Wednesday, 19 October 2016
Guardian article trotts out the old multiplier myth.
The article is by Stephen Koukoulas who according to article is “a Research Fellow at Per Capita, a progressive think tank.” Gosh. He’s “progressive” is he? That’ll impress the rather large proportion of Guardian readers who are cuckolds. (Article title is: "Economic growth more likely...")
And it’s not just me who thinks that self-styled “progressives” are often not all that bright: Bill Mitchell (left of centre Australian economics prof) has made that point over and over.
Anyway, the “multiplier” is simply the idea, widely accepted in economics, that the ULTIMATE effect on GDP of different types of spending varies widely. And the particular instance of that phenomenon that Koukoulas deals with is the different effect on GDP of donating money to the rich as compared to donating it to the poor.
As Koukoulas rightly points out, the poor spend a larger proportion of any increase in income than the rich, thus the ultimate effect on GDP of giving $X to the poor is larger than the effect of giving $X to the rich. Thus it seems that we get better value for money from giving money to the poor.
Unfortunately that argument does not stand inspection and for reasons set out below. I’ve actually set out the flaws in the multiplier before on this blog, but to get any point across one normally has to repeat it ad nausiam, so here goes – for the umpteenth time.
First, to attack the multiplier IS NOT to attack equality enhancing measures Koukoulas refers to. I.e. if it can be shown that higher taxes on the rich and more generous benefits or lower taxes for the poor brings big social benefits, no one can object to that.
But that’s quite separate from the strictly economic claim, namely that there’s a case for giving preference to high multiplier types of spending because the ultimate effect on GDP is higher per dollar of expenditure.
The basic flaw in the latter idea is that stimulus dollars cost nothing in real terms. To illustrate, if stimulus consists of helicopter drops (i.e. literally printing $100 bills and giving a bundle of those bills to every household in the country), the real cost of doing that is minute compared to the face value of those dollars.
As Milton Friedman put it, "It need cost society essentially nothing in real resources to provide the individual with the current services of an additional dollar in cash balances."
Of course stimulus does not normally consist of overt helicopter drops, but conventional stimulus actually comes to almost the same thing. Indeed, conventional fiscal stimulus (government borrows $X, spends $X and gives $X of bonds to borrowers) followed by the central bank printing money and buying back some of those bonds so as to keep interest rates stable or even reduce them, comes to EXACTLY the same thing as a helicopter drop (assuming that by “helicopter drop” one means dropping money onto government departments, state schools, etc as well as households).
Thus given a number of different ways of getting the economy up to capacity, A, B, C etc, the only important point (as far as economic rather than social considerations are concerned) is the effect on GDP. The number of stimulus dollars needed to effect A, B, C etc is irrelevant. I.e. the multiplier is irrelevant.