Continuing Conversation... Hansen on Risk, Ambiguity, and Measurement

EconTalk Extra
by Amy Willis
Hansen on Risk, Ambiguity, and... Michael Munger on the Sharing ...

In this week's episode, Roberts talks about risk and uncertainty with 2013 Nobel laureate, Lars Peter Hansen.

What are your thoughts on this topic? Use the prompts below the fold to join our conversation online, or start your own offline. Let us know what you think. We love to hear from you!

Check Your Knowledge:

1. What is regulatory capture, which Roberts notes as a concern he has about the use of models of risk? To what extent is Roberts's concern justified?

2. What are the different components of uncertainty, according to Hansen?

Going Deeper:

3. Roberts asks Hansen in what areas within economics have we improved the precision of our knowledge. Do you agree with his response? Why? Are there other areas you might suggest that have recently benefited from greater quantitative precision?

4. Nassim Taleb tells the story of the person lost in Paris who finds a map of New York and thinks, well, it's better than nothing. It has to help. His point is that false precision or scientism isn't just unhelpful, it can be dangerous because it leads you to think that know more than you do. Hansen clearly recognizes the limitations of formal models and statistical techniques. Do you think the economics profession has sufficiently internalized this lesson? How might economists insulate themselves more effectively from self-deception?

Extra Credit:

5. Taking points from Hansen's Nobel address and/or Hayek's Nobel address, answer the following question: To what extent is economics a scientific enterprise? Do quantitative models add to our stock of knowledge, or do they deceive us into thinking economics is more scientific than it really is?

Comments and Sharing

TWITTER: Follow Russ Roberts @EconTalker

COMMENTS (3 to date)
Pietro Poggi-Corradini writes:

Have been waiting for the full transcript. I especially want to reread the passage where Hansen said he did not agree with Russ and then Russ brought up the question of where to draw the line between formulas and philosophical heuristics. Clearly people still get interviewed in English and write English words on paper, and don't feel the need to go all out and speak only in formulas.

About 4. I think a map of NYC would be useful to a martian who just landed in Paris and did not know anything about cities, that there are streets and that they have names, mostly meet at straight angles etc...

CG Dong writes:

1. Regulatory capture refers to the tendency that regulatory policy is actually serving special interests and then captured by them. Robert's concern is justified to the extent that if scholars are overconfident on their models, interests group could use them as evidence to support their political arguments and interests at the price of broader social welfare.

2. Three pieces: a first piece that we can quantity and identify with data, like 'unknown' parameters and their distribution; a second piece (related to model selection) that we cannot know for sure but may be able to assign different weights on; and the last piece on that the models we know are wrong, but have to live with it and make important decisions.

3. Hansen based on his arguments on the monetary transmission mechanism, which I am no expert on. But I tend to agree with his response. Scientific advancement could be like an upward spiral, sometimes slow and sometimes fast. But measurement and modeling process are getting better and better. One other area that benefits from such advancement is predicting presidential election results.

4. It has not been enough for quantitative researchers to recognize their limitations of their methods, but more and more people will recognize those problems. Discussions on p-hacking and researcher degrees of freedom are getting popular. Prof. Manski has also talked extensively about model selection and parameter uncertainties. Economists can really benefit a lot from more lectures on model selection and identification issues rather than different estimation techniques.

5. Economics is like a combination of science and art to me, though the scale is more towards the science part right now. I don't think quantitative models add to our stock of knowledge, and it is dangerous to prioritize models over theories. It is always the latter that comes first. Models are developed to test different theories. However, due to the complexities of research topics, we have to make simplifications in modeling process. On the other hand, we have to be very careful about different models out there and potential conflicted predictions on theories. Now, we are back to the model selection problem again. The difference between economics and other sciences is that they know how to and can arrive at a consensus on model selection, but we cannot without precise or complete knowledge.

Brendan writes:

3. I agree with Hansen. Although I admit the limits of models I do agree that models can better visualize information we currently have so economic models evolve more. I can't name other areas of economics that have benefited from quantification but I will give a subject that could use quantification methods is history or more likely financial history.

4. I do think that economists have not learned this lesson completely as evidence by the financial

they haven't yet distinguished physics from real world biology.

5. It's not on either extremes economics is the most scientific of the social sciences as it includes more quantitative information than say psychology. Models do have their strengths as they aid in our visualizing current information and we must think of models like this that our knowledge is never complete and we gain new information as our domain of knowledge increase and our models must demonstrate this.

Comments for this podcast episode have been closed
Return to top