Bryan Caplan  

Existence, Enhancement, CPI Bias, and Progress

Now versus 1973... My Talk at Berkeley...
Almost all economists agree that the official Consumer Price Index is biased upwards.  Two key flaws with the CPI: It imperfectly accounts for (a) quality improvements, and (b) new products.  The Boskin Commission famously estimated that the official annual CPI increase was over a percentage point too high. 

When you measure economic welfare over time, this seemingly academic issue has massive implications:  Year after year, real income rises by over 1% more than official figures claim.  If you overestimate CPI by 1.3 percentage points per year, then after 40 years real income will be 68% higher than you imagine.

In The Great Stagnation, Tyler tips his hat to this argument, then gives the standard response:
[M]ost fundamentally, growth rates are lower today than before 1973, no matter what exact numbers you settle on for the absolute living standard.  Even if the post-1973 era has a lot of unmeasured quality improvements, so does the pre-1973 era.
Perhaps sensing the tension between this standard response and the pessimistic thrust of his e-book*, Tyler immediately adds a much more original argument:
In fact, income measures are most likely to understate growth during times when a lot of new goods are introduced into the marketplace or made more widely available, such as during 1870-1973.  Thinking carefully about measurement biases probably means that earlier decades had even stronger growth, relative to what the diagram shows, compared to the post-1973 period.  It means that our recent relative performance is in reality even worse.
Earlier in the book, Tyler paved the way for this point by contrasting earlier life-altering inventions with subsequent petty tinkering:
We still drive cars, use refrigerators, and turn on the light switch, even if dimmers are more common these days... It makes my life only slightly better to have a larger refrigerator that makes ice in cubed or crushed form.
This admittedly sounds plausible at first.  But by and large, Tyler's wrong.  Merely bringing new inventions into existence typically has little economic benefit; the subsequent enhancement is where most of the value lies.  This is obviously true for the computer; I had a VIC20 in the early eighties that wasn't even a workable word processor.  But the same holds for older inventions.

Take the car.  The earliest internal combustion cars, introduced in the late 19th century, were just toys for millionaires.  It took decades of gradual improvements to make them a viable form of transportation, and decades more to make them reasonably reliable and safe.  The same goes for airplanes.  Thirty years after the Wright brothers, trains and ships dominated long-distance travel.  Planes were still too bumpy and dangerous for the general public to stomach.

Or take television.  The picture and sound quality were horrible for decades.  Even in the Eighties, I had to hold the antenna with my hand or foot just to watch cartoons.  But the deeper problem was the content - three or four channels, all insipid.  As a great cultural critic wrote back in 1998:
Post-war American television, by and large, has not provided cultural riches.  Television programs entertain us and present appealing characters, but a canonic list of the best television programs would not, in this author's opinion, stand up to a comparable list from music, painting, or literature... I concur with Robert Hughes, who notes that several hours of American television provide the best argument against market-supplied culture.
Fifty years after its introduction, television had yet to impress.  But look what's happened since 1998!  DVRs, HDTV, The Sopranos, Battlestar Galactica, Arrested Development, Dexter, and Big Love are just the beginning.  Gradual improvements in technology and story-telling finally won over even the aforementioned cultural critic, who of course is none other than Tyler Cowen himself.

The lesson: Contrary to Tyler, it is during periods of technological consolidation that official CPI numbers understate progress the most.  Flying home for the holidays or watching the latest episode of Glee isn't as awe-inspiring as hearing about the first flight or watching television for the first time.  But it's the cumulative improvements that have transformed mere ideas into concrete progress.

* What kind of a pessimist says, "While progress has been slowing down, it's been even faster than you think for over a century"?

Comments and Sharing

TRACKBACKS (1 to date)
TrackBack URL:
The author at Eli Dourado in a related article titled CPI Bias and Stagnation writes:
    Tyler has been defending his stagnation hypothesis with an intuitive argument about when the CPI is most skewed. I’m not 100% persuaded of Tyler’s intuition. There is a severe conceptual problem that plagues any measure of inflation. It see... [Tracked on February 7, 2011 9:31 AM]
COMMENTS (25 to date)
Elvin writes:

Another area that we ought to consider is the environment. My guess is that overall air and water quality in the US is much better now than in 1973 and that from 1900 to 1973 it became much worse.

What do the data say?

Doug writes:

The notion of trying to "objectively" measure anything like quality improvements is nuts. On this the Austrians are dead right. The reality is there is no "true" CPI.

Case in point, if I'm someone who really enjoys gold and silver jewelry, long car drives in a gas-guzzling truck, eating caviar and living in Manhattan then my inflation rate has been very high over the past 20 years. In contrast if I was someone who likes exchanging messages with their friends, playing video games, watching niche tv shows and movies, don't care about high status clothes or products (i.e. doesn't mind shopping at Wal-Mart) and lives in Iowa city then my inflation rate has been very negative (especially after adjusting for quality improvements).

The point is there's no such thing as a single "basket of goods" that apply to all consumers. And especially when you try to make "quality" adjustments the whole thing goes nuts. A gamer gets a lot more quality improvement from a faster processor than a casual web-browser (who in turn might get more from an iPad).

Matthew Gunn writes:

What a bunch of old world thinking! What about the Internet?

To take a concrete example, how would the CPI account for an innovation such as Wikipedia?

Do you say that encyclopedias, which used to cost a positive amount, now effectively cost zero? That would be a massive underestimate of the welfare gain from the existence of a Wikipedia.

And more generally, how do the national income statistics take into account the proliferation of free content? Wikipedia is created by volunteers and the content is given away for free. Wikipedia is powered by Mediawiki software, which is created by volunteers and given away for free. Pages are served by the Apache web server, which is created by volunteers and given away for free, from computers running the Ubuntu distribution of the Linux/GNU operating system, which is also created by volunteers and given away for free. Wikipedia, Linux, Apache, etc... are not going to show up directly in the GDP or national income data. A huge amount of the value add is going to flow through into consumer surplus.

The only way you're going to measure the welfare gain from free Internet stuff is if these innovations reduce the price level by a sufficient amount.

Charlie Deist writes:

"I concur with Robert Hughes, who notes that several hours of American television provide the best argument against market-supplied culture."

Lowbrow television is a great boon for most "folks," in the sense that Albert Jay Nock used the term. He describes how Gresham's law was beginning to take effect in the early 20th century in literature, with pulp and garbage replacing magazines and books of real intellectual content. The same happened with television vis-a-vis more cultured entertainment, but probably to the benefit of the common man, who wasn't going to be enjoying such high-brow culture anyway. I score this one in favor of market-supplied culture.

Matthew Gunn writes:

Also, as the economy switched over the 20th century from household production to marketplace production (eg. child care etc...), the national statistics may have overstated income growth.

ajb writes:

Since I think both top movies and popular music have actually gotten worse on average than in the 1970s the improvement in TV quality doesn't seem to be a strong enough offset. With things like this, personal opinion is going to be a strong determinant of what constitutes "progress." I still haven't heard anything convincing to refute Tyler's point that things have gotten better but much more slowly since the 1970s. And I know lots of people who don't care about either the internet or the cell phone, and who happily drive cars from the 1970s. Yet I hardly know anyone who would do without antibiotics, telephones, or airplane travel (to compare advancements from 1935 to 1973 vs 1973 to 2011). I just don't see how it's possible to believe that the changes experienced by the average American (especially non-big city Americans in the 1930s) wasn't greater prior to the 70s comapared to the decades since then.

Tyler Cowen writes:

Your point is a good one, but if you look at the enhancement of say autos (your example, and a good one), and the spread of autos, or the other major gains of the 20th century, that is precisely what makes the pre-1973 period so wonderful for growth. On TV, you suddenly switch away from the median and to the Infovore. The median loved standard network TV, and its spread to more or less everyone was again pre-1973.

Steve Miller writes:

Tyler is still being pretty stubborn. If you look at Cox and Alm, it's pretty clear that car ownership increased dramatically post-1973, so the penetration continued. But more importantly, the quality, reliability, features, safety, and overall peace of mind has increased dramatically. In fact, I would say car quality is the perfect example of how standards of living have improved at an increasing rate. My minivan is WAY better than Don Draper's Cadillac.

Why? My best guess is a dramatic increase in global competition, especially in the auto market. I'm not just saying a Camry is better than a Dodge Dart. I'm saying it's WORLDS better. The differences are most dramatic at the lower end of the market. Take an AMC Gremlin versus a Nissan Versys, for example?

Joey Donuts writes:

20 years ago Professor Negroponte originated the meme
"Negroponte Switch". Basically a switch from using wires to move data and voice to wireless transmission and television from wireless to wired transmission. A switch that is still on going.

The implication of this switch, which he discussed in a Harvard Business Review article in 1990 or 1991 (I am recalling this from memory). One of the important implications he discussed in that article was a move to "narrow casting" from broadcasting. This implies more consumer choice and "better" at least more taylored content. This improvement isn't measured by any standard measurement of well being. Yet all of us experience it every day.

The phenomenon of better TV discussed by some posters above is only one aspect of the switch. The rise of the cell phone is another, let alone iPhones, iPads, and who knows what devices yet to come.

Pat writes:

Cowen wants to be the next Tom Friedman. Friedman's big seller had an incorrect "AHA!" title. "The World is Flat" should've been "The World is Small", which isn't an original insight.

Likewise, "The Great Stagnation" should be "The Great Growing a Little Slower Than Before If We Crudely Categorize Everything and Ignore Cool New Stuff". Slower growth for maturing economies is hardly a new insight either.

His yearlong courtship of Ezra Klein led to some pretty nice exposure for the book. Good for him.

Ak Mike writes:

Steve Miller - the question is not whether a Camry is better than a Dart. The question is whether the distance from a Dart to a Camry is greater than the distance from a Ford Model A to a Dart. It is not.

ajb writes:

@Ak Mike -- Very well said!

ajb writes:

I would add one more thing re: transportation.

It is surely true that if even 2-3 percent of flyers could routinely get from San Francisco or LA to Tokyo or Beijing in under 6 hours (**including check in security times!**) for the price of a first class ticket the nature of business would be greatly altered.

Yet that would be a more modest change from the early 70s relative to the shift from 1930s aviation to the 1970s.

Most of the examples people give are about the democratization of early century innovation (planes, cars, phones) not about new innovation.

This matches the pattern of growth. Slowing/stagnation in the USA and other leaders, catch up and rapid growth in the rest of the world.

MikeDC writes:

AK Mike
Perhaps not, but the distance between 1970ish and 2010ish products is pretty freaking huge.
* Cancer treatment, heart care, and a variety of other medical services that both give us more time and better time. For ourselves and with our loved ones.
* Pre-1973 economic reality forced many folks into social arrangements that people now frequently get themselves out of. This is a tremendous improvement in the quality of life for folks at the median and below the income distribution, and it's a direct result of the increased economic opportunities prevalent in the post 73 world.
* Reproductive freedom certainly has a major economic component to it.
* Consumptive opportunities, across the board,are dramatically improved. A more accurate "car question" would be whether the distance from a family sitting at home while Dad drives a Dart to work (73) vs. a family with two Camrys, where mom and the kids can come and go as they please while Dad drives one to work (10) is greater than the distance from a situation in which dad drives a Ford Model A or a Dart.

It's a huge difference. This is one area where I think we men (because I'm not going out on a limb to note that we're predominantly men contributing to this argument) might be missing something big.

Ask not whether you'd be better off in 1973, ask whether your daughter would. If it takes you more than a split second to answer that question, there's something really, really wrong with you.

Badger writes:

Many of the arguments against Tyler sound to me like grasping at straws.
Quality of the median TV program today is definitely not better than it was 30 to 40 years ago.
Traveling 1000 miles by car today is clearly not much less gruesome than it was 30 to 40 years ago.
Playing games or going to the movie theater today is not much more fun than it used to be 30 or 40 years ago (where are all those really cool and fun pinball machines?)
Much of the technical development during the last 40 years is "cool" engineering stuff, but when the push comes to the shove, it's gadgetry, not revolutionary.
For revolutionary, better look at what the Lumiere brothers were doing a century ago.

Steve Miller writes:

Wait, a quibble:I went back to 1973, and named the Dodge Dart. That means the Model B is a better comparison if you go back 38 years, although that would technically be the 1935 Model 48. I'll compromise and go with the Model B.

Is the Dart, er, more better than a Model B than a Camry is than a Dart? (That's poorly worded, but you know what I mean.)

I don't see how. The Camry is much, much, much better than the Dart. Modern tires. ABS disc brakes. Modern transmission. Paint that doesn't fade, and the body has an uncanny resistance to rust. Overhead cams. Trip computers. Peace of mind. That's a few things very quickly off the top of my head. How many miles do you think I would get out of a Dart without a rebuild? What kind of maintenance would it need, and how often? You do realize that you could probably drive a Camry 8 or 10K miles between oil changes, buy new tires and change the coolant every 40 or 50K miles, for a grand total of 250K miles and still have a better car than a new Dart. Is the Dart that much better than the Model B? I doubt it, but maybe. Hard to say, with value being subjective and all.

Chris T writes:

Most of the examples people give are about the democratization of early century innovation (planes, cars, phones) not about new innovation.

My cellphone fits in my pocket, can receive and send calls anywhere, can access the internet anywhere, has more computing power than the Apollo spacecraft, does not require a human operator to direct calls, and can do a wide variety of other tasks, none of which were possible in the 1950s, but represents absolutely no new innovation since then?


Seth writes:

What does the post 1973 growth look like as a percent of income in 1870?

That might give a different view of the progress made since '73.

Comparing percent changes from two different starting points has its merits. But, I think sometimes its helpful to also use the same starting point for more of an apples-to-apples comparison.

After all, my percent change in reading ability probably hasn't increased nearly as much since earning my high school diploma as it did from the K to 12th grade. But, I'm not sure I'd characterize it as a stagnation in my reading abilities.

Yancey Ward writes:
Quality of the median TV program today is definitely not better than it was 30 to 40 years ago.

Well, 30 to 40 years ago, the median television program at any given time is whatever was on NBC.

ajb writes:

Chris T is being deliberately obtuse, as are many commentators. It is NOT that cell phones are not innovative. It's that they don't represent the same giant jump in productivity that matches going from no planes to airplanes, and from biplanes to jet planes. Nor even from no telephones to widely available land lines.

That many/most people in the US today STILL won't pay a premium of a few hundred dollars to upgrade from a basic talk/text only cell phone (if they have one at all) to a smartphone with good web service is evidence of how UN-essential most of the cell phone innovations are.

Since this is an econ blog, this seems a good measure of marginal consumer surplus. This means the marginal phone user is unwilling to spend as little as $200 a year (phone plus pay as you go service) to have an "advanced" phone plus some web surfing. Which means the gains from cell phones to many middle class consumers is small.

I wonder if the same could be said for ALL antibiotics or ALL plane travel?

Chris T writes:

It's that they don't represent the same giant jump in productivity that matches going from no planes to airplanes, and from biplanes to jet planes. Nor even from no telephones to widely available land lines.

It actually took about 60 years before planes were a factor in most peoples' lives (unless you were unfortunate enough to be bombed by one).

The phone took about 35 years to reach 25% of U.S. homes. Since a phone's utility is directly proportional to the number of people who have one, productivity growth wouldn't exactly have been remarkable.


The cell phone went from 12.4 million users in 1990 to 4.6 billion in 2009. The first commercial cell phone was released in 1983; so in 26 years the cell phone went from commercialization to 68% of the world's population while the telephone took 35 to reach 25% of the population of the United States. And you want to talk about productivity growth?

That many/most people in the US today STILL won't pay a premium of a few hundred dollars to upgrade from a basic talk/text only cell phone (if they have one at all) to a smartphone with good web service is evidence of how UN-essential most of the cell phone innovations are.

Good lord, the first official smart phone was released about ten years ago and you're ready to pronounce it unimportant? 15% of the population of the US already uses one.

agnostic writes:

"The Camry is much, much, much better than the Dart."

Well that Toyota Camry won't get you laid or married, while the Dodge Dart will, so from a Darwinian accounting, the Dart destroys that dorky little Camry.

agnostic writes:

Quality improvements in the post-1973 era include computers with more memory, phones that went cordless and then mobile, etc.

Quality improvements circa 1873 meant a railroad car that only caught one acre of farmland on fire, rather than half of the country, from shooting sparks.

Same with mid-20th C. transportation. The Statistical Abstract of the U.S. used to keep records of railroad accidents and deaths, and also automobile-related deaths. I don't recall exactly what these look like, but I do remember that the dramatic decrease took place somewhere between 1920 and 1960 -- not too sure.

Earlier quality improvements were real big and important.

mick writes:

Notice how the CPI is lower than the increase in food, education, medicine, metal, oil and the other vital components of purchasing power? It's mindblowing how niave people are about this.

Babinich writes:

Steve asks: "How many miles do you think I would get out of a Dart without a rebuild?"

The Dodge Dart w/ 225 slant six had the much better engine. As a matter of fact the transmission paired with the 225 was outstanding. :')

I'd also like to see statistics on the durability, or lifespan, of today's durable goods: TVs, refrigerators, washers, dryers, etc...

Comments for this entry have been closed
Return to top