Across the best-available empirical studies focused on publicly traded U.S. firms, “significant” employee ownership is most often operationalized through (i) an ESOP stake as a percent of common shares outstanding and/or voting shares, and (ii) broad-based equity programs (especially broad-based stock options in the 1990s).
The evidence base points to a nonlinear relationship between employee ownership and market performance: a large U.S. Census Center for Economic Studies working paper finds that small ESOPs (<5% of shares outstanding) are associated with materially higher valuation, while larger ESOP stakes show no valuation effect. In contrast, a National Bureau of Economic Research working paper focused on “labor voice” in corporate governance identifies public firms where employees control ≥5% of voting shares and uses a large non-employee-owned control universe; this study’s framing and empirical design emphasize that when employee owners vote a meaningfully sized block, the ownership stake can act as a governance lever that plausibly affects corporate policy (investment, risk, productivity).
For broad-based stock options, a Rutgers working paper (with matched peer benchmarking) reports that option firms show higher Tobin’s Q levels than comparable firms in 1997 (about 0.61–0.62 higher than “otherwise-similar” firms in the authors’ comparisons), while ROA results are more mixed across cross-sectional and pre/post adoption comparisons.
Overall, the literature is constrained by (a) measurement challenges (who owns? how much? voting control vs economic ownership; executive vs non-executive), (b) frequent reliance on selection-prone designs (cross-sectional comparisons even with controls), and (c) relatively few settings with clean quasi-experimental shocks to employee ownership among U.S. public companies. (Accompany your reading with popular opera in WAV, instead of mp3 format. Click below to listen.)
Employee Ownership and Democracy: Author’s Note
This article relied on AI tools. I’ve been interested in this area for about 50 years. I can see I’m not going to live to see the social/economic changes I want, so this post is one of many notes in a bottle that I hope will inspire younger people to carry on. My current vision favors an economy with substantial employee ownership in large publicly traded companies (5-15% to avoid control capture that would disuade outside investors). Employees should be able to vote their shares and also have many of the other “necessary” components of Workplace Democratization outlined by Paul Bernstein in 1976.
As a long-time student of
As you will see in the section of this post entitled “Evidence Gaps and Future Research Directions,” a major obstacle to moving forward is the lack of data. My efforts to obtain disclosure of some of the needed data have been rebuffed. See Incentive Shares: Meta, Amazon & Repligen Refuse to Say How Distributed.
Incentive shares – shareholders have a right to know their distribution. Meta, Amazon, and Repligen argue against disclosure. Working mostly from the same template, they argue that the distribution of shares to employees is “ordinary business” and “does not focus on any significant policy issue that transcends the company’s ordinary business operations.”
Employee Ownership Performance: Primary Sources
The following are the main empirical sources used as the backbone for this report (PDF/official sources where accessible):
- E. Han Kim and Paige Ouimet, Employee Capitalism or Corporate Socialism? Broad-Based Employee Stock Ownership (U.S. Census Bureau Center for Economic Studies working paper CES-WP-09-44).
- Olubunmi Faleye, Vikas Mehrotra, and Randall Morck, When Labor Has a Voice in Corporate Governance (National Bureau of Economic Research working paper 11254).
- James C. Sesil, Maya K. Kroumova, Douglas L. Kruse, and Joseph R. Blasi, Broad-Based Employee Stock Options in the U.S.: Company Performance and Characteristics (Rutgers University working paper).
- Public Companies with Broad-Based Stock Options: Corporate Performance from 1992–1997 (National Center for Employee Ownership report; uses Compustat plus survey/public-plan data).
- Peter Cramton, Hamid Mehran, and Joseph Tracy, ESOP Fables: The Impact of Employee Stock Ownership Plans on Labor Disputes (Federal Reserve Bank of New York Staff Report 347).
- Lilli A. Gordon and John Pound, ESOPs and Corporate Control (abstracted via ScienceDirect).
- Michael Conte, Joseph R. Blasi, and Douglas L. Kruse, Employee Stock Ownership and Corporate Performance Among Public Companies (summary via Rutgers CLEO page; underlying journal access is restricted).
- Lisa F. Borstadt and Thomas J. Zwirlein, ESOPs in Publicly Held Companies: Evidence on Productivity and Firm Performance.
Primary Empirical Studies
Comparison table of study attributes
| Study | Sample period | Sample size | Employee ownership definition and measurement | Outcomes / performance metrics | Identification strategy | Key controls | Main quantitative findings | Robustness checks | Key limitations | Data sources (as described in accessible text) |
|---|---|---|---|---|---|---|---|---|---|---|
| CES WP on broad-based ESOPs and valuation | Observations include ESOP stakes through 2004 (text indicates a window ending in 2004). | 418 ESOPs in sample; 225 reach ≥5% at some point. | ESOP size measured as % of shares outstanding; “small” defined as <5%; “large” ≥5%. Median ESOP 5.93%; for large ESOPs, median 12.18% and mean 16.65%. | Firm valuation (explicitly discussed), wages per employee (incl. “unexplained wages”), and related firm/plant controls. | Matched-control comparisons (ESOP vs matched non-ESOP) plus regressions with additional controls; also uses a “differences-in-differences like approach” for policy confounding checks. | Wage residualization controls include state-year and industry-year mean wages. | Small ESOPs associated with about a 17% increase in firm valuation relative to the sample mean; effects disappear for ESOPs ≥5%. | Tests sensitivity to control-firm selection by dropping controls and estimating on firms that have an ESOP at some point; additional-control robustness in valuation regressions. | Even with matching/controls, employee ownership is endogenous to firm strategy and expectations; causal interpretation remains delicate. (Authors explicitly run multiple robustness checks rather than claiming a clean natural experiment.) | CES working paper setting; text references matched control samples and establishment-level constructs (plants, industry concentration constructs). |
| NBER WP on “labor voice” governance and performance | Baseline universe is proxy statements filed in 1995; design implies firms must survive at least five years after the “event year.” | Start with 3,823 proxy statements; after exclusions 2,165. Identify 277 firms with employee ownership ≥5%; exclude 22 where management votes employee shares; remaining “labor voice” cohort is 255 (139 ESOP-only, 75 non-ESOP channels, 41 mixed). Control sample: 1,888 firms with <5% (or none) employee ownership. | Employee ownership measured as proportion of voting shares owned by employees from beneficial ownership section of SEC proxy statements; threshold ≥5%. Explicit focus on whether labor votes its stock (exclude cases where managers vote labor’s block). | Tobin’s q (shareholder value), labor productivity, and TFP (Cobb-Douglas residual-based), plus corporate policy variables (investment/risk language appears in paper structure). | Mean/median comparisons + multiple regression with control variables to adjust for exogenous factors affecting policy and performance. | Controls are referenced generally as “control for exogenous factors.” (Full list not captured in excerpted lines.) | Quantitative coefficient magnitudes and p-values are presented in tables (e.g., Table 2 referenced for Tobin’s q comparisons), but the extracted snippet does not include numeric table entries. | Paper explicitly describes regression framework and robustness conceptually; excerpt does not expose the specific robustness table outputs beyond setup. | Employee ownership is identified from a single-year proxy-statement universe (1995), creating potential survivorship and selection issues; labor influence may covary with unobserved firm traits. | SEC proxy statements read to determine employee voting-share ownership; COMPUSTAT used for firm variables; productivity derived from sales, employees, and PPE; industry controls via two-digit SIC. |
| Rutgers WP on broad-based employee stock options | Cross-sectional comparisons for 1997 plus pre/post comparisons including 1985–87 vs 1995–97 (as referenced). | Survey-based sample + 385 additional broad-based option firms that did not respond to survey (total “all stock option companies” = surveyed + 385 + “unknown” coverage described). Matched peers include two non-option firms per option firm (next largest and next smallest by employment in same 2-digit industry). | Broad-based stock option programs for non-management employees; intensity proxied by coverage categories (including a 50%+ coverage subgroup referenced in results discussion). | Productivity, total shareholder return, Tobin’s q, ROA, and average labor cost. | Matched-pair benchmarking by industry and employment size + before/after adoption comparisons using multiple windows. | Primary comparisons are “simple comparisons” across groups; excerpt does not list a full multivariate control set (beyond the matching). | In 1997, Tobin’s q is reported as significantly higher for option firms than “otherwise-similar” firms by ~0.61–0.62; ROA is not different overall, but some pre/post ROA changes are statistically significant in specified windows. | Multiple pre/post windows; comparisons to (i) overall Compustat population, (ii) matched peers, and (iii) non-survey broad-based option firms; discussion of mixed results for annual changes in Q and ROA. | Potential selection into broad-based option adoption; incomplete plan details for “unknown coverage” firms; matching on 2-digit industry and employment may not eliminate all confounding. | S&P Compustat (1998 file) plus surveys and public information on option plan adoption; peer matching uses employment in same 2-digit industry. |
| NCEO report on public companies with broad-based stock options | 1992–1997 focus. | Not fully captured in excerpt; described as constructed by merging Compustat performance with survey/public plan data. | Broad-based stock option plans (employee coverage orientation), assembled from surveys and public plan information. | Report frames productivity and shareholder returns as core targets for comparison. | Systematic comparison/reporting; underlying approach aligns with the Rutgers WP’s peer benchmarking logic but is presented as a research report. | Not specified in excerpt. | Not extracted as standalone regression effect sizes in excerpt; serves as an institutional bridge tying plan data to Compustat outcomes. | Not specified in excerpt. | Not peer-reviewed; document is explicit that findings reflect authors and are retrospective. | Compustat merged with survey and public plan information. |
| ILR Review study on public companies with >5% employee ownership (summary-only accessed) | Levels in 1990/91; growth 1980–90 referenced. | Not reported in summary snippet. | Employee ownership defined as employees owning >5% of company stock vs all others. | Profitability, productivity, and compensation (per summary). | Cross-sectional comparison with additional stratification by firm size and contextual variables (takeover threat; wage/benefit concession). | Not specified in summary. | Summary conclusion: differences are not strongly negative; when differences appear, they favor employee-ownership firms, especially smaller firms; takeover/concession context not significantly associated with performance levels or growth. | Authors caution data do not permit clear tests of causality (per summary). | Full paper not accessible here; causal claims limited. | Not specified in summary. |
| ESOP adoption and productivity/performance in public firms (abstract-only accessed) | ESOP adoptions 1973–1986. | 85 publicly traded ESOP adopters with matched size/industry controls. | ESOP adoption among publicly traded firms; matched control by industry and size. | “Several measures of productivity and performance” (not enumerated in abstract snippet). | Matched control design (industry and size). | Not described in abstract snippet. | Abstract reports no evidence of productivity gains or performance improvements post adoption. | Not described in abstract snippet. | Full text not retrieved in tool session; outcome definitions and coefficients unavailable here. | Not described in abstract snippet. |
| ESOP announcements and corporate control (ScienceDirect abstract) | Not stated in abstract snippet. | Not stated in abstract snippet. | ESOPs differentiated by whether adopted amid takeover activity and by control-structure features (e.g., veto power; nonvoting stock). | Stock-price reaction (“shareholder wealth”) at announcement; control vs incentive effects. | Event-study interpretation from announcement returns; heterogeneity by takeover context and structure. | Not specified in abstract snippet. | ESOPs in takeover activity reduce share values by ~4% on average; large nonvoting ESOPs increase share value. | Not specified in abstract snippet. | Abstract-only here; sample construction and full specifications not accessible in-session. | Not specified in abstract snippet. |
| FRBNY staff report on ESOPs and labor disputes (supplementary, but includes shareholder event study component) | 2008 working paper; historical setting for ESOP spread highlighted. | Not captured in excerpt. | ESOP defined as ERISA-sanctioned retirement plan; at least 50% of assets in employer securities; nondiscrimination and diversification features described. | Primary outcome is labor disputes; includes an event study of ESOP announcements to assess shareholder implications. | Structural bargaining model + empirical analyses; event study on new-ESOP announcements. | Not captured in excerpt. | Event study exists (shareholder implications), but numerical CAR outputs not captured in excerpted lines. | Not captured in excerpt. | Not a direct “performance” paper; focuses on labor dispute mechanism with shareholder event study as component. | Not captured in excerpt. |
Analytical notes on identification strategies
The most policy-relevant identification pattern in this evidence base is non-experimental matching with extensive robustness work, rather than clean quasi-experiments:
- The CES working paper’s headline result (valuation gain for ESOPs <5%) is supported by sensitivity checks to control-group choice (dropping controls and estimating within ESOP firms) and robustness with additional controls, but remains observational.
- The NBER “labor voice” study uses an unusually concrete measurement approach (manual parsing of SEC proxy filings, with an explicit focus on who votes employee shares) and a very large internal control universe, but its baseline sample is constructed from a 1995 proxy-statement universe and thus is still subject to selection and survivorship concerns.
- The Rutgers broad-based stock-options study’s matched-peer design (next-largest/next-smallest within 2-digit industry by employment) is transparent and replicable, but adoption into broad-based options is plausibly endogenous to growth opportunities and compensation strategy.
Employee Ownership Performance: Cross-study Synthesis
What the evidence implies about “3–20% ownership” style thresholds
Even though the literature does not consistently target a 3–20% band, several studies effectively hinge on around-5% thresholds that are close in spirit to “meaningful” ownership:
- In one large CES study, the cutoff at 5% separates “small” ESOPs (which are linked to valuation gains) from larger ESOPs (with no valuation consequences).
- In an NBER study explicitly interested in governance influence, ≥5% of voting shares owned by employees is used to define “labor voice” firms; the mean and median employee voting ownership among these firms are ~13% and ~11%.
Taken together, these findings are consistent with a nonlinear story: small, broad-based stakes may operate primarily as incentives (potentially value-enhancing), while larger, voting-controlled blocks can create genuine stakeholder power that changes corporate decision-making, potentially in ways that do not maximize shareholder value in all settings.
Market valuation vs. accounting performance vs. real outcomes
The evidence is uneven across outcome classes:
- Market valuation: The most numerically explicit and interpretable valuation estimate in the sourced set is the ~17% valuation increase associated with small ESOPs in the CES study.
- Tobin’s Q (market-to-assets type measures): The broad-based stock options working paper reports materially higher Tobin’s Q levels in 1997 relative to matched peers (roughly +0.61 to +0.62 in the comparisons highlighted) but also describes mixed patterns when looking at changes over time.
- ROA: The broad-based stock options paper reports that ROA differences are not consistently significant in levels across broad groups, but some pre/post comparisons are significantly positive (depending on window and subsample).
- Productivity and wages: Several studies emphasize that productivity gains (where present) may be partially captured by employees (wage/rent-sharing), making it essential to interpret shareholder returns jointly with wage outcomes.
Publication Timeline
Employee Ownership Performance: Evidence Gaps and Future Research
The existing empirical literature leaves several high-value gaps for a rigorous “public U.S. firms with employee ownership vs without” comparison program:
- Measurement remains the central constraint. The most careful governance-oriented study in this set explicitly excludes firms in which management votes employee-owned shares, highlighting that “employee ownership” can be economically meaningful while not actually giving employees a governance voice. Future research that systematically distinguishes (i) economic ownership, (ii) voting rights, and (iii) who exercises voting discretion (trustees vs employees vs management) would materially improve identification.
- There is a shortage of credible causal shocks in U.S. public-firm settings. The CES study adds robustness and policy-style difference-in-differences-like checks, but the core design is still observational. Stronger designs could leverage quasi-exogenous variation such as discontinuities in plan eligibility rules, staggered adoption driven by regulatory or accounting changes, or plausibly exogenous institutional ownership changes that interact with employee ownership in governance.
- Many public-firm studies emphasize market-based outcomes (announcement returns, Tobin’s Q) but provide incomplete joint accounting of distributional channels (wages, employment, labor share). The CES working paper explicitly frames the question of how productivity gains are split between shareholders and employees, and it operationalizes “unexplained wages” in a way that could be extended into richer decomposition work.
- Finally, there is limited integrated evidence on risk and downside exposure (volatility, tail risk, survival) for U.S. public firms with meaningful employee ownership blocks. The governance-focused NBER study is structurally positioned to examine investment and risk policies, while the broad-based options work highlights risk/return tradeoffs in the background but does not anchor a unified risk framework in the extracted excerpts.
Collaborate with CorpGov.net on Social Media (or Leave a Reply at the bottom)
Linkedin
BlueSky
Facebook
Mastodon
Substack
X (formerly Twitter)
Employee Ownership Performance: Related Posts
Incentive Shares: Meta, Amazon & Repligen Refuse to Say How Distributed
Meta Platforms: Address Inequality Through Ownership Culture
Video Friday: Double Feature on Corporate Governance at ESOP Firms
Comments are closed.