Tuesday, January 27, 2009

Ricardian Equivalence Does Not Imply That Obama’s Fiscal Stimulus Will Be Ineffective

Kevin Quinn noted that the Wikipedia discussion of Ricardian Equivalence had the following error:

Ricardian equivalence states that a deficit-financed increase in government spending will not lead to an increase in aggregate demand. If consumers are 'Ricardian' they will save more now to compensate for the higher taxes they expect to face in the future, as the government has to pay back its debts. The increased government spending is exactly offset by decreased consumption on the part of the public, so aggregate demand does not change.

As noted here, John Cochrane made the same error. I would hope the Myron S. Scholes Professor of Finance at the University of Chicago Booth School of Business does not rely upon Wikipedia for his economic research. Alas, in an otherwise excellent post on fiscal policy, Menzie Chinn sort of falls into this trap as well:

Case 5 (government debts will have to be paid off in its entirety the future): When budget constraints hold with certainty intertemporally, and there is no way to default even partially on government debt (say via unexpected inflation), then increases in government debt due to tax cuts (for instance) induce no change in current consumption because households fully internalize the present value of the future tax liability

Menzie is right about transitional changes in tax policy not being able to change consumption in this Barro-Ricardo model, which is why GOP calls for using tax cuts to stimulate demand are likely not going to be the most effective policy tool. But what about transitional changes in government purchases? It is interesting that Wikipedia noted Ricardo’s 1820 Essay on the Funding System:

Ricardo studied whether it makes a difference to finance a war with the £20 million in current taxes or to issue government bonds with infinite maturity and annual interest payment of £1 million in all following years financed by future taxes. At the assumed interest rate of 5%, Ricardo concluded that "In point of economy there is no real difference in either of the modes, for 20 millions in one payment, 1 million per annum for ever ... are precisely of the same value".

Let’s modernize this example. Suppose we decide to have an additional $100 billion in public investment in 2009. In Ricardo’s example, permanent taxes will increase by $5 billion per year which would have a very modest offsetting reduction in consumption. So if government purchases rise by $100 billion and consumption falls by $5 billion, then isn’t the direct impact on aggregate demand closer to $95 billion for the year rather than zero?

Update: Republicans will oppose more government spending as they prefer tax cuts:

Hours before a meeting with President Barack Obama, House Republican leaders sought to rally opposition Tuesday to a White House-backed economic stimulus measure with an $825 billion price tag. Several officials said that Reps. John Boehner of Ohio, the GOP leader, and Eric Cantor of Virginia, his second-in-command, delivered the appeal at a closed-door meeting of the Republican rank and file. Both men said the legislation contains too much wasteful spending that will not help the economy recover from its worst nosedive since the Great Depression, the officials added ... Senate Republican Leader Mitch McConnell, R-Ky., said in a televised interview that Obama was having problems with Democrats, whom he said favor spending over tax cuts as a remedy for the economic crisis.

If Ricardian Equivalence holds, the GOP is opposing fiscal stimulus that will impact aggregate demand preferring tax cuts that will not increase aggregate demand. Go figure!


kevin quinn said...

Further, suppose it's not for war but for public investment with a greater rate of return than 5%. Then our permanent income is higher, so consumption will increase along with G!

Bruce Webb said...

There is a bizaare idea that floats around all Chicago style economics. Not the one that suggests that all market participants are fully informed, that is just crazy on its face. No I am talking about the one that suggests everyone has a magical tax calculator in their head that resets every bit of economic behavior in reaction to changes in tax level, even where those changes are minor.

These guys are orthodox all right. As in Eastern or Russian Orthodox, they have a belief system and a set of rituals, which don't allow any deviation lest one fall into Heresy. Like thinking that lower income workers benefit from minimum wage, or cut back spending in exact proportion to increases to future deficits that are scheduled to be paid from a revenue stream they currently don't pay.

If I am earning double the minimum wage and have a kid I am not likely to even be exposed to significant income tax. And if I was fully informed and had my magical tax calculator in operation I would know that full well. And so discount most of the future cost of increased deficits to a personal zero.

Which makes Menzies argument collapse on itself. He simply ignores the varying incidence of the income tax.

(BTW this is why proposals to deliver tax relief to workers by lowering the 10% bracket are nonsense. People whose earnings are entirely within that bracket are more likely to be getting EITC than actually paying income tax).

rosserjb@jmu.edu said...


I don't think that Menzie believes in Ricardian Equivalence. Nor does he necessarily buy all the other critical arguments. He is simply laying out how they work.

As for me, I never know what my tax bill will be until I sit down to calculate it. For one thing, lots of the tax changes do not get much pubilicity, and one only encounters them when one actually starts hacking away at the 1040, and if one is using Turbotax or something like that, one may not even quite know what is going on at all, aside from the bottom line.

Anonymous said...

I still don't get what is wrong with the first quotation re: Ricardian equivalence.

kevin quinn said...

Anonymous: here's a simple example. Suppose I make $3000 and pay $1000 in taxes this year. Next year say I will have no income and will pay no taxes. Let's say the interest rate is 0. I decide say to smooth my consumption stream completely by saving 1000 of my after-tax income today, alowing me to consume 1000 in each period. If the government gives me a $500 tax cut today financed by borrowing for one year, I will expect my taxes to go up by 500 next year. So I save the tax cut: I save 1500 total this year (from my 2500 after-tax income) allowing me to consume 1000 in each perios just as before. So far, so good.

Suppose instead of cutting my taxes, the government borrows 500 per person this year to finance increased spending. My taxes will go up by 500 tomorrow, but they won't be lowered today. My life-time after tax income has been reduced, from 2000 to 1500. If my preferences are to smooth my consumption perfectly as before, I will consume 750 in each period. I will save 1250 today in order to do this. My saving has increased, and my consumption therefore decreased, by 250, less than government spending has increased (500), so aggregate demand has increased by 250.

TheTrucker said...

Appeals to Ricardian Equivalence are just like all other conservative appeals in that such appeals assume that government cannot possibly invest in anything. They also assume that demand side stimulus will not occasion investment in the private sector either. What is purposefully denied by conservatives is that improvements in infrastructure cause increases in velocity which cause increases in tax proceeds without increasing tax rates. For brain dead Republicans it does not matter how tax revenues are spent. For them, tax revenues are always burned in a furnace. Note that the Ricardian example is for a war.

Nick Rowe said...

Actually, it depends on whether and how government spending affects people's utility.

If government spending does not affect utility (if it's spent on useless stuff), then what you say is correct.

At the other extreme, if government spending gives equivalent utility to private consumption, so we can write the utility function as U(C+G), then what you say is false. Government spending is like an in-kind transfer of goods that the household would have bought anyway, so Ricardian Equivalence applies to government spending as well as tax cuts.

Both those extreme examples are unrealistic, of course. I haven't done the math for more realistic examples, and I have difficulty figuring it out in my head. But if we assume government spending is useful, I think it all depends on whether it's a substitute or complement for present vs future consumption.

My guess (only a guess) is that for some types of government investment spending, you might even find a negative multiplier. Dunno.

Nick Rowe said...

On thinking about it more, my guess seems to be right. Consider this example:

The government invests in a new highway, which will be completed next year. People today save more of their income in response, so they can buy a car next year and drive it on the new highway. Assuming the rest of Ricardian Equivalence, so that bond-financed spending is equivalent to tax-financed spending, we get a negative government spending multiplier for the current year.

The reason is that in this example, government spending has no effect on utility today, but is a strong complement for future consumption.

It's a cooked-up example of course.

I like Kevin Quinn's argument above. Probably the more general case is where good infrastructure investment increases our future capacity to earn income, so increases our permanent income, and so crowds in current consumption. But it must be *good* infrastructure investment.

MattYoung said...

Ricardo equivalence holds when government is doing its "natural" role, rather than performing some stimulus "countershock".

Yes, Ricardo equivalence implies stimulus does not work, but it also implies that natural rearrangement of government goods and services does work.

The difference is that stimulus supposedly works in unexpected shocks if the stimulus is unexpected.

rosserjb@jmu.edu said...

Ricardian equivalence assumes that people have both rational expectations and a fully generous "bequest motive" for their offsrping, or if they are lacking offspring, for whomever they might wish to leave their inheritance to. Neither of these hold in reality, or at least not fully, although there are some individuals for which either, or even both, of these do hold.

The most blatant empirical counterexamples are the Reagan and W. Bush tax cuts. According to Ricardian Equivalence, the savings rate should have risen after each of them (of course, W. had a bunch of them). Instead they fell. Ricardian Equivalence is an empirical joke.

Dimitri said...

Isn't this criticism slight-of-hand? Let me preface this with the disclaimer that I'm no expert.

The Ricardian equivalence analysis seems to be a present value analysis using perpetuities. The present value of a perpetuity is the annual payment divided by the rate. (So we immediately see that an example with an interest rate of 0% over two periods is not useful).

Examining the "modernized Ricardo's example" in the main post, the aggregate demand in period one is increased at the expense of future aggregate demand. The stream of the change in values looks like this: (+95, -5, -5, -5, ...), where the -5s continue infinitely. The present value of this stream @ 5% interest is zero.

Why is it appropriate to analyze only the first period and ignore the later periods? Specifically, what good is increasing aggregate demand now if we have to decrease it later (in a completely offsetting fashion)?

YouNotSneaky! said...

It's worth reiterating here what PGL's point is: that EVEN IF Ricardian Equivalence holds that does NOT imply that fiscal policy in the form of government deficit spending has no effect on aggregate demand.
If RE holds then both tax cuts and deficit gov spending have the same effect on aggregate demand. If it doesn't they have different effects (MPC/(1-MPC) vs. 1/(1-MPC)). Either way it has an effect.

I corrected this error on Wiki last time Kevin mentioned it but now I notice that it's crept back in there.

Don said...

"households fully internalize the present value of the future tax liability"

I'm puzzled by this statement. Does it purport at all to be about human behavior? Any incentive will work more or less well depending upon a whole number of factors that go into actual human beings making decisions at any given time. Even the historical context makes a difference. I don't think that there's much doubt that massive government spending could influence people's behavior. The question is how much it would take and what might the negative consequences be? These economic equations all look like correlative reasoning or counterfactuals, which are then taken to be like equations that detail the actions of non-teleological objects. Why can't we say say that we're going to spend some and cut some taxes hoping that something might work?

Don the libertarian Democrat

ProGrowthLiberal said...

Dimitri - Keynesian aggregate demand policy is not about increasing spending permanently. It is about moving it forward when the current period has insufficient aggregate demand to achieve full employment.

YouNotSneaky - I noticed that the silly quote noted earlier on this blog was replaced by a similar but somewhat worded differently silly statement. Now I know why. Someone smart (you) removed the first silly version and then someone not so smart replaced it later. Well - you tried to correct the Wiki definition but also it seems Wiki allows too many people access.

Alex said...

I would expect someone trained on the freshwater tradition to know what is implied by Ricardian equivalance, but apparently this is not true.

Plus, the relevance of Ricardian equivalence, at the end of the day, is an empirical issue, which seems rebuffed by research.

Bruce Webb said...

Plus all of this talk ignores taxation feedback. My state (Wash) has a stiff sales tax but no income tax. Our neighbor (Ore) has an income tax but no sales tax. If we each decide to build a $100,000,000 infrastructure project we will get a different mix of tax revenue. That is Washington would get more bang for their buck if it created a huge demand for materials even if the jobs created didn't have particularly high paychecks. Whereas Oregon gets more bang for their buck if it created a huge demand for high-paid engineers who could be taxed on their paychecks. It seems that RE would only hold if these tax effects always magically came out even in all cases under all tax regimes. Whereas as between Wash and Ore it would seem to make a big difference as to which government spent $100 mil repaving roads compared to spending that same money building out a new smart local power grid. Just waving away the material to labor ratio as it relates to tax collections seems very odd indeed.

From the outside RE looks much like the mythical Bed of Procrustes which fit each sleeper just right: you just had to lop the limbs off the big and tall and stretch the small and short so as to fit. A neat solution for the inventor of the bed: "Look a perfect fit!" but a little painful for the test cases.

Orlando Roncesvalles said...

It seems the argument that “stimulus” cannot work rests on rational expectations -- people expect that fiscal/monetary expansion now will require fiscal/monetary retrenchment later. In other words, rational expectations defeat Keynesian prescriptions.

But Kindleberger argued, cogently, that crises are irrational. If we had continuous rationality, we would not observe manias and busts. But we do observe such, so there must be some irrationality out there.

It seems that rationality is useful to analyze the long run where Say’s Law works, and irrationality applies to the Keynesian “short run” (where prices and wages are frozen) and Keynesian policies work.

We are in neither type of “runs,” so there is room for old-fashioned Keynes and also for some skepticism as to its potency.

TheTrucker said...

I look at the wiki article and it is "right" in that it says what "Ricardian Equivalence" is. It is not the job of an encyclopedia to argue whether or not "Ricardian Equivalence" is economically valid. My own opinion (not worth very much) is that the article is correct but that "Ricardian Equivalence" (which was actually predated by Smith and is more correctly called "debt neutrality") is just another installment of neoclassical tap dancing. Both Smith and Ricardo postulated this same theory regarding expenditures for a war and both of them were taking the position that government just "wastes" money. In no way is "Ricardian Equivalence" applicable to invsetments in infrastructure. This sort of "half truth" by rightarded neoeconomists will never stop. They are lying pigs and they will always be lying pigs. The owners of the media will support them in their lying. And we can always count on The Cato Institute to lead the parade.

jrgiguere said...

But for Ricardian equivalence to work you must be at full emplyment which, of course, Smith, Ricardo,the classicals, the Austrians and the freshwaters all presume il always present, whatever the data might say.

ProGrowthLiberal said...

Trucker - if you check the resume of Robert Barro (the fellow who dusted off Ricardo's musings), you will see he did write articles on how increases in government spending were supposed to work in his model and it turns out that they did stimulate the economy. So Wiki's description is inconsistent with the academic writings of Robert Barro. QED!

kevin quinn said...

Nick Lowe has the definitive post on RE and government spending up at his blog, Worthwhile Canadian Initiative.

PGL: yes, Barro has temporary increases in G increasing demand, as you say. From the proverbial horse's mouth. But of course in his classical model that means that the real interest rate rises until C and I fall: there is complete crowding out. So I don't think he'd say the increase in AD stimulates the economy. For those of us who believe that there are involuntarily unemployed resources, this isn't a problem!

kevin quinn said...

Voltaire famously said that you can kill a flock of sheep with "arsenic and oaths." Of course, it's the arsenic that is doing the work. Similarly you can kill a case for fiscal stimulus with Classicism and RE - but it's the Classicism that's doing the work!

Patrick Fisher said...

If "transitional changes in tax policy not being able to change consumption in this Barro-Ricardo model, which is why GOP calls for using tax cuts to stimulate demand are likely not going to be the most effective policy tool" is an accurate statement,

and government spending can increase aggregate demand,

then what is the argument in favor of debt financing over tax hikes?

If cuts don't affect consumption, then barring some magical discontinuity in the rate/consumption curves exactly at the current rate, hikes shouldn't either, and pushing debt to the 12-15 trillion mark is unnecessary.