Abstract

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%.

This is the paper behind a question readers of this site ask often: when a meta-analysis corrects for publication bias and other observable biases, how far does the number move from what the field believed? The answer here is a magnitude, not an anecdote. Across the influential meta-analyses reviewed, corrected effects sit typically 45% to 60% below the conventional wisdom they were tested against, they are much closer to zero, and in some literatures they change sign.

Research revision across 24 economics literatures For each of 24 literatures reviewed by Gechert et al. (2025), the corrected or best-practice mean compared with the conventional wisdom of a seminal study. The median revision is -62%. Two literatures changed sign; one, capital-energy substitution, rose by 7%. no change −100% or beyond Minimum wage and employment. Meta-analysis: Doucouliagos and Stanley (2009), corrected mean 0.04. Conventional wisdom of the seminal study named in the paper’s Table 2: -0.08. The corrected estimate has the opposite sign. Minimum wage and employment Doucouliagos and Stanley (2009) sign flip Gender gap in performance pay. Meta-analysis: Bandiera et al. (2021), corrected mean 0.07. Conventional wisdom of the seminal study named in the paper’s Table 2: -0.28. The corrected estimate has the opposite sign. Gender gap in performance pay Bandiera et al. (2021) sign flip Active labour market policies. Meta-analysis: Vooren et al. (2019), corrected mean 0.004. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.03. -87% revision. Active labour market policies Vooren et al. (2019) Do poor countries catch up?. Meta-analysis: Abreu et al. (2005), corrected mean 0.3. Conventional wisdom of the seminal study named in the paper’s Table 2: 2. -85% revision. Do poor countries catch up? Abreu et al. (2005) Intertemporal substitution. Meta-analysis: Havranek (2015), corrected mean 0.07. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.5. -85% revision. Intertemporal substitution Havranek (2015) Present bias. Meta-analysis: Imai et al. (2021), corrected mean 0.01. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.07. -83% revision. Present bias Imai et al. (2021) Fiscal multipliers. Meta-analysis: Gechert (2015), corrected mean 0.61. Conventional wisdom of the seminal study named in the paper’s Table 2: 2.5. -75% revision. Fiscal multipliers Gechert (2015) Public capital and productivity. Meta-analysis: Bom and Ligthart (2014), corrected mean 0.11. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.39. -73% revision. Public capital and productivity Bom and Ligthart (2014) Business cycles and the euro area. Meta-analysis: Fidrmuc and Korhonen (2006), corrected mean 0.16. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.6. -73% revision. Business cycles and the euro area Fidrmuc and Korhonen (2006) Immigration and wages. Meta-analysis: Longhi et al. (2005), corrected mean -0.04. Conventional wisdom of the seminal study named in the paper’s Table 2: -0.15. -72% revision. Immigration and wages Longhi et al. (2005) Currency unions and trade. Meta-analysis: Rose and Stanley (2005), corrected mean 0.39. Conventional wisdom of the seminal study named in the paper’s Table 2: 1.2. -68% revision. Currency unions and trade Rose and Stanley (2005) Tax rates and FDI. Meta-analysis: Feld and Heckemeyer (2011), corrected mean 1.74. Conventional wisdom of the seminal study named in the paper’s Table 2: 4.79. -64% revision. Tax rates and FDI Feld and Heckemeyer (2011) Capital-labor substitution. Meta-analysis: Gechert et al. (2022), corrected mean 0.3. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.75. -60% revision. Capital-labor substitution Gechert et al. (2022) Skilled-unskilled labor substitution. Meta-analysis: Havranek et al. (2022), corrected mean -0.27. Conventional wisdom of the seminal study named in the paper’s Table 2: -0.67. -60% revision. Skilled-unskilled labor substitution Havranek et al. (2022) Value of a statistical life. Meta-analysis: Doucouliagos et al. (2012), corrected mean 1.66. Conventional wisdom of the seminal study named in the paper’s Table 2: 3.9. -57% revision. Value of a statistical life Doucouliagos et al. (2012) FDI spillovers to local firms. Meta-analysis: Havranek and Irsova (2011), corrected mean 0.18. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.38. -53% revision. FDI spillovers to local firms Havranek and Irsova (2011) Financial education. Meta-analysis: Kaiser et al. (2022), corrected mean 0.13. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.23. -43% revision. Financial education Kaiser et al. (2022) Distance and bilateral trade. Meta-analysis: Disdier and Head (2008), corrected mean 0.8. Conventional wisdom of the seminal study named in the paper’s Table 2: 1.3. -38% revision. Distance and bilateral trade Disdier and Head (2008) Agglomeration and productivity. Meta-analysis: Melo et al. (2009), corrected mean 0.04. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.06. -35% revision. Agglomeration and productivity Melo et al. (2009) Energy price elasticity. Meta-analysis: Labandeira et al. (2017), corrected mean -0.53. Conventional wisdom of the seminal study named in the paper’s Table 2: -0.8. -34% revision. Energy price elasticity Labandeira et al. (2017) Returns to a year of schooling. Meta-analysis: Ashenfelter et al. (1999), corrected mean 0.07. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.09. -24% revision. Returns to a year of schooling Ashenfelter et al. (1999) The wage curve. Meta-analysis: Nijkamp and Poot (2005), corrected mean -0.08. Conventional wisdom of the seminal study named in the paper’s Table 2: -0.1. -23% revision. The wage curve Nijkamp and Poot (2005) Value of lake ecosystems. Meta-analysis: Reynaud and Lanzanova (2017), corrected mean 153. Conventional wisdom of the seminal study named in the paper’s Table 2: 153. +0% revision. Value of lake ecosystems Reynaud and Lanzanova (2017) Capital-energy substitution. Meta-analysis: Koetse et al. (2008), corrected mean 0.46. Conventional wisdom of the seminal study named in the paper’s Table 2: 0.43. +7% revision. Capital-energy substitution Koetse et al. (2008) 24 meta-analyses reviewed in Gechert et al. (2025), J Econ Surveys meta-analysis.cz
What correcting for bias does to a number. Each bar is one of the 24 meta-analyses reviewed in this paper, almost all of them by other researchers — four of the 24 are Havránek’s own. Each compares that meta-analysis’s corrected or best-practice mean with the conventional wisdom of the seminal study named in the paper’s Table 2. Red: the corrected effect is smaller in absolute magnitude than the field believed. Green: larger. Two of them — the minimum wage and gender differences in response to performance pay — came out with the opposite sign, and revised by more than 100%, so their bars are capped at the end of the scale and marked sign flip. The median revision is -62% against the seminal study, -60% against an AI's summary of the prior literature, and -50% against the literature's own simple mean. Numbers from its Tables 2 and 3; 10.1111/joes.12630.

That is the general result which the individual papers collected on this site illustrate one literature at a time. It is also the reason the headline results page reports a corrected figure wherever the paper provides one, and says so where it does not.

Open access. The copy here is the published version, released by the journal under CC BY-NC-ND 4.0 — not the CC BY 4.0 that covers the rest of this site, because the licence on this one is Wiley’s to set and ten authors hold the copyright. Cite it, quote it, redistribute it whole; for a commercial use or an adaptation, ask. The publisher’s page carries the record of version.

Reference: Sebastian Gechert, Bianka Mey, Matej Opatrny, Tomas Havranek, T. D. Stanley, Pedro R. D. Bom, Hristos Doucouliagos, Philipp Heimberger, Zuzana Irsova, Heiko J. Rachinger (2025), “Conventional wisdom, meta-analysis, and research revision in economics.” Journal of Economic Surveys 39, 980–999. 10.1111/joes.12630

Read the paper (PDF) · Published version