Belligerati

I’ve long wondered why it is that the grocery store Trader Joe’s has such educated, clear speaking, skilled, and well, white employees. I thought maybe that the firm had such a great reputation as a cool and fun place to work that they could afford to be discriminating in who they hired. Now I’m pretty sure it is just the pay:

You can’t buy engagement from employees, but the pay at Trader Joe’s helps. Store managers, “captains” in Trader Joe’s parlance — the nautical titles are a holdover from Coulombe (newly promoted captains are commanders; assistant store managers are first mates) — can make in the low six figures, and full-time crew members can start in the $40,000 to $60,000 range. But on top of the pay, Trader Joe’s annually contributes 15.4% of employees’ gross income to tax-deferred retirement accounts.

Inside the secret world of Trader Joe’s

The whole article is interesting. As usual, it isn’t just the seemingly cool attitude that makes Trader Joe’s successful, but a series of business process innovations like reducing the number of products offered, just in time inventory, catering to customer needs in product placement, and alternative packing for fruits and vegetables to speed checkout that makes shopping there cheap and easy. I love shopping there but I find their fruits and veggies the least impressive and most expensive part of shopping there.

A number of other studies have shown that lower-income Americans give proportionally more of their incomes to charity than do upper-income Americans. In 2001, Independent Sector, a nonprofit organization focused on charitable giving, found that households earning less than $25,000 a year gave away an average of 4.2 percent of their incomes; those with earnings of more than $75,000 gave away 2.7 percent.

…

Piff has made a specialty of studying those cultures in his lab at the Institute of Personality and Social Research, most recently in a series of experiments that tested “lower class” and “upper class” subjects (with earnings ranging from around $15,000 to more than $150,000 a year) to see what kind of psychological factors motivated the well-known differences in their giving behaviors. His study, written with Michael W. Kraus and published online last month by The Journal of Personality and Social Psychology, found that lower-income people were more generous, charitable, trusting and helpful to others than were those with more wealth.

…

“Upper class” people, on the other hand, clung to values that “prioritized their own need.” And, he told me this week, “wealth seems to buffer people from attending to the needs of others.”

…

Piff found that if higher-income people were instructed to imagine themselves as lower class, they became more charitable. If they were primed by, say, watching a sympathy-eliciting video, they became more helpful to others — so much so, in fact, that the difference between their behavior and that of the low-income subjects disappeared. And fascinatingly, the inverse was true as well: when lower-income people were led to think of themselves as upper class, they actually became less altruistic.

The Way We Live Now – The Charitable-Giving Divide

It is interesting that they call this the charitable giving divide. The tone of the article might suggest that that the poor give more in total than the rich do. However, since the average person making 25 grand a year or less makes far less than a third of what the average person making 75 grand or more does, even though the giving rates of the poor are 1.55 times higher the much more than three times higher income makes the total given by the $75k+ is substantially greater. The poor are giving much less in total.

That said, the poor exist in an world of self and mutual insurance and non-market trade that that the middle class and rich do not. I watch your kids in emergencies and you watch my kids. I give you ride when your car is broken and vice versa. We give each other rides to the train station or airport instead of taking cabs, that sort of thing. As UCSD economist Eli Berman has pointed out, such poor, non-market exchange communities have reason to resist market work and other market participation. If I know that you are working hard at your job and saving, if you ask me to babysit your kids I know you won’t be able to reciprocate. So higher donations are a product of a higher level of interrelation but also part of a system that resists economic participation and so is part of the reason that these people are poor in the first place. Understood this way, Piff’s experiment doesn’t document a lack of empathy, instead it shows that there are different psychological strategies humans can employ depending on the economic circumstances they face, and that they can shift between them regardless of upbringing. What better way to make people be more charitable or more self reliant then to convince them that these strategies are more ethically desirable. Normally we think of goodness following from ethics, but believing something is morally desirable is a great way to encourage it, so ethics too can be at the mercy of what is adaptive.

I’m skeptical that the framing issue is as powerful as this experiment suggests. Experimental economists often have great difficulty persuading that the behavior evoked in the lab is “externally valid”, that is, useful in a broader context. The link on the article at PubMed has a problem so I cannot check it for myself (Having less, giving more: The influence of social class on prosocial behavior.). That is, making the rich imagine themselves as poor may effect their experimental outcome but do little or nothing to affect charitable giving outside the lab. I doubt it, because religion does appear to make all people more charitable, and one way it does that is by making those folks imagine the fate of the less fortunate. I am skeptical that the imagination part alone rather than the change in beliefs is what causes that shift.

It is possible that this article understates the giving difference between the two income groups. It ignores that being rich is a matter of assets not income. Yet the measure here is fraction of income donated and not fraction of assets. But my understanding of the sorts of people who make heavy charitable donations are that they are later in life with substantial accumulated wealth.  More poor people save nothing, so they are at the zero bound of assets to donate. Accordingly among the poor there is less variation in wealth accumulation and perhaps the charitable giving of the poor is even more impressive.

This measurement issue cuts another way. Most middle class and rich people pay more into the government than they receive. The poor get food stamps, welfare, the negative income tax, subsidized education for their children, rent support and many other forms of government transfer that are means tested (phased out as they earn more). Therefore, the poor have higher income than their incomes state and the rich lower. The right way to do this comparison is to examine income net of transfers. Given a $12k a year income and two kids you’d get the the EITC of $5k and family medicaid benefits worth something like $10k. That means that your take home value of your income net of transfers is more like double your actual income where as for someone making a million a year taxes are more like 50% of their income. Examine charitable income as a fraction of these numbers and not only the number reversed, the 4.2% of the poor becomes 2.1% of income net of transfers and the 2.7% of of the rich becomes 5.4% net of transfers. Deciding on the right thing to measure matters a great deal.

A fascinating article on playing an optimal game of Hangman has a mistake:

I suspect that the 13-game is essentially solvable. There are enough words that are easily guessed that taking more risks with those, to test the harder words, will improve the guessing algorithm from a 99% success rate to 100%. At that point, we are at equilibrium—in the words of WOPR, “A strange game. The only winning move is not to play.” (The WarGames reference is particularly relevant, since the Nash equilibrium was used as the theoretical basis for the Cold War nuclear strategy of mutually assured destruction, and the climax of the film was essentially this kind of simulation—with added computer self-awareness.)

25 Best Hangman Words by   Jon McLoone

He is right that under an optimal mixed strategy all words would be equally to be guessed. However, he’s wrong that the only winning move is not to play. Tic Tac Toe has ties as an outcome. Hangman cannot have ties, the hangman hangs or lives. He’s also wrong about the direction of simple words. Under randomization the odds of guessing an easy to guess word like “difficult” will go down because you want to lower the probability of getting that one right to raise the probability of getting a hard to guess word like “jazz” right. A guessing algorithm can be tuned and improved, but at some point it must be more likely to get a particular word right by being more likely to get another word wrong. Under the optimal strategy, this is tuned so that all words are equally likely to be selected and guessed. That’s going to happen by making easy to guess words less likely to be guessed.