Virulent Word of Mouse

September 11, 2009

Rural broadband and economic growth – a flu shot with vitamins.

Building rural broadband will bring a modest amount of economic growth. That is what economic research suggests.

Building rural broadband is like getting a flu shot. It might help prevent some problems. It might sting at first. It might make some recipients stronger in the long run, but it is just as likely that it will not make any difference. All in all, it is better to get the shot than not.

Hmm, that first metaphor needs modification. Building rural broadband is like a flu shot laced with vitamins.

Why does the metaphor matter? For better or worse, policy in DC gets made by those who find the best sound bites. I am trying to keep up.

More to the point, some people are convinced that broadband is more like a flu shot laced with economic steroids, as if rural broadband will be a big factor in bringing about economic growth. This is partly due to a misreading of research, and maybe due to some willful misrepresentation by lobbyists. This is just wrong, and very misleading. As I have said before, this (mis)perception needs to be fixed in order to generate sound policy.

This is not a simple topic, but for the sake of numerical simplicity let me illustrate the simplest case, the effects of bringing broadband into a community that has not had any, and, crucially, where the residents want it. Households in such a community already have dial-up service (because everybody does), or satellite service (if they can afford it), or, quite possibly, something wireless (offered by a rural WISP) but slower than wireline broadband.

This describes a small part of the population of the US, though a large fraction of the land area of the country. The latest research from the Pew Surveys on the Internet and American Life suggest that less than 5% of the US population lives in such a low density locations, wants wireline broadband, and cannot get it.

For the sake of simplicity I will say that involves 5.5 million homes (out of approximately 115m).  That is almost 14 million people (out of 300m), so it is not a trivial number, though not overwhelming either. (To be sure, it matters if it happens to be you, but forgive me if I abstract away from that for the time being….)

What economic impact would broadband have on those 5.5 million homes? Here is the point of this post: There are one or two big effects, and those happen quickly, benefiting carriers and equipment firms. That is followed by secondary stuff, which helps and hurts retailers, shippers, and local merchants.  Most of the secondary stuff is likely to be modest, but, all in all, highly uncertain in its magnitude. It is exaggeration to forecast anything else.

Gains and losses.

First of all, if all 5.5 million homes take broadband service, it will create some GDP for access services. At $50 per household that could amount to over $260 million dollars if everybody adopts. If the broadband is part of a package, say, a TV/phone/Internet combo, then it might generate something like $130 per household. That could lead to something like $700 million in revenue.

Either case is great, but neither represents economic growth, not as conventionally defined.

Most, if not all, of the households had standard dial-up (after all, by assumption, these are households who want the Internet). That service gets retired. Standard dial-up would have generated something north of $100 million dollars if all the households had a $20 a month account (or more, if they had satellite service). This was the key point Ryan McDevitt and I were trying to make in our research paper, namely, the effect of broadband on the economy had been exaggerated because analysts failed to subtract what would have happened in the absence of broadband.

Economic growth is the additional revenue brought about by deploying broadband. It would be $260m in the first case, and $600m in the second. Of course, in the second case, broadband should only get $260m of credit because the additional revenue is due to adding telephone service and television service. (Once, again, that might sound like a good thing if it happens to your household, but forgive me if I skip that. I will focus on the effects of the Internet).

There is one other effect. It is called a “multiplier effect” in sophomore macroeconomics courses. New additional revenue in final demand (for Internet access) generates new additional revenue for the supply of inputs.

More concretely, new access revenue generates some equipment sales in order to build out the whole system. While it is hard to say how much that will be, it surely will be several hundred million dollars worth. That is one of the things that surveys of the broadband ecosystem show.  Best case scenario: I will say it is another $260m. (And it happens quickly, which is partly why this looks like an attractive stimulus activity). That is great for Cisco and Dell and a few other equipment firms.

BTW, it is a bit churlish to point out, but most of the multiplier effect does not stay in the local community, but goes to equipment suppliers located elsewhere. Only the labor for construction and installation is a local input.

A bunch of indirect effects

More broadband will do a bunch of other things. Some of this is good, some is not, and most of it adds up to a modest impact altogether.

More broadband shapes prices paid by households. Though dial-up Internet already gave these residents access to lots of new goods through an electronic channel, the installation of broadband will reduce the friction of using electronic channels somewhat.

This is not merely idle speculation. Judging from the most recent descriptions of the value affiliated with growing electronic channels, this type of retail commerce is a beneficiary of more or better broadband.

That is great for the households who make purchases. More to the point, it is really great for Fedex and UPS and the Postal Service. They get to carry all those packages and charge for it. Ah, but how much better is it for Fedex and UPS? Methinks this can be exaggerated easily, and for exactly the same reason Ryan and I stated earlier.

Many of the buyers in these rural areas were already using Lands-End and LLBean and other catalogs. Some of the revenue affiliated with more broadband merely shifted revenue from one channel – an off-line one – to the other channel which is on-line. In short, Fedex and UPS would have had plenty of business, even without broadband.

There is growth, to be sure, but not how much more? It cannot be much. Here is why.

If every household of the 5.5m spends an additional $100, then that generates $550m more revenue for goods.  Amazon and a bunch of other retailers get some of the benefit. UPS and Fedex make another $30-60m in revenue.

Now, here is my point. Something as big as additional $550m in retail spending is extremely unlikely.  First off, and once again, economic growth considers all activities, so the proper accounting thinks about the impact on on-line retailers and off-line retailers.

If a household spends more online then it will take away from a brick-and-mortar retailer, in addition to what it took away from catalogues. While nobody will shed a tear for the rural Walmart stores that will lose a fraction of book and DVD sales to Amazon, we might pity some of the other local retailers who will lose out.

Anyway, more to the point, loss of local revenue will hurt employment at those local retailers and it will hurt local sales taxes. After all, a lot of the on-line sales arise because Congress has allowed a loop hole to remain in electronic retailing, not mandating sales taxes on on-line purchases.  (This loop hole is easy to close in principal, but do not hold your breath waiting for Congress to do that. But I digress).

Less friction is good, and fewer sales taxes makes a good cheaper, by definition, but it is an old economic reality applies here: Most households do not change the total amount spent across goods unless price levels come down a lot, which nobody thinks broadband will do. At best broadband lets buyers see goods they might otherwise not have had easy access to. That is an economic gain, but, hey, dial-up did that too. I am not convinced broadband generates a lot more than dial-up did in this respect.

It is also a bit churlish to point out that most of the electronic retailers in the US have their headquarters in places that do not benefit the local area. But, hey, LLBean is located in upstate Maine, and Lands End is located in upstate Wisconsin (though it is owned by Sears, located in Chicago), so not all the revenue goes to the west coast. So there is some distribution of the gains around the country, if that really mattered to you.

The gains and losses for business

So far it does not look as if local business or the local tax base gains very much from the introduction of broadband. That is not entirely fair. There will be gains to some businesses.

Sometimes it is said that specialty businesses will benefit from access to the Internet. The Internet provides a much easier way to for a rural producer to reach a niche user community. The producer might have a specialty product, such as a special food only produced in some area, or furniture that uses a particular type of local tree wood.

In addition, the sense of isolation and cutoff will be gone for producers whose livelihoods depend on national or global prices. Access to faster and better information allows for quicker updates about input and output prices, and forecasts. That can help a savvy farmer avoid costly mistakes, for example.

All this rings true to me, and aligns with much research, as well as personal observation. However, there is one problem with this whole line of reasoning. The savvy businesses already were using the Internet to do that, and frankly, they have been doing it for quite a long time. (Chris Forman, Avi Goldfarb and I documented this a decade ago among medium and large establishments).

Here is a rule of thumb: If it was important, the business used two dial-up accounts, an expensive satellite account, or they picked up and moved a long time ago.

In other words, maybe some small and medium businesses in rural areas will be able to do what they could not previously, but the best case is modest: the Internet connection becomes a little better and a little cheaper for those users who could not afford anything better than dial-up. Look, that is good, but not a productivity miracle in rural locations.

I am not a total pessimist. Other benefits will be more difficult to see, but will arise in quiet ways. The automatic teller machines will work more quickly, and so will the automatic gasoline pumps. Using a credit card will become less of a hassle too, and that will lead to use of fewer checks, and fraud will decline a bit. In other words, any credit running over a wire will work faster, and with more reliability.

Oh, and the You-tube will upload more quickly too. So will i-tunes, Facebook, Wikipedia, Yahoo Sports, X-Box live, Google search, Skype, and every other top twenty site on the web. All in all, these are good things, though it hardly looks like an economic revolution.  It comes closer to resembling the Internet-equivalent to introducing Ed Sullivan to America, a more entertaining evening at home, at least for the time being.

Bottom line: Though I am the type of email addict who can hardly go a day without my fix, the professional economist in me sees what the economic data say. I will not venture a quantitative forecast, but I am on safe ground saying that more broadband will generate a small economic gain in all this indirect stuff.

Judging from the email I got on the last post, let me anticipate a few reactions.

According to recent research, additional on-line advertising seems to be one of the biggest pieces from new broadband (aimed at an additional 5.5m on-line broadband users). What about that?

Over the last decade more than 60 million homes converted to broadband, and on-line advertising grew from approximately $11 to $22 billion a year. By that reasoning, an additional 5.5 million homes will generate an additional $1 billion of on-line advertising a year.

That sounds great until you remember, as Ryan and I continue to emphasize, that that $1B came out of the hides of many other publications. Many newspapers and magazines are going bankrupt, after all, and that started before the economic downturn. On net, there is a gain from shifting advertising to online channels, but, face it, the net gain has to be modest.

But, wait, what about….on-line medicine. Sure, the local hospital also can do some cool electronic things. And electronic records are great. But that goes in the category of quiet gains that are difficult to see.

Not to be dour, but I am still waiting for somebody to show how any of this will make a lot of money. I like medical gadgets too, but is that too much to ask whether electronic X-rays make or save money?

Anyway, speaking of doctors, there is one other thing that should worry observers. Just about every bit of research on information technology in the last two decades finds an association between diffusion of advanced IT and more unequal income. The part of the population with higher education and better sets of skills adjusts to the new IT, and, on average, manages to increase their income as a result of increasing their productivity. Here is the kicker: they increase their more than those with fewer skills do. On net, therefore, income inequality widens, even if everyone is generally better off.

There should not be any illusions. In most prior literature, better information technology is affiliated with better paying jobs, but not everybody gets those. Indeed, it is quite predictable who tends to get them. The likeliest beneficiaries are the smartest kids, the ones who operate the local electronics retail service, real estate agents who put computing to use. A few managers, lawyers, and accountants do ok too, as do the doctors and other professionals in town.

Actually, however, as an honest academic I have to confess that we are not sure what the Internet will do, or whether it will have the same effects as other computing. There is very little research about whether more broadband or better broadband leads to more inequality in a local area. Chris Forman, Avi Goldfarb, and I have tried to look into it, and, well, the evidence so far suggests the Internet widens inequality instead of ameliorates it, but there is a lot more research to be done before we have a firm answer.

The Nebraska Fallacy

The last refuge in this debate is jobs. Jobs, jobs, jobs. Does the broadband create jobs? Look, the direct effect effects on carriers and equipment firms will create jobs. Most will not last long, but a job is a job, and that is not a bad thing during a downtown like the one we presently are in.

There is another more interesting question behind that surface….How about in the long run?

It is possible to have a sensible conversation on this topic, but only if we avoid the Nebraska Fallacy.

What is the Nebraska Fallacy? It goes like this:

Until the mid 1990s the center of the call-center production in the United States was in Nebraska (there is still plenty there, btw). Why there? For one, most Americans do not think Nebraskans have an accent. Other parts of the country hear it as a neutral sound. Second, and more to the point, because the Strategic Air Command located in that part of the country years ago, the fiber connectivity was quite good at a crucial moment, well before the rest of the country had such fiber. It supported lots of telephone calls. Third, as it turned out, there is a pretty good work ethic in Nebraska. So companies liked hiring people from there. The industry grew and grew, at least for a while.

Then most of that industry went away in the mid 1990s. Not because the Strategic Air Command moved. Rather, lots of the activity moved out of Nebraska and into India and Ireland and the Philippines and elsewhere around the world, where call centers are cheaper to operate. It turned out that most people do not mind listening to a mild accent… up to a point.

Now, here is the key question. If tomorrow every business establishment, home, and farmhouse in every Nebraska town had frontier fiber hook ups, would the call center industry return to Nebraska in droves?

No, it will not. No matter how wired up Nebraska becomes, international competition has done its thing. The era of growth in call centers is over. One has to believe the Nebraska fallacy to conclude otherwise.

There is a broad point embedded in the Nebraska Fallacy. Competitive advantage, especially at a regional level, is very context-dependent. It depends on many factors – the presence of present business and whether infrastructure enables new opportunities, the presence of entrepreneurs who take advantage of opportunities, and a gazillion decisions to intelligently take advantage of new capital and new worker discretion.

More generally, just as more broadband will not bring back the call center industry of Nebraska, more broadband is not going to save the manufacturing towns of the Midwest whose prosperity depends on GM’s recovery.  It is not going to save a mining town of the west whose prosperity depends on high commodity prices for metals. It is not going to work economic miracles.

I am sorry about that, but better to speak the plain truth than base policy on misguided beliefs.

So how can broadband help in the long run? I am not sure, but there is silver lining in the uncertainty. It is not for me to say because it is for the entrepreneurs in these locations to decide. If they see an opportunity and can use the broadband to aide it, then they will. Installing fiber gives a location an option, but does not determine how it gets used. It might turn out to be an unrealized option, but in some case it might matter, supporting an entrepreneur’s efforts.


Does rural broadband help economic growth? It might help prevent some problems. It might hurt at first. It might make the recipient stronger in the long run. It is comparable to a flu shot on vitamins, not steroids.

While that is not all bad, a national program to build broadband hardly seems like a panacea for rural locations, or something to generate so much excitement overall.

To be sure, Fedex, UPS, Amazon, Dell, Google, Apple, a few carriers, and maybe Cisco should be ecstatic.  They cannot lose.

More broadly, all this was a discussion about one situation, a setting where households do not have broadband and want it. Broadband creates the most value in that setting, but, alas, that is just a modest amount. And it only applies to a small part of the population.

There are still lots of situations to think through. If it a quick economic shot in a rural area, but not much more, then it is not likely to amount to much in an urban area either.

Bottom line: take your vitamins, but do it in the right dose.


  1. Hi there,
    Good site, I just found it and I am already a fan.

    Comment by Computers & Tech — September 18, 2009 @ 11:56 am | Reply

  2. Redefine what you mean by economic growth and your analysis breaks down. You say “Economic growth is the additional revenue brought about by deploying broadband.”

    If you said “Economic growth is the improvement in productivity brought about by deploying broadband” your analysis and descriptions will go in different directions. Will broadband help make people more productive? The answer to that is an overwhelming yes. Is it “fair” and should the rest of society encourage people to be spread across the country side. I think the answer to that is yes it is a good idea. Does it matter that we give all in a society equal opportunities? I think the answer to that is yes.

    Given this point of view the question now is how to supply broadband for the least cost to the whole community. We know that basic infrastructure – like roads and pipelines and broadband access is supplied most efficiently if there is one supplier to a geographic area so simple competition in supply is not the answer.

    Here is a suggestion on how we can supply broadband to Australia in the most efficient manner. I expect a variation might work for the USA.

    Comment by Kevin Cox — September 21, 2009 @ 1:09 pm | Reply

RSS feed for comments on this post.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at

%d bloggers like this: