You are using an outdated browser.
Please upgrade your browser
and improve your visit to our site.
Skip Navigation

How Amazon Keeps Workers’ Pay Low

New evidence that when Amazon comes to town, warehouse wages go down.


Multiple commentators on economic inequality, including the philanthropist Nick Hanauer and the journalists Noah Smith and Edward Hadas, have suggested that Parker Brothers, the board game company that created Monopoly in 1936, ought to create a new game called Monopsony. I agree.

Monopsony is slightly different from monopoly, though the two usually go hand in hand. Monopoly is the dominance of a market by a single seller. Monopsony is the dominance of a market by a single buyer, especially of labor. The monopolist says: “You don’t like my prices, or my level of services, or the quality of my product, go find somebody else you can buy from. Ha ha!” The monopsonist says: “You don’t like what I pay you, or how I treat you, go find somebody else you can work for or sell to. Ha ha!”

As I see it, a Monopsony board game would have the same game pieces—thimble, hat, iron, shoe, etc.—moving around the same game board featuring the same Atlantic City street names. It could even keep the “Marvin” in Marven Gardens misspelled with an “i,” just like in Monopoly. But whenever a player landed on Baltic Avenue or Park Place or wherever, instead of paying the owner of that property rent, the player would take a job.

You might think the next move would be for the property owner to pay some amount to his new employee. If you think that, you don’t know your Marx. In Monopsony, the employee forks over the surplus value of his labor, i.e., the difference between the wages he receives from the property owner and a higher figure representing what the employee’s labor is really worth. You don’t have to be a Marxist to recognize that the property owner isn’t going to hire the new employee unless doing so will increase revenues and, ultimately, profits.

At the start of the game, a new employee’s surplus value would be low—let’s say, $50—reflecting a competitive market for labor among the other Monopsony players. But as the game progressed and other players either dropped out or fell behind in the purchase of properties, surplus value would rise in $50 increments, reflecting a shrinking number of competitors for labor. That would be true regardless of whether the property your game piece landed on was owned by a player who was winning the game or losing it; monopsony lowers wages paid by everyone within a given industry (in this instance, real estate). As with Monopoly, the game would end when a last losing player or players declared the winning player an asshole and, in a fury, turned the board over, sending all the pieces flying.

Alas, Parker Brothers no longer exists. In a choice example of life imitating board games, Parker Brothers was gobbled up by Kenner in 1985, Kenner was gobbled up by Tonka in 1987, and Tonka was gobbled up by Hasbro in 1991. So it falls to Hasbro to create a new game called Monopsony.

Monopsony would need a mascot, of course. Monopoly has Milburn “Rich Uncle” Pennybags, a top-hatted, mustachioed, portly old gent you probably know as Mr. Monopoly. Mr. Monopoly is really J.P. Morgan; that’s why he looks just like him. Monopsony would need a mascot too. Mr. Monopsony wouldn’t be fat, because girth is no longer a signifier of great wealth. The top hat would obviously have to go as well, and also the mustache, because mustaches are now a rarity among the Forbes 400. (I count only two in the top 10.) Mr. Monopsony would be slender and fit, bald, and clad in a black leather jacket. His right eye would be disconcertingly smaller than his left. That’s right. Mr. Monopsony would be Jeff Bezos.

Writing last week about the Federal Trade Commission’s antitrust lawsuit against Amazon, I observed that Bezos was playing Monopsony against Amazon’s sellers and against Amazon’s mostly nonunion warehouse employees. But I qualified that by saying workers’ beef with Bezos was typically not about wages but working conditions. Now a fascinating new report from the National Employment Law Project, or NELP, argues that Amazon’s warehouse wages aren’t what they’re cracked up to be. An Amazon warehouse’s presence in a given county, NELP says, drives wages down, just like monopsony theory says it should.

Amazon is, after Walmart, the nation’s second-largest private-sector employer, employing roughly one million people in the United States. It might at first seem as though Amazon pays well, because counties with Amazon warehouses have higher average earnings than counties without Amazon warehouses. But it turns out that isn’t because Amazon pays well. It’s because, the NELP report says, Amazon chooses to locate most of its warehouses in high-earning counties (i.e., counties where average earnings exceed the eightieth percentile), apparently to situate them closer to wealthy Amazon consumers.

Amazon doesn’t furnish much detailed information about what it pays its warehouse employees, so NELP gathered its information indirectly through census data. It found that when you take pay for warehouse employees in high-earning “Amazon counties” (i.e., the high-earning counties where most Amazon warehouses are situated) and compare that to pay for warehouse employees in comparably high-earning “non-Amazon counties” (i.e., counties where there are no Amazon warehouses), the warehouse workers in the non-Amazon counties earn about 18 percent more. That translates into about $822 more per month. This pattern is repeated when you compare pay for warehouse employees in all “Amazon counties” (i.e., all counties where Amazon warehouses are situated, not just the wealthy counties) to warehouse employees in all “non-Amazon counties” (i.e., all counties that have warehouses in them, but no Amazon warehouses). Here, warehouse workers in the non-Amazon counties make about 12 percent more.

Warehouse workers in high-earning Amazon counties make 30 percent less than average earnings for all workers in those counties. That’s unsurprising, considering that these are high-earning counties; in a high-earning county, of course a warehouse worker will earn less than the average worker. But in comparably earning non-Amazon counties, warehouse workers make only about 8 percent less than average earnings for all workers. “Warehouse jobs are middle-income jobs or at least closer to middle-income jobs in most of the U.S.,” the NELP report says, “except in the counties where Amazon operates fulfillment centers.”

Remember how in our Monopsony board game a player’s falling wage obliged him to hand over more in surplus-value payments as the game progressed, no matter whose property his game piece landed on? NELP says that’s what’s happening here.

In the first quarter of 2005, Amazon had only three warehouses in the entire country. Today it has more than 300. NELP looked at what warehouse workers were paid before Amazon arrived and after. In 2005, warehouse workers in future Amazon counties made about the same as warehouse workers in future non-Amazon counties. The differences described above between warehouse workers in Amazon counties and non-Amazon counties didn’t exist until Amazon arrived. NELP also found that the gap between what warehouse workers were paid and average earnings in a given county grew after Amazon arrived. In non-Amazon counties, the gap remained about the same.

In his 2021 book Fulfillment: Winning and Losing In One-Click America, Alec MacGillis observed that warehouse jobs “used to be considered somewhat higher-skilled jobs: One could make over $20 per hour and stay years at a time.” But at Amazon, workers have “a more fleeting existence.” They’re younger, and turnover is much higher. Last month Amazon said it was raising its starting minimum pay to $17 per hour, which is pretty good, but, crucially, NELP notes, it didn’t index that minimum to inflation. Indeed, raises will be capped after three years. This is “by design,” NELP says, “as a way to foster high turnover and incentivize workers to leave.”

One way to liven up the Monopsony board game would be to give players the option of joining a labor union. A union can defeat monopsony by compelling management to share more of its monopoly profits with the workers. I propose a Community Chest card that says, “You just won an NLRB-supervised union election! Lower all players’ surplus-value payments back to $50 for the remainder of the game.”

That would add a fun wrinkle. But the Community Chest stack would have to be tall enough that the odds of anyone pulling the union-victory card would be fairly remote in any given game. Otherwise the game just wouldn’t be realistic. MacGillis noted in Fulfillment that the main benefit to Amazon in high warehouse turnover was that “it made it much easier to stamp out union efforts.” An Amazon warehouse in Staten Island is the only one to have organized successfully—ever. That happened 18 months ago, and no contract agreement has been reached there.

Amazon continues to fight efforts to unionize additional warehouses, because it knows that unions represent a mortal threat to monopsony. NELP’s report urges the company to accept the informal collection of union authorization cards as an alternative to NLRB elections at Amazon warehouses, and to bargain in good faith with union locals once they’re formed. A union is the only way to achieve NELP’s other recommendations, such as better pay, safer working conditions, 72 hours’ paid sick leave, and 12 weeks’ medical and family leave.

But I don’t see Amazon ever easing its militant stance against organized labor. The real Jeff Bezos is surely trained by public relations professionals never to discuss such matters. But if you were to query Bezos’s alter ego, Mr. Monopsony, about NELP’s request, I imagine he’d reply: “Don’t be absurd. That’s not how this game is played.” Then he’d roll the dice, land on Free Parking, and collect all the cash in the middle of the board.