Tag Archives: Work

AI Boom Bet Offers

A month ago I mentioned that lots of folks are now saying “this time is different” – we’ll soon see a big increase in jobs lost to automation, even though we’ve heard such warnings every few decades for centuries. Recently Elon Musk joined in:

The risk of something seriously dangerous happening is in the five year timeframe … 10 years at most.

If new software will soon let computers take over many more jobs, that should greatly increase the demand for such software. And it should greatly increase the demand for computer hardware, which is a strong complement to software. So we should see a big increase in the quantity of computer hardware purchased. The US BEA has been tracking the fraction of the US economy devoted to computer and electronics hardware. That fraction was 2.3% in 1997, 1.7% in 2003, and 1.58% in 2008, and 1.56% in 2012. I offer to bet that this number won’t rise above 5% by 2025. And I’ll give 20-1 odds! So far, I have no takers.

The US BLS tracks the US labor share of income, which has fallen from 64% to 58% in the last decade, a clear deviation from prior trends. I don’t think this fall is mainly due to automation, and I think it may continue to fall for those other reasons. Even so, I think this figure rather unlikely to fall below 40% by 2025. So I bet Chris Hallquist at 12-1 odds against this (my $1200 to his $100).

Yes it would be better to bet on software demand directly, and on world stats, not just US stats. But these stats seem hard to find.

Added 3p: US CS/Eng college majors were: 6.5% in ’70, 9.7% in ’80, 9.6% in ’90, 9.4% in ’00, 7.9% in ’10. I’ll give 8-1 odds against > 15% by 2025. US CS majors were: 2.4K in ’70, 15K in ’80, 25K in ’90, 44K in ’00, 59K in ’03, 43K in ’10 (out of 1716K total grads). I’ll give 10-1 against > 200K by 2025.

Added 9Dec: On twitter @harryh accepted my 20-1 bet for $50. And Sam beats my offer: 

GD Star Rating
loading...
Tagged as: , , ,

Conservative vs. Liberal Jobs

My last post got me thinking about the liberal vs. conservative slant of different jobs. Here are two sources of data.

Consider some jobs that lean conservative: soldier, police, doctor, religious worker, insurance broker. These seem to be jobs where there are rare big bad things that can go wrong, and you want workers who can help keep them from happening. That explanation can also makes some sense of these other conservative jobs: grader & sorter, electrical contractor, car dealer, trucker, coal miner, construction worker, gas service station worker, non-professor scientist. Conservatives are more focused on fear of bad things, and protecting against them.

Now consider some jobs that lean liberal: professor, journalist, artist, musician, author. Here you might see these jobs as having rare but big upsides. Maybe the focus is on small chances that a worker will cause a rare huge success. This is plausibly the opposite of a conservative focus on rare big losses.

But consider these other liberal jobs: psychiatrist, lawyer, teacher. Here the focus may just be on people who talk well. And that can also make sense of many of the previous list of liberal jobs. It might also makes sense of another big liberal job: civil servant.

I’m not suggesting these are the only factors that influence which jobs are liberal vs. conservative, but they do seem worth exploring.

Added 20Nov: This post was quoted in full at Marginal Revolution, and commenters pointed to two related data sources.

Added 3Dec: A new article with data.

GD Star Rating
loading...
Tagged as: ,

Firm Inefficiency

Economists are often stereotyped as claiming that firms are very economically efficient, i.e., that they very effectively minimize costs and maximize profits. This is a common source of derision of economists by other social scientists. And it is true that efficiency is the standard assumption made in textbooks and in math models. But over time I’ve been persuaded that it is often far from an accurate assumption. (And I doubt that most older economists believe it.)

I’ve been persuaded by a steady accumulation of plausible examples of widespread persistent inefficiencies. No one example is overwhelmingly obvious – all have stories for why they are only apparent inefficiencies. But added all together, they persuade me. Some examples:

  1. Threats Help Productivity – When firms face more competition, they often have big bursts of productivity. But if increases were possible, why not do them before?
  2. Long-Lasting Deadwood – Firms often keep employees who are widely known within the firm to not be pulling their weight relative to other employees. They tend to be fired during a downturn, or after a takeover.
  3. Not Invented Here – Firms are famously reluctant to adopt changes that appear to have been developed elsewhere, preferring instead changes for which someone internal can take credit.
  4. Shooting Messengers – Many firms greatly discourage passing bad news up to bosses. GM was just exposed as such a firm via a safety issue. Those who do pass bad news up are punished as if they were personally a big cause of the bad news.
  5. Yes Men – If bosses keep quiet about their opinion, they can evaluate subordinates via comparing employee opinions with boss opinion. But bosses consistently forgo this by telling subordinates lots of opinions and punishing those who question such opinions.
  6. Mergers & Acquisitions – Firms that buy and merge with other firms seem to consistently lose money.
  7. Poison Pills – Rules that discourage takeover attempts by financially penalizing such attempts prevent investors from getting more for their shares.
  8. Overpaid CEOs – It is far from clear that firms actually earn more when they hire more expensive CEOs.
  9. Too Many Meetings – It is widely believed that most firms hold too many meetings that go on too long with too many people.
  10. Too Many Interviews – It is hard to find much evidence that interviews add info on job performance. So why do candidates go through so many interviews?
  11. Biased Evaluations – Bosses consistently give lower evaluations to people they didn’t hire, relative to people they did hire. Yet official evaluations don’t correct for this.
  12. Excess Credentials – People consistently feel pressure to hire people whose credentials make them look good on paper, relative to people they believe would do a better job.
  13. Few Experiments – Firms tend to be reluctant to do experiments, such as to find preferred product variations. Experiments would force them to admit they don’t yet know.
  14. Few Track Records – Meetings are full of people making predictions on decision consequences, but firms almost never keep formal track records to rate accuracy.
  15. Reward Braggarts – Firms consistently neglect people who don’t toot their own horn, even when their superior features are widely known.
  16. Allow Info Silos – Groups and divisions with a firm are allowed to keep a lot of info secret within their group. Yet if the firm works together toward a common goal, what can be the benefit of keeping such secrets?
  17. Predictable Consultants – Management consultants are often hired at great expense to give advice that is quite predictable given the opinions of those who hired them.
  18. Little Telecommuting – Telecommuting seems to save big on costs, yet is not adopted much.
  19. Discrimination – Fat women are paid less, tall men paid more, in social jobs.
  20. I’ll add more here in response to suggestions.

My working hypothesis to explain these inefficiencies is that the people and supporting coalitions closest to them tend to gain from them, and that selection pressures on political coalitions are often much stronger than selection pressures on firms.

If many of these inefficiencies are real, then yes government regulators can also see them, and yes it might not be that hard for smart sincere people to design regulations to increase welfare by correcting for them. However, government regulatory agencies are also “inefficient” in many ways, leading them to choose and enforce regulations which differ from those that would most increase welfare. To judge if we are better off giving regulators more powers over firms, we must judge the relative magnitudes of these two types of inefficiencies.

Note that firm efficiency may still be a reasonable assumption to make in models, even if it is not an accurate assumption. Modeling is always a tradeoff between realism and understanding.

GD Star Rating
loading...
Tagged as: , ,

Trustworthy Telepresence

In a recent Ipsos/Reuters poll, which questioned 11,383 people in 24 countries, about half believed that they would be at a disadvantage in earning promotions because of the lack of face-to-face contact. Previous research suggests part-time telecommuters do not communicate less frequently with managers. … After four years of experience, the average male telecommuter will earn about 6.9% less than a non-telecommuter. (more)

Telecommuting requires the use of various types of media to communicate, such as the telephone and email. Emails have a time lag that does not allow for immediate feedback; telephone conversations make it harder to decipher the emotions of the person or team on the phone; and both of these forms of communication do not allow one to see the other person. Typical organization communication patterns are thus altered in telecommuting. For instance, teams using computer-mediated communication with computer conferencing take longer to make group decisions than face-to-face groups. (more)

Decades ago many futurists predicted that many workers would soon telecommute, and empty out cities. Their argument seemed persuasive: workers who work mainly on computers, or who don’t have to move much physical product, seem able to achieve enough coordination to do their jobs via phone, email, and infrequent in-person meetings. And huge cost savings could come from avoiding central city offices, homes near them, and commuting between the two. (For example, five firms might share the same offices, with each firm using them one day per week.)

But it hasn’t remotely happened that way. And the big question is: why?

Some say telecommuters would shirk and not work as much, but it is hard to see that would remain much of a problem with a constant video feed watching them. Bryan Caplan favors a signaling explain, that we show up in person to show our commitment to the firm. But a firm should prefer employees who show devotion via more total work, instead of wasting hours on the road. Yes inefficient signaling equilibria can exist, but firms have many ways to push for this alternate equilibrium.

The standard proximate cause, described in the quote above, is that workers and their bosses get a lot of detailed emotional info via frequent in-person meetings. Such detailed emotional info can help to build stronger feelings of mutual trust and affiliation. But the key question is, why are firms willing to pay so much for that? How does it help firm productivity enough to pay for its huge costs?

My guess: frequent detailed emotional info helps political coalitions, even if not firms. Being able to read detailed loyalty signals is central to maintaining political coalitions. The strongest coalitions take over firms and push policies that help them resist their rivals. If a firm part adopted local policies that weakened the abilities of locals to play politics, that part would be taken over by coalitions from other parts of the firm, who would then push for policies that help them. A lack of telecommuting is only one of a long list of examples of inefficient firm policies than can be reasonably be attributed to coalition politics.

Some people hope that very high resolution telepresence could finally give enough detailed emotional info to make telecommuting workable. And that might indeed give enough info to build strong mutual trust and loyalty. But it is hard to make very high resolution telepresence feel natural, and we still far from having enough bandwidth to cheaply send that much info.

Furthermore, by the time we do we may also have powerful robust ways to fake that info. That is, we might have software that takes outgoing video and audio feeds and edits them to remove signs of disloyalty, to make people seem more trusting and trustworthy than they actually are. And if we all know this is possible, we won’t trust what we see in telepresence.

So, for telepresence to actually foster enough loyalty and trust to make telecommuting viable, not only does it need to feel comfortable and natural and give very high bandwidth info, but the process would need to be controlled by some trusted party, who ensures that people aren’t faking their appearances in ways that make it hard to read real feelings. Setting up a system like that would be much more challenging that just distributing something like Skype software.

Of course eventually humans might have chips under their skin to manipulate their sight and sound in real physical meetings. And then they might want ways to assure others aren’t using those. But that is probably much further off. (And of course ems might always “fake” their physical appearance.)

Again, I have hopes, but only weak hopes, for telepresence allowing for mass human telecommuting.

Added 3July: Perhaps I could have been clearer. The individual telecommuter could clearly be at a political disadvantage by not being part of informal gossip and political conversation. He would have fewer useful allies, and they would thus prefer that he or she not telecommute.

GD Star Rating
loading...
Tagged as: , , ,

Days Of Our Lives

Oedipus famously answered this riddle:

What goes on four feet in the morning, two feet at noon, and three feet in the evening?

The answer: people crawl when babies, walk as adults, and use a cane when old. It seems natural to divide lives into three parts: young, middle, and old. But where exactly should the boundaries fall? One tempting approach comes from the facts that in the US today lifespans average about 29000 days, and people typically marry and have kids at about 10000 days. So maybe we should split life into the first, second, and third 10000 days.

If we split life into 5000 days units, we get:

  • 0 days; 0 years – Birth
  • 5000 days; 13.7 years – Mid-puberty
  • 10000 days; 27.4 years – First marriage & kids
  • 15000 days; 41.1 years – Start to notice body decline
  • 20000 days; 54.8 years – Near kids’ first marriage & kids, own peak of relative income, productivity, 90% still alive
  • 25000 days; 68.5 years – Near when most retire, 75% still alive
  • 30000 days; 82.1 years – Typical death age, 42% still alive
  • 35000 days; 95.8 years – Only 4% still alive

Note that 5000 days is near the doubling time of the world economy.

In my life, I married at 10250, had my first kid at 11500, started grad school again at 12400, started at GMU at 14600, and was tenured at 16540. And today I am 20,000 days old, within a few days of all my kids being employed college graduates. So a lot happened to me in that third 5000 days, and I now enter the last third of a typical lifespan, with expected declining (but hardly zero) relative productivity. Of course if cryonics works I might live lots longer.

GD Star Rating
loading...
Tagged as: , ,

SciCast Pays Out Big!

When I announced SciCast in January, I said we couldn’t pay participants. Alas, many associated folks are skeptical of paying because they’ve heard that “extrinsic” motives just don’t work well relative to “intrinsic” motives. No need to pay folks since what really matters is if they feel involved. This view is quite widespread in academia and government.

But, SciCast will finally do a test:

SciCast is running a special! For four weeks, you can win prizes on some days of the week:
• On Wednesdays, win a badge for your profile.
• On Fridays, win a $25 Amazon Gift Card.
• On Tuesdays, win both a badge and a $25 Amazon Gift Card.
On each prize day 60 valid forecasts and comments made that day will be randomly selected to win (limit of $575 per person).
Be sure to use SciCast from May 26 to June 20!

Since we’ve averaged fewer than 60 of these activities per day, rewarding 60 random activities is huge! Either activity levels will stay the same and pretty much every action on those days will get a big reward, or we’ll get lots more activities on those days. Either you or science will win! :)

So if you or someone you know might be motivated by a relevant extrinsic or intrinsic reward, tell them about our SciCast special, and have them come be active on matching days of the week. We now have 473 questions on science and technology, and you can make conditional forecasts on most of them. Come!

Added 21May: SciCast is mentioned in this Nature article.

GD Star Rating
loading...
Tagged as: ,

Should You Kiss Ass?

Consider two possible work strategies. One strategy is just to try to do a good job. The other is to try to kiss ass and please your boss any way you can. Of course you can try either strategy, both, or neither. Which makes four different kinds of workers. Now ask yourself, of these four kinds of workers, which ones do you think achieve the most career success? Which ones have the most job and life satisfaction?

I came across a fascinating paper (ungated here) from 1994 that asked exactly this question. Looking at 500 ex students of industrial relations, they compared the effect of ass-kissing to doing a good job on success and job satisfaction.

Supervisor-focused tactics … include: agree with your immediate supervisor’s ideas; praise your immediate supervisor on his or her accomplishments; agree with your supervisor’s major opinions outwardly even when you disagree inwardly. Job-focused tactics … include: make others aware of your accomplishments in your job; try to take responsibility for positive events even when you are not solely responsible; arrive at work early in order to look good in front of others.

The result: workers who try to please their boss are more successful in their careers, and workers who try to seem good at their jobs are less successful. Boss-pleasers are also more satisfied with their job and life, while good-jobbers aren’t any more or less satisfied.

The only other thing that predicted satisfaction: being married. Other things that predicted job success: being married, being on the job many years, working more hours per week, and not having a PhD.

We like to act like we just want to do a good job, and would rather not have bosses breathing down our necks. But what if, we actually like kissing ass?

Please speak up if you know of any more recent that might confirm or disconfirm these results.

GD Star Rating
loading...
Tagged as: , ,

More Broken Evals

Back in January I quoted:

In 1980, economists … observed that salaries in companies were more strongly related to age and organizational tenure than they were to job performance. Ensuing research has confirmed and extended their findings, both in the United States and elsewhere. … One meta-analysis of chief executive compensation found that firm size accounted for more than 40 percent of the variation in pay while performance accounted for less than 5 percent. (more)

Part of the reason may be that employee performance evaluations are often political. From an ’87 paper:

Our research approach involved in-depth, semi-structured interviews with 60 executives. … from seven large organizations and represented 11 functional areas. As a group, they averaged more than 20 years of work experience and more than 13 years of managerial experience. ..

Executives admitted that political considerations nearly always were part of the [employee] evaluation process. One vice-president summarized the view these executives shared regarding the politics of appraisal:

As a manager, I will use the review process to do what is best for my people and the division. … I’ve got a lot of leeway – call it discretion – to use the process in that manner. … I’ve used it to get my people better raises in lean years, to kick a guy in the pants if he really needed it, to pick up a guy when he was down or even to tell him that he was no longer welcome here. It is a tool that the manager should use to help him do why it takes to get the job done. I believe most of us here at —- Operate this way regarding appraisals. … Accurately describing an employee’s performance is really not as important as generating ratings that keep things cooking.

Executives suggested several reasons why politics were so pervasive and why accuracy was not their primary concern. First, executives realized that they must live with subordinates in a day-to-day relationship. Second, they were also very cognizant of the permanence of the written document. .. Perhaps the most widespread reason … was that the formal appraisal was linked to compensation, career, and advancement in the organization. …

Executives generally believed the appraisal process became more political and subjective as one moved up the organizational ladder:

The higher you rise in this organization the more weird things get with regard to how they evaluate you. … The process becomes more political and less objective and it seems like the rating process focuses on who you are as opposed to what you’ve actually accomplished … As the stakes get higher, things get more and more political. ..

Although not frequently reported, a few executives admitted to giving a higher rating to a problem employee to the get employee promoted “up and out” of the department. …

A deliberately deflated rating was sometimes used to teach a rebellious subordinate a lesson. … Deflated ratings were also used as part of a termination procedure. First, a strongly negative rating could be used to send and indirect message to a subordinate that he or she should consider quitting. …. Second, once the decision has bee made that the situation was unsalvageable, negative ratings could then be used to build a strongly documented case against the marginal or poor performer.

GD Star Rating
loading...
Tagged as: , ,

Who Gains From Grit?

I’ve often said that while foragers did what felt natural, farmer cultures used religion, conformity, self-control, and “grit,” to get farmers do less-natural-feeling things. But as we’ve become rich over the last few centuries, we’ve felt those pressures less, and revived forager-like attitudes. Today “conservatives” and “liberals” have farmer-like and forager-like attitudes, respectively. I think the following recent quotes support this view.

Tyler Cowen says workers today have less grit:

There is also a special problem for some young men, namely those with especially restless temperaments. They aren’t always well-suited to the new class of service jobs, like greeting customers or taking care of the aged, which require much discipline or sometimes even a subordination of will. (more)

There were two classes of workers fired in the great liquidity shortage of 2008-2010. The first were those revealed to be not very productive or bad for firm morale. They skew male rather than female, and young rather than old. … There really are a large number of workers who fall into the first category. (more)

Alfie Kohn says grit is overrated:

More than smarts, we’re told, what kids need to succeed is old-fashioned self-discipline and willpower, persistence and the ability to defer gratification. … The heart of what’s being disseminated is a notion drummed into us by Aesop’s fables, Benjamin Franklin’s aphorisms, Christian denunciations of sloth and the 19th-century chant, “If at first you don’t succeed, try, try again.” …

On closer inspection, the concept of grit turns out to be dubious, as does the evidence cited to support it. Persistence can actually backfire and distract from more important goals. Emphasizing grit is usually justified as a way to boost academic achievement, which sounds commendable. Indeed, research has found that more A’s are given to students who report that they put off doing what they enjoy until they finish their homework. Another pair of studies found that middle-schoolers who qualified for the National Spelling Bee performed better in that competition if they had more grit, “whereas spellers higher in openness to experience, defined as preferring using their imagination, playing with ideas, and otherwise enjoying a complex mental life,” did worse.

But what should we make of these findings? If enjoying a complex mental life interferes with performance in a contest to see who can spell the most obscure words correctly, is that really an argument for grit? And when kids persist and get good grades, are they just responding to the message that when they do what they’ve been told, they’ll be rewarded by those who told them to do it? Interestingly, separate research, including two studies Duckworth cites to argue that self-discipline predicts academic performance, showed that students with high grades tend to be more conformist than creative. That seems to undermine not only the case for grit but for using grades as markers of success…

Moreover, grit may adversely affect not only decisions but the people who make them. Following a year-long study of adolescents, Canadian researchers Gregory Miller and Carsten Wrosch concluded that those “who can disengage from unattainable goals enjoy better well-being . . . and experience fewer symptoms of everyday illness than do people who have difficulty disengaging from unattainable goals.” …

Finally, the concept isn’t just philosophically conservative in its premise but also politically conservative in its consequences. The more we focus on trying to instill grit, the less likely we’ll be to question larger policies and institutions. (more)

Yes, grit is conservative, and gritty people may not be as playful, open, relaxed, or creative. Grit just helps individuals to succeed, and societies to get ugly things done, like winning their competitions with other societies. But yes, you might be happier to play video games in your parent’s basement, leaving the support of society to someone else.

GD Star Rating
loading...
Tagged as: , , ,

Me On Fox Friday

Friday I’ll appear on The Independents, which airs on Fox Business TV at 9pm EST, discussing “The Rise Of The Machines.”

GD Star Rating
loading...
Tagged as: , ,