Let me begin this post by declaring – without reservation – that I just love what technology and innovation can do to an economy. No, that is not a strong enough declaration. Let me say it this way: I am a card carrying member of the club that believes a substantial amount of economic growth in developed countries build on a wide and solid foundation of innovation and technological change.
What can I say? Among economists I am a technological enthusiast.
It turns out, however, that I am out of step with some of the telecom policy community in Washington D.C.. I have not changed, but the yardstick inside D.C. is just different. I was reminded of this recently after a paper of mine was presented in DC. Several other economists and I were subsequently described as contrarians in a post by Paul J. Feldman, who writes a blog about communication policy.
What is the meaning of contrarian?
At first I was amused by this hyperbolic label, and maybe a little flattered. Reading the post did not offend. Feldman listened well, and got the right substantive message. I would like to meet this guy sometime. He seems to be awake and seem to be thoughtfully paying attention.
But there is something deeper going on. And it deserves a response.
Dictionary.com says this about the definition of “contrarian”:–noun,a person who takes an opposing view, esp. one who rejects the majority opinion, as in economic matters.
Feldman used the label because we all (the economists) insist on interpreting policy choices in light of data and subjecting proposals to scrutiny.
Is that being a contrarian? I beg to differ. That is what any professional analyst should do.
Indeed, that is but a small step in a research agenda. There are many open questions about the economic impact of broadband.
It turns out, however, that conventional wisdom believes the economic issues are already settled. Moreover, conventional wisdom has settled on a view that just ain’t so. There are two parts to that statement — what conventional wisdom says, and what the economists say.
Let me start with my view of things, because it is what I know best. Then after that I will review conventional wisdom.
Ryan McDevitt and I did not intend to take on the role of skeptic. Rather, like patient academics, we took a step in a wide open agenda with our paper, “Broadband bonus: Accounting for Broadband Internet’s Impact on GDP”. We used standard economic reasoning and carefully applied it to actual numbers. Finding the data and interpreting in a sensible way involved many research challenges. We started this more than 18 months ago. We thought others would find the results useful, not because we anticipated the Congressional broadband plan.
We essentially addressed two questions: First, how much additional revenue in GDP did the diffusion of broadband generate? Second, how much consumer surplus did this diffusion of broadband generate among users of broadband? And after a lot of work we finally got some answers. By the norms of economic research, these numbers are large. If you do not believe me, read the paper.
There also is a big unanswered research question in our paper, and we are quite open about it. What is the size of the economic spillovers from broadband into other economic activity? I do not know the answer because it is genuinely hard to answer.
(BTW, in case you were wondering, a “spillover of an economic transaction is an impact on a party that is not directly involved in the transaction. In such a case, prices do not reflect the full costs or benefits in production or consumption of a product or service.” That definition comes from Wikipedia, and it is fine for these purposes.) In other words, how much economic growth does broadband generate for others not involved in the transaction to purchase or offer broadband services, such as Amazon, Google, Intel and Apple?
Let me illustrate why it is hard to answer the key question about spillovers by discussing what happened in advertising this decade. Online advertising has increased from around $10B to just north of $20B over this decade, coincident with the diffusion of the broadband (much of it going to Google, but I digress). Of course, most of that came at the expense of off-line advertising, which has decreased quite a bit in newspapers and magazines (even before the most recent recession). That illustrates why computing the spillovers from the diffusion from technology are hard to calculate precisely. The increase in online advertising has been big and so has the decrease in offline advertising. That is not trivial because these two events are related. A proper economic calculation of economic growth must count them both, both up and down. (Counting only the up is, well, dishonest at worse, disingenuous at best, and incomplete in any event.) On net total advertising has not changed much. That is still an economic benefit, to be sure — we are getting something for cheaper due to a new technology. However, it does not necessarily translate into a direct increase in GDP.
One could do a similar analysis for electronic commerce in retailing and in electronic business transactions, as well as in equipment sales. And it gets real complicated when one starts to look at the music industry and other information-intensive activities.
Anyway, here is the point: Economic outcomes are surely better due to the diffusion of broadband, and business productivity has surely increased as a result. However, on net it is not straightforward to figure out precisely how much growth should be attributed to broadband.
Let me say it another way: Anybody who thinks this can be computed in a few pages is fooling their readers as well as themselves. Answering this question for the entire economy is the type of question that takes hard work, and lots of effort. Nobody has figured out a persuasive answer yet for good reason.
That is also why Ryan and I did one piece of it first. Maybe we will be able to get to the other pieces in due time.
Conventional wisdom versus GDP accounting
That said, nothing prepared me for the type of reception Washington D.C. gave us this time around. On several occasions, when I have presented our paper in Washington D.C., Ryan and I have encountered disbelief and skepticism about the results. We persistently get questions about why our results are not in the hundreds of billions of dollars.
Look, it is not several hundred billion for a straightforward reason. Revenue for all Internet access in the US is only around $40 Billion. For sure, that does not include consumer surplus, spillovers, and, as I mentioned, additional economic growth from productivity gains, but the first cut is still just $40B. It is what it is.
Now here is the straightforward statement: The US economy is NOT going to get several hundred billion dollars of additional economic growth from one investment that amounts to only $40 billion dollars in revenue. Economic spillovers, productivity gains, and consumer surplus might amount to something, but never amount to several orders of magnitude effect greater than the initial economic activity.
But I have learned that I must live in a different world, or, more to the point, most lobbyists in Washington do not bother to talk to most economists. Ryan and I are following standard procedures employed by the Bureau of Economic Analysis to generate GDP, a series of rules for generating the National Income and Product Accounts. While I understand why most people in Washington do not stop in the BEA for a chat to bone up on the latest data, it still seems like the professional lobbyists – especially those who inform policy makers about why more broadband leads to more economic growth – should be required to know the government’s own accounting rules.
Instead, the conversation has been allowed to wander without any grounding in standard GDP accounting. There is a general assumption that the economic benefits from broadband are extremely large, and, frankly, so large as to be out of scale with reality. Lobbyists and enthusiasts have been throwing around numbers in the hundreds of billions of dollars.
Some people have sent Ryan and I these studies. I have had to read some of this stuff, and it is shocking how poorly done some of this is. In public comments I have particularly singled out one of these groups, ConnectedNation, which was brave enough (to their credit) to write down why they believed what they believed. They circulate a statistical study saying that bringing broadband to rural areas will generate $170 billion dollars in economic benefits. These people might be very sincere, and their goals appear to be worthwhile, but this study is so statistically flawed it would not get a passing grade in the MBA statistics classes we teach at my home institution. If a chief marketing officer did this for his or her CEO, that officer would be fired for incompetence.
Look, on face value this is just wrong. An economy cannot make an investment for 15% of the US population (which would be everyone living in rural areas) and generate $170 billion of economic returns. It just can’t happen with broadband, cellphones, or any other miracle drug. And in this case, we can be very precise, because there is enough data to KNOW what happened because 60% of US homes ALREADY use broadband. It only generate $40 billion in GDP — and that is counting both revenue from both home and business users. In other words, even if the US government lays broadband in every rural area and everybody uses it (an unrealistic best case scenario), our starting point for an estimate should be $10 Billion in direct results. And there are lots of reasons to think it might be much lower than that in direct results, since at least a third of rural households will not adopt and some already have it and do not need the additional investments. We can add more to the gains from spillovers, but there is just no way to get a big number like $170 billion. A reasonable first guess is less is less than $10 billion, and that is that.
To paraphrase Will Rogers, It ain’t what you know that scares me. It is what you know and just ain’t so.
Like other professional economists, I will continue to insist that economic reasoning and careful economic accounting needs to inform policy. I will insist on it because I am not a contrarian. I am a technological enthusiast, and I think it would be a good idea for the US government to have a good broadband policy. I want the FCC staff to do a good job, and I think they can.
That is not being a contrarian. It is just asking for policy to use professionally informed opinion.
Let me say that more positively. I still think investments in broadband will yield economic gain. But like all investments in new technologies, the gains will not come easily, and not from broadband alone. The gains will arise after complementary investments in the IT work force, in computing equipment, in business processes, and in many strategy sessions to design clever uses for all those investments. Moreover, most of the gains will not be shared, but, like most IT investments, will tend to generate gains in a few key intensive users, such as retailing, warehouseing, logistics, media, and so on.
More to the point, the economic gains will not be outsized, but will be modest at first because there will be costs too. With better broadband the local car dealer can place online advertising for less, and that will be a gain. But it will also be a loss: it will come at the expense of the local newspapers who will find less demand for ad space in the newspaper. There will be a gain from the availability of a wider range of goods in the online store. But it will come at the expense of the local retailer, who will find a marginally less interested set of local buyers. There will be a gain to the local business who gets it billing done more accurately and with less time, but there will be a loss too, to the local bookkeeper who will get less work in the short run. On net these gains and losses will result in economic growth, because – hurrah! – there will be productivity growth, but it will not be a miracle. It will come with long hard work, like any gain reaped after a thoughtful investment.
It is not contrarian to argue that economic growth is hard.
Society can be better off in the long run with more broadband because it will be possible to do the same activities with fewer resources. Broadband can be a part of that, and can make a contribution. But the policies have to recognize that money spent should yield real and concrete gains. And the numbers thrown around in conversation have to be realistic. Otherwise the policies will waste money and yield fewer gains than possible.