David R. Henderson  

Bruce Bartlett on "Starving the Beast"

PRINT
From the AER... Fiscal Inflation...

I've posted twice now (here and here) about sections of Bruce Bartlett's book, The New American Economy. I found Chapter 6, titled, "Starving the Beast Didn't Work," quite persuasive.

First, Chapter 6. Bartlett deals with the idea, popular with many supply-siders and many Republican politicians in the 1980s, that even if cuts in tax rates didn't increase tax revenues (and, by the way, they didn't increase revenues), they were a good idea because they would "starve the beast." Translation: federal politicians, looking at the revenue constraint, would tame their appetites for federal spending. Even Milton Friedman got on board the "starve the beast" train. I came trepidatiously to this view in the 1980s but I am ultimately an empiricist. And the evidence, which Bartlett presents, is that starving the beast didn't work. Good paragraph:

Bill Niskanen, chairman of the libertarian Cato Institute and a member of the CEA under Reagan, now argues that the starve-the-beast theory has actually been perverse. It led libertarians and conservatives to think that tax cuts are the only thing necessary to restrain the growth of government. Passage of large tax cuts during the George W. Bush administration led them to become "casual about the sustained political discipline necessary to control federal spending." Niskanen says it is a "fantasy" to think that tax cuts have any restraining influence on spending. [italics in original]

Bartlett tells of James Buchanan's switch from being a strong proponent of balanced budgets to being a proponent of tax cuts to restrain spending. I seem to recall that Buchanan once said that people see taxes as the price of government and so when the price falls, people demand more government. Does anyone know where that line is? I think that's what, in fact, happened.

Later this week, I'll discuss his last chapter, "Dealing With Tomorrow's Economic Crisis," which I found much less persuasive.


Comments and Sharing





COMMENTS (22 to date)
Boonton writes:

Let me plug my "ATM Theory of Gov't Fallacy" yet again:

To wit: You go to the ATM intending to take out $40, thinking you have $300. The ATM says your available balance is $340. You have more than you thought so you decide to make it a night on the town and take out $80.

The ATM Theory of Gov't assumes that Congress gets all the tax revenue into one big pot, sees how much is there and spends accordingly. A tax increase simply makes the pot bigger therefore more spending.

Recent history, at least, shows the opposite. Tax cuts in the Bush admin. were followed by massive spending increases both in war and new entitlements (as well as discretionary spending). Tax hikes under Clinton were followed by spending drawdowns. Tax cuts under Obama (half of the stimulus package was tax cuts) had spending increases following. The ATM theory is backward. Gov't sees $340 in the account and takes out only $20, it sees $280 in the account and takes out $80. Why?

A few ideas:

1. Tax hikes are unpopular for obvious reasons. It becomes easy for politicians out of office to point to spending and link the spending with the tax hikes (i.e. "You just got a tax increase and the NEA gave some wacko artist $5000 to display his stupid paintings!"). Hence politicians in office have an incentive to defend their position by making a show of being anti-spending.

2. "One for me, one for you" - People often put great emphasis on status. If I offer all my guests a fine dinner but offer you some mixed nuts, you'd be offended. But if I offered all my guests mixed nuts you wouldn't feel upset at all....even though you're no better off in terms of absolute value. A Republican lowers the top tax bracket. What about everyone else whose incomes are nowhere near that bracket? Well have a child care credit, some college grants, some pork for your local Congressional district etc. If someone is getting something, it's easier for everyone to get something. Likewise the reverse happens with tax increases. If Bob is paying more in taxes, it's easier to sell Rob on the idea that he must 'give up' something like some road construction grants in his district.


3. Keynesian Theory - Very simple, recession economy tax cuts and spending increases, booming economy tax hikes and spending cuts. Blog all you want about the theory, the fact is a huge portion of the gov't is on autopilot with a Keynesian flavor. The income tax means lower income(s) lower tax expenses. Unemployment means higher unemployment benefits. It can also mean higher social security and disability payments if people at the edge of the labor force figure it would be easier to apply for early retirement or disability than to fight it out once again in a tough market. Likewise Medicaid spending goes up as people loose their incomes and health benefits.

4. Republicans - Look Tea Parties aside the fact is the famous quote in this subject here is "Deficits Don't Matter". That was said by our former vice President at a time when the economy appeared to be at full employment and we were not yet at war, not in the depths of a major recession. While today they seek any rhetorical attack on Obama that seems to stick, the fact is the lesson they learned from the 1980's is promise big tax cuts, if you need liberal Congresspersons to go along with you buy them with spending, and all will work out fine. This 'lesson' has become even more esconced in the GOP has time has caused uncomfortable facts to be forgotten from the narrative (such as Reagan only cut taxes in his first two years, he spent the next six years selectively raising them).

kbanaian writes:

I think you're referring to a passage in Democracy In Deficit (Buchanan and Wagner):

As we noted, we are not Ricardians. Pure fiscal policy can exert effects on the behavior of the citizen-taxpayer. To this extent, our analysis is in agreement with the Keynesian. But we must go beyond this and see just how the choice situation is changed. The replacement of current tax financing by government borrowing has the effect of reducing the "perceived price" of governmental goods and services. This "relative price" change embodies an income effect of the orthodox Hicksian sort, and this income effect will generate some attempted increase in the rate of private spending. This is essentially the Keynesian result. Note that we need not require the total absence of a Ricardian recognition of future tax liabilities. Citizens-taxpayers may anticipate the future taxes that are implicit with government borrowing, but, in doing so, they need not value these at the extreme Ricardian limits. To the extent that the costs of governmental goods and services are perceived to be lowered by any degree through the substitution of debt for tax finance, the "relative-price" change will be present.
Boonton writes:

Perception of future taxes:

The problem that bothers me with this tenant of 'rational expectations' is that it doesn't seem very rational. Say the gov't is spending today. I predict tomorrow there will be tax increases for that spending, therefore I increase my savings today. The Keynesian spending is offset by individual savings. Why?

Since taxes are mostly based on income, why wouldn't I spend today on the assumption tomorrow I'll be retired with a low income. It will be 'tomorrows suckers' who will have higher income tax rates. Or suppose I suspect that the tax increases to pay for the spending will fall on consumption with a VAT tax. Again to the degree possible I should still spend today. Instead of, say, saving my money so I can pay the taxes on bottles of wine that will be imposed tomorrow, I can buy the wine today and stash it in my basement leaving the wine buyers of tomorrow as the suckers who will pay for the gov't spending.

Or maybe the idea is that I may not care about tomorrow's tax increases but my kids will. But since my income is, on average, greater than my kids their ability to offset the the spending today by saving more today is limited.

The idea just doesn't really make any sense.

Troy Camplin writes:

Even Ayn Rand said to cut everything first, then cut taxes.

David R. Henderson writes:

@kbanian,
Thanks. I think that's the quote. Funny how I misremembered it, though. He and Dick Wagner use the relative-price reasoning, but they use it get to an increase in private spending. I had remembered it as an increase in demand for government. So maybe I filled in my thinking on top of theirs.
Best,
David

Bob Murphy writes:

David, what exactly do you mean by saying tax rate cuts didn't increase revenues? Tax revenues did in fact go up (eventually), right?

So I assume you mean they didn't go up more than we would have predicted in the absence of the tax rate cut.

But then we get into a really complicated alternate universe. For example, what if the economy didn't grow nearly as much in the 1980s if Reagan had kept marginal income tax rates really high?

I am presumably biased since I worked for Arthur Laffer and saw dozens of his papers making (what appeared to be) very persuasive cases that the Laffer Curve is alive and well.

Boonton writes:

But then we get into a really complicated alternate universe. For example, what if the economy didn't grow nearly as much in the 1980s if Reagan had kept marginal income tax rates really high?

Curious, tax revenue per capita grew 19% from 1980 to 1988. But from 1972 to 1980 it grew 24%. Likewise from 1992 to 2000 it grew 41% (all inflation adjusted).

As was pointed out by 1992 we already had several Reagan and one Bush tax increase as well as one Clinton tax increase. So maybe if Reagan's tax cuts hadn't been so dramatic his era would have lagged behind the others even more.....but I'm not seeing good evidence here for supply side delusions.

Tracy W writes:

Boonton, that's the first degree of analysis - that you could retire first, or buy goods. But the analysis is at a whole-economy level. Consider the second level. Assume that not only you but lots of other people are trying to avoid the future tax. If lots of people retire then governments will probably have to look to retiree's incomes for some of the government revenue (be that by cutting state-provided pensions or raising taxes on your income from savings). If everyone spends up big now and then stashes the goods then governments will just have to raise VAT even further in the future and hit people when they buy non-durable goods, or raise income tax, or some other form of tax.

Some people might be able to avoid future tax increases to pay for government spending, but across the economy as a whole, they're not avoidable.

I'm not a full believer in Riccardian equivalence, but bringing in expectations does explain some things that the standard Keynesian account can't.

Boonton writes:

Whether the tax increases are unavoidable is not really the issue. The question is do individuals think they could avoid the tax increases, either because they will be dead by the time they arrive or because they can keep their future income low or dodge a consumption tax etc. If that's the case then spending increases today will not be offset by savings increases to offset them.

At least from surveys we know that people have wildly untrue ideas about gov't budgets. A huge portion believes, for example, that they will never see any social security even though the worse case project still leaves about 75% of benefits intact. They believe Republican Presidents cut spending and Democratic ones increase it yet the opposite happened with Bush and Clinton. They think things like discretionary spending, foreign aid, 'illegal immigrants' consume a huge portion of gov't spending.

Unlike other areas, there's no market feedback here. If I believed Pets.com was the greatest thing ever created I would be punished by the market for sinking all my savings into it. If I believed computers would go out of fashion and trained to be a typewriter repairman my income would suffer. Yet there's no real feedback for failing to properly project gov't taxes 30 years out. Whether you project well or not you're dead or dying in a nursing home. You don't get to give it another round.

Tracy W writes:

Boonton - this may be so for a 30-year time horizon, but I understand the Keynesian case for fiscal stimulus is about smoothing out fluctuations in the business cycle, so governments spend more in the recession years and less in the growth years. Business cycles are around 3-10 years, not 30 years, which makes the feedback cycle tighter. Furthermore a lot of money in the market is managed by professionals, who do pay attention to government spending.

You are right in saying that if people think they can dodge a future tax increase then they will likely act differently to what Riccardian Equivalence would expect, and also if people think that future tax increases/spending cuts will be larger than expected, they will also act differently to what Riccardian Equivalence would suggest.

Boonton writes:

Head is spinning a bit here :) IMO it isn't plausible. Money managers may pay attention to gov't spending (although the last two years or so has really shaken the idea that big money is all knowing) but they don't really have to worry about taxes except their own. The capital gains they generate are paid by the firms investors or they aren't paid at all for a long time as many mutual funds are held in 401Ks, IRAs or pension funds.

They Keynesian cycle does supposedly run 7-10 years with increased spending in the recession and decreased in the boom. It would only make sense to radically adjust your savings to offset the stimulus, though, if the gov't policy was to not only reduce spending in the boom but pay off the entire cost of stimulus spending. Taxpayers would also have to filter out non-stimulus changes to spending (such as the Iraq War which is something like $1T last I checked) and entitlement changes....or should I say filter that in.

Also the rational taxpayer has to filter out quack theories (such as the Reagan tax cuts generated a huge boom of revenue), which seem to infect even very smart people at think tanks and elsewhere and make forecasts of gov't spending AND political changes going out decades. Or the equity premium idea that stocks should *always* return more than bonds after adjusting for risk.....even though economic theory has no clear case for why that would be so and the recent downturn is starting to make that look like a post-WWII historical fluke. Would a very serious student of economic affairs in 1950 have predicted the decline of unions? The decline of budget balancing as a maxim of the Republican Party? Going off gold for the US dollar?

We may be rational but we spend most of our lives in one state of delusion to one degree or another. For this to work you need superhuman rationality and intelligence.

The problem with Riccardian Equivalence is that it seems like two arrows hitting each other spot on in mid-air. It works for a simple model of the economy but has no real application to the real world.

Tracy W writes:

Boonton - Forecasts need to be perfect to have an impact now. For example, imagine someone in 1950 who mistakenly was certain that the world would be destroyed by a nuclear war by 1960 at the latest and thus didn't save any money for their future. In 1960, that hypothetical person would have had no savings, even though the world was still around. Equally, a person who wrongly believes that there is going to be no public pension available when they retire and thus over-saves will still likely wind up with heaps of savings when they retire (unless of course they do something like invest in a fund run by Madoff).

That is good point about money managers not being normally directly affected by taxes, I was thinking of when I was working for the New Zealand Treasury and we did have financial types deeply interested in things like budget releases, tax rates, the long-term budget projects and tax revenue releases, I can't swear that any of them were market managers, and the ones I can specifically remember were mostly accountants and bankers.

I agree with you that it is massively unlikely that people will be so perfect in their forecasts that they exactly offset fiscal stimulus, and this is one of the reasons why I am not a full Riccardian. But, if people sometimes overshoot or undershoot in their adjustments to fiscal stimulus, that means that fiscal stimulus has a chance of actually worsening the current economic situation, as people over-save. How much a chance depends on how bad people are at forecasting, and which way you think they are biased overall. If people are unbiased bad forecasters then there's a 50% chance that fiscal stimulus would worsen the economic outlook, if people are trying to be Riccardian equivalent.

I also agree with you that if Riccardian Equivalence is right then people will adjust to debt-funded increases in government spending for things like the Iraq War. I think you make an important point, I've noticed a few Keynesians talking about how silly it is to go for fiscal austerity when health and pensions spending swamps the fiscal stimulus, they miss the point that you notice that what matters is the overall fiscal deficit, and if that can be cut by cutting non-stimulus programmes (like the Iraq War or entitlement programmes), then that counts as fiscal austerity. As a general rule, I think governments should cut the least-valuable spending first, not the spending that is labeled stimulus (of course this leaves open the whole question of what is the least-valuable spending).

Barro discusses some related criticisms of Riccardian Equivalence and some of the real world evidence about it in a paper at http://www.ukzn.ac.za/economics/viegi/teaching/uct/barro.pdf.

Boonton writes:

Tracey

How much a chance depends on how bad people are at forecasting, and which way you think they are biased overall. If people are unbiased bad forecasters then there's a 50% chance that fiscal stimulus would worsen the economic outlook, if people are trying to be Riccardian equivalent.

Actually its a lot less than 50%.

1. Stimulus offsetting doesn't require you to offset dollar for dollar. Use 22% as the portion of income taken by Federal taxes. $100B in stimulus in times of unemployment automatically costs only $78B simply due to the fact that the $100B earned by those receiving the stimulus have to pay $22B back in taxes. This is NOT taking into account any additional income due to the multiplier, which would normally be higher in a deep recession. If people save $78B to offset future expected tax increases, tax revenue falls $17.16B (22% of the lower income). So with perfect information consumers will not perfectly offset stimulus but under-offset it.

2. Of course the multiplier reduces the need to save. Some or even most of the stimulus will be paid back by additional income earned that people wouldn't have otherwise earned.

3. Consider that there will probably be taxpayers tomorrow who aren't around yet today. They will most likey be paying a share of the stimulus back meaning your 'share' will have to be lower. (This assumes positive population growth).

This means that even perfectly rational and perfectly knowledgeable consumers would have an incentive to under target a stimulus offset attempt. Of course this also means imperfect information and estimating means there's a 50-50 chance of coming in under or over that target (which is rationally less than the stimulus). That means the odds of overtargetting a stimulus and making things worse is much less than 50-50.

4. Money printing: There's a chance that the gov't will pay off the stimulus by printing money. In that case the best hedge against inflation is real assets. in other words build a house, start a business etc. Now what is the probability you assign inflation as a strategy to pay for the sitmulus? 20%? Then you should put 20% of your 'offset savings' into current investment consumption. That, of course, doesn't counter the stimulus but stimulates it by reinforcing it!

I think you make an important point, I've noticed a few Keynesians talking about how silly it is to go for fiscal austerity when health and pensions spending swamps the fiscal stimulus, they miss the point that you notice that what matters is the overall fiscal deficit, and if that can be cut by cutting non-stimulus programmes

The problem with this thinking is that it's like collecting ice cubes in the winter to fight the summer. If the projections of fiscal diaster are correct then the problem with health and pension programs tomorrow is in the multiple trillion dollar range. Not fighting the Iraq War, cutting back $500B of stimulus, is nothing compared to the coming storm. Krugman is right, the only place to fight tomorrows deficits is health spending tomorrow. Dithering over stimulus today does nothing....although it is of policital advantage since people are totally ignorant of the numbers. Strong arming a $50B unemployment fund today may seem like bold action on the deficit if you're not aware of the scope of things. Saying you want the $50B today but would support modifying cost increases to lower spending growth $500B over ten years sounds like a cop out, even though that's a much more sane economic policy.

I'll try to read the paper later, thank you for the link..

Tracy W writes:

1. I don't follow your figures.

Isn't the point of fiscal stimulus from the Keynesian persepctive not to increase the tax take, but to increase economic activity? So the government spends $78 billion net. Across the whole economy, with perfect Riccardian equivalence, perfect forecasting people would thus increase their savings by $78 billion, in order to pay the future taxes they expect. So net consumption is the same as before the increase in government spending and we don't get out of the recession. I don't see how taxes comes into the analysis of the effect on the economy.

2. Only if you think there is a positive multipler. Riccardian Equivalence is one way that there might not be any multipler. Assuming that Riccardian Equivalence is false doesn't help in arguments against Riccardian Equivalence.

3. This depends on where these future taxpayers come from. If they're the children of existing tax payers, and we assume that parents tend to care about their children's fiscal futures, or it comes from future immigrants who are aware of the policies of the future country they are immigrating to (and I am an immigrant to the UK, who is thinking of heading home as taxes rise here) then the offsetting effect would continue to apply, even in the long-run.

4. This is starting to sound like a monetarist argument. If fear of future inflation drives investment spending and thus current economic activity, why not just introduce the inflation and skip the fiscal stimulus bit? (My own conclusion on business cycle theories, at the end of the my economics degree, was that it seemed reasonable that different recessions and recoveries could be explained by different theories).

Returning to Riccardian Equivalence, what do the people who are hired to build the house and provide business services do with the money?

5. I don't see where you disagree with me on cutting government spending. I have no problems agreeing with Krugman or you that the growth in government healthcare spending needs to be cut (above-inflation increases in government health spending are a world-wide occurrence, not just a US one). For me, dollar for dollar cutting the growth in healthcare spending would be just as effective at cutting fiscal deficits as cutting spending that is labelled "fiscal stimulus", and as you say there are a lot more dollars to be cut in health care spending than in the spending that's labelled "fiscal stimulus". As far as I can tell the reason that governments don't cut health care spending is out of fear of what the voting public will do to them, not because they think that the fiscal deficit is only about spending that is labelled "fiscal stimulus".

Boonton writes:

Points 1-3 are pretty good, I'll let them stand.

#4, Yes stimulus comes in two flavors, monetary and fiscal and there's pros and cons to both. But we are talking about how the gov't will pay off the debt it incurred when it did a fiscal stimulus. If you think gov't will pay for stimulus with taxes then a surefire way to be ok is to simply save an equal amount in the form of bonds.

Gov't borrows $100B to do a stimulus policy. You earn $100B as a contractor to the gov't. Knowing taxes will be raised, you buy $100B in bonds. Ten years from now gov't needs $110B to pay off its debt. Your taxes go up $110B, but since you cash in your bonds for $110B you're position nets to zero. Stimulus was countered by you in the form of savings and anti-stimulus (raising taxes to pay off the debt) was countered by you in the form of dissavings.

But what if gov't decides to just print money to pay it off causing 20% inflation. You cash in your bonds for $110B in cash but your 'tax' is a 20% increase in all prices. You've failed to keep your position neutral.

What if you had known that gov't would print money? Then instead of buying bonds you would have brought something more tied to inflation (let's say gold or real assets like business investment).

So there's two possible 'taxes' to pay the debt. Regular taxes and inflation. However the strategies to counter the stimulus do not overlap. Provided the rational person considers the odds of the inflationary tactic to be non-zero, they will split their 'counter stimulus' portfolio between 'anti-stimulus' orientated savings and 'pro-stimulus' inflation hedges.

Returning to Riccardian Equivalence, what do the people who are hired to build the house and provide business services do with the money?

Again it would seem to depend on what they think is coming. If its inflation they would seek to preserve their wealth by buying real assets. If they think its future taxes they would seek to save. If we are assuming everyone is rational for this exercise it still holds, the 'counter stimulus' target will be less than the actual fiscal stimulus.

For me, dollar for dollar cutting the growth in healthcare spending would be just as effective at cutting fiscal deficits as cutting spending that is labelled "fiscal stimulus", and as you say there are a lot more dollars to be cut in health care spending than in the spending that's labelled "fiscal stimulus".

Cutting a dollar is cutting a dollar. Deficit hawks, though, aren't complaining about billions of dollars in debt, not even trillions. But tens of trillions and just about all of those dollars live not here in the present or in past deficits but hypothetical future deficits. It's not like $1 of today's spend isn't $1 present value of tomorrows spend. It's that $1 doesn't matter compared to the ocean and almost all of the ocean isn't sitting in the world of today but the world of tomorrow. It may sound like a cop out to say 'cut the deficit tomorrow' but that's the only place to cut it (the bulk of it at least). Even if you zeroed out the deficit today AND paid off the entire national debt today you'd still have a sustainability problem tomorrow if forecasts are correct.

Boonton writes:

3. This depends on where these future taxpayers come from. If they're the children of existing tax payers, and we assume that parents tend to care about their children's fiscal futures, or it comes from future immigrants who are aware of the policies of the future country they are immigrating to (and I am an immigrant to the UK, who is thinking of heading home as taxes rise here) then the offsetting effect would continue to apply, even in the long-run

True but notice what you're doing in your 'rational' way. You'll enjoy UK stimulus today but duck out tomorrow when it comes time to pay the taxes. Hence your rational strategy is NOT to save for future expected UK taxes. The model IMO only works where the tax system is some type of univeral head tax that is relatively unavoidable. But income and even VAT taxes are avoidable.

As I pointed out, given the possibility of inflation as a type of tax to pay the debt there are two strategies the individual may use to dodge paying their share of the taxes and offload onto others. For example, you may stash your savings into a Roth IRA thereby you will be immune from income taxes in the future. But what if its a consumption tax? Then you're screwed. If its a wealth tax then your best bet to 'save' would be to put resources into extending your income earning ability rather than amassing wealth (say training for a career that let's you work into old age with ease). The optimal strategy here is to list all the possible ways the gov't may tax to pay the debt off, assign probabilities and then list optimal strategies to counter each possible taxing method and split your portfolio accordingly. As we have seen some strategies are actually pro-stimulative (such as consuming today to counter future consumption taxes or investing today to counter future inflation).

The end point is I think at best you have a drag on stimulus but not enough to overcome it.

2. Only if you think there is a positive multipler. Riccardian Equivalence is one way that there might not be any multipler. Assuming that Riccardian Equivalence is false doesn't help in arguments against Riccardian Equivalence.

It doesn't assume its false (I take back giving you this point BTW). Riccardian Equivalance is assuming the rational citizen will alter his behavior today in the face of stimulus to account for the fact that tomorrow the gov't will ask him to 'pay back' the stimulus. How does he do this? Well to keep it simple if the gov't borrows $100B today and has to pay $110B in ten years he knows $110B needs to be provided in ten years. If he assumes a flat income tax he must divide $110B by total income in ten years to find the tax rate he will face. But if a multiplier is at play then total income in ten years is higher. Tomorrow's tax of $110B will not require as much savings for him today if a multiplier is positive as it would if its zero.

Boonton writes:

The more I think about it, the more flawed it seems to be. The premise here is that fiscal stimulus would work except for the 'rational' taxpayers who see future taxes coming down the line and opt to save the 'stimulus' they receive from the gov't (see the previous hypothetical about getting $100B from the gov't and simply buying the $100B bond in order to pay the taxes in the future).

But this is only rational if there's a single taxpayer. There isn't. If taxpayers don't save, the stimulus works and the multiplier multiplies and the future state enjoys higher income due to the fact that today capacity that is being unutilized (aka unemployment) gets utilized.

The question then is what does everyone else think? If everyone else saves then so should you, if not then the stimulus works and the tax burden in the future will be much smaller. Given that most people overestimate their financial intelligence, given that the political system leaves a lot of room for tax loopholes for the experienced, its reasonable to expect that a lot of other people will believe they can dodge future tax increases to one degree or another. As those other people will then opt to spend rather than save, stimulus becomes more effective and the case for offsetting becomes less rational.

Riccardian Equivalence seems to require a lot of intense game theory analysis.

Tracy W writes:

Again it would seem to depend on what they think is coming. If its inflation they would seek to preserve their wealth by buying real assets.

But for you to buy a real asset someone has to sell it. The money doesn't disappear. If I decide to hedge against inflation by buying £4000 worth of gold, someone else has £4000 in cash to deal with. I've hedged myself against inflation, but the seller has equally unhedged themselves.

You'll enjoy UK stimulus today but duck out tomorrow when it comes time to pay the taxes. Hence your rational strategy is NOT to save for future expected UK taxes.

Which means that perfectly rational people who intend to stay in the UK should save more to offset the disapparence of the footloose like me, along with needing to offset lost earnings from the marginal potential immigrants who are discouraged by the higher taxes. (Perhaps these people are intending to stay in the UK because, unlike me, they don't have passports entitling them to live somewhere with a much lower government debt).

The optimal strategy here is to list all the possible ways the gov't may tax to pay the debt off, assign probabilities and then list optimal strategies to counter each possible taxing method and split your portfolio accordingly.

And if everyone redistributes their taxes optimally, then the government will have to raise future taxes in such a way so that the total raised, taking into account everyone's optimal tax planning policies, is the same as what they would have raised if no one had engaged in any tax-optimising at all.

The question then is what does everyone else think? If everyone else saves then so should you, if not then the stimulus works and the tax burden in the future will be much smaller.

How is the tax burden in the future smaller? What is your counter-factual? My understanding of Keynesian stimulus is that if it works it's about utilising slack resources and returning the economy to full employment sooner than would have otherwise happened, it's not about increasing the long-run growth rate. Economies got themselves out of recession before Keynes was born. (If governments can increase the long-term growth rate by making various investments that only it can adequately do, such as public health programmes, or providing the rule of law, then it will be making those invetments that regardless of whether there's a recession on or not).

As for your argument that people will be biased towards under-saving, what happened to your earlier argument that "a huge proportion believes, for example, that they will never see any social security" which implies to me an offsetting bias towards savings?

I agree with most of what you say about deficits, except I don't know what you mean when you say that "Deficit hawks, though, aren't complaining about billions of dollars in debt, not even trillions." The deficit hawks I know of complain greatly about future healthcare spending and the coming baby-boomer retirement. See for example this post by Arnold Kling on his plans for reforming future entitlements.

Boonton writes:

But for you to buy a real asset someone has to sell it. The money doesn't disappear. If I decide to hedge against inflation by buying £4000 worth of gold, someone else has £4000 in cash to deal with. I've hedged myself against inflation, but the seller has equally unhedged themselves.

True but we hold all else to be equal. Your decision to buy gold was motivated by an attempt to hedge gov't stimulus. It's only offset if someone was equally motivated to sell gold to hedge a gov't stimulus. Otherwise you just brought the gold from someone who was selling anyway.

Same thing works with savings. Your decision to save $4000 in a bank account is matched by someone else who decides to spend $4000. All else being equal your decision to save is counter-stimulative. But if the other person opted to spend (either by borrowing or withdrawing their own savings) because they felt stimulus would have the opposite effect then Riccardian Equilvalence falters.

Which means that perfectly rational people who intend to stay in the UK should save more to offset the disapparence of the footloose like me, along with needing to offset lost earnings from the marginal potential immigrants who are discouraged by the higher taxes.

Or perhaps they are counting on suckers like you ;) After all while you're in the UK you're paying taxes to pay off debts incurred during WWII, their unfortunate experiments with nationalized industries, the Falklan Island War etc. UK residents in the 50's through 90's didn't have to save quite as much as your arrival offset some of what would have been their taxes.

And if everyone redistributes their taxes optimally, then the government will have to raise future taxes in such a way so that the total raised, taking into account everyone's optimal tax planning policies, is the same as what they would have raised if no one had engaged in any tax-optimising at all.

Not quite, the tax structure is relatively independent of the portfolios. You're just taking a bet. For example you are betting that gov't will not inflate the debt away. If gov't does inflate the entire debt away then 100% of your planning would optimally be inflation hedges. You don't know though, but you assign a probability of 20% to inflation, say. But that doesn't mean gov't must inflate only 20% away. They could do 0% or 100%.

How is the tax burden in the future smaller? What is your counter-factual? My understanding of Keynesian stimulus is that if it works it's about utilising slack resources and returning the economy to full employment sooner than would have otherwise happened, it's not about increasing the long-run growth rate.

There's no long run cost to a prolonged period of under utilized resources? Hardly. Long term unemployment destroys human capital. Under utilized capital assets rot away. Three years at 10% unemployment lowers long term income relative to six months.

As for your argument that people will be biased towards under-saving, what happened to your earlier argument that "a huge proportion believes, for example, that they will never see any social security" which implies to me an offsetting bias towards savings?

Economists correctly consider actions louder than words. People say they think social security is less reliable than it was for their parents yet do people act like it? Are they maxing out their 401K's and increasing savings?

Anyway, though, we are talking all things being equal. The gov't does a $100B stimulus how will a world of rational consumers react? If there is only one person and one way to tax I agree it might be rational for him to directly save the $100B stimulus. But we have

1. Multiple people which means the rational person must consider other people's behavior. If other people don't save, stimulus works which means today's income is higher which means if you're saving a % of your income you will automatically have more savings just through higher income. Looking at other people I don't trust they are avid finance watchers who will leap to rachet up savings. Consider that during the Bush years there was no clear increase in savings despite the numerous tax preferences for savings (401Ks, IRAs etc) and the rapid increase in spending during a period of full employment. Spending in full employment yields no increase in income so rational people would have assumed such a policy would result in nothing but future tax increases.

1.1 This is just observing that Riccardian Equilvalence argues that all stimulus will be countered by savings. If people don't do that, then stimulus isn't countered and income today is higher. Is it rational for a rational person to conclude 100% of the US population is rational and will correctly lower their consumption to offset a stimulus? If the rational person assumes many other people will not correctly offset a stimulus, and most other people tend to bias their decisions in favor of more present day consumption, then it becomes rational for him not to fully try to offset the sitmulus with savings.

2. Multiple taxing methods, some of which are pro-stimulative. Unless people can know which method will be used, it will be rational for them to assign at least a portion of their stimulus-tax savings in vehicles that are stimulative.


The deficit hawks I know of complain greatly about future healthcare spending and the coming baby-boomer retirement. See for example this post by Arnold Kling on his plans for reforming future entitlements.

Pretend for a moment that tomorrow you will take on Donald's Trumps debt in 2013. Does it really make sense to worry about your $2,000 Home Depot card balance today? No it doesn't, in 2013 you will have a huge amount of debt and your only rational way to avoid insolvency is to either massively increase your ability to service debt in 2013 or avoid taking on that massive amount of debt. Paying off your Home Depot Card today won't even amount to a rounding error. The deficit hawks have put forth future debt in the range of tens of trillions of dollars. Eliminating today's deficit, even paying off all of today's debt does nothing to counter that. The problem lies in the future, not the present. Serious deficit hawks therefore should care much more about supporting things like the health reforms limits on Medicare growth than opposing short term stimulus bills.

Tracy W writes:

Boonton - the driving assumption behind Riccardian Equivalence is that all people want to hedge against the future government taxes. So if I buy gold, the person I buy the gold from has to figure out some way to hedge themselves against inflation.

Your decision to save $4000 in a bank account is matched by someone else who decides to spend $4000.

Like, for example, the government doing its fiscal stimulus, if say the bank puts my money into government bonds.

But if the other person opted to spend (either by borrowing or withdrawing their own savings) because they felt stimulus would have the opposite effect then Riccardian Equilvalence falters.

Yes, of course, Riccardian Equivalence only holds if people do set to set aside money to offset the future taxes. What's your point here?

Or perhaps they are counting on suckers like you ;) ...UK residents in the 50's through 90's didn't have to save quite as much as your arrival offset some of what would have been their taxes.

On the other hand, higher taxes in the UK as the result of past goverment borrowing would deter the marginal immigrant, thus meaning that UK residents had to pay a larger share of taxes to pay the debts left over from WWII, their unfortunate experiments with nationalised industries, the Falkan Island War, etc, etc, than if the UK hadn't indulged in its spending. If Riccardian Equivalence holds then the perfectly rational UK citizen would take that into account.

You're just taking a bet. For example you are betting that gov't will not inflate the debt away. If gov't does inflate the entire debt away then 100% of your planning would optimally be inflation hedges.

But not everyone in the economy can hedge against inflation. I can buy gold as a hedge against inflation but that means that the seller of the gold is unhedged. I'm sorry, but I don't know how to make this point any clearer. Some people can hedge against inflation, but that's at the price of leaving some other people unhedged.

There's no long run cost to a prolonged period of under utilized resources?

And taxes, by driving a wedge between people's sell price and people's buy price, led to under-utilised resources. So higher taxes to pay for fiscal stimulus mean slightly higher unemployment over the long-term, all else being equal, and under-utilised capital, or capital that doesn't get built.

The gov't does a $100B stimulus how will a world of rational consumers react? ... Looking at other people I don't trust they are avid finance watchers who will leap to rachet up savings.

You are changing definitions here midstream. Riccardian Equivalence is about how rational people will act. You start off with the Riccardian assumption, and then you switch to an argument based on people who are not avid finance watchers. I think the idea that people aren't perfectly rational is a good argument against Riccardian equivalence, as indeed David Ricardo originally felt too. But if you want to understand the basic Riccardian argument then you need to understand it within its assumptions. (This is quite different from forming an opinion of whether the argument is right or not, eg I can understand Kirchoff's Laws in electricity even though I know that the laws are wrong and that Maxwell's Equations are a better explanation in some situations).

2. Unless people can know which method will be used, it will be rational for them to assign at least a portion of their stimulus-tax savings in vehicles that are stimulative.

And what do the people who get the money from those investments do with it?

Consider that during the Bush years there was no clear increase in savings despite the numerous tax preferences for savings (401Ks, IRAs etc) and the rapid increase in spending during a period of full employment.

This is one of the reasons why I am not a full believer in Riccardian Equivalence either.

You appear to be pursuing two threads here. 1) Trying to understand Riccardian Equivalence in its own turns. 2) Forming a view on whether it's right or not as a description of the real world. Doing both together has its advantages, but it does get confusing.

Does it really make sense to worry about your $2,000 Home Depot card balance today?

Having worked for the NZ Treasury, I know what government spending is like. To keep government spending under control you have to spend a lot of time arguing about the little bits and pieces, otherwise they overwhelm the whole. What's the joke, "a billion here, a billion there, before you know it you're talking real money?"

Unless you think that a government is going to get out of financing health care entirely as part of the reforms, then we do need to care about the "little" bits of money when reforming spending. Otherwise all the little bits quickly add up to big bits.

To put it into your analogy, if I am going to take on Donald Trump's debt in 2013, then to avoid insolvency I have to massively increase my ability to service that debt, and for that to happen I have to pay attention to lots of little details (say I start a business. Businesses are about details.). Alternatively, I can seek to minimise the amount of Donald Trump's debt I take on, which would mean lots of arguing with his lawyers about the details.

Boonton writes:

Tracy W

Let me try to make my argument more condensed and coherent.

1. A rational person will ask how other people will behave. If, for whatever reason, you think other people will not save to offset a stimulus, then the argument for you to do the same becomes less. Your 'rational' choice is not to save as much. Other rational people have the same incentive. The offset only works if everyone saves an amount equal to the stimulus.

2. A rational person has to ask how the gov't will pay for the stimulus in the future. Consider the following alternatives and rational strategies:

a. Inflation -- Strategy buy assets like gold, real business investment, real estate etc. Borrow money if you can (at fixed interest) as you'll pay it back in inflated dollars. Note that this does not offset stimulus but increases it.


b. Default --- Do not buy gov't bonds keep your money out of institutions whose health depends on gov't bonds paying back. In this case the 'suckers' who pay for the stimulus are the ones who buy the bonds! This again doesn't offset stimulus.

c. Income taxes -- Try to earn extra income now, retire later when the taxes hit. Extra traditional savings now, 401K, pay off your debts, etc. This does offset stimulus.

Stimulus can only be offset if the rational person puts the odds of c at 100% and offsets accordingly. If the person thinks a and b are greater than 0%, then it doesn't make sense to offset stimulus 100% with savings. You should hedge your bets that paying off the stimulus may not come from c.

3. If stimulus is not offset, then the multiplier works to one degree or another. If the multiplier works then both immediate income goes up and long run expected income is higher. More income makes future debt service easier which again lessens the need for the rational person to save for future tax increases.


Finally:

What's the joke, "a billion here, a billion there, before you know it you're talking real money?"

Except your not. Kling & co. have argued themselves into a box. They've told us a Noah size flood is coming tomorrow. If that's true water proofing your basement today is nothing more than a joke. If the flood isn't coming tomorrow then their predictions of fiscal diaster evaporate. $1T in stimulus today is large but really no big deal over the long term. In other words, Krugman is right. Spend today, address future deficits today (thru entitlement reform that goes into effect years from now). Cutting 'a billion here billion there' makes no sense macroeconomically. It only makes sense to keep control of the efficiency of gov't spending.

Also to bring it back, 'starving the beast' shouldn't work if Riccardian Equivalence is true. If its true then previous gov't deficits have already motivated people to amass piggy banks of money to pay them off.

Tracy W writes:

If, for whatever reason, you think other people will not save to offset a stimulus, then the argument for you to do the same becomes less. Your 'rational' choice is not to save as much.

I don't follow this. Let's say I have worked out my optimal consumption profile and then a recession comes along, the government spends $78 billion in fiscal stimulus, and I expect my share of the additional taxes to pay back the $78 billion to be $1,000. Why would it suddenly become less rational for me to not save as much, if everyone else isn't saving? I'm smoothing out my income. Who cares what they are doing? (If the government is spending the money on something that raises the long-term growth rate of the economy more than the cost of the taxes that's different of course, but they should be doing that regardless of whether there's a recession on).

2. a. Inflation - as I've said before, if you buy assets like gold as hedges against inflation, then someone has to sell those assets, so they have unhedged themselves equally. Borrowing money means that the lenders need to find some way to hedge themselves.

b. So you don't buy government bonds, I don't see what difference this makes to your future tax burden.

c. Income taxes - rational for one person, but if you reduce your future tax burden that way then someone else's future tax burden has to increase, and if everyone is rational then the people's whose tax burden will increase will realise this and increase their savings.

What you are missing is an equilibrium analysis. You keep thinking about this as the case of what an individual can do, but just because one individual can do something doesn't mean they all can. You need to think about what everyone could do, including potential future participants, taking into account the effect that other poeple's actions have on them (like that buying gold requires someone to sell gold).

3. and long run expected income is higher...

Only if long-run expected income is higher. As I stated before, the higher taxes to pay for fiscal stimulus mean higher dead-weight losses, reducing overall income.

There's a basic problem with the multipler argument. It is rather hard to believe that our prosperity can be improved in the long-run just by the government increasing spending, independently of what the money is spent on. Your sudden assertion that government spending raises long-run expected income doesn't convince me that it does.

They've told us a Noah size flood is coming tomorrow. If that's true water proofing your basement today is nothing more than a joke.

The analogy does not hold. Flood-proofing your basement may not be any good against a Noah-size flood, but if you're going to cut down government spending you need to worry about the details.

As for the remainder, did you read my observation at all before writing this? Entitlement reform is necessary, but unless you think that entitlement reform is going to mean eliminating all pensions and healthcare spending (and even Ruth Richardson in New Zealand didn't manage to do that), it's going to still leave massive government spending on healthcare and pensions, which means we need to worry about things like value-for-money of fiscal stimulus. You can make up all the analogies you like in an attempt to dismiss this, but that doesn't mean that the basic problem is going to go away. An extra $1 trillion (or whatever the figure is) in fiscal stimulus now makes the entitlement reforms even harder to do. It's one thing to say that the fiscal stimulus is so important that we should do it anyway, and cut other forms of government spending to pay for it, but it's silly to say that we don't have to worry about it at all.

And by the way, the long-term? It's 2010. That's 65 years from the end of WWII. The babyboomers are starting to retire now. The long-term is the next few years now. In NZ, those neoliberals that Krugman despises put in place the long-term reforms in the 1980s and early 1990s (eg full-funding of ACC, closure of the government employees' pension scheme to new members). It's rather late to be doing that now in the USA.

Comments for this entry have been closed
Return to top