Tomas Havranek’s research while affiliated with Centre for Economic Policy Research - CEPR- and other places

Conventional meta-analyses (both fixed and random effects) of correlations are biased due to the mechanical relationship between the estimated correlation and its standard error. Simulations that are closely calibrated to match actual research conditions widely seen across correlational studies in psychology corroborate these biases and suggest two solutions: UWLS +3 and HS. UWLS +3 is a simple inverse-variance weighted average (the unrestricted weighted least squares) that adjusts the degrees of freedom and thereby reduces small-sample bias to scientific negligibility. UWLS +3 as well as the Hunter and Schmidt approach (HS) are less biased than conventional random-effects estimates of correlations and Fisher’s z , whether or not there is publication selection bias. However, publication selection bias remains a ubiquitous source of bias and false-positive findings. Despite the relationship between the estimated correlation and its standard error in the absence of selective reporting, the precision-effect test/precision-effect estimate with standard error (PET-PEESE) nearly eradicates publication selection bias. Surprisingly, PET-PEESE keeps the rate of false positives (i.e., type I errors) within their nominal levels under the typical conditions widely seen across psychological research whether there is publication selection bias, or not.

Meta-analysis upweights studies reporting lower standard errors and hence more precision. But in observational settings common to much research in social sciences, precision is not given to the researcher. Precision must be estimated, and thus can be p-hacked to achieve statistical significance. Simulations and applications show that spurious precision can invalidate inverse-variance weighting and bias-correction methods based on the funnel plot. Selection models fail to solve the problem, and common cures to publication bias can become worse than the disease. We introduce a novel approach that addresses spuriousness: meta-analysis instrumental variable estimator (MAIVE).

Estimates of relative risk aversion vary widely, but no study has attempted to quantitatively trace the sources of the variation. We collect 1021 estimates from 92 studies that use the consumption Euler equation to measure relative risk aversion and that disentangle it from intertemporal substitution. We show that calibrations of risk aversion are systematically larger than estimates thereof. Moreover, reported estimates are systematically larger than the underlying risk aversion because of publication bias. After correction for the bias, the literature suggests a mean risk aversion of 1 in economics and 2–7 in finance contexts. The reported estimates are driven by the characteristics of data (frequency, dimension, country, stockholding) and utility (functional form, treatment of durables). To obtain these results, we use recently developed techniques to correct for publication bias and Bayesian model averaging techniques to account for model uncertainty.

Research Question/Issue This study identifies the determinants of shareholder value created by investor activism. It quantifies and corrects the pool of published empirical estimates for bias due to the selective reporting of empirical results. It examines how various institutional, investor, and research design characteristics affect the shareholder value created by activism. Research Findings/Insights Using a meta‐analysis of 1973 estimates from 67 studies published between 1983 and 2021, we document that, after adjusting for selective reporting, activism creates a positive shareholder value ranging from 0% to 1.5%, which is smaller than commonly thought. More value is created in settings with better shareholder rights protection and in smaller markets. Activism by individual investors, more confrontational approaches, and campaigns aimed at the sale of the target company enhance firm value more. Theoretical/Academic Implications Our study identifies a specific channel through which the quality of institutional settings mitigates the conflict envisaged by agency theory between firm owners and managers. Our comprehensive synthesis of prior research literature also guides researchers in interpreting and comparing results reported in prior studies and helps them make more informed research design choices in future studies. Practitioner/Policy Implications Exploiting the heterogeneity in prior studies, our study informs investors about the value that specific forms of activism create, allowing them to better trade off the costs and benefits of alternative approaches. It also informs regulators about the relevance of institutional characteristics and offers estimates of the relative effectiveness of activism campaigns by various sponsors. These insights help in seeking optimal ways of regulating corporate governance and financial markets.

Over the past several decades, meta-analysis has emerged as a widely accepted tool to understand economics research. Meta-analyses often challenge the established conventional wisdom of their respec- tive fields. We systematically review a wide range of influential meta-analyses in economics and compare them to “conventional wisdom.” After correcting for observable biases, the empirical economic effects are typically much closer to zero and sometimes switch signs. Typically, the relative reduction in effect sizes is 45%–60%.

We examine the factors influencing published estimates of hedge fund performance. Using a sample of 1,019 intercept terms from regressions of hedge fund returns on risk factors (the alphas) collected from 74 studies, we document a strong downward trend in the reported alphas. The trend persists even after controlling for heterogeneity in hedge fund characteristics and research design choices in the underlying studies. Estimates of current performance implied by best practice methodology are close to zero across all common hedge fund strategies. Additionally, our data allow us to estimate the mean management and performance fees charged by hedge funds. We also document how reported performance estimates vary with hedge fund and study characteristics. Overall, our findings indicate that, while hedge funds historically generated positive value for investors, their ability to do so has diminished substantially.

Common wisdom suggests that beauty helps in the labor market. We show that two factors combine to explain away the mean beauty premium reported in the literature. First, correcting for publication bias reduces the premium by at least a third. Second, controlling for cognitive ability negates the premium for all occupations except sex workers, a point further underscored by the similarity of the beauty effect on earnings and productivity. The second factor implies a positive link, perhaps genetic, between beauty and intelligence. We find little evidence of substantial attenuation bias that could offset publication and omitted-variable biases. The empirical literature is inconsistent with discrimination based solely on tastes for beauty. To obtain these results we collect 1,159 estimates of the effect of beauty on earnings or productivity reported in 67 studies and codify 33 aspects that reflect estimation context, including the potential intensity of attenuation bias. We employ recently developed techniques to account for publication bias and model uncertainty.

We demonstrate that all meta‐analyses of partial correlations are biased, and yet hundreds of meta‐analyses of partial correlation coefficients (PCCs) are conducted each year widely across economics, business, education, psychology, and medical research. To address these biases, we offer a new weighted average, UWLS +3 . UWLS +3 is the unrestricted weighted least squares weighted average that makes an adjustment to the degrees of freedom that are used to calculate partial correlations and, by doing so, renders trivial any remaining meta‐analysis bias. Our simulations also reveal that these meta‐analysis biases are small‐sample biases ( n < 200), and a simple correction factor of ( n − 2)/( n − 1) greatly reduces these small‐sample biases along with Fisher's z. In many applications where primary studies typically have hundreds or more observations, partial correlations can be meta‐analyzed in standard ways with only negligible bias. However, in other fields in the social and the medical sciences that are dominated by small samples, these meta‐analysis biases are easily avoidable by our proposed methods.

Over the past several decades, meta-analysis has emerged as a widely accepted tool to understand economics research. Meta-analyses often challenge the established conventional wisdom of their respective fields. We systematically review a wide range of influential meta-analyses in economics and compare them to ‘conventional wisdom.’ After correcting for observable biases, the empirical economic effects are typically much closer to zero and sometimes switch signs. Typically, the relative reduction in effect sizes is 45-60%.