The blog of Bill Gates

The Failure to Measure Innovation

Filter By
Your search for "", with selected filters, does not match any posts. Please try again with a different search term or reset filters.

Popular searches include: Books, Malaria, and Future of Food.
Sign up to receive occasional updates from the Gates Notes
Privacy Policy
It looks like you're using an older version of Internet Explorer which may not display all the features on this site. Upgrade Now » close
A Missing Statistic

The Failure to Measure Innovation

Quantifying the economic impact of innovation has turned out to be surprisingly difficult. Astrophysicist Lowell Wood, and Nathan Myhrvold, former chief technology officer at Microsoft and founder of Intellectual Ventures, and I talk about how innovation is overlooked in economic statistics.

Bill Gates: One thing I want to understand better is to what degree innovation that improves the human condition is reflected in economic figures.

Let’s say that someone invents a new bug spray that stops mosquitoes from biting people and is super cheap.

Or let’s say someone comes up with a way to replace hearts for people over 80 that enables them to live longer but that costs a lot of money.

I think that innovation does generally map to economic growth over time. But taking a particular innovation and quantifying its direct economic impact is hard.

For example I believe that because the Internet lets people find a lot more information a lot more easily and quickly, it is a much bigger benefit than the economic figures show.

So when people say middle class salaries have not gone up much I think that the “basket of goods” approach to measuring improvement inherently understates how much better life is now than it was in the past.

Lowell Wood: I believe that it’s of enormous—I would even say surpassing—importance to be able to quantify the impact of innovation, at least in limited cases.

I suspect that the objective impact of innovation is huge at least occasionally. Being able to prove this would be of great significance in the general debate over how and how much to foster and encourage innovation.

But it seems to be exceptionally difficult in just about any really big circumstances.

Specifically, just about all attempts through the present have been unable to compel the belief of the reasonable skeptic—for good reason.

The National Science Foundation makes occasional attempts to do so, but these seem quite unpersuasive (even to me, eager would-be believer that I am).

I wholeheartedly agree that the economic contribution of the Internet is undervalued. But that’s because Internet-based capabilities have totally changed my professional life, so much so that I can hardly believe how “crippled” and “isolated” I was even two decades ago.

I also strongly agree with your take on the “basket of goods.” The quality of the basket of goods has advanced greatly and this value-increase is essentially totally discounted by standard modes of accounting.

For example, I bought a ticket for a 5,600-mile airplane flight yesterday for $500, an unfathomable bargain by the standards of any previous decade. And I use a laptop computer that cost a few thousand dollars that has nearly 10,000 times the computing power of a system that cost $10 million less than three decades ago.

I would cheerfully trade two years of professional life a quarter-century ago for a year nowadays. Professional/intellectual life currently is vastly more intense, due to the variety and stunning power of the “mental prostheses” now available—so that I’m enabled to do far more per day than I could back in the “Dark Ages”!

Nathan Myhrvold: I think the core issue here is quality of life—you can call it various other things, but that is what it amounts to. How do you measure the fact that today we have access to goods and services that we didn’t previously. Simply put, most economists don’t attempt this.

Another way to say this is how could an accountant tell? Economic statistics measure employment, income, and so forth and they don’t grasp this.

The baskets of goods and services that are measured are almost invariably inherited from lists made in the past. There’s a good reason to do so—it allows the numbers to be compared. It is not a bad thing to ask what the price of a Kwh of electricity or gallon of milk or gasoline has been over time.

But nobody tracks computing power, access to information, and so forth.

One could create a “digital basket.” In the graphics category it would record how many polygons per second you can draw per dollar. In information it might track how my access to encyclopedias went from Britannica to Encarta to Wikipedia.

When it comes to software and IT spending, economists would say that they sort of track this by measuring worker productivity. For ages there was a “productivity paradox” because spending on IT did not seem to track to worker productivity. Then it caught up very quickly, so economists would say they do capture it. I am still dubious.

But look at Wikipedia. For about 10,000 to 100,000 people it is an obsession that could properly be called entertainment. They edit the thing for fun—or anyway for zero monetary reward, so no economic stats there.

For the rest of the planet it is an amazing resource—incredible information on just about anything. Where does that fit in any economic analysis? 

Sign up to receive occasional updates from the Gates Notes
Privacy Policy

Read previous versions of the Annual Letter

comments powered by Disqus