Silicon Valley Has Not Saved Us from a Productivity Slowdown
nytimes.com
nytimes.com
So Pandora, Facebook, Hacker News, YouTube, Google, Wikipedia, gmail, google maps, Reddit... all of these together are worth less than $50/month? Would any of you give up all the free parts of the internet for $50/month?
And look back 30 years: everything has gotten so much better. Rotary dial landlines have been replaced with smartphones. Houses are better and bigger. Cars are better, more comfortable, safer. Children's toys, strollers, playgrounds are so much safer and more convenient. Music has changed from a few CDs to thousands of songs carried around in almost every pocket. The numbers don't seem to make any sense to me.
Marc Andreessen is onto something. The economic numbers don't make any sense when you compare them to reality.
It's silly to take a fixed point in time and just declare "I think we are all pretty happy here", it's extraordinarily silly to do that as Marc Andreessen.
I just meant to say that the article's claim that slow productivity growth is "real" doesn't make any sense.
We definitely could be doing much better though.
The question shouldn't be: "what more can we have?" then.
It should be: "how can we work less?"
http://groups.csail.mit.edu/mac/users/rauch/worktime/
"An average worker needs to work a mere 11 hours per week to produce as much as one working 40 hours per week in 1950. (The data here is from the US, but productivity increases in Europe and Japan have been of the same magnitude.) The conclusion is inescapable: if productivity means anything at all, a worker should be able to earn the same standard of living as a 1950 worker in only 11 hours per week."
To do that, you must legislate a reduction in work hours as productivity increases. Otherwise, corporations and shareholders see all of the gains while workers continue to slave away.
newspaper -> hacker news, google news: GDP drops, but life is better.
stereo + CDs -> mp3 player: GDP drops, but life is better
GPS -> Google maps, GDP drops, but life is better.
snail mail -> gmail, text, GDP drops, but life is better.
Um, what? Google News doesn't do reporting, it aggregates newspaper content. If the traditional media outlets die, there will be no content left for Google News to aggregate. Unless you are referring the part of the newspaper industry which actually puts ink on dead trees, which is pretty small compared to the newsrooms.
People write posts that go on HN to promote themselves and their business interests. Nothing wrong with a little of that, but it isn't a substitute for journalism.
>it isn't a substitute for journalism.
No true journalist would ever impinge their profession by writing for such trash as a blog.
Google News is sort of similar to the business unit of a newspaper which takes finished copy in digital form from the newsroom and actually delivers dead trees to people's doors. Of course the internet should (and will) kill that entirely.
Journalism is a professional craft. Writing about current events for free because you are passionate about some issue, or for a salary because your benefactor has a vested interest in some issue, etc. is not journalism. Your income has to flow from the fact that people want the news, and you/your bosses/your editors have to be properly incentivized to actually report the news.
You're not even remotely talking about economics at this point. You might as well be arguing:
playing guitar -> heroin: GDP drops, but life is better
How do you know that your life is better because you have an mp3 player? Does it make music better? No. Musicians make music, mp3 players have nothing to do with the music. Honestly, my life feels way better when I pull an old vinyl off the shelf. I also like knowing that the artist is getting a better share of the vinyl sales than from YouTube or Spotify. And I kind of hate using my computer outside of work hours, it stresses me out to think about unread emails. So to me, I already don't agree that my life is better when I use an mp3 player.
The point of economics is to measure real productivity based on the assumption that people need a place to live and they need to live and there's a pretty clear limit to the land and food that we start with.
That you're trying to use economics to argue about subjective opinions on mp3 players? Come on, dude, that's just weak.
Money talks, bullshit walks. End of story. Or maybe in your case you should seek some help with this technology addiction that you're trying to rationalize...
With the exception of Wikipedia, and Hacker News -- which is more of loss leader for Y Combinator -- all of the services you listed are run by for profit companies. Clearly they're making more than $50 month.
We're still underestimating value. Dramatically.
The 1% of GDP ($106 billion) in the study refers only to the chunk of the value that we are not paying for in this way.
You are confused because you are considering both numbers together.
The $106B is closer to $30 than $50 (I rounded up). So the combined ad spending and the 106B estimate together is about $50 per month per person.
It's actually more complicated than that - companies like Comcast also capture a (disproportionate) share of the value created. But even when you add it all together, it's still a big underestimate of the value created.
Rents are a real issue with real consequences. It doesn't matter how much cheaper your phone, or your heroin, or your cigarettes cost unless they make your life more productive. You gotta be able to pay rent.
If you buy a table saw and a tool belt and learn how to use them, you can make money, and then you can pay rent. If your table saw gets cheaper, you get to benefit from that consumer surplus. If your heroin or cigarettes get cheaper, you get no benefits from that consumer surplus, unless you want to argue that the temporary highs from drug use lead to long term productivity.
A cheaper TV does not make you more productive. What we can actually see in these studies is that cheaper phones and more time spent on Facebook do not make people more productive. Sitting around watching cheaper and cheaper videos on your cell phone is about as productive as nodding off on heroin.
---
In some circumstances, policymakers might choose to disregard consumer surplus because they do not respect the preferences that drive buyer behavior. For example, drug addicts are willing to pay a high price for heroin. Yet we would not say that addicts get a large benefit from being able to buy heroin at a low price (even though addicts might say they do). From the standpoint of society, willingness to pay in this instance is not a good measure of the buyer's benefit, and consumer surplus is not a good measure of economic well-being, because addicts are not looking after their own best interests.
- Principles of Economics, N. Gregory Mankiw, p. 140
---
The numbers don't make any sense to you because you're not thinking like an economist. The reason why we care about actual trade value and money changing hands is because that's what economics is all about. It's about real people paying real rents and real people buying real food.
You can't pay rent in Facebook "likes".
UberPool and UberHop are actually productivity increases. A single driver can provide 2-4 trips worth of value with about 1 trip's worth of work. However, that's just going to drive the cost of trips down. If Uber and its drivers end up not increasing their revenue per driver trip, the productivity statistics would look the same even though drivers are moving four times as many people!
My layman's interpretation could be wrong, but it really just looks like we're using productivity wrong. It's useful to compare companies and countries at a particular point in time, but if you use it to compare the 2000 world to the 2016 world, it's probably not going to tell you useful things.
Seriously? UberBlack didn't increase the usage factor[1] of existing black cars? UberX drivers didn't have higher usage factors than regular cabs? Because I'm pretty sure the initial selling point of Uber to limo drivers was "hey, fill in gaps between your scheduled rides", which pretty much only works if it increases the usage factor.
And a higher usage factor means higher productivity.
[1] ratio of "time spent driving a passenger" to "total time working (including idle)", my own term, don't know if there's a standard one
Productivity goes up when customers pay the same price for the same value that can now be produced with less labor, but technology has been driving prices down. Technologies used to take much longer to spread. Cotton gin early adopters made way more money than people who did it the old fashioned way, and it took a while for everyone to get on board and drive the cotton prices down with competition. These days, technology gets adopted so fast that adoption is the cost of survival: prices will drop instantly and you won't be able to keep up. My hypothesis is that the speed of adoption these days hides productivity increases. Productivity growth only measures profit from temporary advantages over competitors, and those advantages have become so temporary that people don't take profit from them.
Another hypothesis is that business strategies have just changed. Instead of using technology to increase profit, we now use it to lower prices and increase market share. Our productivity statistics aren't measuring how much abstract value labor can provide. It's measuring our business practices.
Ditto for UberX: it made greater use of a resource whose fixed costs had already been paid for other reasons, and allowed labor to be "kicked into gear" for very short shifts that labor-employers otherwise aren't able to make use of. That's also a productivity increase, and should rightly appear as such.
In this scenario, you need only 1/4 the drivers, and some fraction of those 3 newly unemployed people will eventually do other work. That is the productivity growth.
In the plain English sense, productivity went way up! We're just measuring something else.
In the meantime, please read my other comments in this post, as I don't need to keep making the same arguments, but also read the comments on Tyler Cowen's blog:
http://marginalrevolution.com/marginalrevolution/2016/03/the...
http://marginalrevolution.com/marginalrevolution/2016/03/pro...
http://marginalrevolution.com/marginalrevolution/2016/03/how...
Marc Andreessen is not an economist and he's also an investor in the companies that he is saying are providing mysterious new productivity that "old economics" can't measure.
Come on, you've got to be able to recognize these kinds of conflicts of interests...
In fact, if anything, the increasing inequity between management, C-level executives, and employees makes companies even less able to handle the complex systems of the present world, which is only growing in complexity.
The real issues are human, and human problems require improvements in management, leadership, internal systems, and knowledge and understanding. When rifts form between management and employees, what happens instead is an individual competitive focus that brings productivity and innovation down significantly.
What we're seeing, effectively, is that the world is changing and becoming more complex, and management and internal systems are not keeping up, and instead in many cases going downhill.
There are small pockets of hope in the Lean movement, Kaizen/continuous improvement, and in those who know and understand Deming management philosophy. For a great modern take, read General Stanley McChrystal's "Team of Teams."
[1] http://faculty.chicagobooth.edu/chad.syverson/research/produ...
Edit: And please don't kill me with semantics of "working" vs "productivity".
> And please don't kill me with semantics of "working" vs "productivity".
Sadly, you still missed the point.
Table 2 and 3 in the article below contain info like what I'm talking about.
https://eh.net/encyclopedia/hours-of-work-in-u-s-history/
I generally consider myself a pro-capitalist rational optimist, but we are a lot richer than we were 100 years ago, and my workweek is still 50+ hours a week. I wonder when we will get those 10 hour workweeks...
The cost effectiveness of providing private offices for knowledge workers, even in the most dense urban areas, has been well known for a long time, yet few organizations do it, and some organizations even spend money to tear down productivity-enhancing privacy features in favor of wasteful open-plan fashion.
Why would anyone look to the start-up world when expecting productivity?
[1] https://research.stlouisfed.org/fred2/series/OUTMS [2] https://research.stlouisfed.org/fred2/series/MANEMP
Unfortunately I cannot for the life of me remember what it's called or where I read it. Good odds it's in Robert Gordon's new book (The Rise and Fall of American Growth) or related discussion. Or somewhere in William Ophuls' Ecology and the Politics of Scarcity discussing technology, or Carlotta Perez's Technological Revolutions and Financial Capital.
Light reading.…
Baumol and Bowen pointed out that the same number of
musicians is needed to play a Beethoven string quartet today
as was needed in the 19th century; that is, the productivity
of classical music performance has not increased. On the
other hand, real wages of musicians (as well as in all other
professions) have increased greatly since the 19th century.
For the past 45 years, Canada has had limited growth in labour productivity and currently stands at $42/hour, contrasting against the US at $52/hour, measured in GDP/hour worked.
It turns out that it's very very difficult to understand productivity. It is against common understanding why Canada's more numerous bachelors degree holders seemingly doesn't contribute to a more productive labour force. If investments in education do not increase labour productivity (as predicted by nearly every model of labour economics), then what measures do affect productivity? It is increasingly unconvincing that countries become more productive by mere accident.
In any given year, 1 in 5 Canadians experiences a mental health or addiction problem.
The "education system" is not an education system. It is a selection system used to choose who gets one of the at any one tine limited number of jobs. Increasing the number of people in that selection system makes it less effective, hence the lowered value of a degree, the need for internships, and increased stress on personal networks for finding jobs.
There are many more plausible reasons that don't have to invoke such strong assumptions on the education and skilled workforce markets.
Cultural, social and political differences, a lower population density et al.
So say everyone needs to pay taxes. Let's say that everyone complies and pays a company $100 for software to do this. Now I make equivalent software which I'm able to sell for $50 (because I'm better at running my company, say). Now people are spending exactly half as much money, but the same job is being done. This looks like an economic productivity loss? If so, surely there are alternative models that capture the intuitive gain from things like this? Or maybe such gains are rare and negligible?
The studies in the article assert that the size of the productivity shortfall is so large that even if we place pretty generous values on all this free stuff we have now it doesn't fully cover the gap.
Certain circles are beginning to realize that the main issues are human and not technological (Lean, Deming, Kaizen, Design thinking, etc).
So there's your answer. Increasing complexity of work, with unchanged or at best slowly increased ability to cope with it.
For a great overview and insight into this shift (and how to tackle it), check out Gen Stanley McChrystal's book, "Team of Teams."
Also, virtually every market is dominated by cartel, duopoly, or monopoly conditions, and companies are often satisfied with that position, so investment is down.
This is a fun thought-experiment.
http://www.mckinsey.com/business-functions/strategy-and-corp...
What he's saying is that all the fuss over Silicon Valley technology is not warranted when it comes to actual productivity gains.
This is a really important point because it suggests two different way of conceiving of technology: 1) how economists define technology, viz something that allows us to do more with less 2) what laymen consider technology: the latest gee whiz gizmo marketed by Silicon Valley
The economic definition is a lot stricter in some ways - and broader in other ways.
The takeaway is that in economic terms we really haven't experienced much technological evolution. I find that to be pretty believable.
What does productivity mean when your economy is increasingly powered by data, not money?
As a wild generalisation, there are two kinds of economic model - the political model, where the point is to use rhetoric to enhance power differentials between social castes, and the evolutionary model, which considers economic activity a proxy measure for collective intelligence and opportunity.
We're in a transition period where the former model is peaking and about to[1] crash, and the latter model is becoming more sharply defined.
In the former model an abundance of data is a bad thing, because the market value of any specific type of data crashes towards zero.
In the latter model an abundance of data is a good thing, because it increases the possibility of invention and innovation. And an abundance of data refinement tools is even better.
Which is a more accurate measure of productivity?
[1] For very poorly defined bounds of "about to". I'm guessing more than decade, less than a century.
What kind of "evolution" are you talking about? I think all you're pointing out is that Silicon Valley is extraordinarily bad at attributing value to data.
It's not the roll of economics to fix that problem. Economics deals with real things like people buying food and having to pay rent and pay actual taxes.
If Silicon Valley can't figure out a way to get translate "data" in to "thing I trade to my landlord so he doesn't call the cops", well, then Silicon Valley is wasting everyone's time.
For example, by 2005 (when the slowdown in Labor productivity began) most of the benefit of computers for most everywhere had been accomplished. Microsoft's dream of a PC on every desk was realized. Email was everywhere. For the average office worker, were there any major productivity gains from tech left? Nope.
Most tech improvement since then has been mostly about entertainment and mobility. Entertainment doesn't enhance the economic definition of technology, and the increase in mobile tech is probably a mixed bag in terms of productivity.
Prior to 2005, the productivity came from a revolution in computers and the Internet. Now we are just refining that revolution. To get high productivity gains, we need another revolution of some sort.
3D printing is nearly at prime time, which should greatly enhance manufacturing and design productivity. I don't know if that will have a great impact or not, but eventually there will be some disruptive technology that will bring real change.
But a car company doesn't need to reinvent production to save money per car. They can do that many ways without a single new robot to replace a person on the factory line. The fact of the matter is that if you look at the last three decades of growth it was all on the backs of cheap Chinese and Asian labor pools. Now those pools are demanding higher wages so we're pushing into ever smaller, cheaper labor pools. Eventually, the whole thing will come to a halt once African nations are at Brazil's level of development since they too will start demanding safer work, better wages, and a welfare state.
Then the real demand for technological productivity gains will be in full swing. Until then, we'll just exploit our fellow human being as best we can to get those sweet sweet metrics for Wall Street!
Derivative of productivity != productivity, NYT.
See, this is what happens with decades of people tolerating media confusion between the GDP and the annualized rate of growth of the GDP.
Now its spreading to everything else.
Yeah, it's slightly disingenuous, but we still need some models. It's like when people take the total revenue of Google and divide by the total number of Google employees (including assistants and chefs and tech support and the PMs and artists and designers and all the actual engineers too) to reach a "revenue per employee" number.
It's not perfect by any means, but anytime you divide one number by another, you get to call your result a metric (then you get to write insights about your new discovery too).