There is a growing number of commentaries by mainstream academic economists which I consider to be revisionist efforts to deny that fiscal policy has had any positive impacts. Some of the more manic offerings assert that the government intervention has actually worsened the recession and will reduce growth in the future because all the debt has to be paid back. These characters never give up trying to assert their twisted notion of self-importance. Some of the notable revisionists come out every recession and say the same thing. They cannot get over the fact that their approach to economics,which has dominated in the last 35 years and finally delivered the World to a state of near Depression, is now without any credibility. They hate the fact that the only way out was of the crisis was to reject their nonsensical policy suggestions – that the market would work it out – and return to fiscal activism. Anyway, today, I consider a notable example of this denial.
From Compact Oxford English Dictionary:
1. Information or signs indicating whether a belief or proposition is true or valid.
2. Law information used to establish facts in a legal investigation or admissible as testimony in a law court.
Which is what I thought the word – evidence – meant.
So when you read the article by Robert Barro in Tuesday’s (February 23, 2101) Wall Street Journal entitled – The Stimulus Evidence One Year On – you know he is using the word “evidence” in a way that doesn’t bear any relation to the way we all use the word.
Recall that Barro’s obsession with crowding out has kept the dead-duck idea of Ricardian equivalence alive in the literature. He continually claims that the government spending has no impact once consumers react and reduce consumption in anticipation of the tax increases. He says the same thing every recession!
So according to Barro, government spending has no real effect on output and employment irrespective of whether it is “tax-financed” or “debt-financed” in his models.
Even ignoring the fact that the description of a government raising taxes to pay back a deficit is nonsensical when applied to a fiat currency issuing government, the Ricardian Equivalence models rest of several assumptions. Should any of these assumptions fail to hold (at any point in time), then the predictions of the models are meaningless.
It just happens that virtually none of the assumptions hold at any point in time. Please read my blogs – Deficits should be cut in a recession and We are sorry – for more detailed discussion on the folly of Ricardian equivalence.
The complete irrelevance of Ricardian equivalence tendencies to the real world hasn’t stopped Barro though from banging on about it for year. He should have stopped in the early 1980s when his last attempt at recession-forecasting was so comprehensively demolished by the data.
Barro and his sycophants in the profession predicted that consumption would not rise after the US Congress gave out large tax cuts in August 1981. Why? They all said that saving would rise to “pay for the future tax burden” – just as they are saying now (read on!). The actual data shows that personal saving rate fell between 1982-84. The fact remains that there has never been a successful empirical application of the theory that they keep using to spread their lies.
And now another recession – some more fiscal intervention – and some more lies.
In his Wall Street Journal article (noted above) Barro claims that his research (yeh, dreaming up numbers and writing them down) shows that:
Over five years … an extra $600 billion of public spending … [comes] … at the cost of $900 billion in private expenditure. That’s a bad deal.
If that was the actual result then it would be a very bad deal. So we better investigate how he gets these results.
Barro says that:
The first anniversary of the Obama stimulus package generated a lot of discussion about whether and how much the package (originally estimated at $787 billion but now priced at $862 billion) moderated the recession. These are complex questions, and their answers require more than merely counting the quantity of goods and services that the government purchased or the number of people that the government hired.
We need to ask whether the government’s spending reduced or enhanced private spending and whether public-sector hiring lowered or raised private hiring.
Yes, the direct impact of the government spending on goods and services is $-for-$. But as Barro says the final impact has to include the indirect effects.
First, the non-government sector has higher income. How do they respond to that?
Second, there may be other policy changes which present off-sets – for example, monetary (rising interest rates) and fiscal policy (tax rate rises).
Thus, in an open economy, there are leakages from the expenditure stream. Accordingly, what is spent will generate income in that period which is available for use. The uses are further consumption; paying taxes and/or buying imports. Imports are considered to be such “leakages”.
So if for every dollar produced and paid out as income, if the economy imports around 20 cents in the dollar, then only 80 cents is available within the system for spending in subsequent periods excluding taxation considerations.
However there are two other “leakages” which arise from domestic sources – saving and taxation. Take taxation first. When income is produced, the households end up with less than they are paid out in gross terms because the government levies a tax. So the income concept available for subsequent spending is called disposable income.
Finally consider saving. Consumers make decisions to spend a proportion of their disposable income. The amount of each dollar they spent at the margin (that is, how much of every extra dollar to they consume) is called the marginal propensity to consume. If that is say, 0.80 – then they would spend 80 cents in every dollar of disposable income.
All these impacts have to be modelled. They are typically summarised by the concept of the expenditure multiplier. This shows the increase in final goods and services (GDP) for a dollar increase in government spending. It takes into account the indirect impacts of induced consumption and the leakages that are also associated with the rising income.
So when government spending directly impacts on income, this induces further consumption expenditure (but some is saved and paid in taxes) which, in turn, produces a “second-round” of income growth and the process proceeds until the the induced effects are neglible. They get smaller and smaller because each subsequent round of spending is smaller as the leakages occur.
Anyway, Barro doesn’t believe any of this. Ricardian Equivalence claims that consumers assume that the increased government spending has to be paid back. When? there is no explicit answer in their models – all the teaching models assume very short time periods so they can fit the graph on a PowerPoint slide! Totally arbitrary like all of the mainstream so-called “rigorous analysis”.
So with consumers armed to the teeth intellectually with their RE-squared New Keynesian model as the government spends, consumers increase their saving and the net effect is zero.
Technical note: RE-squared = Rational Expectations times Ricardian Equivalence. Conclusion: their models are the product of two totally dud assertions that do not hold in the real world.
Barro has extended himself this time though. Now the multipliers are negative – that is $600 billion public spending begets minus $900 billion private spending. This is because he also assumes massive crowding out – why? – just wait for it.
Barro then proceeds to outline his approach. It is in layperson language but we won’t hold that against him. So is most of my own blog writing.
But what the layperson doesn’t know is that the high-tech New Keynesian models that Barro uses are just fronts for ridiculously simple – almost banal ideas. These ideas are hidden behind a wall of second-rate mathematics to beguile the readers and to hold out a sense of authority.
As a trained economist with formal mathematical skills I can tell you that the mainstream models – particularly the ridiculous New Keynesian models – deserves no attribution of authority at all.
As an aside: ask any real mathematician what they think of the mainstream economics approach to formalism – they will just laugh at the idea that the latter is sophisticated from a mathematical perspective – in fact, it is mostly very ordinary indeed. But the public and students do not know that. Therein lies the start of the deception.
Barro claims he calculated the “effects on GDP from increased government purchases (the spending multiplier) and from increased taxes (the tax multiplier)” an analysis of historical data.
He came up with a spending multiplier of 0.4 in the first year and 0.6 over two years. This means that:
… if the government spends an extra $300 billion in each of 2009 and 2010, GDP would be higher than otherwise by $120 billion in 2009 and $180 billion in 2010. These results apply for given taxes and, therefore, when spending is deficit-financed, as in 2009 and 2010. Since the multipliers are less than one, the heightened government outlays reduce other parts of GDP such as personal consumer expenditure, private domestic investment and net exports.
The question you have to ask is why are these multipliers so much lower than even the other mainstream modellers come up with which are usually of the order of around 1.5 total impact?
Further, it is clear why consumer expenditure, private domestic investment and net exports have been contracting and it has nothing to do with the rise in government spending.
This is one of those revisionist efforts that is now going on among mainstream economists. I am increasingly reading that the government intervention caused the recession to be longer. These characters never give up trying to assert their twisted notion of self-importance.
So you have to presume that Barro believes that consumers now are anticipating that tax rates will rise in the near future to “pay back” the stimulus and they are thus “storing up” their pennies now to ensure they can pay the extra marginal tax burdens.
However, why wouldn’t those extremely smart households – who are assumed to be able to compute hamiltonians and solve transversality conditions out into the end spaces of their lifetimes using complex calculus – not also realise that if the economy grows they will all enjoy higher income overall which will not only provide for higher levels of consumption but also allow them to save and pay higher tax burdens?
The reality is that household consumption has collapsed because many have lost their income through unemployment while others who have retained their employment are trying to save more to reduce their debt exposure following the credit binge.
The reason investment has contracted sharply is because consumers are not buying (as per last paragraph) and firms will not continue to increase productive capacity when capacity utilisation rates have fallen below 70 per cent.
In the December quarter National Accounts for the US released by the Bureau of Economic Analysis you read that consumption and investment spending is growing again. That doesn’t look like an economy where the private sector is on strike because they are being crowded out.
Further, it is impossible to assert there is “crowding out” going on at present even if the notion had any credibility at all in normal circumstances (which is doesn’t!).
Short-term interest rates are near zero and not budging and government bond yields across the maturity curve are not budging. The central bank has indicated it will hold rates low for the next extended period.
So why does the increase in government spending of $600 billion in 2009 and 2010 cause a loss of $400 billion in private spending over the same period? No explanation is provided – this is just pure assertion, although we know from his work he would spin a line about Ricardian Equivalence and financial crowding out.
The analysis gets even more crooked when we turn to taxes. Barro says:
For taxes, I focus on a newly constructed measure of average marginal income-tax rates; these rates apply to federal and state income taxes and the Social Security payroll tax. I estimate that an increase in marginal tax rates reduces GDP, particularly in the next year. When one factors in the typical relationship between tax rates and tax revenue, the multiplier is around minus 1.1. Hence, an increase in taxes by $300 billion lowers GDP the next year by about $330 billion.
Barro doesn’t fully explain to his readers what is going on here? Is the rise in taxes the result of increased marginal tax rates or just the increased income? For it to invoke a tax multiplier effect, the tax burden overall has to rise.
The concept of the tax multiplier is straightforward but devilish to estimate in reality.
Assume the overall propensity to consume is 0.80 – which means that overall consumers will spend 80 cents for every extra dollar of disposable income received. So, for example, when tax rates are reduced and disposable income rises, consumption rises by the propensity to consume times the increase in disposable income. The resulting increase in GDP that occurs is defined as the tax multiplier.
Similarly, when tax rates rise and reduce disposable income, the reverse occurs.
The problem that plagues empirical research is to estimate this impact – by isolating the pure tax effects.
One has to first net out the automatic stabiliser effect on tax revenue – which occurs with no discretionary tax regime change at all. It is a common mistake to assume that because tax revenue is rising that tax policy is becoming contractionary. Please read my blog – Will we really pay higher taxes? – for more discussion on this point.
Further, at the individual level, as GDP growth recovers most people will not be paying higher taxes at all while others will be paying a substantial increase – why? Because they move from unemployment (zero taxes paid) to earning an income (some taxes paid). So this is hardly a bad thing.
Christina Romer (Obama’s Chief Economist) and her husband have estimated the tax multiplier to be well over 1 which is contrary to most other research papers on this topic.
That estimate was used in the paper she wrote wtih Jared Bernstein as part of the new US President’s fiscal strategy entitled – The Job Impact of the American Recovery and Reinvestment Program. I talk more about that below.
In a recent paper (September 2009) – How large is the US tax multiplier? – authors Favero and Giavazzi, using careful methodology, find:
… a tax multiplier much smaller than that estimated by Romer and Romer and similar to the size of the multiplier estimated in the traditional fiscal VARs. When we split the sample in two sub-samples (1950-1979 and 1980-2006) we find, before 1980, a multiplier whose size is never greater than one; after 1980 a multiplier not significantly different from zero.
Anyway, Barro has decided on a tax multiplier of 1.1 – higher than the majority of the literature – a bit lower than the Romers. It is clearly very much higher than the state-of-art estimates from Favero and Giavazzi.
Thus, in terms of the spending multiplier Barro is extraordinarily low and in terms of the tax multiplier he is on the high side of the literature.
Clearly deflating the first and inflating the second allow him to reduce the fiscal impact. How convenient!
Barro then proceeds to “assess the 2009-10 fiscal-stimulus package” which he characterises as:
… roughly $300 billion of added government purchases in each of 2009 and 2010 … [and then] … as of 2011, government spending goes back down to its 2008 level … I suppose that taxes do not change in 2009-10, so that the incremental spending is deficit-financed.
The following table summarises Barro’s simulations which are assuming a base scenario of no stimulus.
The yearly totals are explained as follows (all in $US billions):
- Year 1: Government spends 300 with no tax policy change. The spending multiplier is active 0.4 so total GDP rises by 120 which means that private spending plus net exports falls by 180.
- Year 2: Government spends another 300 billion and no tax policy change. The spending multiplier is 0.2 in the second year for the first year’s spending and 0.4 for the second year injection so total GDP rises by 120 plus 60 (the second-year effect on year 1) = 180, implying a further decline in private spending plus net exports of 120.
- Year 3: Budget increase has to be “paid for” by Year 5 so taxes rise by 300 but the impact is felt in the next year. The second-year effect of Year 2 spending is 0.2 times 300 = $60. Curiously there is no crowding out or other loss of private spending in this period. Why? Not explained and irrational.
- Year 4: Spending effects exhausted. Taxes rise by 300 again. With a tax multiplier of 1.1, the loss of GDP is thus -330 and this is all down to loss of private spending.
- Year 5: The lagged tax effect shaves another 330 off GDP and it is all squeezed out of the private sector.
- Totals: 600 spent, 300 lost overall to GDP, private spending plus net exports down 900, budget stimulus fully “paid for”.
Commenting on the first two years, where he claims $US600 billion is bought by a loss of $US120 billion in “tax” (lost private spending) in 2009 and $US180 billion in Year 2 – a “pretty good” deal, he says that:
But these calculations are not nearly the end of the story, because the added $600 billion of government spending leads to a correspondingly larger public debt. These added obligations must be paid for sometime by raising taxes (unless future government spending declines below its 2008 level, an unlikely scenario).
So the assumption is that the US government will raise tax rates (and hence the tax/GDP ratio measured at the full capacity output level) in 2011 and 2012 to fully offset the extra spending.
It is clear that tax revenue will rise as the economy picks up but that is not what the tax multiplier is based on. It requires tax rates to rise to reduce the disposable income available to consumers.
While the US Government is pretty loopy at present and bowing to neo-liberal demands, it is still highly unlikely that the US government will even go close to approximating the tax profile that Barro imposes on his analysis (and upon which his overall conclusion depends). I very much doubt that the US government will raise marginal tax rates by this amount in 2011 and 2012.
So it is clear to me that the private spending losses in Year 1 and 2 cannot be reasonably explained. Year 3 is totally arbitrary – why do private agents not continue to reduce their spending to save for the impending tax rises?
The Year 4 and 5 effects are just made up and will not happen.
What if we assumed the spending multiplier was 1.5 (closer to what the literature indicates) and the tax multiplier was zero (as in the latest research noted above)?
Then over the 5 years, the $US600 billion injection would deliver increased GDP of $US900 billion and extra private spending of $US300 billion. That is more likely what will happen even if the US government followed Barro’s manic tax strategy.
These estimates are much more likely than the deviously confected estimates provided by Barro.
Which means you can take Barro’s conclusion:
The fiscal stimulus package of 2009 was a mistake. It follows that an additional stimulus package in 2010 would be another mistake.
… with a grain of salt.
He is a pure ideologue pushing a fantasy barrow of New Keynesian lies.
Barro’s nonsense was also mimicked in a paper published in September 2009 as an European Central Bank working paper – New Keynesian versus Old Keynesian Government Spending Multipliers – written by noted anti-fiscal mainstreamers John F. Cogan (Stanford University), Tobias Cwik (Goethe University), John B. Taylor (Stanford University) and Volker Wieland (University of Frankfurt).
I provided a critique of that paper in this blog – Spending multipliers.
By way of background, in January 2009, Christina Romer and Jared Bernstein released a working paper as part of the new US President’s fiscal strategy entitled – The Job Impact of the American Recovery and Reinvestment Program.
They modelled the output (income) effects of a permanent stimulus of 1 per cent of GDP (percent) using either a public spending injection and/or an equivalent tax cut (equal to 1 per cent of GDP). The two multipliers estimated were 1.55 for government spending and 0.98 for tax cuts, which are consistent with the usual approach.
So if government spending increases by 1 per cent of GDP the total increase in GDP will be 1.55 per cent.
Unable to cope with the reality that their New Keynesian anti-fiscal policy world was collapsing around them, some of the key ideologues such as Barro and John Taylor then had to trump up some alternative.
In February 2009 Taylor and others released a working paper that was subsequently published in September 2009 as an European Central Bank working paper – New Keynesian versus Old Keynesian Government Spending Multipliers – written by noted anti-fiscal mainstreamers John F. Cogan (Stanford University), Tobias Cwik (Goethe University), John B. Taylor (Stanford University) and Volker Wieland (University of Frankfurt).
I was re-reading that paper today while writing some stuff up for a book I am writing and recalled this gem of a paragraph. As background, these characters purport to refute the fiscal stimulus effects that Romer and Bernstein found by exposing them to “an examination”.
It turns out that they do not actually refute Romer and Bernstein because they invent their own experiment. The difference is that they assume the central bank hikes interest rates to stop hyperinflation arising from the increase in government spending.
Anyway, what do Cogan and all say is their justification for changing the rules of the experiment? This:
Romer and Bernstein assume that the Federal Reserve pegs the interest rate – the federal funds rate – at the current level of zero for as long as their simulations run. Given their assumption that the spending increase is permanent, this means forever. In fact, such a pure interest rate peg is prohibited in new Keynesian models with forward-looking households and firms because it produces calamitous economic consequences. As Thomas Sargent and Neil Wallace pointed out more than thirty years ago, a pure interest rate peg will lead to instability and non-uniqueness in a rational expectations model. Inflation expectations of households and firms become unanchored and unhinged and the price level may explode in an upward spiral.
So because their New Keynesian mathematical models predicts hyperinflation if government increases spending then you clearly would not have this as a policy option. For those not trained in economics, these mathematical models are just made up and reflect the prejudices of the individuals. These particular models used by Cogan etc have no mapping back into reality at all. Garbage-in, garbage-out.
So because their imagination is offended, they refuse to recognise what is happening in the real world.
Pity economics students at Harvard. Among others they have to deal with Mankiw and Barro.
That is enough for today!