Never make predictions, especially about the future.” At the moment I’m typing this, there are 47,600 Google hits attributing that quote to Baseball Hall of Famer Casey Stengel and 56,300 attributing it to Baseball Hall of Famer Yogi Berra. I have no idea who actually said it, which suggests that one should be cautious even in making predictions about the past.
The future is even harder. Around the year AD 100, the Roman senator Sextus Julius Frontinus pronounced that “inventions have long since reached their limit, and I see no hope for future development.” In 1943, Thomas J. Watson, a pioneer in electronic computing and the president of IBM, predicted that “there is a world market for maybe five computers.” In 1955 the entrepreneur and inventor Alex Lewyt predicted that we were no more than ten years away from having a nuclear-powered vacuum cleaner.
As there are failed predictions, so there are failed anti-predictions. My treasured copy of the 1985 humor classic Science Made Stupid (a spot-on parody of the popular-science genre) contains a “Wonderful Future Invention Checklist,” offered tongue-in-cheek, because of course the joke was that there was little chance we’d see any of this in our lifetimes. Just twenty-five years later I was able to check off about a third of the entries: household robot (does my Roomba count?); magnetic train (check!); flat-screen TV (check!); flat-screen 3-D TV (check!); two-way wrist radio (we are so far past this!); two-way wrist TV (ditto); instant access to all human knowledge (check!!!—and as a bonus we also have instant access to all human stupidity); human clones (getting there); first black president (check!); spelling reform (OMG! I cn chk this 1 off 2!).
Of course I’m cherry-picking; history affords many examples of spectacularly accurate predictions as well. (Think of Dmitry Mendeleyev, in 1863, using his Periodic Table to correctly predict the properties of more than forty chemical elements that hadn’t yet been discovered.) In any event, predict we must or we simply can’t go on. I am writing these words, for example, based on the (cautious) prediction that somebody might want to read them.
The puzzles in this chapter call for predictions. Not all of them call for perfect predictions; in some cases the best you can do is argue that one outcome is more likely than another, or that the answer depends on some additional fact beyond what’s given. In some cases, the answer might even be that prediction is impossible. But each problem does call for some genuine insight beyond what you might think is obvious.
Start here.
An experimenter places two pigs in a box—one very large pig and one very small pig. At one end of the box is a lever. At the far end of the box—several pig’s lengths away—is a bowl that fills with food whenever the lever is pressed. Which pig do you predict will eat better?
SOLUTION: When real-life pigs are placed in a real-life box under these circumstances, the big pig does most of the work and the small pig does most of the eating.
That’s because the little pig has absolutely no incentive to press the lever, having quickly learned that if he does, the big pig will steal all the food.
So the big pig, in order to get fed at all, has to do all the lever pushing. While he’s getting ready to push the lever, the small pig is waiting by the food bowl. The big pig pushes the lever and runs to the bowl. By the time he gets there, the small pig has scarfed down most of the food. But the big pig shoves the small pig aside and gets whatever’s left, which is enough to teach him that it’s worth pushing the lever.
The immediate moral is that sometimes it pays to be small. The bigger moral is that even in very simple situations, outcomes are not always what you’d first guess—and if you’re making predictions, a little bit of logic goes a long way.
You’ve just opened an adoption agency in a time and place where it’s well established that most parents prefer sons to daughters. Do you expect to get more requests for boys or for girls?
SOLUTION: I can’t be sure, but my guess is that you’ll get more requests for girls.
That’s because prospective parents, even when they have a strong preference for boys, tend to have a lot of other strong preferences—including a strong preference for healthy, well-behaved children.
Who gets put up for adoption? Disproportionately, kids who are less wanted. And who are those kids? They’re kids with health problems, kids with behavior problems, kids whose parents can’t raise them, and (given what the problem tells us about parental preferences in this time and place) girls.
So if I were looking to adopt a healthy, well-behaved child, I might very well choose a girl even if, like most parents, I had a strong preference for a boy.
As long as parents prefer boys, they’ll tend to cut their sons more slack than they cut their daughters. Therefore (on average, of course), boys have to be really bad to get thrown into the adoption pool, whereas girls are sometimes thrown in just for being girls. If you’re fishing in that pool, it can make good sense to steer clear of the boys, no matter how much you prefer boys in general.
So as long as prospective parents think this through—and when you’re adopting a child, you tend to spend a lot of time thinking things through—your agency might quite likely get a lot more requests for girls than for boys.
You might think this all comes down to which is stronger—the preference for boys or the preference for healthy, well-behaved kids. But that cuts both directions. If you’re an adoptive parent with a very strong preference for boys, you might choose a boy despite the risk. But if you’re an adoptive parent in a time and place where others have a very strong preference for boys, then the risk of choosing a boy is elevated even further, giving you all the more reason to opt for a girl.
Way back in 2003 I wrote a couple of columns for Slate magazine on the subject of whether parents in general prefer boys or girls. I pointed to several bits of relevant evidence: All over the world, parents of boys are less likely to divorce than parents of girls. Again all over the world, parents of girls are more likely to try for another child than parents of boys. Unmarried couples, upon learning the sex of their unborn child, are more likely to marry if the child is a boy. None of these things proves conclusively that parents prefer boys, but I argued that, taken altogether, the evidence seems to point that way. (I’m less convinced of that now than I was in 2003, but that’s not relevant here.)
A great many readers chimed in to say I must be wrong, because adoption agencies get more requests for girls—and this (according to those readers) proves that people on average prefer girls! The preceding problem is dedicated to those readers.
Incidentally, Amazon offers for sale a device that’s advertised to predict the sex of a child in the womb only six weeks after conception. Here’s the distribution of customer reviews for that product:
Apparently this product predicts very accurately just about half the time!
Studies show quite conclusively that, on average, children from four-child families do significantly worse in school than children from three-child families. You already have three children and have just discovered that you’re pregnant with a fourth. Do you predict that your first three children’s school performance is likely to suffer?
SOLUTION: You shouldn’t. It’s true that family size is strongly correlated with educational achievement: once you get past two children, the larger the family, the worse the kids perform. But there’s ample evidence that this correlation is not causal. Parents who choose to have four children are generally less educated than parents who choose to have three, and this (along with other demographic differences), not the family size itself, is why their children do less well in school.
(The effect persists as family sizes get larger; children from five-child families do worse than children from four-child families, and so on.)
But of course your own decision to have a fourth child doesn’t change you into a different sort of person. Even if most four-child parents have less education than you do, having a fourth child won’t cause your alma mater to revoke your MBA. So your first three children will probably do just as well (or just as poorly) with a fourth sibling as without.
How do we know that it’s the demographic characteristics of large families, and not the family size itself, that drives differences in school performance? The simplest test would be a statistical study that controls for demographic characteristics. But that’s an imperfect strategy if you’re not sure which characteristics are important to control for. So researchers have aimed to be a little more clever. For example, they’ve looked at families with four children where the last two are twins. Many of these are families who chose to have three children but ended up with four. It turns out that their children tend to perform a lot like children from three-child families, which suggests that it pays not to come from a three-child family but to come from the sort of family where the parents aim for three children.
Another clever strategy is to look at four-child families with, say, three boys followed by a girl (or the same pattern with the sexes reversed). A lot of these families probably planned to stop at three but then took one final stab at gender diversity. Once again, their kids perform a lot like children from three-child families.
Incidentally, even after we make all the corrections for demographics, children from four-child families still do slightly worse on average than children from three-child families (and children from larger families do even worse), but this can be explained away as a birth order effect: fourth children do worse in school than third children, and therefore bring down the family average without bringing down their older brothers and sisters. Fifth, sixth, and seventh children do even worse. Of course it should go without saying that there’s plenty of variation around these averages. Wolfgang Mozart was the youngest of seven children, and Benjamin Franklin was the youngest of fifteen.
If a new method of birth control is safer, cheaper, more effective, and easier to use than any existing method, what do you predict will happen to the number of unwanted pregnancies?
SOLUTION: It is a safe bet that if a new method of birth control is safer, cheaper, more effective, and easier to use than any existing method, some people will switch from other methods to the new one. It is also a safe bet that some people will switch from abstinence to the new method, or switch from having sex once in a while to twice in a while. So unless the method is perfectly effective, the number of unwanted pregnancies can go either up or down.
Nosmo King is an antismoking crusader who finds that people who don’t recognize him sometimes offer him a cigarette. He always takes the cigarette and throws it away, figuring that if he does this 20 times a year, there will be 20 fewer cigarettes smoked. Is his prediction correct?
SOLUTION: Nosmo is correct that as long as he accepts and discards 20 free cigarettes a year, there will be less smoking. He is wrong to think there will be 20 fewer cigarettes smoked.
After all, once you give a cigarette to Nosmo, you run out a little sooner, and buy a new pack a little sooner. The convenience store where you do your shopping runs low on inventory just a little sooner and reorders just a little sooner, leading its supplier to run out a little sooner. This (ever-so-slight) increase in the demand for cigarettes percolates up the supply chain all the way to the tobacco companies, which notice that buyers want ever so slightly more cigarettes than they’d planned to manufacture—and that they can afford to ask for a slightly higher price.
At the new, higher price, the tobacco companies are willing to produce a few extra cigarettes. If they produce, say, an additional 7, and if Nosmo discards 20, then the number smoked falls not by 20 but by 13.
The same process works in reverse whenever an activist group asks people to go meatless for a day or a week in order to make more food available for others. I’ve seen the slogan “Eat a pound less so someone else can eat a pound more.” But when you and others eat a pound less, the fall in demand leads to a fall in price, which leads producers to provide less meat—so your sacrifice might mean others can eat more, but it does not mean they can eat a pound more.
Over the next few decades it’s expected that technological improvements will vastly increase worker productivity in many industries but not in all. Barbers, for example, will probably be no more productive in the year 2050 than they are today. Do you predict that the wages of the average barber, relative to those of the average worker, will be higher or lower in 2050 than they are today?
SOLUTION: The safe prediction is that wages will rise for more productive workers—and that barbers, even though they’ll be no more productive, will see their wages rise in tandem with everyone else’s.
That’s because higher wages on (say) the auto assembly line tend to lure barbers away from haircutting. The supply of haircuts falls, so the price of a haircut—along with the wage of the average barber—must rise.
And if the wages of barbers rise only a little bit, then the exodus from haircutting to auto assembling will continue, raising barbers’ wages further, until they’re high enough to keep the barbers in the barbershops.
The fact that rising productivity in some industries leads to higher wages in other industries is known to economists as the Baumol effect. The Baumol effect not only predicts the future, it also explains the past. It takes your barber about ten minutes to provide a basic haircut. It took your grandfather’s barber about the same. In other words, your barber is no more productive than your grandfather’s. But your barber is a lot richer than your grandfather’s barber, and the reason is that productivity has risen in other industries.
As a teacher, I’m very glad to report that the same reasoning applies to teachers. While today’s farmer feeds twelve times as many people as a farmer of fifty years ago, I still grade essays at about exactly the same rate as my predecessors in the seventeenth century. Nevertheless, I earn a lot more than they did, because a lot of potential teachers have instead become, if not farmers, then computer programmers or financial analysts, whose productivity has also skyrocketed.
Leopold is a great consumer of mutton kidneys.
a) If Leopold comes into a nice inheritance, what happens to his kidney consumption? Does it increase or decrease?
b) If the price of mutton kidneys rises, what happens to Leopold’s kidney consumption?
c) If I tell you that the answer to (a) is that Leopold’s kidney consumption increases, does that change your answer to (b)?
d) Extra credit: What is Leopold’s last name?
SOLUTION:
a) When I left graduate school and started to earn a decent living, I started eating a lot more steak and a lot less bologna. If Leopold thinks of kidneys the way I thought of steak, he’ll eat more of them. If he thinks of kidneys the way I thought of bologna, he’ll eat fewer of them. That’s about all we can say.
b) There are two factors in play here. First, when the price of kidneys rises, Leopold has an excellent reason to consume fewer: his fourth or fifth kidney might just not be worth it to him at the new price.
But there’s also a second factor: A rise in the price of kidneys means Leopold can no longer continue to live quite as well as he’s accustomed to. In other words, he’s effectively poorer. We saw in (a) that when Leopold becomes richer (or poorer), his kidney consumption could either rise or fall.
Bottom line: The first factor gives Leopold a good reason to cut back on his kidney consumption. The second factor—the effect of feeling poorer—might give him either a second reason to cut back or a reason to do the opposite. We have no idea which reason is more compelling, so we have no idea whether his overall consumption goes up or down.
c) We still have the same two factors in play. The first factor remains a good reason for Leopold to eat fewer kidneys. As for the second factor, the price increase makes Leopold effectively poorer, and we’re now given the additional information that when he’s poorer, he eats fewer kidneys. (Actually we’re told that when he’s richer, he eats more kidneys, but that comes down to the same thing.) Now both factors point in the same direction, so there are two good reasons for his consumption to fall and zero good reasons for his consumption to rise. His consumption goes down.
d) Try asking an English major.
Although this solution is exactly what I’d expect from my students, I should add that there’s one further consideration here: Very few people spend large fractions of their income on mutton kidneys. Therefore, when the price of kidneys goes up, very few people feel a whole lot poorer. That’s a good reason to expect that the second factor in the solution to (b) is ordinarily quite small, which in turn is a good reason (though not a thoroughly compelling reason) to believe that only the first factor really matters. It is therefore extremely likely (though not definite) that Leopold’s consumption falls in (b).
The extremely unlikely (but still logically possible) scenario is that Leopold currently eats kidneys six days a week but splurges on steak every Sunday. When the price of kidneys rises, he feels quite a bit poorer (after all, he buys a lot of kidneys) and therefore gives up steak, eating kidneys seven days a week instead of six—and therefore consuming more than before, in defiance of the freshman economics textbooks that declare a price increase must always lead to a fall in consumption. (Sophomore-level textbooks have always acknowledged that this law might not be universal.)
Something very like this might have occurred in Leopold’s home country of Ireland, in the years shortly before the Great Famine. On a typical day in 1844, the average adult male Irishman ate a staggering 13 pounds of potatoes. At 5 potatoes to the pound, that’s 65 potatoes a day. The average for all men, women, and children was a more modest 9 pounds a day, or 45 potatoes. Foreign travelers in Ireland routinely wrote home about the seemingly superhuman Irish appetite (which was surely driven by a hard life of manual labor in the fields, burning a lot of calories that needed to be replenished). A typical extract from one of those letters observes that “the Englishman would find considerable difficulty in stowing away in his stomach this enormous quantity of vegetable food, and how an Irishman is able to manage it is beyond my ability to explain.”
Economists have long suspected that a small rise in the price of potatoes could have had so devastating an effect on Irish families that they’d have cut back on their small and occasional extravagances and doubled down on potatoes. There is, alas, no reliable historical evidence to support this suspicion. There is, however, some evidence to suggest that something very similar does occur in parts of modern-day Asia, with potatoes replaced by rice and wheat.
If I had to guess, I’d say that the three most common prediction targets are sports, weather, and prices. I doubt I can do much to help you with the first two, but economists are at least pretty good at thinking about prices.
Following is a series of problems that will illustrate the economist’s way of thinking.
Suppose the government imposes a price ceiling on wheat, so that instead of selling at the current price of, say, $4 per bushel, nobody is allowed to charge more than $3 a bushel. What happens to the price of bread?
SOLUTION: Well, you make bread out of wheat, and the price of wheat just fell, so bakers will want to supply more bread, which drives the price down, right?
Hold on. Let’s start over. You make bread out of wheat, and the price of wheat just fell, so farmers will supply less wheat (either switching over to other crops or perhaps, eventually, selling their farms to developers). If there’s less wheat, there’s got to be less bread—which drives the price up, right?
The second argument is correct: the price of bread goes up.
What’s wrong with the first argument? Bakers are indeed happy that the price of wheat went down. But they’re also quite unhappy that wheat has become harder to find. They’re paying less to the farmer but spending a lot more time scrounging around trying to find a farmer who’s willing to sell to them. That takes much of the fun and profit out of making bread, so the statement “bakers will want to supply more” is wrong.
In a town with just one auto mechanic but several bars, a new law requires each business to contribute $20,000 a year toward the construction and maintenance of city parks. Which do you predict will rise more: the price of a car repair or the price of a drink?
SOLUTION: There’s no reason on earth for the price of a car repair to change. Alfred, the lone auto mechanic, has long ago set prices that he believes will maximize his profits. He still wants to maximize his profits, so there’s no reason to change those prices.
Needless to say, Alfred is unhappy about the new law, just as he’d be unhappy if he’d somehow misplaced a suitcase containing $20,000 in cash. But losing your suitcase is no reason to alter your business practices. If raising prices were a smart strategy, Alfred would have raised them long before he lost his suitcase, or long before the new obligation was imposed.
The only way Alfred can avoid this cost is to close up his shop, which he might or might not do. But as long as he stays in business, his price list remains unchanged.
In the short run, you can make exactly the same argument about the bars. There’s no reason any of their prices should change. But over time, we should expect some of those bars to close. After all, there are several of them—and with competition like that, they can’t all be doing exceptionally well in the first place. Chances are excellent that in the face of an extra $20,000 expense, a few of them will simply give up.
That changes everything. Once a few bars close, customers migrate and the demand for drinks at the remaining bars increases. And as everyone knows, it makes sense to raise prices when the demand for your product goes up.
So, in the very short run, no prices change. Eventually, a few bars close and the remaining ones raise their prices. And after you crash your car on the way home from one of those bars, the auto repair will cost just as much as it always did.
A famous Chicago Cubs baseball player demands a $10 million raise, and the management accedes to his demand. What do you predict will happen to ticket prices at the ballpark?
SOLUTION: Paying your pitcher an extra $10 million a year is a lot like being forced to contribute an extra $20,000 a year to park maintenance—it’s unpleasant, but once you’ve decided to keep the pitcher, there’s no way to make it less painful.
Ticket prices are already set to maximize profit. If the Cubs’ owner loses $10 million on an ill-advised stock market investment, he’d be even more ill advised to tamper with those prices. If he loses $10 million to a hard-bargaining pitcher, the story is exactly the same.
The Chicago Cubs play at Wrigley Field, where the ushers have just demanded and received a substantial raise. What do you predict will happen to ticket prices at the ballpark?
SOLUTION: This problem is very different from the preceding one. Once you agree to give your star pitcher a raise, there’s nothing you can do to lessen the pain of paying it. But if your ushers get a raise, you can lessen the pain by hiring fewer ushers. With fewer ushers, you might prefer to sell fewer tickets, as the ushers’ primary job is to monitor the behavior of unruly fans. If you’re willing to sell fewer tickets, you can get away with charging higher prices—so why shouldn’t you?
Likewise, if the employees at Alfred’s car repair shop successfully demand an increase in their hourly wage, or if the suppliers who provide him with auto parts raise their prices, I expect the price of a repair to rise. Unlike a $20,000 annual contribution to the parks department, these are costs he can reduce by choosing to repair fewer cars—whereupon he can afford to drive away a few customers by raising prices.
If you have to pass through two tollbooths to get across a bridge, would you prefer the two booths to be owned by one troll or by two different trolls?
SOLUTION: One troll is better than two.
The only thing preventing the troll (or trolls) from setting astronomical fees is the fear of losing customers, who will find ways to avoid the bridge altogether.
For a troll who owns a tollbooth, a $1 price increase is punished by the loss of, say, 10 customers at that booth. For a troll who owns both tollbooths, a $1 price increase is punished by the loss of 10 customers at both booths. That’s twice the punishment, and hence twice the incentive to keep prices down.
You might pay an exorbitant price in any event, but because of the double-incentive effect, the total cost of getting across the bridge will be less exorbitant with one troll than it is with two.
Microsoft, a well-known monopolist, produces both the Windows operating system and the Office suite of software (including Word and Excel). To most customers, neither product is much use without the other.
As a consumer who plans to buy these products, would you prefer to see Microsoft broken up into two separate monopolies, one selling Windows and the other selling Office?
SOLUTION: One troll is still better than two.
The only thing preventing Microsoft from charging astronomical prices is the fear of losing customers.
For a hypothetical post-breakup company that makes just an operating system, a $1 price increase is punished by the loss of, say, 10 customers for that operating system. For a company (such as the current incarnation of Microsoft) that makes both operating systems and office software, a $1 price increase is punished by the loss of 10 customers for both the operating system and the office software. That’s twice the punishment, hence twice the incentive to keep prices down.
You might pay an exorbitant price in any event, but because of the double-incentive effect, the total cost of Windows plus Office is less exorbitant today than it would be under the proposed breakup.
A monopoly tire company merges with a monopoly rubber company. As a potential car buyer, are you happy about this merger?
SOLUTION: You should welcome the merger. One troll is still better than two.
The only thing preventing the tire company from charging astronomical prices is the fear of losing customers.
For a company that makes just tires, a $1 price increase is punished by the loss of, say, 10 customers for that tire. For a company that makes both tires and rubber, a $1 price increase is punished by the loss of 10 customers for both the tire itself and the rubber that goes into the tire. That’s twice the punishment, hence twice the incentive to keep the price down.
You might pay an exorbitant price in any event, but because of the double-incentive effect, the price of the tire will be less exorbitant with one monopolist than it is with two.
Do you hear an echo? The solutions to the three preceding problems are essentially identical. That’s a cause for celebration. An insight is a precious resource, and the more times it can be recycled, the more precious it becomes.
Three major computer manufacturers are seeking permission to merge into one giant company. A coalition of smaller manufacturers vocally opposes the merger, arguing that if it’s allowed to take place, the new mega-company will exercise vast monopoly power. Do you agree with that prediction?
SOLUTION: It is certainly true in some cases that the merger of three large companies can lead to vast monopoly power. But this is not one of those cases.
I know this because a coalition of smaller companies is opposed to the merger. If the merger were likely to raise prices, the smaller companies would be applauding it. When the big guys raise their prices, the little guys can follow suit.
Instead, the smaller companies must fear that the merger will lead to lower prices, which means they believe the primary effect of the merger will be not greater monopoly power but greater efficiency.
As a general rule, firms welcome monopoly power even when they’re not part of the monopoly. If anyone’s prices rise, then everyone’s prices can rise. So when one firm complains about another firm’s monopoly power, you can be pretty sure they’re dissembling.
If a coalition of well-informed consumers were vocally opposing the merger, it would be a fair guess that the merger is likely to raise prices by creating monopoly power. You can tell a lot about the probable effects of a policy by observing who’s for it and who’s against it.
Notice that the mergers in problems 14 and 15 are quite different. The first is a case of vertical integration, where a company merges with one or more of its suppliers (or one of its customers); the second is a case of horizontal integration, where a company merges with one or more of its competitors. For the reasons explained in the solutions, vertical integration is nearly always a boon to the consumer, while horizontal integration can go either way—but you can often tell which way things will go by observing the reaction of firms that are left out of the merger.
In the 1800s the Pullman Palace Car Company was famous for building luxurious railroad passenger cars. It was famous too for building the town of Pullman, Illinois, as a place for its employees to live and work. Pullman owned all the housing and rented it to the workers. It also owned grocery stores, where the workers shopped.
Elsewhere in Illinois, the demographically similar town of Pushman had many competing employers, many competing landlords, and many competing grocery stores.
Assuming the Pullman Company acts wisely in its own interests, where would you expect groceries to be more expensive—in Pushman or in Pullman?
(Incidentally, the story of Pullman is true. The town of Pushman was invented for this problem, but surely there were many towns very like it.)
SOLUTION: You should expect grocery prices to be the same in both towns.
To see why, imagine that the Pullman Grocery charges the same competitive prices as the groceries in Pushman, and try asking whether Pullman has any incentive to raise those prices.
The key is to note that Pullman, being a profit maximizer, presumably pays its workers just enough to keep them from moving to Pushman in search of a better life. So if prices go up at the Pullman grocery store, Pullman has to compensate the residents through higher wages, lower housing prices, or both.
To see how that plays out, take an example: Initially, the Pullman Grocery charges the competitive price of $10 for a week’s worth of groceries. Suppose Pullman raises grocery prices by 50 percent. You might imagine that Pullmanites now pay $15 a week for their groceries. Not so. Probably they tighten their belts, eat a little less, and pay a grocery bill of, say, $14 a week.
To a Pullmanite, that’s an extra $4 out of pocket and less food on the table. That’s worse than just losing $4, so to keep that Pullmanite happy, the company must give back more than $4 in either higher wages or lower rents.
Bottom line: the company collects an extra $4 at the grocery store and loses more than $4 somewhere else. That’s a losing proposition, and Pullman is presumably too smart to go for it.
This conclusion is perfectly general. Change the numbers in the example any way you like, and the end result is still the same—Pullman does best by keeping grocery prices as low as they are in Pushman.
I teach at a university whose administrators might have been able to learn a thing or two from the Pullman Company. We once had a lovely little on-campus restaurant where the faculty mingled at lunchtime. The restaurant lost money, so the administration shut it down. This struck some of us as shortsighted; anything that makes it easier to recruit and retain faculty can contribute to the larger enterprise in ways that the accountants might not immediately measure. Even if all you care about is the university’s bottom line, losing $10,000 at the restaurant can be worth it if the alternative is to make the campus so unpleasant that you’ve got to start offering higher wages to keep people around.
In other words, insisting that campus restaurants make a profit is shortsighted in exactly the same way that it would be shortsighted for Pullman to insist on extracting monopoly profits from its grocery stores. It overlooks the fact that if you want to keep people around, and you do something that makes their lives worse, you’ve got to do something else—and possibly something more expensive—to make their lives better.*
By way of making this point, I posted a brief message to the all-faculty email list, observing that if we accept the logic that led to the restaurant closure—that each division is to be judged by its individual profits—then the first thing we should do is close the library. The outraged responses from faculty in other departments were unanimous in condemning those damned economists who want to close the library.
Who would have predicted that?
The one sure-fire way to make successful predictions is to predict every possible outcome. Predict both rain and sun for tomorrow, and you’re sure to be right. (You’re also sure to be wrong, but maybe you can learn not to care.) If you can get away with calling both “heads” and “tails” on the same coin toss, you’re a winner every time.
There’s a classic scam that runs along these lines: Pick a stock at random, tell 256 people its price is about to go up, and tell another 256 it’s about to go down. After a week, when one of your forecasts has proved accurate, take the 256 people who got the accurate forecast, divide them into two groups of 128, tell one group that some other randomly chosen stock is about to go up and tell the other that the same stock is about to go down. After another week, you’ve provided 128 people with two accurate forecasts in a row. Divide them into two groups of 64, and repeat. Pretty soon you’re down to 8 people who have heard you make six accurate forecasts in a row. Tell them that if they want your next forecast, they’ll have to pay for it.
Making honest predictions is a lot more difficult, and sometimes a lot less lucrative. But it’s definitely more satisfying.