Innovation and Inequality
The inequality debate has long been a prominent feature of our politics, and many other social and cultural forces are now interpreted in light of it. We have certainly come to understand the progress of technology in that light: Many analysts have argued in recent years that technological advances tend to intensify economic differences, because technology disproportionately increases the incomes of those most able to take advantage of it. Thus, they say, technological innovation is a driver of inequality.
But this analysis is incomplete and misguided. The most important effect of technology on the American economy is not the minting of more millionaires, but the creation of a continual stream of new ideas and products that are quickly enjoyed by everyone. When we surf the web or use mobile computing devices, we are enjoying the value of many hundreds of ideas and innovations. These ideas — not the scarce material resources required to make a tablet casing or to carry packets of information from place to place — constitute most of the value added to our lives by the gadgets in our hands.
Economic value is thus increasingly created not by material things but by the information that arranges the material. And information can be shared equally in ways that material goods simply cannot. As Thomas Jefferson famously put it, "He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me." In the information age, we enjoy ever-greater access to a common pool of ideas that generates more value and consumption for all, substantially tempering the effect of technology's differential boost to incomes.
This shift has important implications for our understanding of inequality. Material resources are zero-sum: Five different people cannot all own the same barrel of crude oil or the same hectare of land. The deployment of these resources is similarly limited. One barrel of crude cannot simultaneously heat thousands of homes and power millions of cars on different continents. Ideas and information, however, are radically different. An infinite number of people can access the same information and then use it in untold numbers of different applications all at the same time. And the accelerating pace of technological change drives down the costs of these information-based products — so far down indeed that some products (like many computer applications and a growing portion of online-education offerings) quickly become essentially free.
The fact that ideas can be shared equally in a way that material goods cannot, and the fact that technology increasingly makes ideas the drivers of our society and economy, means that our circumstances are more equal than the conventional income measures would suggest. And to the extent that inequality should be addressed through public policy, acknowledging these facts will help us craft more promising solutions — policies designed to increase innovation and improve education, rather than to simply transfer resources from one group to another. We would thus be better off thinking not of technology in light of inequality, but of inequality in light of technology.
COMPUTING POWER
It is hard to deny that technological change has accelerated over the broad sweep of history, speeding up the pace of improvements in personal well-being along with it. Anthropologists suggest that humans were hunter-gatherers for 100,000 years. Beginning about 10,000 years ago, human beings transitioned to an agricultural society, which then lasted until a few centuries ago. Through the Industrial Revolution, from 1750 until 1950, the Western world transformed itself into a society based on manufacturing. Since 1950, we have been rapidly entering the information age. As economist Robin Hanson has noted, each of these completed epochs has been marked by a transition to substantially higher rates of economic growth. Each epoch has also become progressively shorter, suggesting that even faster transitions to a more prosperous future may be in the offing.
Technologist Ray Kurzweil has dubbed this phenomenon of ever-faster technological and social change "the law of accelerating returns." He has also suggested that the period between extraordinary advances — huge spikes in knowledge that come from great scientific discoveries and technological inventions — has decreased. Both outside the great epochs of recorded human history and within them, the story of acceleration is much the same.
The acceleration of technological advancement is perhaps best illustrated by the exponential growth of computational power, and by the ways in which this growth is transforming journalism, the military, energy, medicine, and many other sectors. The term Moore's Law — named after Gordon Moore, one of the founders of the computer-chip maker Intel — refers to the observation that the number of transistors that can fit onto a computer chip doubles every 18 months to two years. This theory, offered almost a half-century ago, has correctly predicted the exponential growth since in every aspect of the digital world, from computational calculation power to computer memory capacity.
Over the years, many people have predicted the end of this trend, but the growth has persisted nonetheless. Moreover, Kurzweil has shown that such exponential increases in computational power took place even before the birth of the modern computer. These increases occurred in the development of earlier mechanical computing devices, like the mechanical adding machine, but were too small in absolute terms to draw much attention. Kurzweil therefore expects that new technologies — such as optical or quantum computing — will extend the exponential growth curve even when silicon has given out.
It is difficult to overstate the power of exponential growth. As economist Robert Lucas observed, once a person starts thinking about exponential growth, it is hard to think about anything else. To understand why, consider that the computational power in a cell phone today is a thousand times greater, and a million times less expensive, than all the computing power housed at the Massachusetts Institute of Technology in 1965.
Changes in software and connectivity also multiply overall increases in computation power. A recent study examining the improvements in one benchmark computer task over a 15-year period showed that computer speed had improved by a factor of roughly 43 million. Of that amount, a factor of about 1,000 was attributable to improvements in hardware capacity during that time, but a factor of about 43,000 results from better programming algorithms. Gains in connectivity, too, may increase the effective power of computation. Computers (and, through them, people) are of course increasingly connected through the internet. And one result of this increased interconnection is that more people in nations like India and China — people who were previously removed from the core areas of innovation in the West — can now go online and contribute to the growth of science and technology. A major benefit of this increased collaboration has been exponential growth in the volume of scientific knowledge, which, in turn, creates the platform for further, faster innovation.
While Moore's Law is the best known expression of the exponential growth in computing, the social effects of this growth are better captured by another observation known as Bell's Law. Bell's Law posits that, roughly every decade, hundredfold increases in computational ability create an entirely new class of computers. Each class of machines becomes dramatically smaller in size, but has as much functionality as the class it displaces (or more). The improved class then becomes a new nexus through which people exploit computational power in everyday life.
Most Americans would recognize Bell's Law at work in the experience of the past half-century. In the 1960s, mainframes were the primary locus of computing. In the 1970s, so-called "minicomputers" began their run, only to be succeeded in the 1980s by personal computers. In the 1990s, PCs were joined by laptops, and in the 2000s "smart phones" appeared. It increasingly appears that, by the 2020s, computers will be small enough to be routinely introduced as medical devices into the body, enabling ever-closer interaction between humans and computational machines.
Increases in processing power not only shrink the size of devices but also bring more and more social spheres into the domain of computation. Journalism is increasingly an online enterprise that is ever-faster paced and more open to entry. Higher education is also migrating to the internet, as demonstrated by (for instance) Stanford's recent online course on artificial intelligence, which attracted 160,000 students from all over the world. Similar effects are visible in less peaceful pursuits: Drones are becoming the airpower of choice, and it seems quite possible that the last human fighter pilot has already been born.
Other areas of our lives are in the first stages of being shaped by the computational and information revolutions. Health care is certainly one of them: Genomics (the digital analysis of the gene) promises new kinds of research into the causes of disease and human behavior, with the possibility of significantly more personalized medicine. The potential of this research seems likely to grow as the cost of sequencing genes exponentially declines. And in July 2012, scientists published findings from the first computer simulation of an entire bacterium, a development that will help in generating and testing new cures for diseases.
The energy sector, too, is taking advantage of the computational revolution with smart meters and grids designed to make electricity consumption more efficient. Nanotechnology also progresses more rapidly as computers make design on a molecular scale ever easier. In short, the computer is a revolutionary machine like no other — because almost every device, and nearly every human function, can be improved by taking advantage of the various kinds of additional intelligence and data manipulation that greater computational power can provide. Moreover, these advances are essentially advances in the structuring and distribution of information and thus can be readily shared.
ACCELERATION OR STAGNATION?
Despite these extraordinary advances, some analysts, perhaps most notably George Mason University economist Tyler Cowen, have argued in recent years that technology is stagnating and can no longer be expected to drive economic growth at rates we have come to expect. Much of Cowen's argument rests on the notion that economic deceleration is an indicator of technological deceleration. The median income in the United States has grown much less rapidly since the 1970s than it did in the decades before, and skeptics of technological advancement believe this trend shows that technology is no longer the engine of prosperity that it was in the mid-20th century. In his recent book The Great Stagnation, Cowen writes that "there are periodic technological plateaus, and right now we are sitting on top of one[.]"
This claim relies, of course, on an argument frequently heard in the inequality debate — that the middle class is not doing as well as it once did, and that income gaps are growing. Stagnating median incomes are in fact the central concern of inequality critics. But median income growth in the United States actually provides a poor measure of technological acceleration, for three reasons that have a lot to do with the nature of technology itself.
First, the most common measures of median-income growth do not reflect the growth in fringe benefits. In recent years, total compensation for American workers has risen faster than wages alone, chiefly because of large increases in corporate outlays for health insurance. This increasingly generous insurance coverage, in turn, provides workers with greater access to America's technologically sophisticated health-care system. Cowen, for his part, argues that these increases have not produced real gains in well-being: Life expectancy in the United States is no better than it is in other industrialized nations that spend far less on health insurance. But Cowen's argument fails to take account of various social and demographic factors less common in other developed nations that depress life expectancy in the U.S. For example, as Harvard economist N. Gregory Mankiw has noted, Americans have higher accidental-death rates and are more obese than people in other parts of the world — both because of lifestyle choices and our economy's ability to serve up cheap, high-calorie food. Moreover, survival rates from some cancers and other complicated diseases are higher here, suggesting that Americans benefit from cutting-edge health-care innovations before those innovations become available to the rest of the world. Without these advances in medicine — advances made and paid for in the United States by our higher overall health-care costs — both Americans and people throughout the world would be worse off.
The life-expectancy figures Cowen cites also do not account for the quality-of-life improvements offered by contemporary medicine — such as replacement joints — that allow an active life far into old age. So important are these improvements that, in a 2011 television interview, former Harvard University president Lawrence Summers argued that it was unclear whether people would prefer to have the lower overall standard of living of 1951 but with today's health care, or today's higher standard of living with the lower health-care quality of 1951. Summers's observation suggests that improvements in health care alone have an enormous effect on personal well-being. And as these technology-driven quality-of-life improvements are not measured by the formal metrics used to calculate standards of living, they are easily overlooked in our debates about income growth and inequality.
Second, median income may be the wrong metric to use in the first place. The growth in average (mean) per-capita income would provide a better proxy for growth, because it is unaffected by the distribution of the fruits of technological acceleration. Over the past few decades, mean incomes have increased far more than median incomes, because higher-income individuals have done far better and improved their circumstances far more than the vast majority of Americans. From 1967 to 2008, real median family income and median household income rose approximately 38% and 25%, respectively. But real GDP per capita rose approximately 120% in the same period. Of course, income has taken a hit since the Great Recession of 2008, but historically the economy has rebounded from such shocks.
This increase in domestic income inequality in fact serves more as an indicator of technological acceleration than as evidence against it. By and large, people with more education are better at using technology to improve their productivity than are people who lack education. Moreover, because technological improvement allows superstars within their fields to serve ever-bigger markets, it has boosted the incomes of the most talented and best educated Americans at the expense of the less talented and educated. Income gaps would ordinarily exist between these two different groups, but technology seems to have widened them.
Third and perhaps most important, technological acceleration is distorting measures of economic growth by transforming what one dollar can buy. To assess inflation, economists measure the prices of items in the same "basket" of goods every year. But economists have suggested that these inflation measures are overstated, in part because, while they examine the same goods, they fail to account for improvements in the quality of those goods. A basket might include, for example, a new car. But comparing the price of a new car in 1974 to the price of a new car today is very difficult because, thanks to technology, today's cars are much safer, more comfortable, and capable of vastly more functions than the cars of decades past. "Basket" measures also fail to account fully (or to account quickly enough) for the introduction of major new products into the market. And if technological acceleration is introducing new products and improving their quality faster than ever before, the Consumer Price Index is likely to become more and more inaccurate, leading to the perpetual understatement of economic growth.
These claims are not just theoretical. It took more than 15 years to add cell phones to the price index's basket of goods; that failure resulted in the substantial overstatement of inflation in the telecommunications sector. Computers are getting better all the time, and the CPI tries to account for these changes, but it does so in ways that fail to capture the quality-of-life improvements that such products generate. The best example may be the internet: The unprecedented interconnectivity of the internet is a huge boon to people in their daily lives, both at work and at play, yet it costs very little. This makes it very difficult to compare our material situation today to our circumstances prior to the internet's development. In some cases, it is impossible: Many people, if asked to imagine the experiences of workers doing their jobs before the internet, would be incapable of answering. Without the internet, many of their jobs — as information-technology analysts, social-networking consultants, bloggers, e-commerce entrepreneurs, and so forth — would not exist.
Internet access, moreover, is a service that people of different tastes and capacities can enjoy in different ways. Some surf the web to become more informed about key political and social issues. Others use it to search for products that serve their needs better than the products available in local stores. Still others go online to meet people or to play traditional games like chess or new multiplayer adventure games, often for free. To be sure, some of the claims made about the value of the internet are overblown. But the internet and other inventions — like digital-music devices, mobile phones, and mobile computers — continually improve our quality of life in ways that have not been well captured by the CPI.
Finally, any sense of stagnation may well be an illusion. As discussed above, computation and digitization are just beginning to spread out across some key sectors like medicine, energy, and robotics. As a result, the exponential changes occurring in these technologies are not yet transforming everyday life, although they will do so quickly. Here, it is useful to remember the legendary example of exponential growth in which the king promised to reward the inventor of chess by granting his request for rice: one grain for the first square of the chess board, two for the second, four for the third, continuously doubled for each of the 64 squares. When one does the math, the early doublings do not amount to much. But as the doubling continues, the growth accelerates and the numbers quickly become mind-boggling, reaching 9,223,372,036,854,775,808 on the 64th square. When it comes to growth in a range of technologies — from biotechnology to nanotechnology — we appear to be on just the fifth or sixth square of the board.
CONSUMPTION AND EQUALITY
Much of the evidence regarding technological acceleration also shows that, even if technology widens income gaps between the rich and the poor, it does not necessarily widen gaps in their well-being. Innovations in health care quickly become available to the broader public. Many software programs swiftly become available for free. A billionaire and a member of the middle class have relatively equal portals to the wonders of the internet — certainly far more equal access than the rich and the rest of society would have had to the material goods that defined wealth in centuries past. These facts suggest that inequality is not widening nearly as quickly as we think, because the internet and health care (and other crucial areas being improved by technology) represent important consumer goods or vital components of a flourishing modern life.
There are other ways in which technology in fact reduces inequality of consumption. First, the effects of inflation have been felt less by lower-income Americans than by the wealthy. A major reason is the fact that food — which makes up a greater share of the basket of goods purchased by the lower half of the income distribution than of that purchased by the upper half — has fallen in cost relative to other goods. This decline is not an accident of history: Instead, the costs of manufacturing and agriculture have long been falling as cheaper machines substitute for labor. Genetically modified foods and other advances in biotechnology, meanwhile, have made crops and animals more resistant to disease, yielding greater abundance. Moreover, the distribution of material products has become far more efficient, thanks in large part to the supply-chain innovations of large retail outlets like Wal-Mart. The resulting lower prices have contributed to the relative deflation in the cost of products consumed primarily by people of modest means.
Another result of cheaper manufacturing and distribution processes is that lower-income people enjoy the fruits of the latest technologies much more rapidly. It took hundreds of years after the clock was invented for timekeeping devices to become affordable to the middle class. Even in the last century, for a long time, only the relatively well-to-do had refrigerators and televisions. Today, new technologies circulate throughout the population far more quickly. Five years after the introduction of the multi-touch smart phone, about half of America's population had one.
Technology also helps reduce inequality because it helps generate and harness ideas — and ideas, unlike material goods, are generally free. Once conceived, they can be possessed in common by the rich and poor alike, and ideas serve the rich and poor alike in a variety of capacities. Some ideas are conceptual: People conceive of a new product, like a mobile telephone, or a new mechanism, like a big-box store with its distributional efficiencies. Other ideas are technological: Inventors design new kinds of material technology, like transistors, that are important to mobile telephony. Still others are combinatorial: Synthesizers connect ideas and put them together to make something that is more than the sum of its parts — as when a café is combined with a bookstore or a tablet with a laptop. But all of these ideas produce value for society.
It is of course true that people have been coming up with ideas for thousands of years. But the exponential increases in computational power and connectivity have accelerated the process. And while it is true that most ideas must take material form in order to be converted into usable products, the lower costs of manufacturing (as discussed above) mean that innovative ideas can improve lives at much lower prices.
One possible objection to this argument is that innovations are not really free. In our society, ideas are considered intellectual property — protected by patents and copyrights — and inventors and authors can and do charge for the use of that property. What this argument misses, however, is that only expressions of ideas are subject to intellectual-property laws; the ideas themselves are not. Once the idea of a big-box store or mobile telephony is introduced, it immediately enters the public domain. Our laws also limit protections even for expressions of ideas — requiring, for instance, that patents describing new ideas in detail be fully public in return for their protection, thereby providing the inspiration and foundation for further innovations. And even these limited protections expire; ultimately, the ideas become part of a common pool open to all.
Moreover, technological acceleration itself helps keep the cost of consuming expressions of ideas low (or non-existent). As new products emerge, the values of the slightly older products they replace fall dramatically — even if those products are perfectly usable — regardless of legal protections. Apple was able to charge a large premium for its first iPad, but now tablets with functionality very similar to the iPad 2 are being sold for less than half its price. It hardly seems evidence of significant or harmful inequality that the rich get the latest smartest phone just a year or two before the rest of us do.
Compare our experience in this regard to that of most human beings throughout history, who found economic value chiefly in land or personal property. That value reflected the costs of scarce materials and labor used in the cultivation of the land and the production of the property. As a result, most of the world's wealth could not be jointly consumed. Our digital age, however, offers far more opportunities for common consumption, which means the consumption patterns of the rich and poor converge far more than they used to. This increased common consumption leads to greater equality of living standards than our income measures indicate, because incomes no longer tell as complete a story of our consumption.
Further technological progress may soon make these developments more evident. Nanotechnology aims to radically decrease the cost of manufacturing, with the likely result that more material products now available only to the wealthy will become available to people of modest means. And as virtual reality improves, the difference between a real, material experience and a far cheaper virtual experience will also diminish. One can even imagine a time when a person might virtually have the experience of traveling on a large yacht, formerly the exclusive province of the super-rich. The world of the consumer is dematerializing before our eyes — and the significance of material scarcity to inequality is diminishing along with it.
REDUCING INEQUALITY IN AN AGE OF TECHNOLOGY
Understanding how technological progress reduces inequality of consumption can help us redirect government policy to reduce inequality further still. Public policy aimed at reducing inequality should focus not on direct transfers of wealth, but on improving education and increasing innovation. Innovation creates more ideas that all can benefit from in common; education, meanwhile, allows wider and fuller enjoyment of these ideas, and is itself a spur to innovation.
Education not only helps people increase their incomes by enabling them to better integrate technology on the job, but also improves their consumption by helping them better deploy technology at home. For example, the internet provides an inexpensive mechanism for accessing myriad sources of enjoyment — from finding products that suit one's interests, to playing games, to improving one's understanding of the matters one believes to be most important to the good life. Samuel Johnson once said, regarding travel, that a man must carry knowledge with him if he is to bring knowledge home. The same is true of surfing the web: Those with more knowledge can enjoy it more abundantly. In today's world, the advantages of education are as important as income to raising a person's consumption levels.
Moreover, the pace of technological change now requires education in subjects that would not have been on the curriculum even when today's young adults were in school. Universal access to the internet is the modern equivalent of the 19th century's subsidies for post offices or of the rural electrification projects of the mid-20th century: It is a path to empowerment. Internet-literacy programs for adults (which the Obama administration has rightly promoted) will also be an important priority as the web becomes an ever more essential source of value. As technology progresses, all kinds of high-quality continuing education will be made available online for low cost or even for free. Online courses are already attracting enormous numbers of participants, and ease of access to such resources will be important to enabling people to acquire the skills needed to keep up with technology at work and at home.
When it comes to encouraging innovation, meanwhile, the most important steps the government can take are those that clear obstacles. Most obviously, the government should systematically eliminate regulations that help industry incumbents block the entry of new firms offering disruptive technologies that could transform the marketplace. At the local level, for instance, low-income urban residents would benefit from the introduction of big-box stores that are now thwarted by big-city labor unions. They would also benefit from new car-sharing programs enabled by smart phones — programs that are fiercely contested by taxi companies.
On a global scale, more liberal immigration policies would increase innovation. The free movement of people allows those with talent to collaborate where they will make the most gains. Even in our era of digital communication, innovators in places like Silicon Valley in California and Silicon Alley in Manhattan improve their output by working in proximity to one another.
More generally, a concern with inequality should lead to policies that decrease the social rent-seeking that directs people away from innovation. The too-big-to-fail regime that shields the financial sector has unfairly increased the incomes of some Americans by allowing them to ride to riches on a federal guarantee. Perhaps more important, the artificially vast wealth this regime offers has lured our brightest minds away from producing more innovations in other parts of the economy. The citizenry at large understands the difference between the social benefits of entrepreneurship and the social costs of government guarantees: Steve Jobs and Bill Gates, for example, are not resented for their massive fortunes, because their products have greatly improved the lives of even people of modest means. The bailouts of the big banks, on the other hand, generated hostility toward institutions that lived high on the implicit backing of taxpayers. Of course, we need an efficient financial sector to allocate capital to help drive innovation — but this can certainly be achieved without government support.
At first glance, it might seem like we need to strengthen intellectual-property laws in order to generate more innovation — but that intuition is almost certainly wrong. Excessive protections of intellectual property can actually decrease the rate of innovation. In Launching the Innovation Renaissance, George Mason University economist Alex Tabarrok notes that there has been a great deal of innovation in the field of fashion, where intellectual-property protections are very limited. By contrast, in some fields in which these protections are ample — such as software development — companies use patent litigation to prevent innovation by competitors. Of course, as Tabarrok notes, firms in some sectors — like the pharmaceutical industry — must invest huge sums in developing hard-to-discover products that are extremely easy to copy; they really do need the generous protections they currently have. This suggests the need for greater discrimination in the development and reform of intellectual-property laws to ensure that they promote optimal innovation.
It should be noted that acknowledging technology's capacity to reduce inequality does not mean rejecting government programs intended to alleviate poverty. Even if material goods are becoming cheaper, and even if they occupy a smaller proportional place in a flourishing modern life, there remain some people in the United States (and many more abroad) whose incomes are still too low to comfortably acquire food, shelter, and the means to benefit from the consumption opportunities technology makes possible. But policies targeted at poverty are not the same as those targeted at inequality. Indeed, too great a concern with income inequality may reduce the economic growth needed to produce the surpluses on which anti-poverty programs rely.
A NEW LOOK AT WEALTH AND WELL-BEING
These beginning reflections on the economic value of rapid technological innovation suggest that our political debates about equality and fairness are too often argued using an obsolete vocabulary. Differences in income tell us less than they once did about differences in people's real standards of living. And the immense potential of technology to reshape the lives of the rich and the poor simultaneously has yet to be fully understood and accounted for.
Even when we set out to examine technology and inequality together, we too often get things backwards, considering the effects of technology in terms of who has the money to access it. In reality, the most significant technologies of the past half-century have tended to make money less rather than more important, and have been available more widely, more quickly, than such valuable commodities have ever been before.
In our time, ideas increasingly matter more than things, and information is often more valuable than tangible property. Under such conditions, Americans from different walks of life and in different economic classes have an ever-stronger common interest in promoting innovation. A society with advancing technology is a society in which everyone may rapidly reap the bounty produced by the most talented and entrepreneurial among us. More than ever before, our economy can be a positive-sum game in which gains for some rapidly redound to the benefit of all.