The Metric Fallacy

We often hear some proclamation, usually in support of subsidizing some activity, that some activity is desirable because it leads to success (for example: “Education is an investment in the future”).  But this is what I will henceforth call the “Metric Fallacy,” which is a specific form of the “ad hoc ergo propter hoc” fallacy.

The fallacy goes like this:

Some desirable outcome, X, is highly correlated with broad category Y.

Therefore, to achieve X, one should increase Y.

The first statement is true (X and Y are correlated).  The second statement, however, is fallacious.  A practical example will help:

It is true that higher education is correlated with higher levels of income.  According to the College Board, a person with even just some college education earns approximately 13% more than someone with just a high school degree.  Higher levels of education are correlated with higher levels of income.

But it is fallacious to, therefore, say that education itself is the cause here.  We need far more information.  The type of education matters a whole lot more.  A person in theology earns considerably less than a person in chemical engineering.  Ceteris paribus, a doctoral degree in a STEM field will earn a lot more than a doctoral degree in religion.

Thus, the Metric Fallacy is: to increase income, we need to increase education.

We see the Metric Fallacy all over the place and in high-ranking areas (Education and Economic Growth Theory and subsequently foreign investment, just to name a few).  Many government policies are based off this fallacy, including minimum wage, subsidized college loans, foreign aid plans that involve capital donations, and the like.  Indeed, probably the greatest lie I heard as a child was “if you want a good job, you have to go to college.”

My colleague and professor Chris Coyne writes a lot about this in his excellent book “Doing Bad by Doing Good,” and it is the main thesis of Bill Easterly’s book “The Tyranny of Experts.”

The other day, I wrote about discovering prices vs imposing prices.  Discovering prices means that we see what prices emerge from the price system, people acting in response to relative scarcity.  When prices emerge, they provide information.  A person in a profession who earns a high salary means there is a relative scarcity of that profession.  He is indeed made better off by this, but it also signals that we need more of that profession.  Prices allow us to discover this information.  But it is a fallacy to conclude that his higher wage is in some way desirable for anything other than this purpose (eg, the protectionist who might argue if he is made “worse” off because of competition, we all are made worse off).

Today’s Quote of the Day…

…is from page 99 of Roger Koppl’s 2018 book Expert Failure (original emphasis): 

Many thinkers (and most undergraduate economics students) seem to believe that any “laws of economics” would have to be first devised and promulgated and then followed.  If there is a law, there is a lawgiver.  But the “laws” of a spontaneous order generally function even when nobody is aware of them.  Thus, an increase in the supply of a commodity causes its price to fall in Homeric Greece no less surely than in nineteenth-century London.

JMM: We can broaden Dr Koppl’s statement to more than just economic laws.  Laws, as distinguished from legislation, are followed without necessarily any conscious knowledge of the law.  Some laws are universal (such as the increase of supply noted here), and some may be specific to a time and place (such as the orderly fashion a crowd leaves a stadium after a ballgame), but prior direction of the laws are not necessary to their functioning.  Indeed, we tend to pick up on the laws not necessarily through formal education, but rather through social cues from our interactions.

The fact that laws are not necessarily known beforehand and explicitly leads to the reason why they need to be discovered (see my post here).  Laws are something observed though human action.

 

Costs and Benefits are Culturally Dependent

The above screenshot is taken from Futurama.  Lrrr and his wife, Ndnd, rulers of the planet Omicron Persei 8, are aliens from a warlike race.  Violence permeates their culture.  While Lrrr himself is not particularly warlike (he half-heartedly invades Earth in one episode just to stop his wife from nagging him), the culture still is ingrained in him and much of his confusion regarding Earth is why it is not more warlike (another quote I could have chosen is when he is watching an episode of the in-show spoof of Ally McBeal called Jenny McNeil, he asks “if McNeil wants to be taken seriously, why does she not simply tear the judge’s head off?”).

Lrrr’s confusion stems from the two different cultures of Earth and Omicron Persei 8.  What is “foolish” behavior for him is normal behavior for the people of “ancient” Earth (Futurama takes place in the opening decades of the 3,000’s, so “ancient Earth” to him is the 90’s and 2000’s for us).  For Lrrr, there is a benefit to demonstrating strength before (or to) friends.  For the people of Earth, it is a cost.

This simple fact makes the job of the economist difficult when it comes to regulation.  The typical Econ 101 story for regulation is to do a cost-benefit analysis and, if the costs are less than the benefits, then regulation may be justified.  But this seemingly straightforward process is, as we can see above, actually quite complex.  What counts as a cost and benefit, as well as the relative values of those costs and benefits, are subjective.  It depends on the person observing the situation.  For us, Ross’ behavior is normal and praiseworthy.  For Lrrr, the exact same behavior is confusing and foolish.  And Earth observer might propose a regulation that rewards people who act like Ross.  And Omicron Persei 8 observer might propose legislation that punishes people who act like Ross.

Ultimately, the determination of regulatory behavior comes down to the subjectivity of the observer.  What this means is “optimal” regulation (such as a tariff, tax, wage, etc), while possible in theory, is practically impossible.  What may be an optimal tax for one observer may be too high/low for another and may be a subsidy for a third.

This subjectivity is why I adhere to a status quo bias and a presumption of liberty.  Legislative and legal changes should not be done casually, and even with great thought and analysis, should be undertaken only very carefully.

Today’s Quote of the Day…

…is from page 14 of Mancur Olson’s 1982 book The Rise and Decline of Nations:

Although we should not be satisfied with any theory that fails to explain a lot with a little, we need not of course expect any one theory to explain everything, or even the most important thing.  Absolutely nothing in all of epistemology suggests that valid explanations should be monocausal.

JMM: It’s typical to hear opponents of this theory or that theory be dismissive of the theory because it fails in some way (“Economics is broken because it failed to predict the 2008 recession!” or “Socialism is broken because it failed to predict Venezuela!” etc).  But the world is a complex place.  Few things, if any, are monocausal: Did the Red Sox win the World Series because they had the largest payroll?  The best players?  Luck?  A favorable schedule?  Some kid in Maryland was watching them?  All of these factors played in (well, almost all of them).  They’re all part of the cause (and some even in more ways than one).

When evaluating a theory, we always need to remember to evaluate it on its own terms.  Does it make sense doing what it is trying to do?  Three great examples of this are Adam Smith’s examination of mercantilism in The Wealth of Nations (see Book IV) and Hayek and Mises’ critiques of socialism in the Socialist Calculation Debate.

Today’s Quote of the Day…

…comes from page 236 of Bruno Leoni’s 1961 work Freedom and the Law (3rd Edition, emphasis original):

If we consider it well, there is nothing “rational” in voting that can be compared with rationality in the market.  Of course, voting may be preceeded by argument and bargaining, which may be rational in the same sense as any operation on the market.  But whenever you finally come to vote, you don’t argue or bargain any longer.  You are on another plane.  You accumulate ballots as you would accumulate stones or shells – the implication being that you do not win because you have more reasons than others, but merely because you have more ballots to pile up.  In this operation you have neither partners nor interlocutors but only allies and enemies…The political language reflects quite naturally this aspect of voting: Politicans speak willingly of campaigns to be started, of battles to be won, of enemies to be fought, and so on.  This language does not usually occur in the market.  There is an obvious reason for that While in the market supply and demand are not only compatible but also complementary, in the political field, in which legislation belongs, the choice of winners on the one hand and losers on the other are neither complementary nor even compatible.

JMM: This difference between the market and the political realm is utterly lost on those who advocate for government directing of the market: those who argue for “trade wars” or “minimum wage” or any other government interventions.  Markets are about cooperation, not violence.  When China sells goods to the US, it is not a battle, China is not the enemy.  When Wal-Mart hires workers at a given wage, it is not Wal-Mart exploiting workers or some great battle between labor and management, but rather cooperation between the two.

Political language easily lends itself to conflict and violence.  But to use that same language in the market is to fundamentally misunderstand what the market is.

Local Knowledge, World Series Edition

In Game 4 of the 2018 World Series last night, the LA Dodgers held a 4-0 lead over the Boston Red Sox going into the 7th inning.  Dodgers pitcher Rich Hill was almost unhittable.  No Red Sox player had advanced beyond first base to that point.  However, in the 7th inning, Dodgers manager Dave Roberts pulled Hill from the game for a relief pitcher.  The Red Sox go on to score 3 in the 7th inning, 1 in the 8th inning, and 5 in the 9th inning to win the game 9-6.

President Trump sent out the following tweet during the game:

Watching the Dodgers/Red Sox final innings. It is amazing how a manager takes out a pitcher who is loose & dominating through almost 7 innings, Rich Hill of Dodgers, and brings in nervous reliever(s) who get shellacked. 4 run lead gone. Managers do it all the time, big mistake!

— Donald J. Trump (@realDonaldTrump) October 28, 2018

 

Prima facie, Trump looks like he’s got a point.  Why would Roberts pull a lights-out pitcher from the game?  Especially in retrospect, it seems like a terrible move that cost the Dodgers the game.  But Dave Roberts is no fool.  Let’s ask the man himself:

“I didn’t hear about it [the tweet]. … The president said that?” Roberts responded. “I’m happy he was tuning in and watching the game. I don’t know how many Dodger games he’s watched. I don’t think he is privy to the conversation [had with Rich Hill]. That’s one man’s opinion.”

The conversation Roberts was referencing was the one he had with Hill in the dugout prior to the start of the seventh inning, during which Hill told the manager to “keep an eye” on him, as the lefty was starting to fatigue.

Roberts had local knowledge the President (and the millions of people watching the game) did not: Hill was getting tired.  If Hill stayed in the game, the Red Sox may have scored even more runs.

Roberts’ decision only looks bad in retrospect and without the local knowledge he had.  When incorporating in that local knowledge, the decision makes a lot more sense and it is doubtful the outcome would have changed.

To bring this to economics, we see how important local knowledge is to make decisions.  When governments try to direct activity, they necessarily do not know this local knowledge.  Donald Trump makes the same mistake with his cries of “tariffs” and “losing at trade” as he does with his baseball analysis.

Market Power Does Not Equal Coercive Power

Below is an open letter to Bloomberg:

There is a lot to like in Mark Whitehouse’s op-ed from October 21 (US Labor Markets Aren’t Truly Free) .  However, one place he errs is where he writes (emphasis added): “Economists have offered various explanations, including labor-saving technology, weakened unions, and growing competition from lower-wage countries such as China. More recently, though, they’ve identified another: The job market has become less free. The consolidation of American business has left people with fewer places to work, shifting the balance of power to employers.

The consolidation of American businesses does not necessarily mean the job market is less free.  So long as workers are free to move jobs or relocate, then the job market remains free regardless of its concentration.  So long as transactions are coluntarily entered into and there is no outside interference, the exact structure of a market does not matter when determining freedom.

It is not from the concentration of the market that the lack of freedom arises, but rather the other factors Mr Whitehouse identifies: property zoning, non-poaching agreements, etc.  The industry concentration may be a symptom of these lack of freedoms, but they are not the cause thereof.

The Difference Between Econ 101 and Econ 801

An advantage of teaching undergraduate students as I simultaneously work on my graduate work is I get to go through both undergrad textbooks and graduate books at the same time.  Currently, I’ve been working my way through Cowen and Tabarrok’s Principles of Economics for my undergraduate class and George Stigler’s Theory of Price and Donald Watson’s Price Theory and Its Uses for my research.

We often hear, primarily by people who dislike the implications of Econ 101 models, that Econ 101 is too simplistic, too unrealistic for the real world.  They’ll point to other economic models that better conform to their desired views (eg, the monopsony model for minimum wage).  “We can’t rely on Econ 101!  It’s just too simple!”

But what’s interesting to me is that going through these upper-level texts (Stigler is a high-Masters, low-PhD text), one sees they are not fundamentally different from the undergrad texts.  This holds in other texts, too, such as David Kreps’ Microeconomic Foundations.  There may be more math to formalize the models, but the intuition remains the same; the implications remain the same.

Basic supply and demand analysis gets us a very long way.  It is not complete, of course.  No theory ever is.  But supply and demand is fundamental.  Seeking to overthrow the foundations will not necessarily lead to a more coherent theory.

Be Skeptical of Regulations Because Knowing the Market is Difficult (Even for Experts)

Given how much money business consultants make, one would think they have pretty good insight into a given industry or market.  And sure, they may have lots of information unavailable to most people, but does that necessarily imply they are better?

All data received, we must remember, is context-dependent.  Data never, ever, speak for themselves.  Interpreting and developing models for given market structures is extremely difficult in this regard because it requires certain assumptions.

Consider the following real-world example: when the guys from Xerox (the copy machine company) wanted to start selling their machines to businesses, they met stiff resistance from consultants and financial backers.  “Why would anyone spend thousands of dollars on a copy machine when we have a perfectly good, cheap substitute: carbon paper?  Copy machines will never sell.”  Prima facie, this criticism seemed legitimate.  Firms and experts observed secretaries and typists using carbon paper to make copies for distribution.  There didn’t seem to be any demand for copy machines.

Undeterred, the Xerox guys pressed on.  They decided to give companies the hardware, toner, and paper for free up until the first 2500 copies per month.  Firms jumped at the idea.  After all, how were they ever going to use 2500 copies a month?

Well, the rest is history.  Xerox is still around.  Carbon paper is not.

Why?  What did the experts get wrong?

They got wrong the scope of the market.  Copy machines weren’t for people writing letters.  They were for people receiving and distributing letters!  The market for copy machines was for the owner who got a letter from his lawyer who wanted to distribute it to the rest of the team.  It was for the product manager who needed to itemize his costs for different departments.  Etc etc.

There is an implicit conceit in economics that we know the market, the shapes of demand and supply curves.  But the reality is, we never do.  We have only data points in a certain context which may or may not provide useful information about other contexts.  This implicit conceit becomes vitally important when we start talking about regulation or “optimal taxes,” which require knowledge of the scope and shape of the market.  Knowledge we simply do not have.

A Hidden Danger of Aggregation

I have been verbally outspoken against macroeconomics (not so much writing).  Macroeconomics, that is the examination of aggregate economic data for the purpose of discovering observable phenomena, is quite distinct from microeconomics.

Despite my verbal abuses, macroeconomics does have uses.  Observing these phenomena, such as the impact of interest rates or foreign exchange rates or GDP, can provide useful information.

But they are only useful insofar as the assumptions behind these aggregate measures hold; only insofar as they are used as a signal.  GDP, for example, is used as a measurement for economic well-being.  In a relatively free market, with a working price system and protected property rights, it acts as a pretty good indicator.  Given these assumptions, if GDP rose, say, 4% over the course of a year, it’s reasonable to conclude the people who make up the economy are likely approximately 4% better off than they were the previous year.

However, once those measurements become an object of choice, then they lose all value.  GDP growing at 4% due to market forces tells us important and helpful information.  GDP growing at 4% due to deliberate actions taken by governments to increase that number (say, arbitrarily increasing government spending) does not tell us anything useful.  If anything, it can be misleading.  The signal is distorted.

This argument is the same as discussions on the distortion of price signals.  A price, when determined by supply, demand, and all their components, provides valuable information on the relative scarcities of goods.  A price, when it becomes an object of choice and is thus manipulated to meet some desired level (eg, minimum wage, price controls, tariffs, etc) provides no useful information.  The price no longer functions as a price.

The great sin of macroeconomics is confusing this point; how often do we hear an argument along the lines of “Higher GDP is good.  This action would increase GDP.  Therefore, we should do it”?  But a higher GDP is only useful insofar as it conveys market transactions.  Anything else, and it becomes useless.