Elephantstrunk

Does empirical economics follow the scientific method?
Asks New Economist and more or less answers his own question with:

The authors attempted to replicate the results from 138 empirical articles in the Federal Reserve Bank of St. Louis Review. Of the 138 articles, 29 employed software they did not have or used proprietary data, and were excluded. Of the remaining 117, they were able to replicate only 9 of them: a lousy 8 per cent.

Now

1. When scientists talk about replication, they usually mean generating new data and finding the same result. In the quote above replication targeted is merely repeating the result with the same data and programming!
2. These days it is not hard to provide data and code.
3. Results of the scientific method depend strongly on to what the method is applied, so biology and physics for example are very different. There is no reason to expect scientific economics to look like physics or chemistry.
4. “Scientific” and “valid” overlap but are not the same.

It really is a shocking result. Still my guess is that such things are getting better all the time as software improves and more work gets done but it is not the only frustration with economics. The gap between economics as practised by the illuminati and what gets fed to the public is vast and frustrating for all concerned.
An example. Best sentence I saw on the web last week was Brad DeLong’s:

It is very nice to have, thanks to Golosov, Tsyviniski, Werning, and company, a model that says that you want to tax capital because you don’t want the highly-skilled and productive programmers of Google retiring in their 30s because they are wealth-satiated…

It deftly and vividly interprets an elegant new paper and, I think, shows that DeLong had already been thinking about the issues it raises. One conclusion of the paper is that it is not necessarily right to The paper itself, the talk DeLong describes, the analysis that appears in Delong’s description and indded of other authors on the same topic are not the kind of thing that gives economics a bad name. Indeed it is mainstream economists tackling issues that they are often accused of ignoring. However that paper is not what the world thinks economists have to say about taxing capital. The world thinks economics says that capital incomes should not be taxed at all.

From the outside it doesn’t look like that however. Such claims normally lead to a massacre of straw men but I think here the claim is spot on. Ignoring for now the careful qualifications and statement of results in the papers of Judd and Mirrlees as esoteric, there is the following evidence:

• Exhibit A President’s Advisory Panel on Tax Reform
• Exhibit B From page 146 of “Economic Growth Second Edition” by Robert Barro & Xavier Sala-i-Martin:

• “the imposition of taxes on the income from capital leads to reductions of [capital] and [consumption] in the long run.”

• Exhibit C The dramatic reduction in tax on income from capital, even before the current US administration. Until 1997 effective dividend taxation in the UK was actually negative.
• 
  

Proper economists are thinking vigorously about tax on capital and elaborating subtle ideas. Why wouldn’t they be frustrated if the breadth of their work and the extent of their thought is not recognised.

Unfortunately that is not what the rest of the world sees. Unless they look at things in more detail than a beginning post-graduate course in economics, what they get is as above. A surprising conclusion presented without equivocation but with dramatic policy results. That is frustrating too. Would it be too unkind to say a little knowledge is dangerous? Certainly Exhibit A is a spectacular example of hubris even for one of the perpetrators of Dow 36,000.