The great Excelgate scandal

It is hard to imagine a more embarrassing public humiliation. Two Harvard economics professors, one a former chief economist at the International Monetary Fund, made a basic spreadsheet error in a hugely influential paper.

Carmen Reinhart and Kenneth Rogoff’s 2010 work, Growth in a Time of Debt, helped provide the economic underpinning for the case for austerity. Its headline finding, that economic growth falls rapidly after public debt exceeds 90% of GDP, was taken to identify a tipping point.

For example, George Osborne, then shadow chancellor, cited the Reinhart and Rogoff paper in a key lecture given shortly before the last election. The politician mentioned the two economists by name before going on to argue:

“The latest research suggests that once debt reaches more than about 90 per cent of GDP the risks of a large negative impact on long term growth become highly significant.”

Nor was its influence confined to Britain. In a speech earlier this month the European Commission vice-president, Olli Rehn, argued that:

“Public debt in Europe is expected to stabilise only by 2014 and to do so at above 90% of GDP. Serious empirical research has shown that at such high levels, public debt acts as a permanent drag on growth. If it is not reduced, it will become an ever-heavier burden on our economies, eating resources that could otherwise be channelled into productive investment needed to support job creation.”

With the key finding called into question it is not surprising that many Keynesian critics about the failings of those they often dub “Austerians”. Thomas Herndon, the doctoral student at the University of Massachusetts who discovered the Excel error, has enjoyed glowing profiles. Together with two of his professors he published a paper highlighting several alleged mathematical errors in the Reinhart-Rogoff paper.

Critics of austerity have followed up with what are probably best described as “told you so” articles on Excelgate. A typical example claimed in the headline that: “the argument is over, Paul Krugman has won”. Krugman, as regular readers of this column will know, is one of the most high profile proponents of fiscal stimulus.

Conservatives reacted by downplaying the significance of the mistake. Reinhart and Rogoff themselves acknowledged a spreadsheet error but rejected allegations of two other mathematical mistakes.

The two authors accept that they accidentally left out the first five countries in the alphabet when calculating figures for average economic growth. However, they strongly object to the allegation that there were selective omissions in the data they used. They argue that they used all the data that was available at the time of writing their paper.

The final disagreement relates to how averages are calculated. But here there is considerable room for disagreement about which is the most appropriate approach.

An alternative riposte to austerity’s critics points out that many other authorities have reached similar conclusions to Reinhart and Rogoff. The case for austerity does not rely solely on their one paper.

Although this debate has received enormous attention its importance is exaggerated. The focus on the correlation between government debt levels and growth is itself an illustration of how narrow economic debate has become.

In ascending order of importance here are three reasons the debate should not be taken seriously:

• Economic data can at best only provide a rough guide to economic trends. Figures inevitably vary in accuracy and quality. That is even assuming there are no calculation errors.
• The difference between the Austerians and Keynesians is far less than generally assumed. Although it is presented as a clash over a point of principle – growth versus austerity – it is essentially just a quibble about when austerity would best be implemented. The Keynesians prefer to wait a little later than the Austerians.
• Focusing on just two variables, government debt levels and economic growth, is an extremely crass approach. Simply showing a mathematical correlation between them does not prove that one causes the other to behave in a particular way. The economy is vastly more complex than this approach suggests.

This blog post first appeared today on Fundweb.