The United Auto Workers went on strike in 2023 and won a 25% pay raise over four years. Headlines called it a massive victory for organized labor. Meanwhile, Tesla – with zero union presence – already paid its manufacturing workers competitive wages and handed out stock options that made some of them millionaires. One company needed a six-week strike. The other just paid people.

That tension sits at the center of one of the most emotionally charged questions in economics. Do unions actually raise wages? The honest answer is: sometimes, for some workers, temporarily, at the expense of other workers. That is not the bumper sticker version anyone wants to hear. But it is what the evidence shows.

American Wages Were High Long Before Unions Were Strong

Here is a fact that gets conveniently ignored in every union debate. American workers earned significantly higher wages than their British and German counterparts for decades before the American labor movement had any real power. In the late 1800s and early 1900s, US wages were the highest in the industrialized world. Union membership was negligible.

Why? Because American workers had more capital to work with. More machines, more tools, more factories, more land per worker. A single American farmer with a steel plow and a horse could outproduce five European farmers with hand tools. A factory worker operating a power loom produced more cloth per hour than a cottage weaver could produce in a day.

Productivity drove wages. Not bargaining power. Not picket lines. Not collective agreements. Productivity. This has always been the inconvenient foundation that union advocates have to dance around. If unions are what raise wages, explain why American wages were already the highest in the world before unions showed up.

The Market Already Bids for Your Labor

Employers compete for workers. This is the part of the labor market story that gets zero airtime, because it does not fit the narrative of helpless workers versus evil corporations.

If a software engineer generates $300,000 in value for her employer and is being paid $150,000, a competitor will offer her $180,000. Another will offer $200,000. This bidding process pushes wages toward the worker’s actual productive value. It does not require a union. It requires a functioning labor market with multiple employers.

Look at tech workers. Google, Apple, Meta, Amazon, and Microsoft have spent two decades aggressively poaching each other’s engineers. Salaries in Silicon Valley reached absurd levels not because of collective bargaining but because of collective competition. Companies were literally sued for secretly agreeing NOT to recruit each other’s workers – because the default state of the market is aggressive bidding for talent.

The salary transparency laws spreading across the US right now – requiring companies to post pay ranges in job listings – are doing more to close information asymmetry than decades of union negotiations did. When everyone can see what the market pays, wages converge toward true value without a single picket sign.

Where Unions Actually Help

None of this means unions serve no purpose. They serve several legitimate ones.

First, information asymmetry. A single worker in a factory might genuinely not know what the market rate for her skills is. A union can aggregate that information and negotiate accordingly. This is real value.

Second, working conditions. Individual workers have little leverage to demand safer equipment, reasonable hours, or basic dignity from a large employer. Collective action can fix that. The history of workplace safety improvements is genuinely tied to union advocacy.

Third, bargaining against monopsony. In a small town with one big employer – one mine, one factory, one hospital – workers do not have competing offers to leverage. The single employer can suppress wages below market value because there is no market. A union restores some balance. This is legitimate economics, not ideology.

The Starbucks union drives make sense through this lens. Baristas in many locations face a single dominant employer within their practical commuting radius. Information about pay across locations was opaque. Working conditions varied wildly between stores. Organizing corrected real asymmetries.

When Unions Push Past Market Value, Everyone Pays

Here is where the math turns ugly. When unions push wages above what the market would set – above the worker’s actual productive value – the result is not higher wages for everyone. The result is fewer jobs.

If the market wage for a job is $25/hour and the union negotiates $40/hour, the employer has two choices: automate the position, or hire fewer workers and demand more output from each. Either way, some workers who would have had a $25/hour job now have no job. They get pushed into other industries where they compete with existing workers, driving those wages down.

This is not theory. It is what happened to American autoworkers for decades. The UAW negotiated wages and benefits so far above market rates that Detroit became uncompetitive. Japanese and Korean manufacturers built non-union plants in the American South, paid workers less but still well above average local wages, and ate Detroit’s market share for lunch. The UAW’s “victory” for its members was paid for by the autoworkers who never got hired, the suppliers who went bankrupt, and the communities that were hollowed out.

Tesla’s Fremont factory sits on the exact site of a former GM-Toyota joint venture. Same building. Same region. Non-union. Competitive wages. Producing more cars with fewer workers. The market did not need a strike to set those wages.

The Unequal Raise Problem

Imagine five unions representing five different industries. Farmhands get a 0% raise. Retail workers get 5%. Transport workers get 15%. Construction workers get 25%. Railroad workers get 50%.

The cost of everything made by transport, construction, and railroad workers goes up. The cost of living rises for everyone. But the farmhands got nothing. Their wages are the same, but their groceries, rent, and commute now cost more. They are objectively worse off than before any union negotiated anything.

Even the retail workers with their 5% raise find that prices have risen by 8%. They are also worse off. The only group that actually gained purchasing power is the railroad workers with 50% – and only until prices finish adjusting.

This is not a hypothetical thought experiment. It is the story of every decade of American labor history. The strongest unions in the most concentrated industries – teamsters, steelworkers, auto workers – won enormous raises. Workers in fragmented, hard-to-organize industries – retail, food service, agriculture – got nothing and paid higher prices for everything.

The 2023 Hollywood writers’ and actors’ strikes illustrate this perfectly. The WGA and SAG-AFTRA won significant concessions on streaming residuals and AI protections. Good for them. But the caterers, set construction crews, makeup artists, and location scouts who lost months of income during the strikes had no equivalent leverage. The strongest unions won. The weakest workers around them ate the cost.

The War on Productivity

The worst thing unions do has nothing to do with wages. It is the systematic hostility to productivity improvement.

Police unions protecting officers with documented patterns of misconduct. Teacher unions making it nearly impossible to fire incompetent educators – the infamous “rubber rooms” in New York City where teachers accused of misconduct sat collecting full salary for years while their cases crawled through arbitration. Rigid job classifications where an electrician cannot move a table that is in his way because moving tables is the furniture mover’s job.

These are not strawman examples. They are documented, ongoing, real.

The gig economy fights are the latest battlefield. Uber, DoorDash, and Lyft drivers organized to demand employee classification with benefits. California passed AB5 to force it. The result? Reduced flexibility for drivers, higher prices for riders, and a ballot measure (Proposition 22) where even California voters – not exactly a union-hostile electorate – voted to exempt gig companies. The workers themselves preferred flexibility over union-style rigidity.

Make-work rules, seniority over merit, opposition to automation, deliberate slowdowns, punishing workers who produce “too much” – these practices do not raise wages. They destroy the productivity that wages ultimately depend on. You cannot eat bargaining power. You can only eat what the economy produces.

European Unions: A Different Model, Same Limits

European unions operate differently than American ones. In Germany, workers sit on corporate boards through codetermination. In Scandinavian countries, unions negotiate sector-wide agreements rather than company-by-company contracts. These models avoid some of the worst American union pathologies.

But they do not escape the fundamental constraint. German wages are high because German productivity is high – because of German engineering, German capital investment, German infrastructure. The union structure distributes the output more evenly, but it does not create the output. When German automakers face competition from Chinese EVs, the union seats on the board do not make the cars cheaper to produce.

Scandinavian countries have high wages and strong unions. They also have small, highly educated populations, massive natural resources per capita, and deep capital markets. The unions negotiate the distribution of enormous productivity. They did not create that productivity.

What Actually Raises Wages

If you want to know what actually raises wages across an entire economy – not for one group at another’s expense, but for everyone – the answer is brutally simple and completely unexciting.

Capital accumulation. More machines, better tools, newer equipment, modern factories. A worker with a $500,000 CNC machine produces more value per hour than a worker with a hand file. That difference in output is what allows the higher wage.

Technological advancement. New methods, new materials, new processes. The reason a software developer earns more than a medieval scribe is not that developers have a better union. It is that the tools are incomparably more productive.

These two forces – capital and technology – have raised real wages more in the last two centuries than every union negotiation in history combined. It is not close. It is not debatable. The entire standard of living improvement that separates modern life from 1820 is the product of investment and innovation, not collective bargaining.

Amazon warehouse workers in 2025 earn $20+ per hour with benefits. Not because of a union. Because the robots in those warehouses make each worker so productive that Amazon can pay that wage and still profit. The productivity came first. The wage followed.

The Takeaway

Unions are not evil. They serve real functions: correcting information gaps, improving safety, balancing power against monopsony employers. Those are legitimate roles.

But unions cannot permanently raise real wages for all workers. They can redistribute wages – giving more to organized workers at the expense of unorganized ones. They can temporarily push wages above market value – until the resulting unemployment, price increases, and capital flight bring reality back into focus.

The uncomfortable truth is that wage growth comes from one place: producing more value per hour of work. That requires investment, technology, education, and an economy that rewards efficiency rather than punishing it.

Every dollar of wage growth in history was produced before it was negotiated. The negotiation decides who gets the dollar. The productivity decides whether the dollar exists at all.