Last week, economist Salim Furth testified before the Senate Budget Committee on the issue of austerity. During the Q&A period, Senator Whitehouse accused him of misleading the committee by using fake OECD data. The senator’s evidence: his own set of OECD data, which showed very different results. Furth’s experience, I assume, was unpleasant and made even worse by uncalled-for attacks by several journalists who rushed to accuse Furth of dishonesty without, it seems, checking the validity of the original attack.
Once again at the heart of the issue here is not whether European governments have tried to reduce their deficits and debt levels since 2010. They have. Furth was looking was looking at what kind of “austerity” was implemented: spending cuts, tax increases, or a mix of both. He used OECD data from 2007 to 2012, and found that, over that period, European governments had mostly raised revenue rather than reducing spending. Whitehouse argued that these data were wrong and hence Furth was dishonest (“meretricious”). The senator proceeded to quote his own numbers, which showed that some European governments had implemented austerity packages composed entirely or mostly of spending cuts.
As it turns out, the two data sets are quite different, and for good reasons. First, as mentioned before, Furth’s data looked at actual evidence from 2007 to 20012 (all his charts are properly and clearly labeled to that effect) while the senator used data from 2009 to 2016 (meaning actual evidence from 2009 to 2012 and then planned policies from 2012 to 2016). In other words, half of the senator’s data comes from projections about what European governments will do in the future. Steve Landsburg explains on his blog why such data should be taken with a grain of salt:
The numbers Whitehouse quoted come from an OECD report on what various countries plan to do (or say they plan to do) over the next few years. Because these numbers differ from what these countries have actually done, Whitehouse wants you to believe first that Furth’s accurate account of what’s actually happened is irrelevant, and second that Furth is a liar for reporting the truth.
But surely Whitehouse realizes that he can’t actually get you to believe that, which is, I’d guess, why he didn’t bother to tell you that the numbers he was reading were largely nothing but pie-in-the-sky projections from politicians with extremely poor track records of living up to their promises.
It’s as if I’d announced plans to lose 30 pounds over the next year, and then promptly gained 10 pounds. Furth comes before Congress and says “Landsburg tells me he just gained 10 pounds”. Whitehouse says: “I can’t imagine where you got that number, because I have a number here from Landsburg that refers to a loss of 30 pounds — and that comes directly from Landsburg, who you say is your source. This makes me very concerned about your testimony, very concerned about where you’re getting these numbers . . .” followed by eight minutes of innuendo suggesting that only sheer dishonesty can account for a discrepancy like this.
Well, yes, only sheer dishonesty — or perhaps an extraordinary failure of competence — can account for a discrepancy like this. But the dishonesty is not on Furth’s part.
And, I’d point out, the data error was on the Senator’s part. On Friday, Furth responded to Senator Whitehouse in this space. He wrote:
But even when I tried to replicate the result provided by Senator Whitehouse with the data from the OECD report on which they were based, I found that the numbers didn’t add up. Actually, they added up too much, counting some austerity measures six and seven times over!
The OECD tables on which Senator Whitehouse’s data are based have footnotes that explain the data they represent. The key word in these footnotes is “cumulative”:
Fiscal balance is the general government financial balance and gross debt is general government gross financial liabilities (Maastricht basis) as a per cent of nominal GDP projected by the government. Fiscal consolidation is the cumulative consolidation volume as a per cent of nominal GDP projected by the government. The composition of consolidation is expenditure reductions and revenue enhancements in the actual year (cumulative, total 100%). Nominal GDP growth forecasts are the government’s estimates of nominal GDP. Fiscal consolidation in the national currency is based on the government’s quantified measures (cumulative), except the latest bank rescue package.
“Cumulative” here means that if a government cuts spending (or raises taxes) by $5 in 2010, $4 in 2011, and $3 in 2012, the table will record $5 in 2010, $9 in 2011, $12 in 2012. Make sense?
But Senator Whitehouse added up those already added-up numbers, getting $26. He triple-counted the first $5. So for the countries with seven data points, he counted the first-year cuts seven times over!”
Yesterday, Landsburg added some comments on this point and notes that, for instance, according to the senator, Ireland has done fiscal consolidation amounting to 95 percent of GDP, when the correct number is 18 percent.
Now, we can debate whether it is better to look at policies implemented in Europe starting in 2007 or 2010. There is no doubt, as Furth noted in his written testimony, that most European countries initially responded to the financial crisis by increasing spending. Should that data be excluded? Maybe, but maybe not. For one, the need for “austerity” was amplified by the large increases in spending that took place during the previous years. But one thing is sure, no one should present data that lumps together numbers from 2009 to 2016 without noting that these numbers are only promises of policies that may or may not take place.
Over at Heritage, on Friday Furth also responded to Dylan Matthews of the Washington Post, who had suggested that Furth’s measurements of government spending were inaccurate. In fact, Furth had used OECD data that measured spending as a proportion of potential GDP, adjusting for the recession, when Matthews suggested otherwise. Since then, Matthews acknowledged that error and retracted his erroneous accusation (though he stands by the OECD’s projected austerity numbers, and has not acknowledged the senator’s huge calculation errors).