Eight charts that reveal how economic rewards have shifted from workers to capital owners, machines, and foreign competitors
Updated July 13, 2026 - Corrections to China-shock figures, union-membership labeling, income-tier data, and tax-policy provisions
The following analysis examines peer-reviewed economic literature on the structural forces behind labor's declining share of national income.
In the 1970s, U.S. labor compensation made up roughly 50% of gross value added. By 2024, it had fallen to about 44% – a headline decline of roughly 6 percentage points. But this figure somewhat overstates the structural shift: when self-employment income is properly allocated between labor and capital (following Elsby, Hobijn, and Şahin 2013), the decline is closer to 4 percentage points. Even this more modest decline represents a significant shift in economic rewards, driven by technology, globalization, and institutional changes.
Labor's share of income has declined modestly but persistently, with cyclical variations around the trend
Economists have long noted this decline in labor's share, but only recently understood its full implications. After remaining roughly stable for decades – one of Kaldor's "stylized facts" – labor's share began its downward trend in the 1980s. Leading economists like Thomas Piketty and Emmanuel Saez have documented how this shift coincides with surging inequality and wealth concentration.
While the raw numbers show a decline from 50% to 44%, measurement matters. The headline series treats all proprietor income as capital, but self-employed workers earn labor income too. Elsby, Hobijn, and Şahin (2013) show that properly allocating proprietor income reduces the decline by about one-third.
Additionally, labor share is cyclical – it typically falls late in expansions and rises in recessions. The series has hovered near 44% since the mid-2010s (44.5% in 2019, 44.0% in 2024), showing no clear rebound. Even a 4-point structural decline represents tens of billions annually shifting from paychecks to profits.
Worker productivity has far outpaced average hourly compensation since 1948, opening a persistent gap
For most of American history, productivity and pay rose in tandem – a social contract ensuring that as workers became more productive, they shared in the gains. But starting in the 1970s, this link was severed. The chart above tracks average real hourly compensation, which grew 212% since 1948 versus 388% for productivity; because averages are pulled up by rapid pay gains at the top, the gap facing the typical worker is wider still. The Economic Policy Institute finds that net productivity rose 61.8% from 1979 to 2020 while compensation for the median worker rose just 17.5%. Using consistent deflators (both CPI or both IPD) reduces the raw gap by about 40%, but a significant divergence from historical norms remains.
Karabarbounis and Neiman (2014) attribute much of this to the declining price of capital goods – as computers and machinery became cheaper, firms substituted technology for workers. Bergholt, Furlanetto, and Maffei-Faccioli (2022) in the American Economic Journal concluded that automation is the single biggest driver of this divergence.
The gap between these lines, while partly reflecting different price deflators, still represents a real shift in economic rewards. Using consistent deflators shows compensation lagging productivity by 50-70 percentage points rather than 176 points – substantial but less dramatic than the headline numbers suggest.
After-tax corporate profits as a percentage of GDP have reached levels not seen since the 1960s
The surge in after-tax corporate profits from ~6% of GDP in the early 1990s to nearly 11% as of 2024 represents what economists call "financialization" – the increasing dominance of financial motives in the economy. The International Labour Organization (2013) found that financialization was the single largest factor in falling wage shares worldwide, accounting for 46% of the decline.
William Lazonick's research (2014) revealed the mechanism: from 2003-2012, S&P 500 companies used 91% of their net income on stock buybacks and dividends. Critics note this doesn't preclude investment (firms can use debt or retained earnings), but the shift toward shareholder payouts over worker compensation is clear. This "downsize-and-distribute" model replaced the post-war "retain-and-reinvest" approach.
Notice how profit spikes often coincide with recessions – when workers lose bargaining power, corporate profits tend to surge.
Total union membership has plummeted, tracking closely with labor's declining share (private-sector density has fallen even more steeply)
The decline of unions and labor share show correlation, though the relationship is complex. Western and Rosenfeld (2011) found that union decline explains 20-33% of rising wage inequality among men. However, it's worth noting that some countries experienced union decline without similar labor share drops, suggesting U.S.-specific factors like weaker labor laws and employer opposition played crucial roles.
The visual correlation is striking: as the overall union membership rate fell from 20.1% in 1983 to 9.9% in 2024 (BLS), labor's share of GDP declined alongside it. The private-sector figure fell even more steeply, from about 16.8% in 1983 to 5.9% in 2024. Unions didn't just raise wages for members – they set standards that lifted all workers. Their collapse left individual workers with diminished leverage against increasingly concentrated corporate power.
While federal statutory rates fell from 52% to 21%, effective federal rates (what corporations actually pay) dropped even more dramatically
While the statutory federal rate cut from 52% to 21% was dramatic, the effective rate story is even more striking. In the 1950s, corporations actually paid about 40% of their profits in federal taxes. By 2024, despite a 21% statutory rate, corporations pay just 11.9% in federal taxes – benefiting from loopholes, deductions, and creative accounting. Including state and local taxes adds roughly 3-5 percentage points to these figures.
The decline is staggering: effective corporate tax rates fell from 39.6% in the 1950s to just 8.9% in the 2020s. This means corporations today keep over 90% of their profits, compared to just 60% in the post-war era. Kaymak and Schott (2023) found that this tax decline can explain about half of the labor share decline in manufacturing.
Most shocking: even during the 2017 tax cut, the effective rate was already at historic lows (11.7% in 2010). The Tax Cuts and Jobs Act pushed an already minimal tax burden even lower. In 2020, corporations paid just 8.9% – the lowest effective rate in modern history.
This represents hundreds of billions annually that once funded schools, infrastructure, and social programs now flowing directly to shareholders. It wasn't inevitable – it was a choice.
As corporate tax rates fell from roughly 52% to 21%, labor's share of income declined over the same decades (both series as used in Chart 1 and Chart 5)
Corporate tax rates and labor's share of income fell over the same decades. As statutory corporate tax rates dropped from around 52% in the 1950s to 21% today (as of 2024), labor's share slipped from roughly 50% to 44%. The two trends move together – but, as we note below, shared timing does not by itself establish that one caused the other.
Each major tax cut coincided with a further erosion of worker power: Reagan's 1986 cuts, Bush's 2001 cuts, and Trump's 2017 cuts all accelerated the transfer of income from workers to capital. The mechanism is straightforward: lower taxes increase after-tax returns to capital, incentivizing automation and offshoring while reducing the tax revenue needed for public investments that support workers.
While correlation doesn't prove causation, the mechanism is clear: lower corporate taxes increase after-tax returns to capital ownership, shifting incentives toward automation and shareholder payouts over wage growth.
The middle class shrank from 61% to 51% as Americans moved to both extremes in roughly equal measure
America's economy has split into two tracks. On one side: capital-intensive firms like Google, Goldman Sachs, and biotech companies that hire fewer, highly-paid workers. On the other: labor-intensive service businesses – restaurants, retail, care work – offering low wages and few benefits.
The casualties? Middle-income jobs that once sustained the American Dream. Traditional retailers like Sears provided stable employment with benefits and opportunities for advancement. Today's dominant employers in retail and logistics offer fewer pathways to middle-class security. Travel agents, bank tellers, manufacturing workers – entire occupations that provided pathways to the middle class have vanished or been transformed.
As MIT economist David Autor documented, automation eliminates "routine" middle-skill jobs while complementing high-skill work and leaving low-skill service jobs untouched. The result: job polarization that mirrors income polarization.
The numbers tell a more balanced story than a simple "winners" framing suggests: as the middle class shrank by 10 points between 1971 and 2023, the upper tier grew by 5 points (14% to 19% of adults) and the lower tier grew by 5 points (25% to 30%) – movement up and down was roughly symmetric. But that symmetry masks who moved where. Those who lost middle-class jobs often did not move up; new college graduates claimed many of the high-income spots while displaced workers competed for low-wage service jobs. Same economy, different worlds.
Upper-income households captured nearly all income gains, now holding 48% of total income despite being only 19% of adults
While the middle class shrank by 10 percentage points in population, their share of total income collapsed by 19 points – from 62% to 43%. Meanwhile, the upper-income tier nearly doubled their share from 29% to 48%, despite growing from just 11% to 19% of the population.
This reveals the true nature of inequality: it's not just about how many people are in each tier, but how much of the economic pie they control. Today, the richest 19% of Americans control nearly half of all income – more than the entire middle class.
Most striking: lower-income Americans, despite growing from 27% to 30% of the population, saw their already meager share of income shrink further from 10% to just 9%. The American economy hasn't just polarized – it has fundamentally redistributed rewards upward.
Upper-income households have increasingly concentrated in a handful of coastal metros while many interior metros lost ground. The map below is illustrative of this geographic sorting; bubble sizes reflect a relative wealth index, not precise dollar figures.
America's wealth hasn't just concentrated – it has physically migrated. In 1980, upper-income households were distributed across manufacturing centers like Detroit, Cleveland, and St. Louis. Today, they cluster in a handful of superstar cities: San Francisco, New York, Boston, Seattle, DC.
Research from the Brookings Institution and Pew Research Center documents this geographic sorting: high-income households have increasingly concentrated in major metropolitan areas, particularly coastal cities, while smaller metros and rural areas have seen their share of high earners decline substantially over the past four decades.
This isn't natural economic evolution – it's economic abandonment. When a factory closes in Ohio, its engineers move to Silicon Valley. When a bank consolidates in Charlotte, its executives relocate to Manhattan. The middle of America hasn't just lost jobs; it has hemorrhaged its highest earners.
Most devastating: this creates a feedback loop. As high earners leave, tax revenues collapse, schools deteriorate, and infrastructure crumbles – ensuring the next generation of talent will also flee. The American interior is becoming an economic colony of its coasts.
Each robot per 1,000 workers reduces employment by 0.2% and wages by 0.42% (Acemoglu-Restrepo)
The quadrupling of robot density from 1993 to 2015 represents a new phase in the labor-capital struggle. Acemoglu and Restrepo (2020) found that each additional robot per 1,000 workers reduces employment by 0.2 percentage points and wages by 0.42%. However, these estimates remain debated – some studies find smaller effects, and robots may create offsetting jobs in other sectors.
These are what Acemoglu calls "so-so technologies" – just good enough to replace workers but not transformative enough to create new opportunities. While concerning, history shows that technological disruption often eventually creates new types of work, though the transition can be painful for displaced workers.
The impact compounds: regions with more robots saw larger declines in employment and wages, creating "robot zones" of economic distress. And this is just the beginning – artificial intelligence threatens to automate cognitive work next.
The 0.61% cumulative wage decline represents one estimate of automation's impact. While significant for affected workers, the overall effect on labor share remains under study, with automation explaining perhaps 10-20% of the total decline according to various estimates.
Chinese import penetration surged 12-fold, destroying millions of manufacturing jobs in America's heartland
The "China shock" identified by economists Autor, Dorn, and Hanson represents the most sudden and devastating trade impact ever experienced by American workers. Chinese imports as a share of U.S. spending exploded from 0.6% in 1990 to 7.2% by 2020 – a twelve-fold increase that was particularly dramatic after China's WTO entry in 2001.
The human cost was severe. Autor, Dorn, and Hanson (2013) attribute roughly 985,000 manufacturing jobs lost to Chinese import competition between 1990 and 2007 – about a quarter of the total decline in U.S. manufacturing employment over that period. Counting knock-on effects across the wider economy, a later study by Acemoglu, Autor, Dorn, Hanson, and Price (2016) estimated that Chinese import competition cost the United States on the order of 2.0-2.4 million jobs in total between 1999 and 2011. Unlike previous trade shocks, these manufacturing jobs never came back. Workers in affected regions experienced persistent income losses, with each $1,000 increase in import exposure per worker associated with a 0.5-0.8 percentage point (central estimate: 0.6pp) drop in the local manufacturing share of employment.
The scatter plot below shows the devastating correlation: regions with higher Chinese import exposure (horizontal axis) experienced dramatically larger manufacturing job losses (vertical axis). Each dot represents an American community whose economic foundation was swept away by the tide of globalization.
Each $1,000 increase in Chinese import exposure per worker led to a 0.5-0.8 percentage point decline in manufacturing employment share of working-age population (central estimate: 0.6pp)
The Hickory-Lenoir-Morganton, NC commuting zone sat near the bull's-eye of the China boom, with among the highest per-worker import exposure in the country and a roughly five-point collapse in factory employment. New York City, barely touched by Chinese manufacturing competition, lost essentially no factory share and kept adding high-paying service jobs. The same national trade policy produced opposite local destinies.
The figures below are approximate and illustrative of the patterns Autor, Dorn, and Hanson document across commuting zones, not exact published values for these two areas.
Furniture/Textiles hub
Finance, Media, Tech hub
What makes the China shock particularly cruel is its permanence. Autor and colleagues found no evidence of recovery in manufacturing employment or workers' lifetime earnings even a decade later.* Unlike the economic theory that promised new jobs would replace old ones, these communities experienced persistent unemployment, rising disability claims (Autor et al., 2014), and declining marriage rates (Autor et al., 2018).
The political consequences were profound: regions hit hardest by Chinese import competition shifted dramatically toward political extremes, expressing their economic despair through the ballot box (Feigenbaum & Hall, 2015; Autor et al., 2020). The bipartisan consensus on free trade, built over decades, shattered in these hollowed-out factory towns.
This wasn't creative destruction – it was just destruction. The China shock proves that when global capitalism moves too fast, it doesn't create opportunity; it creates wastelands.
*Note: A 2021 follow-up by Autor et al. found that while local GDP eventually rebounds through growth in health care, education, and retail sectors, manufacturing jobs and the lifetime earnings of displaced workers never recover.
These twelve charts tell a unified story: over the past 50 years, America has witnessed one of the largest transfers of wealth in human history – not between individuals, but between economic classes. The transfer from labor to capital, accelerated by automation and globalization.
Leading economists have reached a consensus: this wasn't accidental or inevitable. It was the result of deliberate policy choices, technological changes that favored capital, and the systematic dismantling of worker protections.
The question for America's future: Will we continue down this path, or will we build an economy that once again rewards work, not just wealth?
Current tax incentives systematically favor automation over employment
| Feature | Description | Automation Impact |
|---|---|---|
| Bonus Depreciation (100%) | Firms can write off 100% of equipment/software investment in year one (permanently restored by the One Big Beautiful Bill Act, July 2025) | Encourages heavy upfront capex—robots, servers, automation infrastructure |
| R&D Credits | Permanent tax credits for software, AI model development, robotics | Subsidizes capital-embodied innovation, not training or hiring |
| No Payroll Tax on Machines | Labor comes with 15.3% FICA payroll tax; machines do not | Makes labor more expensive on a relative basis |
| Stock-based pay deductibility | Firms deduct stock comp as an expense; helps subsidize lean, equity-heavy comp structures in tech | Incentivizes firms to grow value with fewer people, more IP and automation |
The 2017 Tax Cuts and Jobs Act (TCJA) introduced provisions that accelerate automation incentives, and the One Big Beautiful Bill Act (OBBBA), signed July 4, 2025, made the most generous of them permanent – restoring 100% bonus depreciation and immediate expensing of domestic research costs. Together these provisions further shorten payback periods for capital investment:
| Provision | Current Law | Recent & Proposed Changes | What it means for automation |
|---|---|---|---|
| Bonus depreciation | 100% permanent (restored by OBBBA, July 2025; had been phasing down to 60% in 2024, 40% in 2025, zero by 2027 under the TCJA) | Enacted July 2025 – 100% first-year write-off is now permanent | Cuts after-tax cost of robots, servers, CNC machines by ~21% in the year of purchase. Speeds pay-back periods and raises IRR on cap-ex projects. |
| Section 179 expensing | $1.22m cap, $3.05m phase-out (2024) | Proposals to expand caps would increase limits | Higher caps would allow small manufacturers to write off entire automated production lines in year one. |
| Structures | No bonus; 39-/27.5-year depreciation | Some proposals include accelerated depreciation for qualifying facilities | Would make fully automated plants cheaper to build relative to hiring labor in legacy space. |
| R&D costs (§ 174 / § 174A) | Immediate expensing of domestic R&D restored under new § 174A (OBBBA, July 2025); foreign R&D still amortized 15 yrs | Enacted July 2025 – reverses the 2022 TCJA rule that required 5-yr domestic amortization | Would lower the cost of developing automation software, AI models, and robotics; encourages domestic R&D over offshoring. |
Leading economists attribute labor's declining share to multiple reinforcing factors:
Note: These percentages come from different studies using different methodologies and time periods. They are not additive and should not be summed – factors interact and overlap. The key insight is that no single cause dominates; labor's declining share resulted from multiple reinforcing forces.
All data pulled from official FRED (Federal Reserve Economic Data) API: