The Economic Impact of Telework (Part 2)
Decentralizing Industrial Geography and Spawning New Industries
In part 1 of this study of the economic impact of telework, my main theme was how telework is allowing, and will continue to promote, the outsourcing of mental work from creative class metros to other cities, small towns and rural areas, thereby raising economy-wide productivity and non-metro area mental work wages, while mental workers in creative class metros will face formidable new competition from teleworkers, but also new opportunities.
We now turn to the impact of telework on industrial geography. Telework opens up new possibilities for how production can be distributed over space. This effect varies a lot depending on the type of production. At one extreme, the production of software can (it stands to reason) be completely place-independent, even if some software firms still like a bit of physical presence. The work is done on computers and the product consists of information that lives in computer memories. Many other productive activities– growing food, making widgets, fixing houses, and so forth– have physical products and clearly require some physical presence, yet there are always some abstract, remote-capable aspects of production if you know where to look– accounting, contracts, scheduling, sales, payroll, and so forth. The mix varies a lot.
In the age of telework, activities requiring physical presence will continue to pin down the locus of production to some extent, but a whole array of supporting mental work activities can float off into the cloud. The model presented here elucidates how that will change site selection and industrial geography.
Industry, Firm, or Plant: Which Organizational Level Drives Site Selection?
What before how. Before we model how site selection decisions are made, we need to wrestle what does the deciding. Is it the industry? The firm? Or the plant?
In one sense, it has to be the firm. Firms have executives and management structures. They’re a locus of decision-making. Plants are subordinate units within firms. They do as they’re told. And industries are more or less decentralized and have no governance. Firms decide.
But firms may not have much freedom of action. Their decisions may be more or less dictated by a variety of circumstances, especially prices and resource availability. They can’t do what’s infeasible, and they can’t sustainably do what’s uneconomic. Firms make mistakes, of course, but their mistakes tend to be self-correcting. If a firm starts a factory in the wrong place, it may lose money and need to be shut down. If the firm is too stubborn to shut down money-losing factories, eventually the firm has to shut down.
In other words, firms are often pawns of market forces.
And one way to understand the market forces is to set aside real firms in favor of idealized firms. Firms may not be inclined, or may not be smart enough, to maximize profits. But in a competitive market, firms that operate close to the profit-maximizing ideal will tend to survive and grow, while firms further from that ideal will tend to shrink or disappear. Firms may be rationally aware of what roughly maximizes profits, or firms that hit on profit-maximizing behaviors by luck may become normal through survival-of-the-fittest. Either way, rational self-interest or evolutionary dynamics, or some mix of the two, make firms operate in fairly profit-maximizing ways. So economists can often get insight by jumping over the evolutionary dynamics and assuming profit maximization.
These generalizations should be borne in mind when we think about regional economics.
A good starting place for regional economics is to recognize, and seek explanations for, certain stylized facts about what industries operate where. We can start with a list of well-known regional economic stereotypes, such as the following:
New York – Finance
San Francisco – IT and software
Los Angeles – Film and entertainment
Nashville – Music
Houston – Energy
Detroit – Cars
Las Vegas – Gambling
Boston – Top universities
Washington – Politics and bureaucracy
Seattle – Airplane manufacturing
Never mind how accurate these stereotypes are for the moment. There’s clearly enough truth in some of them to constitute a pattern of regional economic concentration, though it’s very rough and imperfect. Why does regional economic concentration occur?
Sometimes natural resources are a factor. Houston is in oil-rich Texas. Sometimes it’s political. Washington has a lot of government bureaucracy and political lobbying because it’s the nation’s capital. Las Vegas has a lot of gambling because it’s legal there. Logistics can be important, too: a seacoast with a good harbor, a central location at the meeting of many roads, or a river may determine where an industrial hub springs up.
But usually, industrial concentration seems very underdetermined by the location and natural features of a site. Often, history matters most, or to use a term of art, there is path dependency. The industry is concentrated there now because it was concentrated there ten, twenty, thirty, forty years ago. It’s not because firms feel duty-bound to honor old traditions that history matters. History matters in part because it’s costly for industries, firms, and plants to move, but also, because firms often want to go to where other firms in the same industry are located. There are “economies of concentration.” Finance firms go to where the finance workers are, and finance workers to where the finance firms are, so finance stays in New York. Software start-ups go to where the software venture capitalists and the software engineers are, and vice versa, so software stays in San Francisco.
That’s not to say that industrial geography is immutable. Big changes are hard, but they happen. Rather, the point is that sometimes firms don’t have much choice. The site selection can happen at the level of a whole industry, and firms are compelled by market forces to follow the economies of concentration. But in other cases, an apparent industrial concentration is mostly just the footprint of one big firm, or two or three.
Taking some examples from the list above:
New York’s leadership in finance, and San Francisco’s leadership in software, reflect industry-level economies of concentration, with many firms trading with and investing in each other while competing for workers.
But Seattle’s leadership in airplane manufacturing is just the footprint of one firm: Boeing.
And Detroit’s leadership in carmaking is largely the footprint of two firms: Ford and GM.
Moving from famous cities to mid-sized cities and small towns, the pattern of regional economic concentration persists– it’s common for a city, town, or area to be disproportionately dependent on one or two or three industries or firms for its work– but rather than a corporate headquarters, plants or factories under the aegis of some larger corporate entity often supply the jobs. Again, geography often helps, but history and path dependency are usually as or more important.
For more exploration of this, Michael Porter’s classic The Competitive Advantage of Nations is a good, practical guide, albeit focused on nations rather than regions. For present purposes, I want to abstract from the distinction between industries, firms, and plants. These issues of industrial organization– How big are firms? What are the boundaries of an industry?-- are often underdetermined and not all that impactful. I’ll talk about “industries,” but readers can keep firms and plants in mind as substitute concepts. The point is that whatever the locus of site selection decisioning is, that’s what we’re modeling.
But before we go to the model, a brief meta-theoretical ramble is needed, to serve as a kind of expectations management, and dispel vague impressions that some readers might have, especially if they’ve studied economics in grad school, that the graphical arguments presented below are mere placeholders for a real theory. There’s a reason I’m going about my inquiry in this hasty DIY way rather than writing down lots of equations and citing a boatload of arcane academic papers.
Probing the Mystery of Economic Theory’s Strange Spatial Blind Spot
The sad fact is that economists are not as insightful as might be hoped about industrial geography, for deep reasons that were very astutely diagnosed in Paul Krugman’s book The Spatial Economy. Economists are addicted to the idea of perfect competition, and to the tidiness that it enables economic theory to achieve, at the cost of a fatal lack of realism.
Industrial clusters are not news. The phenomenon was studied, for example, in Alfred Marshall’s Principles of Economics as long ago as 1890, and I think it was a well-known stylized fact about the economy even then, albeit far from equally applicable to all places and industries. But Marshall’s explanation of industrial clusters as a result of “knowledge spillovers,” though not without merit, was in important ways a misstep. Marshall had to envision industrial clusters as not a market-driven phenomenon, exactly, because he, and the whole neoclassical school of which he was an eminent spokesman, were busy reorganizing and unifying economics around the themes of perfect competition and competitive equilibrium.
For generations now, economic theory has become strongly habituated to solving for “equilibrium.” An equilibrium is interpreted as a conclusion with a high degree of approximate generality, a state towards which the economy persistently converges. And who brings about the “equilibrium?” On paper: the economic theorist. In the real world: the “invisible hand” of the market, which is a placeholder for all sorts of individually rational, utility-maximizing or profit-maximizing households and firms whose detailed behavior is omitted from the model in the interests of simplicity.
But what if there is no equilibrium? Or what if there are multiple equilibria? It’s easy to write down game-theoretic situations in which there are multiple equilibria, or in which the only (Nash) equilibria are “mixed strategy” equilibria in which players randomize their behavior, which would hardly look like equilibria in a real world context. To get the well-behaved stable equilibria that economists like, you have to make a lot of assumptions that are very frequently not true in the real world.
In particular, any kind of increasing returns to scale, economies of scale, agglomeration efficiencies, decreasing marginal costs, productivity gains from team production, unexhausted gains from division of labor and specialization and trade, etc., break any neoclassical economic model based on stable, perfectly competitive equilibrium. For more on this topic, there are a lot of things you could read, such as Paul Krugman’s Development, Geography and Economic Theory, David Warsh’s Knowledge and the Wealth of Nations, or Xiaokai Yang’s Economics: Classical and Neoclassical Frameworks. Or my dissertation, Complexity, Competition and Growth. But be warned: it’s a very nerdy topic!
History shows that intellectuals can go down rabbit holes and get lost there for a very long time. Alchemy, the medieval science– or we might call it a pseudoscience, but there’s an element of luck in the science vs. pseudoscience distinction, since a bona fide investigator might pursue an unlucky hypothesis– that tried to transform base metals into gold by adding fire, has become the poster child of a diligent intellectual endeavor doomed to futility because of faulty premises. Neoclassical economics isn’t as bad as that. There’s actually a lot that can be learned about the operations of real, live capitalism using methods derived from a grand theory that is indefensibly biased in favor of seeing competitive markets and stable equilibria everywhere. But there’s a lot that cannot be learned. Economics has huge blind spots that arise from its original sin of modeling the world as a system of perfectly competitive market equilibria.
One of those blind spots is regional economics.
The classic Econ 101 supply-and-demand chart begs the question of where. Where are the suppliers? Where are the customers? Where do they trade? A “market,” in natural language, is first and foremost a physical place dedicated to buying and selling. Economics abstracts from that and envisions markets as an aspatial domain of buying and selling, which is a triumph of insight insofar as commercial competition and the incentives it gives rise to operate, in advanced capitalist conditions, over a wider range and often pretty independently of the physical “markets” of natural language. Yet capitalist competition isn’t so immaculately aspatial as the models. Drive times affect where you buy. Many customer-facing businesses obsessively value location, location, location. You’d think economists could embrace that, but “perfect competition” requires that commodities be perfect substitutes, and that breaks down as soon as you realize that widgets-at-place-A are never perfect substitutes for widgets-at-place-B.
Krugman’s The Spatial Economy surveys some efforts to apply economics to space in spite of this handicap, but stresses that it’s all been kind of a failure. Integration with mainstream economic theory remained elusive, and refinement and application of the various models that have been tried have been limited.
Mine may fare no better. But at least, now you know why I’m starting from scratch with some rough-and-ready graphs, rather than invoking some authoritative tradition for how to think about regional economics. There isn’t one. My graphs wouldn’t stand up to much critical scrutiny, as most theories coined by the human race haven’t, in the long run. I could refute them if I chose, and make partial defenses against my own refutations, through several iterations that I won’t explain here. But my considered judgment is that the graphs can deliver some valuable insight about the economic impact of telework before they get knocked down.
End of digression.
Starting Point: Labor Shed, Land Prices, and Site Selection
Returning to the main topic, we’re thinking about an industry, or firm, or plant– but we’ll call it an industry from now on– and about how it selects a location.
I should mention first of all that we’re dealing primarily with tradables industries here. If an industry’s products can’t be transported, it has to stay close to customers, and it can’t go hunting for an optimal production site. The distinction is not absolute. Even a business whose products can be costlessly shipped anywhere, such as software, may like to use face-to-face meetings with clients to build rapport and talk through requirements. And even very customer-facing businesses like K-12 schooling or residential home repair have distance modalities, such as boarding schools or traveling technicians. Grocery stores seem very localized, and yet people also buy some groceries from Amazon. Still, as a rough guide, think of this as a model of production site selection by tradables industries, which are supplemented by locally operating non-tradables industries that are not shown.
Some requirements are industry-specific. For the drydock business, you need to be by a coast. For a water park, you want lots of hot summer days. For an automotive cold-weather testing facility, winter needs to deliver some very cold days pretty reliably. Natural resource extraction industries– mining, fishing– have to be close to the resources and may not be able to be city-based at all. But other industries may need (depending on the state of transportation logistics) to be near the resources being processed, though nowadays less often than you’d think. But let’s skip that and turn to more general factors.
Two of the most consistent decision factors in site selection are:
Labor shed. How many (and what kind of) workers live in the vicinity?
Land prices/availability. Can land adequate for the production process be secured, and at what price?
Except for maybe a totally place-independent business like software, it’s hard to think of a business that doesn’t care about both of these things. If you have any need at all for physical presence work in your production process, you need to secure some kind of physical footprint where the production will occur, and you need to hire some local workers to come work there. We could get nuanced and add that affordable land might be indirectly necessary because the kind of workers you want to hire need backyards to be happy… but never mind. We’ll keep things as simple as we can. There will still be plenty of complexity to keep us busy. That said, it’s probably worth thinking of “land prices” more broadly than just the physical price of acreage to include the costs of overcoming local resistance based on the negative externalities that some production activities generate. One of the advantages of building factories in out-of-the-way places is that fewer people are around to object, and they might be more easily compensated, whether directly in cash for land or easements, or indirectly through their propensity, as voting citizens, to welcome taxes and jobs that big cities would disdain.
Not only are labor shed and land prices always relevant, but the desired direction is always the same: lower land prices and larger labor shed.
Here’s the problem: lower land prices and larger labor shed don’t generally go together. Where there are lots of people, land prices rise. And so industries face a trade-off. They can go where population density is high, and pay a lot for the land. Or they can go to where land is cheaper, but they may have trouble finding the workers they need. And this trade-off is shown in Figure 1, though it will take an astute reader to see how, until I’ve explained it:
Figure 1
In Figure 1, the two axes are (a) land prices and (b) labor shed. Up means more expensive land. Right means more people. The dots are cities, not real cities of course, but an arbitrary and illustrative distribution of cities in a two-dimensional space of land prices versus labor shed. Two simplifying assumptions here are that there’s just one value for land prices, and just one value for labor shed, per city. But that’s not too crippling of an oversimplification.
The subtlest part of Figure 1 is the curves. I’ll call these “isoprofit curves,” a term which has some precedent in economic theory although they’re usually represented in a different abstract space. “Iso-surplus curves” might be more precise (since what we really care about is the sum of the industry’s profit and the consumer surplus, which comprise the industry’s total value-add) but never mind. They are a form of “indifference curves,” where all the points in a given curve show values of land prices and labor shed between which an industry is indifferent. The point is that these curves show what an industry needs to operate, and what it prefers– what it seeks out, what makes it more profitable, what makes it expand– in terms of land prices and labor shed. The solid curve shows the minimum; the dotted curves show the preferences.
The trade-off between land prices and labor shed is shown in the upward slope of the isoprofit curves. To see why, imagine yourself in the place of an industry that is moving up and to the right along an isoprofit curve. With each increment of motion, land prices get a little bit more expensive, but the labor shed increases as well. The land prices raise costs, but easier staffing and more access to qualified specialists raises productivity. How does that affect the industry’s profits, or total surplus? There’s no net effect: that’s what the isoprofit curve means. Along an isoprofit curve, the gains and losses exactly offset each other.
By contrast, if you move down and to the right, profits rise. In that direction, you’re getting cheaper land and a more liquid workforce with more qualified specialists. And so each industry will pick the city that’s on the highest-profit (lowest/right-most in chart position) isoprofit curve. But that’s different for each industry, because while all industries want cheap land and abundant workers, the way industries value these things differs a lot. In Figure 1, the green industry chooses cheap land at the expense of a small labor shed; the purple industry chooses the biggest labor shed despite high land prices; and two other industries land somewhere in between and choose mid-sized cities.
There are two industries shown in Figure 1 that don’t exist at all. Then why show them? Because unrealized potentialities are important. Figure 1 shows why the industries don’t exist, namely, that the combinations of labor shed and land prices which could sustain them are not available anywhere. In the right environment, these industries would emerge. Another interpretation is that these industries may exist somewhere, but they can’t profitably expand. They may be able to survive on legacy capital where past investments are already sunk costs, but they can’t justify new capex investment.
How Telework Changes Industrial Geography
Now we’re ready to bring telework into the scene. What is its impact? Telework reduces the labor shed requirements of all industries. It does this because an important component of the workforce that industry requires, mental workers who use computers, can now work from anywhere. An industry choosing a site doesn’t need to consider whether it can recruit enough mental workers locally, because they can telecommute. It only needs enough labor shed to supply the physical presence workers, especially manual workers but maybe some kinds of face-to-face persuasion and management roles. Figure 2 shows the effects on site selection.
Figure 2. How telework changes industrial geography by reducing industries’ labor shed requirements
In Figure 2, relative to Figure 1, every industry’s isoprofit curves shift to the left, in the direction of a smaller labor shed, reflecting the fact that mental workers no longer need to be recruited locally but can be recruited and telework from anywhere. Labor shed still matters because manual and other physical presence workers still need to be local. But you need less labor shed to support any given industrial site selection, with the reduction in required labor shed varying depending on how much mental workforce vs. physical presence workforce an industry needs. The degree of decentralization varies by industry.
In Figure 2, the purple industry, which is presumably mental work intensive, exhibits a huge reduction in its labor shed requirements, and shifts from the largest metropolis to far smaller mid-sized cities, with a huge increase in profits. The light blue industry, by contrast, which is presumably more physical presence work intensive, shifts only modestly in the direction of smaller cities.
The prediction of the model here is not only or primarily that mental work will get more geographically decentralized. That was the work of the model in Part 1. This model assumes, rather than demonstrates, not so much that mental work will become more geographically decentralized as that mental workers will be pooled into what might be called a “virtual labor shed” that reaches as far as the internet itself. What the model predicts is, instead, that physical presence workers will become more decentralized, because a critical mass of complementary mental workers was part of what used to tie manual workers to expensive big cities, and now that constraint has been removed. Hybrid teams of teleworking professionals and local manual workers will enable all sorts of production activities to move to mid-sized cities and small towns, which would hitherto have been too sophisticated for them to support.
Hollowing Out vs. New Frontiers: The Fate of Big Cities in the Age of Telework
Both these models of the economic impact of telework, in Part 1 and here, would seem to have some bad news for big cities. And it’s probably true that expensive creative class metro areas will experience a certain degree of hollowing out in the next few years as a result of telework.
But there are upsides, too. In Part 1, we saw that while mental workers based in creative class metros will feel downward wage pressure due to competition from out-of-town teleworkers, organizations in creative class metros will employ a larger share of the national or global mental work labor force. And this growth will probably create promotion opportunities for some metro-based mental workers based on whatever advantages still accrue to physical presence and/or to seniority.
In this model, the upsides of telework for big cities are even more fascinating, because whole new industries can emerge thanks to the blessing of a globalized virtual labor shed of teleworkers to support whatever cutting-edge physical presence activities the big cities might dream up. Or, again, by another by not incompatible interpretation, stalled industries that have some foothold in the world economy but have been unable to expand could gain a sudden new growth momentum in the telework age.
Figure 2 shows two examples of new industries that emerge, or become investable and start to expand, because of telework.
First, think of the gold industry as an industry that didn’t exist at all pre-telework because no site in the world offered a large enough labor shed. This is counter-intuitive, but I would ask readers to wrap their heads around the reality that technological industries frequently depend on teams of one-in-a-million specialists. That’s not flamboyant rhetoric but a sober fact that a lot of organizations’ teams are selective enough that if you removed certain people from their jobs and substituted a randomly selected draftee instead, the odds that the random draftee could do the job as well as his predecessor are really one in a million or less. If you need a lot of such people, even New York City might not offer a big enough hiring pool. Even the New York City metro area, after all, has only ~6% of the US population, and ~0.25% of the world’s population. There’s a lot of room to grow the hiring pool from that.
When telework lets you recruit the best people from anywhere in the world, new things might become possible that are the stuff of engineers’ fantasies today. And so the gold industry emerges for the first time in the age of telework.
Second, think of the red industry as an industry that either didn’t exist at all pre-telework, or wasn’t investable at the margin, not because it’s too sophisticated for any metro to support, but because it’s uneconomic in the places where it can get the workers it needs to launch effective operations. The red industry is apparently somewhat land-intensive, and it can’t launch if land prices are too high. But it also requires some specialists and sizable teams, and it can’t make do with a very small labor shed like the green industry. And so, pre-telework, it either doesn’t exist or doesn’t expand, because all candidate sites either can’t provide the workers it needs, or can’t provide the land at a price it can afford.
But telework is a game changer for the red industry, because now it only needs to recruit physical presence workers locally, while mental work specialists can telecommute in. That enables the red industry to recruit the teams that it needs even in towns and smaller cities where the land is cheap enough to let them afford the physical footprint they need.
There’s a value in keeping things abstract, because while my expertise as an economist enables me to see that there almost surely must be some industries that will be enabled to emerge or expand by the great virtual labor shed of teleworkers that have now become available thanks to changing capitalist norms, I don’t have the domain expertise in marketing and engineering that would give me insight about the prospects for specific candidates to be the industries that are just close enough to competitiveness that telework will push them over the edge into runaway success. Still, abstraction can be stultifying, so I’ll put forward a few concrete candidates to play the role of the gold and red industries in Figure 2, just for illustrative purposes and to get the imaginative juices flowing. Keeping the disclaimer in mind, let me offer a glimpse of a few cutting-edge industries that the telework economy may be about to deliver for us.
First, drone delivery.
Wouldn’t it be nice if drones could deliver things to your door from stores and restaurants? You might still need to go to the store to get three bags of groceries, but if you just want a taco or a toothbrush, the drone will come and deliver. Save energy. Save your own time. Reduce traffic on the roads. Get it faster.
It would be complicated to implement, from both the technical and regulatory perspectives. Technical: how do you load the drone, control it, and make it deliver the drop? Regulatory: how do you get the right to use the airspace that way? The high overhead of setting it up would need to be justified on the back end by a large critical mass of concentrated demand. It will take a big city at first.
Telework helps in a lot of ways. Expert teams can be assembled from anywhere. And they can be nimble and opportunistic to play regulatory arbitrage. If New York says no, try Cleveland, or Detroit, or Fort Worth. You’ll need some local recruitment, but your teleworking masterminds of the drone delivery revolution can live anywhere.
Also, telework should free up a lot of commercial real estate, which a drone delivery startup could acquire cheaply, and repurpose as a warehouse to stock the drones. And telework gives cities a new incentive to compete on cutting-edge quality-of-life dimensions like drone delivery. Suppose your city is the first to get a drone delivery infrastructure up and running. Don’t you think quite a few footloose, high-income teleworkers would choose to settle in your futuristic mecca?
Second, mass customization with 3-D printability.
I have a lamp in my living room that we need to get rid of, because we lost a tiny knob, and now we can’t turn it on. The lamp works, and the knob would be easy to screw back on. This sort of thing happens all the time. Wouldn’t it be nice if manufacturers put a QR code where you could look up the model in CAD, and all the parts would be 3-D printable? Repair would be easier. And you could upgrade and customize more. Parts-seeking visitors would be great upsell opportunities for the manufacturers.
Something along these lines seems like another great business for creative class metro areas in the telework age to grow into. You’d need teams able to combine savvy online engagement with machine operator competence. Former office workers in creative class metros, looking for a career change, as out-of-town teleworkers undercut their wages, into something that better leverages their physical presence in capitalism’s innovative hotspots, might do great at bringing manufacturing to an Information Age maturity of digital everything.
Maybe Information Age manufacturing should optimize products for repairability by virtual reality headsets.
Third, telehealth factories.
Health care mostly operates as a non-tradable industry. Doctors need to be close to patients. Yet there is such a thing as medical tourism. Telehealth is an emerging field with obvious advantages– no need to drive to the doctor’s office– and obvious disadvantages– the doctor can’t see your body, draw blood, etc. A further disadvantage, at present, is that a local doctor you’ve visited before will have your chart, whereas the teledoc probably won’t. But this advantage could flip. Lots of other records live in the cloud. Why not medical records? And medicine is a knowledge-intensive field, where specialized teams and masterful use of AI, such as a small town doctor’s office probably can’t implement but big urban corporate hospitals maybe can, could add a lot of value.
So it’s possible to envision a medical future where people telecommute to big, urban, corporate hospitals for medical consultation and diagnostics, and sometimes fly to those same hospitals for scheduled preventive medicine and major procedures that aren’t super time-sensitive, while locally-provided medical services focus on physical presence components of the job, under supervision by urban hospitals, sometimes stabilizing the patient and escalating to the urban hospital when the patient needs more than they can give. Telehealth to service the hinterland could become big business for big cities as mass office work fades.
Fourth, concentrated solar power is an interesting future industry that mid-sized cities might thrive in.
Electricity supply is a semi-tradable industry, since while energy can move a long way over power lines pretty efficiently, it’s not as movable or storable as wheat or widgets. Coal, natural gas, nuclear, and hydropower supply most energy today, but all have serious environmental (and sometimes geopolitical) downsides, and many conscientious citizens and policymakers have long aspired to transition the economy to meeting its needs with renewables, of which the most promising are solar and wind. Meanwhile, AI is escalating demand for electricity to support its high-power computing. So we need more electricity generation.
Now, solar power most often takes the form of photovoltaic solar panels, which are being installed on rooftops at a rapid pace. But rooftop photovoltaics can’t supply any electricity at night. Up to a point, solar power helps the grid by coming online just when demand is highest, in the daytime. But solar self-suppliers have to rely on the grid, and its constant flow of non-renewable coal or nuclear or hydropower electricity, to keep the lights on at night.
That’s where concentrated solar power (CSP) comes in. CSP power plants configure mirrors to reflect light to towers and heat salts to extremely high temperatures where they can be used to power turbines. And when the sun sets, the salts stay hot for a while, and keep generating electricity. So CSP is a form of solar power that can stay on at night. There’s a big CSP power plant near Clark Mountain in the Mojave Desert in California, but in general, CSP hasn’t seen the growth that rooftop photovoltaics have. Why not?
I don’t know the economics ins-and-outs of that, but installing a CSP power plant sounds hard. Doubtless, you need electrical engineers and financial analysts, but you also need a lot of construction workers to build, and then there’s maintenance. At the same time, in a given area, sunlight is proportional to land, so this is land-intensive. It sounds like a business that could thrive in the outskirts of a mid-sized city, partially hollowed out by telework, with a decently large labor shed of physical presence workers, but affordable land, and of course, teleworking teams of engineers and financial analysis and project managers to design and monitor the technical and business plans.
Fifth, urban farming is a counter-intuitive but promising business for mid-sized cities to embrace.
People think of farming as land-intensive, but the reality is that land often isn’t the constraining factor on how much you can grow. To control the growing conditions for optimal plant health and crop quality can be a lot harder. It depends on the crop. But high-tech greenhouses can often out-perform sprawling fields, and that’s why the Netherlands is the world’s 2nd-largest agricultural exporter in spite of being quite small and densely populated. “High-tech greenhouses could be the future of agriculture,” writes Forbes magazine.
Again, this sounds like a good job for mid-sized cities suffering a mild hollowing out due to telework. They have reasonably affordable land, a sizable labor shed, and an infrastructural basis for efficient logistics. Let them build industrial greenhouses and grow food. Maybe they can grow some things that are too tricky to grow out in the countryside in the midst of all the natural weather.
Now, to reiterate the disclaimer, these examples aren’t meant as serious technoeconomic forecasts about what specific industries will thrive in the age of telework. To attempt such forecasts seriously would require more domain expertise in the candidate industries than I have. What my theory of the economic impact of telework indicates is that there will be some industries which will be empowered to emerge or expand because the virtual labor shed of teleworkers makes it possible to assemble strong hybrid virtual/physical teams to ideate and execute projects couldn’t be done until the great norm shift that virtualized the office, either anywhere, or in particular places. Telework will decentralize mature existing industries, and create economic pressures tending to hollow out big cities, contributing something to equality, efficiency, family formation (kids need backyards) and quality of life, especially for those of us lucky enough to be able to telework. But the big payoff of telework, in the long run, may be that by taking away the need for big cities to be organized around mass office work, we free them up to do fancy new kinds of physical presence work that push the technological frontier forward.