Footynomics: the economics of the AFL's future

In a recent book titled Soccernomics, the authors, a finance writer and an economist, stated that sporting leagues like the NFL and the AFL were likely be overwhelmed by soccer. “But Aussie rules can exist side by side with soccer. We said in the book that it may be a subsidised folklore festival so it is not my bet but I do think it is a distinct possibility," says one of the authors according to SBS's Matthew Hall. One must worry at the outset that people who love soccer enough to write a book about it might be slightly biased in their opinion, but be that as it may. Does such an idea make sense? Does the economics of it make sense?

Papau New Guinea celebrates winning the 2008 International Cup

Read more at USFootyNews.com

Do Crowds Learn Wisdom?

Over the last few years, academic economists have debated whether prediction markets, such as Intrade or the Iowa Electronic Exchange actually predict.  In particular, the question has been raised about whether prices from prediction markets can be interpreted as probabilities (see Manski (2006)).  The paper, "Do Crowds Learn Wisdom?  Theory and evidence from trading during the World Cup", suggests that while the price may not be always equal to the probability of the event it may converge to it as the new information is incorporated into the market.

Wall Street Journal on Decision Making in Sports

How Efficient Are Cricket, Tennis Players?

How efficient are sports participants, anyway? Among economists, the jury seems to still be out on whether football teams should go for it, rather than kick, on fourth down more often than they do. But football — highlighted in today’s Journal — isn’t the only sport they’re watching.

... more at WSJ Economics Blog.

Small Band of Economists Trumpet Sports Betting for Insights

By Justin Lahart

Word Count: 998

New Orleans

Academic economists got a clue that this year's meeting of the American Economic Association would be different when Louisiana State University's football team -- in town for today's championship game against Ohio State -- streamed through the lobby of the main conference hotel. As economists gathered Friday to discuss papers on business cycles, the evolution of checks and the factors that make for good schools, the LSU Tigers belted out "The Star-Spangled Banner" over breakfast nearby.

...more at Wall Street Journal.

My paper discussed in the article is Estimating the Value of Going For It (When No One Does).

Going For It on Fourth Down

Rationality and optimizing behavior are often a central assumption in economic models. Are economic agents rational? Can economic agents solve complicated optimization problems? A body of literature has grown up testing this assumption in the laboratory (Camerer (2003) for example). There has been less work testing this assumption in the field. One recent exception is Levitt (2006), which argues that even an MIT trained Ph.D. economist is unable to choose bagel prices optimally. In a similar vain, Romer (2006) argues that NFL coaches fail to optimally solve dynamic programming problems. Romer (2006) argues that NFL coaches fail to go for it enough on fourth down and in particular, these coaches leave expected points on the table. If true, Romer’s result suggests that economic actors may not be the rational optimizers often assumed in economic models. In particular, Romer argues the results suggest firms do not maximize profits. This paper takes a closer a look at the analysis and presents results that suggest NFL coaches behave in a way that is consistent with optimizing behavior and thus casts doubt on the conclusion that firms do not maximize profits. One concern with empirical analysis of going for it on fourth down is that we almost never observe teams going for it on fourth down. To overcome the lack of data, Romer (2006) assumes that success rates on third down can proxy for success rates on fourth down. This paper takes a closer look at this assumption and presents results from three alternative strategies for estimating fourth down success rates. Two of the strategies lead to estimated success rates consistent with Romer (2006), the third approach (the structural model) leads to estimates that are quite different from Romer (2006). Moreover, the structural estimates suggest that NFL coaches may act as if they can solve complicated dynamic programming problems.

The paper presents three alternative strategies for estimating success rates on fourth down. First, the paper uses actual fourth down attempts. The paper presents alternative approaches for identifying success rates at various to go distances and field positions from actual fourth downs. The paper uses data from early in the game and assumes there is little selection problem in regards to observed fourth down attempts. However, this data has few observations at the longer to go distances so functional form assumptions are needed for identification. The paper also uses all the fourth down attempts and a selection model (a “Heckit”) to account for selection into fourth down and therefore estimates “adjusted” success rates. Second, the paper estimates success rates at various field positions and to go distances using simulated data. Simulations are run on Madden NFL 07 using random assignments of teams into field positions and to go distances. Offensive play calls are also selected at random but the success rates are calculated using “optimal play calls”. Third, the paper estimates a game theoretic structural model of third down situations and fourth down attempts. Third down data on play calls and success rates and information on expected points from Romer (2006) is used to estimate parameters of the model. Success rates on fourth down from various field positions and to go distances are calculated assuming a mixed strategy equilibrium using parameter estimates and expected points.

The results from using actual fourth down data and the results from the simulated fourth down data are generally consistent with the results presented in Romer (2006). Using third downs as a proxy for fourth down, Romer (2006) determines the value of going for it versus kicking (kicking a field goal or punting) from various field positions and to go distances. Romer’s results suggest that NFL coaches should go for it much more often than they do, particularly in their own half of the field. Results from the structural model are not consistent with Romer’s findings. In particular, the success rates on fourth down from the structural model suggest teams should generally not go for it in their own half, a result consistent with observed behavior. It should be noted that this prediction is based upon the same model of decision making as Romer (2006) and the same expected points for various first down positions and kicks as Romer (2006). The different predictions are due to the different estimated success rates from the structural model. Moreover, the results contradict the statement in Romer (2006) that equilibrium outcomes on fourth down would not differ from equilibrium outcomes on third down.

Knowledge and the Wealth of Nations: A review

(Written for the Society of Government Economists Newsletter)

David Warsh's Knowledge and the Wealth of Nations is ostensibly about a relatively short research paper written by Stanford's Paul Romer and published in the Journal of Political Economy in 1990. Actually, the book is an amazing tour through the history of economic thought on the question of what determines the wealth of nations. The book takes the reader through the seminal contributions of Smith, Maltas, Ricardo, Mills, Walras, Marshall, von Neumann, Keynes, Robinson, Ramsey, Samuelson, Solow, Arrow, Debreu, Dixit, Stiglitz, Summers, Mankiw, Krugman, Romer, Romer and Romer (to name a few Romers). It shows how these people came to the question of what determines the wealth of nations and how their contribution increased or in some cases decreased our understanding. Warsh argues that Romer 1990 solved one of the great questions of theoretical economics: how to combine Adam Smith’s Pin Factory with Adam Smith’s Competitive Equilibrium. In the Pin Factory the manufacturer enjoys increasing returns through the virtuous cycle of more sales leading to greater productivity and lower costs leading to lower prices and more sales, i.e. Walmart. Increasing returns suggests large monopolies will dominate their markets. In a competitive equilibrium thousands of small firms compete on price to provide consumers with what they want at the lowest possible price, i.e. the Five and Dime. How is it that an economy can have both? Doesn’t one necessarily contradict the other? According to Warsh, Romer’s model allows increasing returns to lead to growth while still having a general equilibrium competitive framework.

I thoroughly enjoyed the book. Warsh provides an amazing insight into the history of economic thought of how an economy grows. For a non-economist, Warsh has an amazing understanding of economics and economists. I only came across one factual error. In a discussion of how Buzz Brock changed the Chicago Economic’s Department, Warsh states that Buzz is a tap dancer. Now, as every Wisconsin grad knows, Buzz is a clogger. But if you are willing to live with such variations on the truth and you are looking for an excellent overview of the economics of growth, then I recommend this book.

My biggest disappointment with the book is that I still don’t really understand the contribution of Romer 1990. I was exposed to Romer 1990 and its brethren in first year Macro. However, my understanding of the paper (to the extent I had any) was expunged as I walked out of the Macro prelim. By reading this book, I was hoping to get some insight into what I had been taught in Macro and possibly get some understanding of what it is that Macro economists do. Warsh provides all the pieces of the puzzle. He explains the inherent contradiction in Smith’s Wealth of Nations. He argues the importance of knowledge to understanding economic growth. He describes how Romer allows knowledge to be created endogenously in the market. He explains the need for market power to explain why investment in knowledge is profitable. What Warsh does not do, is put all the pieces together to explain how and why Romer 1990 works. I guess I will have to break open the seal on my Macro notes or, heaven forbid, actually read Romer 1990.

The most interesting thing about the book is the discussion of the policy implications. Warsh argues that traditional macroeconomic policies such as monetary policy, may be of secondary importance relative to education policy and NIH funding. To these policies I would add immigration policy, particularly in relationship to visas for scientists and students, and federal policy regarding stem cell research. I was in New York City recently and while I was having breakfast I witnessed two examples of how economies grow. The first example was a New York Times article about Singapore government’s strategy of developing the biotechnology industry. The article argued that Singapore’s less stringent laws regarding research on stem cells are encouraging scientists to move their. In this example, a country attempts to “pick winners” and encourage growth in a particular industry or sector. The second example was provided by the two people sitting next to me in the restaurant. One was an entrepreneur working with a firm developing a new technology to teach reading to children, while the other was an executive in a publishing house. The two were discussing the possibility of the publishing executive coming over to the new firm and using her experience to develop the product. In this example, the growth policy is to let New York City be a Very Large City. In New York City growth happens because people can get together over breakfast and create new firms and new technologies.

What leads to the wealth of nations? Thomas Friedman argues in the The World is Flat that India’s huge recent gains are the result of two things. One was India’s government’s long term investment in education. This policy has provided India (and the United States through immigration) with a large number of extremely well educated people. The second was the overinvestment in fiber optic cable by private firms during the Tech Bubble. The fiber optic cable brought India closer to the United States and Europe and the education policy provided know how to take advantage of the opportunity. Warsh argues that Romer 1990 provides a framework for thinking about growth and the importance of knowledge in creating the wealth of nations.

Learning in Prediction Markets

Recently a number of papers have attempted to pin down whether prediction markets like the Iowa Electronic Exchange can provide information on the beliefs of market participants and the true state of the world (Manski(2006), Wolfers and Zitzewitz (2004)). If market prices can be used to provide information about market participants' beliefs and the true state of the world, then market prices could be used for a number of purposes. Some have suggested that they could be used to estimate probabilities of election wins (Knight (2004), Wolfers and Zitzewitz (2005)). There was even a proposal to use such a market to estimate the probability of a terrorist attack. See Hahn (2006) for a recent overview of the literature. One open question is whether antitrust authorities could use this type of information in merger analysis. In particular, if a court case is going to significantly alter the competitive effect of a merger, could stock market prices be used to determine the likely outcome of the case and thus the expected effect of the merger? This paper presents a model of updating beliefs in an electronic market in which traders receive information about the true state of the world on which they are betting. The market clearing condition may provide information to the traders allowing them to update their beliefs and allowing the market to converge to a price that reveals the true state of the world. That is, with learning, the market price identifies the true state of the world.

This paper builds on the analysis in Manksi (2004) and Wolfers and Zitzewitz (2005) and shows when traders can learn or update their beliefs based on observing the market price, the market price will converge to the true state. In particular, the paper shows that when learning is added to the model used in Manksi (2004) the result that the market price does not identify the mean of the distribution of beliefs is overturned. The implication is that data from electronic markets may identify the true state of the world, although care needs to be taken before assuming the state is identified.

Manksi (2004) shows that in a simple model with risk-neutral price-taking traders the market price is equal to a particular quantile of the distribution of trader beliefs. The paper also shows that the mean of the distribution of trader beliefs is bounded. However, the mean of the distribution of beliefs is not identified in general. Wolfers and Zitzewitz (2005) shows that if traders have log-utility then the market price is equal to the mean of the distribution of beliefs. Wolfers and Zitzewitz (2005) and Ottaviani discuss other assumptions and evidence that suggest the market price will not be "too different" from the mean. In this paper, I show that if learning is added to the model used in Manski (2004) the market price will converge to the mean of the distribution of beliefs. While the paper does not require a particular form of learning it does require a degree of rationality such that traders do not hold beliefs about the true state that cannot be true.

The paper also assumes a fairly stylized version of a dynamic market. There are a countable number of periods and in each period the traders behave as if they are in the static model presented in Manksi (2004) and Wolfers and Zitzewitz (2005). That is, traders take their current beliefs as given and trade until the market is in equilibrium. Once the market is in equilibrium the period ends and the traders may update their beliefs given the observed market clearing price. In the next period, the traders take their updated beliefs as given and trade based on those beliefs.

Given the stylized dynamics and the assumption that traders do not hold beliefs which cannot be true given the observed market price, the market price will converge to the true state. Almost. Convergence does require that traders can use the observed prices to bound the true state. For example, traders can use the relationship between the equilibrium price and the mean (as stated in Manski (2004)) to update their beliefs about the true state of the world. The paper shows that the bounds calculated by Manksi (2004) allow the support of the distribution of beliefs to shrink. Eventually the support converges to a point which is equal to the true state (and the mean of the ex ante distribution of trader beliefs). The paper also considers less restrictive assumptions such as allowing traders to have more general preferences over risk and allowing for the mean of the beliefs to vary (in some sense) from the true state.

Full paper is here.

Links:
Chuck Manski
Justin Wolfers
Marco Ottaviani
Robert Hahn
Iowa Electronic Market

Curriculum Vitae

Christopher P. Adams, Ph.D.
Bureau of Economics
Federal Trade Commission
cadams {at} ftc {dot} gov


Previous Positions:
Assistant Professor (Visiting), University of Vermont 2000 to 2001
Teaching Assistant, University of Wisconsin 1995 to 1999

Academic:
Ph.D. (Economics), University of Wisconsin 2001
Masters of Science (Economics), University of Wisconsin 1996
Masters of Commerce (Economics), University of Melbourne 1993
Bachelor of Commerce (Honors), University of Melbourne 1991

Research Areas: Pharmaceutical R&D, Empirical Industrial Organization, Internet Auctions, Incentive Contracts

Published Papers (Refereed):
1. Spending on New Drug Development with Van Brantner, Health Economics, Forthcoming.
2. Estimating Demand from eBay Prices, International Journal of Industrial Organization, December 2007.
3. Estimating the Cost of New Drug Development: Is it really $802m? with Van Brantner, Health Affairs, March/April: 420-428, 2006.
4. Optimal Team Incentives with CES Production, Economics Letters, 92: 143-148, 2006.
5. Pharmaceutical Development Phases: A duration analysis, with Rosa Abrantes-Metz and Albert Metz, Journal of Pharmaceutical Finance, Economics and Policy, 14(4): 19-42, 2005.
6. The Use of Profit Sharing When Workers Make Decisions: Evidence from a Survey of Manufacturing Workers, in The Determinants of the Incidence and the Effects of Participatory Organizations, Takao Kato and Jeffery Pliskin (eds), chapter 7, volume 7, 2003, pp. 173-209, 2003.

Other Research Papers:
2009
Identification of the Joint Value Distribution from Auction Data.
2008
eBay aNtitrust, ABA Antitrust Section Economics Committee Newsletter, Spring.
2007
Introduction to Symposium on Online Auctions, with Patrick Bajari in International Journal of Industrial Organization December.
Tests of the Sealed-Bid Abstraction in Online Auctions, with Robert Zeithammer, under review at Marketing Science.
Roundtable on the Economics of Internet Auctions: A summary
2006
Estimating the Value of “Going For It” (When No One Does), under review at Journal of Sports Economics.
‘Vettes and Lemons on eBay, with Laura Hosken and Peter Newberry, under review at Quantitative Marketing and Economics.
Learning in Prediction Markets
New Dealer Laws and eBay Motors, with Laura Hosken and Debra Holt
Moral Hazard, Time Preferences and Stock Options
Empirical Facts and Innovation Market Analysis, with Rosa Abrantes-Metz and Albert Metz, Antitrust Source, March.
2004
Private Information and Ex Ante Contracts
Identifying Demand in EBay Auctions, FTC Working Paper #272
Is it Always Optimal to “Sell the Firm” to a Risk Neutral Agent? FTC Working Paper # 268
Digital Demand: Demand for new digital cameras on eBay, with Bill Vogt and Hao Xu
2003
New Drug Development: Estimating entry from human clinical trials, with Van Brantner, FTC Working Paper #262
Focusing on Demand: Using eBay data to analyze the demand for telescopes, with Laura Hosken, FTC Working Paper #257
Agent Discretion, Adverse Selection and the Risk-Incentive Trade Off, FTC Working Paper #255
2002
Does Size Really Matter? Empirical evidence on group incentives, FTC Working Paper #252
Selection of “High Performance Work Systems” in U.S. Manufacturing, FTC Working Paper #247

Recent Presentations:
2009
Econometric Society Summer Meetings, Boston MA, June
University of Wisconsin, Madison WI, April
International Industrial Organization Conference, Boston MA, April
2008
University of Minnesota, Minneapolis MN, October
American Course on Drug Development and Regulatory Sciences, UCSF, San Francisco CA, October
Chicago Graduate School of Business, Chicago IL, September
American Economic Association Meetings, New Orleans LA January
2007
Cornell University Policy Analysis and Management, Ithaca NY October
American Course on Drug Development and Regulatory Sciences, Washington DC September
Competition Bureau Canada, Ottawa ON Canada September
International Industrial Organization Conference, Savannah GA April
2006
University of Maryland Smith Business School, College Park MD September
2nd Annual Statistical Challenges in Ecommerce Research, Minneapolis MN June
University of Virginia, Charlottesville VA March
International Industrial Organization Conference, Boston MA April
2005
University of Maryland, College Park MD November
Federal Reserve Board of Governors, Washington DC November
Southern Economic Association Meetings, Washington DC November
INFORMS, San Francisco CA November
Department of Justice, Washington DC October
National Bureau of Economic Research Summer Institute, Boston MA July
International Industrial Organization Conference, Atlanta GA April
George Washington University, Washington DC April
Bureau of Economic Analysis, Washington DC May
2004
Charles River Associates Inc., Washington DC July
International Industrial Organization Conference, Chicago IL April
2003
Interdisciplinary Workshop on Reputation Systems, Boston MA May
North American Summer Meeting of the Econometric Society, Chicago IL June

Other Recent Experience
Rising Star Sessions Organizer, International Industrial Organization Conference 2008-2009
Organizer for the FTC/Northwestern Microeconomics Conference 2008 (with Dan O'Brien)
Session organizer, Economic Behavior: Evidence from Sports, American Economic Association Conference 2008
Local Co-Chair of 2008 International Industrial Organization Conference (with Ken Button)
Co-Editor of Symposium on Online Auctions in IJIO (with Patrick Bajari) December 2007
Organizer for the FTC Grocery Store Antitrust Conference (with Michael Salinger) 2007
Organizer for the FTC Economics of the Pharmaceutical Industry Roundtable 2006
Organizer for the FTC Economics of Internet Auctions Roundtable 2005
Session organizer, International Industrial Organization Conference 2004 and 2005

FTC case work: Realcomp II, Synopysis/Nassda, RJR/Brown and Williamson, Pfizer/Pharmacia, Meade/Celestron, Amgen/Immunex

Referee for American Economic Review, American Journal of Agricultural Economics, Health Affairs, Health Economics, Journal of Business, Journal of Econometrics, Journal of Economics and Management Strategy, Journal of Health Economics, International Journal of Industrial Organization, Journal of Political Economy, Management Science, Review of Industrial Organization.

Reviewer for the Sixth Edition of Michael Parkin’s, Microeconomics

Other
FTC’s Paul Rand Dixon Award 2007