Consensus in economics and biology


Jason Collins


January 9, 2013

Despite the common public brawls, a paper presented at the American Economic Association annual meeting by Gordon and Dahl shows high levels of consensus between economists on most economic issues. Based on questions to a 41 economist panel established by the Chicago Booth School of Business, on average only 6 per cent disagree with the “consensus” answer to a question, with around 25 per cent uncertain. This observation holds for what might be viewed as relatively controversial questions, including the effect of stimulus on jobs (although the phrasing of the questions could be changed to increase dispute - such as asking whether the benefits of the stimulus outweighed the costs. Some of the questions seem benign). In some areas the consensus is weaker, such as on labour issues where 17 per cent oppose the consensus and 29 per cent are uncertain, leaving a 3 to 1 ratio between those in the majority and those against.

There has been plenty of reaction to these findings in the blogosphere (such as Noah Smith and Paul Krugman), and Justin Wolfers’s comments on the role of  ideology are worth a look. But the question in my mind is what should we consider to be a high level of consensus?

As a benchmark, a raft of climate and earth scientist surveys put disagreement among climate scientists on whether humans are behind climate change at six per cent or less.

What if we posed some questions to biologists? Group selection seems an obvious area to generate a division, but my instinct is that most professional biologists would coalesce around the same view. Advocates of the general importance of group selection are still a small minority. If we expanded the question to include cultural group selection (and similarly increased the surveyed group), we might be closer to discovering a split. However, I suspect one of the features increasing consensus in the case presented by Gordon and Dahl is that it is a small group of high-profile economists from top institutions. We could probably also increase the measured level of dispute between economists by increasing the sample to include those less tied to the middle.

Evolutionary psychology might be fertile ground for a dispute, with there still some reluctance to apply evolutionary theory to the human mind. However, I expect that it would be specific evolutionary psychology hypotheses that would be more controversial and not the general concepts. Other possible areas of dispute might be the rate of recent human evolution and the question of race, which sees some disagreement among biologists (16 per cent in the minority), although anthropologists seem more divided. On that note, I would expect a stronger split if I started quizzing anthropologists about issues such as human nature. I am also  relatively confident I could predict which anthropology faculties would be on each side of the debate.

On a historical note, would Stephen Jay Gould have been able to rally more than 10 per cent of evolutionary biologists to his side in his famous debates with Dawkins or Maynard Smith? I suspect not (is this hindsight bias from the perspective of the victors?), although he might have succeeded at exceeding that benchmark in the sociobiology debates.

In this light, it’s probably right to consider the level of consensus in economics as being strong, but to note that some of the areas have less consensus than we’d normally see in a field such as biology. Compare the consensus in economics to anthropology, however, and economists could appear quite unified.

But what of the future? Gordon and Dahl noted that consensus is generally strongest where the area in question has a larger body of literature behind it. That makes sense, but for the areas of dispute (particularly the public disputes that created the perception of a rift between economists), I don’t expect to see much short-term change. I have never been very confident on the rejection of controversial theories in economics through data and Wolfers shows evidence of ideology behind the division for some issues, so I suspect that at a minimum we may need to wait for a few funerals to occur.