Comment by lkrubner | Hacker News Reader

Comment by lkrubner | Hacker News Reader

Parent Full thread

Those are PPP numbers, which means they are a total joke. The graph you post does not show the post war boom, nor does it show the productivity stagnation after 1973. The USA went from 3% per year productivity growth, before 1973, to 1.5% productivity growth after 1973. None of that is visible in the chart because the chart is using PPP numbers.

But, as a separate issue, the above essay is focused on computers, and so it implicitly raises the Solow Productivity Paradox: why don't computers seem to add anything to productivity?

See:

The Solow Productivity Paradox: What Do Computers Do to Productivity?

https://www.brookings.edu/articles/the-solow-productivity-pa...

McKinsey wrote this in 2018, predicting a productivity boom was coming soon, though of course this prediction has been made often since 1973, and the only time it was even slightly true was during the boom in big box retail 1995-2005:

https://www.mckinsey.com/business-functions/mckinsey-digital...

It's well known that in the 1800s the telegraph and then the telephone allowed a large scale re-organization of capital, and capital invested heavily to take advantage of the new technologies, but it was impossible to see a productivity benefit. These technologies allowed new systems of control, but not necessarily any productivity benefit. Since 1970 computer network technologies have offered a similar reorganization, but they are vastly more powerful than the telephone, and they allowed a re-organization that was vastly more profound than any previous reorganization of capital. Millions of jobs were outsourced to India, China, Hong Kong, Singapore, and then Brazil, Russia, Romania, etc.

> But, as a separate issue, the above essay is focused on computers, and so it implicitly raises the Solow Productivity Paradox: why don't computers seem to add anything to productivity?

Just going by eye, there is a pretty sharp trend upward from about 1990-2005, which I think roughly corresponds to the period when the bulk of computerization happened in homes, offices, schools, as well as being put to use in many of the most important embedded applications -- factories and machinery, automobiles, etc.

Eyeballing the trend before 1990 has it continuing to low-$50s, instead it's mid-$60s, a 20% increase.

So why do you think this graph shows computers don't seem to add anything to productivity? I'm not saying they have, but I can't see how that graph demonstrates they haven't.

I can not take credit for Solow Productivity Paradox, all credit needs to go to Solow himself.

The uptick in productivity between 1995 and 2005 is mostly attributable to 2 sectors:

retail

finance

In retail, it was the era of big boxes, lead by WalMart. Basically, it was the era when big retail learned how to take advantage of big databases to better manage inventory. Sadly, this did not seem to be the beginning of a sustained period of higher productivity trends. Rather, it was a one-off innovation.

In finance, much of the uptick involved HFT. Some economists feel this contributed to the crash of 2008, so how much we should celebrate this particular uptick is open to question.

One of my thoughts is the hard return on efficiency isn't linear.

140 years ago generators were 60% efficient. 20 years later 96% efficient. That's big close to 40% better. 20 years after that 98% efficient. 4% better... Now maybe 99.5% so 1.5% better.

I feel like a lot of things they through computers at to improve productivity are kind already efficient. My mom's first washing machine washed clothes. Mine with a microcomputer also washes clothes. Parents first microwave had a power dial and a mechanical timer. You reheat food in it. Mine with a computer also reheats food.

There is no real fundamental change here.

Do you have a source for that? You're saying a productivity increase that amounted to around 20% across the entire workforce is mostly attributable to advances in HFT and logistics in the retail industry?

Even if we ignore that both of those were probably only made possibly with computers, that seems astounding considering retail and finance makes up less than 20% of the economy by GDP.

The link should be here, if Google was working:

solow hft retail site:krugman.blogs.nytimes.com

At the moment Google is acting as if Krugman never mentioned HFT or retail.

Adding to the problem, the New York Times is giving 503 errors over most of Krugman’s old weblog, so perhaps the Times has given up on maintaining the old blog.

Am I allowed to make the joke that Solow’s Productivity Paradox is mostly explained by rampant link rot? But I’m only half kidding, because the rapid way technology breaks down is an obvious hindrance to productivity.

So here is a link that is a few years earlier than the one I want, but which covers the productivity numbers in the same way:

https://krugman.blogs.nytimes.com/2009/11/09/reagan-mythbust...

But that was written in 2009, before it was clear that productivity slowed again after 2005.

In a later post, Krugman analyzed the higher productivity of 1995 to 2005 and shows much of it was simply big box stores and HFT.

In general, the period since 1973 has been disappointing for productivity. And as yet, computers have not brought us the kind of sustained productivity growth that we saw in the early to mid 20th Century.

I can't load that page for some reason, and although I don't really like Krugman and certainly if it just comes from a blog with his opinion and vague handwaving in it I don't think it's worth much. But that surprises me that he would have changed his tune so quickly after being famously against HFT, e.g.,

https://www.nytimes.com/2009/08/03/opinion/03krugman.html

> It’s hard to imagine a better illustration than high-frequency trading. The stock market is supposed to allocate capital to its most productive uses, for example by helping companies with good ideas raise money. But it’s hard to see how traders who place their orders one-thirtieth of a second faster than anyone else do anything to improve that social function.

> What about Mr. Hall? The Times report suggests that he makes money mainly by outsmarting other investors, rather than by directing resources to where they’re needed. Again, it’s hard to see the social value of what he does.

> And there’s a good case that such activities are actually harmful. For example, high-frequency trading probably degrades the stock market’s function, because it’s a kind of tax on investors who lack access to those superfast computers which means that the money Goldman spends on those computers has a negative effect on national wealth.

"he would have changed his tune so quickly after being famously against HFT"

He was still against HFT, that was part of his point. He was saying that even some of the surge of 1995-2005 might have been an illusion, because it was "only" HFT. If you feel that HFT really does raise the USA standard of living, then of course you are free to view the surge of 1995-2005 as entirely real.

“You're saying a productivity increase that amounted to around 20% across the entire workforce is mostly attributable to advances in HFT and logistics in the retail industry?”

No one ever suggested that. It’s the excess productivity growth of 1995-2005 that is largely attributable to retail and what improvements there were in the productivity of trading financial instruments.

I'm not talking about the total increase of productivity, I'm talking about the roughly 20% it increased beyond where it would have been if the pre-95 trajectory had continued. I.e., the excess. It just doesn't seem like a plausible explanation.

EDIT: going back and looking at the actual numbers rather than eyeballing it's more like 10-12% excess. That's across the entire economy of course, so it would still be a massive increase to attribute that mostly to a subset of retail and finance sectors.

I have a research assistant I hire to deal with FRED for me, since it is really its own specialty, and at some point we will get to the productivity numbers, but I'm not going to engage them simply for a comment on Hacker News. I probably should not be writing on Hacker News, I keep thinking that I should focus my efforts elsewhere. But eventually I'll have more complete work to share with you and the world. Feel free to follow my blog, you'll see an announcement at some point in the next 6 months.

There are internal corporate studies out there that try to determine which manager or technology brought productivity gains.

The math is so complex they could not draw any conclusions.

Understand; we listened to lesser educated men of the previous generation about building business and economy. Men who missed out on our early education in science, and who were often far more religious. Men who had much different political narratives to satisfy; IBM was gifted government research. How many of our startups are?

Brookings conveniently avoids discussing the weakening political power of the US over the decades; easy to grow when the world needs you to manufacture for it after WW2.

It’s not hard for me to believe much of our economic theory is bullshit, and output is driven by political arrangements. Humans will generate economic activity naturally to survive. One can only induce so much superfluous activity beyond that.

Humans have a history of ignorantly extrapolating the wrong conclusions from reality.

Reply on news.ycombinator.com