The Economic Impact of Telework (Part 1)

Outsourcing Mental Work Away from Creative Class Cities

When history looks back, the most important economic effect of the Covid-19 pandemic is likely to be the telework breakthrough. Mass telework seemed to be a technical possibility for twenty years before the Covid-19 pandemic, but somehow, it never seemed to take off. Then suddenly, in April 2020, government-imposed pandemic lockdowns turned the vast majority of white-collar work remote overnight. And when the lockdowns ended, telework mostly stayed.

Multiple Equilibrium, and Why Telework is Here to Stay

Doubtless there were some organizational learning hiccups in emergency telework adoption. But all in all, the transition seems to have been very smooth. Lots of things broke in the US economy in 2020, but white-collar work wasn’t one of them. Organizations kept functioning. In the aftermath, office attendance mandates look hard to justify. We know the old suit-and-commute routine isn’t actually necessary. So the return-to-office movement has been rather halting. Many bosses and some employees prefer the office for a variety of reasons, good and bad, and as the pandemic lockdowns recede into memory, some organizations have tried, with varying success, to opt out of the telework revolution. 

Telework has obvious upsides. (No time is wasted on commuting. There are fewer carbon emissions and traffic jams.) There are also arguable downsides. (Does it make work less sociable?) But rather than discuss them here, I’ll assume what I've concluded from thinking and reading, namely, that telework is an almost perfect substitute for office work. My real interest here is to study what follows for the future of the US economy. 

Essentially, I think there are multiple equilibria, including a “telework is weird” equilibrium where adverse selection turns telework preferences signal low worker commitment so normal professionals go to the office without understanding why, and a “telework is normal” equilibrium where teleworkers are trusted so lots of normal professionals telework. The Covid-19 pandemic flipped us from “telework is weird” to “telework is normal.” Once established, “telework is normal” has staying power, because a critical mass of high-quality professionals are doing it, so the adverse signal of telework is destroyed, and the cost advantages of telework– less office space maintenance costs– assert themselves. 

If anything, I’d expect the telework revolution not only to last, but to deepen over time, as office space leases run out and companies save costs by removing an under-utilized office from the budget. If so, how will US economic geography change? To understand that better, I’ve developed an economic model, or rather, two interconnected models.

A Methodological Digression: Why Economists Should Theorize with Graphs

Here I have to stray, with apologies, into some methodological soul searching. Feel free to skip ahead to the next section if you're interested in the economic impact of telework but not in methodological reform of economics.

“Economic models” have a mixed reputation. For some, they have great mystique and authority. For others, they’re a laughingstock, or sinister propaganda. The economist’s habit of making assumptions for the sake of a theory has inspired an old joke about economists. Here’s a version of it:

A physicist, a chemist, and an economist are shipwrecked on a desert island, and getting hungry. Fortunately, a crate of canned food washed ashore. But they have no tools to open the metal cans. What should they do? 

First, the physicist suggests finding a piece of curved glass, and directing the sun’s rays to a single point until the metal of the can heats up to the point of melting. But it sounds difficult to find glass of the right shape.

Second, the chemist suggests some ways to make an explosive from substances they might find on the island. But it will take a lot of searching and digging and mixing in the best cases. 

So the economist speaks up. “You guys are overthinking it,” he says. “It’s easy. Just assume a can opener!” 

Ha ha. It’s not quite fair. Economists’ assumptions aren’t as arbitrarily convenient as that. And it’s not just economists who rely on simplifying assumptions to interpret the world. If you’re planning your day, and you’ll need to make a trip to the store, you might assume a car speed in order to calculate how long you’ll take to get there. Your actual car speed will vary. But the assumed constant car speed makes the calculation feasible, and yields a not too inaccurate result. 

Still, economic theorists have given assumptions a bad name by making assumptions that are too unrealistic, without enough intuitive buffering. To some extent, economists themselves have lost faith in theory for this reason, provoking an overreaction into “empirical” economics. The word “empirical” might suggest that economists have strayed into the fallacy of thinking a science can just be a massive heap of facts. Sometimes. But the real problem is a little different. An odd metastasis of the computer revolution with the publish-or-perish obsessions of academia has spawned a massive “literature” of papers founded on the conceit that the human power of pattern recognition can be simulated by statistical tricks. It has become routine to look for accidentally emergent natural experiments, and then launch wild, irresponsible extrapolations in a desperate reach for relevance. This “literature” fails to cohere into any larger understanding of the world, but it does get a lot of people tenure, which is why it keeps being produced. Statistical obsessions also provide a certain security, in woke academia, from being a politically incorrect lightning rod. “Don’t blame me,” the econometrician protests. “The data made me say it!” It’s a mess.

Meanwhile, what should you do if you don’t care about the academic rat race, but you just want to think clearly about the economy, and discover and share some useful truths?

I think the answer is to avoid both “theory” and “empirics,” as the pride of the profession now conceives those things, and instead, to borrow one’s methodology from the humble undergraduate economics classroom. 

Lots and lots of people have taken economics courses in college, and come away feeling like they understand capitalism better because of it. There, at least, economists do some good. How? Above all, with theory that uses graphs. Not equations. Undergraduate economics classrooms feature few equations. And not regressions. Regressions themselves rarely put in an appearance in the undergraduate economics classroom, and while plenty of truths are reported there that have some empirical support from regression analysis, rarely if ever are undergraduate economics students illuminated by a counterintuitive truth that is known only or mainly because of regression analysis. There is some game theory, too. Graphs and game theory are the methodological toolkit of the undergraduate economics classroom.

I think that graphs and game theory should have pride of place in the methodological toolkit of the academic economist’s research, too, for reasons, both of practical persuasion, and of lofty epistemology that I don’t have time to deal with fully here. Academic research methods should be more continuous with the classroom methods of teaching by which economists earn their bread. A pivot that takes place when the bright econ undergrad turns into a PhD student, away from graphs and game theory, in favor of equations and regressions. It’s a mistake. There are reasons for it, of course. Translating theories from graphs to equations allows the dimensions to increase from two or three to n, but it also requires additional unrealistic assumptions and makes the model more opaque. The price is too high. And “empirical” researchers must always remember that data tests theories, that generality is the prerogative of theory, and that no empirics can be better than the theories that it tests.

Meanwhile, I’ll cut this methodological discussion short and get back to my model. But now you have a glimpse of why it features graphs.

Outsourcing Mental Work from Metro to Non-Metro Areas 

Clearly, not all jobs can telework. Many jobs, probably most, inherently require physical presence, but others don’t, and an ancient distinction executes the sort pretty well. People have long recognized that some work with their hands and some work with their minds. It’s the manual workers versus the mental workers. Mental workers can telework, manual workers can’t. The phrases “white collar” and “blue collar,” though sartorially obsolete, have a useful familiarity, and map decently well onto the mental work vs. manual work distinction. The contemporary buzzword “laptop class” is helpful here as well, for computers are the principal job tool of the mental worker today. College degrees are strongly associated with being a mental worker, too.

So the first step in my model is to focus on mental workers by drawing a graph whose lower axis is the economy’s total supply of mental workers. Left out are a lot of blue-collar workers who are unlikely ever to work full-time at a computer. But included are a lot of people who could do fine working full-time on a computer, and might like to, but they do manual work instead because it pays better where they are, or because they can’t find a mental work job. They’re also part of the supply of mental workers even though they’re not doing mental work at the moment. With that, it’s probably easiest to show the whole chart, whose features I'll continue to explain.

Figure 1

Second, note that the chart distinguishes “metro” and “non-metro” mental worker populations. I’m assuming for now that workers are place-bound. We’ll relax that assumption later. Metro vs. non-metro is a bit different from the more familiar, though not very well-defined, urban vs. rural distinction. I want to capture the difference between (a) a few “creative class” metro areas where the economy's cutting-edge, creative, innovative, visionary, cognitively demanding work is disproportionately concentrated, and (b) everywhere else, including rural areas and small towns, but also some less fashionable metro areas. High housing prices are typically (not always) a symptom of being a creative class metro. Most of the mental workers live in “non-metro” areas. 

Third, Figure 1 shows the demand curves for mental workers in metro and non-metro areas. Each demand curve, as usual, is a combination of quantities and prices, indicating how much someone, in this case employers, would pay for a given quantity, in this case of mental workers. The prices here are wages, indicated on the vertical axes. The left vertical axis shows the mental work wage in metro areas, the right vertical axis, in non-metro areas. Of course, it’s a major simplification to show a single wage for each area. For one thing, some mental workers are more productive than others, but also there’s a lot of history and happenstance that gets built into wage structures in a way that’s far from fully rational. Still, market forces have always caused some convergence pressure within the class of metro area mental workers, and within the class of non-metro area mental workers. The metro demand curve slopes down from left to right, as the quantity of metro mental workers increases. The non-metro demand curve slops down from right to left, as the quantity of non-metro mental workers increases.

Fourth, let’s take a look at the pre-telework equilibrium. Metro area mental worker wages were determined by the intersection of the metro demand curve with the metro mental worker labor supply curve. They are relatively high. Non-metro area mental worker wages, likewise, were determined by the intersection of the non-metro demand curve with the non-metro mental worker labor supply curve. They are relatively low. 

Now we’re ready to see the impact of the telework revolution. With telework, metro-based organizations can seamlessly hire mental workers from non-metro areas. And so the equilibrium moves to the intersection of the two demand curves. In the new equilibrium:

• Metro area mental work wages are lower than before

• Non-metro area mental work wages are higher than before

• Mental work wages are the same across metro and non-metro areas

• Metro-based employers of mental workers expand and end up with an increased share of total mental work employment

• Also, metro-based mental work employers see their surpluses increase (which might mean greater profits or be passed on to consumers as increased consumer surplus)

• Non-metro based mental work employers shrink, and see their surpluses fall both because they're smaller and because they have to pay higher wages

• Total product increases

• Total productivity increases because the increase in total product doesn't depend on an increase in the factors of production, but n improved job matching between mental work employers and employees

Let's think a little more about the meaning of those shifts. What will the future world be like, in light of this model? Here are some predictions.

1. Rural and small town college grads will work remotely for offices in creative class metro areas and make more money. 

The telework revolution is great, above all, for people like me, cognitively endowed people living for personal reasons in out of the way places, for whom telework opens up magical new possibilities of combining what has traditionally been an urban career with a rural lifestyle. It doesn't just benefit us, either. Teleworkers bring funds into the local economies where they live. They earn elsewhere and spend near home, and their spending has multiplier effects and creates jobs for others.

2. Rural and small town organizations will struggle to retain college educated workers. 

A disadvantage for telework exporting regions is that local organizations face new labor market competition from far away. The key question here is whether this represents a dangerous downside of the telework revolution, or whether it's little loss because those organizations didn't make good use of mental work skills anyway. Doubtless, there are small town school teachers who can now, or will be able soon, to make more money as proofreaders or project managers for far away corporate offices, and who, if they leave for new opportunities, will be sorely missed. There are also a lot of college grads living near home for the sake of family and community and working construction or fast food jobs that don't use their degree at all. Telework will fix a lot of “mal-employment.”

3. Metro-based mental workers will face new competition, but also new opportunities. 

A very interesting feature of the model is the impact on metro-based mental workers. In the pre-telework equilibrium, they look privileged, enjoying high wages because of where they live. Prima facie, they are big losers from the telework transition. Suddenly, they’ll find that a whole lot of newly-empowered rural and small town mental workers are hot on the heels of their jobs. They’ll start feeling severe downward wage pressure. As this model underscores, they have reason to be worried. 

But there are several considerations to weigh against this.

First, some metro-based mental workers might prefer a rural lifestyle, but pre-telework, that wasn’t an option professionally. Now, it probably is. In the best case, they’ll be able to take their jobs with them, and get the best of both worlds: the high-flying urban career, and the pleasant country estate. It’s good for family formation, since it’s easier to raise kids if you live where a backyard doesn’t cost a million bucks. It’s nice to spread out a little bit and have more space. And for those who left a beloved hometown under economic duress for the sake of a faraway city job, telework could be a godsend, the ticket to a joyful homecoming that was almost beyond hope. 

Second, the model shows metro-based organizations expanding impressively. For those who are trusted insiders in these expanding organizations, such expansion should generate some opportunities for promotion. 

Third, strong norms of downward nominal wage stickiness make it unlikely that metro-based mental workers will face explicit wage cuts. A period of 0% annual raises might be on the cards as organizations use their cash to hire remotely instead of giving raises to expensive urban professionals.

But fourth, the outsourcing of mental work may ease pressure on metro-area housing prices. Metro-based mental workers who stay put might see it become easier to buy or rent.

Finally– fifth– cities are resilient places, whose economic function has kept adapting over time. Entrepreneurs may find that if mental work is getting outsourced from cities, it’s a good time to innovate new jobs that combine some physical presence value-add with the intelligence and professionalism. Metro-based mental workers, even as they see the office jobs filter out to the countryside, may be first in line for the jobs of the future. More about that in the next post.

4. In the long run, higher productivity will be amplified by more accumulation of human and physical capital.

How long these trends will take to play out is a complex question. Organizations can take a long time to adjust their staffing levels. Teams have institutional memory. New staff turn over a lot. A future where a substantial share of the population lives in rural areas and small towns while teleworking in faraway metro-based office jobs will surely take a decade or more to mature. 

Meanwhile, it’s very exciting to a growth economist like me that the equilibrium shift in Figure 1 raises “total factor” productivity. Robert Solow showed long ago how economic growth can be powered by capital accumulation up to a point, but eventually gets overtaken by diminishing returns and the burden of maintaining a larger capital stock. But whenever “total factor” productivity (that is, productivity for a given quantity of the factors of production) increases, there are knock-on effects over time through capital accumulation. Solow was thinking mainly of physical capital, but the same logic applies to human capital as well.

Now, telework raises total factor productivity. We know this from Figure 1 because the combined area under the demand curves is greater. The productivity boost comes from better matching of employers and workers. But that also means society can afford to maintain a larger capital stock, both physical and human. Since it's making better use of educated workers, it can justify educating more workers. Higher profits will get reinvested. And high-income teleworkers will maintain and expand rural homes.

Telework should be a great stimulus to education and professional development in rural and small town America, and in some of the less creative-class metro areas. Hitherto, if people from those places got top-notch skills, they'd have to go away in order to make full use of them. No longer! Now higher education might be the best way to make sure that you can stay in your hometown. 

So Figure 1 forecasts a brighter future, where the wealth of metro-based creative capitalism gets spread to a lot of left-behind places, leavening society for a general quickening of broad-based economic growth for a while. But there’s more to be said. For while so far we’ve assumed that companies stay put in the age of telework, while jobs move, telework actually changes the economic logic of site selection, and there are good reasons to think it will transform industrial geography, driving a certain kind of corporate decentralization while also opening the way for the emergence of new industries. Read more in Part 2 (upcoming).

By the way: let me note that this model spins off a lot of apparent forecasts, some of which could be checked right now, since the telework transition is well underway, and some of which might make sense to check in five or ten years. The forecasts aren’t very exact, and if they’re not vindicated, that doesn’t necessarily disprove the model, since other factors could offset. Still, the predictions keep the model honest. I’ve said something meaningful because I could turn out to be wrong.