
What this covers
This is a one-hour EconTalk conversation in which Russ Roberts questions Tyler Cowen about his e-book on the slowdown in American prosperity. The argument runs through low-hanging fruit and innovation waves: gains from fossil fuels and machines came quickly, while later ones, in education, life expectancy and even track records, come harder. Cowen cites Julian Simon to say ideas are not exhausted, but he holds that growth arrives in spurts. He stresses how much a gap between 3 and 4 percent annual growth matters, since it changes how fast incomes double.
The middle of the discussion is about measurement and demographics. Cowen argues that a rise in divorce and single-person households distorts snapshot comparisons of median household income, and that the CPI likely overstates inflation, which would widen the gap between early and later decades. Roberts adds that more women entering the workforce was a one-time boost with unmeasured costs. They also consider health care, education and government, roughly 20 to 30 percent of the economy, where spending may be recorded as output without being final value.
The more contested turns concern distribution and politics. Cowen says the internet mainly helps high-skill users, so recent innovation tilts toward the top 10 percent. He treats rent-seeking in finance, with its heads I win, tails you lose innovations, as real but not the main cause. He also argues that slower median growth across the OECD undercuts a crude story of the rich capturing government. He ends on jobless recoveries, the Walmart puzzle of productivity gains that faded, and a forecast that internet-enabled science will eventually pay off in biosciences.
Cowen argues that median American living standards have stagnated since about 1973 because the easy gains from past innovations are exhausted, the big sectors of health, education, and government absorb resources without raising productivity, and recent innovation favors the already skilled rather than the median worker.
- Post-1973 median income, productivity, and education gains slowed, while 1917-1957 saw rapid growth in living standards
- Innovation comes in uneven waves, and the easiest gains (electricity, toilets, cars, the Industrial Revolution) have already been picked
- Health, education, and government consume a growing share of output without measurable productivity gains, and the internet's benefits flow mainly to the cognitively advantaged
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Some market-based innovation in health care and finance is value-destroying rather than value-creating: by 2006 people became masters of the 'heads I win, tails you lose' game, which is innovation that transfers rather than creates value.
“in nineteen sixty three people were not very good at playing heads I win tails you lose by 2006 they become masters at it that's the kind of innovation but it's more value destroying than value creator”
Brains are the ultimate resource that is not subject to physical limits, and innovation from better ideas, such as backpacks and the Kindle, allows people to carry thousands of books in digital form, so Julian Simon's view that innovation will never be exhausted is correct.
“I don't think we'll ever exhaust innovation in that sense I think he's absolutely correct”
Innovation proceeds in uneven bursts followed by stagnation (the pattern of fire, the wheel, the Industrial Revolution), and this is the more common historical pattern, so the process is always uneven.
“that's the more common historical pattern like man invents fire this means a lot of benefits”
The internet will eventually deliver large productivity gains through easing scientific communication, with payoffs in biosciences and genome-related medicine, but as with electricity and computers, these gains take decades to show up in productivity statistics.
“the internet also its biggest benefits will not come within a few years but they will come over decades”
Rent-seeking is when people spend energy, resources, and money to take value from others rather than create it; it can operate through politics, lobbying, or fraud, and is notable in finance and the top of the Harvard class, but it is not the main driver of stagnation.
“rent-seeking is a term that refers to when people invest their energies resources and money in taking things from others rather than creating value”
Left-leaning people do not trust the system to steer growth toward the average person, so they tell a story where the rich stole from everyone else; Cowen argues that median growth slowdowns across most OECD nations since 1973 contradict the crude political story that the rich control government.
“keep in mind there's been median growth slowdowns in most of the OECD nations since 1973 and that to me militates against what I would call the crudely political stories”
The rise in divorce and single-person households from the late 1960s through the 1970s increased the number of households far faster than population growth, so median household income measured across snapshots falls, distorting comparisons of living standards over time.
“there was a rather large increase in the divorce rate in the United States which means that the number of households or families increased much more dramatically than a population growth”
Part of the recent income gains came from more women working outside the home rather than from productivity gains, and this is a one-time shift that cannot be repeated and involved costs in stress and forgone household production that go unmeasured.
“some of the gains we have gotten have come from more women working outside of the house not from productivity gains those are to some extent one-time changes”
The third reason for stagnation is that education gains are harder now: when almost no one went to college in 1900, it was easy to educate the marginal person and get a large gain, but now the marginal student (such as those who would not get into George Mason) is much harder to make more productive, and existing institutions are distorted.
“if you think of hardly anyone is going to college at the beginning of the 20th century it was very easy to educate the marginal person”
The internet has been the most important recent technological marvel because it makes capital more substitutable for labor, which benefits high-skill users like Roberts and Cowen enormously but produces smaller benefits for the median American.
“the Internet has a few implications a one is that it is made capital more substitutable for labor so this has helped people like you and I”
Post-2001 recoveries are jobless: recoveries from the 1990, 2001, and current recessions are slower and increasingly jobless, and this reflects a weaker background rate of job-creating innovation rather than a normal cyclical pattern.
“it's because there are fewer job-creating innovations going on in the background than in previous recoveries”
The internet has been the single most important recent technological marvel, and it has made capital more substitutable for labor, which has greatly benefited people like Cowen and Roberts, yet its gains reach the median American much less, because the most famous internet firms (Facebook) employ remarkably few people.
“the Internet has a few implications a one is that it is made capital more substitutable for labor so this has helped people like you and I”
The left's pessimistic story that rich people stole from everyone else is not supported by the evidence, because median growth slowed in most OECD nations after 1973, which is hard to explain by crude political capture and better explained by demographic and technological change.
“there's been median growth slowdowns in most of the OECD nations since 1973 and that to me militates against what I would call the crudely political stories”
Comparing real income across a century is very hard because the quality change in goods is enormous: portable music went from violinists following you around in 1900 to an iPod holding thousands of songs, so simple nominal price comparison fails.
“portable music in 1900 was for violinists following you around portable music in 2000 was an iPod”
The main gains in life expectancy from 1900 to the 1960s came from penicillin and reductions in infant mortality, and since then progress has continued at a slower pace because later gains are harder to achieve, the low-hanging fruit having been plucked first.
“most of the gains in life expectancy over that period were reductions in infant mortality which are some it's a low hanging fruit and then it's harder and harder to get people to live longer”
Low-hanging fruit: things that have been exhausted in their possibilities are easy to pick at first, and once exhausted, progress must move to something else. Some areas of innovation, such as physical resources, are constrained, while the brain is not, so innovation in clever ideas can keep going.
“the idea of low-hanging fruits appealing that there's certain things that you exhaust there possible they're easy you exhaust their possibility or their potential and then you have to move on”
Health care gains (penicillin, life expectancy rising from about 40 to 70 by the 1960s) were rapid, but since the 1960s progress has slowed because most of the life-expectancy gains came from reducing infant mortality, and further gains are harder, which is simply the low-hanging fruit effect.
“even in health care the big gains are like penicillin life expectancy goes up from around 40 to 70 by the 60s”
Physical and biological limits mean that human performance improvements (such as track-and-field records) fall at a decreasing rate, because quality can be measured and progress eventually exhausts improvements in nutrition, training, and knowledge.
“I assume that if you look at say x for the mile in the world record for the mile or other measures of human physical achievement that they fall at a decreasing rate”
Growth rate differences matter enormously: at 4 percent a year the average living standard doubles in about 18 years, at 3 percent in about 24, and at 2 percent in about 35.
“4% a year you double yours in 18 years roughly and at 3% it takes 24 and if you down at 2 it's 35 years”
The main measurement bias to worry about is that large spending in health and education is recorded as value added even when it may not be final output, so measured growth overstates real productivity in those sectors.
“the bias I worry about the most is we're spending a lot of money and we're simply writing it down as value-added when it may not be final”
Implementation of ideas matters as much as clever ideas: Walmart, which does what stores have always done of buying and selling goods, made cheap goods far more accessible through technology, computers, and the Internet in its retailing.
“an emphasis on clever ideas versus implementing ideas Walmart which is in some sense does just what any store did in 1950 or 1920”
There are still plenty of creators, entrepreneurs, and geniuses, and much innovation even during the Great Stagnation; the question is where that innovation is directed, since the goods that most benefit the median (like basic household items) have already been filled out.
“there are still plenty of creators plenty of entrepreneurs plenty of geniuses an enormous amount of innovation even during what I am calling the great stagnation the question is where is that innovation geared”
Some market-based innovation in finance has questionable real returns, as in the 'heads I win, tails you lose' game: in 1963 people were not good at it, but by 2006 they became masters, so this is innovation that is more value-destroying than value-creating.
“in nineteen sixty three people were not very good at playing heads I win tails you lose by 2006 they become masters at it that's the kind of innovation but it's more value destroying than value creator”
The biggest bias in measuring output is that large spending in health, education, and government gets written down as value added even when it does not produce final value, so national income overstates real output in those sectors.
“the bias I worry about the most is we're spending a lot of money and we're simply writing it down as value-added when it may not be final”
The third reason for the stagnation is that the easy marginal gains in education are gone: when almost no one went to college early in the 20th century, educating the marginal person was easy and produced big gains, but educating the marginal student today (someone who would not get into George Mason) is much harder.
“if you think of hardly anyone is going to college at the beginning of the 20th century it was very easy to educate the marginal person”
Innovation is still abundant during the stagnation: there are many creators, entrepreneurs, and geniuses; the question is where innovation is directed and whether it targets the goods and services that benefit the median, since many everyday domains (such as toilets) are already filled out.
“there are still plenty of creators plenty of entrepreneurs plenty of geniuses an enormous amount of innovation even during what I am calling the great stagnation the question is where is that innovation geared”
Innovation over the last decade has been less geared toward the median and more toward the top 10 percent, producing a change in the distributional pattern of benefits.
“innovation itself at least over the last 10 years has been less geared toward the median and more geared toward a mix of the Infowars and the top 10%”
The current innovation wave (Google, Facebook, eBay) is about information processing and ordering, and the people who benefit most are cognitively better prepared; the lag in educational achievement in the US is part of why the benefits are skewed away from the median American.
“the people who benefit from those tend in cognitive terms to be better prepared and it's not quite the median American who's the main beneficiary”
Post-1970s divorce growth meant household and family counts grew much faster than population, so the median shifted to the left among the existing population, and snapshot median-income comparisons understate growth for the same people followed over time.
“there was a rather large increase in the divorce rate in the United States which means that the number of households or families increased much more dramatically than a population growth”
The last decade's weak job creation reflects a slowdown in productivity and fewer job-creating innovations in the background, which is why recent recoveries are 'jobless recoveries' that are much slower than past recoveries (2001 slower than 1991, and the post-2007 recovery slower than both).
“it's because there are fewer job-creating innovations going on in the background than in previous recoveries”
Brains are the ultimate resource with no physical limits, so innovation in ideas (from backpacks to the Kindle to digital libraries) can overcome physical constraints indefinitely, and Julian Simon was right that we will not run out of innovation or physical resources.
“the human brain allows us to overcome lots of physical limitations I still admire Simon's work greatly”
Health care (about 17 percent of GDP), education (about 6 percent of GDP in major-GDP terms), and government consumption (15 to 20 percent) together make up roughly 20 to 30 percent of the economy, and these sectors have massive subsidies and distorted incentives, and their value is hard to measure.
“about 17% of it is health care about 6% of it in terms of major GDP is education and with some overlap 15 to 20 percent is what we call government consumption”
Productivity gains in the late 1990s raised typical wages, and analysts attribute part of these gains to retailing innovation like Walmart, but those gains faded in the data, which is a puzzle.
“there's one period post 73 where you have big productivity gains and they show up in typical wages and that's the late 90s wages at the median go up quite a bit”
Growth tends to arrive in spurts after major technologies, so the stagnation reflects a technological plateau: after the Industrial Revolution's fossil fuel and machine gains were harvested quickly, the next fundamental improvement (replacing the car) is very hard and will take time.
“there are technological plateaus if we look at the broader sweep of history growth tends to come in spurts so the car was a big deal”
The demographic change of rising divorce, single-parent families, later marriage, and more single people entering the workforce unprepared is part of the stagnation story across the developed West including Europe and Japan, and reflects a social adjustment.
“the developed nations they also went through the same demographic change that we did they went to a big increase in divorce a big change in family structure”
Low-hanging fruit explains the slowdown in physical achievement: track and field records fall at a decreasing rate because improvements in nutrition, training, and knowledge eventually get exhausted.
“I assume that if you look at say x for the mile in the world record for the mile or other measures of human physical achievement that they fall at a decreasing rate”
The Industrial Revolution's combination of fossil fuels and sophisticated machines produced big, quick gains, and growth comes in spurts, so the economy is now waiting for the next major technological breakthrough, and replacing the car with something fundamentally better is a very hard problem that will take time.
“if we look at the broader sweep of history growth tends to come in spurts so the car was a big deal”
Some of the post-war income gains came from more women working outside the home rather than from productivity gains, and this is a one-time change that cannot be repeated and carries costs (stress and forgone household production) that the statistics do not measure.
“some of the gains we have gotten have come from more women working outside of the house not from productivity gains those are to some extent one-time changes”
Comparing real income across a century is very hard because quality changes are enormous: portable music went from violinists traveling with you to an iPod holding thousands of songs, so nominal prices of the two cannot be directly compared.
“portable music in 1900 was for violinists following you around portable music in 2000 was it was a an iPod”
Health care (about 17 percent of the economy), education (about 6 percent of GDP), and government consumption (15 to 20 percent, roughly 20 to 30 percent combined with overlap) form a large share of output dominated by subsidies, distorted incentives, and poorly measured value.
“about 17% of it is health care about 6% of it in terms of major GDP is education and with some overlap 15 to 20 percent is what we call government consumption”
The Western developed nations (Europe and Japan) went through the same demographic change as the US, including rising divorce, more single-parent households, later marriage, and more single people entering the workforce unprepared, which Cowen sees as part of the stagnation through social adjustment.
“they went through the same demographic change that we did they went to a big increase in divorce a big change in family structure”
Consumer price index measurement likely overstates inflation, and the bias is largest when fundamentally new goods are introduced and spread through the market, so correcting it would make the growth gap between early and later periods larger, not smaller.
“those biases will be largest in periods when you have fundamentally new Goods and during the marketplace and spreading”
The financial sector grew dramatically without clear productivity gains, and its share of corporate profits peaked in 2006, at over a third of corporate profits, which in retrospect reflected rent-seeking rather than productivity.
“the financial sector of course really blossomed in the numbers in 2006 I think is the peak it was I think over a third of corporate profits”
Price index measurement is biased in a way that understates real income growth. The bias is largest when fundamentally new goods enter the market and spread, so inflation is probably overmeasured across all of history, and correcting it would make the gap between early and later periods larger, not smaller.
“those biases will be largest in periods when you have fundamentally new Goods and during the marketplace and spreading”
Implementing ideas (not just clever ideas) drives living standards: Walmart does what any store did in 1920 or 1950 (buy goods and sell them), but its use of technology and computers made cheap goods far more accessible.
“Walmart which is in some sense does just what any store did in 1950 or 1920 they buy a bunch of stuff and they sell it to people but because of their ability to use technology”
A 3 percent annual growth rate versus 4 percent makes a huge difference: at 4 percent a year incomes double in about 18 years, at 3 percent in about 24 years, and at 2 percent in about 35 years.
“4% a year you double yours in 18 years roughly and at 3% it takes 24 and if you down at 2 it's 35 years”
Historical growth follows an uneven pattern: after an innovation like fire, benefits play out and the economy stagnates for a while before the next wave, so slowdowns are the more common historical pattern rather than the exception.
“that's the more common historical pattern like man invents fire this means a lot of benefits”
Government can make one-off changes in the distribution of income (for example, Sweden has a more equal distribution than the United States), but such one-off transfers are swamped over time by changes in growth rates that stem from more fundamental factors such as productivity and technology.
“government I think is pretty good at making one-off changes in the distribution of wealth but those one-off shifts over time get dwarfed or swamped by changes in rates of growth”
Government can make one-off changes in the distribution of wealth (such as Sweden's more equal wealth distribution), but such transfers are partly lost to bad incentives, and over longer horizons productivity and growth rates swamp these one-off distributional shifts.
“government I think is pretty good at making one-off changes in the distribution of wealth so in Sweden the distribution of wealth is more equal”
Cowen does not blame the shift in income distribution mainly on politics as Paul Krugman or Jacob Hacker do; he thinks innovations come in waves, and the current wave favors information-processing innovations.
“I don't blame the shift and distribution on politics the way say Paul Krugman would or Jacob hacker would I simply think innovations come in waves”
Much education spending is consumption rather than investment: modern college dorms look like hotels and high school spending on administration, facilities and gyms does not raise future productivity or growth.
“a lot of what we call education spending is consumption not investment”
Living standards for the average American family grew very rapidly from 1917 to 1957 (electricity, flush toilets, automobiles), but growth slowed after 1973, and that post-1973 slowdown is what Cowen means by the Great Stagnation.
“there's a slowdown in growth in living standards for the average person or the average family and that's what I mean when I refer to the great stagnation”
General-purpose technologies such as electricity and the computer took decades to have a significant productivity impact, so the internet's biggest benefits will come over decades rather than within a few years, and the stagnation is likely to end within our lifetimes.
“electricity took actually decades to have a significant impact on productivity the computer took decades to have the significant impact on productivity”
Medicare should be restricted to catastrophic expenses rather than covering the elderly per se, and the best health system is Singapore's, which uses health savings accounts for smaller expenditures, single-payer coverage for catastrophic costs, and a savings mandate integrated with retirement.
“I think ideally Medicare should be for catastrophic expenses and not for the elderly per se I think the best health care system in the world is that of Singapore”
Cowen says he does not blame the shift in income distribution on politics the way Paul Krugman or Jacob Hacker do; he believes innovations come in waves, and each period favors one kind of innovation.
“I don't blame the shift and distribution on politics the way say Paul Krugman would or Jacob hacker would I simply think innovations come in waves”
Major internet firms such as Facebook employ remarkably few people relative to their output, so the benefits of internet innovation are large but do not translate into job creation or median income gains.
“if you look at the most famous internet-based firms like Facebook the employee remarkably small numbers of people”
Education is a poor example of progress: high school graduation rose from very low levels to about 60 percent between 1900 and 1960, but the rate has since stagnated and is now declining, and so K-12 education may be moving backward even though college attendance rose.
“rates of graduating from high school went from very low to 60% the rate has stagnated now for several decades and it's actually declining”
Healthcare should use catastrophic-only Medicare coverage with health savings accounts for smaller expenses, and Singapore's system, with savings accounts, single-payer catastrophic coverage, and a savings mandate integrated with retirement, is the best model; but such reforms are not politically practical in the US.
“I think the best health care system in the world is that of Singapore where you have health savings accounts for smaller expenditures single payer coverage for catastrophic”
A self-propelling process explains strong recoveries: new ideas, new businesses, and firms hiring new workers generate income, employment, and output that feed upon themselves, so one does not need to be a Keynesian to see why recoveries are strong when innovation is high.
“you get the self propelling process you don't have to be a Keynesian to think that this feeds upon itself and it makes the recovery strong”
Bank leverage went from about twelve to one to about thirty to one before the crisis, which was a mistake; returning to twelve to one is painful but was workable before, and the people who gained from the move up to thirty have a cushion to absorb the pain.
“it used to be twelve to one and it went up to thirty to one that was a mistake”
Measurement bias in the consumer price index is not clearly worse in 1946-1973 than after 1973, and the post-war gains were partly a matter of diffusion: by 1973 virtually all homes had basics like electricity and plumbing that many lacked in 1946.
“in 1946 in this country you still have a lot of homes that don't have absolute basics and by 1973 virtually all of them do”
Much of the spending on higher education is consumption rather than investment: dorm rooms look like hotels, and high school administration, facilities, and gyms do not increase productivity or growth.
“a lot of what we call education spending is is his consumption not investment”
Government policy is a second cause of the stagnation: policies distort incentives, and economists who criticize such policies for lowering growth are inconsistent when they doubt that average-family growth has slowed.
“a main criticism of these policies is that they will lower our economic growth so I'm coming along and saying the rate of economic growth for the typical family cleaned”
The financial sector grew dramatically without necessarily adding productivity, and defense spending may protect against threats but does not make the country better off, so both should be added to the list of reasons for stagnation.
“you've left out the financial sector which is the sector that also has grown dramatically without much perhaps absolutely”
The financial sector's share of corporate profits peaked around 2006, at over a third of corporate profits in that year, and in retrospect this growth reflected rent-seeking rather than productivity.
“in the financial sector of course really blossomed in the numbers in 2006 I think is the peak it was I think over a third of corporate profits in that year”
Bank leverage rose from about 12-to-1 to 30-to-1 and that was a mistake; moving back to 12-to-1 is painful, but the world did not end at 12-to-1, and the people who enjoyed the move up have cushions to absorb the pain.
“it used to be twelve to one and it went up to thirty to one that was a mistake”
Financial reform should have minimal explicit regulation and tight restrictions on leverage, because leverage has proven dangerous and micro-regulating banks is very hard; Cowen would also repeal much of Dodd-Frank, and he doubts we can commit to no bailouts.
“the best way to proceed is to have a minimum of explicit regulation but really pretty tight restrictions on leverage”
Education policy should favor experimentation at the state and local level, including charter schools, school choice, and the ability to fire bad teachers; the Obama administration's unpopularity with teachers unions signals a shifted political equilibrium.
“charter schools are good I favor more experiments with school choice mostly I think we need more experimentation”
A psychological inability to accept slower growth is a big part of the problem: politicians cannot fix the economy or promise fixed growth rates, and fiscal discipline and a sense of limits are needed, which the political system lacks from either party.
“a big part of the problem is simply we are not psychologically equipped to deal with this slowdown”
The illusion of a smooth 3 percent annual growth rate for the American economy is a dangerous bird's-eye view, because innovation proceeds by trial and error and growth is uneven; the illusion was less dangerous in 1985 but is now dangerous because fiscal spending and taxing decisions are leading toward a precipice.
“we have this illusion that it's 3% a year but that's the growth rate of the American economy”
Harvard graduates' career choices illustrate misdirected talent: striking numbers of the smartest, most ambitious graduates want to go into finance, which Cowen thinks is unhealthy and not the best form of human-capital investment in innovation.
“if you look at a graduating class from Harvard it's striking how many of the smartest most ambitious ones want to go into finance and I don't think that's healthy”
The Great Stagnation refers to a slowdown in growth of living standards for the average person or family after roughly 1973, in contrast to the very rapid gains in living standards between 1917 and 1957 (electricity, flush toilets, automobiles).
“there's a slowdown in growth in living standards for the average person or the average family and that's what I mean when I refer to the great stagnation”
Stagnation shows up in phases with different starting points: the high school graduation rate rose from very low levels to about 60 percent between 1900 and 1960, then stalled for several decades and is now declining.
“rates of graduating from high school went from very low to 60% the rate has stagnated now for several decades and it's actually declining”
Panel data that follow the same people over time show much larger income increases than snapshot medians, and studies such as Duncan, Becker and Gordon using PSID panel data find income stagnation results similar to Census results, which weakens the idea that household change overturns the stagnation finding.
“the increases are quite dramatic and certainly much larger than just looking at snapshots of the median over time”
Government policies that critics say reduce growth are the same ones that have been enacted, yet economists who criticize them for lowering growth are skeptical that the typical family's growth rate has slowed, which is an inconsistency Cowen points out.
“typically a main criticism of these policies is that they will lower our economic growth so I'm coming along and saying the rate of economic growth for the typical family slowed”
Current spending and tax decisions are leading the US off a fiscal precipice, and within about 20 years bond markets will force a fiscal adjustment that burst the illusion of steady growth and require reorganizing government.
“our combination of spending and taxing decisions is leading us off the precipice fiscally sometime in the next 20 years well before then given bond markets something we'll have to give”
The CPI measurement problem is probably no worse in the 1946-1973 period than after 1973, and the post-war period had a big spread effect: in 1946 many homes lacked basics that virtually all homes had by 1973, so the measured growth gap is genuinely different across periods.
“in 1946 in this country you still have a lot of homes that don't have absolute basics and by 1973 virtually all of them do”
The fiscal reckoning is coming: spending and taxing choices will force a reorganization within about 20 years, well before then given bond markets, and the illusion of steady growth will be burst.
“I'm finally reaching the point fiscally sometime in the next 20 years well before then given bond markets something we'll have to give and that illusion will be burst”
Realistic policy responses to stagnation include restraining the growth of Medicare spending (the main driver of fiscal problems), introducing healthy incentives, and taking on special interests, and Cowen favors government subsidies to pure science research because it is a public good.
“the rate of increase of Medicare spending would be slowed that's really the driver of our current fiscal problems probably unlike you I absolutely favor subsidies to pure research”
Productivity statistics, median income, and education and test-score data are broadly independent measures that all point toward some kind of slowdown, which makes the stagnation claim more robust than any single statistic.
“we have a bunch of different facts which are broadly speaking independent pointing in the direction of a slowdown of some kind”
Financial regulation should be minimal in explicit rules but with tight restrictions on leverage, since leverage has proven dangerous, and banks are hard to micro-regulate; Cowen would also repeal much of Dodd-Frank.
“the best way to proceed is to have a minimum of explicit regulation but really pretty tight restrictions on leverage”
The belief that the American economy grows at a steady 3 percent a year is a dangerous illusion: the actual process of innovation is trial and error and uneven, and the illusion persisted because it was cheap to believe in 1985, while fiscal pressures from spending and taxing decisions will force a reckoning within about 20 years.
“we have this illusion that it's 3% a year but that's the growth rate of the American economy and that's a bad bird's-eye view”
Treating the three big sectors of health, education, and government as lacking market prices is the core of the argument, and critics who cherry-pick sectors that contradict their own worldview (left or right) to blame a cause are making a clever debating move.
“it's so easy for left and right to cherry-pick things that are go against their sides worldview and said well that's the cause of whatever's going wrong”
The political economy problem is that the great stagnation would be easier to live with if politicians recognized that politics cannot make the economy grow at a set percentage each year, and if fiscal discipline and a sense of limits existed; the main problem is that Americans are not psychologically equipped to deal with a slowdown.
“a big part of the problem is simply we are not psychologically equipped to deal with this slowdown”
Innovation over the last decade has been geared less toward the median American and more toward the top 10 percent, producing a shift in the distributional pattern of benefits.
“innovation itself at least over the last 10 years has been less geared toward the median and more geared toward a mix of the Infowars and the top 10%”
Median family income since 1973 has risen only modestly (on one plausible estimate from 1973 to 2004, only 20 to 30 percent), and in the most recent decade median income has actually declined and net job growth has been roughly zero.
“under one plausible estimate from 73 to 2004 it's only gone up 20 to 30% in the last decade it's actually declined there has not been net jobs growth in the last decade”
Several largely independent data sources (median income, productivity statistics, education, and test scores) all point toward some kind of slowdown, so the stagnation thesis does not rest on a single measurement.
“we have a bunch of different facts which are broadly speaking independent pointing in the direction of a slowdown of some kind”
Cowen would decriminalize most or all drugs, cut defense spending, and eliminate all farm subsidies, and he argues that because we do not know in advance which reforms raise the long-term growth rate, the right approach is to institute as many policy improvements as possible and let growth reveal the effects.
“we should institute as many policy improvements as we can we won't know in advance which ones really help the growth rate”
Progress is deeply uneven: productivity bursts are now closer together than before, owing to science, capitalism, decent governance, and the internet, which makes Cowen pessimistic about the left's story but optimistic about the economy overall.
“my view is that progress is deeply uneven we're in a slow patch it will come back I don't pretend to know when I see a lot of positive developments”
Pro-growth policies such as smart regulation, price-based regulation, lower marginal tax rates and policies that encourage savings are the traditional tools, but many on the left do not trust the system to steer growth toward the average person.
“the tradition pro-growth policies which would be quote smart regulation price based regulation lower marginal tax rates policies that encourage savings”
The internet will ease scientific communication and produce significant productivity gains in the longer term, and the next technological wave may come in the biosciences and genome-related medicine.
“one of the great benefits of the internet is how much it eases scientific communication this will in the longer term lead to significant productivity gains”
The time distance between productivity bursts is shrinking because of science, capitalism, decent governance, and now the internet, which makes Cowen optimistic even though he expects the slow patch to end.
“I think the time distance between productivity bursts is smaller than it used to be in large part because of science and capitalism in the modern world and decent governance and now the Internet”
Cowen would decriminalize most or all drugs, cut defense spending, and eliminate all farm subsidies, and he argues that policy improvements should be instituted in bulk because policymakers cannot know in advance which ones will improve long-run growth.
“we should institute as many policy improvements as we can we won't know in advance which ones really help the growth rate and which ones just give us one off improvements”
Market prices capture the value of new consumer innovations like the smartphone only partly: most Americans do not yet find smartphones worth the money, but adoption is growing rapidly, and the 20-to-30-year horizon may see a climb out of stagnation.
“it is my view growing like crazy that in the next twenty to thirty years there's a very good chance we will climb out of the great stagnation”
Market prices do capture the value of new consumer technology to some extent: most Americans do not yet find the iPhone worth the money, but smartphone adoption is growing rapidly and will likely lead the country out of stagnation within 20 to 30 years.
“in the next twenty to thirty years there's a very good chance we will climb out of the great stagnation”
Many left-leaning people do not stress economic growth as much as they should, because psychologically it is hard for growth to become a ruling idea when distribution and social justice seem more important, though the pro-growth progressive coalition since the Clinton years is a positive sign.
“a lot of people who would call themselves as being on the Left don't stress the importance of economic growth as much as they ought to”
Median family income in the US has risen only about 20 to 30 percent from 1973 to 2004 under one plausible estimate, and it has declined in the last decade, with no net job growth over that decade.
“under one plausible estimate from 73 to 2004 it's only gone up 20 to 30%”
Many great 20th-century devices (central air conditioning, dishwashers, microwaves) became widely available by the 1970s and 1980s because rising wealth let most households afford them, so the mid-century gains reached the average American broadly.
“they did get access to say central central air air-conditioning in the south and in the north dishwashers a lot of microwaves”
Physical performance records such as track and field measures fall at a decreasing rate, reflecting physical limits, so the decline in progress is a real and measurable phenomenon and not only a matter of how we define quality.
“I had a blog post on marginal revolution a few weeks ago about track and field and there it's definitely true”
Median income is preferable to mean income for measuring the typical American's affluence, because mean income is pulled upward by very high earners and does not represent mass affluence.
“mean income mean income would be the average and that would include very high income people who would pull the average up and would not represent the typical American”
Rent-seeking is investing energy, resources, and money in taking things from others rather than creating value, often through politics, though simple fraud is also an example; it is a notable factor in the stagnation, though not the main driver.
“rent-seeking is a term that refers to when people invest their energies resources and money in taking things from others rather than creating value”
Facebook and other leading Internet firms employ remarkably few people, which Cowen says is a feature rather than a bug because it lets a few people finance great outputs.
“the employee remarkably small numbers of people this is a benefit in many ways”
Median income data omit employer-paid benefits such as health insurance, and because those benefits grew substantially in the post-war period, the census-based income measure understates total compensation growth, though the consumer value of those benefits may be smaller than their cost.
“doesn't include compensation folk measures of compensation doesn't include benefits which of course over the period again post-war there's a lot of increase in the value of those benefits”
Most major 20th-century devices such as central air conditioning and dishwashers became widely available by the 1970s and 1980s because rising incomes let many people buy them, so the average American's access to these gadgets was broadly achieved.
“a lot of the great devices of the 20th century are now widely available right even in 1970 and again I think the 70s were in 80s were pretty good times”
Labor accounts for almost 70 percent of economic output, so the stagnation of educational attainment matters for the economy as a whole, and it is not obvious that the US is doing better at the K-12 level.
“labor accounts for almost 70 percent of economic output”
The census median income measure omits employee benefits, and the value of those benefits (largely health care, whose value to consumers is questionable) has grown in the post-war period, so the income statistics understate total compensation growth.
“doesn't include compensation folk measures of compensation doesn't include benefits”
Mean income overstates typical living standards because very high earners pull the average up, so median income is the better measure of mass affluence, though median measures still have problems.
“mean income mean income would be the average and that would include very high income people who would pull the average up and would not represent the typical American”
Measurement bias in the CPI is not clearly larger in the 1946-1973 period than in the post-1973 period, so the measured growth gap between those eras is substantively different despite measurement problems.
“it's harder to make the case that CPI problems in say the 46 to 73 period are more or worse than they are say in the 73 or onward period”
Whether recoveries are driven by new firms or by existing firms expanding can be tested with recent data, though existing firms also innovate, so the test may not map cleanly onto the innovation hypothesis.
“how much does the increase in employment after a recession come from new firms coming into existence versus existing firms expanding”
Measured education is overrated as the sole determinant of growth, because many creative people learn outside classrooms and change lives in extraordinary ways, even though education has in fact improved.
“measured education is a little bit overrated as a is the only determinant of growth”
The government should subsidize pure research, since research is a public good that speeds technological progress, and NIH funding has been a good investment, though subsidies alone will not pull the economy out of its rut.
“probably unlike you I absolutely favor subsidies to pure research I think it's a public good I think it speeds technological progress I think the NIH for the most part has been a good investment”
Defense spending may protect against new threats, but it is not obvious what the country gets for that money, and spending is not the same as being better off.
“defense spending that's trickier but it's not obvious what we're getting with a lot of that money”
Cowen favors government subsidies for pure research as a public good that speeds technological progress, and he thinks NIH has been a good investment, though he does not think subsidies alone pull the economy out of its rut.
“I absolutely favor subsidies to pure research I think it's a public good I think it speeds technological progress I think the NIH for the most part has been a good investment”
The pessimistic story on the political left, that the rich have taken over the government and must be stopped with 90 percent marginal tax rates and a 50 percent unionized economy, is politically and economically impossible, while Cowen sees progress as uneven but temporally shorter between productivity bursts than before.
“the time distance between productivity bursts is smaller than it used to be in large part because of science and capitalism in the modern world and decent governance and now the Internet”
A striking share of the smartest and most ambitious Harvard graduates go into finance, which Cowen does not regard as a healthy allocation of human capital to innovation.
“if you look at a graduating class from Harvard it's striking how many of the smartest most ambitious ones want to go into finance and I don't think that's healthy”
A better economic policy regime is needed, but Cowen's stagnation thesis is not an argument for government policies that cannot fix the economy, and the main policy problem is the political expectation that presidents can set growth targets.
“the president can't always fix the we're not going to grow at X percent every year”
The education sector should be reformed through state and local experimentation under federalism, including charter schools, school choice, and the ability to fire bad teachers; the Obama administration's unpopularity with teachers unions signals that the political equilibrium has shifted.
“the Obama administration is very unpopular with teachers unions and I take that to be a very good sign”
A late-1990s productivity boom raised typical wages significantly, with many analysts attributing the gains to retailing innovation like Walmart, but those gains faded in the data.
“there's one period post 73 where you have big productivity gains and they show up in typical wages and that's the late 90s wages at the median go up quite a bit”
Cowen is not trying to persuade people to be pro-growth; he is trying to figure out what he thinks is correct, and he accepts that his thesis may make good policy harder to achieve in the political arena, because the connection between good policy and good results is looser than thought.
“I don't think of myself is trying to convince them or anyone for me it's trying to figure out for me what I think is correct”
Books should be shorter and cheaper: most books are stretched-out magazine articles, and Cowen's short e-book was priced at $4 to test a model of concise books, which he regards as a marker of a new publishing world.
“people liked short books most books are way too long stretched out magazine articles I thought let's have a book cut to the chase tell people what I think and then end it and charge people a commensurate price”
Measured education is overrated as the main determinant of growth, because many creative people learn outside classrooms and change lives in extraordinary ways, though education has improved.
“measured education is a little bit overrated as the only determinant of growth”
The last ten years are atypical as a recovery and innovation period: the 2000s include a recession and a financial crisis, so they may not represent what is going on in innovation, a point Roberts disputes.
“the last ten years are and not so typical right sure you got a you had a recession”
Venture capital has been slow lately after the financial crisis, and that slow patch is another sign of a stagnant period, though whether a given innovation comes through venture capital or government will be seen over time.
“venture capital has been pretty slow lately it's recovered from the crisis but it's not impressing people”
Median household income kept growing robustly after 1989 in the same way as the unadjusted series, and in the 1960s household size shrank even faster, so most researchers do not think changing household size overturns the finding of slower living-standard growth.
“most people who work in this area they don't see changing household size as fundamentally overturning that”
Panel data studies such as Duesenberry-type work using PSID data (the Becker and Gordon paper) show income stagnation results similar to census data, and so panel evidence does not overturn the stagnation thesis.
“if you look at the du Becker and Gordon paper which uses panel data they get results on income stagnation which are very similar to what the census gets”
Household size shrank even faster in the 1960s, yet median household income still grew robustly then, so changing household size does not overturn the finding of slower growth in average living standards, according to most researchers in the area.
“in the 1960s households are shrinking even more rapidly than they have been in recent times and yet you still have robust growth or the median household”
Government's role in the economy has grown since the 1980s: the overall trend is toward a larger, more active government, with growth in spending and regulation, and since the 1980s much of that growth has been in state and local rather than federal dollar terms.
“if you look at the overall trend it's toward a larger and more active government and since the 80s that growth has mostly been stayed in not federal in dollar terms”
Roberts argues the last ten years are not typical, since the 1980s and 1990s had good job growth overall and the 2000s had a recession and a major financial disaster, so the recent innovation pattern may not be a trend.
“the last ten years are and not so typical right sure you got a you had a recession in the beginning of the of the two thousands”
Rising Medicare spending is the main driver of current fiscal problems, and reforms to health care should introduce healthy incentives and take on special interests to slow the rate of increase.
“the rate of increase of Medicare spending would be slowed that's really the driver of our current fiscal problems”
The politics of the US has become uglier, stupider, and more polarized, but American politics in the 19th century was also partisan and polarized, so the current politics is not a new innovation and is likely to survive; Cowen is somewhat pessimistic about it.
“if you look back at the broader history of the American Republic politics especially in the 19th century has been very partisan and polarized and stupid and based in lies”
The share of employment growth after a recession coming from new firms versus existing firms is a testable way to see whether innovation is driving recoveries, though existing firms also innovate, so the comparison may not be exact.
“how much does the increase in employment after a recession come from new firms coming into existence versus existing firms expanding”
Publishing innovation: the e-book version of the book is cheap and short (about 2,000 words, the length of 17th-century economics pamphlets), because most books are too long and prices of books should be lower, shorter, and priced commensurately.
“most books are way too long stretched out magazine articles I thought let's have a book cut to the chase tell people what I think and then end it and charge people a commensurate price”
Pro-growth policies that are unpopular with the left, such as price-based regulation, lower marginal tax rates, and policies that encourage savings, are needed to steer capital toward productive uses rather than rent-seeking.
“the tradition pro-growth policies which would be quote smart regulation price based regulation lower marginal tax rates policies that encourage savings”
Politics has become uglier, stupider, and more polarized, but American politics in the 19th century was similarly partisan, so the current political dysfunction is not new, and the large number of smart, hardworking people wanting to innovate is a major reason for optimism.
“if you look back at the broader history of the American Republic politics especially in the 19th century has been very partisan and polarized and stupid”
Many people on the political left do not stress economic growth enough, and it is psychologically hard for growth to become a ruling idea when distribution and social justice dominate, though a pro-growth progressive coalition has emerged since the Clinton years.
“a lot of people who would call themselves as being on the Left don't stress the importance of economic growth as much as they ought to”
Government has grown more active since the 1980s: there has been more spending and regulation, with some deregulation, and the growth has shifted from state to federal spending in dollar terms, and the overall trend toward a larger government is clear.
“if you look at the overall trend it's toward a larger and more active government and since the 80s that growth has mostly been stayed in not federal in dollar terms”
Cowen's book is distributed freely or cheaply as an e-book, and this distribution is an innovation: he argues that books should be shorter and cheaper, with about 2,000 words that echo 17th-century economics pamphlets, and that the Kindle is clunky but benefits Infowaris first.
“I thought let's have a book cut to the chase tell people what I think and then end it and charge people a commensurate price I think prices of books should be lower books should be shorter”
Robin Hanson's prediction that brains will be uploaded to computers within 50 to 100 years is far off, and civilization may end before anything like a technological singularity arrives; Cowen is a skeptic of the singularity.
“I'm a big skeptic about the singularity my colleague Robin Hanson his version of the singularity is that within fifty to a hundred years our brains are uploaded”
Cowen is not trying to persuade anyone to adopt pro-growth policies; his goal is to figure out what he thinks is correct, and the stagnation thesis may make good policy harder to achieve because the link between good policy and good results is looser than people think.
“the connection between good policy and good results is looser than we used to think and this may make it harder to achieve good policy”
Cowen is optimistic because more smart people than ever before work hard and want to innovate, and the number of capable innovators is a major reason to expect the stagnation to end.
“what I do see is like more smart people than ever before and they work really hard and they really want to innovate and that absolutely is true of 2011 and it's a big reason to be optimistic”
Future health care gains may come from understanding of genetics and the genome, which could produce a new wave of progress.
“I think we will see some other wave at some point due to understanding of genetic in the genome and other facts about how human beings work”
Cowen is a skeptic of the technological singularity: he thinks we are very far from it and that civilization as we know it would more likely end before reaching a singularity.
“I'm a big skeptic about the singularity”
Roberts argues the 1973 cutoff is a little strange, and that the stagnation should be dated and examined in phases with different starting points rather than as a single break.
“the nineteen 73 cutoff though I find a little bit strange”
Cowen says he worries more about what people study than about how many attend college, treating the question of curriculum as separate from attainment.
“I would also worried about what people study I think that's a separate question”
Economists who accept that markets are productive are being cherry-picked: left and right each select sectors that fit their worldview, and the three big stagnating sectors (health, education, government) are ones where prices are used little or not at all.
“the three sector is just the three bigs there are three big sectors of the economy big and growing health education and government where we use prices less or not at all”
The progressive story is pessimistic because it holds that the rich have taken over government and must be pushed back with 90 percent marginal tax rates and a 50 percent unionized economy, which Cowen says is never going to happen politically or economically.
“the progressive story in my view is the pessimistic one it's like the rich have taken over the government and we need to seize it back and Institute 90 percent marginal tax rates and a 50% unionized economy”
Venture capital is important for innovation but has been in a slow patch since the crisis, and this slowdown is another sign of a stagnant period, although it is not clear whether a given innovation comes through venture capital.
“venture capital has been pretty slow lately it's recovered from the crisis but it's not impressing people”
There are more smart people working hard and wanting to innovate than ever before, which is a big reason to be optimistic about the future.
“what I do see is like more smart people than ever before and they work really hard and they really want to innovate and that absolutely is true of 2011 and it's a big reason to be optimistic”