Pith. sign in

REVIEW 3 major objections 5 minor 116 references

Fighting discrimination with reputation: The case of online platforms

T0 review · 3 major / 5 minor · reviewed 2026-07-11 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read On a French ridesharing platform, new minority drivers earn 11.6% less until reviews erase most of the gap.

desk verdict Strong reduced-form and strike evidence that reputation closes minority revenue gaps; absolute prior-bias euros rest on normalizations that are not independently tested. read the letter →

arxiv 2607.05627 v1 pith:YZ522O3G submitted 2026-07-06 econ.GN q-fin.EC

Xavier Lambin , Emil Palikot This is my paper · ORCID
classification econ.GNq-fin.EC
keywords discriminationreputationsystemsridesharingstatisticalcareerconcernsbelief-basedpartialityonlineplatformsreviewaccumulation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper shows that ethnic pay gaps on a large French ridesharing platform are not a fixed taste-based tax but an information problem that reputation largely corrects. Minority drivers with almost no reviews earn about 11.6% less revenue than similar nonminority drivers; the gap shrinks toward statistical insignificance once they have roughly twenty reviews. A railway strike that temporarily tripled demand and sped up review accumulation caused minority entrants to gain the most afterward, which the authors read as causal evidence that reviews, not just selection, close the gap. A structural model of passenger choice and driver career concerns recovers the mechanism: passengers start with overly pessimistic priors about minority entrants, so those drivers cut introductory prices and work harder to force belief revision. Counterfactuals imply that correcting the prior would raise minority entry value by about 5.8%, that locking the bias in would widen the lifetime wedge, and that a sharper rating system helps but cannot replace getting the prior right.

What carries the argument

A dynamic career-concerns model of passenger choice and driver pricing and effort, with group-specific possibly biased priors, Bayesian updating from noisy reviews, and an oblivious equilibrium for competition. Demand coefficients map residual minority utility penalties into grade-scale belief gaps; post-burnout grades recover types, noise precision, and effort; counterfactuals re-solve the equilibrium under correct priors, permanent bias, or sharper or noisier ratings.

What would settle it

If, after matching on displayed ratings and experience, an independent measure of passenger expected quality before booking (surveys or booking experiments that hide identity) showed no systematic minority under-expectation relative to realized post-ride ratings, the recovered prior gap and the correct-prior welfare gain would collapse.

Watch

Extended reading notes

Core claim

Passengers hold systematically too-pessimistic priors about minority entrants: the market expects quality around 3.06 on a five-point scale before the ride, while realized early ratings are about 4.92. That belief gap, not only a small true type difference of about 0.1 grade, drives the 11.6% entry revenue penalty. Minority drivers respond by discounting prices more and exerting more effort; as reviews accumulate, posteriors converge and discrimination fades. The reputation system therefore strictly benefits minority drivers relative to a world of permanent bias.

Load-bearing premise

The absolute size of the market's pessimism about minority entrants is pinned down only if one accepts that passengers are unbiased about nonminority drivers and that any residual minority penalty vanishes for drivers with many reviews.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper studies ethnic discrimination and reputation on BlaBlaCar. Reduced-form evidence shows minority entrants earn about 11.6% less revenue than similar nonminority drivers (AIPW ATT), with the gap shrinking as reviews accumulate and becoming small beyond roughly 15–20 reviews. A 2018 railway-strike DiD (Sant’Anna–Zhao doubly robust) finds that strike-day exposure raises post-strike revenue more for minority entrants than for nonminority entrants (about 61% larger). Descriptive patterns show early-career price discounts and elevated five-star shares that decline with reviews. A structural career-concerns model with group-specific priors, competition, and strategic price/effort is estimated via demand logit with SmartStop/fuel control functions, post-burnout types and effort residuals, static cost inversion for experienced drivers, and an oblivious equilibrium. Estimates imply passengers under-expect minority entrant quality (prior ~3.06 vs early realized grades ~4.92); minority entrants discount more and exert more effort. Counterfactuals attribute roughly half the entry-value wedge to incorrect priors and show that more informative ratings narrow but do not eliminate the wedge.

Significance. If the reduced-form and structural results hold, the paper makes a substantial contribution to discrimination and platform design. It links career-stage outcome gaps to a natural experiment on review accumulation, recovers supply-side reputation investment (introductory discounts and effort), and quantifies how much of the entry gap is unwarranted pessimism versus a statistical-discrimination floor. The combination of AIPW gaps by experience, strike DiD with balance checks, and an estimated OE with explicit counterfactuals on priors and rating precision is rare in this literature and would be of clear interest to applied micro, IO, and digital-platform audiences. Strengths include transparent double-robust reduced form, documented regime changes in the rating system, and sensitivity of supply primitives to the burnout cutoff.

major comments (3)
  1. [§7.1, §8.2, §9, Table 5] Section 7.1 and Section 8.2: absolute prior bias (market expects 3.06 for minority entrants; 1.86-grade underestimation) and the correct-prior / persistent-bias euro magnitudes in Table 5 and Section 9 rest on two normalizations—unbiased nonminority beliefs at every experience bin (expected quality equals displayed grade) and zero residual minority penalty for 21+ drivers. Relative within-bin minority penalties are identified from demand; the absolute grade-scale map κ̂1/α_grade and the 5.8% entry-value gain are not. The paper itself reports nonminority first-grade means of 4.94 against an anchored nonminority entry belief of 4.71; that residual is attributed to effort/noise only under normalization (ii). Please either (a) provide independent validation or bounds under alternative anchors (e.g., allow nonminority entry bias, or a nonzero 21+ residual), or (b) reframe the headline structu
  2. [§4, Table 3] Section 4 and Table 3: the strike design is a reduced-form ITT of strike-day exposure, not an instrumented effect of an additional review. The manuscript interprets the 61% larger minority-entrant ATT as the causal value of accelerated review accumulation. That reading is plausible given balance (Figure 4, Appendix K Table 14), post-strike common demand, and null effects for experienced drivers, but alternative channels (learning about the platform, network effects, route selection after the shock) are not fully ruled out. Please either instrument review counts (or seats sold during the window) more explicitly, or state more carefully that the estimand is exposure to strike-induced demand and that the review channel is the preferred interpretation rather than a directly identified review effect.
  3. [§7.2–7.4, Appendix P, §8.4, Table 5] Section 7.2–7.4 and Appendix P: types and effort are recovered from post-t*=20 grades under negligible post-burnout effort; Appendix P shows the type gap is stable over t*∈[10,30], which is reassuring, but the joint F-test of post-t* residuals on competitors, relative price, hours-to-departure, and strike (F=0.37) is underpowered if residual strategic effort is small. Because equilibrium effort, introductory discounts, and entry values inherit these primitives, please report how the main OE paths and Table 5 counterfactuals move when t* and the min post-cutoff review threshold vary over the ranges already explored in Appendix P, not only the type moments.
minor comments (5)
  1. [§8.2] Clarify early in Section 8 that the 1.86-grade underestimation combines the demand-side grade-scale penalty (~1.65) with the nonminority anchor and realized early grades, so readers do not treat 3.06 as a pure demand object.
  2. [§6.1, Assumption 1] Assumption 1 (no expected future competition against the same rivals) is strong for a platform with repeated route-day markets; a short discussion of when OE is a good approximation would help.
  3. [Figure 2, Table 2] Figure 2 and Table 2 use slightly different experience bins (deciles vs 0–5 / 6–15 / 15+); aligning labels would reduce reader friction.
  4. [§9.3, Appendix B] Appendix B’s regime-change evidence is useful for the τ_ϵ counterfactual; a one-sentence pointer in Section 9.3 to the R² informativeness results in Table 6 would strengthen the link.
  5. [§7–9, Appendices] Notation: both τ_ϵ and h_ϵ appear for review-noise precision across main text and appendices; standardize.

Circularity Check

1 steps flagged · score 2.0 of 10

No load-bearing circular derivation: absolute prior levels rest on explicit normalizations, but minority demand penalties, grades, and strike effects are independent data objects.

  1. self definitional [Section 7.1 (From coefficients to beliefs); Section 8.2]
    "First, the omitted-reference choice for the 21+ bin sets the residual minority coefficient there to zero, so the market’s expected quality for an experienced minority driver equals her displayed grade. ... The grade-scale penalty thus declines from 1.65 at entry to 1.00 at 6–20 and, by the normalization, to zero at 21+."

    In the structural demand equation the residual minority utility penalty at 21+ reviews is not estimated; it is set to zero by making 21+ the omitted reference. The model’s statement that the grade-scale belief gap falls to zero for experienced drivers is therefore true by that normalization, not by a free estimate of late-career residual discrimination. (Reduced-form attenuation is still independently estimated; only the structural residual-at-21+ object is definitional.)

full rationale

The paper’s core chain is not circular. Reduced-form minority revenue gaps (AIPW), reputation attenuation (Figure 2/Table 2), and the railway-strike DiD are estimated from distinct outcome and treatment variation. Demand recovers minority×experience utility penalties (κ̂1, κ̂2) from choice data with a price control function; post-burnout grades separately identify type means and noise precision; the strike experiment is external. Mapping κ̂/α_grade into an absolute entrant prior of 3.06 and a 1.86-grade bias does require two normalizations (zero residual minority penalty at 21+; nonminority expected quality equals displayed grade at every bin)—so absolute euro counterfactuals scale with those anchors—but that is identification, not a prediction forced by fitting the same object twice. The structural claim that the residual belief gap is zero at 21+ is partly by the omitted-category normalization (explicitly acknowledged), which is a minor self-definitional step for the structural path only; reduced-form convergence is still independently estimated. Effort residuals (early grade − type) and OE-solved effort are standard structural recovery, not renaming of a fitted target as an out-of-sample prediction. No self-citation uniqueness theorem or ansatz chain carries the result. Score 2 for the one explicit normalization that sets structural late-career residual discrimination to zero by construction.

Assumptions & free parameters 6 free parameters · 7 assumptions · 2 invented entities

The reduced-form strike result needs exogeneity of strike-day driving and interpretation via review stocks. The structural claim about incorrect priors and reputation value needs logit demand with control-function price, Bayesian updating with group priors, burnout identification of types/effort, static FOC costs for experienced drivers, and oblivious equilibrium with calibrated entry/exit. Free parameters (discount factor, burnout cutoff, noise precision, demand coefficients, λ_g, ζ) pin magnitudes; domain assumptions (no dynamic multi-driver strategic feedback; nonminority priors unbiased; post-t* effort ~0) pin identification of absolute bias and effort.

free parameters (6)
  • discount factor δ = 0.96
    Calibrated at 0.96 for dynamic effort/pricing; shapes reputation investment and entry values.
  • burnout cutoff t* = 20
    Chosen at 20 reviews so post-cutoff mean grade identifies type and residual identifies effort; Appendix P varies 10–30.
  • review-noise precision τ_ϵ = 2.74
    Estimated from within-driver post-burnout grade variance; sets Bayesian weight of each review and rating-design counterfactuals.
  • demand coefficients (γ, α, ψ, κ1, κ2) = γ≈-0.106, α≈0.352, κ1≈-0.145, κ2≈-0.088
    Estimated by conditional logit with control function; κ/α_grade maps to grade-scale belief gaps central to the prior-bias claim.
  • exit hazard ζ and entry rates λ_g = ζ≈0.59 per listing-period
    Calibrated from listing activity for OE steady-state competitor distribution; affect long-run ¯s and welfare levels.
  • effort cost function g(a) = g(a)≈0.0604a+0.1230a²
    Quadratic approximation fitted to map incentives to observed effort residuals (Appendix O).
assumptions (7)
  • domain assumption Passengers have Type-I extreme value shocks and choose by multinomial logit on expected quality and price (McFadden).
    Section 6–7 demand; standard but rules out flexible substitution patterns that could change minority penalties.
  • ad hoc to paper Market prior about nonminority drivers is unbiased at every experience bin; residual minority penalty is zero for 21+ reviews.
    Section 7.1 normalizations that convert relative utility gaps into absolute grade-scale priors and bias ˆμ_g−μ_g.
  • ad hoc to paper Post-burnout effort is negligible so mean post-t* grade equals intrinsic type η_i.
    Section 7.2; joint F-test and t* sensitivity offered as support, still a substantive restriction.
  • ad hoc to paper Assumption 1: drivers do not expect to compete against the same rivals in future periods when choosing p_t and a_t.
    Section 6.1; enables single-agent OE career concerns without dynamic oligopoly feedback.
  • domain assumption Strike-day driving is as-good-as-random conditional on observables; post-strike effects operate through review accumulation.
    Section 4; balance checks support exogeneity, but treatment is exposure not instrumented review count.
  • standard math Bayesian normal-normal updating of driver type from quality reports net of equilibrium effort.
    Section 6.2; standard conjugate learning used to define posteriors and Proposition 3 convergence.
  • domain assumption Unconfoundedness for AIPW minority effects given rich covariates (causal forest propensity).
    Section 3.1; authors note it is unverifiable and complement with the strike design.
invented entities (2)
  • Group-specific market prior ˆμ_g (belief-based partiality, possibly biased) independent evidence
    purpose: Separates passenger priors from true type means and defines discrimination D(p,q) vs incorrect beliefs.
    Standard statistical-discrimination object, but the paper’s contribution is recovering a large biased ˆμ_m from platform data under stated normalizations; independent handle is comparison to realized grades and post-burnout types.
  • Oblivious-equilibrium competitor distribution ¯s with listing-period driver clock
    purpose: Compute dynamic prices, effort, and entry values under many drivers without full MPE.
    Adaptation of Weintraub et al. OE; computational device rather than a new physical entity, but welfare numbers depend on it.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fighting discrimination with reputation: The case of online platforms." pith.science (2026). https://pith.science/paper/YZ522O3G

@misc{pith2026260705627,
  author       = {Pith},
  title        = {Pith review of: Fighting discrimination with reputation: The case of online platforms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YZ522O3G}},
  note         = {Machine review of arXiv:2607.05627}
}
read the original abstract

On a large French ridesharing platform, new minority drivers earn 11.6% less revenue than otherwise similar nonminority drivers; the gap nearly vanishes as they accumulate reviews. Reviews drive the convergence: when a railway strike exogenously raised demand and sped up review accumulation, minority entrants gained the most. We explain the pattern with an estimated model of passenger choice and driver career concerns. Passengers hold overly pessimistic priors about minority entrants - expecting substantially lower quality before the ride than they report after it. As a result, minority drivers cut introductory prices and exert extra effort to overturn those beliefs quickly. Counterfactuals show the cost of incorrect priors is high, and the reputation system strictly benefits minority drivers.

Figures

Figures reproduced from arXiv: 2607.05627 by the authors.

Figure 6
Figure 6. Market prior beliefs and realized entrant grades Note: Shaded densities – distribution of driver intrinsic types (post-burnout mean grade, t > t ∗ ) for minority (red) and nonminority (blue) drivers. Solid vertical lines – market prior belief about entrant quality for each group, computed from the final demand specifica￾tion. Dashed vertical lines – mean of the first two grades received by entrants of each group. Th… view at source ↗
Figure 7
Figure 7. Marginal costs Note: Marginal costs recovered from equation (14) on markets where every listed driver has at least 21 reviews. Selected routes from Paris. 8.4 Equilibrium prices, effort, and value Effort over the career. Equilibrium effort at entry is a ∗ min,0 = 0.43 against a ∗ non,0 = 0.35 grade points, a 22% minority gap. Averaged over t ∈ [0, 5], minority effort is 0.32 against 0.26 (+24%); it is largely exhaus… view at source ↗
Figure 8
Figure 8. Equilibrium effort and prices over a driver’s career (A) Effort (B) Price Notes: Group-flow-weighted equilibrium effort a ∗ g,t (left panel) and prices p ∗ g,t (right panel) across 128 representative markets at the estimated OE, t = 0, . . . , 20. Red — minority; blue — nonminority. The discrimination wedge in entry value. Expected discounted profit at entry is Vmin,0 = 248.15 for minority entrants against Vnon,0 = … view at source ↗
Figures from the paper (6 more)
Figure 9
Figure 9. Figure 9: Navigating BlaBlaCar: search, listing, and driver profile (A) Listing offered on a given route (B) Details of a posting (C) A driver’s profile A1 [PITH_FULL_IMAGE:figures/full_fig_p047_9.png]
Figure 10
Figure 10. Figure 10: Mean latent score and average actual rating by month Notes: Monthly means of the displayed star rating (mean actual rating) and of the review-text sentiment score (mean latent score), 2010–2019. Shaded bands mark the three reputation-system regimes. Displayed ratings …
Figure 12
Figure 12. Figure 12: Classification process for gender and ethnicity (A) Gender (B) Ethnicity Notes: Alluvial diagrams of the classification flow across three stages—by name (Name), by facial recognition (Face), and the combined Final label—for gender (Panel A) and ethnicity (Panel B). Ba…
Figure 13
Figure 13. Figure 13: is consistent with this mechanism. It plots the minority share of visible listings against the number of days until departure at the moment each listing is observed. The share is highest close to departure — approaching 18% a day or two before the ride — precisely the…
Figure 17
Figure 17. Figure 17: Number of drivers with type estimates across cutoff combinations Notes: Each cell reports the number of drivers satisfying the indicated cutoff combination. Darker shading corresponds to larger samples. Discussion. The type gap µˆnon − µˆmin is remarkably stable acros…
Figure 18
Figure 18. Figure 18: Within-driver price calibration: observed versus predicted deviations Note: Listings on the 128 solved routes with n ≤ 20 reviews (96,033 listings from 31,700 drivers observed at least twice). Each listing is assigned the OE-predicted price for its (route, g, n) cell;…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

116 extracted references · 6 canonical work pages

  1. [1]

    American Economic Review , volume=

    The dynamics of discrimination: Theory and evidence , author=. American Economic Review , volume=

  2. [2]

    1995 , journal =

    Automobile Prices in Market Equilibrium , author =. 1995 , journal =

    1995

  3. [3]

    Brookings India , title =

    Yaraghi, Niam and Ravi, Shamika , doi =. Brookings India , title =

  4. [4]

    Achenfelter, A

    The theory of discrimination, discrimination in labor markets , author=. Achenfelter, A. Ress (eds.), Princeton--New Jersey , year=

  5. [5]

    Information Economics and Policy , year=

    What drives pricing behavior in Peer-to-Peer markets? Evidence from the carsharing platform blablacar , author=. Information Economics and Policy , year=

  6. [6]

    Journal of Economics and Management Strategy , number =

    Roger, Guillaume and Vasconcelos, Lu??s , doi =. Journal of Economics and Management Strategy , number =

  7. [7]

    Does standardized information in online markets disproportionately benefit job applicants from less developed countries?

    Ajay Agrawal and Nicola Lacetera and Elizabeth Lyons. Does standardized information in online markets disproportionately benefit job applicants from less developed countries?. Journal of International Economics. 2016. doi:https://doi.org/10.1016/j.jinteco.2016.08.003

    work page doi:10.1016/j.jinteco.2016.08.003 2016

  8. [8]

    Journal of Public Economics , volume=

    Racial discrimination in transportation network companies , author=. Journal of Public Economics , volume=. 2020 , publisher=

    2020

Show all 116 references
  1. [9]

    Review of Economic Studies , keywords =

    Jullien, Bruno and Park, In Uck , doi =. Review of Economic Studies , keywords =

  2. [10]

    Journal of the European Economic Association , pages =

    Rochet, Jean-Charles and Tirole, Jean , doi =. Journal of the European Economic Association , pages =. arXiv , arxivId =:/dl.acm.org/citation.cfm?id=2017212.2017217 , isbn =

  3. [11]

    International Journal of Industrial Organization , keywords =

    Jolivet, Gr. International Journal of Industrial Organization , keywords =. doi:10.1016/j.ijindorg.2016.01.003 , issn =

    2016 doi

  4. [12]

    Journal of Economic Perspectives , number =

    Darity, William A and Mason, Patrick L , doi =. Journal of Economic Perspectives , number =

  5. [13]

    Available at SSRN 4360029 , year=

    Riding together: eliciting travelers' preferences for long-distance carpooling , author=. Available at SSRN 4360029 , year=

  6. [14]

    and Tadelis, S

    Nosko, C. and Tadelis, S. , doi =. NBER Working Paper Series , keywords =

  7. [15]

    and Pavan, Alessandro , doi =

    Garrett, Daniel F. and Pavan, Alessandro , doi =. Journal of Political Economy , number =

  8. [16]

    Stata Journal , keywords =

    Blackwell, Matthew and Iacus, Stefano and King, Gary and Porro, Giuseppe , doi =. Stata Journal , keywords =. arXiv , arxivId =:The Stata Journal , isbn =

  9. [17]

    Review of Economics and Statistics , keywords =

    Livingston, Jeffrey A , doi =. Review of Economics and Statistics , keywords =

  10. [18]

    and Rubin, Donald B

    Rosenbaum, Paul R. and Rubin, Donald B. , doi =. Biometrika , keywords =. arXiv , arxivId =:/www.jstor.org/stable/2335942 , isbn =

  11. [19]

    Industrial and Labor Relations Review , number =

    Aigner, Dennis J and Cain, Glen G , doi =. Industrial and Labor Relations Review , number =

  12. [20]

    The Quarterly Journal of Economics , volume=

    Employer learning and statistical discrimination , author=. The Quarterly Journal of Economics , volume=. 2001 , publisher=

    2001

  13. [21]

    American Economic Journal: Applied Economics , number =

    Edelman, BG; and Luca, Michael; and Svirsky, Dan , doi =. American Economic Journal: Applied Economics , number =

  14. [22]

    Heckman, J. J. and Ichimura, H. and Todd, P. E. , doi =. The Review of Economic Studies , pmid =

  15. [25]

    and Myerson, Roger B

    Baron, David P. and Myerson, Roger B. , doi =. Econometrica: Journal of the Econometric Society , number =

  16. [27]

    Foundations and Trends

    Bar-Isaac, Heski and Tadelis, Steven , doi =. Foundations and Trends

  17. [28]

    Journal of Political Economy , number =

    Charles, Kerwin Kofi and Guryan, Jonathan , doi =. Journal of Political Economy , number =

  18. [29]

    Working paper , pages =

    Zervas, G and Proserpio, D and Byers, J , doi =. Working paper , pages =. arXiv , arxivId =:1606.07138 , isbn =

    arXiv

  19. [30]

    International Journal of Industrial Organization , keywords =

    Spagnolo, Giancarlo , doi =. International Journal of Industrial Organization , keywords =

  20. [31]

    Econometrica , volume=

    Matching on the estimated propensity score , author=. Econometrica , volume=. 2016 , publisher=

    2016

  21. [32]

    and King, Gary and Porro, Giuseppe , doi =

    Iacus, Stefano M. and King, Gary and Porro, Giuseppe , doi =. Journal of Statistical Software , title =

  22. [33]

    and Luca, Michael , doi =

    Edelman, Benjamin G. and Luca, Michael , doi =. SSRN Electronic Journal , title =

  23. [34]

    Altonji, Joseph G. , doi =. American Economic Review , number =

  24. [35]

    Econometrica , number =

    Maskin, Eric and Tirole, Jean , doi =. Econometrica , number =

  25. [36]

    The Journal of Industrial Economics , number =

    Fudenberg, Drew and Tirole, Jean , doi =. The Journal of Industrial Economics , number =

  26. [37]

    Transportation Research Part F: Traffic Psychology and Behaviour , keywords =

    Goddard, Tara and Kahn, Kimberly Barsamian and Adkins, Arlie , doi =. Transportation Research Part F: Traffic Psychology and Behaviour , keywords =

  27. [38]

    Sarsons, Heather , journal =

  28. [39]

    The Journal of Industrial Economics , keywords =

    Cabral, Luis Lu. The Journal of Industrial Economics , keywords =. doi:10.1111/j.1467-6451.2010.00405.x , isbn =

    2010 doi

  29. [40]

    doi:10.1257/aer.104.8.2421 , eprint =

    Mayzlin, Dina and Dover, Yaniv and Chevalier, Judith , booktitle =. doi:10.1257/aer.104.8.2421 , eprint =

    doi

  30. [41]

    The RAND Journal of Economics , number =

    Caillaud, Bernard and Jullien, Bruno , doi =. The RAND Journal of Economics , number =

  31. [42]

    Quarterly Journal of Economics , number =

    Farber, Henry S and Gibbons, Robert , doi =. Quarterly Journal of Economics , number =

  32. [43]

    American Economic Journal: Applied Economics , volume=

    Can information reduce ethnic discrimination? Evidence from Airbnb , author=. American Economic Journal: Applied Economics , volume=. 2022 , publisher=

    2022

  33. [44]

    Econometrica , keywords =

    Laffont, Jean-Jacques and Tirole, Jean , doi =. Econometrica , keywords =

  34. [45]

    Journal of Public Economics , keywords =

    Castillo, Marco and Petrie, Ragan and Torero, Maximo and Vesterlund, Lise , doi =. Journal of Public Economics , keywords =

  35. [46]

    International Journal of Industrial Organization , keywords =

    Fan, Ying and Ju, Jiandong and Xiao, Mo , doi =. International Journal of Industrial Organization , keywords =

  36. [48]

    BlaBlaCar , file =

    Mazzella, Fr. BlaBlaCar , file =

  37. [49]

    Coate, Stephen and Loury, Glenn C. , doi =. The American Economic Review , number =

  38. [50]

    The Review of Economic Studies , volume=

    Managerial incentive problems: A dynamic perspective , author=. The Review of Economic Studies , volume=. 1999 , publisher=

    1999

  39. [51]

    Dellarocas, Chrysanthos and Wood, Charles A. , doi =. Management Science , number =. arXiv , arxivId =:1 , isbn =

  40. [52]

    Journal of Economic Theory , keywords =

    Liu, Qingmin and Skrzypacz, Andrzej , doi =. Journal of Economic Theory , keywords =

  41. [54]

    and Alm, James , doi =

    Melnik, Mikhail I. and Alm, James , doi =. The Journal of Industrial Economics , number =

  42. [56]

    Journal of Political Economy , keywords =

    Laffont, Jean-Jacques and Tirole, Jean , doi =. Journal of Political Economy , keywords =

  43. [57]

    American Economic Review , number =

    Phelps, Edmund S , doi =. American Economic Review , number =. arXiv , arxivId =:arXiv:1011.1669v3 , isbn =

    arXiv

  44. [58]

    The RAND Journal of Economics , pages=

    Estimating discrete-choice models of product differentiation , author=. The RAND Journal of Economics , pages=. 1994 , publisher=

    1994

  45. [59]

    2013 , institution=

    Estimating demand for differentiated products with error in market shares , author=. 2013 , institution=

    2013

  46. [60]

    Journal of Political Economy , volume=

    Early starters versus late beginners , author=. Journal of Political Economy , volume=. 1999 , publisher=

    1999

  47. [61]

    Proceedings of the National Academy of Sciences , volume=

    Reputation offsets trust judgments based on social biases among Airbnb users , author=. Proceedings of the National Academy of Sciences , volume=. 2017 , publisher=

    2017

  48. [62]

    Review of Economics and Statistics , volume=

    Statistical discrimination or prejudice? A large sample field experiment , author=. Review of Economics and Statistics , volume=. 2014 , publisher=

    2014

  49. [63]

    Upubliceret paper

    The Gender Earnings Gap in the Gig Economy: Evidence from over a Million Rideshare Drivers , author=. Upubliceret paper. Tilg

  50. [64]

    European Sociological Review , volume=

    Ride with Me -Ethnic Discrimination, Social Markets, and the Sharing Economy , author=. European Sociological Review , volume=. 2018 , publisher=

    2018

  51. [65]

    Journal of Political Economy , volume=

    Ethnic inequality , author=. Journal of Political Economy , volume=. 2016 , publisher=

    2016

  52. [66]

    University of Chicago Press Economics Books , year=

    The Economics of Discrimination , author=. University of Chicago Press Economics Books , year=

  53. [67]

    The Review of Economic Studies , volume=

    How efficiently is capital allocated? Evidence from the knitted garment industry in Tirupur , author=. The Review of Economic Studies , volume=. 2004 , publisher=

    2004

  54. [68]

    The Quarterly Journal of Economics , volume=

    Ethnic divisions and production in firms , author=. The Quarterly Journal of Economics , volume=. 2014 , publisher=

    2014

  55. [69]

    American Economic Review , volume=

    Attention discrimination: Theory and field experiments with monitoring information acquisition , author=. American Economic Review , volume=

  56. [70]

    Journal of international Economics , volume=

    Does standardized information in online markets disproportionately benefit job applicants from less developed countries? , author=. Journal of international Economics , volume=. 2016 , publisher=

    2016

  57. [71]

    Handbook of social economics , volume=

    Theories of statistical discrimination and affirmative action: A survey , author=. Handbook of social economics , volume=. 2011 , publisher=

    2011

  58. [72]

    2019 , institution=

    Best practices for differentiated products demand estimation with pyblp , author=. 2019 , institution=

    2019

  59. [73]

    Journal of Econometrics , volume=

    Improving the performance of random coefficients demand models: the role of optimal instruments , author=. Journal of Econometrics , volume=. 2014 , publisher=

    2014

  60. [74]

    2005 , publisher=

    Optimal statistical decisions , author=. 2005 , publisher=

    2005

  61. [75]

    American Economic Review , volume=

    The economics of predation: What drives pricing when there is learning-by-doing? , author=. American Economic Review , volume=

  62. [76]

    Handbook of industrial organization , volume=

    A framework for applied dynamic analysis in IO , author=. Handbook of industrial organization , volume=. 2007 , publisher=

    2007

  63. [77]

    Management Science , year=

    Reducing discrimination with reviews in the sharing economy: Evidence from field experiments on Airbnb , author=. Management Science , year=

  64. [78]

    , author=

    Inaccurate Statistical Discrimination: An Identification Problem. , author=. Review of Economics & Statistics , volume=

  65. [79]

    The Quarterly Journal of Economics , volume=

    Discrimination as a self-fulfilling prophecy: Evidence from French grocery stores , author=. The Quarterly Journal of Economics , volume=. 2017 , publisher=

    2017

  66. [80]

    American Economic Review , volume=

    Inefficient hiring in entry-level labor markets , author=. American Economic Review , volume=

  67. [81]

    1974 , journal=

    Conditional logit analysis of qualitative choice behavior , author=. 1974 , journal=

    1974

  68. [82]

    The American economic review , volume=

    Economic growth and income inequality , author=. The American economic review , volume=. 1955 , publisher=

    1955

  69. [83]

    Population, Space and Place , year=

    Who can ride along? Discrimination in a German carpooling market , author=. Population, Space and Place , year=

  70. [84]

    Annual Review of Economics , volume=

    Reputation and feedback systems in online platform markets , author=. Annual Review of Economics , volume=. 2016 , publisher=

    2016

  71. [85]

    Journal of Housing Economics , volume=

    The visible host: Does race guide Airbnb rental rates in San Francisco? , author=. Journal of Housing Economics , volume=. 2018 , publisher=

    2018

  72. [86]

    Journal of Econometrics , volume=

    Doubly robust difference-in-differences estimators , author=. Journal of Econometrics , volume=. 2020 , publisher=

    2020

  73. [87]

    American economic review , volume=

    Are Emily and Greg more employable than Lakisha and Jamal? A field experiment on labor market discrimination , author=. American economic review , volume=. 2004 , publisher=

    2004

  74. [88]

    Sociological Science , volume=

    How black are Lakisha and Jamal? Racial perceptions from names used in correspondence audit studies , author=. Sociological Science , volume=. 2017 , publisher=

    2017

  75. [89]

    The Quarterly Journal of Economics , volume=

    The causes and consequences of distinctively black names , author=. The Quarterly Journal of Economics , volume=. 2004 , publisher=

    2004

  76. [90]

    com , author=

    What’s in a Picture?: Evidence of Discrimination from Prosper. com , author=. Journal of Human resources , volume=. 2011 , publisher=

    2011

  77. [91]

    2022 , institution=

    Smiles in profiles: Improving fairness and efficiency using estimates of user preferences in online marketplaces , author=. 2022 , institution=

    2022

  78. [92]

    International Journal of Hospitality Management , volume=

    Reducing racial discrimination in the sharing economy: Empirical results from Airbnb , author=. International Journal of Hospitality Management , volume=. 2022 , publisher=

    2022

  79. [93]

    Social and Personality Psychology Compass , volume=

    Automated classification of demographics from face images: A tutorial and validation , author=. Social and Personality Psychology Compass , volume=. 2020 , publisher=

    2020

  80. [94]

    Observational studies , volume=

    Estimating treatment effects with causal forests: An application , author=. Observational studies , volume=. 2019 , publisher=

    2019

  81. [95]

    Journal of the American statistical Association , volume=

    Estimation of regression coefficients when some regressors are not always observed , author=. Journal of the American statistical Association , volume=. 1994 , publisher=

    1994

  82. [96]

    Journal of the American Statistical Association , volume=

    Estimation and inference of heterogeneous treatment effects using random forests , author=. Journal of the American Statistical Association , volume=. 2018 , publisher=

    2018

  83. [97]

    grf: Generalized random forests , author=

  84. [98]

    Biometrika , volume=

    Dealing with limited overlap in estimation of average treatment effects , author=. Biometrika , volume=. 2009 , publisher=

    2009

  85. [99]

    , title =

    Train, Kenneth E. , title =. 2009 , address =

    2009

  86. [100]

    Halton sequences for mixed logit , author=

  87. [101]

    Journal of marketing research , volume=

    A control function approach to endogeneity in consumer choice models , author=. Journal of marketing research , volume=. 2010 , publisher=

    2010

  88. [102]

    Econometrica , volume=

    Markov Perfect Industry Dynamics with Many Firms , author=. Econometrica , volume=

  89. [103]

    Operations Research , volume=

    Computational Methods for Oblivious Equilibrium , author=. Operations Research , volume=

  90. [104]

    Marketing Science , volume=

    The evolution of discrimination in online markets: How the rise in anti-Asian bias affected Airbnb during the pandemic , author=. Marketing Science , volume=. 2026 , publisher=

    2026

  91. [105]

    The Review of Economic Studies , volume=

    The gender earnings gap in the gig economy: Evidence from over a million rideshare drivers , author=. The Review of Economic Studies , volume=. 2021 , publisher=

    2021

  92. [106]

    Proceedings of the National Academy of Sciences , volume=

    Ride-hailing technology mitigates effects of driver racial discrimination, but effects of residential segregation persist , author=. Proceedings of the National Academy of Sciences , volume=. 2024 , publisher=

    2024

  93. [107]

    Management Science , volume=

    The role of beliefs in driving gender discrimination , author=. Management Science , volume=. 2021 , publisher=

    2021

  94. [108]

    Management Science , volume=

    Explicit and implicit belief-based gender discrimination: A hiring experiment , author=. Management Science , volume=. 2025 , publisher=

    2025

  95. [109]

    The Quarterly Journal of Economics , volume=

    Systemic discrimination: Theory and measurement , author=. The Quarterly Journal of Economics , volume=. 2025 , publisher=

    2025

  96. [110]

    Journal of the European Economic Association , volume=

    Stereotypes and belief updating , author=. Journal of the European Economic Association , volume=. 2024 , publisher=

    2024

  97. [111]

    Journal of Economic Behavior & Organization , volume=

    Discrimination with inaccurate beliefs and confirmation bias , author=. Journal of Economic Behavior & Organization , volume=. 2023 , publisher=

    2023

  98. [112]

    Proceedings of the 2018 ACM Conference on Economics and Computation , pages=

    Reputation inflation , author=. Proceedings of the 2018 ACM Conference on Economics and Computation , pages=

    2018

  99. [113]

    Nature , volume=

    Scale dichotomization reduces customer racial discrimination and income inequality , author=. Nature , volume=. 2025 , publisher=

    2025

  100. [114]

    Management Science , volume=

    Market design choices, racial discrimination, and equitable microentrepreneurship in digital marketplaces , author=. Management Science , volume=. 2026 , publisher=

    2026

  101. [115]

    Science Advances , volume=

    Fighting bias with bias: How same-race endorsements reduce racial discrimination on Airbnb , author=. Science Advances , volume=. 2023 , publisher=

    2023

  102. [116]

    The RAND Journal of Economics , volume=

    Buying reputation as a signal of quality: Evidence from an online marketplace , author=. The RAND Journal of Economics , volume=. 2020 , publisher=

    2020

  103. [117]

    Production and Operations Management , volume=

    Does greater visibility benefit minority businesses? Evidence from an online review platform , author=. Production and Operations Management , volume=. 2025 , publisher=

    2025

  104. [118]

    Persistent Employer Biases, and Discrimination (March 2, 2021) , year=

    Endogenous learning, persistent employer biases, and discrimination , author=. Persistent Employer Biases, and Discrimination (March 2, 2021) , year=

    2021

  105. [119]

    Available at SSRN 4155065 , year=

    Learning to Discriminate on the Job , author=. Available at SSRN 4155065 , year=

  106. [120]

    Management Science , volume=

    On statistical discrimination as a failure of social learning: A multiarmed bandit approach , author=. Management Science , volume=. 2026 , publisher=

    2026

  107. [121]

    Review of Economic Studies , volume=

    Hiring as exploration , author=. Review of Economic Studies , volume=. 2026 , publisher=

    2026

  108. [122]

    Available at SSRN 6171007 , year=

    Leveling Down: Competition and Discrimination in Service Markets , author=. Available at SSRN 6171007 , year=

Pith tools

Reviewed July 11, 2026 · model on record in the stance chip above.