Economists Mustn’t Forget the Fallacy of Decomposition
By Gary Galles • Tuesday August 9, 2016 2:29 PM PDT •
Because of their field’s stress on logic, economics teachers commonly emphasize some fallacies for students to avoid in the early stages of principles courses. One widespread example is the fallacy of composition—the mistaken notion that what is true of an individual is necessarily true of a group. In reality, of course, what is true of an individual need not be true of a group.
Since so much of a person’s economic intuition derives from his or her understanding of what they would do when facing particular circumstances, this focus makes a great deal sense.
Perhaps the best-known illustration is: standing up at a football game. If an individual fan stands up, other things equal, he or she will be better able to see the action on the gridiron. But if everyone stands up, that will no longer be true. The view may be just as obstructed as before.
An example using money shows how the fallacy of composition could apply to economics. If one person had more money, other things equal, they would have more wealth (at others’ expense), because they could then command a larger share of society’s resources. But if all people had proportionately more money, prices would also rise proportionately and society would not have more wealth.
The fallacy of composition has important applications in economics, but there is a complementary fallacy that gets less attention than it deserves. That is the fallacy of decomposition—the false notion what is true of a group is necessarily true of an individual. In reality, what is true of a group may not be true of an individual.
Consider the monetary illustration above. The “all get more money” story has several implicit assumptions: It assumes, for example, that every person gets proportionately more money at the same time, and that prices have had time to completely adjust to the new level of spending made possible by the larger amount of money. But that is not how monetary expansion actually works in the real world. Everyone does not get proportionately more money at the same time. In reality, some people will receive the monetary increases before others do, and some prices will change faster than others. Those realities, in turn, imply that some people will get wealthier at others’ expense.
They also imply that more mistakes will be made during the process of economic adjustment, mistakes which will make society poorer by undermining the social cooperation that is coordinated by changes in relative prices. Economics students who focus solely on the fallacy of composition may forget these implications and thereby fall prey to the fallacy of decomposition—especially in a world where public policies are often driven by the short-run redistribution of wealth. (This redistribution is often disguised with rhetorical cover insisting that the policies are “really” about efficiency or growth or fairness.)
Cartels and other collusive arrangements offer an important illustration of the fallacy of decomposition. Cartels form because it can be in a group’s interest to jointly restrict output, as a means to higher prices and mutually higher profits. Hence Adam Smith’s reference to group meetings that end in a “conspiracy against the public.” Fear of that joint interest is behind a great deal of antitrust policy. However, it ignores the fragility of private-sector cartels, because what is true of the colluding group need not be true of individual producers.
The three main reasons cartels do not form or break down are all applications of the fallacy of decomposition.
While it is in the joint interest of a colluding group to restrict output, an individual producer could do still better by not joining the group, and thereby benefit from higher prices as others restrict their outputs and without needing to restrict its own output. Similarly, for those who join a colluding group, individual members can often do better by cheating on the agreement (e.g., by not abiding by commitments to reduce output or offering secret price reductions). And even if a cartel group was successful in raising a good’s price, that success gives others who are not cartel members an incentive to enter the market in search of profits, thereby undercutting the cartel’s ability to generate profits.
Inattention to the fallacy of decomposition can spur policies to deter something that would seldom happen under sustainable voluntary agreements. Another risk is that it distracts attention from the fact that government is by far the most effective cartelizing agent, and therefore the primary actor that consumers should actually fear.
Governments can solve each of the breakdown mechanisms more completely by employing coercive powers that firms do not possess. They can force firms to join a cartel, they can punish cheaters, and they can keep entrants out—actions all backed by civil and/or criminal penalties. In consequence, we get antitrust policies that interfere widely in market arrangements, thereby distorting what individuals would voluntarily do, yet accomplishing little consumer protection. At the same time they distract attention away from the adverse effects of government cartelization in areas ranging from labor markets (e.g., unions’ exemption from antitrust laws) to agriculture (e.g., price floors and agricultural marketing orders) to the almost innumerable instances of barriers to entry (e.g., licensing restrictions).
Many other policy arguments similarly commit the fallacy of decomposition. For instance, a common justification of higher minimum wages is that they must be good for “the poor,” because low-skilled workers as a group may earn more money. But even if that were so, many in the supposedly benefitted group would lose income by losing jobs, hours of work, or on-the-job training. And one cannot justify harming those poor people in the name of helping the poor.
There are good reasons for economists to emphasize avoiding the fallacy of composition. However, avoiding the fallacy of decomposition is also important, and in many cases, even more important to our understanding of economic relationships and policies. That is ample reason for economists to pay more attention to that common violation of logic, if we wish to train our students to accurately evaluate policies.