For much of the 20th century, workers at big companies were paid better than workers at small ones. An employee of a company with more than 500 employees historically earned 30%–50% more than someone doing the same job at a firm with fewer than 25 employees, for instance. But the pay gap between large and small companies has narrowed in recent years, and that decline is one reason for rising inequality in America.

It’s also a reminder that inequality is deeply intertwined with the day-to-day decisions companies make, say, about outsourcing manufacturing, or contracting with a caterer, or aiming for vertical integration, or focusing on the core. Big firms began doing countless things differently over the last few decades, for just as many reasons. But one major difference in big companies today compared to 40 years ago is that today’s giants pay less generously than the giants of the past, especially when it comes to their lowest-paid employees.

There are multiple reasons why big firms historically paid better than smaller ones. Part of it was related to the people who worked there. Bigger companies could attract qualified, sought-after employees who could demand higher wages. And big companies tended to be more efficient than smaller firms, which meant their workers were more productive and therefore better paid.

Bigger firms also seemed to resist too much inequality developing between pay at the top and bottom. That may have been because of unions, social norms, or the belief that equal pay would make employees work harder or stay longer. Whatever the reason, firms couldn’t get away with paying their top employees less — or else they’d leave — so they ended up paying their less-well-paid employees more. That meant the primary beneficiaries of higher pay at big companies were the lowest-paid people who worked there.

But all of that appears to have changed. Since the late 1980s, the gap between how well big and small firms pay has shrunk, according to a recent paper by J. Adam Cobb of the University of Pennsylvania and Ken-Hou Lin and Paige Gabriel of the University of Texas at Austin. But the gap hasn’t shrunk equally for everyone. Highly paid workers at big companies continue to make a bit more than their counterparts at smaller firms, and this gap hasn’t changed. The shift has been in the pay premium for their colleagues further down the pay scale. Mid- and low-wage workers at big companies still make more than their counterparts at small ones, but nowhere near as much more as they used to.

The researchers estimate that this decline in how much more big firms pay explains 32% of the rise in inequality between the 90th and 10th percentiles of income distribution. In other words, if big companies today paid as generously as they did in the past, incomes would be substantially less unequal.

Why did big firms stop paying so much more than smaller ones? It’s hard to say, since a lot of other things happened in those same decades: the decline of unions, an explosion of information technology, a new round of globalization, and the dramatic rise in CEO pay. In a forthcoming study Nicholas Bloom of Stanford suggests that part of what’s going on is a shift from a manufacturing to a services economy. Large services firms historically have paid better than smaller competitors, but the gap wasn’t ever as large as it was in manufacturing. (It’s not clear why there is a difference between sectors.) Even so, Bloom finds that the big-firm pay gap is shrinking in services, too.

The most interesting explanation is managerial. Cobb and his coauthors suggest that big firms stopped paying lower-level employees as much because they chose to reorient around their core competencies:

Beginning in the early 1990s, many large firms, particularly those involved in the production of goods, sent their manufacturing overseas, and many back-office services found new homes at home and abroad. Spurred by new theory emphasizing the importance of firms to focus on their “core competence,” as well as financial markets rewarding firms for generating profits while harboring fewer physical assets, externalizing their workforce through the use of contract work, temporary work, and outsourcing emerged as the au courant approach for firms attempting to lower labor costs and maintain flexibility in uncertain product market environments.

The theory here is that the big-firm pay premium was partly a consequence of having lots of different kinds of workers at the same company. For example, if a big firm had some cafeteria workers on payroll, it felt at least some pressure not to let their wages fall too far, because inequality was bad for morale. But when corporate catering companies came along, two things happened. First, the catering companies hired employees at the going market rate, without any wage premium. Second, the big companies that still had cafeteria staff started comparing how much it paid those workers to the alternative of contracting with the caterer. As firms restructured around one or a few competencies or occupations, the thinking goes, wages converged toward the market rate.

Last year the Obama White House released a report with a very different theory of why big companies pay what they do. The authors cited “growing concern” among economists and policy wonks over lack of competition, which can shift “the balance of bargaining power toward employers.” The result could be “monopsony,” meaning an economy in which powerful companies don’t have to fear competition, and can therefore get away with paying “a lower wage than they would in a competitive labor market.”

There’s an important disagreement here. The Obama White House thought big companies were paying lower-level employees less than a competitive market would demand. By contrast, the research cited above suggests that those workers historically were paid more than market rate in the first place.

The White House’s argument isn’t fully at odds with the research I’ve cited: “We both argue that large firms play a central role in setting wage[s] and [that] the rising inequality shouldn’t be viewed as purely a ‘market-driven’ phenomenon,” said Ken-Hou Lin.

And both sides agree that something has changed. In 1950 the economist S.H. Slichter wrote: “When management can easily afford to pay high wages, they tend to do so.” That no longer seems to be the case.