Scott Sumner  

The Keynesian model is not a "big government" model

PRINT
The End of British Coal... The Keynesian Model is not a B...

[This is a sort of follow-up to my most recent Econlog post.]

Before getting into this post, let me remind you of previous pointless debates, such as whether "Islam" is "really" a this or that sort of religion. I hope my readers can see the pointlessness of those debates. Similarly, I do understand that many Keynesians, and also many anti-Keynesians, think that Keynesianism is a sort of big government ideology, at least when compared to new classical economics, monetarism, Austrianism, etc.

So there is a sense in which the title of this post is wrong. If Keynesians act as if their model has big government implications, then in a sense it does.

So what do I mean by the title of the post? I mean that the technical aspects of the model have no big government implications, and that people who claim otherwise are simply using bad logic. To use the Islam analogy, I'm looking at the Koran. Here's The Economist:

The fundamental reason for Europe's low interest rates and bond yields is the fragility of its economy. Its unemployment rate is stuck at 10%. While the ECB has been doing what it can to press down the accelerator, however, the austerity preached by the likes of the German and Dutch governments has slammed on the brakes. For years, Mr Draghi has been saying that monetary policy alone cannot speed up the economy, and that creditworthy governments must use fiscal policy as well, ideally by raising public investment. If Mr Schäuble wants higher yields for German savers, he should be spending more money. Instead, his government is running a budget surplus.
I have lots of problems with this:

1. The Keynesian model is wrong, fiscal stimulus is not a good idea. The Economist is flat out wrong when they claim, "the ECB has been doing what it can", that doesn't even pass the laugh test. If they don't know how to do more, then please call me in for an interview and I'll show them.

But in this post I'm not interested in bashing the Keynesian model, and for the rest of the post I'll assume I'm wrong and the Keynesians are right. Fiscal stimulus is needed.

2. Fiscal stimulus does not mean "more spending" as The Economist claims, and as hundreds of other recent news articles claim, it means a bigger budget deficit. I don't think fiscal policy ever makes much sense, but if I were wrong I'd always favor tax rate changes, never spending changes. There are no big government implications from pro-cyclical tax rates.

3. Some claim that spending changes are more powerful, but that's probably wrong, and irrelevant even if it's right. Romer and Romer showed that tax changes had a surprisingly large multiplier effect, which you'd expect due to their supply-side effects. (There's still no agreement as to the relative size of tax and spending multipliers.) Some people oppose tax cuts because they are regressive. But that's also wrong; tax cuts can involve lump sum rebates, as with the 2008 Bush stimulus, or higher EITC, etc. The regressivity or progressivity is totally up to the government.

4. And even if I'm wrong and spending changes have a bigger multiplier effect, you should still never use spending increases. Just make the tax cuts bigger. Spending changes should always reflect classical cost/benefit considerations; they should never be used to boost AD.

5. Some Keynesians argue that government investment is more justified when interest rates are low. I hope by now that everyone sees that as a basic EC101 error, reasoning from a price change. Interest rates are often low precisely because the productivity of new investments is lower than usual.

6. Some Keynesians suggest that more infrastructure can be justified right now on cost/benefit (NPV) considerations. Fine, but don't pretend that that has anything to do with Keynesian stimulus. Even anti-Keynesian economists agree that investments are justified if the cost benefit exceeds the benefits costs, and that's true even when there is no need to boost AD. If the term "stimulus" is to have any coherent meaning, in the sense that Keynes intended, then it must mean boosting the deficit in a way that would not be justified under classical cost/benefit considerations, but only becomes worthwhile when taking account of the expected impact on AD. Otherwise you are just doing classical economics.

7. And even if everything I said above is totally 100% wrong, I'm still right and the Keynesians and anti-Keynesians are still wrong, Keynesian economics is not inherently a big government model. It doesn't call for bigger government, it calls for more countercyclical government. Thus if under classical assumptions you favored government spending 36% of GDP, with 4% of that infrastructure, a Keynesian who rejected all of my arguments above might favor spending of 34% during booms and 38% during recessions. The infrastructure spending would be 2% of GDP during booms and 6% during recessions. (In practice, other spending is also somewhat countercyclical; I'm just trying to simplify my example.)

So even if everything I said in points 1 through 6 is completely wrong (and they aren't all wrong), the assumption that Keynesianism is a big government model is not based on anything in the model itself, but rather the (false) perception of both Keynesians and anti-Keynesians. What can we infer from this widespread misconception? One possibility is that the model is being misused. Another is that people are confusing battles over the size of government with battles over stimulus. This might explain, for instance, why Paul Krugman favored higher taxes in 2013. He believed that a tax increase would lead, in the long run, to higher spending---which he prefers.

But I would caution readers that there are many more conservative Keynesians than most people assume. When people use the term "Keynesian" they often instinctively think of progressives, but unless I'm mistaken most conservative economists (Mankiw, Feldstein, etc.) also think of macro issues using a Keynesian framework. There is nothing necessarily "big government" about that theoretical framework.

I am a libertarian who dislikes Keynesian economics, but not for "anti-government" reasons. If you could convince me that it worked in a technical sense, I'd immediately favor tax cuts in recessions and tax increases in booms. I'm eager to be converted; I want to be a Keynesian. So show me the evidence.


Comments and Sharing






COMMENTS (19 to date)
Jon Murphy writes:

This is an excellent post, thank you!

I think you're right that the model is sometimes misused and that leads to the "big government" label. There are politicians and pundits who use the model to justify any and all government spending, regardless if it's countercyclical or not. Similarly, there are those who react to every penny of government spending as "big government," regardless of whether or not said spending is part of a necessary government function or not.

BTW, I'm focusing mostly on deficit spending in this comment because I think that's the part of the model most abused in this manner.

mobile writes:

Keynesian models get a lot more attention during economic downturns, so it shouldn't be surprising that in most peoples' minds it has no other associations than deficit spending / big government.

Ally writes:

Scott,

In point 6. you say:
"Even anti-Keynesian economists agree that investments are justified if the cost exceeds the benefits"

I think you have costs and benefits mixed up in that sentence.

Brian Donohue writes:

Well done. Lucid.

Andrew_FL writes:

In my comments on your earlier post I was not saying the formal Keynesian Model is a Big Government or Socialist model. I was saying Keynes The Man leaned towards those things. The formal Keynesian Model has very little to do with Keynes himself, except to the extent he inspired it. We owe the formalizations of Keynes, we owe Keynesian Hydraulics, to people like Alvin Hansen and John Hicks, and the many who came after them.

Also:

Fiscal stimulus does not mean "more spending" as The Economist claims, and as hundreds of other recent news articles claim, it means a bigger budget deficit.

Strictly speaking, isn't it the case that it means not just bigger deficits but specifically government deficit finance of purchases of actual final goods? Oh, sure, you can develop a detailed consumption function for effects of transfer payments but the basic, macro 101 Keynesian Cross is all about the big G in GDP.

Roger Sweeny writes:

I think it was James Buchanan in Democracy in Deficit who said that though the Keynesian model did not imply bigger government, it inevitably led to it. The model says run deficits in bad times and surpluses in good times, so the budget is "balanced over the business cycle." But good times are rarely considered good enough to run surpluses for long.

And since government agencies and advocacy groups always have wish lists, and since any legislator can say she is "creating jobs" by funding some of those wishes, it is easier to expand government when "stimulus" is needed than it is to send money back to the people.

MikeDC writes:

To me the problem is that Countercyclical policy is a political suicide device.

Everyone wants to be countercyclical in a recession, but no one wants to be countercyclical during an expansion. Because (under Keynesian logic) you're quite literally standing up and saying "I wish to enact policies that will reduce economic growth".

Michael writes:

Amen! Except for Point #5, "Interest rates are often low precisely because the productivity of new investments is lower than usual."

Yes, they are often low for that reason. But government borrowing rates can also be low if there is abnormal demand for safe assets, while private firms must pay abnormally large risk premia at the same moment. Many believe we have been in that circumstance since 2008, and if so the low rates should indeed factor into public-sector cost-benefit analysis.

Scott Sumner writes:

Thanks Jon and Brian,

Mobile, Actually, I'd say it's surprising that people don't advocate tax cuts in recessions. The Keynesian model does not have big government implications, even during recessions.

Thanks Ally, I fixed it.

Andrew, You asked:

"Strictly speaking, isn't it the case that it means not just bigger deficits but specifically government deficit finance of purchases of actual final goods? Oh, sure, you can develop a detailed consumption function for effects of transfer payments but the basic, macro 101 Keynesian Cross is all about the big G in GDP."

No, I don't agree. G plays no special role; AD is all that matters.

Roger, I've heard that argument, but I don't agree with it. Government expands over time (when it expands) for reasons that are mostly unrelated to the Keynesian model.

MikeDC, You said:

"Everyone wants to be countercyclical in a recession, but no one wants to be countercyclical during an expansion."

I could argue that the Fed does.

In any case, it's mathematically impossible to be countercyclical in recessions but not booms. If spending is above average in recessions, then, ipso facto, it's below average during booms. So again, no big government bias.

Michael, I think you misunderstood me. Low rates should always factor into public sector investment decisions. But the productivity of public investment today is far lower than in the 1960s. Compare the Interstate Highway System and California high speed rail today.

Andrew_FL writes:

@Scott Sumner-"No, I don't agree. G plays no special role; AD is all that matters."

But if you define AD as GDP, then G is the government spending component, so if government deficits are to be simulative, don't they need to impact G? Put another way, is it actually true, from your perspective, that in the Keynesian Model, stimulus can be done entirely through transfer payments while no actual purchases are made by the Government?

ThaomasH writes:

Is it "Keynesian" if during a recession in which the central bank is allowing NGDP to remain below trend, for the government to step up expenditures on activities that pass an NPV test? [I'm assuming that a) real LT interest rates will have fallen sharply relative to the level when NGDP was on trend; b) marginal costs of many inputs into public expenditures have fallen below their market prices; and c) the recession has not yet reduced the long run growth potential of the economy.]

If instead, the government reduces expenditures to reduce the deficit, is that not properly called "austerity" in the pejorative sense?

Greg G writes:

Roger,

Yes, Buchanan did argue in "Democracy in Deficit" that a countercyclical fiscal policy might work in principle but could never work in a democracy. He had a point, but there is a lot of irony in that argument.

First of all Buchanan's ideas proved even harder than Keynes' to implement in a democracy. And he was freaking out about the debt at exactly the point debt to GDP reached its post WWII low.

And roughly Keynesian policies were implemented with good success in the generation after WWII. Debt to GDP was reduced by 2/3 before being exploded again by Supply Sides Economics under Reagan who certainly did not consider himself a Keynesian.

[link added--Econlib Ed.]

Roger Sweeny writes:

For many on the right, "Keynesian" is a dirty word, so Reagan may well not have been willing to assign that label to himself. But he certainly believed that tax cuts "stimulated" the economy. He also believed that eventually they would lead to smaller government--which was a major reason many opposed them. He was, of course, wrong about that.

Since then major deficit spending has been rather bi-partisan, the major exception being divided government during Clinton virum.

Scott Sumner writes:

Andrew, In the Keynesian model, tax cuts boost C, and in some cases I.

Thaomas, You asked:

"Is it "Keynesian" if during a recession in which the central bank is allowing NGDP to remain below trend, for the government to step up expenditures on activities that pass an NPV test?"

No, that's classical economics. That's not to say Keynesians would be opposed, it's sound reasoning, but that's not really Keynesian economics.

Roger, Reagan favored tax cuts for supply side reasons, not as a way of boosting demand. Demand was certainly the least of our problems back then, the Fed was aggressively trying to reduce demand, and Reagan approved of their anti-inflation program.

In my view the Reagan tax cuts did reduce spending, relative to the no-tax cut alternative. If he had not cut taxes, I believe spending would have grown more under Clinton.

Cullen Roche writes:

Scott, this is a very excellent post. Well done.

I've talked about this point before and how the term Keynesian is misconstrued as meaning procyclical persistently larger govt.

Nice to see that we don't disagree nearly as much as I might have previously thought.

Best,

Cullen

WC Varones writes:

"tax cuts can involve lump sum rebates, as with the 2008 Bush stimulus, or higher EITC, etc. The regressivity or progressivity is totally up to the government."

Not really. The EITC handout is spending, not a tax cut. And with the poor paying zero taxes and the upper-middle to wealthy paying the vast majority of taxes, governments are constrained in making tax cuts progressive.

Now you can certainly write big stimulus checks to lower-income households, and that's in-line with Keynesian prescriptions. And it's big government.

Marc Sargen writes:

I agree that that Keynesian does not mandated a big government but you since we're taking the True model then you point 3. is plain & simply wrong. Keynesian theory clearly shows that a tax cut is less efficient in causing a fiscal stimulus than direct fiscal spending because is does not have the initial fiscal push that spending does.

In many ways the best part of Keynesian theory, the counter-cyclicle balance is already baked into government programs. Welfare/Transfer spending like unemployment increased during problem times when people need support get them past hard times & then recedes are jobs increase & needs decline. Progressive taxes takes in more revenues during boom time trimming euphoric growth but drop as earning become lean. Unlike most stimulus spending or growth cooling costs they happen automatically & without the time delay that sabotages any other fiscal support program.

But while these programs would give a smaller, more fiscally stable government, they are view as not enough by most Keynesians but then that's not surprising because the biggest cheerleader to do more was Keynes himself.

It's clear that Keynes voted for big government, & it's his models so I'd have to say that is clear that Keynesian economic requires advocating for a massive government. It's not just the nuts & bolts of the model that make the social science of economics but also they basic assumptions of the model. Keynes assumed big government spending is preferred.

Floccina writes:

Another great post!
Seems like the best Keynesian policy would be to temporarily have the Fed Gov. pay the FICA and matching FICA + Medicare taxes up to some amount for each employee. So say the 1st $15K a person earns is tax free.

Benjamin Cole writes:

Excellent blogging.

Comments for this entry have been closed
Return to top