It’s the Productivity, Stupid (or Why Rising Wages Drive Economic Growth)

WalmartI agree 100 percent with Forbes blogger Tim Worstall. No, not with his blind faith in the divine wisdom of the free market, or with his pioneering use of intra-word semicolons. I simply agree with Worstall’s belated acknowledgement that higher wages lead to higher productivity:

Costco pays very much higher wages than Walmart. It also employs about half the amount of labor per volume of sales. Costco is therefore simply in a different model along that same spectrum of trading off hourly productivity against hourly wages. And note the obvious point here: Costco pays much better but also uses many fewer labor hours. And as Walmart moves along that same spectrum of possible labor models it is facing exactly the same calculation. Raise the wages in order to get more productive staff and the very point and purpose of what you’re doing is to reduce the number of labor hours you must purchase.

Of course, this is part of what we’ve always argued when we’ve insisted that a higher minimum wage won’t inevitably lead to some combination of higher prices, lower profits, worse service, and/or shuttered businesses. There are real benefits that accrue to employers in moving toward a higher wage business model, two of which include lower rates of turnover and higher rates of productivity.

Higher productivity—even according to traditional economics, that’s a good thing, right? Indeed, throughout the course of US history (if in fits and starts), the steadily rising productivity of the American worker has helped to lift living standards for us all.

And yet, from his headline—”Of Course Walmart Cut Hours After Raising Pay–What Did You Expect?”—Worstall seems to imply that rising productivity is actually bad for Walmart workers? Weird.

The implied threat is that if workers are more productive, Walmart will need less of them, meaning a higher wage will at least indirectly cost some Walmart workers their jobs. It’s a more nuanced argument than the old “if the cost of labor rises businesses will purchase less of it” meme that minimum wage opponents usually put forth, but it still oversimplifies what is actually happening on the ground at companies like Costco and Walmart.

You see, Costco’s higher-wage model doesn’t just deliver substantially higher sales per employee than Walmart, it also delivers substantially higher sales per store and sales per square foot. But more importantly for the purpose of this discussion, year after year, Costco’s same-store sales growth is consistently higher than Walmart’s—7 percent versus 3.5 percent in the most recent quarter. And even though Costco requires fewer employees per unit of sales, the faster Costco grows sales, the faster it adds workers. And the same dynamic should hold true for Walmart, at least once it fully realizes the productivity enhancements that naturally come from employing a more experienced, loyal, and competitively compensated workforce.

Thus when Worstall concludes that “the very point and purpose” of what Walmart is doing is “to reduce the number of labor hours you must purchase,” he’s actually missing the point entirely. Yes, a more productive workforce means Walmart will need fewer labor hours per unit of sales. But sales are not static. In fact, rather than lowering labor costs, the very point and purpose of what Walmart is doing is to increase same-store sales.

Worstall describes the wage/productivity calculation as a “trad[e] off” between wages and jobs, but this isn’t a zero-sum game at either the micro- or macroeconomic level. After all, there’s a reason why Republicans are running on economic growth: because economic growth is how we generate jobs. “That will be my goal as president,” Jeb!™ absurdly promised back in June, “4 percent growth, and the 19 million new jobs that come with it.” Okay, whatever. Still, GDP growth is mostly a function of change in the size of the workforce plus change in productivity. The more we increase productivity, the more the economy grows. And the bigger the economy, the more jobs.

So if (as Worstall correctly argues) higher wages drives higher productivity, and higher productivity drives economic growth, and economic growth drives the creation of new jobs—then logically, higher wages must drive the creation of new jobs. Worstall can’t have it both ways. Rising productivity can’t simultaneously increase and decrease net employment. And since over the history of market capitalism rising productivity has generally led to more jobs, higher wages, and higher living standards, I’m guessing it’s the former.

Interestingly, the whole implied threat in Worstall’s wages argument is identical to the threat at the heart of the “robots are coming for your jobs” meme: rising productivity is somehow bad for workers. (That is the whole point of automation, after all—to increase productivity.) But in the long term, at the macroeconomic level, this dystopian vision of productivity as a job-killer simply cannot come true… at least not in any economic, social, or political system remotely similar to the one we live in today.

In our current economy, productivity drives growth, and growth creates jobs. That’s how it works at Costco. That’s how it works at Walmart. And that’s how it works in the economy at large.

AEI’s Mark Perry Inadvertently Makes the Case that $15 is 31% Below the Peak Minimum Wage


Over at the American Enterprise Institute’s Carping Diem blog, Professor Mark J. Perry flexes his Excel spreadsheet muscles by producing a chart (above) tracking Seattle’s minimum wage since 1938. The dark blue line is the real (inflation-adjusted) minimum wage in 2015 “dolalrs” [sic]. The light blue line you can ignore; nominal dollars are just plain stupid.

Pointing out that Seattle’s  2017 minimum wage of $15 an hour will be about 36 percent above its 1968 peak (actually, closer to 32 percent when you account for two additional years of inflation), Perry asks: “Economic death wish? Free lunch?”

Well, to help Professor Perry answer his own question, I’ve annotated his chart (below) to put his numbers in some actual, you know, context:

anotated_perry_chartTurns out that Seattle did quite well during most of the first three decades of Perry’s chart, quickly growing to become our nation’s 19th largest city. It was only after the federal minimum wage was unpegged from productivity growth that Seattle’s economy hit the skids. Then, starting in the late 1980s, right around the time Washington State started boosting its minimum wage substantially above the eroding federal rate, Seattle’s economy started to boom! Today, Seattle not only has one of the highest minimum wages in the nation, it is also one of the fastest growing big cities!

Sure, correlation does not equal causation. But it’s hard to look at past performance and conclude that a $15 minimum wage is some sort of economic death wish.

Still, the larger problem with Perry’s chart is that when he talks about our 2017 minimum wage being 36 percent above the 1968 peak (well, 32 percent), he glosses over the question, peak of what? For example, from 1938 through 1968—through Democratic and Republican administrations alike—there was a bipartisan commitment to raising the real value of the minimum wage roughly in step with rising productivity. Just take a look at the chart below (rather than the lazy straight lines I pasted into Perry’s chart):

Productivity vs Min. Wage (via

Productivity vs Min. Wage (via

That’s productivity growth versus the real federal minimum wage, rather than Seattle’s, but the point remains the same. Through 1968, the minimum wage largely tracked productivity—not inflation or median wage or anything else. So if Perry truly believes that the 1968 minimum wage is a meaningful standard by which to compare Seattle’s 2017 minimum wage, wouldn’t it be more accurate to describe $15 as 31 percent below the productivity-adjusted peak rather than 32 percent above?

If Perry says “no”—that productivity is not an index by which we should adjust the minimum wage—then what’s his point? That $15 is economic suicide because it’s 32 percent higher than some arbitrarily arrived-at number? That’s not much of a persuasive argument. But if Perry says “yes”—that the productivity-adjusted 1968 minimum wage is a reasonable standard by which to judge future minimum wage increases—isn’t he just validating the very notion of a productivity-adjusted minimum wage?

The 1968 metric is either bullshit or it isn’t. Which is it, Professor Perry?