The most useful way to read the past century of American labor history is as a single 90-year arc of bargaining power: rising sharply from the Wagner Act in 1935 through roughly 1973, declining steadily for 50 years afterward, and finally — perhaps — bottoming out and beginning to recover between 2021 and now. That framing isn't original to us; it's the consensus position of a generation of labor historians and economists including Nelson Lichtenstein, Jefferson Cowie, Lawrence Mishel, and more recently Anna Stansbury and Larry Summers. The strength of the framing is that it survives the test of measurable variables: union density, real-wage growth, share of national income going to labor, and the prevalence of stable full-time employment. All four track the same arc.
The argument of this piece is that the standard "decline of unions" narrative both understates the structural causes of that decline and undersells the small but real signs that the labor share has stopped falling. What workers do next depends heavily on which version of the story policymakers, organizers, and employers internalize.
The construction of the modern workplace, 1935–1973
Three legal interventions defined what work in America looked like for most of the 20th century. The National Labor Relations Act of 1935 (the Wagner Act) created the legal infrastructure for private-sector union recognition and collective bargaining and is the single most consequential piece of labor legislation in U.S. history. The Fair Labor Standards Act of 1938 established the federal minimum wage, the 40-hour workweek, overtime, and child-labor protections. The GI Bill of 1944 created the financial infrastructure for mass homeownership and college education for returning veterans — which, as Ira Katznelson's 2005 book When Affirmative Action Was White documents in detail, was largely structured to exclude Black veterans, with effects that compounded through the rest of the century.
By the early 1950s, U.S. private-sector union density peaked at around 35%. Real wages for production workers grew steadily; the productivity-pay link — wages growing roughly in step with productivity — held until about 1973. Jefferson Cowie's 2010 book Stayin' Alive: The 1970s and the Last Days of the Working Class identifies that decade as the inflection point, when the postwar settlement broke under the combined weight of oil shocks, intensifying foreign competition, deindustrialization, and a political shift that began to treat organized labor as an antagonist rather than a stabilizer.
The long decline, 1973–2021
BLS data show U.S. union density falling from 24% in 1973 to 10.0% in 2024 — and below 6% in the private sector. Over the same period, the productivity-pay gap opened sharply. EPI's analysis, the most-cited summary of this data, shows that net productivity grew 80% from 1979 to 2023 while average compensation for nonsupervisory workers grew about 30%. That is the wedge from which most other distributional changes flow.
Anna Stansbury and Larry Summers' 2020 Brookings paper, "The Declining Worker Power Hypothesis," argues that the most parsimonious explanation for the long decline in the labor share, the rise in profit margins, and stagnant wage growth is the erosion of worker bargaining power — not technology, not globalization in isolation, but the joint effect of weakened unions and intensified employer pricing power in labor markets. That hypothesis is now broadly accepted in the empirical labor-economics literature.
The 1973–2021 period also saw the deliberate construction of the modern "non-employee" workforce: independent contractors, temp-agency placements, and platform workers, all of whom sit largely outside the National Labor Relations Act framework. David Weil's 2014 book The Fissured Workplace (Harvard University Press) is the standard reference for how this happened — large firms shifted employment responsibility to subcontractors, franchisees, and staffing agencies, eroding the implicit social contract of the postwar period while preserving the brand-level firm.
The reversal, 2021 onward
Three things have shifted measurably since 2021, and each of them constitutes evidence that the long decline has at least paused.
First, tight labor markets. The post-pandemic recovery produced quit rates the BLS had not measured in decades — peaking above 3% of employed workers monthly in 2022. That gave workers real exit power, and the data on bottom-quartile wage growth reflects it: David Autor, Annie McGrew, and Arindrajit Dube's 2023 NBER paper "The Unexpected Compression" documents real wage growth for low-wage workers outpacing high-wage workers for the first sustained period since the early 1970s.
Second, organized labor activity. NLRB election petitions hit 2,594 in fiscal year 2023, the highest figure in nearly two decades. The UAW won contracts at Ford, GM, and Stellantis in 2023 with roughly 25% wage increases over the contract life. The Teamsters won a UPS contract that raised the part-time wage floor to $21 an hour. The Hollywood writers' and actors' strikes produced AI-use protections now being templated across other industries. Each of these would have been inconceivable in 2015.
Third, policy. The IRA, BIL, and CHIPS Acts collectively unlocked more than $1 trillion of public investment with prevailing-wage and registered-apprenticeship requirements attached to most of it — a structural pro-labor design choice with no real precedent since the New Deal. The FTC's 2024 non-compete rule, partly blocked but partly active in state-level versions, is another. The NLRB under the Biden administration issued the most pro-worker decisions in 40 years, several of which are now being tested under successor administrations.
What the historians get right that the headlines miss
The single most consistent point that serious labor historians make about the current moment is that durable gains require durable institutions. A wave of organizing energy that does not consolidate into long-term bargaining structures will dissipate the way the 1970s wave did. Nelson Lichtenstein's 2002 book State of the Union: A Century of American Labor remains the best single treatment of why this matters: episodic worker mobilization, however moving, is not the same as institutional power.
The second point is that the bargain underlying the postwar settlement — broad-based wage growth in exchange for managerial control of the workplace — has not been reconstructed. Neither has any clear successor. The American workplace of 2026 is a hybrid of the old industrial model (for manufacturing, healthcare, and skilled trades), the fissured model (for retail, hospitality, and logistics), and the platform model (for delivery, ride-share, and increasingly knowledge-work freelancing). Each has different rules; few of those rules have been deliberately designed.
For a more current look at how the bargaining-power question is playing out, see our flagship piece on The New Labor Movement →.
The long decline in U.S. worker bargaining power has paused, possibly reversed, for the first time in 50 years. Whether that becomes a generation-defining shift or a brief tightening depends on whether the institutions get built this decade.
What the next 25 years might look like
Three possible trajectories are visible. The first is a Scandinavian-style sectoral bargaining model — where wages and conditions are set across industries through tripartite negotiation between unions, employer associations, and government — which is being seriously discussed in U.S. policy circles for the first time in 40 years. The second is a continuation of the fissured-firm model with new platform protections layered on top, which is roughly what the EU Platform Work Directive looks like in practice. The third is regression: a return to the pre-2021 status quo if the current organizing wave fails to consolidate.
Which trajectory we end up on will be decided by demographic, political, and policy choices over roughly the next decade. The historians' contribution is to remind us that what looks inevitable in retrospect was, at every prior inflection point, deeply contingent — and that the most consequential labor-policy decisions in American history were made when most observers did not realize they were being made.
Updated May 21, 2026. This piece was substantively rewritten as part of NWLB's 2026 editorial refresh.
