Experts Surveyed on the Future of Work

From Pew, which says

We call this a canvassing because it is not a representative, randomized survey. Its findings emerge from an “opt in” invitation to experts who have been identified by researching those who are widely quoted as technology builders and analysts and those who have made insightful predictions to our previous queries about the future of the Internet.

Respondents gave their answers to the following prompts:

The economic impact of robotic advances and AI: Self-driving cars, intelligent digital agents that can act for you, and robots are advancing rapidly. Will networked, automated, artificial intelligence (AI) applications and robotic devices have displaced more jobs than they have created by 2025?

Please elaborate on your answer: Describe your expectation about the degree to which robots, digital agents, and AI tools will have disrupted white-collar and blue-collar jobs by 2025 and the social consequences that will emerge from that.

Bonus question: To what degree will AI and robotics be parts of the ordinary landscape of the general population by 2025? Describe which parts of life will change the most as these tools advance and which parts of life will remain relatively unchanged.

Here is how I would have answered (note the PSST slant):

1. In a market economy, trade takes place in order to gain from specialization and comparative advantage.

2. There are always such gains to be had, so in principle everyone can always participate in the market economy. In a frictionless economy, meaning that all adjustment to change takes place immediately and costlessly, everyone would do at least some market work.

3. In practice, the economy is not frictionless. One of the big frictions is that market work is taxed, whereas non-market work is not taxed. Among those who face the highest tax rates for market work are poor households who receive subsidies that disappear rapidly as income rises. This is not to say that the subsidies are bad policy (although in my opinion they are very poorly designed) but simply to report the fact.

4. Another source of friction might be termed discovery-process friction. Sustainable patterns of specialization and comparative advantage are only revealed gradually, through a trial-and-error process of new ventures launched and old ventures shut down.

5. Holding the discovery process and other frictions constant, more rapid technological change may lead to higher rates of unemployment, on average. With more rapid technological change, the economy is likely to spend more time out of adjustment than in a state of near-full employment.

6. Looking ahead, I would not rule out technological changes that help to improve the discovery process and thus reduce economic friction. For example, it might be the case that as more business data is collected and analyzed in real time, entrepreneurs will become more accurate in choosing profitable business opportunities and avoiding unprofitable ones.

7. However, I would not rule out the possibility that government policy will become increasingly dysfunctional and thus increase economic friction.

8. I think that the more fundamental question is to what extent valuable skills and capital in the economy will tend to become highly concentrated among a small elite rather than dispersed. I certainly think that this is one scenario. Thus, even though everyone may participate in the market economy, some of us could be Vickies and some of us could be Thetes.

9. Which brings me to my answer to the bonus question. I would suggest that Neal Stephenson’s The Diamond Age is more insightful than any science fiction that I might write.

Davis and Haltiwanger

It’s the paper everyone is talking about. I am on travel so I have yet to read the paper. James Pethokoukis has a helpful post. He concludes

On the downside, less labor dynamism “goes hand in hand with a slower arrival rate of new job opportunities” which “increases the risk of long jobless spells” and hampers the ability to “switch employers so as to move up a job ladder, change careers, or satisfy locational constraints.” When Americans are on the move, America is on the move. And right now, we aren’t — with evidence to suggest bad government policy shares a large chunk of the blame.

Pethokoukis also posts on a paper by David Autor that sounds interesting.

The Fed and Money Markets

Jon Hilsenrath writes,

Because banks have so much cash, the fed-funds market where they tap reserves experiences very little day-to-day trading. One New York Fed study shows daily trading volume in the market has contracted from an already-thin $200 billion before the financial crisis to nearly $50 billion. Moreover, traditional U.S. commercial banks are especially inactive. The most active players are government sponsored Federal Home Loan Banks and foreign banks.

“The fed funds market is but a shadow of what it was prior to the crisis,” Raymond Stone, an analyst at Stone McCarthy Research, said in a note to clients Wednesday. “It is no longer clear that the funds rate is the key determining factor of the behavior of short-term interest rates.”

Read the whole thing.

Hookernomics

Subtitled The Business of Sex, it is a recent book, which the author encouraged me to review. He profiles the contemporary prostitution industry and offers his views on the best policy for dealing with it. One of his suggestions is for a walled-garden approach, in which there are some areas where prostitution is at least de facto legal while there are other areas where laws against prostitution are enforced.

He offers some simple theoretical hypotheses, e.g. that better birth control technology shifts the supply curve to the right and that desirable mates will prefer marriage to prostitution, so that the median prostitute will not be a desirable mate. He offers some back-of-the-envelope estimates of various sorts of prostitution activity. Although the estimates are not based on any formal research by the author, they seem to offer useful, reasonable bounds on the true numbers.

I read the book on part of a plane ride, and I found it breezy and worth my time. However, the content is heavily weighted toward the author’s opinions, which I often found unpersuasive. The passage that stuck with me the most was this:

The obvious thing is that both prostitutes and their customers are notorious liars. Prostitutes are paid to lie. Their job is to flatter the customer and play any role he wants them to play…On one level it’s all about showbiz, and, in that context, there is nothing wrong with it.

I find myself reacting mostly to the “notorious liars” phrase, which is a complete turn-off for me. It probably explains why I do not share the author’s positive outlook on prostitution. The “showbiz” phrase suggests that one could be entertained by it, just as one can be entertained by going to a play or movie where you know that people are just acting. But when you go to a play you remain separate from what takes place on stage. When I interact with other people, I have a strong preference for authenticity.

Related: Scott Sumner discusses the daughter test, and the author takes the position that he would be ok with it. I would not.

Edward Lazear on JOLTS

He writes,

During the typical month when jobs increase by about 100,000, 5.1 million workers are hired and five million separate from their jobs, resulting in a net change of +100,000 jobs.

…more hires are needed, specifically about 5.2 million in the average month. We are just past the halfway point in getting hiring back to normal levels.

Pointer from John Cochrane.

The AS-AD story is that there is still not enough aggregate demand to stimulate enough new hires per month to overcome (new job separations per month + labor force growth + excess workers accumulated over the past six years of weak AD).

The PSST story is one of huge frictions that keep people from working in the market economy. One friction is that as technology and other factors change in the economy, the discovery process of new ventures does not keep pace. Another source of friction is government policies that reduce supply and demand in the labor market.

I doubt that there is an empirical method of distinguishing between these two stories. However, I would note that in the post-2008 period there are relatively few cases of workers getting laid off and returning to their old jobs. If there were large numbers of such cases, that would suggest that the problem of discovering new patterns of sustainable specialization and trade was relatively unimportant, and it would incline me toward the AS-AD story.

This Looks Interesting

Andrew Moseman and Carl Davis write,

Frankly, learning to play a song the Rocksmith way is exhilarating. If I (Carl) had looked up the chords online, I could have played the song just as easily. But I might have stopped to scroll down on the computer screen or to relearn the first half of the song until I got it down pat. After a few progressively more difficult play-throughs on Rocksmith, I’d memorized the song without even thinking too hard about it.

First of all, I always wish I had been a teenager at a time when you could learn guitar from YouTube, or from something like this.

Second, isn’t this the way a lot of learning could be? Teaching equals feedback, and what this product attempts to do is greatly accelerate the feedback process. My guess is that the biggest advances in education will come from something like this rather than from a MOOC.

Is the Aging of Firms Good News?

Amidst all the discussion of the trend toward fewer young firms in the economy, this question occurred to me.

I think of entrepreneurship as a trial-and-error learning process. If we had perfect knowledge in the economy, we would not need new firms. Is it possible that information about business prospects is getting better, so that we do not need as many new firms?

The analogy might be with inventories. In the 20th century, a major factor in recessions was that auto and steel manufacturers often were caught with large excess inventories, resulting in layoffs as the companies reduced production to work off the unwanted inventories. Over the last 25 years, computer and communication technology has caused the inventory cycle to be much more subdued.

What if “big data” in the economic and business environment has a similar effect on starting new ventures? Instead of having to spend years to determine whether a new business idea is viable, suppose it takes months, or can be predicted with reasonable accuracy even before it is launched?

In that kind of environment, I think that some of the advantage shifts away from entrepreneurs and toward existing businesses. As Amar Bhide pointed out, ambiguity deters large established businesses but attracts entrepreneurs. We think of rapid technological change as adding ambiguity to the environment, because change creates uncertainty. But if one of the effects of change is to make the business environment easier to read, maybe the ambiguity is lower today than it was 20 years ago.

Put this post in the “ideas I put out there but to which I am not committed” category.

UPDATE: That didn’t take long. from the comments:

I think if this were the case we would expect to see fewer of the types of entrepreneurs that don’t grow/fail quickly, but the types of entrepreneurs who grow rapidly and create a lot of jobs would not be affected (because predictions of success are becoming more accurate). This may have been the case prior to 2000, but we show that since 2000 the top of the growth rate distribution for entrepreneurs has fallen. Entrepreneurs aren’t growing as fast, and/or the high-growth types aren’t entering the market, which I think reduces the plausibility of the “lower ambiguity” theory. See http://updatedpriors.blogspot.com/2014/08/entrepreneurial-decline-mom-n-pop-or.html

The Courage to Desist

Robert Shiller writes,

I wrote with some concern about the high ratio in this space a little over a year ago, when it stood at around 23, far above its 20th-century average of 15.21. (CAPE stands for cyclically adjusted price-earnings.) Now it is above 25, a level that has been surpassed since 1881 in only three previous periods: the years clustered around 1929, 1999 and 2007. Major market drops followed those peaks.

Pointer from Mark Thoma. Here is my take:

1. I think investors look at the zero interest rate on short-term money-market instruments and say, “I can’t possibly settle for that.” As they reach for yield, long-term interest rates fall.

2. At low interest rates, long-term bonds become very speculative investments. A small decline in market interest rates, and the market value of your bonds shoots up. Conversely, it takes only a small increase in market interest rates to create a negative return on a bond mutual fund (holding the actual bond rather than a bond fund avoids marking your losses to market, but on an opportunity-cost basis, an actual bond still gives you a negative return in a rising-rate environment.)

3. If Shiller is right and stocks are over-priced, your best strategy may be to sit on those low-yielding short-term instruments and wait for prices to come down. This is hard to do. It is my strategy in fantasy baseball auctions–watch the first 50 players get chosen at prices that I think are too high, and then wait for prices to come down. If prices are too high early, this has to work, because the fixed budgets given to owners mean that prices have to come down eventually. I am not saying you win a fantasy league that way, because luck tends to dominate any advantage you might appear to gain in the auction. But if you wait, you can get good value cheap. I think that we are in that type of stock market.

4. Having said that, we know that for every credible theory that stocks are over-priced there must be an equal and opposite theory that they are under-priced. (See Brad DeLong’s response to Shiller. See also this comment that Tyler Cowen found on his blog) Otherwise, prices would not be as high as they are. The thing is, most of the movement in stock returns is due to changes in the taste/toleranace for risk, and there is no guarantee that this parameter will head toward one particular value.

Related: James Hamilton on the San Diego public employee pension fund reaching for yield.

The Problem of Monopoly

Tim Harford writes,

No policy can guarantee innovation, financial stability, sharper focus on social problems, healthier democracies, higher quality and lower prices. But assertive competition policy would improve our odds, whether through helping consumers to make empowered choices, splitting up large corporations or blocking megamergers. Such structural approaches are more effective than looking over the shoulders of giant corporations and nagging them; they should be a trusted tool of government rather than a last resort.

Pointer from Mark Thoma.

Unfortunately, I can imagine “assertive competition policy” creating more monopoly power. The problem is that government is by far the most secure monopolist. The more “assertive” is government, the more assertive is this monopolist. One can hope that this monopoly power will be exercised wisely. I, for example, hope that the government could break up big banks wisely.

But the public choice of the matter, as Harford points out, is not reliable.

Steven Braun vs. Brad DeLong

Braun as a co-author writes,

the US has enjoyed a sustained economic recovery that has exceeded most contemporaneous and historical financial crisis benchmarks. Up until a year ago, the unemployment rate was falling by an average of 0.7 percentage points per year, roughly tracking the more successful historical experiences, and well exceeding the norm following a financial crisis. In the past year, the pace of the decline in the unemployment rate has doubled. As a result, the official unemployment rate is 83% of the way back to its pre-Global Crisis average. A variety of other indicators–such as broader measures of labour market underutilisation that include discouraged workers or marginally attached workers as well as unemployment rates for different demographic groups–show similar magnitudes of recovery

DeLong writes,

But I look at 25-54 [year-olds]. The employment rate is down from 79.9% in 2007 to 76.6% in July 2014–3.3%-points less, compared to 4.0%-point a fall from 63% to 59% over the entire population. The participation rate is down from 80.8% in July 2014 compared to 83.1% for 2007–2.3%-points, compared to the 3.1%-point fall from 66.0% to 62.9% over the entire population.

Pointer from Mark Thoma.

Ordinarily, I would side with Braun, for two reasons. First, he is a gentle, soft-spoken guy. Second, he knows labor market data better than anybody I know.

In fact, right now I find DeLong’s take more compelling. But if Braun were to convince me that I am wrong, it would not be the first time he has done so.

UPDATE: DeLong strengthens his case.