Comments on economics, mystery fiction, drama, and art.

Thursday, December 28, 2006

Losing to win?

Several weeks ago I got involved in an extended discussion on a sports economics listserv about whether teams ever intentionally plan to lose a game in order to advance their chances of winning a title. Now comes an article in the Chicago Tribune in which the writer, David Haugh, argues that, if the Chicago Bears lose to the Green Bay Packers this weekend, the Bears' chances of winning the Super Bowl will be enhanced:

If the Packers win, it could reduce significantly the number of plausible playoff possibilities for the Bears, especially if coupled with a road loss Saturday night by the dissension-torn Giants at Washington. If Green Bay still has a chance to clinch a postseason berth at kickoff—a good chance—the Bears knocking the Packers out could mean letting the Panthers, Falcons, Rams or Giants in.

Every one of those teams poses a potentially bigger threat to the Bears in January than Green Bay would because of big-play, quick-strike skill players who can turn a playoff game around on a single snap. Those other teams rely on the types of explosive players the Bears have struggled to stop in recent games against the Bucs, Lions and Rams.

I'm guessing that if a newspaper columnist can figure this out, so can the Bears' coaching staff. And they don't even have to do anything odd--just rest most of their most important players fr most of the game against the Packers. After all, the Bears have home-field advantage throughout the playoffs, and by this point in the season lots of players have little aches and pains that some down time could help with. And no one would think it odd--a couple of years ago, the Eagles essentially decided not to go for an undefeated season in order to rest their regulars for the playoffs.

But, in this case, the larger agenda is to increase their chances of winning the Super Bowl by changing who participates in the playoffs. The issue here is a set of "conditions of contest" that can make playing to lose--not simply not caring whether you win or lose, but preferring to lose--the optimal choice for a team. I think we'd all agree that there's something odd about this, but not omething odd about a team taking advantage of it.

Wednesday, December 20, 2006

Gift-Giving: An Economic Analysis?

It's Christmas-time once again, and once again, economists weigh in on the rationailty of, and the utility derived from gift-giving. The usual conclusion is that, leaving aside any benefits derived from knowing that someone cares enough about you to give you a gift, you will probably derive more satisfaction (utility) if people simply give you cash.

Now that exception might be major. And as Daniel Kahneman demonstrated over and over again, people aren't necessarily rational decision-makers even when they are choosing for themselves--they change their minds frequently and rapidly, they choose both sides of the same offer when it's framed differently, they express regret, and so on.

But almost no economists go the next step. If I get more utility from my choices, and if you get more utility from your choices, why even give each other money? Why not just keep it and spend it on ourselves? Cutting out the middleman, so to speak?

This is a case, I think, in which economics tries to do too much. I think gift-giving continues because it is a powerful symbol, and that both the giver and the recipient see it as such. Giving is a sign of affection (and sometimes power--"See how much I can give you. See how much I can withhold"); accepting what someone gives you is also a sign of affection. Knowing that someone cares enough to spend time thinking about what you might want does have an effect, as well.

Maybe it pays to restrict your giving to people you care deeply about. Maybe it makes sense to alter your gift-giving (and receiving) habits.

But pay no attention to those economists who urge you not to give. They haven't really thought carefully enough about it.

What is the largest cash crop for US agriculture?

The answer may or may not surprise you. According to the LA Times, it's


(Thanks to Marginal Revolution for the tip).

Friday, December 15, 2006

As the grading season comes to an end

An annotated and illustrated update of the age-old grading system, courtesy of Daniel Solove at Concurring Opinions: the staircase method. Check it out. (A tip of the hat to Michael Froomkin.)

Calculating the value of leasing the Hoosier Lottery

I decided to do a calculation, to make a judgment about whether leasing the lottery would "pay off." Here are my assumptions: The lease runs 20 years. The current state general fund share of the lottery (about $190 million) will grow at a 5% nominal rate. The gross revenues of the lottery will also grow at 5% per year. The sale of the lease will generate an up-front payment of $1 billion, plus an annual payment of $200 million, plus 5% of the excess of gross revenue over $700 million (current gross revenue is about $725 million). I use a 5% nominal discount rate.

The current revenue flow ($190 million growing at 5%, discounted at 5%) for 20 years is simply $190 million times 20, or $3.8 billion.

The discounted revenue flow from leasing the lottery (annuitizing the $1 billion up-front payment over 20 years--about $76 million pre year) is about $3.7 billion.

So it's close to break-even.

For the state to come out ahead, the lessor will have to generate faster revenue growth than 5% per year. At 10% per year revenue growth, the PV of the general fund revenue would be $4.1 billion. But 10% revenue growth would almost certainly lead the state's share to rise in any event (let's say by 10% per year as well). Oops--the PV of that revenue flow would be $5.8 billion--not so close to break-even.

The lessor would have to generate faster growth in revenue that the state could on its own for this to be a good deal.

I'm still not sure what to make of it.

Thursday, December 14, 2006

The Indiana Lottery

Well, this makes more sense. I'm still not sure how I feel about it, and the finances for a bidder get a bit murky.

The proposal is to lease the operations of the lottery for an up-front payment (which the state apparently expects to be at least $1 billion), plus $200 million a year, plus a percentage of the gross (the percentage mentioned is 5% of the gross above $700 million per year).

The up-front payment goes (60%/40%) into two scholarship endowments aimed at encouraging students to remain in Indiana after then grduate.

As with the toll road lease, I'm not sure I see how an outside operator could make much of a profit, if any, on this deal. Right now, the distribution of the revenue from the lottery is as follows:

61% Prizes
26% State of Indiana General Fund
10% Retained by ticket sellers
2% Administrative expenses
1% Advertising and promotion

Based on the current payments to the general fund (about $190 million), total ticket sales are between $725 million and $750 million per year. At $750 million, and assuming $200 million to the state, $450 million to prizes, and $75 million to ticket agents, we have $25 million left over for administrative expenses and advertising. The proposal calls for the $200 million annual payment to continue. Reducing the payout rate is possible, but would make the Hoosier Lottery less competitive compared to other states, and is therefore not real likely.

So we're left with about $100 million to play with. The ticket agent 10% rake-off might be cut, let's say even in half. That's $37.5 million more for the lottery operator. Assuming that the $25 million does in fact cover operting costs and promotion, that's a 5% rate of return on revenues, which isn't bad.

Except--if you pay $1 billion for the opportunity to do this, you're giving up $50 million or so in interest income. Forever. Suddently, it's not so good a deal.

How do you make a profit? You could boost promotion, in order to sell more tickets. You could create new lottery possibilities. I don't think you can cut out the ticket agents, because you do need a retail network in order to sell tickets. If you could boost lottery sales to $1 billion, maintaining the 60% payout, the $200 million (plus 5% of revenues over $700 million--about $15 million, in this scenario), cut the agent's share to 5%, and hold administration and promotion to $50 million, what would you have? Expenses of $915 million and revenue of $1 billion...might be worth it.

Wednesday, December 13, 2006

First, the Indiana Toll Road. Now, the Indiana Lottery

From an article in today's Indianapolis Star:

12:26 PM December 13, 2006
Lottery lease might bring in $1 billion

A top House Republican says leasing out the Hoosier Lottery could generate a one-time payment of around $1 billion...That money would go toward funding higher education initiatives, such as scholarships and professorships.

In addition to that payment, Indiana would keep some of the lottery's annual revenues, which presently help fund pensions for teachers, firefighters and police and provide relief from auto excise taxes. Lottery distributions to the state totaled $189 million in fiscal 2005. The arrangement would allow the state to collect the one-time windfall, while continuing payment for pensions and the auto excise tax, Espich said.
Copyright 2006 All rights reserved

Let's assume we annuitize this, with a 5% rate of return on the remaining principal. Then to generate the current $189 million annual lottery revenues, we'd run out of money in, um, 6 - 7 years.

And then? How would we plan to replace that $189 million in the state's budget? OK, there's supposed to be some ongoing revenue flow, but how much?

I could make sense of the lease of the toll road, but, right now, this makes no sense whatsoever. (Of course, replacing the lottery with a less-regressive way of raising revenue for the state would be an even better that'll ever happen.)

UPDATE: It's worth noting that proposals to "earmark" one or another revenye stream for education (or almost any other specific expenditure type) rarely generate much new spending. Almost always, the legislature sooner or later (usually sooner) concludes that we've provided a dedicated revenue stream for education (or whatever), so we can cut back on our other appropriations for education...

Friday, December 08, 2006

David Card on Skill-Based Technical Changes and Wages

David Card is one of the best economists around at looking closely and carefully at the data we have available to us; his specialty is what's happening in labor markets. Over the last 10 - 15 years, a primary hypothesis for explaining growing inequality in wage and salary income has been the Skill-Based Technical Change hypothesis, which argues that much of the growing inequality (e.g., the rising gap between wages of college-educated workers and others) is a growing skill gap. In an interview by the Minneapolis Federal Reserve Bank and published in The Region, Card makes the following observations about the SBTC hypothesis:

John DiNardo (of the University of Michigan) and I were troubled by the fact that there are a lot of patterns and trends in the labor market that don't fit in very well with a skill-biased technical change explanation. We were motivated to embark on a Don Quixote mission, a noble cause that wasn't going to go anywhere [laughs].

One thing we pointed out, for example, is that women are lower skilled than men, if you take the fact that they have lower wages as evidence of their skill. The SBTC theory says that people with lower skills should have slower wage growth than people with higher skills. But over the 1980s, women did much better than men. It's also the case that over the 1990s, women's relative wages were fairly stable again. So there was a long period of stability of women's relative wages, then a period of convergence of women relative to men that ended in 1991-92, and then stability again. That's an important set of trends that SBTC doesn't address. SBTC might be consistent with it; it might not be, but the theory needs a lot of auxiliary hypotheses to work.

The same thing is true with respect to the black/white wage gaps. Blacks earn less than whites, and many people believe that the reason they do so is because they're less skilled. Nevertheless, during the 1980s, the black/white wage differential was stable. It didn't widen as people had predicted it might.

Another trend that didn't fit with the SBTC hypothesis concerns the relative wages of people with different bachelor's degrees. There are a couple of different data sets that collect starting salaries for newly minted B.A.s. What these data show is quite remarkable. Everyone knows that the average wage of young college graduates went up over the 1980s. It wasn't the case, however, that the gains were most pronounced in engineering or science. They were actually greater for graduates in the humanities, which doesn't seem consistent with the idea that there is increasing demand for technically proficient, computer-savvy people.

...A final puzzle concerned the age structure of the increases in the relative wages of college versus high school graduates. Wages of young college-educated workers rose relative to young high school workers, but for people over age 40 or so, there really wasn't any change in the high school/college premium.

The moral for economists is that theory is a useful thing, but the real issue is whether the theory is in agreement with reality. If it isn't, then it's the theory that needs to be revised.