How much of the value of the internet is not captured in gdp?

by on May 27, 2015 at 1:33 am in Economics, Web/Tech | Permalink

I have been hearing this question more and more lately, even in China.  Overall I think it has gone from an underrated effect to an overrated effect.  Tim Worstall offers an introduction to this debate.

Let’s not forget that you do in fact pay for Facebook access, indirectly, when you pay for your cable connection, your iPad, and your smart phone. including the monthly bill, all of which are part of measured gdp.  The more value Facebook brings you, the more you would be willing to pay for these goods and services.  The same is true for Google and the like.  So Facebook and other internet services are part of a bundled package of market value, but that is very different from claiming they are not measured in gdp at all.

There is of course consumer surplus from the internet and Facebook, just as there is from Dunkin’ Donuts.  Might that consumer surplus be especially high?  Well, we don’t know, but don’t assume it will be.  I did some casual googling, and found a number of estimates suggesting that smart phone demand is relatively price elastic, with the iPhone a possible exception to that regularity.  That implies consumer surplus isn’t especially high, because many people aren’t willing to buy at the higher price.  I thus think Brad DeLong is far too optimistic in his estimates of ratio consumer surplus to market price.

You also could look at the literature on the demand for cable internet services.  The results are mixed, but again I don’t see a strong case for a disproportionately high consumer surplus from these services, if anything the contrary.

Now maybe these estimates are wrong, or looking at the wrong margin in some way, but the fact that I hear them mentioned so rarely gives me pause.  Cowen’s Third Law.

There is also advertising over the internet.  Let’s say Facebook is a profit maximizer.  Insofar as Facebook is of value to consumers, the company can get away with putting a lot of ads on the site.  These will spur additional market purchases, and so part of the value of the site is again captured in gdp.  Obviously some of these ad effects are simply expenditure-switching, and so there is no full capture of value, but still Facebook shows up in gdp statistics in yet another way.

Here are some previous posts on this topic.

1 Michael May 27, 2015 at 1:53 am

Cowen’s Third Law–finally something I can agree with you on, Professor!

2 Steve Sailer May 27, 2015 at 2:02 am

Has anybody figured out how much GPS is worth by people not getting lost so much?

Here’s a question: does GPS make a convoluted Renaissance city like Boston more attractive relative to a rationalized easy-to-navigate city like Salt Lake City?

3 Doug May 27, 2015 at 3:49 am

I’d definitely say the effect is real. Even within cities. How much of the downtown Manhattan real-estate boom is driven by Google Maps making South of 14th street easily navigable. I have a theory that grids are naturally bad for neighborhoods. The large avenues naturally limit any potential micro-locales to no more than a few straight blocks. Cities like London or Boston, allow for many different niches and hidden away gems. It allows for both better hyper-localization (“these seven square blocks are known for their stationary stores or Laotian food”), as well as makes aimless strolls much more pleasant and less oppressive.

4 Jeff R. May 27, 2015 at 9:48 am

Yeah, GPS has helped me navigate to niche retailers I might not be bothered to try to find, otherwise. But then, the resulting sales figures would show up in GDP stats.

5 Marcos May 27, 2015 at 4:50 pm

Not so fast. If you didn’t find those niche stores, you’d have replaced the items you brought with imperfect substitutes. The difference in price between what you brought an what you’d have brought is captured by the GDP, but it’s not always big, and odds are it’s even negative. The difference in value you’ve got from the items you brought to the items you didn’t isn’t captured at all, and this it the one that’s normally big, and never negative.

As I said last week, this kind of thing was always a problem for the GDP. But almost the entire value proposition of IT is creating those situations, and neither IT nor the niche merchants can capture most of the value, thus none of that is on the GDP.

6 Thomas May 27, 2015 at 2:05 am

People are looking in the wrong direction. We’re mis-measuring productivity improvements, not consumer surplus. Check your traffic stats during working hours.

7 mulp May 27, 2015 at 2:53 pm

Tyler is refusing to explain that GDP is a bogus measure of anything relative to human productivity or well being.

Picking say America in year 1000, how can you argue GDP was greater than zero?

No goods were bought and sold, but were shared and perhaps traded. There was no hiring of workers so no pay with which to buy goods for sale.

When talking of GDP in much of Africa, one might assume the people must be starving and totally lacking in anything produced by industry in the past two hundred years. Instead, you find they have cell phones and solar panels and trade internationally. But 90% of their living is not part of GDP. Hunter-gatherers-farmers do not get tallied in GDP. Mothers produce capital assets for society without it being counted in GDP: workers. (One will not call a hunting cabin you inherit that was built by your grandfather with his own hands by cutting trees, sawing them, and so on, a non-capital asset simply because its construction was not included in GDP. Yet you can rent it and earn money from it, just like you can from the gift of life and upbringing by your mother.)

If you value things being cheaper and cheaper, then you need to value declining GDP as well. If cheaper goods from robots and cheaper food from technology, and cheaper energy and cheaper clothes from imports and cheaper housing by infringing on others (blocking their views, taxing them out of the block based on their single house 2000 sq-ft being more productive as a 1000 unit highrise condo at 500 sq-ft each, then you are arguing for lower GDP. If cheaper is better then lower GDP is better because cheaper goods means lower incomes.

8 Larry Siegel May 27, 2015 at 11:58 pm

Economists, or at least that subset of economists that care about economic history, have ways of adjusting GDP numbers to reflect the value of subsistence living, barter, trade without money, etc.

9 carlospln May 27, 2015 at 2:48 am

How can there be ‘consumer surplus’ from confectionery made from refined white flour, sugar & grease [trans fats]?

Only in America.

10 Doug May 27, 2015 at 3:56 am

1) Dunkin Donuts doesn’t use trans fat in the vast majority of their items. Grease and trans-fats are two very different things.

2) Europeans eat as much white flour and sugar as Americans if not more. At the very least the French certainly do, and they derive as much consumer surplus if not more. The primary difference is that American meat consumption is much higher. This goes to why American obesity rates are relatively overestimated on a BMI basis. Americans have the highest lean body mass and upper body strength of any major first world citizenry.

11 The Engineer May 27, 2015 at 6:05 am

Well, we do have Michelle Obama. Girl Gots Guns.

12 Axa May 27, 2015 at 6:16 am

Consumer surplus is the difference between market price and the price you will and are able to pay for something. Ironically, experiments in sugar tax that lead to no decrease in sugar consumption, indicate that there’s a lot of consumer surplus in sweet & greasy products 😉

If you dislike the “consumer surplus” definition, you’re free go ahead and invent a new concept of consumer welfare. Just name it different.

13 Urso May 27, 2015 at 11:11 am

Yes, Frenchmen and Spaniards are famous for avoiding pastries. As Don King would say, only in America!

14 Cooper May 27, 2015 at 1:51 pm

Americans consume 2.5 kilograms of butter per capita, per year.

Germans consume 6.2 kilograms.

The French lead the world with 7.9 kilograms

If eating fatty foods is bad for you, the French should be far less healthy than the Americans. That doesn’t appear to be the case.

15 Nathan W May 27, 2015 at 2:29 pm

This can probably be explained by margarine more than anything else. I’ve never met a French person who wouldn’t go to some length to get butter over margarine. Also, Americans eat a lot more fried food, which leaves less opportunity to consume butter.

16 Steve Sailer May 27, 2015 at 4:22 am

“Americans have the highest lean body mass and upper body strength of any major first world citizenry.”

So I’m betting on America in the next world war (assuming it’s fought solely through Greco-Roman wrestling).

17 Morgan Warstler May 27, 2015 at 6:31 am

Tyler you are wrong.

Period. The End. Please stop pushing your tired old man noise on us.

Horse Buggy -> Car

Boat -> Train -> Plane

Now invent teleportation. See we invented it, it is called VR. EVERYONE eventually ends up in a VR rig sucking down only calories and electrons (the Matrix).

The path goes from here to there, so everyday is one step closer to no human physical activity.

So we see clearly, there is NO way to measure GDP, as we wind down from 100% atomic to none, how many virtual logs were produced in how many milliseconds in your virtual mansion? How many virtual props, virtual actors, in your virtual movie?

The only way to get s accurate measure is a New Economic Metric – a top line number that makes tech guys look like gods. Watch:

Alex Gaudino – Destination Calabria music video uses virtual set to generate a music video that is human actors would cost how much?

The consumer surplus of this music video is what atomic version would cost using humans minus the real cost.

Or as a topline Tech Guys are gods consumer surplus metric how about using the luxury atomic price minus the current price, so a 60* TV at $50K is now $500 – consumer surplus $49,500 on each TV sold.

The problem gents is that this reality isn’t going away. No stagnation story will ever survive as we wind down to your VR pod, so start putting the scaffolding around some kind of wild metric that throws our 2000 years of atomic economic theory and remakes it in digital end of scarcity.

18 Mesa May 27, 2015 at 10:40 am

GDP measures nothing, ultimately.

But economists love it, much like a cat loves chasing a laser spotlight around the floor; the illusion always fascinates them as some perceived slice of reality.

19 mulp May 27, 2015 at 3:25 pm

No, GDP measures what workers are paid for working for someone else.

Free lunch economics, what HW called voodoo economics, what Reagan sold America, holds that paying workers is a great burden on growing the economy based on GDP as the measure.

So, with falling wages for large sectors of work and overall stagnant wages, how has growth been accomplished: by borrow and spend. Some borrowing is effectively a tax, which is the price of civilization, in that government debt became permanent with Reagan as a means of funding consumption. Other borrowing is wealth redistribution – Donald Trump borrows millions, gets over extended and goes to court to get the wealth of workers who saved given to him in debt forgiveness – banks hide the loss from workers by charging workers high fees and interest on worker debt to provide the cash they give to people like Trump.

But the endless debt that everyone assumed would go on forever was cynically part of Starve the Beast, but the “beast” is GDP, not just government which is simply added to GDP based on government spending being equal to consumers buying food or cars. The private sector has relied on both government and private debt to boost GDP beyond what workers are paid. But by using debt to fund consumption, capital assets – roads and bridges, water and sewer, as well as the US industrial capacity in both factories, modern technology, and skilled factory workers – has been pillaged and plundered in the quest for cheaper goods through lower wages BUT higher GDP.

With free lunch economics, the “value” of something is the “price” derived from creating monopoly scarcity of things to drive up their price or the rents that input price. Rather than having uncopyrightable folk songs sung by a million performers, the model is to have a hundred singers who have monopoly by copyright who sing a few times for really high prices, with the “industry” coming up with all sorts of ways to block most other performers.

To allow every bar to hire bands to play from a large body of popular music that was new 20 years ago when written and is now freely performed and interpreted today, free lunch economists argued that by extending copyright the value of new songs will go up as more money is collected for a century and that will produce more music and art than ever. Instead it has produced less and artists pretty universally believe they have been screwed. They and many others think that performing once should be compensated as if they performed a million times because the recording being played is equivalent to the artist working. Thus all the unpaid listens is theft of the artist’s pay for performing the work and thus a loss to GDP. bs

20 Mike Sax May 27, 2015 at 3:45 pm

In picking sides here, I’m definitely closer to Morgan than Tyler. I remember the 80s-prior to cable tv. and the Internet and life was very boring. Unless you were a social butterfly which as kid I wasn’t you didn’t have much to do.

When thinking about the Internet that makes this entire conversation possible, it’s tough to know where to start in qualifying the improvements of what we had before. I’ve always been a reader, but going to an old brick and mortar library today is a field trip to a quaint museum compared to the intellectual experience of being on the Net. With Amazon you can read most any book you could want with a couple of clicks of the keyboard.

Now I don’t know how you quantify this in the terms that economists prefer to use but I would say that ‘quality of life’ has improved for me and millions of others exponentially-25% of all marriages now start on the Net.

However the trouble is that it’s also made so many jobs obsolete. We’ve become a nation of burger flippers and the Internet is a big part of it, as much as I love it.

21 Hmm May 27, 2015 at 6:40 am

What if Facebook actually delivers negative utility to people by being a distraction bomb though?

If crack cocaine grew on trees and everyone used it constantly, would we ask whether it was generating value not captured in our GDP?

I’ll bet if you polled people and asked if they wanted to spend a) less b) same c) more time on Facebook, you’d get a) more than c). (I expect you’d get b) for most goods.)

22 A Definite Beta Guy May 27, 2015 at 9:59 am

Revealed Preference FTW

23 Nathan W May 27, 2015 at 2:30 pm

Perhaps in some cases. But I’m generally inclined to think that he might be on to something in quite a large number of cases.

24 Cooper May 27, 2015 at 1:55 pm

I’ll just leave this here for you

25 Bob W May 31, 2015 at 10:31 am

You might consider adding D as a choice. D being more cocaine.

26 rayward May 27, 2015 at 6:48 am

Overestimating the value of the internet derives from an overestimation of what the internet is used for. No, it’s not used primarily to access great works of literature or to conduct research in pursuit of a cure for cancer, it’s used to shop; and Google, Facebook, and the rest aren’t futuristic tech companies but platforms for advertising, advertising simple products made in the world of atoms. That’s not to say that the internet isn’t valuable, it’s just to recognize that it’s use by the vast majority of people is essentially as an updated version of a newspaper, magazine, or television. What about airbnb and uber and vrbo? The internet makes them more efficient, not possible, as anyone could have advertised a room for rent or a car for rent or a house for rent in the newspaper or a magazine; indeed, the internet has captured much of the business that at one time was the backbone of newspapers and magazines, weakening newspapers and magazines and, arguably, weakening “community” and democracy. Not only that, but airbnb and uber and vrbo succeed by flaunting local zoning and licensing laws, disrupting stable businesses and neighborhoods; single family residential neighborhoods in the coastal area of the Florida panhandle have been converted to de facto commercial areas, and war zones between owners who expected a quiet residential neighborhood and non-residents who wish to cash in on the chaos. For those of a libertarian persuasion, the disruptive influence of the internet must be a welcome development and must have “value” far above what’s included in gdp. Or not, depending on whether the libertarian suffers the consequences of the disruption.

27 Alain May 27, 2015 at 11:05 am

a) LOL
b) if there is no surplus on the ‘Internet for shopping’ why are you *posting* on an economics website?

28 Urso May 27, 2015 at 11:13 am

I agree entirely. I used to subscribe to the local newspaper, now I read the NY Times online. Is that better? Maybe not, if I’m less informed about local events.

Ah, but you can still read the local paper online as well, you cry. Sure, except they fired over 50% of their reporting staff and have replaced them with fresh out of college bloggers and vloggers in a cost cutting measure, and their online site is a train wreck anyway.

29 Nemo May 27, 2015 at 7:32 am

Having been a software developer for 30 years, I would say that my entire year’s internet costs are easily saved in a single day – maybe in a single hour – when calculated against the time wasted looking up the answer to some problem that can now often be trivially found in a quick search.

30 The Engineer May 27, 2015 at 8:25 am

We have reached “peak search”.

Where do those answers lie? Generally in forums of some sort, perhaps in a blog post.

Both forums and blogs are being decimated by social media, which is not searchable and doesn’t serve the same purpose as forums do.

Facebook and Twitter are dumbing down the internet as a result.

31 Lord Action May 27, 2015 at 9:44 am

A lot of people have moved to Twitter as the vehicle for getting technical questions answered. It does require some effort to remain connected to the right people.

32 oli May 27, 2015 at 10:34 am

Stackoverflow is the most popular online forum for programming Q & A. It’s answers are close to exhaustive and of consistently high quality. It’s less than 8 years old, growing fast and (as far as I can tell) faces no threat from social media. Google’s ability to take you directly to the relevant paragraph in online documentation and being able to see other developer’s code on github are both serious time-savers. I’d say access to google + SO + docs + github is worth a good $10K + to most professional software developers before you even get into the open-source languages and libraries we use.

33 Steve J May 27, 2015 at 12:04 pm

It is hard to estimate how much the internet improves software development productivity. Like you I would guess my yearly internet usage fees are often “paid for” by the time I eat lunch.

34 AB May 27, 2015 at 7:38 am

But most smart phones are bought by people who already have some sort of internet access, usually on an existing, no-longer-quite-state-of-the-art smart phone. So price elasticity doesn’t capture much about the value of internet access.

35 Patrick May 27, 2015 at 8:50 am

I wanted to make this point. At this stage, smart phone demand is about the consumer surplus to new smartphone features, not about the internet, except to the extent that one of those new smartphone features is commonly faster internet access.

36 Robert Simmons May 27, 2015 at 9:00 am

Also, he’s looking at price elasticity for individual smartphones, not for the category. That people have options within Android makes us more price sensitive when Samsung tries to price at a premium, but we still buy some kind of smartphone.

37 ladderff May 27, 2015 at 7:46 am

I used to stick up for GDP as better-than-nothing. No longer.

38 derek May 27, 2015 at 8:09 am

In my small business the price each month for communications has been around $500 per month since I’ve been doing it. When I started, it was a landline and one cell phone, with a fax line. There wasn’t internet. Now I have no landline, four phones, a g4 modem for the office for internet for all the desktop machines, and cable internet at another location. We use a web based accounting and ticket tracking system that we can all access via mobile devices or desktop. Google email and their whole suite. Software costs are another $150 or so per month.

On a percentage of sales it is quite reasonable and easily scalable. Hardware is about a three year cycle on average.

So what are they worried about? Too cheap? No value? In my four man shop I have access to software solutions that a decade ago cost millions to implement and are only done by large mobile service operations. Mapping is really handy, including street view for dispatch. A movie camera in effect hand is very nice, I cursed many times sticking my head into small holes to try to read some numbers through bifocals, a quick photo is accurate and quick, easy to send off to a vendor.

There is no replacement for yellow pages yet. Google could do it but they need work.

39 derek May 27, 2015 at 8:10 am

Every hand.

The great stagnation in comment editing continues.

40 Mesa May 27, 2015 at 10:46 am

Concise communication is an acquired skill, typically untaught in formal education.

41 JG May 27, 2015 at 2:33 pm

No replacement for the yellow pages? I haven’t used the yellow pages in over 15 years, I found a replacement, it’s just not in individual things.

42 Lee A. Arnold May 27, 2015 at 8:11 am

Tyler: “How much of the value of the internet is not captured in GDP?”

Consider my perfectly analogous question:

Lee: “How much of the value of the hammer and saw out in my toolshed is not captured in GDP?”

They are not consumer surplus, they are personal capital to do things. (Not “human capital”, which is defined differently.)

Refine this, for the internet/telecoms: here it is BOTH. Sometimes it is consumer surplus (watching movies, contacting friends) and sometimes it is personal capital to perform work which is nonmonetized, as well as work which IS monetized (e.g. people use Facebook to conduct business for the GDP).

The consumer surplus in a smartphone is mostly in the communication, which is already provided for most people at low price. iPhones are for beta-testers, trendsetters, status seekers.

Tasks no longer monetized? The do-it-yourselfers and handy types can perform home-improvement work without entering into the GDP (except to buy repair parts, etc.) I hypothesize that there may have been a recent surge in this activity, to save money in desperate straits, because the Great Recession followed the financial crash AT JUST THE SAME MOMENT when things like detailed instructional how-to’s (e.g. YouTube videos) have come on line. (You can even contact some of these instructors, e.g. via YouTube comments, and ask them remarkably detailed technical questions.) If you’re in a pinch, it’s the way to go. Someone really ought to do a thorough study of the economic effects of on-line DIY (I’ll bet you that Google is way ahead of you.)

To earn incomes for that which still requires money (to buy food, pay the rent and transportation, buy computer/phone, buy repair parts, etc.) people must enter the workforce. But more and more people (both labor and small entrepreneurs) can only find occupations which do not admit of much productivity growth (hands-on, personal services, etc.) and therefore do not admit of much incomes growth. You can work your butt off, and not get anywhere. Which everyone pretty much observes, at this point.

You can also start-up an online business, and gains incomes growth, but the massive competition to any idea makes it head in the direction of a perfect market, i.e. zero rents. Indeed, intellectual property rights are likely to be shortened, because many of these ideas could occur to anyone (one-click shopping? really!?) and priority rights will seem increasingly unfair.

Anyway, the puzzle of the productivity slowdown may have a tripartite solution: 1/3 because after you have consumer satisfaction and comfort, you then become limited by your available time; 1/3 of it because infotech allows continuous nonmonetized satisfaction, repair, and improvements; and 1/3 of it because a larger portion of available employment does not allow of productivity increase out there in the market, and is thus a drag on the (monetized) GDP.

Cheer up. It’s a good thing. We go to a ten-hour work week, and everybody can get really cookin’!

(This is just about what half the 19th-century thinkers expected to happen, though it was delayed by a century of wars and technological advance.)

However, that leaves us, at present, with the problems of the very, very few things which are currently increasing in price because of EXOGENOUS factors: 1. because they are TRULY scarce (e.g. desirable real estate at the beach) or 2. because technical advance has NOT reduced costs in humanly-necessary sectors (e.g. primarily, healthcare and medicine).

So in these exogenous items, I think we’ll see a transition period: Desirable real estate slowly ends up in state parks, or in timeshares or lotteries. And I’ll bet that the costs of healthcare are about to start coming down, because biotech-nanotech-genomics-masscomputation is the right tool for the job, after millennia of drifting and chance. Until then, a gov’t monopsony could fund the private medical sector, and even print the money as necessary.

43 The Engineer May 27, 2015 at 8:27 am

It’s weird for an economist to say that my cable internet is capturing the value created by Facebook. That’s like saying that the interstates are capturing the value of interstate trucking.

I guess we’ll see. With all these cable mergers, and with cable TV going the way of the compact disc, there will be pressure on cable companies to capture more of that value through price increases on broadband.

44 Patrick May 27, 2015 at 8:51 am

It isn’t weird at all, is it easier if you replace “interstates” with “toll roads”?

45 Lord Action May 27, 2015 at 9:43 am

Easier, for sure. But obviously false, right?

46 Nathan W May 27, 2015 at 2:42 pm

If the tolls were higher than the price of maintaining the road, then indeed the toll company would be capturing some of the value provided by the trucking service.

47 Los Ranchos May 27, 2015 at 9:18 am

If the consumer surplus was really large you would see subscription or other fee based models for Google, Facebook, etc. You don’t.

48 Steve J May 27, 2015 at 12:13 pm

Search/social charge fees? It doesn’t seem possible when there will always be free alternatives.

49 Cooper May 27, 2015 at 1:57 pm

Not necessarily. Facebook and Google make mountains of cash from advertising. Why would they want to risk that revenue stream by introducing a new set of fees that potentially opens the door to competitors?

50 JG May 27, 2015 at 2:41 pm

A) there are subscription fees, but they’re not paid directly by the end users (advertisers, promoters, app developers).
B) There are network effects. Let’s say 80% of facebook users receive relatively little consumer surplus, but the remaining 20% receive a lot of surplus. It may be difficult for facebook to charge the second set of users without also charging the first set of users for their consumer surplus. If facebook discourages marginal users they may also drastically decrease the value of the service to “power users”.

If you replace facebook with linkedin and imagine the first group is most currently employed linkedin members and the second group is recruiters and firms looking to hire you can see why this example is tricky. Now recruiters do pay a lot to linkedin so you could argue they’re capturing it, but there’s also a group of people looking for work they don’t charge that might get lots of value because if you only had people looking for work then it would reduce the value to recruiters and also decrease the long term value to “happily employed” users (even if they were expected to only pay later).

Single pricing models are difficult (or even tiered pricing) when there are legitimate network effects and a long tail of users that help provide the value of those network effects.

51 Nathan W May 27, 2015 at 2:43 pm

That’s because the marginal activity costs very close to zero and competitors could come up.

When Google started charging you $0.0001 per search, how long would it take how many users to switch to Yahoo, etc.? That’s why they stick with the ad based revenue models.

52 Los Ranchos May 27, 2015 at 9:30 am

Above is not true if those services are easily replaced by other competitors with low marginal costs of production.

More generally, can you have a service or good with a really low marginal cost of production and a giant consumer surplus survive in that state? Clearly consumers would demand more of that good until the consumer surplus went to zero for the last increment of consumption. So maybe the story is that there is a large consumer surplus, but each consumer sets his consumption so that his marginal surplus for the last unit of consumption is zero. Remember we have fixed prices for internet access and zero cost for many web services here. In this type of market once paid up, the consumer keeps consuming until his marginal utility is zero. Since his marginal costs are zero, he consumes until his lost time is a break-even trade for utility. Would it be possible for some people to consume this good 24 hours a day? Some may approach that.

Would charging per bandwidth usage changes these markets a lot?

53 o. nate May 27, 2015 at 12:35 pm

Excellent comment. Best analysis of this puzzle that I’ve seen.

54 Alex Godofsky May 27, 2015 at 9:43 am

There is of course consumer surplus from the internet and Facebook, just as there is from Dunkin’ Donuts. Might that consumer surplus be especially high? Well, we don’t know, but don’t assume it will be.

Yes, we should believe it’s especially high. We should believe so because while the price of Internet access hasn’t increased substantially in 30 years, the actual stuff available on the internet has absolutely exploded.

55 Morgan Warstler May 27, 2015 at 9:48 am

Another way to see Tyler wrong is the coming of ad supported Internet access. Facebook is literally providing the access for free.

Alex, tell Tyler he needs to admit he’s wrong.

56 Cooper May 27, 2015 at 1:59 pm

My local FM radio stations are ad supported. They seem to do alright.

Why shouldn’t the same be true of Facebook?

57 Floccina May 27, 2015 at 10:46 am

People buy less music (and some other things, maps) now that is an indication of the value of the internet.

58 Nathan Taylor May 27, 2015 at 10:51 am

I agree “Overall I think it has gone from an underrated effect to an overrated effect.” But I believe this is really is a case of Amara’s law: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” Short term tech gets hyped, and by the time it really happens we’re bored.

If we just project out another decade or two, it’s pretty clear the marginal cost and value of physical goods will continue to their current relative decline compared to virtual digital goods. And as this continues it’s not clear how we’ll calculate the human value of digital goods poorly captured by existing GDP. Implied value by attention? Somehow Gary Becker’s insights on the allocation of time have to come into play here since in a digital economy goods have zero marginal cost. Not clear how this will work, but I think the theorists have their work cut out for them. And to that extent this problem is not overhyped at all. It’s overestimated now, but we should be preparing better economic measurement tools today.

59 KevinH May 27, 2015 at 11:18 am

The price-elasticity of smart phones wouldn’t matter for public goods.

I think the really hard part about the internet is that it is messing with the basic metric of modern economics, transfer of money. Basically, economics is a way to measure vast interconnected groups of people. To measure the impacts of war and peace, to understand the human connections that help build a pencil, you used to be able to just follow the money. The internet allows for a mostly money-less form of very broad and very deep human connections.

I’d say a better metric would be time-elasticity factoring in wages. Do people spend more or less unpaid time on the internet as their wages go up? That might get you at an opportunity cost model of it’s value. Probably someone has already thought of this, but I’d love a link if anyone know about one.

60 Biil Harshaw May 27, 2015 at 12:17 pm

The Post today ( reports that parking tickets in DC are way down. Why, because of smart meters and apps which find vacant parking spaces for you. How is that going to be captured in any economic statistics?

“Smartphones and new technology have made it easier to pay for parking in the District, and the result has been a decline in the number of tickets issued, an analysis shows.

In just four years, the number of parking tickets issued in the District has dropped by 300,000, according to city records. More than half of drivers who use on-street parking are estimated to be using the Parkmobile app, an average of 600,000 per month. Add drivers who use credit cards to pay at meters or pay stations, and about 70 percent of on-street parking transactions are cash-free, D.C. records show.”

61 o. nate May 27, 2015 at 1:09 pm

I admire Tyler’s determination to be contrarian on this, but I think the evidence he adduces is weak. I feel stupid for having to point this out but demand is only one side of the equation. If the marginal cost of producing a good is extremely low, then even very high demand will not appreciably raise the price. The marginal cost to Facebook of an additional user is very low, so we can’t conclude that the demand for Facebook’s product is low by the fact that they give it away for free (with ads). Yes, you need some kind of connection or device to access it, but again the marginal cost of that connection or device is fairly low and getting lower all the time, even despite the poor regulation that allows entrenched cable monopolies to perpetuate themselves. It’s hard to measure the price elasticity of people’s demand for any internet access whatsoever, because there are so many ways of obtaining it (even for free, such as by visiting any public library).

62 Dan Hanson May 27, 2015 at 1:11 pm

Consider a person 20 years ago who buys $1000 worth of music in a year. Music professionally created and distributed. Wages are paid, profits are made, consumer money is spent. All this shows up in the economic data.

Now consider a person making a music video for fun, and posting it on youtube. He or she gets paid nothing for this, unless they are signed up as ad-supported providers. Nonetheless, they are spending time in a recreation that is not tracked, and which replaces old recreations which were.

Now consider the consumer of that music, who is substituting free content for purchased music. Both producer and consumer are better off, but in terms of GDP this would look like a reduction because no money is changing hands.

There are large numbers of transactions which have moved from being financial transactions captured in economic statistics to transactions where the motivations are driven by reputation, fun, peer acceptance, etc. It’s a different and very large form of exchange which is not captured in GDP numbers.

What is the cumulative value of all the open-source software that was created for free and distributed for free? It has to be in the many billions of dollars, but does any of that show up in GDP?

63 Economist May 27, 2015 at 5:17 pm

Some of the open source software may show up. It is used by commercial enterprises and they may be able to charge a higher price (If the quality of their product improves as a consequence). This is, to the best of my understanding, is the type of argument being made here. But they are several missing pieces to this argument and it is unclear if it useful or empirically relevant.

But I agree with you, the odds are that satisfaction from non-monetary activities has probably gone up and is not reflected in prices (and consequently in GDP).

64 Economist May 27, 2015 at 5:11 pm

Your argument may work in theory: in a world with perfect competition. Many technology-intensive industries are oligopolistic and set product prices that are not competitive (in the textbook sense of being in a long-run equilibrium with economic returns equal to the risk-adjusted rate of capital) for reasons involving market share/entry deterrence. Amazon’s very low profits may be a consequence of a predatory pricing strategy or it may not. Both positive and negative biases exist when trying to relate prices to welfare in this context.
I have not thought through this. But I suspect very few people have.

65 Los Ranchos May 27, 2015 at 7:52 pm

Let’s see there are a huge number of other private services that cost <= $2 per day that people spend 4-5 hours per day on average using, ranging up to 24 hours/day. Right? There must be. Maybe not. What else. Roads? Cable? Electricity? What else? It's a public or mixed public/private utility model I think. All those have network effects as well.

66 Geogra May 28, 2015 at 3:46 pm

Value for internet is definitely massive and I don’t think we will be able to survive without that, it might be said for food and water, but believe me that internet is as important now a days, I am doing Forex trading and for me without internet means no earning and that is not easy for survival. I am also lucky that I am working with OctaFX broker; it’s a multi award winning Brokerage Company with having ideal mobile platform, so that helps me trade from any corner of the world.

67 Boonton May 28, 2015 at 3:47 pm

I suspect a lot of the value of the Internet is ‘leaking’ into demand reduction. I used this example in a previous post. Say you used to go out to eat a lot trying many differen tplaces….say 3 times a week. GDP is 3 meals served. Now with Yelp you are able to zero in on the best places in your area. Now you only go out twice a week. Some of your consumption has decreased because it was devoted to ‘experiments’ to try to find things you want that ended in failure. Now you do not have to do as much of that.

So now GDP falls by 1 meal per weak in town. This might be slightly offset by higher prices since the two places you do hit have higher demand and may respond with higher prices. Likewise since the Internet has saved you some serious money you are more open to paying a bit more to your cable company and/or cell phone provider but this demand is not sufficient to offset the lost GDP. This applies to other consumption areas. Facebook means you don’t have to go out as much to check in with your friends. Ebay/Craigslist means you don’t have to buy as many new products. Easy to find reviews means you waste fewer nights spending $30+ on movies you hate. And so on.

Of course this implies you can test this theory by increasing demand stimulus. If the internet allows consumers to ‘demand smarter’, then you can have more stimulus before you exhaust supply and start incurring inflation. If monetary/fiscal policy makers don’t see this they will continue with demand policies set from the days when consumption was ‘dumber’.

68 phil May 28, 2015 at 9:03 pm

I am wondering:

Assume we manage to slow down global population by means of birth control and eduction.

Assume we manage to slow down energy consumption by means of more efficient technologies and education.

Assume we slow down transportation by means of more advanced communication systems.

Assume we slow down food production by means of education.

Assume we slow down weapon production by means of education.

I guess then we would have a bad balance sheet with regards to “economic growth” but a brilliant balance sheet with regards to humanity.

69 Michael Pettis May 29, 2015 at 10:35 am

What about the Google search function, especially for those of us who are much better at processing and synthesizing information, or recognizing patterns, than at remembering? Thirty years ago, if you had offered me a massive Ivy-League library just outside my office with 100 brilliant, energetic young assistants whose only purpose in life was to chase down at any time I chose any bit of information, however vague, and however momentarily, that piqued my curiosity (hmmm, didn’t Alexander Hamilton once say something about the role of debt in centralizing elite support, and by the way just how many times did Phillip II default on asientos, and did he ever default on juros?), I wouldn’t have been able to find any way to express my astonishment and gratitude.

But now, thanks to the Google search function, I have something far better than the library and the 100 assistants. But I still have no idea how to value the enormous impact it has had on the quality of my life and work. Every time I use the Google search function, I am adding value. Some of this value might be captured in GDP, for example if it helps me write a good paper that gets me a speaking engagement. Most of it isn’t. Yesterday I stumbled upon a very nice reminiscence, by some young man who I don’t know, about the influence my father had on his life. That will never add to GDP. Nor will my having cleared up earlier today some confusion about one of Lou Reed’s lyrics.

In one sense this is a trivial problem, and is just one of the standard and well understood flaws of GDP. It is no different than, say, the selected works of Yeats that I bought in my grad student days for fifteen dollars, which I can still pull out and, by reading a few pages, add to the quality of my life any time I want, always without adding a penny to GDP.

But as I tried to explain in my last blog entry, GDP doesn’t have to measure value with anything close to accuracy for it nonetheless to be useful. As long as the biases that make it “useless” are consistent, GDP can be great for comparing things — for example if we compare US GDP in 2013 with US GDP in 2014, we can be pretty sure that as measures of value created in either year, both numbers are wrong, and maybe even massively wrong, but as long as their “wrongness” is biased consistently, and in the same way, the comparison of the two GDP numbers tells us something quite real and useful, and maybe even very accurate, about real US growth. It is true that the extra value created by my being able to pull out the Yeats book is real, and unrecorded, but I would guess that its hidden value is pretty consistent from year to year, so the GDP comparison still works.

I can also compare US GDP with Canada’s GDP, and say, with some level of confidence, that the US produces 9.1 times as much as Canada does, because I suspect Canadian GDP is “wrong” in the same way US GDP is “wrong”. Not all countries can be compared, however. China, if only because of the very different rates at which it recognizes and write down its very large amount of bad debt, has a GDP that is biased very differently than US GDP, so that while neither captures real value creation, they are “wrong” in different ways. I do not think for that reason that I can meaningfully compare the two (which, by the way, is why I would argue that the PPP adjustment makes no sense when it is applied to China).

To return to Google search, I think that things like Google search, GPS, word processing, widely available smart phones, etc. (remember, for a tiny example of the changes it has brought, how incredibly complicated, and often useless, it was before we all carried mobile phones to organize spending New Year’s Eve with different groups of friends if you lived in a big city?) have caused a major dislocation in the GDP biases and as our lives are increasingly organized around these technological changes, the result is not so much a one-off shift in the biases, but rather continuously changing biases, probably understating GDP growth (unless you think that smart phones are destroying the young or some variation thereof).

For that reason I am not sure that we can meaningfully discuss the growth in US wealth from 1990 to 2010 with the same confidence that we can discuss the growth from 1970 to 1990. You can probably say the same about the 1960s and the 1920s, I guess.

Sorry, this is a little long, so I will stop, and I have no idea if I have even added anything useful that isn’t obvious, but to summarize: First, I am pretty certain that while GDP might be a very badly biased and therefore useless measure of whatever it is that we think we are measuring, it is nonetheless useful and even accurate for one, and perhaps only one, reason. It allows us to compare the relative size of two economies, or of one economy over two periods, as long as we believe that the biases are stable. Second, there are things, like the Google search function, that have transformed the old biases by adding tremendously to our lives (certainly mine) in news ways that we cannot capture, and this transformation may be continuous over some period as our lives and institutions adjust. If that’s the case, the one useful thing about GDP is rendered useless.

If you are a well-trained economist, this won’t matter too much because GDP can be quite precisely measured against a consistent set of rules, and you will have learned to accept that precision and consistency matter far more than accuracy. If not, GDP becomes a mess.

Comments on this entry are closed.

Previous post:

Next post: