regression analysis
A statistical method for estimating relationships among variables; running many regressions enables cherry-picking results that confirm a bias.
As defined in Nassim Nicholas Taleb on Black Swans
Other definitions
A statistical technique estimating how a dependent variable (e.g., earnings) responds to inputs (currency, oil, rates), used here to forecast forward earnings.
Stan Druckenmiller | Podcast | In Good Company | Norges Bank Investment Management
Where it comes up · 9
- Jim Manzi on Knowledge, Policy, and Uncontrolled3 claims
- Campbell Harvey on Randomness, Skill, and Investment Strategies3 claims
- Future of Finance1 claim
- Conversations with Tyler - Arthur Brooks1 claim
- Nate Silver1 claim
- Twenty Years of Freakonomics (with Stephen Dubner)1 claim
- Nassim Nicholas Taleb on Black Swans1 claim
- David Boaz, P.J. O'Rourke, and George Will on the State of Liberty1 claim
and 1 more
Most important claims · 10 of 13Search all claims for “regression analysis” →
- Ed Leamer's response to the cherry-picking problem in econometrics is sensitivity analysis: examine all combinations of plausible right-hand-side variables, and only trust a result (e.g., that capital punishment deters murder) if the estimated effect remains in a narrow band across specifications rather than flipping sign with variable choice.Campbell Harvey on Randomness, Skill, and Investment Strategies
- George Stigler noted that young economists once ran only three or four regressions by hand and thought hard about which variables to test, but today you can run hundreds of regressions and easily convince yourself the 'good' ones are valid, so cheap computation lets us be fooled by randomness and select results that confirm our bias.Nassim Nicholas Taleb on Black Swans
- A researcher at a top investment bank presented the 17th monthly lag of US Industrial Production as a significant market predictor, but admitted it was 'the only one that worked' out of roughly 24 lag tests, so adjusting significance for the number of tests reveals it as insignificant—a finance version of the jellybean problem.Campbell Harvey on Randomness, Skill, and Investment Strategies
- In regression analysis of charitable giving with covariates including political ideology, religious behavior, income, and demographics, the large and highly significant coefficient is religious activity, not politics—so conservatives gave more in 2006-2008 mainly because they were more religious, and as the conservative wing becomes more secular this giving gap is disappearing.Conversations with Tyler - Arthur Brooks
- Economists never disclose how many regressions they ran before obtaining a dramatic result—it is treated as personal—but requiring authors to publish or at least count all the specifications that failed or were unconvincing would have an interesting and salutary effect on the profession's reliability.Twenty Years of Freakonomics (with Stephen Dubner)
- Macroeconomic empirical work is fundamentally unreliable because cherry-picking of sample size, time period, and variables can prove that Keynesianism or monetary policy works or doesn't, and since researchers never disclose how many regressions they ran, many findings will not replicate going forward.Campbell Harvey on Randomness, Skill, and Investment Strategies
- Raw regression analysis across states over 50-year periods shows that states with higher minimum wages have lower teen employment, supporting the traditional consensus; however, revisionist literature adds statistical controls that make the minimum wage's explanatory power disappear, which Murphy argues may represent over-controlling and assumption-making rather than genuine causality.Top 9 Economic Lessons of the Past 35 Years | InFi #54
- Agent-based modeling—simulating underlying mechanisms, as weather forecasting builds a physical model of the atmosphere resolved mathematically rather than statistically—can yield insights regression cannot when you truly know how agents behave as a system, but if your assumptions are even somewhat wrong things can go haywire fast, which is why testing models against weird edge-case inputs matters.Nate Silver
- In systems with high causal density—many simultaneous causes acting on outcomes—plausible causal stories are usually true for at least some individuals, but millions of other causes act simultaneously, so non-experimental methods like regression cannot actually hold everything else constant and isolate the effect of one intervention; we kid ourselves when we think we have isolated cause and effect.Jim Manzi on Knowledge, Policy, and Uncontrolled
- The relevant practical standard for an analysis is not whether it is the best available analysis but whether it adds value over alternative analyses or expert judgment; in many policy cases regression analysis fails to clear the hurdle of practicality because it produces no useful information beyond what is already available without it.Jim Manzi on Knowledge, Policy, and Uncontrolled