Thursday, June 13, 2019

MMT and the Theory of the Government Budget Constraint 1


One of the key arguments which MMTers all seem to share is the view that there is no such thing as a government budget constraint, at least, not for a government which can print its own money.  There’s a certain irony to this, since the bit of traditional theory with which MMT seems to have the most in common is the Government Budget Constraint literature of the 1970s, as represented by Blinder and Solow (1973), and extended by Takayama (1980).


MMTers argue that the fact that a government can print its own money (so we’re only looking at the national level, not at the level of the province) means that there is no limit on its ability to spend, and hence nothing to stop it raising aggregate demand to the full employment level.  (The first part of this statement is true in nominal terms, although their full policy argument rests critically on the argument that expanding the money supply does not cause inflation, or, as Stephanie Kelton seems to have framed the case, if inflation does start to appear, the appropriate policy is to raise taxes to pull money out of the economy.)  In interpreting the term “government budget constraint” as they do, though, the MMTers miss the key element of the 1970s literature.  What the GBC does is not put a hard limit on government spending, rather it emphasizes that government spending has to be paid for somehow and shows that the decision to run a deficit automatically converts a static economic problem into a dynamic one.

Consider the case where the economy is in Aggregate Demand/Aggregate Supply (AD/AS) equilibrium and the government’s budget is balanced, so that the equilibrium level of aggregate income is generating just enough revenue to pay for the current level of government spending.  Assume, however, that the current equilibrium level of aggregate income is not the full employment (potential) level.  The government decides to increase G to shift the AD curve to the right. 
That increase in G could be paid for by an increase in taxes, T, in which case the government’s budget will still be in balance.  In the story we’re telling, using a simple Keynesian model, this would require an increase in the income tax rate.   In intro macro theory terms, the balanced budget multiplier comes into play.  It may be, however, that the government wants to set off a larger short run multiplier effect than the balanced budget multiplier yields, so it decides to run a deficit – to increase G with no increase in the tax rate so that G rises by more than T.  It has to acquire the funds necessary to pay for the additional G.

Our government has two options open to it (we ignore the mixed option, as did the GBC literature).  It can either borrow the funds it needs from the public, by printing new government bonds and selling them to the public, or it can finance the increased spending by printing money – technically by printing new government bonds and selling them to the central bank.  The first approach was referred to as bond financing of the deficit and the second as money financing of the deficit.  MMTers come down in favour of money financing the deficit as, interestingly enough, did a lot of conventional Keynesian economists of the 1970s.  What the 1970s GBC literature was about, though, was what happens next. 

Suppose the increase in G succeeds in raising equilibrium income to the full employment level.  Government spending will have risen, and tax revenue will also have risen by an amount equal to the product of the (assumed proportional) tax rate times the ISLM multiplier (which allows for any increase in the rate of interest) times the increase in G.  (The 1970s literature, for the most part, assumed that the price level was constant through all of this, another point of commonality with MMT.  For a diagrammatic exposition in AD/AS terms, which allows for price increase, see Mayer (1984).)  So the increase in income caused by the increase in G will yield an increase in T which will pay for part, but not all, of the increase in G.  The government will have gone from having a balanced budget to running a deficit, although the deficit won’t be as large as the increase in G.

That’s what happens in the period in which G is increased.  Now consider the next period.  We want to hold income at the full employment level, which means that we have to keep G at its current, higher level.  That means that in the second period income will remain unchanged at the full employment level, G will remain at the level which pushed income to full employment and T will remain at the level generated by full employment income.  The catch is that G will still be larger than T, so the deficit will also be unchanged.  In other words, there won’t be enough tax revenue coming in to hold G at the level which will keep Y at its full employment level.  So there will still be a deficit and the government will have to come up with funds equal to that deficit.  If we think in terms of money financing, Y will remain unchanged, but the money supply will have had to be increased by an amount equal to the amount of the deficit. 

The key here, obviously, is the fact that G has not somehow been ratcheted up permanently – each period’s G is a new episode of spending and each has to be paid for in its own period.  This is why we said that the GBC literature emphasized the dynamic element of government spending – Y, G and T might remain unchanged over time, but the money supply will keep growing as long as the deficit persists. 

Interestingly, the 70’s GBC literature doesn’t say that Y, T and the deficit will remain unchanged over time, even though G will, but we’ll take that up in another post.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.