The Wall Street Journal has described Hal Varian as the Adam Smith of Googlenomics. As the tech giant’s chief economist, he revolutionized Google’s business strategy, and is known now as perhaps the most prominent skeptic of America’s official, sluggish productivity numbers. He joined the podcast to discuss the tech industry, the future of the economy, and much more.Possibly also of interest, some of our related posts:
In addition to serving as Google’s chief economist, Hal Varian is a professor emeritus at the University of Berkeley and a fellow at the Guggenheim Foundation, the Econometric Society, and the American Academy of Arts and Sciences. He’s also the author of two economics textbooks, and the co-author of the bestselling business strategy book, Information Rules: A Strategic Guide to the Network Economy.
You can subscribe to my podcast on iTunes or Stitcher, or download the podcast on Ricochet.
JAMES PETHOKOUKIS: As my listeners know, frequently I’ll have guests on and we will talk about the productivity paradox (including here, here, and here). It’s the idea that we seem to be witnessing all sorts of tech advances, not least of which is the ability to find all manner of information via a bit of glass in our pockets. But yet the official productivity numbers, as measured by the government, are terrible. So how do you square the circle on this?
HAL VARIAN: It’s a tough problem. I think one of the issues that I’ve dug into quite a bit is the measurement issue — whether we’re measuring the right things.
And there’s two aspects of that. One is are we measuring welfare or consumer well-being with these numbers correctly, which is a different thing than GDP? And two is, given that we want to measure GDP, are we measuring GDP correctly? So let me say a word or two about the first one.
Let’s take an example of photography, all right? So in 2000, there were 80 billion photos produced. We know that because there were only three companies that produced film. And fast-forward to 2015, there are about 1.6 trillion photos produced. Back in 2000, photos cost about 50 cents apiece. Now they cost zero a piece essentially. So any ordinary person would say, wow, what a fantastic increase in productivity, because we’ve got a huge amount of more output and we’ve got a much, much lower cost.
But if we go look at that from the GDP lens, it doesn’t show up in GDP for the most part because those photos are typically traded among friends and put in albums and things like that. They’re not sold on the market. GDP is the market value of transactions out there, and anything that’s not sold or has a zero price isn’t going to show up in GDP.
So we are missing important things in the GDP measurement, but that’s not the whole story to the productivity paradox that you describe because it’s not just the technology sector that’s showing this gap. It’s all over; every sector pretty much is showing us.
It’s across sectors and it’s also just not the United States.
Absolutely. It’s international as well. Of course, now, a lot of these technologies are available internationally so they’re still seeing that.
Another example, same general theme, is look at the GPS. Back in 1990, GPSs were expensive, $1,000. Trucking firms used them, commercial vehicles, but they were above consumers’ price point.
What’s happened is the prices come down, down, down, real GDP is going up, up, up, because the price of the goods has gotten cheaper until it hit zero. When it hit zero, it’s out of GDP so now everybody has GPS on their phone and we’ve got, again, what any normal person would say is a big productivity improvement, but it doesn’t show up in GDP because of the zero price.
Now those are two compelling stories that you’ve told. And the idea, which Joel Mokyr from Northwestern has said, is we have all these measures meant for a wheat-and-steel economy and we have a digital economy.
But do you think that’s enough to explain why productivity growth has been so sluggish? You could absolutely be correct, but, one, even if you add all that stuff in, we’d still have low productivity growth, and, two, that’s always been the case. There’s always been this uncounted surplus from technologies, so that’s always been the case and it’s not getting any worse.
So do you think that something is different and it’s a bigger piece? And does it get us all the way back to a healthy productivity number if you included all this?
So it definitely does not get us all the way back. It’s a component. It’s a feature of what’s going on because we’ve got all these new technologies, new business models, new capabilities and we’re still working on figuring out how to measure them appropriately. So it’s a piece of the action, but it’s not the entire action.
If you ask what’s left over, again, I can’t say I’m going to give you an entire solution, but I’ll give you a good place to look. A good place to look is to look at the leaders and laggards. If you look at the leading companies that are doing the best, that are the most advanced at using these new technologies, they’re doing pretty well in terms of productivity, where we think of output per worker, output per hour worked.
But then there are still a lot of laggards who aren’t really adopting the new technologies and aren’t as productive as the leading firms in their industry. And there, I think, what we rely on or we hope for is diffusion of this knowledge through the different indices, and we need to take advantage of the potential productivity gains that are there.
So we’re now in a situation which we’re on our way, we believe and hope, to a more productive economy, but it just hasn’t sprung up everywhere.
To some degree this has always been the case, but is it worse now that you have this frontier leading edge of companies that are benefiting greatly from this new technology, and then you have everybody else? So we have productivity and innovation, it’s just locked up in not enough companies? Is there something different this time about that process, and do you see the diffusion beginning to happen in any of the data?
So absolutely I see the diffusion beginning to happen. And it’s not so different than 100 years ago when you look at the assembly line: Henry Ford comes out with the assembly line, they are hugely productive compared to other automobile assemblers, which is really what they were at that point. But the technology, the assembly line, permeated through the economy and now that’s how we do most manufacturing.
So if we go look at the online world, a lot of these technologies, like machine learning and artificial intelligence and robotics, those are being used in frontier-level firms but they have not diffused widely through the entire economy.
But we’re seeing it happen. We’re seeing more and more companies becoming interested in adopting these technologies. We’re seeing a huge change in the universities with students all wanting to get a degree in data mining or data science or computer engineering. So the expertise is being dialed up.
We want to see that technology spread beyond only the leading group of companies that we’re all aware of. But then there’s the issue of really using that technology and optimizing the use where you have to figure out the best way to use it.
And the classic example is with electric motors — that even when we had electricity, the motor took a while for factories to figure out how to use them. They had to rearrange the factory floors, and it took quite a long time. And part of the Robert Gordon argument is that this information technology has already been around for a while, so we’ve already figured out how to use it and we’ve already gleaned whatever productivity gains we can have from it.
So do you think, especially now that we see artificial intelligence advancing, that there’s just a long ways to go before we really figure out the best way to use these technologies, perhaps even at these leading-edge firms?
Right. Well, some of it is just failure of imagination — that people aren’t quite sure what they can do with the technology, they don’t know, they don’t have an example they can see and so it’s hard for them to really envision making that change.
One of the nice places to look at this is on Kaggle. It’s a startup that was acquired by Google about six months ago. They run competitions in machine learning. So real businesses come with real problems and real money and want a real solution to their problem. So you can see these examples of how the technology can be used in ordinary businesses.
So I’ll give you a couple of examples. One example is the Heritage Health Prize. That was a health foundation. What they did is they offered a $1 million prize to the teams that could produce the best forecast of hospital readmissions. So somebody’s discharged from the hospital and has to come back in a couple of months later; that’s viewed as a problem for the health care industry and they would really like to be able to forecast that more effectively.
Zillow, the online real estate site — they have a house price forecasting system which estimates the market value of houses based on the characteristics. They came up with a $1 million prize for whoever could improve on their forecasting system, and so on and so on and so on....MUCH MORE
GDP ‘Not Fit For Purpose’ in Digital Age, Says Digital Giant
A Romp Through Information Security Economics and Society With A Cambridge Professor
Cleveland Fed: Hope for U.S. Productivity Growth
Google's Chief Economist Talks Intelligent Technology
"Silicon Valley Doesn’t Believe U.S. Productivity Is Down" (GOOG)
One of these days I'll get around to doing a post on capital deepening. I promise.
"Facebook, Google, and the Economics of Time" (FB; GOOG)
Although this story is not about intertemporal arbitrage I'm sure that's the first thing some of our readers thought of.
"The Ray Kurzweil Show, Now at the Googleplex" (GOOG)
Google Chief Economist: "Micromultinationals Will Run the World"
For Today’s Graduate, Just One Word: StatisticsStill some time to run for the forecast in that August 6, 2009 post.
DealBreaker tips us to this New York Times story by saying:
Tired of getting laid? Have a penchant for hating yourself and others? Then statistics might just be for you, and it turns out people need the poor bastards to do it.“I keep saying that the sexy job in the next 10 years will be statisticians,” said Hal Varian, chief economist at Google. “And I’m not kidding.