A Multiplier of Zero?

In an op-ed in the Wall Street Journal John Cogan and John Taylor expand on the point they’ve been making over the period of the last two years—the 2009 stimulus package has been ineffective:

The economic data rolling in confirm that recent temporary, targeted stimulus programs have not worked, and that their enactment was a triumph of Keynesian wishful-thinking over practical experience.

As I have said before, I am prepared to believe that a properly constructed, well-timed fiscal stimulus package might indeed stimulate the economy and I’m eager to see the empirical evidence of the effects of the 2009 stimulus package. To date the evidence of effectiveness I’ve seen has come in the form of cranking the inputs through various economic models. That’s not empirical evidence.

Moreover, I’m skeptical that the Congress is capable of producing a properly constructed, well-timed fiscal stimulus package and probably wouldn’t vote for one if it were presented to it.

Is the latest tax proposal an example of a properly constructed, well-timed fiscal stimulus package? Not hardly. As I said on OTB Radio last night, if you’d deliberately set out to construct an ineffective stimulus package it might look something like the proposal that’s on the table. Among the questionable features:

  • Focus on the highest income earners
  • Short duration of cuts
  • Special interest subsidies, e.g. ethanol
  • Extending unemployment benefits rather than expanding them

The Congress tends to use the Christmas tree approach to legislation: the more ornaments on the tree, the better it looks. I strongly suspect that what we’ve seen so far is only the tip of the iceberg.

Meanwhile, Ezra Klein has interviewed Moody’s Mark Zandi on the likely effects of the things in the proposal:

“This will make a difference,” he says. “It will add a lot to growth in the first half of next year, when the recovery will be at its most vulnerable. It really seals the deal for the recovery evolving into a self-sustaining economic expansion.”

Other models, like Macroeconomic Adviser’s, rank it lower than that. I can’t help but wonder if the various effects of the increased debt won’t overwhelm whatever modest stimulus effect the proposal might bring.

16 comments… add one
  • john personna Link

    I’m inclined to believe most values between 0 and 1, but 0 itself begs incredulity.

    (If there were ever a domain for Fuzzy Logic, this is it.)

  • sam Link

    I’d like some instruction here, because it seems to me that arguments for and against the effectiveness of a stimulus package always involve counterfactuals in some way.

    For: Without the stimulus, we would never have been able to get out of the hole.
    Against: We would have gotten out of the hole without the stimulus.

    How can it shown that a stimulus did in fact do the job it was designed for in such way as to defeat the claim it was unnecessary?
    What kind of evidence would you say shows that the stimulus was key to the (a) recovery?

  • sam:

    See here for Cogan and Taylor’s prior work. The short version is you’d need to see actual measured performance above trend that persisted.

  • john personna Link

    paraphrasing – “how can you tell?”

    You need to find US regions which are the same in all respects, except for the stimulus they received.

    The more universal a stimulus, the harder it is to find the reference case.

  • See also the graph here. Everyone, Keynesians included, recognizes that if Ricardian equivalence holds then there is no multiplier. That means that if those who receive the funds save them rather than spending them there is no multiplier.

    The evidence in the graph shows that state and local governments decreased their borrowing as a consequence of receiving stimulus funds. That’s not increased spending, it’s just an exchange between bank accounts.

  • john personna Link

    OK, so I did go look at this new article. It has absolutely nothing to do with multipliers. Well, more on that in a sec.

    Here’s what it’s really about:

    “The bottom-line is the federal government borrowed funds from the public, transferred these funds to state and local governments, who then used the funds mainly to reduce borrowing from the public. The net impact on aggregate economic activity is zero, regardless of the magnitude of the government purchases multiplier.”

    Duh. We’ve been watching that unfold for a year, and we should certainly know that net stimulus was offset by anti-stimulus in state contractions.

    So, isn’t it kind of bullshit to measure a multiplier on the federal amount, without counting the state reductions as what they were, reductions to the stimulus?

  • john personna Link

    We posted at the same time, but my comment does dovetail with yours pretty well, Dave.

    Just to call out one paragraph though:

    “The evidence in the graph shows that state and local governments decreased their borrowing as a consequence of receiving stimulus funds. That’s not increased spending, it’s just an exchange between bank accounts.”

    No way. No one seriously believes that states cut back spending in response to stimulus (except perhaps professional pretenders at the WSJ). Everyone else knows that they cut back spending in response to falling receipts, and growing budget gaps.

    I can’t believe you claimed that.

  • PD Shaw Link

    What is the muliplier effect of a 2% reduction in the payroll tax if it’s offset by an 18% increase in the healthcare insurance taken out of the paycheck?

  • Here is what the CBO estimates:

    On that basis, CBO estimates that ARRA’s policies had the following effects in the third quarter of calendar year 2010:

    * They raised real (inflation-adjusted) gross domestic product by between 1.4 percent and 4.1 percent,

    In the third quarter of 2010, GDP increased by 2.5% (that’s still an estimate I think). Compared to CBO’s estimate, that means without the ARRA, CBO is saying that GDP growth would have been somewhere between 0.8% and -1.6%.

    For the second quarter of 2010, GDP growth was 1.7%. CBO’s estimate of the ARRA for that quarter is:

    aised the level of real (inflation-adjusted) gross domestic product (GDP) by between 1.7 percent and 4.5 percent

    So CBO is saying there that without the ARRA, GDP growth would have been between 0% and -2.8%.

    Personally, I’m skeptical of those estimates, but there’s no way to know.

  • CBO’s estimates are based solely, deliberately, and admittedly on inputs. They are not empirical, at least not in the sense of collecting data and making a determination based on data.

    That said they might be right or they might be wrong. I’m still looking for empirical evidence.

  • PD Shaw Link

    CBO is also estimating short-term impacts in Andy’s example (the effect on the most recent quarter). The CBO also has (at least for tax cuts) estimated the long-term impacts of stimulus.


    Extending the tax cuts for everyone through 2012:
    Increases GNP .3 to .9 percent in 2011
    Increases GNP .3 to 1.1 percent in 2012
    . . .
    Decreases GNP .3 percent in 2020

  • PD Shaw Link

    To rebeat a drum, a large part of the problem with the stimulus is that it is not sensitive to state-by-state differences. CBO appears to assume a uniform multiplier rate, which would mean a dollar sent to North Dakota has the same multiplier as a dollar sent to Michigan, a dollar sent to the state of California, the same as a number of states that aren’t experiencing budget shortfalls. This study concludes that the stimulus money is not going where it is needed (where economic activity is below potential and labor is idle), but to states with “accumulated political apparatus and skills aimed at capturing federal funds”:

    “We find that the ARRA is, in practice, poorly-designed countercyclical stimulus. After controlling for political variables, coefficients on Keynesian variables are often statistically insignificant. When they are statistically significant they are often the “incorrect” sign. On the other hand, statistically significant effects associated with political variables are almost always of the sign predicted by public choice theory.”


  • steve Link

    “That said they might be right or they might be wrong. I’m still looking for empirical evidence.”

    You can only get that with an alternate universe. I do find it interesting that those who doubted the calculations made by Macroeconomic Advisors or Moody’s, now cite their numbers showing that tax cuts will increase growth. They used the same models.

    FTR, I do agre with you on the more modest growth estimate. Too much excess inventory and capacity. Not enough savings. Not devalued enough to export.


  • You can only get that with an alternate universe.

    Horse hockey. If data in the absence of a control are useless, so are quantitative models. They don’t spring full grown from the forehead of Zeus, you know. They’re based on data.

    To acknowledge that something is imperfect is not to say that it is useless. There’s a lot of distance between metaphysical certainty and metaphysical doubt.

  • Dave,

    I agree, I’m just putting those out there. Generally I’m skeptical that we would potentially see GDP contract by up to 2.8% were it not for that one piece of legislation.

    On the question of evidence, are we able to establish causality of effects in a complex system like the economy with any reasonable precision? Say unemployment goes down or GDP goes up. How does one empirically show the reason(s) or the relative influence of various causes for such changes, much less trace them back to specific policy actions?

  • john personna Link

    “To acknowledge that something is imperfect is not to say that it is useless. There’s a lot of distance between metaphysical certainty and metaphysical doubt.”

    It’s a little like those “best guesses” about the future, isn’t it? Some people hold on to them because they are the best they have, not because they are particularly good.

Leave a Comment