The white collar job apocalypse did not happen

In a follow-up paper released Friday, another economist, Adam Ozimek, revisited Mr. Blinder’s analysis to see what had happened over the past decade. Some job categories that Mr. Blinder identified as vulnerable [to offshoring], like data-entry workers, have seen a decline in United States employment. But the ranks of others, like actuaries, have continued to grow.

Over all, of the 26 occupations that Mr. Blinder identified as “highly offshorable” and for which Mr. Ozimek had data, 15 have added jobs over the past decade and 11 have cut them. Altogether, those occupations have eliminated fewer than 200,000 jobs over 10 years, hardly the millions that many feared. A second tier of jobs — which Mr. Blinder labeled “offshorable” — has actually added more than 1.5 million jobs.

But Mr. Blinder didn’t miss the mark entirely, said Mr. Ozimek, who is chief economist at Upwork, an online platform for hiring freelancers. The new study found that in the jobs that Mr. Blinder identified as easily offshored, a growing share of workers were now working from home. Mr. Ozimek said he suspected that many more were working in satellite offices or for outside contractors, rather than at a company’s main location. In other words, technology like cloud computing and videoconferencing has enabled these jobs to be done remotely, just not quite as remotely as Mr. Blinder and many others assumed.

Here is more from Ben Casselman (NYT).


Aren't the disemployment effects of automation/AI bounded by those of globalization? In the worst case scenario, a general AI is built that, perhaps after some up significant front costs, has small marginal costs (electricity, licensing fees for the IP maybe), has roughly the same or superior intelligence compared to the average American laborer, never takes vacation, doesn't need fringe benefits, works 24/7 on demand... but you have just described the urban Asian workforce! So in kind, globalization is already the worst case outcome of automation and AI.

Obviously, there may be differences in degree in how widely GAI would be adopted versus how wide globalization has been adopted.

And perhaps the worst case for GAI isn't roughly human level intelligence, but rather a Kurzweil/Hanson singularity type accelerating intelligence... fair enough, I don't know how to price that.

AI luddites are lame.


AI is inevitable, best to keep an agnostic approach and see where it naturally permeates and where it flops.

extracting, sorting and coding qualitative data may be a nightmare, but algorithms in quantitative data can capture unique patterns pretty quickly. I suspect the job market will see growth in opportunities in data science as "add-on" or supplementary to most white-collar jobs.

You cuckolds need to worry more about the black penis than the white collar.

don't implicate me; I'm innocent.

As is so frequent, I see no mention of "immigration" in Ozimek's paper, even though it should be obvious than many white collar jobs, such as coding, have been substantially inshored by use of immigrants.

But why are foreign born software developers more valuable in the US than in India and China? I think more is going on than just a story of immigrants holding down wages, because if that were the case, than why would they be paid a premium over their fellow countrymen doing development overseas?

One of the biggest hurdles in software development is clear communication of the needs of the end user to the coder. Adding thousands of miles and 8 time zones offset does nothing helpful.

In other words, technology like cloud computing and videoconferencing has enabled these jobs to be done remotely, just not quite as remotely as Mr. Blinder and many others assumed.

Cultural, language, and time-zone differences are real.

+1. I've seen some non-trivial coding outsourced to Asia where managing the project was painful and the results mediocre, precisely because of the reasons you cite.

Agreed, my wife works in software UX (working directly with clients’ customers to understand their needs and how to improve users’ software experience, so tougher to offshore). Her experience working with Indian developers has been consistently negative, that it frequently ends up costing more than if they had US-based programmers

A lot of Indian software developers are actually just terrible as well, trained to just copy-paste code.

I used to hire programmers for industrial automation work. It's hard enough to get Americans who live in the same time zone to understand what you want and deliver what you've asked for. I can't imagine trying to do so with language, cultural and time zone barriers. The price of failure is too high.

The principle barrier to off-shoring work is the difficulty in writing a really good specification, particularly for software. Because so few really good specifications are written, the quality of the resulting work depends highly on the contractors ability to communicate effectively with the client.

Writing specs is rather old-fashioned, the "waterfall model". Better to "just code" and figure out the spec on version 2.0. At least that's one enlightened "Tier V" level programming approach. It's how I code too. As for copy-and-paste, everybody who codes does that to a degree.

Burden trivia: static instantiation maybe milliseconds faster than dynamic instantiation, but at some point with a higher level language you have to do OOP, no? Static instantiation is for Pascal, C, not higher level languages. That said, they once spent $100s of millions coding up the LSE, using C# an OOP with automatic garbage collection and then the algorithmic traders said that a millisecond faster was all the difference in the world, and they had to scrap the whole project and go to C++ (native) or even C machine level code. Happened about 10-15 years ago.

Old joke, “you guys start coding, and I’ll go find out what they want”.

Code first, spec later may work for some things, but if you are Boeing, or doing medical equipment, or have security requirements , or ... it’s maybe not the way to go.

"Code first spec never" works for companies like Google, Microsoft, Apple, Amazon, you know trillion dollar companies. Boeing does waterfall with offshore Indians and their planes still fall from the sky. Corruption and quality of management matter more than methodology.

Intellectual jobs will forever be done in OECD+East Asia.

Many jobs are being outsourced, just not entirely offshored. Being outsourced has advantages (flexibility) and disadvantages (no benefits). I'd point out that mfg. output in America has spiked since the great recession while GDP has not. Somebody has to record the mfg. output even if the intermediate goods are produced offshore.

You 'point out' something patently untrue. GDP was $14.72T in 2008, $20.5T in 2018. Pretty nice spike.

If by outsourced you are referring to contract jobs where a US company "rents" workers to another firm, those jobs may well have benefits, depending on the skill levels needed, albeit inferior to the standard benefits the contractee employer would offer. I've done contract/consulting work in the past and the jobs came with health benefits, paid holidays and vacation and sick days. Also, such jobs are often used as trial employment periods and the contract person may be hired as a full employee eventually, which is how I started at my current job.

Actuaries are protected by occupational licensing. That's what stops them being offshored. That's why it's suicide for people to do what libertarians want and give up occupational licensing.

While there are probably actuaries who are well compensated to rubber stamp work just because they have a credential, I don't think this is sufficient.

For one, it's a very very hard credential to get. You've got to have high IQ and high conscientiousness to slog through all those exams. It's not like this is a barbers license.

In addition there is no "limit" on the number of actuaries. Med Schools may ration doctors, but anyone that wants to take the actuarial exam can. The only filter is that they are hard to pass. Moreover, nobody has to attend expensive schooling to be an actuary. Most people self study and are even paid to do so while they are employed.

For another, insurance is an industry where there is always a lot of pressure to fudge the numbers to make risks seem smaller than they are. It's best to have a group of highly paid highly intelligent professionals whose job is to make sure that doesn't happen. Undercapitalized insurance is a fraud that effects people when they need it most.

Finally, bad actuarial work can easily cost 1,000x whatever the actuaries salary is. These people are typically decision makers on big financial transactions, do you really want to skimp on that?

Slogging through exams is a perfectly acceptable way of life in many Asian countries.

95%+ of actuarial work is not protected by occupational licensing. Legally, anyone can do my current job. The only real thing that can preserve the earning power of actuaries is their own competentcy. Data scientists and other such professions could replace us if they could do the job better (though having inertia in our favor is a huge advantage).

Isn’t there defacto occupational licensing through legal statutes about professional competence and via case law? That is, either you are required by law to hire certified actuaries, or if you don’t you will be in big legal jeopardy.

Mfg is up 16% and GDP is up 30% since 2009

To Rayward ^^

GDP per capita is up 14%.

Industrial production peaked at 105.1 in 2007 and is now at 109.2

I personally would not call a 4% gain over 12 years a spike.

TMC-- where did you get your data?

The Fed says manufacturing peaked at 110 in 2007 and is now at 104.7, of below its 2007 peak.

I went with right after, Q4 2009 at 92, though Q2 was 88.

Also, this link has Q2 2019 as 107.

Here is the data you linked to for manufacturing output.
The formatting probably will be bad, but it shows output in 2007 at 107.4 and 107.7 in 2019. This is quarterly data and I was using monthly data.

2007-01-01 106.147
2007-04-01 107.797
2007-07-01 108.110
2007-10-01 108.361
2008-01-01 107.338
2008-04-01 105.200
2008-07-01 101.590
2008-10-01 96.113
2009-01-01 90.577
2009-04-01 88.745
2009-07-01 90.915
2009-10-01 92.504
2010-01-01 93.923
2010-04-01 96.210
2010-07-01 97.153
2010-10-01 97.455
2011-01-01 98.408
2011-04-01 98.320
2011-07-01 99.107
2011-10-01 99.498
2012-01-01 100.075
2012-04-01 99.929
2012-07-01 99.698
2012-10-01 100.298
2013-01-01 101.696
2013-04-01 102.132
2013-07-01 102.284
2013-10-01 102.864
2014-01-01 102.560
2014-04-01 103.613
2014-07-01 104.018
2014-10-01 104.165
2015-01-01 103.328
2015-04-01 102.981
2015-07-01 103.099
2015-10-01 102.580
2016-01-01 102.810
2016-04-01 102.412
2016-07-01 102.872
2016-10-01 103.574
2017-01-01 104.333
2017-04-01 105.254
2017-07-01 104.867
2017-10-01 106.333
2018-01-01 106.838
2018-04-01 107.539
2018-07-01 108.610
2018-10-01 109.037
2019-01-01 108.550
2019-04-01 107.716

Yep. rayward just said the recession. You went for the beginning, I used the end.

Yes, you are right.

I went for the more standard convention of comparing to the previous peak father than the bottom. My way will always show less growth than you way. But we are both right in trying to correct
Rayward This far into an expansion I would prefer to measure growth from the prior peak.

Blinders, like most economists, have forgotten, or mislearned what zero sum means.

Economies are zero sum.

Here's a question economists never ask, much less answer.

Who puts money in consumer pockets?

Milton Friedman argued businesses, employers, should not put money in consumer pockets, and in fact, managers must do everything possible to NOT put money in consumer pockets.

According to Friedman, profit, money not put in consumer pockets, but from dipping deeply into consumer pockets is their moral duty.

Profit is the revenue never paid to workers.

Friedman did argue for government putting money in consumer pockets, but he supported politicians who called for ending all government putting money into consumer pockets, eg, eliminating Social Security, the biggest government program putting money in consumer pockets. Gas taxes always put money in consumer pockets quickly, but that is opposed by those who were inspired by Friedman.

Trump seems to argue the Fed should put lots of money in consumer pockets by increasing consumer debt faster, and also supporting faster government putting money in consumer pockets.

In my experience, managers aare not idiots so they work hard to put money in their consumer coworker friends, creating a network of current and future managers who will keep putting money in their consumer pockets.

So, by hook or crook, managers will not abide by economists who tell them to cut costs by not putting money in consumer pockets.

Managers, who are white collar, keep paying their peer white collar workers, because white collar workers can be good consumers only if businesses put money in their pockets.

Everyone wants lots of consumers with lots of money in their pockets to spend.

I've never heard many economists calling for cost cutting by cutting consumer spending year after year.

Someone needs to put more money in consumer pockets year after year to have GDP growth.

If not employers, who?

You will mulp, when you become president in 2020.

No, economies are NOT zero sum. We've only had two centuries of experience and study proving that beyond any doubt.

TMC -where do you get your data?

The Fed says manufacturing peaked at 110 in 2007
and is now at 104.7, still below its 2007 peak.

"Alan, because of his stature, provided a degree of legitimacy to what many of us had come to feel anecdotally―that the anxiety over outsourcing and offshoring was a far larger phenomenon than traditional economic analysis was showing."

sounds like famed economist Blinder (and many others) was just simply wrong.

faulty expert economic predictions are so very rare (?)

We are not done yet. Blinder was likely early.

In the long run, Blinder's dead.

I think one lesson of the past decade is that lowered bargaining power / worse conditions for employees does not necessarily take the form of unemployment.

Leftists always predict their hopes.

Leftist they ain't. When the cities are being bombed, oil production bombed and the banks liquidated, you will see a "leftist".
Capitalism basically has 3 body parts:
1.The Banks are its brain
2.Cities are its heart
3.Oil is its blood.
You kill one, you kill them all.

Is that why productive growth has been slow?

I don't think it matters. Capitalism is a jewish debt ponzi scheme that the Church unleashed and the Monarchs had no choice but to accept after the black death. Capitalism is the debt based system. Without debt it doesn't work. Offshore, onshore, it doesn't matter. Whole fields of jobs have surged due to offshoring and its supply chain network. Do the chicken little's realize that? Debt expansion into investment is what matters.

Sadly the ole Germanic tribes have become slaves of the jewish debt system, living a "pampered" like to the non-Germanics as house pets. What a disgrace we have become. Germanic people need to return to the tribal ways of before Christ myth ravaged our souls. No more bourgeois leg humping, just savage tribal warfare. That is why I can respect the black and halfbreeded hispanic gangs in the cities. They live a tribal code that we have long forgotten. Death is simply nature. Tribal warfare is what leads to scientific progress on large scale.

Someone sane and rational posting here.

This is a 6/10 troll post. Next time, leave off the "Tribal warfare is what leads to scientific progress on large scale" sentence at the end.

anonymous finally went over the edge?

Did I miss something? Don't invest too much in your fantasy versions of "anonymous."

As long as managers are paid in proportion to the number of people they manage, they will work hard to increase the size of the heirarchy reporting to them.

The idea that managers generally think about creating demand in the economy by paying their workers well is almost as silly as the incoherent bs The Rage spews.

Comments for this post are closed