Eli Dourado

How do we move the needle on progress?

This summer, Patrick Collison and Tyler Cowen called for a new science of progress. Let’s study, they say, “the combination of economic, technological, scientific, cultural, and organizational advancement that has transformed our lives and raised standards of living over the past couple of centuries.” By understanding these factors better, we may be able to accelerate further progress in living standards.

A lot of people have responded to Patrick and Tyler’s article, covering different aspects of the challenge. In this post, let me bite off just the economic, forward-facing portion, recognizing that this is only a small part of the overall “progress studies” objective. Starting from where we are now, what would it take from an economic perspective to move the needle on growth?

To answer this question, we need to know where to focus. I think there are three criteria we can use to select focus areas.

The first criterion is economic significance. Obviously, progress in an area that is a huge component of GDP is one way to achieve progress overall. But not all of human welfare is captured by GDP or spending. Let’s also consider stuff we spend a lot of time on; after all, we have limited time budgets just as we have limited money budgets. We also need to consider undercompensated factors, stuff the full economic value of which isn’t reflected in the money or time we spend.

Not all technological advances are equally valuable. For example, imagine a technology emerged that revolutionized the lemonade industry. Lemonade simply isn’t that economically important. We don’t collectively spend a lot of time or money on it, and its value is not systematically excluded from our existing economic metrics, so we can ignore lemonade technology for this exercise.

A second criterion is low productivity over the last few decades. We aren’t going to move the needle on economic growth by focusing on industries like television manufacturing that have already experienced massive productivity gains. We need to address the laggards, those with low relative productivity gains. Fortunately, Mark Perry has already done a bit of homework for us by looking at BLS data to show areas where we see relative price increases versus decreases. These relative price increases are essentially relative productivity declines.

To move the needle on human welfare, then, we need to focus on items like hospital services that have gone up in relative price. Combining this criterion with the economic significance criterion helps us prioritize. For example, college textbooks have gotten a lot more expensive over the last 20 years, but one can hardly say that’s a cause of overall economic stagnation. If college textbooks suddenly received a massive productivity boost (“textbooks too cheap to meter?"), that would maybe save US consumers a few billion dollars per year, which is less than 0.1% of GDP. That isn’t important my purposes in this post.

A final criterion is the existence of a path to much higher productivity. I don’t think there are many parts of the economy that will permanently suffer from what Helland and Tabarrok call the Baumol effect, but childcare may be one of them for the foreseeable future. Robots could plausibly be designed to care for children, but it seems far-fetched to think we would deprive our children of human contact for, say, a whole work day. We should focus instead on industries where there are plausible large productivity gains to be had.

We are looking for economically significant sectors with lagging productivity and a plausible path to improved productivity in the future. So what meets these three criteria? I think a careful look at the economic landscape reveals four areas from which we could derive massive increases in human welfare: health, housing, energy, and transportation/logistics. If we want progress, we need a relentless push for innovation and dynamism in these sectors.


The health sector is the epitome of economic significance. According to the Centers for Medicare and Medicaid Services, health care spending in the US is around \$3.5 trillion per year, almost 18 percent of GDP. More than one in six dollars goes to the healthcare-industrial complex. As Perry’s chart shows, medical care and hospital services are two areas where relative productivity has stagnated.

More broad than healthcare is health itself. What we are really after is wellness—a long life without illness and maybe without even needing treatment. Think about the amount of time a seriously sick person spends worrying about her illness. Or about the cost of the disruption in one’s life of even a minor illness—when my kids get a cold and have to stay home from school, someone needs to stay home with them. These are undercompensated elements of health, factors not fully accounted for in the GDP spending data.

These economic impacts of health are pretty mainstream. But there is even more scope for progress when we bring in the possibility of life extension and rejuvenation technology. The monetary value of a quality-adjusted life-year is often estimated to be on the order of \$100,000. With more than 300 million people in the US, every QALY we add to average US lifespan adds around \$30 trillion in value. There’s a debate among scientists over whether we can extend lifespans much beyond 120 years, but even if we “only” raise today’s life expectancy from around 80 to around 110 (and keep people in good health during the extra years), that’s 30 QALYs per person, or \$900 trillion in possible welfare gains for people in the US who are alive today, and \$3 million for every new person born from now on. At around 4 million births per year in the US, that’s \$12 trillion/year or an extra 57% of GDP in welfare produced on an ongoing basis after the one-time \$900 trillion gain.

Given the scope of the potential gains to improved health, we are remarkably complacent about improving health productivity. Politically, nearly all of the discourse is centered on from whose pockets today’s 18 percent of GDP spent on healthcare will come. To make progress, we need to get much more aggressive on reducing costs and improving outcomes. There is a lot of room for progress.

Perhaps the first element of any plan to generate progress on health should be radical consumer empowerment. Today, we are largely held hostage to the health establishment, but at least some of today’s routine medical care is or could be made obsolete if we tried. When I get my annual physical, my doctor listens to my heartbeat. I also have a computer on my wrist all day every day monitoring my heart for an irregular rhythm, and capable of doing a quick and dirty ECG. Given that my primary doctor is not a cardiac specialist, maybe the watch actually provides better medical diagnosis.

In addition to heart rate data, my devices collect data on my exercise, weight, body fat, and more. What is the evolution of this trend? Imagine consumers empowered by more advanced medical devices with comprehensive real-time medical data and software. Such equipment would spot health issues early. You could eliminate much routine care and reliably address non-routine problems before they become acute and more expensive to treat.

What is more, let’s not discard all the data being produced by these devices or even by today’s medical records. What could happen if you pooled all this data and made it available for research? Today, we don’t share medical records because of medical privacy concerns. But what if we got over ourselves? What if we gave researchers access to every consumer’s comprehensive real-time medical data, including outcomes? How much faster could medical science progress?

We should also be doing more to disintermediate medical treatment. For example, most hormonal birth control medication is quite safe, but we still make women see their OB/GYNs every year to get a prescription. The OB/GYNs like this because it keeps women in the system and doctors' paychecks coming. Moving birth control over the counter would eliminate expense and hassle from the system for millions of women.

More generally, we could consider moving most prescription drugs over the counter. Philosopher Jessica Flanigan argues that giving patients the right to choose their own drugs is morally obligatory. But aside from being the right thing to do from a moral standpoint, it could also lead to much lower medical costs, as doctors would have to compete with nurses, pharmacists, and yes, the Internet and apps for medical advice.

To disintermediate further, 3D chemical printers could make custom pharmaceuticals in exactly the right dose based on consumer-owned biomarker data. If we are sharing medical records in a public database available to researchers, the response to these printed drugs could be studied in detail to determine how safe and effective they are with statistically-constructed control data, reducing the need for costly clinical trials.

We should also seek progress on public health. The evidence is mounting that air pollution, especially from diesel vehicles, is worse than we thought. The electrification of trucks and school buses and construction equipment could therefore pay large dividends in terms of a healthier population. And there continues to be well-founded concern that federal nutrition advice has been motivated by agricultural interests rather than or in addition to health. How many health problems could be averted with more nutritious diets?

Funding for medical research needs to be improved. Top medical researchers spend up to 50 percent of their time writing grant proposals. This fact has led some to propose distributing grant money by lottery. It is not a crazy idea. If we could divert the time spent applying for grants to time spent doing research, we could double research output.

Lastly, given the overwhelming value associated with life extension, more of our research dollars should be focused on aging. In addition to diverting existing medical research funds to aging, we might consider spending less on heroic treatments for patients who aren’t likely to survive in order to spend those medical resources on aging research. Cut Medicare dollars from today’s elderly in order to make the future elderly live much longer and healthier lives. The payoff in terms of human welfare to cracking aging could be unimaginably high.


Another big ticket item with lagging productivity is housing. Between rent, imputed housing services, and new construction, we spend about 16 percent of GDP on housing, just below healthcare’s 18 percent. Most of the variation in annual spending is due to construction, which is highly sensitive to the business cycle.

Aggregate statistics understate the scale of the problem, however, as low housing productivity is concentrated in a handful of otherwise-productive cities in the US. For example, housing prices have roughly doubled in San Francisco (median home price \$1.35M) in the last decade, while in the US as a whole (median home price \$229k) they have only gone up 40 percent.

Aside from high expenditure in a pure accounting sense, high housing prices also distort all kinds of other markets. They increase commute times. They make people risk averse. They affect job decisions. They delay childbearing age.

It’s useful to divide housing prices into two components: land and structures. The division in value between land and structures varies according to market conditions of course. There are some key economic differentiators: for example, land doesn’t usually depreciate while structures do, and supply elasticity for land is zero, while it is positive for structures. This latter factor means that in areas with lots of housing demand, it is the land component of housing that is going to be the bigger factor in driving up housing prices.

This elasticity analysis matches reality. Consider this property, currently listed for sale in Palo Alto. The asking price is \$1.65M. The house, built in 1931, has 636 square feet, and the lot is 1,942 square feet, or 0.04 acres. As hard as it is to fathom that a 0.04-acre lot could be worth well over a million dollars, it’s certainly clear that a 636-square-foot structure could not be driving the price here, as the replacement cost of building another 636-square-foot dwelling.

Since housing costs in the most problematic markets are driven by land scarcity, there is an obvious solution. We can stack dwellings on top of each other so that they use less land. From a technological standpoint, we know how to do this. Many cities around the world have engaged in such dwelling-stacking.

The cost of housing in the most expensive cities is an entirely self-imposed problem. If we built a large number of apartment buildings in Palo Alto we could enable many more people to live in an area that is so otherwise productive that people today are willing to pay 7 figures for a cottage. In addition, by soaking up demand for housing into apartment buildings, we would lower the price of land so that the cottage wouldn’t be as expensive in the first place.

The free market would certainly build higher-density housing in high-rent cities. The reason it isn’t happening is zoning, permitting, parking requirements, and other land-use regulations. Not everybody wants high-rises in their city, preferring the charm of the million-dollar cottages or of whatever else happened to be there first. Challenging this NIMBYism is an enormous opportunity to generate economic progress, and I am pleased to say we have seen some limited success lately. Much more reform is needed.

Because of the inertia imposed by NIMBYism, it may be politically impossible to put residential high-rises in every expensive housing market. It is worth, therefore, figuring out how to the package greater density with the charm that NIMBYs seem to want in order to make density less of an uphill battle. It’s certainly possible to combine density and charm. Paris’s Haussmann buildings are up to six stories tall, and contribute to the city’s quaintness while affording greater density than the buildings they replaced in the mid-1800s. NIMBYs would still oppose new, more palatable options for density, but perhaps they would not oppose them quite so hard.

City governments should liberalize land use regulations like zoning and permitting. But they should also incentivize density by replacing today’s property taxes with land value taxes. Instead of taxing the value of the land and the structure as we do today, tax only the land. Change the rate of the tax to ensure that total revenues stay the same. On average, tax bills stay the same, but at the margin, the incentive to improve homes and neighborhoods increases, as does the incentive for the city to provide higher amenity value and more efficient services.

Although construction costs are not the major culprit behind today’s high housing costs, it’s worth thinking about how they could be reduced through technology. Through better use of robotics, new materials like self-healing concrete, more advanced production techniques like 3D printing and greater use of pre-fabrication, and drones for job site inspection, it’s possible we could get the cost of structures down, too.

Pre-fabrication is particularly promising for tall buildings, leading to cheaper buildings that can go up dramatically faster. As José Luis Ricón points out, the Chinese firm Broad Sustainable Building is capable of “insane building rates” through the use of factory-produced modules. This video shows them putting up a 57-story mixed-use building in 19 days, for an average of 3 stories per day.

There are more exotic building ideas that remain fictional. I for one have always wanted to see an arcology after playing with them in SimCity 2000. Neal Stephenson’s short story Atmosphaera Incognita features a 20-km-high tower with active stabilization using airfoils (and even turbofans in one section).

Although I will discuss transportation and commuting in a later section, it’s worth noting here that remote work is a popular solution to the housing crisis. By living in cheaper parts of the country, remote employees can pocket the gains from working in a big city while incurring lower living expenses. Remote work has clear benefits when housing markets are dysfunctional, but I am not so bullish on it in general. We are still very interpersonal beings, with finely honed microexpression detection skills. It remains impossible to fully replace in-person interactions. Maybe with computer-brain interfaces that will change, but for now I think we should focus on getting housing markets working again.


The Energy Information Administration tallies energy expenditures at only 5.8 percent of GDP, a much lower share than health or housing, which we examined above. But unlike healthcare or housing, energy is an intermediate good, and it features high complementarity with other production goods.

Another way to look at the significance of energy is through energy intensity of the economy. In the US, we use about 1.5 kWh to produce \$1 of GDP. Looking only at electricity, the average price across residential, commercial, and industrial uses in May 2019 was 10.43¢/kWh. A gallon of gasoline embeds 38.7 kWh of energy content, so at today’s price of \$2.66/gallon, that’s 6.87¢/kWh. Let’s average the electricity and gasoline numbers and say a kWh costs about 8¢ on average. That means energy counts for about 12¢ out of every dollar of GDP produced in the US, or 12 percent of GDP.

An undercompensated factor in this analysis is pollution. Fossil fuels made up 92.7 percent of the US’s primary energy mix in 2016. Burning fossil fuels produces both particulate matter that harms local public health (and is especially crippling in a number of Asian cities) and carbon dioxide emissions that warm the planet. Globally, although energy prices have fallen modestly over time, we are not yet making much progress on moving away from fossil fuels.

The goal for energy progress is actually pretty easy to define: electrify all the things and produce unlimited clean energy too cheap to meter. In such a world, people would use significantly more energy than they do today. In the past 50 years, energy consumption per capita has gone up, which means electrons are doing more stuff for us than ever before. And when cheap electricity exists, we should expect the energy intensity of the economy would increase rather than decrease. For example, thanks to abundant geothermal and hydroelectric power, Iceland has some of the cheapest and cleanest energy on the planet, and they use it in energy-intensive industries like aluminum smelting. Their economy is three times more energy-intensive than the US’s based on kWh/\$ GDP.

Achieving our energy goal requires electrifying ground transportation. The US uses around 9 million barrels of gasoline per day. I am fairly convinced a transition to electric cars will happen eventually, as they are superior to gasoline cars in terms of maintenance costs, operating costs, and performance characteristics (especially acceleration and torque), and they are still improving rapidly. But to get there faster we need to overcome the inertia associated with inadequate charging infrastructure and the scientific and engineering challenges of improving battery performance (energy density, charging time, durability, stability). And of course, lower electricity costs can only help as they would further reduce electric vehicle operating costs relative to gasoline or diesel vehicles.

Progress means producing a lot more clean energy, enough to equal current electricity usage, replace 9 million barrels a day of gasoline energy content, and grow our energy budget. Most sustainable energy investment is going into solar and wind development. Even coupled with batteries for addressing intermittency, these sources are increasingly providing electricity at competitive rates. Ramez Naam notes that new solar and wind (combined with storage) plants are getting to be cheaper than operating existing fossil fuel plants.

That’s great, but it might not be enough. To produce clean electricity at the scale we’ll need, we will also need to look at non-solar and non-wind sources of clean energy. The elephant in the room, of course, is nuclear energy. In the US, nuclear power provides less than 20 percent of our electricity. France, which invested heavily in nuclear power plants, gets more than 70 percent of its electricity from nuclear power. Nuclear power has several advantages. As Hannah Ritchie demonstrates, relative to fossil fuel sources, it is dramatically safer, despite assuming conservatively that the nearly discredited linear-no-threshold model of radiation exposure is true. The nuclear fear-mongers are wrong.

Relative to solar and wind power, nuclear doesn’t have the intermittency problem. It can consistently provide base load energy to the grid. Nuclear power plants are also available at much higher scale than solar power plants. The largest solar plant in the world, Tengger Desert Solar Park in China, has only a 1.5 GW power output. That’s the capacity of a smallish nuclear plant; larger nuclear plants produce 5 GW or more. (Wind power plants are usually below 1.5 GW, but the world’s largest wind farm is Gansu in—where else?—China, with a current output of 8 GW and a planned output of 20 GW. China also has the world’s largest hydroelectric plant by far, Three Gorges Dam, which produces 22.5 GW).

The US consumes 24 PWh of primary energy (i.e., including gasoline) per year. With 8,760 hours in a year, that is 2.7 TW of electricity at a constant load. If we restrict ourselves to plants that produce 1.5 GW, the size of the world’s largest solar farm, we would need 1,800 of them to fully decarbonize. If we combine that with rooftop solar, the total number of solar farms needed would be lower. There are some advantage to generating power on rooftops, where it doesn’t have to be transmitted to retail customers. But solar power is still much cheaper when it is produced at scale than when it is produced on roofs, so rooftop solar only gets you so far.

Nuclear power has the advantage of being safer and cleaner than fossil fuels, while being non-intermittent like solar and wind, and able to deliver energy at a significantly greater scale than solar and all but one (Chinese) wind farm in the world.

New kinds of reactors would only sweeten the nuclear deal. Thorium reactors would be safer and produce less waste while producing a lower risk of weaponization. Nuclear fusion could be an even safer and cheaper technology, if we can get it to work. I cannot evaluate the merits of these claims, but smart people say we are likely to achieve energy-positive fusion reactions in the next five years, with commercialization in the next 15. Lockheed Martin in particular seems to be getting close, and there are also today a number of fusion startups, something that was unheard of a decade ago.

Aside from nuclear, one underrated energy source is geothermal. 62 percent of Iceland’s primary energy comes from geothermal, including both electricity generation and home heating. Another 19 percent comes from hydroelectric. I was in Iceland in June, and I can tell you they have abundant wind resources as well. As Iceland has abundant clean energy, it is a great place to locate energy intensive industries like aluminum smelting.

Iceland is fortunate to be sitting on a volcanic hotspot, which makes it easy to tap into geothermal resources, but lots of other countries have untapped geothermal reserves. In Yellowstone alone, according to a NASA study mainly focused on supervolcano risk, there is enough geothermal energy to power the entire continental US. Unfortunately, §28(c) of the Geothermal Steam Act of 1970 essentially bans geothermal energy leases in national parks, ostensibly to protect their natural beauty. This is bananas. Iceland too has lots of natural beauty, almost none of which is marred by their geothermal investments. If in a misguided fit of 1970s regulatory zeal they had banned geothermal energy in their beautiful landscapes, they would not be the clean energy powerhouse they are today.

Another underrated aspect of geothermal energy is that it doesn’t have to use volcanic heat. Dandelion is a home heating and cooling company that graduated from Alphabet’s X lab. The company’s key insight is that ambient temperature in the ground 10 feet below the frost line is a stable 55º F, regardless of the time of year or outside temperature. Instead of using a heat pump to exchange energy with outdoor air, use one to exchange it with the ground, which is relatively cool in the summer and warm in the winter. The system is cost-effective, saving homeowners on average \$2,250/year, according to the company. We need more of this kind of entrepreneurial innovation to achieve our goal of clean energy too cheap to meter.

Transportation, logistics, and infrastructure

Transportation and especially aviation are near and dear to my heart. There is no single line in the National Income and Product Accounts that captures the value of transportation to the economy. If you add up some of the relevant transportation components, you get to a number around 9 percent of GDP.

There are a number of reasons why the 9 percent figure understates transportation’s importance. The first one is mathematical, a wonderful insight from Nick Szabo. Starting from a central point, the number of places that you can economically travel to is an inverse square function of the transportation cost. We can consider the cost to include both time and money costs. If you could travel in half the time and for half the money, the number of places you can access given your time and money budgets goes up by a factor of four. Think $A = \pi r^2$, where $A$ is your travel opportunity set and $r$ is how far you can afford to go in any one direction. What Szabo points out is that the value of transportation networks, like other networks, increases according to Metcalfe’s Law, proportionally to the square of the number of nodes in the network. The social gains to lower travel costs (or higher speeds), therefore, are doubly quadratic. The social welfare from transportation technology is an inverse fourth-power function of transportation costs. A halving of transportation costs, therefore, raises the value of the transportation network 16-fold.

Additionally, there is a strong undercompensated component to transportation. Transportation takes up a large amount of time, which is particularly relevant when we consider transportation of human cargo. The American Community Survey says workers spend on average 26.4 minutes commuting in each direction. That’s 52 minutes per day, 5 days per week, or more than 5 percent of your waking hours on workdays. In a survey, Kahneman and Krueger find that commuting is one of our most unpleasant experiences. Happiness research suggests that although you can adapt to most negative shocks in your life, you can’t adapt to a long commute. Long commuters are more likely to divorce and less likely to exercise. Adding 20 minutes to your commute leads to as much job dissatisfaction as a 19-percent pay cut.

The time and money cost of travel also has an unseen effect in the form of trips not taken. Think about the job you don’t take because it’s just out of commuting range. Or the friendships that wither when one person moves away. Or the places that you don’t go on vacation because the trip is too long. We have not seen a meaningful speedup in car or plane transportation in decades, so the unseen effect of trips not taken is not usually top of mind, but it is a big one. Research shows that high-speed rail in China leads to more research collaboration across cities, that international travel facilitates international business relationships, and that the switch from sailboats to steamships caused the first wave of globalization in the 1800s.

Because they are so interlinked, it’s worth discussing logistics and infrastructure in the same section as transportation. Logistics services (food or grocery delivery, packages from Amazon) are a substitute for traveling somewhere yourself, and of course they often include a cargo transportation component. According to the American Time Use Survey, we spend on average 1.78 hours per day (11 percent of our waking hours!) on household activities, many of which could be replaced by logistics services if the price would only fall. Cheap laundry pickup and drop off, instant availability of high-quality food options, and faster package delivery would reduce the time we spend on chores and errands.

And what can we say about infrastructure stagnation? We have it. Americans are spending \$2.2 billion per kilometer to expand New York’s subway system, while Paris expands theirs for \$230 million per kilometer. The United States seems completely incapable of deploying high-speed rail, keeping the DC Metro from catching fire, or making money on Amtrak. Projects spend years in environmental review.

Transportation, logistics, and infrastructure in the US could be enormously improved. Progress would naturally mean going faster, bringing back supersonic travel and then going further to hypersonic speeds.

China is killing us in high-speed rail, having added the lion’s share of its 18,000 miles of high-speed track in the last decade. Progress would mean replacing our current rail lines with the faster trains available in China, or even leapfrogging that technology and building out a Hyperloop network.

For intra-city travel, we need fleets of autonomous cars and urban air mobility. Robotaxis would save us from the inefficiencies of car ownership—your car is a wasteful capital asset, sitting idle 95% of the time. With all the investment going into autonomy, we will actually probably get them some time soon. Even more exciting than robotaxis is urban air mobility technology, which would provide the first major speed up in getting across town since the introduction of the car. UAM could provide a benefit even if only a few people use it. Ground traffic is highly nonlinear with the number of cars on the road, so taking even 10 percent of the passengers and putting them in the air could dramatically reduce ground traffic.

Lastly, we need progress in cargo services. Drone delivery seems ready to go once FAA allows it. Without a waiver, current regulations still restrict drone pilots to operations within visual line of sight, which makes drone delivery uneconomical. If you have to have a human close enough to see the drone, why not have the human just deliver the item? When fully allowed, drones may be great for short cargo hops, but we also need new long-haul cargo technology. Unmanned airships could come back for this purpose, beating today’s cargo ships on speed, and not being limited to coastal ports. China is investing heavily in airship technology.

Concluding observations

My aim in this post has been merely to think through what it would take, economically speaking, to generate dramatic progress. If we had large productivity gains in health, housing, energy, and transportation, I think we could safely declare that progress had arrived. Human welfare would be significantly higher.

Progress studies, as Patrick and Tyler have defined it, goes way beyond what I’ve done in this post and into the study of cultural and historical factors that lead to progress. I don’t claim to be an expert on these factors, but we can make a few observations that are useful to this expansive version of progress studies.

First, health, housing, energy, and transportation are all highly regulated sectors of the economy. It is simply impossible to take significant steps toward progress without addressing regulatory obstacles. We are not going to get progress by tweaking the tax code or by business as usual. In some cases, regulators will need to allow more safety risks to be taken.

Second, the US tech industry as represented by Silicon Valley venture capitalists dramatically underinvests in highly regulated sectors. A lot of the disinterest may actually be due to the limited partners, not the general partners. The LPs, perhaps pension funds or university endowments, want exposure to “tech stuff,” but not really to hard technology with potential regulatory barriers. This leads many VCs to focus on pure software companies, or on companies that simply apply better software to do something that is otherwise clearly legal and straightforward. We need new companies operating in regulated spaces so that regulators feel a sense of urgency to reform.

A third observation is that China is killing us in all the areas where we are stagnating. China is still a middle-income country, yet they already have longer healthy lifespans than Americans do. They are killing us on housing construction, putting up 57-story mixed-use buildings in 19 days. The world’s largest hydro plant, solar farm, and wind farm are all in China, and they are investing more in renewable energy technologies than Europe and the United States combined. China has newer and better infrastructure than the US, with gleaming modern airports and thousands of miles of high-speed rail. Of the areas I’ve discussed, the main one where China clearly lags is air pollution, which is still intolerable.

Fourth, the progress described in this post represents an agenda that should appeal to the entire political spectrum. It is difficult to see how we could achieve massive productivity gains in these four key areas and anyone could be left behind. It doesn’t matter whether you are rich or poor, urban or rural. Everyone’s living standards would rise if we could achieve the milestones I have set out.

I am grateful to Patrick and Tyler for being strong advocates of progress. Their article got a lot of people to think more seriously and systematically about how to overcome our present stagnation. I hope this post can also contribute to their burgeoning movement.