The Trouble With Picking Winners

Three problems that college football and American economic policy have in common

---

As the college football season gets in full swing, controversies are already swirling as to who is the best team in the country. While such disagreements tend to inhere in every sports league from youth soccer to pro baseball, NCAA football has a uniquely acute case. Along with changing leaves and Halloween festivities, it seems that college football rankings scandals are one of the perennial features of autumn's dawn as another team claims a denial of due process.

So why is the system so uniquely inefficient?

As any college football fan will tell you, the first reason is that college football lacks a true playoff system. Unlike professional football and every other major college sport, there is no tournament system for establishing a champion at the end of the season. Instead, college football's top teams are determined by an awkward and arbitrary human ranking system, where coaches and professional football critics rank each team according to their esteemed opinion.

This problem is generally exacerbated by the second major problem, which is the advent of a computer ranking system called the Bowl Championship Series (BCS). The BCS was invented to help alleviate Problem 1. Its strategy is to create a complicated mathematical formula -combining human polls along with various "objective" computer-generated evaluations - to determine the best teams in the country.

The problem is it doesn't work. Instead of quelling disagreement, the BCS has amplified it. By this point, every college football fan hates the system. My own team, the University of Miami Hurricanes, were denied a spot in the 2000 national championship game because a team we beat, Florida State, was ranked higher in the BCS. The problem has become so bad that even President Obama said that he wanted to change the system.

Now the President's interest in correcting these unique inefficiencies is noble and greatly appreciated by this frustrated college football fan. But I would rather he address the unique inefficiencies that plague our government policy, because the rules of the economic playing field need to be changed as well.

In my view, the there are 3 major problems that college football and American economic policy have in common:

1) Emphasis on theoretical analysis instead of raw competition

The problem with college football rankings (especially the BCS) is that they inevitably must rank teams by factors other than actual competition. As there are over 100 major college football programs, and each team generally only plays 10-11 games in a season, the generation of the Top 25 teams inevitably means comparing teams that have never met on the field.

Not surprisingly, this comparative system ends up with bizarre results. According to current polls, for example, Penn State is ranked higher than the Iowa team that just beat them, and Oregon is ranked lower than the California team that they shellacked over the past weekend.

Government policy - especially the stimulus package - ranks projects with a similarly flawed methodology. Because one can't compare a specific project with every other possible use of that money, one must rely on arbitrary comparisons. It's one thing to say that a given road construction project would improve more people's lives than another road construction project. But how does one compare a road construction project to a private office building complex, for example? You can't, unless you let people choose for themselves in every possible situation. (This would never happen, of course, since would violate the stimulus package's fundamental premise that government can spend the money better than the people can.)

In a perfect college football system, rankings would be determined by competition on the field of play. Likewise, in a perfect economy, project funding would be determined by competition in the open marketplace.

The reason is the same in both cases: competition is the wellspring of value. Let it play out.

2) Valuing effort over skill

In the Myth of the Rational Voter, economist Brian Caplan coined what he called the "make-work" bias, which he defined as "tendency to underestimate the economic benefits of conserving labor." One could also redefine it as the tendency to overestimate the benefits of working hard. Caplan sees the make-work bias skewing voter preference toward inefficient and misguided government programs.

While the make-work bias is very pervasive in the human rankings system, it is also at the core of the BCS rankings system.

For example, the BCS ranks teams higher according to their margin of victory. While that's undoubtedly important in demonstrating a team's competitive capacity, it also can be misleading just by itself. As a single statistic, it fails to take into account any injuries or personnel changes a team may have incurred; it also fails to take into account intangibles like home field advantage or a team's momentum by virtue of the previous week's games. Teams are rewarded for running up the score rather than, say, overcoming adversity by coming back from a large deficit.

If the BCS has fallen for work-effort bias, the recent stimulus package suffers from a terminal case of it. The package was designed to minimize one number - the unemployment rate - without regard for the complexities behind it. The policy was wholly designed to "create jobs", whether or not the jobs would last, were cost effective, or were even needed in today's economy. But while one may be able to create a job by paying someone to hammer away at concrete, there's no saying that person couldn't be doing something better somewhere else. We don't need more jobs for every worker. We need more quality jobs and more skilled workers. There's a big difference.

The make-work bias is a dangerous fallacy. Rewarding the University of Florida for beating up on the equivalent of junior college football squads is no more defensible than rewarding the construction of a bridge to nowhere. Unfortunately, we're currently doing both.

3) Heeding past calculations over present reality

The BCS is rightly derided for its excessively quantitative rankings of Byzantine complexity. While there is no doubting the comprehensiveness of those calculations, the fact is that as noted in the aforementioned examples, the BCS spits out a result that is rife with absurdity.

The stimulus package suffered from a similar fate, thanks to its like-minded allegiance to quantitative factors. In order to find the most promising stimulus projects, the government consulted number-crunching economists, who evaluated a project according to its "Keynesian multiplier".

The "multiplier" is an amorphous concept referring to a project's ability to stimulate additional economic activity. This figure is generally calculated by reference to past performance data, an reasonable approach that satisfies professors but actually enlightens few. Transformative growth comes from spontaneous entrepreneurship, unpredictable trends and black swan innovations. Just because a specific industry created a lot of jobs in the past doesn't mean it will continue to do so in the future. Fast-growing industries are where job growth happens, but they are also the most difficult to predict.

Much like a BCS analysis, the value of Keynesian analysis can be summed up by removing the middle letter in the acronym. Also just like the BCS, it succumbs to too much egg-head calculation, and not enough break-neck competition.