Monday, 19 July 2010

Keynsian “borrow and spend” is crazy.

Definitions. 1. The word “borrow” is used here to refer to where a government borrows in the market, as distinct from borrowing from its own central bank. The latter operation typically involves the creation of bonds, payments of interest, etc. But this is just a paper and money shuffling exercise within the “government – central bank machine”. It has no effect on the real economy (assuming government bonds are not sold in the market by the central bank).

2. Having referred above to two separate entities, government and central bank, the word government will henceforth be used to refer to the two entities combined.

3. The word “print” as in “print money” does not refer simply to printing banknotes: it refers to creating money in the broadest sense of the phrase, e.g. creating money by pressing computer keyboards, which in practice is how banks create most of the money they create nowadays.

4. The phrase “Keynsian borrow and spend” is unfair to Keynes in that towards the end of his life he turned against borrow and spend and in favour of having government simply print extra money and spend it, in a recession. However, since his death, plenty of economists who call themselves Keynsians have stuck to the “borrow and spend” option (for reasons that baffle me). So to that extent, the phrase “Keynesian borrow and spend” is fair.


When government borrows, it borrows something, i.e. money, which government itself created in the first place. It borrows something which government can create in limitless quantities any time. Thus borrow and spend is as absurd as a dairy farmer buying milk in a shop when there is a thousand gallon tank of milk a few yards from his front door. It is absurd as carrying coals to Newcastle.

As to whether printing money is inflationary, it is not inflationary (as pointed out by David Hume over two hundred years ago) until it is spent, and spent in sufficient quantities.* The latter equals stimulus (or possibly too much stimulus). Moreover “print money in the right quantities” equals “stimulus of exactly the right amount”. Ergo stimulus can be achieved without any Keynsian “borrow and spend”, that is, without any borrowing.

Wright Patman (Chair of the House Committee on Banking and Currency for 40 years) took a similarly dim view of borrow and spend. He said that when government “pays interest for the use of its own money” government has got involved in an “idiotic system”.

(Congressional Record of the House of Representatives (pages 7582-7583), Sept 29, 1941. A fuller quote is here.)

A second farcical aspect of Keynsian borrow and spend is thus. The latter borrowing is bound to push up interest rates, which crowds out private sector borrowing and spending. The latter is of course exactly what is not needed. So what do governments do to make sure this crowding out does not occur? They make sure that interest rates DO NOT rise! Indeed, since cutting interest rates is allegedly stimulatory, governments are very likely to CUT interest rates.

And how do they cut interest rates? Well they wade into the government bond market and buy back debt (with money created out of thin air), which raises the price of government debt, which in turn cuts interest rates.

So what is really doing the work here? That is, what is that has the stimulatory effect: the “borrow and spend” or the money printing?

Indeed, does it really matter which of the two latter has the real effect? The answer is “not really”. That is, whether an economy is stimulated by money printing or by borrow and spend does not make a scrap of difference in that money has to be printed in both cases! So why bother with “borrow and spend”? Answer: there is NO point! Borrow and spend is a farce.

A third reason for thinking that borrow and spend is crazy is that it involves governments in expanding the national debt and paying interest on that debt when there is no need to pay interest. Reasons are thus.

One of the reasons for unemployment is “paradox of thrift”: the belief by private sector entities that they hold insufficient financial assets, and thus that they need to save money. This saving, or “failure to spend” results in less aggregate demand than would otherwise be the case, which in turn means unemployment.

The obvious solution is to feed additional assets to the private sector. In the case of borrow and spend, this “feeding of additional assets” comes about as follows. 1, government borrows $X, 2, government gives $X of bonds to lenders, and 3, government spends $X. The net result is that the private sector is up to the tune of $X.

But for a government to borrow and pay interest on sums borrowed in this scenario is absurd because the private sector actually WANTS or NEEDS additional assets, and there is no need to pay anyone to take something they WANT or NEED.

Conclusion: so far as stimulus goes, it is CASH that ought to fed into private sector pockets, not bonds.

A fourth daft aspect of borrow and spend is thus. Borrowing is deflationary because it withdraws money from the economy. But that is in direct conflict with the purpose of borrow and spend which is REFLATIONARY or STIMULATORY.

In other words the borrow part of borrow and spend is a bit like throwing dirt over your car before cleaning it! Totally pointless!

In this connection, there is a possible objection to print and spend, namely that it is almost certainly a more potent weapon, dollar for dollar, than borrow and spend (because as just pointed out, much of the “spend” part of “borrow and spend” is cancelled out by the “borrow” part). That might induce some readers to conclude that print and spend is more inflationary, dollar for dollar.

Now there is a simple solution to this problem as follows. If you use a more potent fuel in an engine, but want constant power output from the engine, what do you do? The average ten year old knows the answer: use less fuel!

A fifth daft aspect of borrow and spend is that it expands the national debt and the larger the national debt, the more of such debt is likely to end up in the hands for foreign entities. Now borrowing from abroad CAN sometimes make sense. But paying interest to foreign lenders when (as pointed out above) no interest should be paid at all, is plain daft.

Conclusion so far: for stimulus purposes, borrow and spend is unnecessarily complicated. That is, simply printing money and spending it would make more sense, as pointed out by Abba Lerner, arguably the founding father of Modern Monetary Theory. And for those impressed by Nobel Prize winning economists, at least two agreed with the above money printing idea: Milton Friedman and William Vickrey. (The Friedman paper is also available here.)

A counter argument.

It could be argued that the costs of borrow and spend are not all that great, in that while this policy IS a pointless paper chase, the costs of paper pushing as a proportion of GDP are small. (Not a strong argument in view of the astronomic costs of the average country’s bureaucracy, but never mind!)

However, against that, there is a real and more serious problem as follows. An ever expanding national debt, or a national debt that expands too fast, causes economic illiterates to go berserk, and campaign for cuts in government spending (or tax increases). The two latter DO HAVE serious economic consequences: the result is a decline in demand in real terms, and means unnecessary unemployment. And this is a very real problem in the U.S. at the time of writing.

Another possible criticism.

A popular idea, which is also a possible criticism of the anti-borrow and spend argument here is that borrow and spend “gets an economy going”, and that come the recovery, a government can pay back the sums borrowed.

The problem with this argument is that if borrow and spend provides stimulus, then the most plausible prima facie assumption regarding the opposite of borrow and spend (i.e. repaying debt) is that the effect is ANTI-stimulatory, i.e. deflationary. And that in turn leads to the conclusion that debt should NOT be repaid (or at least not repaid till it looks like deflationary measures are called for, e.g. because of looming inflation).

And indeed, this “debt is never repaid” scenario is very roughly the reality for most countries. That is, while national debt as a proportion of GDP varies widely between different countries and as between different decades for any country you like to mention, most countries debt as a proportion of GDP is in the 30% - 100% range, just as it was a century ago.

Borrowing as a substitute for tax.

Government borrowing for stimulus purposes is quite different from government borrowing as a substitute for tax. And the above anti-borrowing arguments are not in themselves an argument against borrowing as a substitute for tax. (Though as it happens the arguments for the latter borrowing are pretty feeble. See here.)

Nor are the above arguments about the lack of any necessity for governments to pay interest in the case of stimulus borrowing applicable to “substitute for tax” borrowing. Indeed, given adequate PNFA and given full employment, a government which wants to borrow as a substitute for tax will HAVE to induce the private sector hold MORE than its desired PNFA. And inducements cost. In this case the cost is called “interest”.

Using interest rates to regulate economies.

A possible objection to the above arguments is that they imply an abolition or near abolition of government borrowing. And since governments adjust interest rates by buying or selling government securities, the above print and spend policy would make it difficult for governments to regulate interest rates.

To be more accurate, the fact of not engaging in borrow and spend for stimulus purposes does not rule out borrowing as an alternative to tax. But when politicians borrow rather than raise taxes, the motive is usually political cowardice, or (much the same thing) the desire to buy votes. So there is a good argument for abolishing or reducing government borrowing here as well.

So let’s assume a near ideal world in which politicians abstain from borrowing in order to buy votes: that is, let us assume there is no government borrowing either for stimulus purposes for “substitute for tax” purposes. Does the resulting paucity of government securities matter?

There several reasons for thinking it would not matter.

First, using interest rates to adjust demand is distortionary, since it works only via private sector entities that are significantly reliant on variable rate loans. It is true that less or more activity by these entities ultimately affects other entities. Given for example an interest rate cut, it is true that additional activity by those entities will sooner or later trickle down to other entities. But that is far from ideal.

In particular, by the time the trickling down is complete, it is possible the economy is suffering excess demand and inflation, and stimulus from the latter “trickle down” is exactly what is not needed.

The ideal is to induce as many private sector entities into additional activity as soon as possible, and print and spend is a superior measure in this respect to altering interest rates: at least, a well designed print and spend policy would be less distortionary. For example, a payroll tax cut would benefit every employer and employee in the country. That of course leaves out pensioners and others on social security. But with a little ingenuity the latter could be catered for.

Second, the fact of not borrowing for stimulus or “substitute for tax” purposes does not rule out borrowing specifically so as to influence interest rates. Indeed, where a government wanted for example to damp down demand by raising interest rates, the effect would come not just from the increased rates. Such a government would announce a willingness to borrow at a higher rate than the prevailing rate. That in turn would withdraw funds from the economy, which (as pointed out under the second objection to borrow and spend above) has a deflationary effect. At least that would be the effect, assuming the money borrowed is not spent.

The conclusion is that print and spend is a superior policy to borrow and spend, plus any shortage of government securities resulting from a print and spend policy would not make it significantly more difficult for governments to raise interest rates, and even if it did make it more difficult, that would not matter, because governments can always regulate by altering their net spending.

Conclusion: Keynsian borrow and spend is crazy.

Endnotes:1. it is quite likely that Keynes realised that Keynsian borrow and spend is crazy, to judge by what he said about Abba Lerner. (Lerner favoured “print and spend” rather than “borrow and spend” money”.) Keynes said "Lerner's argument is impeccable but heaven help anyone who tries to put it across to the plain man at this stage of the evolution of our ideas.”

In other words possibly Keynes realised that “print and spend” was beyond the comprehension of those surrounding him, and that he’d be better off going for the second best and daft alternative, namely borrow and spend.

Also Keynes in a letter to Roosevelt in 1933 said that government should “create additional current incomes through the expenditure of borrowed or printed money.” Thus it seems Keynes was well aware of “print” as an alternative to “borrow”.

Incidentally, Keynes in this letter also makes the point (mentioned above) that the quantity of money is irrelevant: rather it’s the rate at which it is spent that is crucial.

2.Note, dated 17th Aug. 2010. This post has been regularly added to and revised since it was first published on 19th July. Also I have expanded the above points into something more like a formal academic paper. See here.


* See David Hume’s essay “Of Money”, in particular the two paras starting “It seems a maxim….”

No comments:

Post a Comment

Post a comment.