Believe those who are seeking the truth. Doubt those who find it. Andre Gide


Friday, April 10, 2009

Does Fiscal Spending "Work"?

I just returned from a visit to the Bank of Canada and had the opportunity to speak to a few high-level officials at the Bank along with several leading academics concerning a variety of issues relating to the current financial crisis.

In some of my conversations, I brought up the subject of fiscal spending; in particular, the desirability of many of the so-called "fiscal stimulus" packages that are currently being proposed. I am not sure why I was surprised, but almost everyone I spoke to thought that there was clear merit in the idea of "fiscal stimulus;" especially in the form of infrastructure investment.

Once their view was made known, I asked the question "why?" What evidence can be brought to bear in support of their view? What was it that convinced them of their belief on this matter?

Judging by the delay in the responses I received, I got the impression that many were not used to being asked such a question. They had to pause, after the initial shock I suppose, of having to collect their thoughts on the matter. The responses I received were not entirely convincing.

The responses could be divided into one of three categories:

[1] There is econometric evidence which suggests that the government spending multiplier is greater than one;
[2] The big jump in U.S. fiscal spending during WW2 and corresponding increase in GDP constitutes clear evidence of the efficacy of fiscal stimulus; and
[3] These are unusual times; the market is screwed up and *something* of this sort must be tried.

Personally, I am skeptical of having one's belief on the matter so firmly rooted on the basis of [1]. But perhaps one of you might persuade me otherwise.

The second world war is one data point that most people like to point to for confirming evidence. But is it wise to base one's beliefs on this one data point? The war experience was rather unusual. True, there was a massive buildup in military hardware (and hence, measured GDP), much of which was destined for destruction. But there was also a sharp drop in personal consumption expenditures (e.g., foodstuffs and material were rationed to the population). And while employment surged, much of this was by taking women out of the home sector and into the production of war materials. I can see why a population might want to make such sacrifices during a period of war; but would they be willing to do so today? And if so, would such a diversion of resources make society better off in any meaningful sense? I am not sure.

I found [3] the least convincing. My own view is that if we do not have convincing evidence that a particular policy will "work," then how do we know that doing "something" might not leave things even worse off than doing "nothing" at all? Since doing "something" involves the mass appropriation of resources from private citizens, would it not be better to err on the side of doing "nothing?" (Evidently not).

Of course, there may be better answers to the questions I asked, but these were the answers I received. My surprise, I suppose, was not in the answers themselves; but rather, how firmly people seemed to believe in the value of fiscal stimulus on the basis of their reported answers. Is this some sort of new religion? Perhaps they might have provided more convincing answers if they had a little more time to think about it, but I am not sure.

My own view is that people believe that fiscal stimulus works because there is no question that it does work at the microeconomic level. That is, when the government commissions a large number of workers to build something, one can actually see it being built, and one can actually see workers being employed in the act of construction.

But of course, what appears to work at the microeconomic level does not necessarily mean that it "works" at the macroeconomic level. It is harder to "see" the general equilibrium effects. It is harder to estimate net employment creation (rather than gross). What would these workers have been doing absent the government project? Would they have sat at home "idle;" or would they have been employed in some other sector (or perhaps engaged in retraining?). And how does the tax bill levied on the population at large affect their desired spending? If government spending is so successful in one area, then why not have the government coordinate all production activities in the economy? What is the optimal level of government spending? How does one calculate it? These are much harder questions to answer; and so, I suppose it is easier if they are not asked.

I leave you with a couple of articles on Japan's fiscal experience:
Bloated Bureaucracy Exposed (Japan Times)
Japan's Big-Works Stimulus is a Lesson (NY Times)

For the time-being, I remain agnostic on the subject (although, I confess that the evidence from Japan leans me more in one direction than the other). If someone can provide me with clear evidence one way or the other, I'd really appreciate it!

18 comments:

  1. You ask some very good questions..

    ..I'm surprised the Bank of Canada didn't give more thought to why a stimulus would not work and discuss the reasons. At least I would expect any crown corporation (I pay for) would consider all sides of the story. And this only leaves me feeling more uneasy about the current problems. Why can't they just say they don't know what the best policy is?

    Anyways this leaves me with a few questions:

    1. It seems that the developed countries are dealing with the current crisis by co-ordinating their efforts and pursuing fiscal spending. That is, they all appear to be doing essentially the same thing. Isn't that risky? Shouldn't the different countries try different things to see which works better? Or does it, in this case, not pay to diversify?

    2. Wouldn't infrastructure spending help out developing countries? If we went into a developing country and built roads and bridges, would that help them increase their standard of living? I suspect this isn't completly true and that there are other factors such as formal and informal institutions that are important. I don't know what the evidence is but maybe I should find out.


    3. If the "economy" is changed by large fiscal stimulus, how would we know what will change, if, as it appears, that economists are not in agreement on the effects of stimulus.

    To me, the analogy is: if we changed an ecosystem, we know species will adapt and evolve, but can we say beforehand how they'll adapt? Isn't it too complex to know specific outcomes.

    For example, how do we know how formal and informal institutions will change in response to large fiscal stimulus? What else could change due to fiscal stimulus?

    ReplyDelete
  2. It seems to me that the burden of proof is on those who say that fiscal spending doesn't work. As you note, the visible evidence suggests that it works: the roads and bridges get built; there are many people looking for work, some of whom can presumably be employed to build those roads and bridges. If you want to argue about the general equilibrium effects, present a model in which those do erase the advantage of the stimulus, and someone else will present a model in which they don't. Then we can test the two empirically. Since, as your responders noted, there is econometric evidence which suggests that the government spending multiplier is greater than one, I'm guessing that the empirical analysis will tend to reject your model.

    Regarding the Japanese evidence, as interpreted in Times article, I don't think it suggests that fiscal spending "doesn't work" in the sense of not being beneficial. Rather, it suggests that fiscal policy (at least of the type and scale used in Japan) is not enough. That proposition is one with which many pro-fiscal economists will agree: few if any are recommending that a fiscal stimulus comparable to that used in Japan should be our only response to the economic crisis.

    Regarding WWII, your discussion seems to center on what happened after the US entered the war (e.g. taking women out of the home sector). But a fiscal stimulus was already taking place prior to that point, as the build-up to the war involved a lot of government purchases (or foreign purchases financed by the US government). My understanding is that the evidence indicates that stimulus to have been effective in raising employment and total product.

    ReplyDelete
  3. Great blog post David. Many insights here. I'm equally sceptical with regards to the positive impacts of discretionary fiscal, typically counter-cyclical stimulation. Unless one is a strong believer in roll-your-own VAR models, the empirical evidence is weak at best.

    Question imposes: Did these Bank of Canada and leading academic economists spend any air time on the risk of deflation and exactly what kind of threat that poses to the Canadian economy?

    That's the one bullet-proof argument that supports discretionary fiscal stimulation. That argument rests upon two strong premises. 1.) Deflation is indeed a serious threat to economic stability. 2.) Expansionary fiscal regimes compliment expansionary monetary policy and contribute to higher rates of inflation.

    -----------------

    More questions:

    Does wasteful rent-seeking by bureaucrats increase during periods of 'fiscal stimulation'? And if so, by enough to quash the whole idea of 'fiscal stimulation' or simply to scale back the magnitude of the projects?

    ---

    The popularity of fiscal stimulation can be explained by the continuing stranglehold on the public by earlier Keynes-inspired disequilibrium models. But the immediacy of public spending projects probably does play an important role as you so aptly point out.

    I believe one observes similar information effects at work in renewable natural resource environments. Many attempts to fiddle with nature and improve productivity enjoy documented support at the micro level but zero support at the bigger picture level. Many of those attempts (fish hatchery stocking, mono-culture tree planting) have lead to decreases in overall productivity.

    Politically popular, in-stream 'habitat improvements' for Pacific salmon appear completely overwhelmed by unfavourable ocean regimes and poor survival.

    ReplyDelete
  4. David writes, "My own view is that people believe that fiscal stimulus works because there is no question that it does work at the microeconomic level...But of course, what appears to work at the microeconomic level does not necessarily mean that it "works" at the macroeconomic level. It is harder to "see" the general equilibrium effects."

    Eggertsson has an interesting paper on the combination of monetary and fiscal policy under FDR and their effect on people's perception of FDR's commitment to reflation and expectations of future policy:

    http://www.newyorkfed.org/research/staff_reports/sr234.pdf

    I think it has also been published in the AER. What I fine nice about it is that he is able to do so in the context of dynamic stochastic general equilibrium model with sticky prices (grimace) and rational expectations (grin).

    Pani Pani writes, "Why can't they just say they don't know what the best policy is?"

    I don't know. This kind of agnosticism seems appropriate for the type of discourse going on in the blogosphere. But imagine the reaction of Joe Six Pack if Bank of Canada officials got on TV and said (to quote Homer Simpson), "Don't ask me how the economy works?"

    ReplyDelete
  5. David,

    By the way, nice blog. I found the link on Mark Thoma's site today.

    ReplyDelete
  6. "What would these workers have been doing absent the government project? Would they have sat at home "idle;" or would they have been employed in some other sector (or perhaps engaged in retraining?)."

    The point of fiscal stimulus is to increase employment. Whether or not some people happen to move from one job to another is irrelevant. They will have left behind their old jobs for a new one.

    The reason there is unemployment is that there is not enough labor demand. With an increase in government spending, the number of jobs still increases, labor demand increases, and employment increases.

    "And how does the tax bill levied on the population at large affect their desired spending?"

    Even if Ricardian equivalence were to hold, increased government spending will still have a stimulative effect in the present.

    "If government spending is so successful in one area, then why not have the government coordinate all production activities in the economy?"

    Hello straw man. No one is arguing for a centrally planned economy; we are arguing for a temporary increase in deficit spending. Whether government spending is more or less efficient than private spending is not the question; we need to create jobs now, and expansionary fiscal policy is the best and only way to do that.

    "What is the optimal level of government spending? How does one calculate it?"

    Sigh. When the economy is once again producing at full employment levels, then we can debate the optimum level of government spending; but right now, the question is not relevant. We must increase the deficit now and worry about questions of levels or efficiency later.

    There are good arguments against increasing government spending in our current situation; unfortunately you don't make any of them.

    I hate to be the one to break it to you, but you probably got so many blank stares and hesitations because people were so shocked that you actually had a degree in economics.

    ReplyDelete
  7. Andy:

    You say that the burden of the proof resides with those who say that fiscal stimulus does not work. My libertarian tendencies lead me to disagree with you. The cost of fiscal action is a reduction in individual liberty (each person keeps less of what they earn). Hence, if there is an argument to be made in favour of fiscal stimulus, the burden of proof falls on the agency who wishes to impose their will on its citizens.

    I think that the proponents of fiscal stimulus would have a stronger case to make if they would instead sell the program as a redistributive one. Here, the evidence is much stronger. But of course, a politician would much rather say that "everyone" will benefit; rather than just some at the expense of others.

    ReplyDelete
  8. Igloo53:

    You say that the reason there is unemployment is because "there is not enough labor demand."

    Thank you for this insight; it reveals much about the depth of the crater occupying the upper portion of your skull cap.

    I am rather dense myself it appears. You have made nothing but bald statements that reveal your convictions; it was a poor exercise in persuasion. But I encourage to you to keep trying...

    ReplyDelete
  9. Fiscal spending can help if the problem is consumer debt. If you believe that the cause of the recession is that people were living beyond their means, and we are likely to see a change towards more responsible behavior, then fiscal spending by the government should help the process along by providing employment. Note that this did not help Japan because people had significant savings.

    It is also possible for the government to lower costs across the economy by building infrastructure. Time spent waiting in traffic adds up to a lot of wasted money for commuters and businesses alike. At some point this type of investment becomes ineffective, as in Japan, but in Canada it should work fairly well at least initially. The key is to invest in things that promote long term growth.

    A third strategy (which Japan never tried but probably should have) is to throw open the doors for immigrants. People will whine and bitch that they are stealing jobs, but population growth and an influx of money is sure to grow the economy over time. This requires more fiscal spending by the government in the form of schools, hospitals, etc.

    ReplyDelete
  10. Pani Pani: good questions...but there is not enough memory on this system to record an adequate reply!

    Westslope: I do not buy into the premise that deflation is a destabilizing force. Michael Bordo and Angela Redish are two economic historians that have documented how deflations are frequently associated with economic booms, for example.

    Roman Pearce: Not sure what you have in mind with your first point; but the next two points seem sensible enough.

    ReplyDelete
  11. "I do not buy into the premise that deflation is a destabilizing force."

    Assuming by deflation you mean price deflation, you are my new hero. I've been trying to make that argument for so long but all I hear in response is crazed fear mongering involving spirals. An economics professor who will remain unnamed asked me a few weeks ago how else demand will increase (if not by government spending)... I answered, let the prices drop. Isn't that high school level logic? I want to buy many things right now, but the prices are too high. Making them higher will not increase the likelihood I will buy them. The whole "a little inflation is good" religion needs to die along with its fiscal stimulus cousin.

    ReplyDelete
  12. I think a lot of the confusion comes from 'old' macro theory putting GDP and employment as what is being maximized in their models. Of course, with that objective just mechanically, in an accounting sense, fiscal stimuli are 'good' in the sense of maximizing this (ridiculous) objective. As I keep saying here, the USSR achieved great GDP growth and full employment for many years but somehow you guys in the West did not agree. How come when the US or Canadian government is doing the same it is fine, that's what puzzles me.

    The first thing any student learns these days (I hope) is that a 'social planner' or an ideal/fictional/benevolent government cares about maximizing welfare, not income. If I don't want to spend now, how come if the government steals a fraction of my future income and spends it instead against my will makes me better off?

    ReplyDelete
  13. To summarize - the lesson is that there's some psychological treat in many people to believe in 'action', 'miracles' and 'mom-figures' which gets exploited by crafty politicians who win elections and spend other people's money. but that's the world we live in, friedman tried to change it but he's dead.

    ReplyDelete
  14. David writes: Westslope: I do not buy into the premise that deflation is a destabilizing force. Michael Bordo and Angela Redish are two economic historians that have documented how deflations are frequently associated with economic booms, for example.

    westslope continues: Well, the delaying of purchases argument makes some intuitive sense, though I never fully understood some small positive rate of inflation was regarded as optimal.

    I'll look up Bordo and Reddish at some point. Thanks. Appreciated.

    ReplyDelete
  15. David:

    Perhaps we need to be more careful with the phrase "burden of proof." Perhaps "proof" is the wrong word, since none of this will ever really be "proven."

    I'm no libertarian, but I'm conservative (and liberal) enough in principle to agree that there should be a presumption against any new policy that involves an increase in government power -- much like the modest presumption in favor of the defendant in a civil trial, but not at all like the extreme presumption in favor of the defendant in a criminal trial. I think the visible evidence is sufficient to reverse the initial presumption in this case. Plus I think there is other empirical and analytical evidence that supports the effectiveness a stimulus and is not sufficiently countered by contrary evidence.

    ReplyDelete
  16. Andy,

    I agree with what you say about "proof." Let us take this to mean "compelling evidence."

    You say that the "visible evidence" is sufficient to support your belief. I'm sure this is the case. But what is the "visible evience?" Why don't you share it with us?

    ReplyDelete