Tuesday, May 22, 2012

Robert P. Murphy Gets it Wrong on Stimulus in Sweden and the US

Robert P. Murphy has a post here criticising a comment of mine on fiscal policy in Sweden (as compared with the US):
Robert P. Murphy, “Lord Keynes Beautifully Illustrates Why We Get Nowhere in the Stimulus Debate,” Free Advice, 21 May.
Unfortunately, his response is flawed:
(1) the links I cited were to demonstrate that Sweden implemented a stimulus from 2008, not what Murphy says.

The first remarks of Murphy’s post are therefore of no value: it is only Murphy’s erroneous assumption that is at fault here. Murphy assumed, falsely, that my links were meant to prove this idea: “that Sweden is running a budget surplus now is a demonstration that their stimulus worked.” In fact, they were there to prove my assertion that Sweden “passed a large stimulus package in 2008, which continued in 2009 and 2010.” Does Murphy deny this?

Nor did I deny that “the US under any plausible metric ran a bigger Keynesian stimulus than Sweden” – of course it did. That is not the point.

The inference that Sweden’s stimulus worked is my inference, easily confirmed by the fact that
(i) the Swedish stimulus has resulted in real output growth in 2009, 2010, and most of 2011 (which, of course, the links confirm; see here as well) and
(ii) rising tax revenues.
Does Murphy deny either of these two facts?

(2) The whole assumption underlying Murphy’s comparison of the size of the stimulus in Sweden and the US is flawed for the following simple reason: what kind of naive or ignorant person believes that the global recession of 2008-2009 was exactly of the same scale, depth and magnitude in all nations?

What kind of naive person believes that the financial crisis and resulting debt deflationary effects were exactly the same in all countries? Or that the asset bubbles and private debt levels (and resulting private sector deleveraging effects and knock-on effects on the real economy) were all the same?

This is a nonsensical assumption: different countries had different economic conditions, and different crises; consequently, there is no reason why different levels of stimulus will have worked in some nations and not in others. Or why a stimulus of a certain level in Sweden was appropriate there, but not in America. Or why America’s stimulus, even though it was larger than Sweden’s, had different effects too (e.g., not as great an affect on employment).

America had a financial crisis and credit contraction of much greater severity than Sweden. America’s housing bubble and private debt levels are much higher than Sweden’s.

(3) Murphy shows himself incompetent in even understanding basic elements of Keynesian economics. He asserts:
“First let’s consider the deficit as a % of GDP, which is how Keynesians typically evaluate stimulus in the 1930s.”
Um, no, they don’t, Murphy – at least not serious Keynesian economists. How Keynesians “evaluate stimulus in the 1930s” will be find in E. Cary Brown, 1956. “Fiscal Policy in the ’Thirties: A Reappraisal” (American Economic Review 46.5: 857–879): it does not evaluate stimulus in terms of some crude citation of deficits. There is a reason why. It is not the size of a budget deficit per se that will show you if a budget is expansionary or contractionary in terms of fiscal effects. It is perfectly possible to have a budget deficit and have contractionary fiscal policy (as in Ireland and Greece today).

In order to stimulate an economy back to its growth path and potential GDP, one has to do the following:
(i) calculate potential GDP and estimate how severely GDP is likely to collapse by,
(ii) estimate the Keynesian multiplier and
(iii) then design fiscal policy to expand demand by tax cuts and/or appropriate level of discretionary spending increases to hit potential GDP via the multiplier.
A great deal of any budget deficit during a recession is merely the result of maintaining spending because of tax revenue collapse.

In both theory and practice, you could have a budget deficit, yet impart zero stimulus to an economy. You can even contract an economy and run a deficit. It beggars belief that a person like Murphy, who sets himself up as some great critic of Keynesianism, appears ignorant of this.

One will need to look at the overall expansionary effect of a budget in terms of its addition to aggregate demand, the most important part of which is how high increases in discretionary spending were.

Sweden and the US both had different recessions. The US had a severe financial crisis. Sweden had no serious financial crisis (see under the heading “Do we have a financial crisis in Sweden?”). America had a huge housing bubble; in Sweden there has been a much smaller real estate bubble and it has not yet burst. Develeraging and debt deflationary effects in America and Sweden have been different. The state of the private sector in both countries is different.

Comparing the size of budget deficits in Sweden and the US does not even show us comparable data for the size of the stimulus in each nation. As a matter of fact, the US stimulus was about 2% of GDP in both 2009 and in 2010. Sweden was much smaller: additional fiscal spending was about 0.38% of GDP (David Saha and Jakob von Weizsäcker, “Estimating the size of the European stimulus packages for 2009,” 20th, February 2009, p. 17).

But then Sweden’s financial sector was not crippled, nor was its private sector in such a bad state as America’s in 2008, 2009 and 2010. It is not surprising that a differently-sized stimulus to that in America worked well in Sweden’s case.

(4) Murphy then cites the overall size of government spending in the economies of Sweden and US, and comes to conclusions so bizarre it so difficult to take him seriously. Here are his data:
Swedish Gov’t Spending as % of GDP
2007: 51.0%
2008: 51.7%
2009: 55.2%
2010: 53.0%

US Federal Gov’t Spending as % of GDP
2007: 19.7%
2008: 20.8%
2009: 25.2%
2010: 24.1%
The fact that Sweden has government spending of over 50% means that its economy was already cushioned from private sector shocks and falls in real output in the 2000s long before the great recession, and certainly to a far greater extent than an economy where it is on the order of 20-25% (like the US).

Sweden’s recovery is thus partly a function of the high degree of government spending (G) in its GDP already in 2008 when its recession struck.

Nor is the particular degree to which government spending rose in each country relevant here: for the US and Sweden experienced different types of recession and thus the degree of stimulus necessary was different in each case (horses for courses, so to speak).

(5) And what is this?:
“Since Sweden handled the crisis much better than the US did, I would say the case of Sweden is prima facie evidence for the Austrian / austerian camp. As always in these matters, these particular data don’t prove anything; maybe there are confounding factors.”
What!? A nation that got out of recession after implementing a stimulus, and where government spending was 51.7% of its GDP in 2008, which then increased to 55.2% in 2009, is “prima facie evidence for the Austrian ... camp.”

Then the whole thing collapses with the words “these particular data don’t prove anything.” What? So what was the point of citing them?
Finally, some questions for Murphy:
(1) Do you dispute that Sweden implemented a stimulus, with expansionary fiscal policy in 2009 and 2010?

(2) Do you dispute that the Swedish recession ended about the middle of 2009 after this stimulus was implemented, and real output growth resumed? If “yes,” then what in your view caused the end of the recession and real output growth that Sweden has had subsequently? Magic?

(3) Do you dispute that the Swedish recovery led to rising tax revenues? That the budget deficit fell?


BIBLIOGRAPHY

Cary Brown, E. 1956. “Fiscal Policy in the 'Thirties: A Reappraisal,” American Economic Review 46.5: 857–879.

12 comments:

  1. You = wrong.

    http://scholar.harvard.edu/barro/files/wsj_stimulus_12_0510.pdf

    ReplyDelete
    Replies
    1. This article does not refute me.

      And, incidentally, the author, Robert Barro, is not even an Austrian, but a Chicago school New Classical fool who believes in the joke of Ricardian equivalence:

      http://socialdemocracy21stcentury.blogspot.com/2011/09/ricardian-equivalence-is-myth.html

      In fact, his claims are laughably false:

      (1) Sweden pursued expansionary fiscal policy in 2009 and 2010.
      If it didn't, then why did Swedish government spending rise? Why did government spending as a % of GDP rise from 51.7% in 2008 to 55.2% in 2009?

      Sweden might well have drawn back on fiscal expansion in 2011, but the stimulus had done its job by then, and the private sector was providing the necessary AD again. Much the same thing has happened in Australia.

      (2) Greece, Latvia and Ireland rejected Keynesian stimulus and had full blown depressions.

      Delete
  2. LK said

    In order to stimulate an economy back to its growth path and potential GDP, one has to do the following:
    (i) Calculate potential GDP and estimate how severely GDP is likely to collapse by,
    (ii) Estimate the Keynesian multiplier and
    (iii) then design fiscal policy to expand demand by tax cuts and/or appropriate level of discretionary spending increases to hit potential GDP via the multiplier.

    Could you give any examples of institutions or governmental bodies which did the above? What I am looking for is a forecast, not a historical analysis which is quantitative in nature.

    ReplyDelete
  3. LK -> "Sweden’s recovery is thus partly a function of the high degree of government spending (G) in its GDP already in 2008 when its recession struck."

    Really ? Look at these numbers :
    http://menghusblog.files.wordpress.com/2012/04/general-government-total-outlays-united-kingdom-oecd.png
    France -> 52,9% (2006) - 52,6% (2007) - 53,3% (2008) - 56,7% (2009) - 56,7% (2010) - 56,2% (2011)

    As V. de Rugy explained : "Sweden has significantly cut government spending without equivalent increases in taxes" (as opposed to France).
    http://mercatus.org/publication/gdp-growth-rates-swedish-approach

    ReplyDelete
  4. "Sweden has significantly cut government spending without equivalent increases in taxes"

    When?

    Two facts:

    (1) Sweden had government spending of 50% oof GDP in 2007.

    (2) Sweden passed a stimulus in 2009 and 2010 and government spending increased. Then the recovery began.

    ReplyDelete
  5. "When?"

    General government total outlays, % of nominal GDP (years)

    Sweden - 55,0% (2009) - 52,9% (2010) - 51,8% (2011)

    Compare these numbers with that of France.

    France - 52,6% (2007) - 53,3% (2008) - 56,7% (2009) - 56,7% (2010) - 56,2% (2011)

    France increased its government spending, without slashing spending after 2009.

    ReplyDelete
  6. Swedish Gov’t Spending as % of GDP
    2007: 51.0%
    2008: 51.7%
    2009: 55.2%
    2010: 53.0%

    Note the huge surge from 2008-2009 (partly, of course, a function of falling GDP from 2008-2009, though also the stimulus).

    Then the % falls once the recovery occurred and the stimulus did its work, causing private sector growth, and rising GDP and tax revenues.

    All precisely predicable and consistent with Keynesian economics.

    All you're doing is confirming what I said.

    ReplyDelete
  7. Have you read Rugy's post ?
    Her graph shows that Sweden experienced the best recovery, and France the worst. France did not cut its spending, as opposed to Sweden.

    ReplyDelete
    Replies
    1. Rugy's post is rubbish:

      http://socialdemocracy21stcentury.blogspot.com/2012/05/did-anders-borg-pursue-austerity-in.html

      Delete
  8. LK,

    "It is not the size of a budget deficit per se that will show you if a budget is expansionary or contractionary in terms of fiscal effects. It is perfectly possible to have a budget deficit and have contractionary fiscal policy"

    I find this a bit unclear.

    Say the economy is in recession, and the govt increases the budget deficit via discretionary spending, that would be 'expansionary' fiscal policy, right? But if the increase in deficit spending was too small in relation to the fall in private spending, then the economy might continue to contract... so is the policy then 'contractionary' fiscal policy instead?

    Maybe its clearer to talk in terms of tight and loose fiscal policy? For example, if the govt increases discretionary deficit spending, this would constitute a 'loosening' of fiscal policy, though fiscal policy might still be too tight, given economic conditions.

    Alternatively you would have to say the govt expanded fiscal policy, but fiscal policy was still too contractionary, given economic conditions, which is kind of confusing.

    See what I mean?

    ReplyDelete
    Replies
    1. Yes, I see your point, but perhaps it is better phrased as below:

      (1) "Say the economy is in recession, and the govt increases the budget deficit via discretionary spending, that would be 'expansionary' fiscal policy, right?"

      Yes, it should normally be, as long as they do not raise taxes.

      (2) "But if the increase in deficit spending was too small in relation to the fall in private spending, then the economy might continue to contract... so is the policy then 'contractionary' fiscal policy instead?"

      In that case, it would be weak expansionary fiscal policy that is not sufficient to stop the continued contraction, say, like Hoover's federal budget in 1932.

      Delete
  9. This is relevant to the discussion over at Free Advice.

    Note that Krugman calculates the degree of fiscal 'austerity' or contraction by estimtating the 'change in structural balance as a percentage of GDP':

    http://krugman.blogs.nytimes.com/2014/02/23/the-myth-of-german-austerity/

    Also, you need to know more about the situation regarding taxation in both Netherlands and Germany to have a better picture of the difference between the two. Just looking at spending ignores the other side of fiscal policy.

    ReplyDelete