The Daily Cow

This comes up often because it is human nature to make judgments. It is human nature to use past experiences (and others' past experiences) to make inferences about the future. Those inferences may not be 100% accurate, but statistically they're the best you got given what you know. One of the main things you learn in history is that forgetting the past leads to repeated history. So why do we as individuals choose to forget the past?

Out of the desire to appear politically correct, people often reject summary statistics as "generalizations". Those who use them are "being judgmental". Just because there is a correlation between one thing and another doesn't mean the former implies the latter. But given nothing else, that's close to the best you can get--which is that the former is linked to the latter with probability r (which is the statistical correlation).

Would you fault me to claiming that tall fathers tend to have tall sons? Would you fault me for claiming that black fathers tend to have black sons? Would you fault me for claiming that the same father probably scored lower on the SAT than the national mean? Would you fault me for claiming that those who score lower on the SATs tend to be less intelligent than their peers, and vice versa? Would you fault me for claiming that those who are less successful are more likely to be involved in crime?

If you said no to any of them, ask yourself: If you had to bet 100 dollars for or against my proposition, which would you bet on for the higher expectation? Your sense of political correctness is tickling your senses, but when faced with a completely mathematical option, you have nowhere to insert your OWN judgment. Either the expectation is or isn't greater than zero.

So where's the problem with using data to back up judgments? People disregard that as stereotyping. Do I stereotype if I assert that smart people tend to have smart kids? No? What if I say that in people of a certain race tend to be less economically successful than people of the dominant race? No? What if I say that those who are more strapped for money are more likely to commit crime? No? But what if I combine them and assert that those same people of that certain race are more likely to be criminals than people of another race? (Which means if the police has (a) limited resources--enough to arrest only one--and (b) equivalent evidence against two suspects of a crime, it should choose the better guess based on past data)

Why does the PC-meter signal red on the last statement? Is it because people can see only one step in an equation? If a => b and b => c, can't people see that a => c? Or do they recognize that but refuse to acknowledge it publicly out of the desire to appear PC? Have we been so inculcated with the messages of PC-ness that we now fail to see the meaning behind it? If everything is indeed equal (level of education, past criminal history, people of association, economic standing, etc)--and that's impossible to assure--then sure, these politically sensitive criteria should play no role in people's decisions and judgments.

But not everything is equal. Therefore, we use data we know to guess at the variables we don't know. That's called the empirical mode of thinking and is the basis of the method of scientific inquiry.

So don't be so judgmental of your more perceptive peers. It's only human nature.

Many politically-sensitive topics become trivial when considering only their immediate outcomes--the poor deserve a decent life too, so give them the basic necessities like food, shelter and some spending money. But when considered as a subset of the entire picture, that line of reasoning is disastrous. The poor will have no incentive to work (if they can't earn reasonably more than what they're given for free to cover for their time and effort) and to limit their family size. With a smaller tax-base and a much larger tax-burden, the rest of society will crumble underneath the growing weight of taxation, pushing the society further into poverty.

The tension isn't just economic. It can also be social. Take abortion for instance. There are arguments to be made for both sides, but the most convincing of them are that (a) women have the right to control their own bodies (an appeal to liberty) and that (b) making it easier to get an abortion disincentivizes people from being more careful through preventive methods (an appeal to distortion of incentives). The former argument (a) preaches the outcome side while the latter argument (b) criticizes the incentive side. Neither side is willing to break the tension by crossing the border between incentives and outcomes. Therefore, neither argument is complete.

As an economist, I believe that there is a net benefit (benefit - cost) to everything, which we call "utility", an abstract value that captures the trade-offs people are willing to make. Simply capture the entire net social cost of an abortion in a price, make alternative forms of contraception more affordable (such that the cost directly varies with urgency), and the entire argument (b) becomes null while retaining women's rights to their own bodies. Perhaps make preventive methods (condoms and birth control pills) free, morning-after methods inexpensive, and abortions expensive (Sufficient data is necessary to correctly calibrate their optimal prices). This construction maintains the liberty while not distorting incentives.

The same principles apply to almost instance of tension between outcomes and incentives. Simply capture the net cost of the issue in a flat "tax", and things will naturally tend toward a more efficient equilibrium.

In a NYT article esteemed NYU Stern professor Nouriel Roubini discusses the falling prestige of the dollar and the inevitable elevation of the Chinese RMB to the world's de facto reserve currency. In attempting to adapt his explanations to the general public, he completely butchers his reasoning, and instead says something completely wrong.

He claims that "we have been able to issue debt in our own currency rather than a foreign one, thus shifting the losses of a fall in the value of the dollar to our creditors," forgetting that an external cost such as currency exchange loss or gain is incorporated into the demand-and-cost calculations of the buyer and the seller, in expectation. That is, since foreign investors knew that there was a possibility of a currency loss, the demand was adjusted accordingly to the expected return with currency fluctuations priced in (plus some factors for risk aversion), so the seller (US Government) would be facing a weaker demand than would have been the case for a safer environment.

He then claims that "waving commodities priced in dollars has also meant that a fall in the dollar’s value doesn’t lead to a rise in the price of imports," which is another form of the same mistake--forgetting that in expectation, the currency is just a veil. In economics terms, if it costs me 20 Pounds to produce a barrel of oil when teach dollar bought exactly a pound (making oil 20 dollars/barrel in cost for me), it does not still cost me 20 dollars/barrel when the dollar is worth 50 cents. Assuming other costs are constant, the cost of oil production in a local currency still being 20 Pounds/barrel, the cost has risen to 40 dollars/barrel.

Therefore, Americans are not immune to the effects of dollar devaluation and currency loss, as economics professor Nouriel Roubini claims.