Weekend reading links

1. Amazon's Mechanical Turk is the latest in the cost-minimising innovations with massive negative externalities on society.

Employers, known as requesters, post batches of what are called Human Intelligence Tasks, or HITs, on Mechanical Turk’s website. A task could be transcribing an invoice, or taking part in a study, or labeling photographs to train an artificial intelligence program... Most tasks pay a dime or less, and there is a daily churn of tasks that pay only a penny... People turk to save for a motorcycle. They turk to buy insulin. They turk to pay off debt or pass the time profitably while on the clock at a boring job. Some do it because there are few decent-paying jobs that can be done at will. People who are confined to their homes by disability or social anxiety or who live where there are few jobs do it because, despite lousy wages, it seems like the best option. Plenty turk full time. In a 2016 Pew Research Center survey of nearly 3,000 American turkers, a quarter said they made most or all of their earned income on the platform. More than half the turkers surveyed said they earned under $5 an hour. 

As little as turking appears to pay on paper, in practice it often pays less because MTurk, as it is known, is a sloppy, shoddy free-for-all. Turkers spend their time fighting requesters over an unfair 10-cent rejection or a missing 60-cent payment. They waste minutes filling out bubbles on defective questionnaires that cannot be submitted. They abandon “10-minute” surveys after half an hour. They swap horror stories and warnings on turker message boards (“rejection on a $0.50 hit,” read one recent bulletin, “reason is ‘funds were not allocated’”). They leave scathing reviews on the turker-run site Turkopticon (“unfair and wild use of the rejection button”). Just how much turkers make is the subject of considerable scholarly debate, but one paper published last year analyzed millions of tasks done by thousands of turkers. Though they probably overrepresented novice turkers like me who do the lowest-paying tasks, the paper’s authors concluded that if you count time spent looking for tasks and working on tasks that came to nothing, the median turker’s hourly wage was $1.77... 

Presiding over this production is the world’s biggest tech company, feet firmly planted on the sidelines. Amazon usually declines to get involved when turkers say requesters rip them off, even as it lets requesters hide behind aliases that can make them impossible to track down... Amazon even finds ways to recoup some of the pennies turkers earn, a reminder of the days when miners were paid in scrip redeemable only at the company store. While American turkers can get their wages direct-deposited, thousands of turkers overseas have only one way to get paid without incurring third-party fees: on an Amazon gift card... Minimum-wage laws generally do not apply to piecework jobs like turking. Mechanical Turk is now one of a handful of big players in the field known as crowdwork or microwork. (One crowdwork company, Prolific, used by academic researchers, enforces a minimum wage: $6.50 an hour.)

Crowdwork’s proponents see a gleaming future — a borderless, no-overhead labor market where task-creator and task-doer meet at the intersection of supply and demand. Its critics see a throwback to something more Dickensian, where the lack of regulation and accountability keeps workers in the dark and on the defensive. Mechanical Turk, in particular, combines the inconsistency and precariousness of gig work with Big Tech’s tendency to dodge liability for the icky things that happen on its platforms. 

2. The Economist has a nice article on the $3.5 trillion US healthcare industry, perhaps the most famous example of why unfettered markets do not work. Sample this,

The merger wave has increased concentration and pricing power. Brent Fulton of the University of California, Berkeley, found that 90% of America’s hospital markets, representing a population of over 200m, were highly concentrated. Zack Cooper of Yale University, whose team looked at insurance claims covering over a quarter of Americans with employer-provided health insurance, discovered that prices at hospitals with a local monopoly were 12% higher than in markets with four or more rivals. A study by an insurance-industry body concluded that consolidation cut costs by 15-30% at acquired hospitals, but average prices for hospital services still rose by between 6% and 18%.

According to the American Hospital Association, a lobby group, operating margins in the industry rose from 4.4% in 2007 to 6.4% in 2017. But many hospitals in rural areas, which suffer from undercapacity, and in poor urban areas, which have lots of uninsured patients, barely break even or lose money. Big for-profit chains like hca Healthcare, with around 180 hospitals, can enjoy high (if volatile) margins. Non-profit institutions often plough those gains into expansion or salaries.

And the incentives facing the different participants,

Patients are often not price-sensitive. They are either in need of urgent care, with no time to shop around, or have insurance, and so pay a fraction of the full cost (often nothing beyond an annual out-of-pocket limit). Insurers, for their part, care less about prices because they now make more money by managing health plans for self-insured employers than by managing risk. They may even like to see inflation rise, since they can take a bigger cut from a bigger base. A well-intentioned Obamacare rule forces insurers to pay out at least 80% of their revenue from premiums. But by capping margins, it encourages raising revenue, not efficiency—and higher costs can be used to justify higher premiums.

3. Excellent essay by Paul Graham on what goes behind the creation of a genius (HT: Rajeev Mantri),

An obsessive interest in a topic is both a proxy for ability and a substitute for determination. Unless you have sufficient mathematical aptitude, you won't find series interesting. And when you're obsessively interested in something, you don't need as much determination: you don't need to push yourself as hard when curiosity is pulling you. An obsessive interest will even bring you luck, to the extent anything can. Chance, as Pasteur said, favors the prepared mind, and if there's one thing an obsessed mind is, it's prepared. The disinterestedness of this kind of obsession is its most important feature. Not just because it's a filter for earnestness, but because it helps you discover new ideas... The popular story is that they simply have better vision: because they're so talented, they see paths that others miss. But if you look at the way great discoveries are made, that's not what happens. Darwin didn't pay closer attention to individual species than other people because he saw that this would lead to great discoveries, and they didn't. He was just really, really interested in such things. Darwin couldn't turn it off. Neither could Ramanujan. They didn't discover the hidden paths that they did because they seemed promising, but because they couldn't help it. That's what allowed them to follow paths that someone who was merely ambitious would have ignored.

The important thing is to cultivate disinterested obsessive interest in something which matters!

4. Livemint points to the tanking of one of the last remaining drivers of economic growth, the state governments capital investments. 

The value of new state government projects fell 75% over last year, and is the lowest in 15 years. This, at a time when across states, government funded capex has overtaken private capex as the major source of investment. 

The article also points to how UDAY, without the complementary power sectors reforms to lower losses and raise tariffs, has both weakened state government and discom finances and lowered investments in the power sector. 

5. A stunning snippet capturing the demographic shifts taking place,

Harvard has more students at its Division for Continuing Education (for mature and retired students) than it does at the university itself.

6. Much gloss has been applied on the recent Business Roundtable Declaration by over 180 leading US corporate executives pledging their allegiance to beyond shareholders to cover employees and customers.  Sample some of the reality behind it,

Most of the CEOs who pledge to fight climate change do not run firms that are responsible for it. Take the biggest 200 Western firms that disclose emission figures. Of these, the top 20 are responsible for 70% of all emissions: the other 180 don’t matter much... the accusations of hypocrisy: it is not hard to find. Nike, which has pushed virtuous branding, has been embroiled in a doping scandal. BlackRock, a fund manager that pushes other firms to invest more, spent over 100% of its own cashflow on buybacks in the past 12 months. Visa signed the Roundtable letter championing customers, but is part of a payments oligopoly. 

7. Finally a call for industrial policy to recover America's lost manufacturing competitiveness,

This national effort—call it Industrial Policy 2.0—should focus on ensuring that hardware innovations are manufactured in this country. The idea is not to recover lost industries but to rebuild lost capabilities. The U.S. needs to leverage its dominance in science and technology to create future industries, to provide us with first-mover advantages and reclaim American leadership in manufacturing. The first step would be to create a new federal agency responsible for the health of U.S. manufacturing. A number of agencies currently have manufacturing-related programs, but there is little or no coordination or strategy. Defense alone cannot solve this challenge because defense procurement needs are dwarfed by commercial markets, and defense-specific technologies may have few commercial applications. A new agency is needed to signal new priorities. This National Manufacturing Foundation, as it could be called, would be a cabinet-level agency focused on rebuilding America’s industrial commons and translating our scientific knowledge into new products and processes.