For 125 years, practically since the beginning of the industrial age, General Electric Co. has been the pole star among US technology companies, second only to DuPont Co.” and, for a time, Bell Telephone.

Founded in Manhattan in 1883 by Tomas Edison himself, buttressed before long by the acquisition of rival Thomson-Houston Electric Co. of Lynn, Mass. (and subsequently led by its executives), GE, with the Bell System (which was organized in Boston in 1977), invented industrial research in the US.  The company then scaled up in the 1920s to become a major manufacturer of industrial equipment and consumer appliances. Edison’s one-time secretary, Samuel Insull, was pivotal in the creation of General Electric, then founded Chicago’s Consolidated Edison utility and invented the power grid before going broke in the Depression.

GE emerged from World War II as the pioneer in manufacturing the first jet engines.  In the 1950s it hired economist Martin Shubik to plan for it and actor Ronald Reagan to communicate its views. The company touted Boulwarism as an alternative to collective bargaining. After a price-fixing scandal in the early ’60s, the company turned to Reginald Jones in the ’70s and to Jack Welch in the ’80s and ’90s and returned to the top of the heap.   The news last week was that, after a “lost decade,” in which the prices of its shares fell well behind the market, GE is moving its headquarters to Boston’s new Seaport neighborhood. Forty years in suburban Fairfield Connecticut is enough.

It is easy to see why GE preferred Boston to the other possibilities (Atlanta, Dallas, Providence, Manhattan and New York’s Westchester County were among them). Despite Massachusetts’ loss of semiconductor manufacturing to the San Francisco Bay area, beginning in the 1950s, resulting in the rise of Stanford University as its most formidable competitor, greater Boston remains a major North American intellectual hub, home to a dozen universities, including Harvard and the Massachusetts Institute of Technology. Its medical, biotech, and financial industries remain strong; IBM is among the largest employers in the state, thanks to a string of quiet acquisitions over the last dozen years. The airport is a ten-minute drive from the neighborhood GE is said to have in mind.

Little noted is that Boston – or rather, Cambridge, across the Charles River – is the world capital of economic global trade theory.  Perhaps the city (and nearby Salem) have been ever since the days of the Yankee traders.

What’s changed since 1974, when GE left its Manhattan skyscraper for the distant  Connecticut suburbs? In a word, globalization. US companies had become multinational long before Japanese autos began to enter US markets in the1970s. With China’s entry into the global market system after 1978, the dis-integration of production processes expanded dramatically. To take a now-standard example, when Boeing Corp set out to build its 787 Dreamliner, its development team included fifty suppliers in ten countries and non-American suppliers accounted for nearly 70 percent of the airplane’s parts.

Economics has come a long way since 1979, when a 26-year-old MIT economist named Paul Krugman opened the door to a more sophisticated view of trade than the technology- and resource-based view that had dominated textbooks since the Napoleonic Wars. Already in 1961, Staffan Burenstam Linder had noticed that much international trade took place within particular industries (steel, say) rather that between sectors (steel for wheat). Krugman and the other young economists who quickly created the “new trade economics” showed how, using simple workable models: firms were differentiating their products, creating brands, and enjoying larger profits than they might otherwise.  The strategic consulting business, too, has come a long way from those days, when Bruce Henderson, of the Boston Consulting Group, expounded the gospel of market share.

Today, Krugman is an influential newspaper columnist and author, Nobel Prizewinner, too, but Elhanan Helpman, his co-author on the 1985 monograph that firmly established the new view, turned Harvard into a powerhouse of trade theory by hiring Pol Antràs and Marc Melitz.  Only Gene Grossman an Esteban Rossi-Hansberg, both of Princeton University, are figures of comparable magnitude.

I spent a few hours last week with Antràs’s new book, Global Production: Firms, Contracts, and Trade Structure, adapted from a series of lectures he gave at Pompeu Fabra University in Barcelona, in 2012.  The book isn’t easy reading; it shows how a first-year graduate student, or maybe a talented undergraduate learns to talk about globalization these days. Firms face two key decisions when setting their production strategies, Antràs explains: the first has to do with where to locate each stage of production, from research and development to assembly of the final product.  The second has to do with how much control to attempt to assert over each stage along the way:  within the firm, out-sourced to some other entity, and, if so, where.

Setting those boundaries would be hard enough in a world in which every agreement could be secured by easily enforced contracts, he continues.  But in a world in which a contract  is often only the first stage of negotiation, and many contracts are seriously incomplete, decisions of how and where to manufacture quickly become highly intricate. (One figure shows the duration of a legal procedure aimed at collecting a bounced check in some forty countries:  from a few days in the Netherlands and Singapore to more than 500 days in Colombia and Italy.) Antràs writes,

Practitioners (and perhaps some academics, too) might react skeptically to the idea that low-dimensional models may be able to capture the reasoning behind the complex and idiosyncratic decisions of firms in the world economy.  Business school cases often highlight the peculiarities of particular organizational decisions, making it hard to envisage that much can be gained from extrapolating from these particular cases.  The fact that comprehensive data sets on the integration decisions are not readily available might have only compounded this belief, as most empirical studies of integration decisions rely on data from specific industries or firms.

And yet, he says, with the customary presbyopia of the theorist, much can be learned from the theoretical and empirical investigation of the fundamental forces affecting decisions of where and how to organize affairs. (Antràs cedes much ground to the organizational economists; MIT, meanwhile, has lured John Van Reenen, of the London School of Economics, to Cambridge to pursue such matters.)

Practice?  That’s what business schools and consulting firms are for. Michael Porter, of Harvard Business School, and others of his tribe, have had far more influence on practice over the last three decades than all the theorists put together.  These days, there are activist investors,too, to administer an additional nudge. In the white paper that accompanied the news last autumn of its $2.5 billion investment in GE (making it the firm’s ninth-largest investor), Nelson Peltz’s Trian Fund argued that simplifying and de-layering the company could eventually raise its share prices to $45 from its current level of a little less than $30.  With most of GE Capital, its huge financial subsidiary, carved up and auctioned off to banks, and its appliance division sold last week to a Chinese manufacturer, much of that has been accomplished. The move to town from its suburban campus should help as well.

Everybody in Boston is about to get a little smarter, thanks to the positive spillovers that GE’s presence will produce,  a set of mechanisms that Garrett Jones, of George Mason University, describes to good effect in Hive Mind: How Your Nation’s IQ Matters So Much More than Your Own.

.                                                              xxx

Nothing discombobulates EP like a succession of four-day holiday weekends – unless it’s flying across the country.  I made two errors last week, one of them embarrassing. The week before, I spiked a busted column.

Describing Robert Gordon’s national accounting arithmetic in The Rise and Fall of American Growth, I chose a figure from the wrong column in a table in the book. Gordon’s estimate of potential GDP growth for the foreseeable future is 1.6 percent annually, compared with the CBO’s projection of 2.2 percent, not 0.5 percent as I wrote (and subsequently corrected on the Web). And Gordon’s estimate of median real income growth per person is even less that that – 0.4 percent annually for many years to come.  I’ll straighten this out properly in a future column.

Also, it was the very familiar John Campbell, Harvard professor and a founder of Arrowstreet Capital, an asset management firm for institutions looking after some $54 billion of client money, who gave the Ely lecture, not Jeremy Campbell, of the London Evening Standard, a journalist for whom I have a special admiration, or any other Jeremy.

The omission: The Big Short, an excellent but somewhat misleading film, was a little too complicated to write about on the road   EP plans to go to the movies on February 28, in time for the Academy Awards.

As for the meetings, I might have added that Roland Fryer, of Harvard University, received the John Bates Clark medal; that Paul Romer, of New York University, gave the Omicron Delta Epsilon John R. Commons Award Lecture (“The Trouble with Macroeconomics”), that Daniel Ellsberg was the dinner speaker at the annual meeting of Economists for Peace and Security, and that that Nobel laureate Alvin Roth, of Stanford University, is president-elect of the American Economic Association.

.                                                          xxx

Here I append, as I have for the past several years, an account of the meeting by James W. Fox, a veteran USAID economist, now retired, who is a shrewd an energetic observer of the meetings. He provides a more granular view than do I.

.                           What I Saw at the 2016 Economics Meetings

.                                                       by James W. Fox

.                                                     jameswfox@cox.net

Atmospherics

The meeting was the biggest ever, with 13,350 registered participants.  Plenty of Asians, particularly Asian women.  Women were evident everywhere except – as usual — at the head tables for the two plenary luncheons.  For the Nobel luncheon, there were two women among 17 men. At the AFA/AEA session, women were two of 20.  This continues the observation I have noted for these luncheons for the last decade.  Occasionally, there might be a third woman at the head table.

Weather was warm, though with light rain much of the time.  A far better situation than the last two, or the next, AEA meetings.  Unfortunately, the meetings are again scheduled next year for Chicago (where snow might mean that thousands of participants arrive a day late, having been bused in from airports in warmer climes.  (Maybe global warming might resolve this problem.)  The future lineup:

• Chicago
• Atlanta
• Philadelphia (again after three years ago)
• San Diego
• Chicago (again)
• Boston

This is not my idea of rationally thinking through the logistical issues relating to a meeting in early January.   Clogged airports, canceled flights, presenters missing because their flight didn’t go.

Angus Deaton, Nobelist and past president of the AEA, who has graciously served as my quality-control reviewer for a number of years before my summary reaches a wider audience, noted that there are few cities that can manage 12,000 people with a headquarters hotel with a nearby conference center, and that the AEA tries to hold hotel rates to a modest ($100/night) level because of the thousands of graduate students flocking to the meetings in search of a job. So it is a case where economic rationality prevails.  In January, hotel rates are high in good places and low in those with more-questionable (i.e., bad) weather.  The AEA used market forces this year, charging people who chose to stay in the headquarters hotel an extra $40 to subsidize costs in the other hotels.  I paid for the convenience.  Deaton also thinks Chicago gets a bad rap for weather, as New York and Boston are often worse.

Richard Thaler gave the presidential address.  His presentation, summarized below, put a stake in the heart of homo economicus.  The president-elect, Robert Schiller, also a behavioral economist, helped shape the sessions, with numerous sessions on behavioral issues.  His successor as president-elect, Alvin Roth of Stanford, will shape next year’s sessions for those who manage to get to O’Hare in mid-winter.

I would note that numerous sessions, including the Presidential Address and a dozen others are available to people from the AEA website.  I recommend the Presidential Address, as I will be unable to capture its sweep and humor in my summary.  Also, the session on Robert Gordon’s book and the Salvatore session are good listening.

Finally, I would note two big features of many of the individual sessions I attended: randomized controlled trials, and big data.  One presenter used NASA satellite images of light around the world to estimate per capita income; another used frequency of cell phone calling to predict rioting in the Ivory Coast.  Another theme ran through some sessions: “Are our best days behind us?” All are discussed below.

Plenary Sessions

1. The AEA Presidential Address

Richard Thaler gave this on the topic of “Behavioral Economics, Past, Present, and Future.”  This was a tour de force on how human behavior often departs dramatically from the economic model of rational maximization that is the basis for most economic analysis.  His numerous examples drive a stake into the heart of homo economicus.  His bottom line: “If everyone adopts an evidence-based approach to economics, then behavioral economics will cease to exist.”

Thaler asserts that good economics needs good psychology, instead of its tradition of inventing bad psychology.  He makes short work of various defenses of the rational maximizer model, including “if you raise the stakes, people will get it right,” which studies contradict.  Some examples of irrationality:

• Markets cater to consumer biases rather than correct them. Extended warranties generate $27 billion/year in revenues, even though economists and Consumer Reports agree that one should never buy them.

• Financial intermediation is 9% of GDP, despite the ease (and almost certainly higher returns for most investors from) index funds.

• A closed-end mutual fund investing in the Caribbean, not including Cuba, had the ticker name CUBA. Long trading significantly below net asset value (itself an irrationality), its name made the shares spike when President Obama announced diplomatic relations with Cuba, and only slowly retreated over the next few months to net asset value.

• Housing prices have historically averaged 20 times rents. In the 2000’s they ballooned far above that ratio.  A bubble, or what?

• Economists’ theory of labor markets says that workers earn their marginal product. But when workers shift from high-wage firms to low-wage ones, or vice versa, their compensation falls or rises according to the success of their firm.  Janitors make far more at Goldman-Sachs than at Wal-Mart.  Is that because they are more productive?

• Co-pays for health care are an economist’s remedy for excessive health care spending. Economists want the consumer to have some “skin in the game.” But experience shows that many people with high-risk problems (diabetes, heart disease) will skimp on medicines if they have to co-pay, with resulting higher health costs.

• There is much more. I recommend that people watch the webcast.

2. The Richard T. Ely Lecture

This was given by John Campbell of Harvard on “Restoring Rational Choice: The Challenge of Financial Markets.”  The bottom line from this presentation – and one that seems to have gained considerable currency among economists – is that the economics profession needs to promote paternalistic interventions in financial markets to promote the general welfare.

Economists should interfere with revealed preference?  How does he get there? He argues that the financial consequences of consumer decisions may be too important to be left to each individual:

• the world makes greater long-term financial demands on individuals than ever
• people have longer retirements, and many have lost defined benefits
• consumers are not up to the task, and
• their mistakes have social consequences

Campbell notes that rich people are more diversified, and take more risk.  Both lead to higher returns.  Others tend to make unambiguous mistakes:

• Failing to refinance when interest rates fall
• Not participating in retirement plans when the employer will match them
• Repeated use of high-interest credit card debt when unused home equity would be better
• Etc, etc.

He divides consumers into rationalists and behaviorists.  He admits that greater regulation may be sub-optimal for rationalists, but that cost should be far more than offset by the improvements in the financial choices of the behaviorists.

Campbell argues that financial education and disclosures by financial agents – the two easy ways to improve financial literacy – are too weak to do the job.  The literature on financial education as a solution is not encouraging, and financial agents can make disclosures too opaque for most consumers to understand.  As a result, consumers, particularly for saving for retirement, pay agents too much in hidden fees and commissions, and can find themselves in debt traps.  He recommends viewing on YouTube the ad attacking the Consumer Finance Protection Bureau (CFPB Denied) as evidence of the financial industry’s sophistication in tricking consumers, or as Nobelists Akerlof and Shiller put it in their new book, advertisers are “Phishing for Phools.”  The book is summarized in the Annex.

3. The AEA/AFA Luncheon

The speaker at the AEA/AFA luncheon was Bengt Holmstrom, with the topic of “Why are Money Markets Different?” Holmstrom sees money markets as the latest manifestation of the thousands-year-old tradition of the pawnbroker.  Money markets should provide risk-free liquidity, just like pawnbrokers.  Money is lent for less than the value of the collateral, so there is none of the risk or potential asymmetric information that characterizes risk-based investments.

The Repo market is the modern equivalent of the pawn shop, where the collateral offered is enough above the amount lent so that no price discovery is needed.  Such opacity is usually good for liquidity.

Nevertheless, financial crises do happen, and the collapse of Bear Stearns caused the Repo market, previously smooth, to get the shakes. The dark side of opacity is that it hides systemic risk.  Whatever increases liquidity in normal times increases risk during a crisis.

In such a crisis, government needs to step in to recapitalize banks, and do “whatever it takes” to calm the markets. It can then use stress tests, with corrective actions, improve capital/liquidity requirements, and generally return to a situation where lenders no longer worry about the collateral provided for Repos, and the opaqueness of the money market returns.

4. The Nobel Prize Luncheon.

This honored Jean Tirole.  As usual, this featured a parade of former students lauding his virtues.  Eric Maskin led the way, mentioning Tirole’s 180 journal articles, 13 books, and various other activities.  Maskin argued that Tirole transformed theoretical industrial organization, was a major force in corporate finance and banking theory and in behavioral economics, and even wrote a book on game theory.  Tirole is apparently all scholar, all the time.

5. The Econometric Society Presidential Address

This was given by Robert Parker on the topic of “Empirical Analysis and Auction Design.”

Parker is convincing that auctions of things like electronic spectrum are complicated.  The seller’s objective (whether profit maximization or some other public purpose), what exactly is being sold, who should be able to bid, and the auction rules need to be tailored to be optimal.  His modest conclusion is that “optimal design depends on the features of the economic environment, and empirical analysis can indicate which features are salient”.  In other words, hire an econometrician to design your auction.

Individual Sessions I Attended

6. Robert Gordon’s book, The Rise and Fall of American Growth: The U.S. Standard of Living since the Civil War.

This new book, just out, was reviewed by four economic historians, two of who called it magisterial.  The book certainly is massive, coming in at 655 pages, and is far more readable than last year’s book sensation, Picketty’s Capital in the 21th Century, which I could only slog my way through half. I am sure I will finish this book, and write up a summary for my friends.    The main focus is on the century from 1870 to 1970, when American people emerged from “an unremitting grind of painful manual labor, household drudgery, darkness, isolation, and early death,” into a world beyond their possible recognition.  This was particularly the case for the 75% of the population who lived on farms in 1870.  So far, it is a great read, and his three-page summary of the inadequacies of per-capita GNP to adequately measure the dramatic change in the standard of living of the American people is the best I have read.  The book notes the slowdown in productivity growth since 1970, and predicts an even slower-growth future.  This produced much of the discussion and doubt on the part of the reviewers.

Most reviewers challenged his dismal prediction, citing a variety of different sources of future growth.  Most cited the uncertainty about what innovations are likely to be in store, and with what impacts.

One notable complaint came from James Robinson, who is working with Acemoglu on a study of the impact of the U.S. Postal Service on life in the country over the early periods that Gordon studies.  Robinson noted that postmasters were 79% of all federal civilian employees in that era; for most Americans, the postmaster was the only federal employee they knew about.  The Post Office provided postage-free submissions to the Patent Office, which apparently led to more patents in areas with Post Offices.  Gordon does give some attention to the importance of the Post Office in linking isolated farmers to the urban world (e.g., postmen would take cash from farmers ordering from Sears or Montgomery Ward’s catalogues, buy a money order in town, and send it off with the order to the proper place.)

Anyway, a great session.

7. The Salvatore Session

The topic here was “The American Economy: Where Do We Go from Here?”  Dominick Salvatore provided the usual lineup of heavyweights:  Martin Feldstein, Joe Stiglitz, Olivier Blanchard, Stanley Fischer and John Taylor.  Everybody had some interesting insights.

Blanchard thought we might be back to the 1960’s, with the Phillips curve.  He argued that inflationary expectations are now anchored, so that the Fed can stimulate employment by easy money without worrying about falling into 1970’s stagflation.

Martin Feldstein worried, as usual, about the deficit, but offered an interesting approach that he thought would make it manageable.  Besides his usual recipe of raising the retirement age for Social Security and slowing benefit growth, he proposed a limit on “tax expenditures” to 2% of income.  He thought Republicans would like this because it reduced government spending, and Democrats would like it because it effectively raised tax rates on the highest-income groups.  He also proposed a gas tax, which, sadly, no candidate of either party has been willing to endorse—when gas is cheaper than ever, and gas guzzlers are flying out the doors of auto dealerships.  As it can be sold as repairing needed highway infrastructure rather than dealing with global warming, who would oppose it?  Apparently, everybody who wants to lead the country.

Stanley Fischer suggested that we may be in an era of a permanently lower long-run interest rates, but, in effect, said “nobody knows.” In response to Fischer’s noting of the zero lower bound for interest rates, Blanchard reported that some European Central Banks have gone negative, with interest rates as low as -0.75%.  He noted that this was easier in Sweden, where cash has almost disappeared, than in countries where citizens hold lots of currency.

Stiglitz thought the economy was unhealthy, with lending to SME’s down big time, government having shed 2.5 million jobs, and demand weak. He offered a detailed agenda:

• A climate-change tax
• Investment in infrastructure and technology
• End government austerity
• Fight economic inequality
• Reform the financial sector, stopping their rip-offs of the rest of us
• Industrial policy, and
• Reform of global systems

John Taylor was his usual self, arguing that the Great Recession was a policy problem, presumably due to not following the “Taylor Rule.”  Tax reform, regulatory and entitlement reform, and trade agreements would fix everything.

8. Development Economics

There were several interesting papers here.

From Proof of Concept to Scalable Policies.  

This was co-authored by Duflo and Banerjee, but presented by James Berry.  This was a defense by the founders of Randomized Controlled Trials (RCT) of their approach against doubts about its generalizability and scalability.  The paper addressed four such experiments in India, which attempted to deal with the reality that the syllabus for Indian grade schools was rigid and challenging, but provided automatic promotion.  The result was that half of 5^th grade students couldn’t read at the 2^nd grade level.  The first project involved an Indian NGO that demonstrated outside of schools that teaching at the right level with competent teachers could produce learning.  A second project sought to implement this concept in two Indian states, which largely failed for lack of buy-in by the teaching profession to the approach.  Two more experiments followed, the most successful coming from using the NGO to use a mobile group to have four 10-day bursts of work with students on the teaching at the right level concept, which seems to have produced some results.

Nevertheless, the six years of effort over four projects seems to me not to have produced a scalable product.

 National Accounts vs. Household Surveys.

This paper notes anomalous differences between measurements from household surveys and national accounts data.  For India, over some period, the national accounts data showed a 100 percent increase in GDP per capita, while household surveys found only a 20% increase.  It was worse in Angola, where a doubling of GDP per capita produced a -5% result from household surveys.  Which to believe?

The methodology was to use nighttime lighting as a proxy for GDP.  NASA provides such data around the world, with daily measurements computable into annual averages.  By looking at variations over time in the amount of light in the evenings in different countries, they seek to mediate between the two measurements.  For the three countries shown, Angola, India, and Zimbabwe, the results seem to better approximate the GDP trends.  Lights went out in Zimbabwe, and came on in various parts of India and Angola (not just the capital).  They conclude that a proper measure of income increase would be weighted 85% from national accounts, and 15% from household surveys.  They argue that household survey forms are complicated, and that non-responses are a large and growing problem.  They hypothesize that people with rising incomes are less willing to allocate time for such an effort than poor people.

Social Identity and Aspirations

This paper interviewed 1,000 adolescents in 19 villages in Rajasthan, India for evidence of conformation to stereotypes regarding gender and caste, particularly with regard to aspirations. The study concluded that caste identity by the interviewer was, in general, positive for high castes, but not negative for lower castes.  For girls, however, caste identity was generally negative, especially for higher castes.  (As I know little about such matters, I cannot interpret this result.)

9. Measuring Social Indicators with Mobile Phone Data

There were two interesting papers here.

Predicting Wealth and Poverty from Passively-Collected Mobile Phone Data

This paper used supervised machine-learning, along with a survey of 1,000 mobile phone subscribers to estimate wealth and poverty in Rwanda.  It seems to work, and much more can be done.

Violence and Cell-Phone Communication Patterns:  Evidence from Cote d’Ivoire

This paper uses mobile phone data to predict violence in Cote d’Ivoire during a particularly violent period, when uprisings and rioting characterized some areas at particular times.  Notably, cellphone chatter increased substantially in particular nodes when groups were planning an uprising.  There was much more, but the finding looks like an invitation to Big Brother government to monitor mobile phone communications in real time.

10. Vocational Education

I stayed for two papers, both of which reinforced my suspicion of donor projects for vocational education.

One paper dealt with internships for university graduates in Yemen, who suffered from 40% unemployment.  There were many STEM graduates, but few STEM jobs.  So the World Bank thought internships for graduates might reduce unemployment for them.  (The Bank may not have noticed that university graduates there were probably children of the elites, so could weather 40% unemployment more easily than the average Joe.)  In any event, the country descended into civil war after placements after the first interns had been put in place, and a planned second year was abandoned.  Needless to say, there was little to show for the effort.

Another project in Kenya provided up to $460 to 2,700 carefully-chosen new labor market entrants to obtain further education or technical specialization.  (The country has lots of training institutions, and $460 would typically pay for two years of training.)  Sadly, the effort seems to have produced no increase in earnings or workhours.

11. Other Stuff

The John Bates Clark Medal went to Ronald Fryer of Harvard, for his work on education and race.

Distinguished Fellows awards went to:

Hal Varian of Google, who may have reduced fears of a Big Brother government by providing a private-sector alternative.

Theodore Bergstrom of University of California

Gary Chamberlain of Harvard, and

Thomas Rothenberg of University of California

Prizes for the best papers in each of its four subsidiary publications, made necessary by the explosion of economic research and, not coincidentally, by the need of academic economists to publish or perish, went to:

The AEJ Applied Economics prize was awarded to Enrico Moretti, for the paper “Real Wage Inequality,” January 2013, pp. 65-103.

The AEJ Economic Policy prize went to Alan Auerbach and Yurity Gordodnichenko for “Measuring the Output Responses to Fiscal Policy,” May, 2012, pp. 1-27.

The AEJ Macroeconomics prize went to François Gourio for “Credit Risk and Disaster Risk,” July 2013.

The AEJ Microeconomics prize went to Kenneth Hendricks, Alan Sorenson, and Thomas Wiseman for “Observational Goods and Demand for Search Goods,” February 2012.