Month: August 2013

Computers which magnify our prejudices

As AI spreads, this will become an increasingly important and controversial issue:

For one British university, what began as a time-saving exercise ended in disgrace when a computer model set up to streamline its admissions process exposed – and then exacerbated – gender and racial discrimination.

As detailed here in the British Medical Journal, staff at St George’s Hospital Medical School decided to write an algorithm that would automate the first round of its admissions process. The formulae used historical patterns in the characteristics of candidates whose applications were traditionally rejected to filter out new candidates whose profiles matched those of the least successful applicants.

By 1979 the list of candidates selected by the algorithms was a 90-95% match for those chosen by the selection panel, and in 1982 it was decided that the whole initial stage of the admissions process would be handled by the model. Candidates were assigned a score without their applications having passed a single human pair of eyes, and this score was used to determine whether or not they would be interviewed.

Quite aside from the obvious concerns that a student would have upon finding out a computer was rejecting their application, a more disturbing discovery was made. The admissions data that was used to define the model’s outputs showed bias against females and people with non-European-looking names.

The truth was discovered by two professors at St George’s, and the university co-operated fully with an inquiry by the Commission for Racial Equality, both taking steps to ensure the same would not happen again and contacting applicants who had been unfairly screened out, in some cases even offering them a place.

There is more here, and I thank the excellent Mark Thorson for the pointer.

Is “outside” overrated?

People are going to read this as a kind of #Slatepitch but as I live my life I see that despite all the talk about how great it is to be outside, people don’t really put their money where their mouth is. Property owners are much more likely to build an addition to their house (thereby increasing the inside/outside ratio of their property) when they’re feeling flush than to orchestrate a subtraction in order to get more open space. Even in places like Southern California where there’s really great weather all the time, the landscape is covered with buildings and people spend a very large sum of their income on purchasing or renting inside space in which to live over and above the inside space in which they work.

Here is much more, from Matt Yglesias, who explains why “inside” has “crucial advantages.”  For a useful discussion on this matter I thank also Claire Hill.

Fiscal Sentiment and the Weak Recovery from the Great Recession

That is a new paper (pdf) by Finn Kydland and Carlos Zarazaga, the abstract is here:

The U.S. economy isn’t recovering from the deep Great Recession of 2008-2009 with the anticipated strength. A widespread conjecture is that this weakness can be traced to perceptions of an imminent switch to a higher taxes regime. The paper explores quantitatively this fiscal sentiment hypothesis. The main finding is that the hypothesis can account for a significant fraction of the decline in investment and labor input in the aftermath of the Great Recession, relative to their pre-recession trends. These results require, however, a qualification: The perceived higher taxes must fall almost exclusively on capital income.

In general I don’t buy it, mostly because I don’t think the higher taxes on capital income are coming that soon.  I see the higher risk premium from the crisis itself, the fracturing of the intangible capital behind credit relationships, slow underlying productivity growth, the labor-saving nature of recent innovation, and a growing sense of the dysfunctionality of U.S. government as the main culprits behind the slow recovery.  Still, the Kydland and Zarazaga paper brings us further down its stated path than I would have thought possible.

Why are new unemployment claims so low?

Scott Sumner reports:

…the ratio of new [unemployment] claims to pop is roughly back to the boom levels of 1999-2000 and 2006-07.   And yet the other indicators (total jobs, unemployment rate, etc), remain deeply depressed.  I can think of two ways to interpret this data:

1.  Casey Mulligan is right, we have lots of structural issues that are causing high unemployment right now.  The job market’s not that bad, it’s just that lots of people don’t want to work at the wages being offered, or are frozen out by the 40% rise in minimum wages during the housing bust.

2.  AD is still the main problem, but since 1975 there’s been a long term downward trend in the claims/pop ratio, for some mysterious reason.  That trend would explain why (according to new claims) the labor market looked as good in 2006 as 2000, even though most people think it was not.

On this topic, here is a kebko post of interest; he argues that employment has more or less recovered, once we adjust for various obstacles.

My favorite things Java

1. Novel: Pramoedya Toer, The Buru Quartet.  I like this better than say Mahfouz or Rushdie.  It focuses on the key question of what a life may be said to consist of, and each of the four volumes makes the others richer.  Another famous Indonesian novel is Andrea Hirata, The Rainbow Troops, it is good not great.

2. Gamelan music CD: Java: Court Gamelan.  Dreamy and beautiful, this is one of my favorite world music CDs.

3. Popular music: The Smithsonian has a very good 17-CD set of Indonesian popular music.

4. Classical pianist: Eduardus Halim.

5. Movie: Raid: The Redemption.  Martial arts reign, after the police storm an apartment fortress.  If you like the genre at all, this one is special.  Note that The Act of Killing is set mainly in Sumatra and thus does not count.

6. Movie, set in: The Year of Living Dangerously is the obvious choice.  I believe the old “Krakatoa, East of Java” is set…east of Java.

7. Painter: I can’t say I love Affandi, so I must be missing someone.

Xaviera Hollander deserves mention, although in what category should she be my favorite?  Old Javanese Gold is excellent, plus there is puppetry.  The best textiles are from Sumatra, or from the smaller islands.

Overall I have long felt a kinship to this culture, and I am excited to be visiting Java for the first time.

Jakarta fact of the day

…9.9 million cars, motorcycles, trucks and other vehicles take to the capital’s streets each workday, according to the Jakarta Transportation Agency, nearly two million of them from neighboring municipalities in the provinces of West Java and Banten.

The city has no mass transit system.  There is more here, and for the pointer I thank Claire Morgan.

Last week, however, was that of the yearly mass abandonment of the city:

Around 9.7 million people from Jakarta “abandoned” the capital and left for their hometowns during the annual Idul Fitri exodus, an increase of 8.6 percent from last year, resulting in empty roads in almost all parts of the city.

That story you will find here.

Raising Rival’s Costs and Reform in the Public Interest

How can we achieve reform in the public interest when the public is rationally ignorant and unorganized while the special interests are informed, organized and well funded?  Matt Yglesias draws some interesting lessons and hope (!) from my paper on The Separation of Commerical and Investment Banking: The Morgans vs. the Rockefellers (pdf).

Yglesias offers a brief summary of the paper:

The basic story is that the Depression led to a lot of public outrage about the financial system and the outrage was—as outrage tends to be—a little bit inchoate and not really focused on the fine-grained details of public policy. Meanwhile, the Rockefeller family and the Morgan family had some longstanding business conflicts between their respective empires. And the Glass-Steagall bill was essentially an effort by the Rockefellers to channel that inchoate public outrage in a direction that would harm the Morgans:

More than anyone else, Winthrop Aldrich, representative of the Rockefeller banking interests, was responsible for the separation of commercial and investment banking. With the help of other well-connected anti-Morgan bankers like W. Averell Harriman, Aldrich drove the separation of commercial and investment banking through Congress. Although separation raised the costs of banking to the Rockefeller group, separation hurt the House of Morgan disproportionately and gave the Rockefeller group a decisive advantage in their battle with the Morgans.

He then draws an interesting conclusion:

Tabarrok notes that when this kind of regulatory strategy is pursued in a given industry “the industry as a whole will shrink” even while one firm gains an advantage over its rivals. And here we have actually an answer to a question that’s troubled me for years: How, given political realities, can the financial sector ever be brought to heel?… It shows a way that smart and savvy would-be regulators can find ways to undermine sector-level political solidarity. Not just in ways that favor one firm against another (which would be pointless) but even in ways that shrink the sector as a whole.

Here’s the big picture. Under certain conditions, free markets channel self-interest towards the social good – that is the meaning of the invisible hand theorem. Unfortunately, there is no invisible hand theorem for politics. There are institutions, such as democracy, checks and balances and an independent judiciary, which help to channel political self-interest if not to the public good then at least away from the public evil. Even given the right macro institutions, however, breaking the iron triangle of politics is difficult. Industry self-interest and the public interest will typically align only accidentally. Universities are not less self-interested than any other actors but support for basic research is (arguably) in the public interest. The usual situation, however, is that industry self-interest pushes well beyond the point of alignment with the public interest. At current spending levels, lobbying by defense firms does not benefit the public even if national defense is a public good.

Yglesias is interested in the most difficult case when the public interest favors not a larger but a smaller industry. Will industry self-interest every align with a smaller industry? Rarely but if public anger against an industry is high then some industry participants may see that a smaller industry is consistent with their self-interest if their share of the industry grows enough as the industry shrinks–a bigger share of a smaller pie. This is the theory of raising rival’s costs that I argue led to the Glass-Steagall Act. Other examples of raising rival’s costs are firms like Costco, that already pay high wages advocating for increases in the minimum wage.

Could we apply this strategy to the provision of other public goods? Here’s an idea. The best political strategy to combat global climate change may be to bring the cleaner parts of the energy industry into a coalition with environmentalists to support a carbon tax. That means bringing the environmentalists together with the nuclear, hydro-electric and fracking part of the energy industry in a play to raise the relative costs of coal and foreign oil. Could this happen? It is unlikely but not inconceivable. As Yglesias says it would take “a smart and savvy” regulator and, I would add, a public-interested regulator (a small but let’s be charitable and say not a zero intersection) to bring the coalition together. You can see why I am less optimistic than Yglesias that the theory can be used to support the public interest but no one said that smart, savvy, public interested regulators would have it easy.

The Vietnam War was worse than most people think

That is the central message of the new and excellent Kill Anything That Moves: The Real American War in Vietnam, by Nick Turse.

For the entire course of the war, Turse considers an estimate of 2 million civilian Vietnamese dead and 5.3 million civilian wounded.  Of course by no means were all of those the result of U.S. military action but many were.  Here is a staggering estimate:

…between 1965 and 1968, thirty-two tons of bombs per hour were dropped on the North.

And yet, in fact, even more tonnage was dropped on the South, the ostensible ally!  The total quantity of explosives dropped is estimated to have been equal to 640 Hiroshima-sized atomic bombs.

There is a very good review of this book here.

What I’ve been listening to

These are some CDs which have stayed in my active listening pile for six months or maybe more:

1. The Roots of Drone.  Usually I hate collections, and listen to them only once, but on this virtually every track is good and the order is very well arranged.

2. St. Vincent, Strange Mercy.  I don’t like the more recent CD with David Byrne nearly as much.

3. Dabke: Sounds of the Syrian Houran.  Powerful stuff, music for a revolution or civil war.

4. Laura Marling, I Speak Because I Can.

5. Brian Eno, Lux.  As good as any of his older albums, believe it or not.

6. P.J. Harvey, Let England Shake.

7. Alela Diane, Wild Divine.

8. Continuous Beat, Rez Abbasi Trio.  Guitarist born in Karachi, this is probably my favorite jazz album over the last few years.

9. Earl Hines in New Orleans.  I’ve spent a lot of time looking for the best Hines CD and this seems to be it.

10. My Bloody Valentine, Mbv.  Get the LP version for the proper sound.  It’s amazing how good this comeback is, after a twenty year hiatus.

There is also James Blake, Taylor Swift’s “Stay, Stay, Stay” and classical music I will save for another day.  If you can find on iTunes Cecile McLorin Salvant’s “Jitterbug Waltz,” with the excellent Aaron Diehl, buy it, it is the most musical fun I have had all year.

*Ninety Percent of Everything*

The author is Rose George and the subtitle is Inside Shipping, The Invisible Industry that Puts Clothes on Your Back, Gas in Your Car, and Food on Your Plate.  Here is one excerpt:

The biggest container ship can carry fifteen thousand boxes.  It can hold 746 million bananas, one for every European..

And here is one other part:

EU-NAVFOR releases 80 percent of detained pirates because it can’t find willing courts.

Here is the book’s home page.

The future of higher education in Alabama?

The video that was posted online appeared to be a tour of the spa area at some swanky new hotel.

There were cascading waterfalls into hot and cold pools. There was an arcade section. A smoothie bar. Flat-screen TVs adorned every open space. There were lockers the members at Augusta National would find acceptable.

This was luxury, no doubt. But it was not at a hotel.

Instead, this shaky video tour was of the inside of a college football team’s training and lounge area. Specifically, it is the training, weight room and lounge area within the Mal Moore Athletic Complex on the campus of the University of Alabama.

Pricetag: $9 million. (And that’s just for the upgrades. The original facility, which opened in 2005, cost about $50 million.)

The host school, the University of Alabama, raised tuition seven percent last year.

Elon Musk’s hyperloop

I cannot speak to the technical issues (a very critical and very interesting take is here) but I wonder where the real gain is (there is a Musk pdf overview of the hyperloop here).  You can already fly LA to San Francisco in about an hour.  Is saving that time so important, noting that in the best case scenario the hyperloop still takes 35 minutes?

You might object that such a plane trip takes far more than an hour, all things considered.  But a hyperloop trip also will involve getting to the station, parking, waiting for departure, and perhaps TSA security as well.  What is supposed to be the net time gain, all things considered?  It would be cheaper to let up on the TSA procedures a bit, and even if we don’t do that surely there is no “regulatory arbitrage” case per se for building a hyperloop (“the TSA won’t budge, so I have this new and easy-to-implement idea…”).

Flying is carbon-negative, but of course building and running a hyperloop would be too.  In any case, it is hard to believe that a hyperloop is the marginally cost effective way to reduce carbon emissions, compared to say shutting down some more dirty coal or pricing traffic congestion.

How about putting working wireless, or maybe cable TV, into all LA to SF plane flights?  That would make time in the plane a lot more like time on the couch and in essence lower the time cost of flying.  We could even teach people to read books or let them keep their Kindles on during take-off, a rumored change which may even be in the works.

In other words, the whole Hyperloop thing seems to me like a publicity stunt.  I’m still waiting on that 24-7 Kindle access thing.

The binding constraint for a lot of transport improvements is getting the land rights (can it all be on top of I-5?  how much do pylons cost?) and overcoming local opposition.  I say the hyperloop proposal only makes that harder.

When it comes to transportation at least, There is a Great Stagnation.

p.s. under the counterfactual where the thing is built, what would the price be?  Here is some skepticism from Brad Plumer.