisition of long-term assets ‘liquidity creation’. I assume liquidity refers to the the ease, speed and cost (including effect on the market price) which which a financial or real claim can be sold/realised. However, the maturity of a claim need bear no relation to its liquidity. A newly issued 10-year US government TB is more liquid than my 1-month maturity loan to my 16-year old son, because there is a deep liquid secondary market for the former but not for the latter. Maturity transformation is therefore neither necessary nor sufficient for liquidity creation.
Paul then says it may be more accurate to refer to the maturity transformation/liquidity creation by hedgefunds as credit creation. That cannot be right either. You create credit if you run a net financial surplus. Mucking about with the asset and liability composition of the balance sheet, and/or expanding the gross size of the balance sheet, as hedgefunds have done on a significant scale, can be done without running a financial surplus or deficit.
If hedgefunds don’t create either liquidity or credit, what do they do?
Posted by: Willem H. Buiter | July 16th, 2007 at 11:22 am | Report this commentMartin Wolf: This was a fascinating, albeit deeply technical, discussion. I was disappointed that Paul has not (yet) responded to Willem and Patrick.
I would like to make a few comments, recognising that I am certainly not up to debating contemporary monetary economics with Patrick, Paul or Willem.
First, we really do know remarkably little about how our economies work, including how intermediation affects them. So, it seems to me, we need to accept that rules of thumb - lessons of experience - matter. So there is an empirical question: do the various M measures help tell us what is going to happen to spending, or not? If the answer is yes, we should pay attention to them. If it is no, we should not.
Second, Patrick wants to stick with the “rational agent” model, with a rich enough account of uncertainty. I have no doubt that models of this kind can be saved in this way. But how different are they then from the behavioural models he despises? With sufficient uncertainty, how can you know whether behaviour is rational, or not? Obviously, one can find some assumptions that make it so. In the same way, Ptolemy found sufficient epicycles to preserve circular orbits for planets. But what is the point?
Third, Patrick repeats his view that it is wrong to prick bubbles. Possibly. But one can just as convincingly argue that it is wrong to let them emerge, in the first place. In the two cases he mentions the mistake was surely the failure to act decisively after the bubble burst (which would surely have happened, anyway).
Fourth, are the broader measures of M worthless, as Patrick and Willem agree? I have a relatively simple-minded view on this. I think of money as an asset that can be converted into means of payment both quickly and at par. In other words, it has a fixed nominal value. Broad money has this characteristic. Now what happens if people find themselves holding more of this stuff than they want (perhaps because banks created too much credit and so “money”, on which more below). Then they must spend it either on assets (whose prices rise) or goods and services. Alternatively, they may be persuaded to hold it by higher interest rates. All these alternatives matter for monetary policy. So it seems to me sensible for central banks to pay more attention to money than to, say, long-term bonds. If people do not want to hold the latter, their prices will fall, which will solve that problem pretty quickly.
Finally, I am struck by disagreement on what terms mean. Of course, it does not matter what they mean so long as we understand what we mean (and so do readers).
Liquidity, for example, seems to me to refer not just to the ability to buy and sell assets easily, but to the properties of certain assets themselves. Bank deposits are highly liquid, because they can be converted into means of payment (or are means of payment) at once and have (as stated above) a predictable nominal value. In this case, the maturity of an asset is an aspect of its liquidity, so defined. Of course, there are highly illiquid short-term assets (as Willem points out). But, for any given issuer, short-term liabilities are more “liquid”, in the sense generally used, than longer-term ones.
Another word whose meaning seems to be unclear in this discussion is “credit”. The grant of credit means turning something illiquid (a promise to repay, for example, or, maybe, an old watch) into today’s purchasing power. Some institutions offer credit. But other institutions are able to create it, because their liabilities are universally accepted means of payments (i.e. banks). I do not see how hedge funds create credit, though they presumably do offer it, as Paul argues. But they do surely make markets more liquid, in Willem’s sense. They do this by increasing the number of players on both sides of markets.
Posted by: Martin Wolf | July 30th, 2007 at 7:04 pm | Report this commentMartin Wolf: I would like to add a point to the definition of “credit”. Willem has defined credit as the consequence of running a financial surplus. This is a perfectly reasonable definition, in terms of the flow of funds. Under this definition, if households spend less than they earn and buy corporate equity with the difference, they are creating credit. But my definition was in terms of balance sheets in the economy and particularly the balance sheets of banks. If someone receives a loan from a bank in return for a promise to repay at some point in the future, credit is created (and so, in my view, is money). Again, I think this is a perfectly reasonable definition, just different from Willem’s.
One of the lessons I have learned from this discussion and Paul’s column is that it would be really helpful if we tried to define much more precisely what we mean by “liquidity”, “money” and “credit”. It seems that Paul and Willem mean different things and so, again, do I.
Posted by: Martin Wolf | July 31st, 2007 at 12:46 pm | Report this comment