« Revisiting Ontario's Decline | Main | Does tweeting signal success? »


Feed You can follow this conversation by subscribing to the comment feed for this post.

"Neither output, prices, nor interest rates, will change at all. "
Yup, 1974, Robert Barro, "Are government bonds net wealth?"

Do you think that this is an example of the dangers of working on novel, pushing-the-frontiers-of-knowledge research when instead economists should be repeating a centuries-old message:

If people are rational, and care about their children and grandchildren, simply cutting taxes won't promote growth.

Isn't holding the path of government spending constant assuming the conclusion? Constant in what sense, nominal, real, percentage of gdp, or percentage of expected gdp? Expectations are tricky things. The convention is the government act is unanticipated but the result is fully anticipated, but it could just as well be both are fully anticipated in which case a failure of the government to act is unanticipated which causes a change in expectations and a non action could cause a reaction.

I am a bit surprised nobody notices the close analogy between Ricardian equivalence and the liquidity trap. Both concepts rely on the anticipation of a future reverse of the policy. I also see another hidden hypothesis made by Barro: fiscal policy has no influence on monetary policy. Of course, it is legitimate to do so, but the debate would be clearer if fiscal policy proponents mentionned that (Scott Sumner already made that point earlier.)
By the way, Ricardian equivalence does not exclude some second order effects: fiscal policy has still an impact on deadweight losses and there is still a case for a deficitary budget during recessions to lower deadweight losses (but much weaker).

Antonio Fatas: "The assumption that I was questioning in my post was the fact that models where Ricardian Equivalence applies are models where the economy is in equilibrium. For simplicity you can think about it as the economy always being in full employment."

I don't think he's assuming full employment. He's just trying to simplify his explanation. But having said that, I'm not at all sure what other kind of equilibrium notion he has in mind.

If Ricardian Equivalence required full employment, the original rationale for the debate wouldn't even exist. Who cares about the effectiveness of fiscal policy if you're always at full employment anyhow? Either Fatas is confused, or he's making his point in a confusing and misleading way. He doesn't actually say that Ricardian Equivalence requires full employment, but this is the impression left by his post. If he means something quite different, then he needs to clarify it.

It's true, as I understand it, that the canonical New Keynesian model does have full employment in the sense that there is no unemployed labor. But I take the model as using underproduction as a metaphor for unemployment because it's easier to model. If Fatas wants to throw out the canonical model because it's too unrealistic, I can't say I entirely disagree with that goal, but I would say that Ricardian Equivalence should be the least of his worries. I'm guessing, too (based on the same logic you use in this post), that a model that treated the supply side more realistically would still exhibit Ricardian Equivalence if it maintained similar demand-side assumptions. I'm not certain about that, though: models sometimes surprise you when you write them down, and the supply/demand side distinction is not as clear cut as we pretend. I would, however, say that the burden of proof is on those who suggest otherwise.

Frances: yep. That was the paper that resurrected REP. (Ricardo had mooted it as a curiosity, but rejected it).

Lord: the standard keynesian assumption is to hold real government spending constant. But if REP is true, the effects of a bond-financed tax cut would be zero whether we held, real or nominal government spending constant, or G/GDP constant. Because they would all mean the same thing, if P and GDP aren't affected by the cut in T.

jean: hadn't thought of that analogy. I always think of the analogy to the Modigliani Miller Theorem. Yep. REP assumes future money supplies stay the same. And lump-sum taxes. Etc......

Kevin: Here's more from Antonio:
"What happens in a world where unemployment is high and there is a large output gap? Depending at which speed the economy returns to potential, we have a different path for income and potentially for private spending. If fiscal policy raises employment and income by closing the output gap faster, private spending will increase even if the other assumptions are still valid."

"If fiscal policy raises employment and income...." But it won't, under REP, if it's a tax cut.

Andy: agreed. But Buchanan and Barro, for example, would still argue about REP even under full employment. Will deficits raise interest rates and/or exchange rates and crowd out investment and/or net exports? Will they create a burden on future generations?

But just write down a standard ISLM, for example, hold P constant, and replace the keynesian consumption function with the PIH, substitute the government budget constraint into the household budget constraint, and REP pops right out. (That's implicitly what I'm doing here).

Is the economy on the IS curve at all? Isn't the IS curve itself an equilibrium locus - both the Old and New Keynesian IS curves?

Confused: The (old keynesian) IS curve is a *semi*-equilibrium locus. It's a set of points in {r,Y} space such that output *demanded* at r,Y equals output Y. It says nothing about output *supplied* -- the level of output firms (and workers) would *like* to sell. The standard (old) keynesian model does assume the economy is "on" the IS curve, but may be off the LRAS curve.

See my old post:

The standard keynesian argument for fiscal policy in a recession does assume the economy is on the IS curve (or gets on to it fairly quickly), and that fiscal policy works by shifting the IS curve to the right.

To clarify:

Full equilibrium requires Yd=Y=Ys. The semi-equilibrium of the IS curve requires only that Yd=Y.

No disagreement at all about the Old-Keynesian IS being, as you say, semi-eq condition. My main point was if it makes sense to assume the economy "on" the IS curve( esp. the NK IS). I'll need to think about that. Thanks for the old link. You write with great clarity.

"Ricardian Equivalence says that a bond-financed increase in government spending is equivalent to a tax-financed increase in government spending. That's where the "equivalence" comes from."

Hmmm. So let's take the case of the U.S. If we had maintained the Federal debt at the level of 1835 (basically zero) and had relied solely upon taxes to fund our ever-increasing gov't spending, things would have unfolded pretty much as they did? Really? Does anybody believe that?

Doesn't this go against the idea that a moderate deficit is a good thing?

Min: nobody believes that REP is the whole, literal, exact truth. It's a starting point.

Nick: "nobody believes that REP is the whole, literal, exact truth. It's a starting point."

Scuze me, but that does not inspire confidence.

Min: My view is just the opposite. I would lose confidence in any economist (or anyone?) who thought that any economic (or other?) theory was the whole, literal, exact truth. The old saw: "All theories are false; some are useful".

Thanks, Nick. :) Then I misunderstood your previous comment.

However, I have seen no explication of the Ricardian Equivalence/Rational Expectations hypothesis that seems to meet what Freud called reality testing. Oh, it is a plausible negative feedback mechanism, but it seems to claim too much.

Min. I'm really not someone who has been following this literature, but econometricians have been testing REP (or PIH+RE). Trouble is, when you test it, like any theory, if the test is powerful enough you will almost certainly find it false. Which isn't very interesting.

The interesting tests are those that test PIH+RE against some alternative theory. The natural alternative in this case is the Current Income Hypothesis (consumption depends on current income only). IIRC, the answer those tests give is "the truth is about halfway between those two", or "it depends". Which sounds plausible to me. But again, I haven't been following this literature.

My take-away is that REP is not something we should believe. But it is something we should think about. We shouldn't ignore those effects. They very probably matter.

Nick: "My take-away is that REP is not something we should believe. But it is something we should think about. We shouldn't ignore those effects. They very probably matter."

Thanks, Nick. :)

It could well be that consumption depends more on permanent income for each individual, but in aggregate it depends more on current (national) income, as most income smoothing cancels out at the aggregate level.

For the individual, there is no reason to believe that the discount rate used is the risk-free rate, as expected future wages are not guaranteed by the full faith and credit of the government. There will certainly be a risk premium that will be constantly updated as more information becomes available. It's plausible that it would increase with income uncertainty during downturns and decreases during booms.

But even with constant risk premium for each person, someone with secure work, or someone rolling over treasuries would have a low risk premium, but a marginally attached worker would have a high risk premium, and would appear to be consuming mostly out of current income, e.g. feast or famine. Then, as the marginally attached workers lose current income during the downturn, the economy appears to be consuming less because of a decline in current income.

I just realized the above was muddled.

What I *meant* was, assume that a marginal worker is earning $1 in period 1, there is a 50% chance he will earn $3 in period 2, and a 50% chance he will earn $1 in period 2.

With a 0 rate, no discount, and log utility, he will maximize expected utility by borrowing only $0.22 in the current period, consuming 1.22, even though his average expected income is $1.5. That is what I meant by discounting permanent income by idiosyncratic risk. Ex-post, the consumption in the economy will be either (1.22, .78) or (1.22, 3.78), with expected consumption = (1.22, 2.28).

Ricardian Equivalence follows from the Permanent Income Hypothesis only if a current increase in spending leads to a future increase in taxes. But if we're not at full employment, we can increase spending without increasing future taxes. Thus, following the P.I.H., consumers will foresee that the policy will not cause a future tax increase and will not need to decrease their consumption to compensate.

We can see that increasing spending when not at full employment will not increase future taxes without resorting to the IS-LM model. At full employment, if we produce more consumption goods today we have to invest less today. Our supply of labor is finite, and each worker can be doing one thing or another but not both. If we invest less today, that means we can produce less tomorrow which means that we can consume less tomorrow.("consume less", in this case, means expect higher taxes) If we're not at full employment, however, we don't need to decrease current investment to increase current production. We can increase current production by employing unemployed workers. Since those workers were not producing anything before, employing them doesn't reduce our level of investment. Since our investment is not reduced, our future consumption is not reduced and we have no reason to expect that we'll need to consume less in the future(ie no need to expect higher taxes). In fact, if we employ a fraction of the unemployed workers in investment activities rather than consumption activities, we should expect that our future production and consumption will increase.

I suppose Fatas derived his conclusion from Krugman's post (to which Nick commented "Thanks Paul. Yep.")

Essentially, Krugman asserted that the Ricardian equivalence does not hold in the liquidity trap. I think Fatas took the contraposition of the assertion, which is, if the Ricardian equivalence holds, then the economy is not in the liquidity trap. (That is, not being in the liquidity trap is necessary condition of the Ricardian equivalence.)

zosima: "But if we're not at full employment, we can increase spending without increasing future taxes."

I misses seeing your comment earlier.

I don't see how that follows. REP does assume that the rate of interest exceeds the growth rate of the economy. (Otherwise, the government could simply roll over the extra debt forever, and the debt/GDP ratio would not grow). This implies, from the intertemporal government budget constraint that current debt + PV(Government spending = PV(Taxes). So a tax cut today implies a tax increase in the future, so that PV(Taxes) stays the same. This says nothing about whether the economy is at full employment or not, either today or in the future.

himaginary: I hope you are OK. Very sad news out of Japan. But from what I hear the Japanese are coping very courageously with a terrible problem.

I had to look up "contraposition"! I don't think that Paul Krugman asserted that. Instead, he asserted that even if Ricardian Equivalence holds, temporary increases in government spending will increase employment in a liquidity trap.

Nick, thank you very much for your kind words. I too believe we can somehow muddle through these problems. I hope that is sooner rather than later.

"Instead, he asserted that even if Ricardian Equivalence holds, temporary increases in government spending will increase employment in a liquidity trap."

Now I realized that maybe you and Fatos are using different definitions of the Ricardian Equivalence.

I suppose Fatos thinks of the Ricardian Equivalence as defined in investopedia, which is "when a government tries to stimulate demand by increasing debt-financed government spending, demand remains unchanged." In Krugman's model, the government increases spending and the private consumption remains unchanged, so the total demand increases. Therefore, the Ricardian Equivalence in that definition does not hold any more. My previous comment was also along that definition.

On the other hand, I re-read your post and realized that your definition is "a bond-financed increase in government spending is equivalent to a tax-financed increase in government spending", which doesn't impose any particular constraint on total demand. Therefore, "That still leaves open the possibility that an increase in government spending (whether bond- or tax-financed) will have an effect." I suppose this is correct definition in academic context, while investopia definition is what usually comes to people's mind about the Ricardian Equivalence.

One must consider the public choice aspects of taxation. Tax increases invariably lead to permanently higher average expenditures. Tax cuts are seldom matched by budget cuts, and almost never result in permanent decreases in average spending.

Revenue volatility increases with tax rates in a progressive tax system. Demand for social services increases precisely at the same time revenues are falling. Special interest groups lobby hard to retain benefits enshrined in structural deficits, and grudgingly accept only temporary clawbacks when cyclical deficits necessitate it. In the recovery period, the special interests use their "sacrifices" during the recession as leverage for recovering all their "losses" as if they were entitled to growth in benefits.

A period of rising revenues is a good time for special interests and government bureaucrats to lobby for more money. Politicians, beholden to special interests, may grudgingly cut budgets but do so at a slower pace than revenues are falling, hoping that an ensuing recovery will make further cuts unnecessary. Because of this, surpluses are completely spent and, even if they were saved, would not fully offset subsequent deficits.

Policy makers and their motivations need to be put into these models, as should rent seeking (or rent retaining) behavior. To ignore these impacts on taxes, expenditures, and debt would be to wave our hands at approaching calamity.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad