ICYMI, Simon Wren-Lewis wrote a post wondering what MMT contributes to macroeconomics that isn’t either false or already well-known. I answered with a post arguing that MMT is really a philosophical theory about the purpose of currency and the role of the state.
Wren-Lewis acknowledged my argument but remains unconvinced by the theory of currency. Nick Rowe wrote me a detailed reply, arguing against that theory. It gave me a lot to think about, though I’ll save a reply to his reply for a later post.
I should thank both of them. Rowe engaged me in a detailed and informative discussion. Wren-Lewis paid me the compliment of saying that my post plus discussions allowed him to pick which parts of MMT he does and doesn’t agree with, which is all I’m really after. I think we might be getting somewhere.
I’d like to continue the dialogue a bit more, so I hope at least one of them reads this at some point.
MMT and IS-LM
Rowe suggests the relevant bit of MMT for policy is substantially reducible to IS-LM analysis with a flat LM curve and a vertical IS curve (Rowe points out that this reduces to an even more basic ‘Keynesian Cross’ analysis). Wren-Lewis admits that some MMTers propose something more complicated than this – more an IS curve of unknown shape.
Strictly speaking, of course, we can’t reduce MMT to any sort of IS-LM. MMT, like Post-Keynesian economics in general, rejects IS-LM analysis altogether as simply not applicable to the real world.
But I don’t want to quibble about that now. Maybe Rowe’s point could be restated as follows:
IS-LM with a flat LM curve and a vertical IS curve captures as much of the MMT position as can be represented within standard macroeconomics.
I want to dispute that claim.
The Flat LM Curve
The “flat LM curve” part I definitely agree with. As Rowe points out, this just represents the operational reality of a floating fx system with no actual reserve constraints on banks. Mitchell and Wray’s MMT textbook explicitly states that if we were to use IS-LM, we should have a flat LM curve:
The fact that the money supply is endogenously determined means that the LM will be horizontal at the policy interest rate.
So far, Rowe is right.
Next, as he notes, New Keynesian macro generally accepts this revision to classic IS-LM. Thus Rowe draws the “New Keynesian or Neo-Wicksellian version” of IS-LM as follows:
This is a picture of an economy in recession. The current equilibrium is at (Y0, r0). But ‘full employment’ output (or output at the ‘natural rate’ of unemployment) is Yn. The ‘natural’ rate of interest, given the current level of spending, is rn. In terms of a policy response, you can have a fiscal expansion to push the IS curve outwards until rn rises to the level of r0, or you can have monetary policy lowering the interest rate (currently r0) until it falls to rn. (Inflation expectations are left out for simplicity.)
We should pause here to remind ourselves that one main policy disagreement between MMTers and New Keynesians is that the latter believe, and the former do not, that monetary policy on its own can bring the economy to a ‘full employment’ level of output. For New Keynesians this is true except in the special case where rn<0. For MMTers it isn’t true at all.
Wren-Lewis asks: “What about a world where monetary policy did successfully control demand and inflation, which is the world I’m writing about?” The MMT reply would be that such a world doesn’t exist. But why not?
The Vertical IS Curve?
To get to this MMT conclusion, Rowe proposes a vertical IS curve. If the IS curve is vertical, then raising and lowering rates quite obviously can have no effect on output, and so the world in which monetary policy controls demand and inflation doesn’t exist:
But I disagree with this reading of MMT.
The key point for MMT is not that aggregate spending is unresponsive to changes in the interest rates, so that the IS curve is vertical. Rather, the point is that monetary policy has fiscal effects, and these undermine the intended effects of monetary policy.
Monetary Policy is Fiscal (whether it wants to be or not)
Here is how Warren Mosler puts it:
The only way a rate cut would result in increased borrowing to spend would be if the propensity to spend of borrowers exceeded that of savers. The economy, however, is a large net saver, as government is an equally large net payer of interest on its outstanding debt. Therefore, rate cuts directly reduce government spending and the economy’s private sector’s net interest income.
The first bit suggests that the propensity to spend of borrowers and savers might be such as to have an IS curve with something other than a negative slope. But Mosler doesn’t follow that up. Instead, he points out that monetary policy has fiscal results: cutting rates reduces interest payments from the government to the private sector. Raising rates increases such payments.
This isn’t an MMT point as such. It’s just a recognition of the interest-income effect. But MMT, which is particularly focussed on existing institutions and operations, takes fairly seriously the size of the interest-income effect in a context where the government is very big and a net payer of interest.
My point here is just that this idea can be represented in the New Keynesian version of the IS-LM diagram. There is no need for any vertical IS curve.
MMT in IS-LM
In the New Keynesian version of IS-LM analysis, monetary policy is represented as upward or downward shifts of the LM curve, as the central bank adjusts rates. Meanwhile, fiscal policy is represented with leftward or rightward shifts of the IS curve.
To see Mosler’s point, imagine that the central bank lowers rates during a recession. Here is the key: In the first instance, the result of that will be an decrease in total government spending, since interest payments on government debt are reduced without any other adjustments in spending. Multiplier effects follow. This shifts the IS curve leftwards, so that the end result can be a fall in output. It depends in part on how far the IS curve shifts. But MMT, again, takes seriously the institutional features that make the interest-income effect likely to be big. Thus in the diagram below, the lowering of rates, combined with the interest-income effect, provokes a move from (Y0, r0) to (Y1, r1):
These effects occur only in the first instance, as I highlighted above. In the above example, the government can take the low rates as an opportunity to cut taxes or increase other spending, thus countering the negative interest-income effect. On the other hand, it might not do so. That’s a matter of fiscal policy.
The moral is that the effect of monetary policy hangs entirely on the response of fiscal policy. Monetary policy on its own can’t control demand and inflation because it never operates on its own. Interest adjustments always have fiscal effects, and where the government is a large net payer of interest, these will be big enough to matter a lot.
Importantly, lowering the rate when the government is committed to fiscal austerity is likely to reduce aggregate demand, not raise it.
Given our current institutional framework, that’s the world we’re in. And it is the world that MMT is writing about.
To be clear: I don’t think IS-LM is the right way to express this point. But if you insist on forcing MMT into an IS-LM framework, there’s no need to make the IS curve a different shape from what it is in the standard New Keynesian version.
It’s not about the shape of the curves. It’s about the way they mooooove.