Source: Google Street View
2 / 17
Why Business Owners Disclose Marginalized Identities
Questions, pushback, and wild ideas welcome.
East Coast Doctoral Conference · Columbia Business School · May 1, 2026
Kyle Emory McCullers · Ross School of Business, University of Michigan
Source: Google Street View
Black-owned businesses · 2020
"How The Black Lives Matter Movement Boosted Local Black-Owned Businesses"
GBH · 2021"Black businesses still face systemic racism — even from well-meaning customers"
Washington Post · 2020Asian-owned businesses · COVID-19
"The Best AAPI-Owned Businesses to Shop Right Now"
NBC News Select"Reports of Anti-Asian hate crimes rose nearly 150% in major U.S. cities"
CBS News · 2021Palestinian-owned businesses · 2023
"A Palestinian-owned business guide to show solidarity"
USA Today · 2023"Jewish groups call for boycott of pro-Palestinian businesses"
Forward · 2023Latino-owned businesses · 2024–2025
"Support Lake Street Latino businesses as immigration fears keep customers away"
Star Tribune · 2025"Hispanic businesses in Montgomery feeling 'hunted' after triple-slaying"
NBC News · 2024These are not parallel tracks — the solidarity signal is made necessary by discrimination exposure. Visibility attracts aligned customers and invites retaliation simultaneously.
Research Question
"Why, then, would an entrepreneur voluntarily disclose a characteristic that has historically served as a liability, thereby inviting potential bias in the pursuit of competitive advantage?"
Identity Theory
Goffman (1963)
Stigma management — individuals control information about discreditable attributes. The default is concealment or passing. Disclosure is a cost to be managed, not a strategy to be deployed.
Tajfel & Turner (1979)
Positive distinctiveness — group membership is claimed when it enables favorable social comparisons. Identity expression is driven by in-group psychology, not market calculation.
Bertrand & Mullainathan (2004)
Audit evidence: résumés with Black-sounding names received 50% fewer callbacks — racial identity functions as a market liability, motivating concealment as the rational baseline.
These traditions explain how stigmatized identities are experienced and managed in market contexts. Extending them to settings where disclosure is an active competitive signal — rather than a liability to manage — is the work this paper does.
Competitive Strategy
Porter (1980) · Barney (1991)
Advantage from resources that are valuable, rare, and hard to imitate. If identity is such a resource, owners should deploy it.
Hsu (2006)
Audiences penalize producers that span categories — identity signals generate returns only where a legible, local audience for that category exists.
Aneja, Luca & Reshef (2025) · Younkin & Kuppuswamy (2018) · Sharma, Frake & Watson (2025)
Empirical work documenting what disclosure produces — gains in some markets, penalties in others, symbolic support that rarely converts to revenue. Takes the disclosure decision itself as given.
This work explains what disclosure produces. The prior question — what structural conditions make the option available to an owner — is where this paper enters.
Identity theory explains how stigmatized identities are managed. Strategy explains when resources produce advantage. This paper asks the question that precedes both: when does an owner treat their own identity as a resource at all — and what structural conditions make that choice possible?
In conversation with: Hwang & Phillips (2025) · entrepreneurship as exit from labor market discrimination among stigmatized populations — a supply-side complement to the disclosure decision examined here.
Disclosing minority-owned status on a platform boosted traffic, calls, orders, and visits — demand effects concentrated in more Democratic, less racially biased areas (Aneja, Luca & Reshef, 2025).
BLM drove a surge in symbolic support (reviews, ratings) for Black-owned businesses but produced no meaningful increase in substantive support (revenue, foot traffic) — outrage translated into likes, not dollars.
Together, these studies establish that disclosure can pay — and illuminate the economic logic behind it. They leave open the prior question: under what structural conditions do owners perceive that logic as applicable to them?
But Aneja et al. show heterogeneous returns — two owners in the same city, same market, can face identical expected returns and still make opposite choices.
The missing piece
Disclosure is a two-stage decision.
Enabling conditions make it possible — activation triggers make it happen. Classic signaling theory only models the second stage (Spence, 1973).
Stage 1 — All three required
Possesses a marginalized identity
Retains positioning discretion (owner-operator)
Holds legitimacy beliefs — views identity as a claimable competitive dimension
Stage 2 — Either suffices
Perceived community support — bounded solidarity (co-ethnic density) or generalized solidarity (values-aligned allies)
Personal identity commitment — discloses as moral obligation independent of market calculation
Stage 1 is structural — any absent condition forecloses disclosure before Stage 2 is reached. Stage 2 is perceived — owners differ even when structural conditions are identical.
J1 — Bounded Solidarity → H1
"I know my neighborhood will show up"
Portes & Sensenbrenner, AJS (1993)
J2 — Generalized Solidarity → H2
"The community broadly values what I'm doing"
Author's theoretical contribution (Greenberg & Mollick, 2017)
J fires when either form of solidarity is perceived. H3 tests whether the two forms amplify each other — the interaction is the paper's theoretically decisive test.
Disclosure rates increase with Black population share. (Bounded solidarity — J1)
Disclosure rates increase with Democratic vote share. (Generalized solidarity — J2)
The J1 effect is conditional: significant in Democratic counties, null in Republican counties. (Attenuation)
The interaction does not predict Veteran-led disclosure. (Discriminant validity)
What we're measuring
Identity tag system — opt-in, permanent until removed, publicly visible
Google's identity tag system allows business owners to voluntarily self-report attributes. Tags appear as colored pill labels on the listing. Approximately 14,000 (0.28% of all listings; ~8.7% of 161,000 Census-identified Black-owned employer businesses) use the Black-owned attribute.
Unit of analysis: business listings geocoded to county. Outcome: Black-owned tag (0/1). Note: Census RDC application in progress for county-level Black-owned business counts — the correct at-risk denominator.
Scope: This analysis captures voluntary platform disclosure — a theoretically distinct act from MWBE certification or disclosure driven by government contracting eligibility, both of which carry strong and interpretable selection pressures not present here.
US Census — ACS 5-Year Estimates
County-level Black population share (pct_black_alone) and poverty rate drawn from the American Community Survey 5-year estimates. Population data anchors the bounded solidarity proxy and serves as a primary control.
MIT Election Lab — Presidential Returns
County-level Democratic vote margin from the 2020 presidential election (dem_margin_2020). Operationalizes generalized solidarity — political context as a signal of community alignment with equity norms.
Brooklyn Blooms
Does not disclose any identity
Brooklyn Tea
Discloses race
The Lit Bar
Discloses race, gender, and ethnicity
Q1 is 3× the rate of Q3. The intersection of community size and political support matters.
| H1 Bounded |
H2 Generalized |
H3 Attenuation |
H4 Discriminant |
|
|---|---|---|---|---|
| Black pop. share (%) | 0.090*** | — | 0.140*** | — |
| (0.010) | (0.023) | |||
| Dem. vote margin | — | 0.015*** | — | — |
| (0.002) | ||||
| Dem. majority (binary) | — | — | 0.481*** | — |
| (0.076) | ||||
| Black pop. × Dem. majority | — | — | −0.090** | — |
| (0.031) | ||||
| Veteran-led × Dem. majority | — | — | — | 0.002 |
| (0.022) | ||||
| State + industry FE | ✓ | ✓ | ✓ | ✓ |
| N (businesses) | 152,794 | 152,794 | 152,794 | 152,794 |
Logit coefficients. Standard errors clustered by county in parentheses. *** p<.001, ** p<.01. H4 interaction is not significant (p=.93), establishing discriminant validity.
• Black population share: For every 10 percentage-point increase in a county's Black population share, the local disclosure rate rises by ~0.09 pp — roughly a 30% increase over the baseline of 0.29%.
• Democratic vote margin: For every 10-point swing toward Democrats, the disclosure rate rises by ~0.04 pp — political climate matters independently of community size.
• Interaction: The political-climate effect is roughly 25% larger in high-Black-population counties than low-population ones — solidarity and political environment amplify each other, not just add.
What predicts retention?
Black population share (β=0.925, p<.001) and Democratic vote margin (β=0.608, p<.001) — the same solidarity structures that activate initial disclosure also sustain it. Extends H1 and H2 to persistence.
Where do owners drop off?
Highest drop-off: NH, OR, CO, IA, NM, UT — not the Deep South. The pattern is structural thinness: where the co-ethnic and ally customer base is too sparse to sustain the signal, disclosure quietly lapses.
Drop-off correlates with both lower Black population share and lower Democratic margin (r ≈ −0.21, p<.0001 each) — consistent with the bounded and generalized solidarity mechanisms operating jointly on persistence, not just entry.
"I didn't realize how radical it was to say, I'm creating a black space for black people by black people... my being here, even being a successful entrepreneur, was an act of resistance."
Entrepreneurship educator · Detroit Metro · K: identity as resistance"I'm going to drive that angle. But then there's also the dangerous side — I'm Jewish, but I'm not going to drive on that... that can be dangerous for business."
Boutique owner · Ann Arbor · Same owner — opposite disclosure across two identities"I want to win the blind taste test every time. I don't want people to be like, 'Yeah, I liked hers because she was a white girl from Michigan and I identify with that.' I don't care about that. I think that is a crutch to me."
Food & beverage owner · Ann Arbor · Product Purist — identity as competitive dimension, rejected"When we opened we were only 26... Koreans would come in from the suburbs. It was like a cult of party Koreans. And it did play a big role."
Restaurant owner · Ann Arbor · J₁: bounded solidarity through co-ethnic visibilityIdentity as a Conditional Strategic Resource
Brings identity theory and competitive strategy into contact: extends Goffman's stigma management — where the default is concealment — to market contexts where disclosure is an active competitive signal. Theorizes the decision that precedes any performance claim. This paper does not measure outcomes; it explains why some owners ever enter that calculation at all.
Solidarity as a Geographically Structured Mechanism
Bounded and generalized solidarity are distinct mechanisms — not substitutes — that activate disclosure under different structural conditions. Disclosure rates are more than 3× higher where both are present. The "valuable" dimension of identity-as-resource is geographically contingent, not portable across contexts.
Where this goes next
Census RDC
County-level Black-owned business counts — the correct at-risk denominator for estimating the true disclosure rate.
Companion Qualitative Study
In-depth interviews examining individual strategic orientations — the individual-level complement to the community-level quantitative patterns.
Longitudinal Tracking
2021→2026 panel reveals which owners sustain disclosure over time — extending H1 and H2 from initial decisions to strategic persistence.
Industry-Level Analysis
Disclosure norms vary substantially across industries. Within-industry tests will sharpen the solidarity mechanism and address heterogeneity concerns.
This paper responds to calls to center marginalized entrepreneurs in management research — Phillips et al. (Research in Organizational Behavior, 2024); Phillips & Ranganathan (ASQ, 2025).
Questions, pushback, and suggestions on theory, data, or framing are all welcome.
Kyle McCullers · kylemcc@umich.edu