Harvard’s Robert Barro, who is good at this stuff, estimates (in round numbers) the effects of last year’s stimulus package (numbers represent billions of dollars):
The executive summary is that income (that is, the total income of all Americans) rises in 2009 and 2010 (while the stimulus money is being spent), and continues a bit higher for another year after that, but falls in later years (when the taxes, with their accompanying disincentive effects, come due). (Of course, the day of reckoning can be delayed, but not forever—so the arithmetic still rules). Adding up over five years, income falls by $300 billion, or about $1000 per American.
These numbers confirm my prejudice that the stimulus package is a bad idea, but they still make me uncomfortable. Let me first add a few remarks about what the numbers mean and then I’ll tell you what I don’t quite get about them.
Interpreting the numbers. The first thing to observe is that spending by one party generates income for another. That’s why the spending changes (in columns 1 and 2) add up to the income change (in column 3). Barro makes a few further observations:
- The chart implicitly assumes that all stimulus money is spent wisely; it assumes this by not accounting for losses due to wasteful spending. Correcting for this would mean replacing the blue numbers (actual government spending) with lower numbers that represent the value of what gets bought, and would make the stimulus look like an even worse deal.
- Christy Romer, the head of President Obama’s Council of Economic Advisors, is famous among economists for research suggesting that the green numbers should be substantially more negative than Professor Barro’s own estimates. If we replace Barro’s estimates with Romer’s, the stimulus looks like an even worse deal.
- Romer also claims the red numbers should be much less negative, or even positive. This would make the stimulus look like a better deal. But Professor Barro claims that in contrast to the green numbers, which she’s written about extensively, Romer has never given a substantial justification for her guesses about the red numbers.
- Barro doesn’t mention this, but it’s worth pointing out that if we’re earning less in 2012 and 2013, that means we’ll also be working less. Accounting for the value of that additional leisure would make the stimulus look like a somewhat better deal.
What I’m leery of. Here’s what bothers me: According to this chart, in the years 2009 and 2010, household incomes rise but household spending falls. Why would people spend less when their incomes rise? The usual answer is that government spending causes the interest rate to rise, and that’s what chokes off private spending. But here we’re talking about a period when interest rates were near zero. If the interest rate doesn’t choke off spending, what does?
One possible answer is that interest rates for household borrowing haven’t really been that low. Maybe that’s the whole answer, but I’d like to see this addressed. (Or you could reconcile this if consumption—and/or investment—is extremely sensitive to small interest rate changes, but I think you’d need implausibly much sensitivity for this to be the whole story.)
Another partial answer is that people spend less because they’re squirreling away income in anticipation of future tax hikes. But that can’t account for numbers of the magnitude we’re seeing in this chart.
Or maybe the argument is that people can afford to spend less when the government does their shopping for them. If President Obama buys you a bowl of corn flakes every morning, you’ll stop buying corn flakes for yourself. But this works only if the stimulus money is spent on stuff people were going to buy anyway (or on good substitutes for that stuff). Is that what Barro is assuming? And if so, on what basis? I’d love to see the details of this.
Those are my questions. I have substantial confidence that Professor Barro (who has thought about this stuff at least 3000 times harder than I have) has good answers. But from his Wall Street Journal column, I can’t tell what those answers are.
Some thoughts,
1. Even if the nominal interest rate is 0, the real interest rate could be somewhat higher.
2. If the risk free real interest rate is 0 and if this is a considerably more risky environment, few people will be able to find the 0 interest rate. Some will find it hard to get loans at all.
With the FED paying interest on reserves, what is the point in a bank lending more money in a risky environment. My personal experience has been that credit is being held stable or contracted rather than expanding.
The multiplyer used was based on wartime defense spending, and he says it is probably generous as it neglects increased productivity due to patriotism. Perhaps he is correct that if the Govt in 2004 had suddenly announced a huge fiscal stimulus, it would result in these figures, i.e. the multiplier is correct. However, he says that he cannot understand the use of a multiplier above 1 by the Govt. economists. Surely, these are extraordinary times, and the estimates based on normal years do not apply over this peroid. I presume that the greater than 1 multiplier derives from estimates of “saving” industries that would otherwise close due to the short term effects of recession / depression. This large multiplier would only be accurate for this short period of time in these circumstances. Is in necessary that the multiplier must always be the same?
You make some interesting observations — and a lot of assumptions.
The most major is what might have happened without the stimulus? Of course it’s virtually impossible to demonstrate that, but it’s not unreasonable to expect that the income figures would have gone down faster. That could easily have led to a deeper and more long-lasting recession/depression. Determining value requires a comparison and the logical comparison would be with greater spending or lesser spending.
It was predictable that a stimulus would be followed by a decline. It was also predictable that a decline would happen even without the stimulus. The only real question, then, is whether the decline would have been greater, less or the same if spending were adjusted. Economics, unfortunately, does not give us a laboratory in which we can repeat the experiment and see.
In public spending, there is also the issue of consumer confidence. It’s very possible that people’s income has stabilized and gone up, but they have cut back in spending because of uncertainty.
One example is alcohol — folklore suggests alcohol sales are stable in a recession, but recent news stories say that while volume has remained fairly constant, people are buying fewer premium brands. Spending continues, but at a lower rate.
Perhaps the stimulus is a bad deal, but we could be worse without it. It could be the real mistake is that we didn’t spend more.
Looking again, I see that I have sort of re-stated Steven’s concerns over the reduction in private spending. This reduction is defined by the multiplier, thus a multiplier of 0.4 means that private spending is reduced by 180 in year 1. If spending is not reduced by as much, then the multiplier is larger. I suppose it again comes down to the question, is the multiplier derived from wartime defense spending valid for peacetime banking crisis? If not, how do we calculate the multiplier?
I have no way to formalize the theory, but I believe people spend less because of uncertainty. That is separate than spending less on account of future tax hikes, as that is saving because of a certain expense, not uncertain ones.
With uncertainty, people want to increase their cash holdings more than they want new stuff. Not too hard a case to imagine, especially when many households are awash in several years stock of new stuff.
There is also some evidence that while interest rates are not too high, credit card debt is tighter, and, I think, as part of general zeitgeist and falling house prices you’d see much less HELOC activity going on, which also contributed to a great deal of consumer spending.
I am veering into the territory of animal spirits, but it isn’t that hard to imagine circumstances where, at any positive interest rate, many people may not want to take out loans just to buy shirts at the mall and Bluray players, instead preferring to put a little aside for a “rainy day.”
In a Mankiwian world this cannot be allowed, under any circumstances, so we must lend at negative interest or simulate it through punitive inflation on cash.
It does make one wonder how an entire profession can go mad, each practitioner (decently exemplified by Romer and Mankiw) and with their own perfect scheme to coerce people into behaving the way they think that they should. All for their own good, of course.
ThomasL, but why does Govt. spending cause the uncertainty?
Harold:
but why does Govt. spending cause the uncertainty?
The right question exactly. All of these numbers, remember, are estimates of the difference between the no-stimulus and the stimulus worlds, holding everything else constant.
Eh, don’t bother me (or yourselves) with all of this highfalutin’ economic analysis; the important thing is that we stimulate the economy (or be seen to be attempting to do so) in the short term, ie before the next congressional or even presidential elections. As long as the bill comes due afterward, who cares about it?
You wrote:
“The first thing to observe is that spending by one party generates income for another. That’s why the spending changes (in columns 1 and 2) add up to the income change (in column 3).”
but you also wrote:
“The chart implicitly assumes that all stimulus money is spent wisely; it assumes this by not accounting for losses due to wasteful spending. Correcting for this would mean replacing the blue numbers (actual government spending) with lower numbers that represent the value of what gets bought.”
Are you sure this second part is correct? As you said, spending by one party means income for another. That’s true even if the spending is “wasteful” spending. If the Defense Department buys a $9,000 toilet seat, that’s $9,000 extra income for somebody, somewhere.
It seems that the proper place to account for the “wastefulness” of the spending, is a fourth and fifth column to account for “added value” due to government spending and “added value” due to private spending. (Where these columns would contain negative numbers — “value lost” — when spending of either kind goes down.) In theory, if private sector spending does down by X, then added value should go down by more than X, because people are foregoing the opportunities to buy things that they valued by more than X (otherwise they wouldn’t have bought them). Whereas if government spending changes by X, “added value” can change by more or less than X, depending on how wasteful the spending is.
Harold,
Steve’s question was: “Why would people spend less when their incomes rise?”
My answer is that they don’t know what is coming down the road in the general sense of how stable is their job, their customer base, their credit, etc. Their uncertainty, I would argue, is greater than their income increase. Particular when the income increase is diffuse, difficult to calculate, and not always monetary.
That can easily become an argument for why fiscal stimulus won’t work, whether or not the govt. is the cause of the uncertainty.
Combine it with something like the permanent income hypothesis, and you have an answer for why a few hundred or a few thousand more dollars in income, one time, won’t move my spending much. Even a worker hired purely on stimulus money to, for example, put down new sod in the center median of a highway (yes, a real stimulus job) knows that is short term work. You can plan for next year on that. Better to put some of it at lest in piggy bank for down the road.
So, my case is that uncertainty is so high, the govt. does no need to introduce more in order for stimulus not to have much effect.
With that said, I do think the govt. can introduce a lot of uncertainty.
Should I buy a car now or wait to see if there is another cash for clunkers? A house now? No, the rebate expired. Oh, wait! it didn’t expire; I can still get one… but, hmm, the “second stimulus” that people are talking about may even have a better rebate… I do need a new washer, but, what’s it going to take to really get my energystar rebate? I know the new washer qualifies but is my old washer on the list? What if the Bush tax cuts do totally expire? I make less than $100k, but I still saw my taxes drop a lot with them after 2003. And Orzag just said that “everything is on the table” including hikes (above the lapses even!) on people making well less than 100k [head nod to Steve’s previous mention of taxes]. And this health care thing. They are still insisting they’ll pass something, but they don’t know what. Are my rates going to go up or down? Is my company drop insurance and I end up in some public plan? What about the HSA that I use for vitamins and chiropractor visits. My insurance already doesn’t cover that, and Congress has sworn they’re going to abolish HSAs. And what about the FSA my work has for dental? I have no idea what is happening to that, and I think my kid will need braces in a couple of years, but the dentist says they need to wait till he’s older. Bah, and my credit card fees just went from $0/y to $75/y because of the new credit card regulations, and now they charge interest from the day of purchase! I’ll just cancel the card and use my debit. I’ll have to economize a bit, but… $^!%^!^%^@! it all, my tires just jumped from $120/ea to $165 because of Obama’s anti-Chinese posturing. I guess I can get by for a few more months, but I’ll need them for sure before the end of the year. Maybe they’ll drop again… Maybe I should get that new car, you don’t need new tires on a tradein… but what if they do start another cash for clunkers and I just traded it in…
That should read, “You _can’t_ plan for next year…” not, “You can…”
Bennett Haselton: Yes, you’re right. My point is that the chart does not have a separate column for “added value”. It presumes that added value is already measured by income. If
you assume spending is wasteful, the income column does not change but the (implicit) “added value” column does.
Ah OK so when you said “replacing the blue numbers (actual government spending) with lower numbers that represent the value of what gets bought”, you meant replacing the numbers AND changing the definition of the column :) I thought you meant changing the numbers but still labeling it “Government spending”, which I assumed would be incorrect.
Bennett: You have got it exactly right. Sorry I was confusing.
What about the impact of people rebuilding their savings? Between the depreciation in home values, and loss in investments (the DJIA is about 30% below its 2007 peak), many people’s savings have been severely depleted. These people may have reduced spending to rebuild their savings for retirement. This would not be caused by government spending, but rather government spending is compensating for the loss of consumer spending.
Al V.: If it’s not caused by govt spending then it won’t show up in the chart. The chart shows only the *change* in spending that’s due to the stimulus package.
The numbers follow from Barro’s assumption of the government spending multiplier. But presumably, Barro is using an empirical estimate of this multiplier obtained from data over time periods where the sort of interest rate crowding out you describe did occur, even though it may not occur now with the Fed holding interest rates around zero. This calls into question is the appropriateness of Barro’s government spending multiplier in the current situation.
Steve identifies several good questions about Barro’s analysis, but since I’m not a libertarian, I’m not as inclined to give the benefit of the doubt to a Hoover Institute guy as Steve is. :)
Here are some others question I have:
* For the sake of argument let’s take Barro’s defense spending multiplier as a reasonable proxy for non-defense spending. Elsewhere, Barrow estimates…
http://www.economics.harvard.edu/faculty/barro/files/wsj_09_1001.pdf
…the defense multiplier at 0.7, increasing “by around 0.1 for each two percentage points by which the unemployment rate exceeds its long-run median of 5.6%.” So, taking an average unemployment rate of 10% for 2009/2010, the estimated multiplier reaches 0.9, instead of the 0.4 (2009) and 0.6 (2010) Barro uses in the article Steve cites here.
That’s one hell of a difference.
* Next, how reasonable is the assumption that an estimated defense spending multiplier is a good proxy for a non-defense spending multiplier? Barro glides right by this issue.
In fact, we (and he) don’t know that it IS a good proxy because no one has estimated a reliable non-defense multiplier, and it’s not for trying. Estimating a defense multiplier is easier, so that’s what he uses. He offers no evidence that it’s a good proxy for non-defense spending.
* While there are not reliable estimates of non-defense multipliers, there are some that are suggestive and are considerably higher than Barro’s (0.4 in 2009; 0.6 in 2010). Romer and Romer estimate a non-defense multiplier of 1.6, and Stevans and Sessions estimate a multiplier of 3.33.
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1424446
So, with Barro-revised (based on a 10% unemployment rate), Romer& Romer, and Stevans and Sessions, we have a range of estimates between 0.9 and 3.33.
Obviously, using any of these estimates would have a profound effect on Barro’s conclusions.
* Next, let’s step back from theory and ask a few practical questions: how does a multiplier of 0.4 – 0.6 make sense if the stimulus is entirely financed by deficit spending and interest rates remain unchanged?
Perhaps these are naïve questions, but how is it that when unemployment benefits are spent, there is a < 1 multiplier? When home sales rise due to the tax incentive, how is it that the multiplier is < 1? When infrastructure dollars are spent creating new demand for labor and materials, how is it that the multiplier is < 1?
What happens to the 40% to 60% of the $850 billion stimulus that is missing in Barro’s analysis? Does it disappear? And are there NO secondary effects of this spending?
Since there's no increase in taxes in 2009/10, that can't account for the 2009/10 reduction in private spending.
Is it higher interest rates in 2009/10 driven by increases in Treasury bond sales to finance the additional deficit spending? There’s been no rise in interest rates and none is expected in 2010.
Is it because federal spending "crowds out" private spending? How? Where?
In fact, private spending is likely to INCREASE not decrease in response to the stimulus package because the stimulus positively affects consumer and business confidence thereby loosening their wallets. Stimulus spending (combined with monetary and fiscal support to the banking sector and jawboning by the White House) is likely to support private spending by increasing confidence of bankers and encouraging bankers to loosen credit restrictions.
I have more, but this is already too long.
Harold” “but why does Govt. spending cause the uncertainty?”
Damn good question. Seems like private spending is likely to INCREASE not decrease because the stimulus REDUCES uncertainty and increases confidence (compared to govt inaction),and thereby leads them to loosen their wallets. Stimulus spending (combined with monetary and fiscal support to the banking sector and jawboning by the White House) is likely to further support private spending by increasing confidence and encouraging bankers to loosen credit restrictions.
Harold:
“I suppose it again comes down to the question, is the multiplier derived from wartime defense spending valid for peacetime banking crisis? If not, how do we calculate the multiplier?”
Yes, this is one of the $64,000 questions. But before we turn to that question, we should take a quick look at Barro’s estimate of the defense multiplier.
Elsewhere…
http://www.economics.harvard.edu/faculty/barro/files/wsj_09_1001.pdf
…Barro estimates the defense multiplier at 0.7, increasing “by around 0.1 for each two percentage points by which the unemployment rate exceeds its long-run median of 5.6%.” So, taking an average unemployment rate of 10% for 2009/2010, the estimated multiplier reaches 0.9, instead of the 0.4 (2009) and 0.6 (2010) Barro uses in the article Steve cites here.
That’s one hell of a difference.
Now, how reasonable is the assumption that an estimated defense spending multiplier is a good proxy for a non-defense spending multiplier? (In fact, Barro argues that the non-defense multiplier should be lower.)
Barro glides right by this issue.
The truth is we (and he) don’t know that it IS a good proxy because no one has estimated a reliable non-defense multiplier, and it’s not for trying. Estimating a defense multiplier is easier, so that’s what he uses. He offers no evidence that it’s a good proxy for non-defense spending.
While there are not estimates of a non-defense multiplier that are as reliable as there are for a defense multiplier, there are some that are suggestive and are considerably higher than Barro’s (0.4 in 2009; 0.6 in 2010).
Romer and Romer estimate a non-defense multiplier of 1.6, and Stevans and Sessions estimate a multiplier of 3.33.
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1424446
So, with Barro-revised (based on a 10% unemployment rate), Romer& Romer, and Stevans & Sessions, we have a range of estimates between 0.9 and 3.33.
Obviously, using any of these estimates would have a profound effect on Barro’s conclusions.
ThomasL-
“I believe people spend less because of uncertainty…. With uncertainty, people want to increase their cash holdings more than they want new stuff. Not too hard a case to imagine, especially when many households are awash in several years stock of new stuff.”
I don’t understand how this bears one way or another on the differential effect of the stimulus package (vs no stimulus)). That is, unless you’re also arguing that stimulus spending will somehow increase uncertainty.
Above, I’ve argued the opposite. Private spending is likely to INCREASE not decrease because the stimulus REDUCES uncertainty and increases confidence (compared to govt inaction),and thereby leads them to loosen their wallets. Stimulus spending (combined with monetary and fiscal support to the banking sector and jawboning by the White House) is likely to further support private spending by increasing confidence and encouraging bankers to loosen credit restrictions.
Steve-
“The first thing to observe is that spending by one party generates income for another.”
How is this statement consistent with Barro’s claim of a 0.4 (2009) and 0.6 (2010) multiplier. 60% and 40% of the stimulus seems to “disappear” in this chart.
If I understand the argument here, the answer is that the “missing” stimulus is captured in the 2009/10 reductions in private spending.
But how can that be? Since there’s no increase in taxes in 2009/10, that can’t account for the 2009/10 reduction in private spending.
It can’t be higher interest rates in 2009/10 driven by increases in Treasury bond sales to finance the additional deficit spending. There’s been no rise in interest rates and none is expected in 2010.
Is it because federal spending “crowds out” private spending? How? In what sectors?
In fact, I don’t undertand why private spending isn’t likely to INCREASE rather than decrease because the stimulus increases consumer and business confidence leading them to loosen their wallets.
Neil-
“But presumably, Barro is using an empirical estimate of this multiplier obtained from data over time periods where the sort of interest rate crowding out you describe did occur, even though it may not occur now with the Fed holding interest rates around zero. This calls into question is the appropriateness of Barro’s government spending multiplier in the current situation.”
It’s a good point. Barro’s time series is heavily weighted with periods of moderate to strong economic growth (where. as you observe, crowding out is more likely) and no periods of contraction comparable to the current one. Therefore, his model significantly underestimates the multiplier during severe recessionary periods.
Elsewhere…
http://www.economics.harvard.edu/faculty/barro/files/wsj_09_1001.pdf
Barro acknowledges the multiplier is considerably higher during periods of high unemployment. He estimates the defense spending multiplier at 0.7, increasing “by around 0.1 for each two percentage points by which the unemployment rate exceeds its long-run median of 5.6%.”
So, assuming an average unemployment rate of 10% for 2009/2010, the estimated multiplier reaches 0.9, instead of the 0.4 (2009) and 0.6 (2010) Barro uses in the article Steve cites here.
Why does government spending cause uncertainty?
Cash for Clunkers subsidized the purchase of fuel efficient cars as long as they were bought in a certain time period.
Similar programs that subsidize the purchase of energy efficient home appliances have been enacted.
As a consumer, I see that the government has recently been subsidizing the purchase of energy efficient goods. Uncertainty arises over which goods the government will subsidize next – this results in a decrease in spending. If I think that the government might subsidize energy efficient ovens next summer, I won’t purchase my new oven now, but will wait until next summer.
So it seems that:
The multiplier used to calculate the figures was 0.4-0.6
This multiplier was calculated empirically from defense spending not during recession.
There are other published estimates of the multiplier which are much larger.
Barro himself acknowledges elsewhere that larger multipliers are appropriate in these type of situations.
The reasoning behind assuming a higher multiplier appear on the surface to be sensible and reasonable.
There would have to be a very good case for selecting the lower multiplier in these circumstances. This case has not been made by Barro or anyone else here. (That is not to say that the case doesn’t exist).
It is looking to me that Barro has used an inappropriate value for the multiplier which has exagerated the case against the stimulus.
Here’s another look at Barro’s work from different economists. The question everybody has locked on is what happens when interest rates are up against the zero bound. We only have 1930’s US and mid 90’s Japan as any kind of precedent as far as I know (which isn’t very far).
http://www.voxeu.org/index.php?q=node/4227
In the meantime, there seems to be some consensus that the stimulus has created 1.8ish Million jobs (Moody’s economy.com, Congressional Budget office). And yes, there probably could have been more etc., but God knows we need jobs right now, and sometimes you just have to kick a problem down the road a bit and deal with it later – especially when there is hardly consensus about how what is happening today is going to affect tomorrow (although I am gloomy about tomorrow but for probably different reasons than most of the posters here).
“Seems like private spending is likely to INCREASE not decrease because the stimulus REDUCES uncertainty and increases confidence (compared to govt inaction),and thereby leads them to loosen their wallets. ”
I don’t agree with that at all. Stimulus increases uncertainty because no one can see what the “true” level of market activity is. The gov’t stimulates some sectors and not others, and each of those at different rates. That has secondary effects on other industries as the money ripples out. This is generally considered the “good” effect of stimulus. Rippling out is what it is supposed to do.
Ignoring the moral problems of the Cantillion-style effect of giving freshly printed money to some favored groups of people first, I still have no idea what the real demand for my goods and services is as long as that is going on.
Some day, some how the gov’t will stop subsidizing stupid projects on the grounds that they help every one to buy stupid things. They may decide to subsidize non-stupid projects instead, or they may decide to subsidize an entirely new, discrete set of stupid things, or they may decide to subsidize nothing. Whatever they decide, the one thing I know it won’t be this way for long.
To repeat it in a differnt way. As long as they keep stimulating, I’ll never know what the demand for my products and services would be outside of the stimulated environment, and by the very nature of stimulus I know that the stimulated version is not stable.
That doesn’t help me plan for the future at all, or reduce my uncertainty, it does the exact opposite. It does give me more cash in my pocket in the short term, but as long as I can’t even kind of see what is coming down the road, I’m going to squirrel some away, probably all that I can. Even if I think I _might_ know what is going to happen, I can’t tell if Bernanke, or Obama, or Pelosi, or some Republican congress in 2011 will decide to entirely shift the field and make all the positions I took unprofitable. It is safer to wait and see.
That is not what I call reducing uncertainty. That even has its own special term, “regime uncertainty.”
I wonder if the multiplier could be turned into some form of financial instrument, like a bond. Tie the payout to the coming period’s tax revenues. If GDP really does grow based on a multiplier effect (and thus tax revenues also grow), people will be chucking bags of money at this investment, and the government can afford to reward the investors. Better yet, if you don’t believe there’s a big multiplier effect (as I don’t), you don’t have to fund the government through such an instrument. A true Keynesian can put his money where his mouth is. They don’t need to confiscate from us skeptics and promise everything will turn out as predicted.
Is there anything about the multiplier effect that requires the requisite government spending to be funded by unwilling future taxpayers, as opposed to willing investors? (Not a rhetorical question; there may well be a reason that I’m not thinking of.)
I guess my larger point is, why haven’t our financial markets already tapped into this multiplier effect? And I’m starting from the assumption that investment companies are incredibly greedy profit maximizing corporations. If sudden cheap credit and spending splurges create a lot of wealth, why are investment bankers missing that potential?
GregS,
I like the idea. In response to:
“I guess my larger point is, why haven’t our financial markets already tapped into this multiplier effect?”
I think only the USG could do such a thing, because only they, through income and other taxes, have such a broadly based income stream that they can tap into a gigantic aggregate like GDP.
I wish more gov’t activity were funded through voluntary means. It is usually tricky to work out, but in the case you proposed, if the multiplier is true and significant like they claim, it should be fairly simple.
GregS, I think the problem with the bond idea is that the growth is compared to what would have happened without the stimulus. Since we can never know what that would have been, we will never know for cirtain what the multiplier actually was.
Good point, Harold. I didn’t mean to imply it was easy, but I’m convinced it could be done. If you can put in money and extract a greater value of wealth, you can create a financial instrument around that investment. If the government is competent enough to engineer an economy, it ought to be clever enough to create an instrument that takes advantage of the multiplier effect (err, and instrument that you can purchase willingly). The financial markets already create complicated instruments based on uncertain future and past data.
There are probably teams of bankers workng on it as we speak!
JLA-
“As a consumer, I see that the government has recently been subsidizing the purchase of energy efficient goods. Uncertainty arises over which goods the government will subsidize next – this results in a decrease in spending. If I think that the government might subsidize energy efficient ovens next summer, I won’t purchase my new oven now, but will wait until next summer.”
Where’s the evidence that this type of “uncertainty”, if it has any significant effect on consumer behavior, is anything but a very short-term effect?
Very few consumers who need ovens NOW are going to wait six months, and eat frozen dinners cooked in a toaster oven, on the off-chance the feds MIGHT enact a rebate in six months. Now, if I’d like to have a new oven sometime in the near future, I might wait to see what happens, but my decision to wait has nothing to do with “uncertainty”. It has everything to do with “opportunity”.
Harold-
You got it.
And one more thing to add:
Barro’s methodological decisions are all tilted in the same direction–to minimize the multiplier he uses to demonstrate his economic case against the stimulus package. A conclusion that just happens to fit nicely with his ideological proclivities and those of the Hoover Institute with which he is affiliated.
This will draw a cry of “foul” from Barro’s defenders, but considering the consistency in the effects of his choices, it bears noting.
ThomasL-
“Stimulus increases uncertainty because no one can see what the “true” level of market activity is.”
At the microeconomic level, I don’t think this is a big deal. This stimulus package puts cash into the hands of wage-earners through tax cuts, the unemployed through extending uempl. compensation, companies and workers who build infrastructure, etc.
Consumers respond to these increases in income by either buying or saving. If they buy, companies respond by hiring workers and, once excess capacity is absorbed, by new capital investments. If consumers save the income, more money is available for capital investment and more jobs are created.
In short, the increase in economic activity and consumer/business confidence generated by the stimulus decreases uncertainty compared to the uncertainty of falling economic activity and the related fear it induces in the economic environment of a severe recession.
Philip-
Saving doesn’t always equal investment, only when it is lent back out to do some development. Investment is also a long term solution, not the short term one that stimulus was meant to address. I salute you for having an entirely different view of saving than most Keynesian/stimulus supporters though. As a rule, Keynesians seem to despise savings and start muttering about paradoxes which no one can quite comprehend.
I think the over all economic and political environment is such that as a consumer, I’m not going to buy anything I don’t absolutely have to buy, and as a business I’m going to hunker down and concentrate on increasing efficiency rather than on new lines of business until all the dust settles.
The stimulus with its cash for clunkers, odd ball projects that no one would have funded with private money, appliance rebates, green renovation rebates, etc. just keeps stamping around kicking up more and more. I had a rant at the end of one of my earlier comments. All of that was stuff done purely in calendar 2009. If you can just wave you hand and say “none of that matters” than I suppose would agree that stimulus can’t create uncertainty.
If, however, you happened to be with me the other day overhearing the tire store owner grumbling about how he can’t get any tires in from overseas, and when he can get them the prices are up and people can’t pay, and even domestic prices are up because it affects Chinese rubber. All of this because one day, without any new warning, Obama decided it should be that way, you might acknowledge that government fiddling does promotes uncertainty. Expand that a bit an imagine you made tires. If I were a business that just signed an offshore production contract, I got slammed. If I just signed a state-signed one, I made out well relative to my competitors that didn’t, at least. If I haven’t signed anything, I’m likely not to want to do anything for a while. I want to see how the whole thing settles out. Check out the private investment numbers and I think you’ll see a lot of that.
You mention new hiring, but very few companies actually need to hire or expand in anyway to deal with the increase from stimulus business because of a compensating drop in regular business–else, why would they need stimulus at all. They may not lay people off. Yet. When the stimulus runs out they are back where they started. What good did it do to push it off a year? If I’m a business man making doors, I need to know exactly how much demand there is for doors. Presenting a kind of false demand for doors, because otherwise the bottom would have dropped out of the door market, only hands me money today. It doesn’t help me plan my business, because eventually you’ll stop, and I still need to figure out how many workers, machines, and factories I need to supply the real market demand for doors. You just kicked the can down the road.
The whole idea then of stimulus is that people have simply gotten in a funk and won’t spend at the “right” level. The government will spend instead, and then people will gradually get out of their funk and spend and the government won’t have to.
That theory suffers tremendously from over-aggregation. It treats the economy as one big corporation, that takes dollars on one side and spits out GDP on the other. If consumers stop feeding it a certain amount of dollars, the USG can feed it instead and all is well.
In the real world, it matters whether dollars are spent at a construction company, a software company, or at clothing company. They aren’t all the same thing, and it is quite possible to have developed too much construction capacity and not enough software capacity. It is also quite possible not to know which of those is true, even once one moves past the aggregate and acknowledges that is possible to have differences.
Despite its rhetorical emphasis on aggregation, because the USG can’t (or won’t) hand money to everybody equally, stimulus must favor someone. The process assumes either or both of these two things:
1) The downturn is improperly affecting certain parties. We know who those parties are and the level of consumption that they should have, and can restore that consumption through USG spending.
2) The downturn is affecting everything, but there are certain promising parties we know will provide an excellent RoI. We can subsidize those parties now, to push up development of those sectors and lift the entire economy with them.
There aren’t a lot of other options when some people get the cash and some don’t, and people had to make decisions on who would be which. The first assumes complete omniscience. The second assumes omniscience, clairvoyance, and better business sense than all the people that cannot be persuaded to invest private funds and reap that tremendous RoI for themselves.
In the bill that we got, this doesn’t help you at all if you run a restaurant or a clothing store, but it does if you run a construction company, make “green” windows, or just managed to pitch your new idea to __fill in the blank__ spectacularly well to your congressman (perhaps along with a little $$ on the side).
People claim secondary effects. Those effects are necessarily weak as long as people save (hence, why Keynesian hate savings). This is all new money, so what you have is favoritism and Cantillon effects. Even in the best case, the people that are getting the new money first are being helped on the backs of the people that get it last.
There is a centrally planned transfer in the distribution of discretionary income. The construction worker that gets the stimulus money directly can choose to eat at my restaurant or not. I can’t choose not to pay my taxes, and I can’t choose not to have my dollar weakened relative to what it would have been, by all the new money that he and his compatriots (but not I) are getting. He is obviously in a better spot compared to me for no other reason than that the mood at the moment the bill passed happened to be in his favor. Of course, his stimulus money is bound to dry up sometime. Then I can comfort myself that at least we’re both out of business now.
That is bad politics and bad economics all at once.
ThomasL-
“Saving doesn’t always equal investment, only when it is lent back out to do some development. Investment is also a long term solution, not the short term one that stimulus was meant to address. I salute you for having an entirely different view of saving than most Keynesian/stimulus supporters though. As a rule, Keynesians seem to despise savings and start muttering about paradoxes which no one can quite comprehend.”
Yes, I agree that saving behavior frustrates the goal of stimulus to generate immediate economic activity. But of course, it produces economic activity later and that can reinforce the secondary effects of stimulus down the road, say in 2010/2011 in this case.
But what you’re argument suggests about effective stimulus packages is that (1) they should focus on putting money into the hands of consumers most likely to spend it immediately, which tends to low- to oderate income households, and business with “shovel-ready” projects that will quickly gnerate secondary business activity and create new jobs, and (2) they should avoid tax cuts for businesses in order to encourage capital investment. Avoiding (2) is reinforced by the fact that it fails to address one of the basic problems of severe recessions/depressions: over-supply and excessive productive capacity.
“I think the over all economic and political environment is such that as a consumer, I’m not going to buy anything I don’t absolutely have to buy, and as a business I’m going to hunker down and concentrate on increasing efficiency rather than on new lines of business until all the dust settles.”
Again, that’s why you focus the stimulus on consumer households most likely to immediately spend the money and business sectors most likely to immediately ramp up production–like builders of infrastructure with excess production capacity. Your point isn’t an argument against stimulus in general, but for targeting it.
ThomasL-
I meant to respond to some of your other statements before posting:
“If, however, you happened to be with me the other day overhearing the tire store owner grumbling about how he can’t get any tires in from overseas, and when he can get them the prices are up and people can’t pay, and even domestic prices are up because it affects Chinese rubber. All of this because one day, without any new warning, Obama decided it should be that way, you might acknowledge that government fiddling does promotes uncertainty.”
Demand in excess of supply is a GOOD problem to have for any business in a depression, and certainly better than have no demand at all. Moreover, the fact he can’t get tire supplies has nothing to do with the stimulus. And by stimulating demand for tires, the stimulus will lead to new production to meet that demand. This is the very secondary effect you deny occur in your later discussion.
“If I were a business that just signed an offshore production contract, I got slammed.”
Why? If domestic demand increases for your foreign-produced product, how does that hurt you?
“If I just signed a state-signed one, I made out well.
OK. That’s my case.
“If I haven’t signed anything, I’m likely not to want to do anything for a while. I want to see how the whole thing settles out.”
Well, you might until the stimulus package is enacted. After that, the uncertainty is resolved. This is at most a very short term effect, overcome once the stimulus money hits the street.
“Check out the private investment numbers and I think you’ll see a lot of that.”
Of course private investment is down. We’re in a severe recession. Why invest when there’s over-supply and excess capacity?
“You mention new hiring, but very few companies actually need to hire or expand in anyway to deal with the increase from stimulus business because of a compensating drop in regular business.”
What drop in “regular business”? You mean the drop caused by the recession itself? If so, why have weekly/monthly net employment losses dropped so dramatically in the months since the stimulus began to hit the markets? First, you have to stop the bleeding; then you revive the patient.
“When the stimulus runs out they are back where they started.”
No, they aren’t, not if the stimulus is large enough to have more than a de minimis effect. Markets are driven by greed and fear. Severe recessions are periods of pervasive, widespread, persistent fear. An effective stimulus reduces the fear by quickly putting money in the hands of individuals and businesses most likely to spend it immediately. When the money is spent, economic activity rises which slows down the layoffs, creates more jobs and business income, and reduces the fear and pervasive pessimism.
“If I’m a business man making doors, I need to know exactly how much demand there is for doors.”
That’s what demand does. It sends market signals to produce more doors. Door makers respond to those signals. They stop layoffs to meet demand. They order more wood products and hinges. The stimulus has had its desired effect.
In a depression businesses are desperate for more business in order to survive, and they’re happy to have the demand NOW. They’ll worry about tomorrow later. If I’m correct, as fear subsides and economic activity rises, the demand for doors will increase to its “natural” level as the depression eases.
“The whole idea then of stimulus is that people have simply gotten in a funk and won’t spend at the “right” level.”
It’s not a “funk”. It’s raw, widespread, persistent fear. Your language trivializes what happens in severe recessions.
“Despite its rhetorical emphasis on aggregation, because the USG can’t (or won’t) hand money to everybody equally, stimulus must favor someone.”
That’s right. It should favor those who are going to spend the money fastest. As with all economic generalizations, there’s an implied “all other variables being unchanged”. You’ll object to that but in doing so, you’ll eliminate the ability to talk in any meaningful way about economic policy making.
“1) The downturn is improperly affecting certain parties. We know who those parties are and the level of consumption that they should have, and can restore that consumption through USG spending.”
No, there’s no such assumption re: the appropriate level of consumption of certain parties. See above.
“2) The downturn is affecting everything, but there are certain promising parties we know will provide an excellent RoI. We can subsidize those parties now, to push up development of those sectors and lift the entire economy with them.”
No, the case for a stimulus has nothing to do with determining an RoI for any component of the package or the package as a whole. See above.
“The second assumes omniscience, clairvoyance, and better business sense than all the people that cannot be persuaded to invest private funds and reap that tremendous RoI for themselves.”
No, it doesn’t assume anything of the sort. It’s based on the observation (supported by theory) that in depressions economic incentives (and fear) at the micro-level reinforce behavior that worsens and prolongs the depression. The purpose of stimulus is to change those micro-level incentives (and reduce the fear) so that rational actors act in ways that gererate new economic activity.
“People claim secondary effects. Those effects are necessarily weak as long as people save.”
That’s why you target consumers and business with the highest propensity to spend in depressions.
If you choose to respond to this post, please respond to specific points I make, quoting those points as I do here, rather than engaging in a more generalized discussion.
Thanks.
Philip-
I certainly see why you favor stimulus. If people spend it, it’s a win, if they save it, it’s a win. :D
I’ll make an impartial appeal to authority for a moment, in this case Milton Friedman:
“The more stubborn consumers are with respect to how much they will spend on consumption out of a given income, and the more stubborn purchasers of capital goods are with respect to how much they will spend on such good regardless do cost, the nearer the result will be to the Keynesian extreme of $300 rise. On the other side, the more stubborn money holders are with respect to the ratio they wish to maintain between their cash balances and their income, the closer the result will be to the rigid quantity theory extreme of no change in income.”
Your view, the best I can see, is that by targeting people with strong consumption preferences you get maximum stimulus. Their additional spending reduces uncertainty, because it gives businesses more stable income.
My view is that almost everyone’s consumption preferences are elastic and tend toward holding more cash/savings when they are uncertain. By centrally planning who gets stimulus to do what, you are likely to either crowd out private investment (if the project was worthwhile), or waste money entirely (if the project was not worthwhile). That increases uncertainty because the markets are distorted based on the [changing] whims of the legislature rather than on demand, while entrepreneurs and businesses need a true read on demand to plan and invest for the future wisely.
I doubt there is an unambiguous answer to these different views. I think it is probably an agree to disagree situation.
I didn’t see your second post before responding.
I don’t have time to do them all, but here:
“Demand in excess of supply is a GOOD problem to have for any business in a depression, and certainly better than have no demand at all. Moreover, the fact he can’t get tire supplies has nothing to do with the stimulus.”
No. There isn’t demand in excess of supply. The reason he can’t get tires is because demand existed at the pre-tariff price, and Obama slapped a 35% increase on the same tire. No one wants a $120 tire for $165. Or, actually, in my case a $250 tire at $340. Why? It promotes cheap tires into competition with mid-level, and mid-level to high-end. Why buy a mid-level $250 tire for $340 when I could get a premium tire for $340? Overseas production falls, off, imports fall off, etc. Now he can’t get his regularly stocked tires even if he wanted them (which he doesn’t), and he has to look for alternatives. He also has to deal with a decimated low-end market, since this established a new price floor around the previous mid-level. All in the name of stimulating employment in domestic tire mfgs. (Not part of the main stimulus bill, btw, but done for stimulative effect.)
“Well, you might until the stimulus package is enacted. After that, the uncertainty is resolved.”
No, not at all. This is actually a perfect example, because Obama made this decision [almost] out of the blue, ~8 months after his stimulus bill. How was anybody supposed to foresee and plan accordingly for that? What about 3y contracts I signed 18mo ago? Was I supposed to have known years in advance? That kind of regime uncertainty pushes my planning horizon closer and closer, which limits the kind of investment I’m willing to do to much shorter terms.
For a few others. The house credit expired, then it didn’t. Cash for clunkers was expiring, oh, wait, it’s extended, oh, wait it ended early, etc., etc. Do check out the litany of those things at the end of one my first comments. Those weren’t even all…
“No, there’s no such assumption re: the appropriate level of consumption of certain parties. See above.”
Oh, yes there is. If I stimulate demand in a certain sector rather than some other sector, it is because I think this sector is unfairly drooping and that I can hold demand constant in it until popular demand comes back up. That means I think popular demand is too low in that sector, rather than that the sector is overdeveloped.
“No, the case for a stimulus has nothing to do with determining an RoI for any component of the package or the package as a whole. See above.”
Then why all the VC like funding the stimulus bill has for X alternative energy research company and Y electric car company? Isn’t it because the USG thinks they will pay off down the road and make us all better off for it? These engineers aren’t the “likely to spend the money because they have to” workers you advocated before.
It seems from above that if the stimulus was to be most effective, it should have been in the form of time limited “spending vouchers” given to everyone on low incomes. I have heard this idea, but don’t know if it has been tried anywhere
Harold-
The Bush “stimulus” circa ’08 is the closest thing that comes to mind.
It was criticized because, as cash, it was as likely to be saved as spent. Some recommended an expiring debit card this time round, but it didn’t go anywhere as a plan, because I don’t think Keynesian stimulus was ever really the point. Keith Hennessey has a rather length post describing what the Bush stimulus. He should know, as I think he was the architect.
I don’t favor either of those routes, but as bad ideas go, Bush’s seems less bad to me than Obama’s, because it has far less in it of picking winners and losers from the top down.
In terms of getting money in the hands of those likely to spend it, I would also read it as closer to the idea that Philip is recommending.
I put mine in Bush stimulus in savings just for the kicks, but I’m kind of misanthropic.
ThomasL-
“No. There isn’t demand in excess of supply. The reason he can’t get tires is because demand existed at the pre-tariff price, and Obama slapped a 35% increase on the same tire.”
We’re talking about the stimulus package here. As you say, this isn’t part of that package so it’s irrelevant to this discussion.
Moreover, it wasn’t done for stimulative purposes. It was done to stop Chinese dumping of tires at prices below the cost of production which the Chinese did to stimulate THEIR economy. In other words, the Chinese govt was subsidizing these tires.
“No, not at all. This is actually a perfect example, because Obama made this decision [almost] out of the blue, ~8 months after his stimulus bill. How was anybody supposed to foresee and plan accordingly for that?”
This doesn’t apply for the same reasons I provide above.
“If I stimulate demand in a certain sector rather than some other sector, it is because I think this sector is unfairly drooping and that I can hold demand constant in it until popular demand comes back up.”
This is incorrect because the sectors are being selected based on which ones are ready to spend the money now to get the biggest, fastest bang for the buck. It has nothing to do with judging which ones are “drooping unfairly”.
Moreover, much of the money didn’t go to any particular sector, but instead went to individuals who will spend it in multiple sectors.
“Then why all the VC like funding the stimulus bill has for X alternative energy research company and Y electric car company? Isn’t it because the USG thinks they will pay off down the road and make us all better off for it? These engineers aren’t the “likely to spend the money because they have to” workers you advocated before.”
It wasn’t like VC funding at all, and I’ll eat my hat if you can find any evidence that the administration or Congress resorted to RoIs to analyze these provisions.
These were policy decisions designed to boost manufacturers that are ready to spend the money now and have the potential to reduce our dependence on foreign and carbon-producing energy sources. In other words, it was a twofer: a stimulus and a way to address energy issues.
Philip-
“It was done to stop Chinese dumping of tires at prices below the cost of production which the Chinese did to stimulate THEIR economy.”
For one thing, that isn’t true. It is more of a dispute between labour and management, with labour wanting protected from outsourcing so that their domestic jobs could be “created or saved.”
For another, that is completely nonsensical. You don’t stimulate your economy by consuming $100 in real resources in return for $60 in benefit. With every one you produce you’re instantly poorer for it.
“This is incorrect because the sectors are being selected based on which ones are ready to spend the money now to get the biggest, fastest bang for the buck.”
I don’t buy that for a second, because you could stimulate a lot more by cutting payroll taxes, for example, with the added bonus of lowering the cost of labour relative to capital, helping tame unemployment.
More importantly though you are back to over aggregation. Your view holds that if construction is down, construction has slack, which provides a good opportunity to draw on that slack with stimulus, since those idle resources can gear up quickest for maximum spending/stimulative effect. Unless you believe that private investment in construction will rebound to that stimulated level naturally, it is just kicking the can down the road because as soon as the stimulus is withdrawn construction must slacken again and there will be another round of unemployed workers and slumping output. If the stimulus was hyper-effective, there may even be _more_ unemployed workers: all the originals plus all the new ones drawn into the “hot” construction sector. The only way that won’t happen is if you assert (as you implicitly are) that construction as a sector is not overdeveloped vs. some known “right” level, which brings the argument all the way back to the beginning.
“It wasn’t like VC funding at all, and I’ll eat my hat if you can find any evidence that the administration or Congress resorted to RoIs to analyze these provisions.”
So they intentionally invested billions in things that they didn’t even bother to analyze would be worthwhile? That is supposed to make it better?
Even your analysis concedes the point, however, “[T]he potential to reduce our dependence on foreign and carbon-producing energy sources,” is roughly analogous to my own statement, “[W]ill pay off down the road and make us all better off for it.” In that line, it shouldn’t take you too much Googling to find quotes of politicians ca. 1Q09 claiming that the stimulus funding of green tech will make energy cleaner, cheaper to produce, lives richer, great for the future economy, etc. That is positive RoI.
The other half, of green companies being more “ready” to spend. What evidence do you have that particular industry was more ready to spend money than some other industry? There isn’t any; they are just more politically vogue.
Ok, new version, the whole argument in a few lines:
Stimulus distorts price signals. Distorted prices hinder planning. The inability to plan causes private money to sit on the sidelines or be malinvested.
That is my whole argument in nutshell, and I’ll leave it there.
Here’s my question: What assumptions could you adopt, even in theory, that would lead you to assume that government *spending* in a recession is better than simple government handouts (i.e. stimulus checks)?
It seems you can prove fairly simply that government spending is always less efficient than handouts.
Suppose the govt. buys a chair from Alice for $10 and gives it to Bob for free. The cost to Alice of building the chair (the minimum she would sell for) is x, and the value Bob places on the chair is y. So the benefit to Alice (her consumer surplus) is (10-x), and the benefit to Bob is y. Total benefit: 10 – x + y, which can be rewritten as 10 – (x – y).
But we know that x – y is positive. Because if x had been less than y — if Alice’s costs had been less than the value that Bob places on the chair — the transaction would have happened without the government buying the chair. Therefore x is greater than y, and 10 – (x-y), the total benefit to Alice and Bob, is less than 10.
So, it would have been more efficient for the government to simply hand out $10 in cash to Bob and Alice (depending on how you split it between them, you could make them both individually better off than they would be under the spending plan, but you can certainly make them *collectively* better off).
Advocates of government spending say that once Alice has the money, she’ll go out and spend it and stimulate the economy further — but it seems that should be true whether the cash was a handout or payment for something the government bought. So if you’re comparing government spending to government stimulus checks, that consideration weighs equally on both sides.
And of course the $10 will have to be repaid using taxes raised in the future, which is a downside of both plans — but an equal downside for both of them.
So I can’t think of any conceivable model where government spending would be preferable to simple government stimulus checks. What would be one?
Basically, government spending to buy something from A to give to B, amounts to a cash handout followed by an inefficient transaction (if it were efficient, it would have happened anyway), and it would always be better just to do the cash handout. If there’s a counterargument to this — a benefit to government spending that would not also be true for government handouts — then what is it?
(Government spending can be useful if it’s for something that is a public good, like public safety — but that’s true regardless of whether or not you’re trying to “stimulate” your way out of a recession.)
I think this ties in with the post on voting – it would be a better economic way to hand out stimulus checks, but a poor political one. Giving away spending vouchers to poor people is not going to go down well in the polls.
Bennett: The usual argument is this: Say the equilibrium wage rate is $10. In a recession, the wage rate gets stuck (for some reason) above equilibrium at $15. At that rate, there are enough unemployed workers that the social opportunity cost of hiring is very low (say $3)—-there’s not much opportunity cost to hiring someone who was going to spend all day sitting on his doorstep anyway. When firms evaluate the costs of projects, they look at that $15 wage rate, but when the government evaluates projects, it should look at that $3 opportunity cost, which is lower than the “usual” $5 opportunity cost—so a lot of govt projects that would be ill-advised in normal times become bargains in a recession.
Bennett-
In addition to Harold’s point (about the politics of passing a stimulus; it’ll never fly if all the money goes to the poor; you have to include incentives for the middle class and business, and tax cuts along with cash disbursements) and Steve’s point about opportunity costs, I can think of two other reasons for government spending:
(1) there are products that individual and business don’t buy or are far less likely to buy (ever or during a recession, for a number of reasons). If businesses that produce these products can respond quickly to government incentives, if makes sense to include them in the package. Three examples from the Obama package immediately come to mind: infrastructure construction, automobile manufacturing, alternative energy products, broadband services to underserved markets.
(2) in the Alice example, the government can tap economies of scale that allow Alice to produce more efficiently then pass part of the savings on to Bob while Alice keeps a part of it. Both parties get more value from the transaction than they would if Alice depended on many individual transactions.
There’s a weakness I often see in arguments by opponents of the stimulus and have been meaning to address.
There’s an assumption that a cost borne in the present is equal to the same cost borne in the future (in constant dollars). In a recession, this isn’t the case.
The easiest way to think about this, especially if you’ve lived paycheck to paycheck, is: a bill that comes due today requires payment with dollars that are much more valuable than the same bill that comes due immediately after you’ve received the next paycheck.
Likewise, in a recession a tax bill that comes due today is paid with dollars that are much more dear to taxpayers than the same tax bill that comes due after the recession.
I promise this will be my last blog on this subject, though it will be a bit extended. :)
I don’t find Barro persuasive for the following reasons:
(1) Barro’s data is weak (for this particular research application) because they’re heavily weighted with years of strong growth or economic stasis and applied to a period of severe recession when we’d expect the multiplier to be significantly stronger, I understand there’s insufficient data to analyze periods of depression/severe recession, but that doesn’t mitigate the problem with the data he uses.
(2) Barro uses a defense multiplier as a proxy for a non-defense multiplier data without making a strong case for why its a good proxy. I understand that data and analytical tools available to Barro are inadequate for developing a non-defense multiplier, but that doesn’t mitigate the problem with using the defense multiplier as a proxy.
(3) the stimulus package Barro analyzes is very different from the typical annual non-defense appropriation package and is likely to have a different multiplier from either a defense multiplier or a generic non-defense multiplier.
(4) these weaknesses in Barro’s analysis should lead him to be cautious in his conclusions about the Obama package. In fact, he is anything but cautious. He has been actively promoting his study as the basis for a strong criticism of the merits of enacting the package and, now, for challenging its effectiveness.
Barro is the most prominent academic economist-critic of the stimulus as demonstrated by his ubiquity. (Google “stimulus+non-defense+mutliplier”; 18 of the first 20 results are articles that focus on Barro’s research or are Barro’s articles. I didn’t look beyond the first 20.)
(5) considering those weaknesses, Barro’s lack of caution in interpreting his results, and his role as a primary critic of the stimulus, it’s reasonable to speculate that his analysis was undertaken largely for the purpose of weighing into a political debate. It’s also reasonable to question the data and methodological choices he made, all of which lean in favor of the results he obtained and support his political and policy preferences.
The bottom line: none of this research (Barro’s or those on the other side of the issue, eg., Romer and Romer) seems to be strong enough to play a legitimate role in a rational debate about whether to enact a stimulus package or in determining its effectiveness. Economic analysis is not currently in a position to contribute much to this particular policy decision: therefore, the decision should be made on other grounds.
Economists who push weak conclusions too aggressively in policy advocacy are suspect. They’re probably pushing a political agenda and relying on their academic credentials while posing as a dis-interested social scientist.
I’ve discovered something even worse. Barro’s data includes and perhaps is dominated by the effect of defense spending during World War II. Well, in WWII it would not be hard to answer Steve’s question of why private spending should fall–the government actively suppressed private spending during the war. Why would anyone apply that to the current situation?
Philip:
It’s also reasonable to question the data and methodological choices he made, all of which lean in favor of the results he obtained
This isn’t entirely true. His estimate of the tax multiplier, compared with the Romer estimate, leans *against* the results he obtained.
Steve-
You may have missed this sentence:
“The bottom line: none of this research (Barro’s or those on the other side of the issue, eg., Romer and Romer) seems to be strong enough to play a legitimate role in a rational debate about whether to enact a stimulus package or in determining its effectiveness.”
What data and methodological choices does Barro make that lean against the results he obtains and seems to seek?
I’m curious about your response to the larger point:
“(5) considering those weaknesses, Barro’s lack of caution in interpreting his results, and his role as a primary critic of the stimulus, it’s reasonable to speculate that his analysis was undertaken largely for the purpose of weighing into a political debate. It’s also reasonable to question the data and methodological choices he made….
Economists who push weak conclusions too aggressively in policy advocacy are suspect. They’re probably pushing a political agenda and relying on their academic credentials while posing as a dis-interested social scientist.”
Note that I’m not questioning Barro merely on the grounds of his policy and political preference. I’m first building a case against his choice of data and methodology and then looking “behind the veil” at why he might have made those choices. This seem entirely appropriate since Barro isn’t simply acting as a social scientist but has chosen to enter a political debate (and to do so aggressively) with conclusions that are highly questionable for all the reasons stated above.
Neil:
“I’ve discovered something even worse. Barro’s data includes and perhaps is dominated by the effect of defense spending during World War II. Well, in WWII it would not be hard to answer Steve’s question of why private spending should fall–the government actively suppressed private spending during the war. Why would anyone apply that to the current situation?”
Right. Moreover, during much of the war we were at full employment and interest rates rose, both of which would lower the multiplier. Neither condition applies in this recssion.
There’s another major weakness I overlooked:
Barro himself says that the defense multiplier is higher in times of high unemployment, higher by .1 for every 2% above the baseline of 5.6%. That would put the multiplier at about .9 at an unemployment rate of 10%. Yet Barro uses multipliers of .4 (2009) and .6 (2010) in the analysis Steve cites here.
Why?
Philip,
I think a multiplier of .9 is still too low when interest rates are up against the zero bound.
http://www.tcd.ie/iiis/pages/publications/discussionpapers/IIISDP303.php
@Steven — that would indeed be consistent with my objections, if there’s an efficient transaction (someone pays the worker $10 for their labor) that is not taking place somehow, then the government could enable an efficient transaction by paying the worker. But that just begs the question, what are circumstances — either in theory OR in practice — where the “wage rate gets stuck at $15” even though the equilibrium wage rate is $10?
@Harold — agreed that handing out cash just to poor people would be bad politics, but I thought most stimulus plans handed out checks mainly to middle-class voters. In my example, since the government spending plan benefited both Alice and Bob, it could be replaced by mailing checks in corresponding amounts to Alice and Bob.
@Philip — You said there are products that people won’t buy during a recession, but that if a business can respond to government demand for these products, it “makes sense” to include them in the package — but why? If people are not buying the products, it’s because they value the money more than the products. So if government spends money to buy those products and then gives people the products, it would have benefitted those people more just to give them the money. If not, why not?
The examples you listed — infrastructure, alternative energy, broadband to underserved areas — seem like good deals mainly because they produce positive externalities. But in that case, they would be a good deal at any time, not just during a recession. So I still don’t see the logic that makes those expenditures especially compelling *during* a recession.
Finally, you mentioned the economy of scale when government steps in to buy things and redistribute them. In this case I’d still say that you’re better off by just giving the money to those potential customers, and here’s why: If they all would have bought that good anyway, then the economy of scale will kick in by itself after you give them the money. On the other hand, if many of them wouldn’t have bought that good and they’d rather have the money, then buying that good on their behalf was wasteful. So, yes, you created an “economy of scale”, but only by buying things that the recipients didn’t want or need. Generally, you’re not going to come out ahead by creating an “economy of scale” through buying extra things that nobody wants — that’s like buying two pizzas to get 50% off the second pizza, even if I’m going to throw away the second one because all I can eat is one pizza :)
So I still don’t see a precise model that would lead you, by a series of rigorous steps, to the conclusion that government spending is a better deal than government stimulus checks to get you out of a recession– unless it’s on projects that, because of the positive externalities (like infrastructure), would be a good deal at any time, with or without a recession. Are there rigorous mathematical models that lead to that conclusion, or is it more of an empirical thing — it “seems to work”, even if all of the loose ends aren’t tied up theoretically?
Bennett-
“I thought most stimulus plans handed out checks mainly to middle-class voters.”
There are more than 37 million low-income people in the country and virtually all of them are reached by one provision or another of the stimulus package. The most important provisions affecting the poor are the extension of unemployment compensation, an increase in funding for Supplemental Security Income, an increase for food stamps, expansions to the Child Tax Credit and the EITC, a new tax credit for low income workers, and increased funding for Medicaid,
There are a variety of other provisions that benefit the poor directly or indirectly, including child care subsidies for the poor, funding for primary schooling of poor and disabled students, a boost in Head Start funding, funding for medical services for low-income mothers, incentives for retaining high quality teachers in low income districts, Pell Grants for low income students, funding for school lunch and breakfast programs for low-income students, funding for expansion of low-income housing, and incentives to low-income consumers to purchase energy efficient appliances. There are probably others.
I think the package covers the poor pretty well. As discussed above, in order to enact it, spending and tax cuts that benefit the middle-income and business are also included.
“You said there are products that people won’t buy during a recession, but that if a business can respond to government demand for these products, it “makes sense” to include them in the package —
but why? If people are not buying the products, it’s because they value the money more than the products.”
When I said “people don’t buy” certain products, I was referring to “products” like infrastructure which are almost exclusively purchased by govt.
I didn’t say people “won’t buy” certain consumer products during a recession; I said they’re they’re “less likely to buy” them during a recession. Government stimulus incentives do not induce consumers to buy products they don’t want. They lower the costs of products consumers want or need and are “less likely to buy” because of the recession, making them more affordable and making consumers “more likely to buy” them.
“The examples you listed — infrastructure, alternative energy, broadband to underserved areas — seem like good deals mainly because they produce positive externalities. But in that case, they would be a good deal at any time, not just during a recession.”
Irrespective of what Congress SHOULD do re: funding these projects, they don’t. In good economic times, Congress doesn’t fund all the infrastructure projects worthy of being built because of budgetary constraints. When a severe recession occurs Congress becomes much more willing to do so in order to get the stimulative affect.
“Are there rigorous mathematical models that lead to that conclusion, or is it more of an empirical thing — it “seems to work”, even if all of the loose ends aren’t tied up theoretically?
I don’t know. I’m not an economist. I’ve sppent most of my career around economic and other public policy-making, I read and I try to reason things out.
DividedLine_
Yes, the multiplier may be, probably is, above .9. My point is that, even if we accept Barro’s estimated multiplier despite all its flaws, he has used the wrong multiplier (by his own calculations) and profoundly biased his results in favor of his political and policy preferences.
My opinion, for what it is worth…
1) Barro used a completely inappropriate multiplier and his analysis should be totally discounted.
2) Much of the stimulus is utter government waste, at least how it has been used in my state. This debt financed waste is a future tax burden, and a negative. I do not want to see a repeat of it.
Re giving to poor prople, in order for the stimulus to work, it must be spent. If you give as cash to better off people it might be saved. If you restrict the range of goods the “voucher” can be spent on, or time limit it, the better off may simply not bother to spend it.
Harold-
Both good points. Plus, there’s a significant cost to distributing “vouchers” to the poor and it’s not so easy to reach them as you’d think. There are already well-established channels for reaching the vast majority of the poor: unemployment compensation, Medicaid, SSI, the EITC, food stamps, WIC, the school lunch program…
They get stimulus money out fast and to the right people in a way that make it likely it will be spent immediately.
Why not use them?
Philip,
Good post, and good point. The WSJ used to be the best source of business and economic news, and it’s deteriorating.
http://blogs.reuters.com/felix-salmon/2010/02/26/the-sensationalist-wsj-2/
Also, funny post about economists. I personally don’t believe in efficient market hypothesis, but the larger point that almost everybody agrees upon is that incentives are important. Apparently gold medals are a perverse incentive (don’t buy it myself), but then so too would be tenure, exhorbanent Wall St bonus packages, and CEO compensation that has left the surly bonds of this earth.
Stimulus money in the hands of those who will spend it right away is an excellent place to put it. Since 40% of those who are now unemployed have been so for more than 6 months, there are a great many people who can use the relief and there is little danger they have the luxury of saving it. State budgets are a good place to send it, since most have a balanced budget law. With the high unemployment rate the tax reveues are down, leaving only two choices for those states. Cut programs (i.e. jobs) and/or raise taxes, both of which are contractive in their effects.
“Apparently gold medals are a perverse incentive (don’t buy it myself), but then so too would be tenure…”
So true re: tenure. It’s even more perverse than the incentives of government bureaucrats which the laissez-fairers love to go on about.
Here’s something on the subject I wrote on Steve’s “Arsenic and Gold Medals” blog post (it’s long):
Q: “Do we need additional economics professors as well?”
A: Steve’s original post inspires the following answer in the same vein:
First, we must answer the question: where do economics professors come from? Of course, they’re supplied by other economics professors. (Although there’s another theory that they are spawn of the devil, but the multiplier used in these studies is questionable.)
Now, whenever you want to know if you’ve got too much or too little of something, you should start by asking: “What incentives do the suppliers face?”.
Economics professors, for example, face pretty perverse incentives: Their rewards are independent of either the value students receive from their educations or the price (salary) paid to economics graduates when they get their first jobs.
Unlike oranges, grow an economics graduate worth $100,000, you’re paid X; grow an economics graduate worth $10,000, you’re still paid X. Grow 10 economics graduates, you’re paid X; grow 100 economics graduates, you’re still paid X.
We can stop here for orange-growers, but the economics professor industry is more complex.
Economics professors have a second line-of-business: published research. Unlike their education line-of-business, there are strong incentives operating here: tenure based on the quality and quantity of published research.
Since there are no incentives on the education side of the business and a powerful incentive on the research side, economics professors invest heavily in conducting and publishing research and ignore the education line-of-business.
That is until, they receive tenure. Then the incentive for production of research disappears.
So, where does this leave us?
* In the education line-of-business, since there’s no incentive to educate students, no education occurs. Therefore, there’s no return on economics education investment; so, even one economics professor is too many.
* In the research line-of-business, there’s a powerful incentive for as-yet untenured professors to publish research. Furthermore,there are more untenured professors than tenured positions awaiting them, and there’s an inadequate supply of siginificant questions to be researched (with corresponding data and analytical tooks) to meet demand. As a consequence, much of their research efforts will be wasted, and too much irrelevant, flawed, and trivial research will be produced. Therefore, there are too many economics professors.
* As for tenured economic professors, since there’s no incentive to produce anything, one tenured economics professor is too many.
* Finally, in a market in which there are too many economics professors, some will explore other markets in which to sell their wares, such as public policy-making and partisan politics. Fortunately for them, there are long-established (some of them 60+ years) and well-funded buyers in this market (think tanks, the oldest and best-funded being conservative or libertarian). Unfortunately for them, the analytical tools and data sets available to economists are too crude to shed light on most public policy and political issues. Also, most policy makers, politicians and the public lack sufficient economics knowledge to know what these economists are talking about.
Since in this market economics professors’ products are of little value to consumers, there are too many economic professors, especially of the conservative and libertarian variety.
Bottom line: there are WAY too many economics professors.
First, why are these incentives perverse just for economics professors? Tenure etc affects all professors. Why are there too many ECONOMICS professors (relative to other professors)? Second, maybe you mean there are too many professors of all kinds. But even then, wouldn’t the benefits of tenure simply result in lower wages? People who are professors could make more as non-professors, but they trade off the security of tenure against lower wages. Perhaps, we simply have selection going on–the most risk averse people (of those capable of being professors) are professors and the others are rich (if they would have been economics professors).
I meant “compared to if they would have been economics professors”.
Yes, without thinking about it, I’d say the logic probably applies to all professors. I’d didn’t apply it to all professors because the question I responded to referenced economics professors.
My major point, however, was to apply Steve’s logic regarding the over-supply of athletes (under the Arsenic and Gold post) to the question of whether there were too many economics professors, just to see where I ended up.
I don’t take very seriously the results of pursuing chains of logic like this as is so common with reasoning of economists infatuated with mathematics. An economist will have deep first-hand knowledge of the factual errors and flawed assumption that blow this reasoning apart. But they don’t tend to question their own depth of knowledge when applying logic like this to other areas of life (like athletics at elite levels).
Actually, I don’t know how seriously anyone on this site takes an exercise like this. I enjoy it; maybe it’s all just for entertainment.
But I doubt it.
Why? Because I see it applied to policymaking, an area I know deeply and where I immediately see the factual errors and flawed assumptions that blow the reasoning apart. And the conclusions being reached, discussed and defended here seem to be taken very seriously.
Philip — you wrote:
“Government stimulus incentives do not induce consumers to buy products they don’t want. They lower the costs of products consumers want or need and are “less likely to buy” because of the recession, making them more affordable and making consumers “more likely to buy” them.”
The trouble is that my argument applies just as much to subsidized products as it does to free products — the government subsidies are still less efficient than if they had just handed out cash.
Suppose 1,000 chair makers are willing to make chairs and sell them for any price above $10 (that’s their cost to make the chair). And 1,000 chair buyers are willing to buy chairs for any cost below $5 (that’s exactly how much they value the chairs). No trades take place.
So the government steps in, buys 1,000 chairs from the chair makers for, say, $12 apiece, and resells them to the chair buyers at a price of, say, $2. Government spent $12,000 and made back $2,000, so spent $10,000.
The chair sellers, who made the chairs at a cost of $10 each and sold them to Uncle Sam for $12 for a $2 profit, are better off by 1000 x $2 = $2000. The chair buyers, who valued the chairs at $5 and got them for $2, are better off by 1000 x $3 = $3000. So the total improvement in everyone’s lives is $5000.
Thus the government spent $10,000 to make people better off by a total of $5000. Fail! It would have been more efficient just to hand out cash.
No matter what numbers you use, there’s a simple proof that the government subsidies are always going to be less efficient than just handing out cash: When an exchange only takes place with the aid of a government subsidy, that means the cost to the producers was greater than the benefit to the consumer, so you’re getting less benefit than your costs, and the exchange is wasteful. A government subsidy (of a purchase that otherwise would not have happened) is equivalent to handing out cash and then mandating a wasteful exchange. This is always less efficient than the handout of cash would have been by itself.
@Bennett Without trying to argue for or against your conclusion, I think that your example is too simplistic. What about the wood supplier that’s better off since he made a sale to the chair manufacturer? What about the chair factory worker who kept his job because of the order rather than being laid off? They are both better off yet not accounted for.
Stephen Coy: The benefits to the wood supplier and the chair factory worker are called “pecuniary externalities”. There is a well-developed theory of pecuniary externalities that tells us that for every positive pecuniary externality there is an equal and opposite negative pecuniary externality. The wood supplier is better off for making the sale; this slightly pushes up the price of wood, which damages everyone else who’s trying to buy wood. It is not obvious (but nonetheless true) that under quite general circumstances pecuniary externalities cancel each other out, which is why economists usually don’t bother totaling them up in the first place.
Steve-
“The wood supplier is better off for making the sale; this slightly pushes up the price of wood, which damages everyone else who’s trying to buy wood.”
Does this assume there is no excess supply of wood as you’d expect in a recession? If supply and demand are balanced or if demand exceeds supply, I can see how the pecuniary externalities would cancel each other out. But when supply exceeds demand, it’s not so clear.
Does this analysis also assume that the negative and positive pecuniary externalities are both internal to the domestic economy? If so, wouldn’t the “export” of any part of the negative externality leave the US economy with a net gain?
@Stephen Coy I would say that in addition to the theory of pecuniary externalities (which I’ve never heard of), you could make this simple argument:
If the government buys a chair from the chair manufacturer, then the chair manufacturer passes on some of that money to the wood supplier, and that benefits the wood supplier, fair enough.
But if the government had just given cash to the chair maker (as I was advocating), then he probably wouldn’t have spent it on wood from the wood supplier, but he would have spent the money *somewhere* (what good is money if you don’t spend it eventually), and now that money recipient is better off too.
Therefore this, by itself, does not seem to weigh in favor of government spending as opposed to government stimulus checks.
Bennett-
“If the government buys a chair from the chair manufacturer, then the chair manufacturer passes on some of that money to the wood supplier, and that benefits the wood supplier, fair enough.
“But if the government had just given cash to the chair maker (as I was advocating), then he probably wouldn’t have spent it on wood from the wood supplier, but he would have spent the money *somewhere* (what good is money if you don’t spend it eventually), and now that money recipient is better off too.”
But if the government buys the chair, the manufacturer must produce the chair now and therefore must hire the workers and buy the materials, exactly the response you want in a recession.
If the govt gives the manufacturer a check with no requirement to make chairs, why wouldn’t the manufacturer simply save the check in order to build his reserves to better ride out a recession of uncertain duration? Sure, “he’ll spend it eventually”, but the purpose of the stimulus expenditure is to have it spent immediately, not saved.
Brad Delong also responded to Barro’s op-ed. I am surprised no one has brought it up in the comments section so far:
http://delong.typepad.com/sdj/2010/02/a-stimulus-opponent-who-can-actually-find-his—-.html
Dilip,
Thanks for the link. Good article.
Gee. DeLong’s points all look familiar.
Bennett-
I believe your analysis is flawed because it leaves out crucial factors relevant to the effectiveness of a stimulus and contains some flawed assumptions.
I’d prefer to start with the premise of your argument by can’t because the chair producer you describe couldn’t be in business, recession or no. A producer with a cost of $10 a chair could not survive in a market where consumers place a $5 value on the chair.
But let me proceed in the spirit of your analysis:
In good times, we have a chair producer whose cost is $12/chair, which he sells for $13 because that’s what the market will bear and consumers value the chairs at $13 (or more).
A severe recession hits and demand for chairs at $13 collapses. Prices throughout the economy,, including wood supplies, drop due to the widespread fall in demand. The chair producer negotiates a lower price for wood, finds other ways to increase efficiency, cuts his profit margin and reduces prices by 20% to $10.
Consumers, also affected by the recession, reduce the value placed on chairs by 30% to $9 or more, but less than $10. Chairs still don’t sell.
The govt steps in with a program that gives consumers a $1 rebate on chairs purchased within 6 months. 1,000 chairs are sold.
Here are the results:
* The govt. spends $1,000.
* Consumers are induced to buy chairs they value at $9. They’re no worse off, and if they value the chairs at more than $9, they’re better off. (Note that this calculation is very different from yours.)
* The chair producer covers both his fixed and incremental costs, makes a small margin (i.e., whatever he retained at $10) and stays in business. He’s better off by the amount of his fixed costs and small profit margin, and he lives to fight another day.
* Employees of the chair producer keep their jobs or are rehired, as are employees up the producer’s supply chain.
* Employees use their income to continue spending, boosting the economy above where it otherwise would be without a stimulus.
With a broad stimulus package, this happens across multiple sectors of the economy, reducing fear and pessimism. As the precipitous fall in the economy eases, consumer and business confidence stabilizes and private spending gradually increases.
If, instead of rebates, the govt sent checks to the chair producer or to consumers, there’s no reason to believe all, most or any of it would go toward buying chairs (or anything else for that matter). Consumers might be just as likely to save the windfall to increase their financial reserves in the face of a recession of uncertain duration. I certainly would. Hell, I’ll buy the chairs later if the recession ends soon.
If there’s no stimulus, the recession deepens, and instead of triggering the positive feedback we seek to induce, the negative feedback generated by a severe recession is sustained. In fact, it may be worsened because people lose confidence that the govt knows what its doing in economic policy making.
I can address Steve’s point about pecuiary externalities, if you’re interested, but I’ve taken enough space already.