Sunday, January 27, 2008

Begging to kick the habit

I'm not a smoker, so when a fashionably dressed young man asks me for a cigarette on the street, he has exactly zero chances of getting a free nicotine fix. A poorly dressed beggar asking for cash has much better odds of tapping my generosity. Presumably he wants to buy food and correctly assumes that I do not walk around with a stash of mixed nuts in my jacket. I can strongly relate to someone who is hungry: might my fortunes suddenly change, I could see myself on the street begging to stave off my hunger. But there is no way I would give anything to someone who is not hungry, so even had I been carrying a pack of cigarettes at the time, I would not have handed one to that nicotine craving chap out of charity.

Most beggars make a sufficient amount of money on the streets to eat very well; in fact, with a typical 25$/hour donation rate they can easily support such habits as smoking, drinking and using illegal drugs. So by giving money to beggars, we are helping out a small minority whose main purpose is to pay for their next meal (they need to beg only one hour a day to eat at McDonald's), and hurting the rest by paying for their substance abuse habits (75$/day can easily buy food, cigarettes, alcohol and—depending on market prices—various other illegal drugs). Unfortunately, it's very hard to estimate the charitable income that a given person on the street is getting. Might no one be giving them any money, their income could well be zero: they could be starving.

If the state gave each poor person enough cash to buy three meals a day, I would be disinclined to give money to buy better food, let alone to subsidize substance abuse. This would remove the possibility of getting any income above 25$ a day without getting a job. And it would ensure that everyone is eating well enough to get by. Some issues need to be addressed in order for this subsistence income program to work as intended.

1. How do you know who is eligible and who is not?
I would suggest making every resident of Canada eligible, so that there is little bureaucratic overhead. Well-off persons will just pay the equivalent amount of income tax to compensate for the subsistence income they are receiving. If their fortune changes overnight, the subsistence income would be immediately available, since they are getting it all the time. Persons with accumulated capital on which income tax has already been paid (such as retirees) would benefit unduly from this subsistence income, since the income tax hike could not be retroactive. All forms of changes in tax structure have this problem.

2. How do you make sure that people are not using fake identities to receive multiple subsistence income streams?
This is a tough one. Population records are fairly accurate, so massive fraud would be difficult (in terms of percentage of population). Births are quite easy to control, but it is likely that people would take a bit longer to die than usual.

3. How can we make sure that everyone on the street is covered, illegal immigrants and all?
This is another difficult one to answer. Perhaps the effects of immigrants would be minor.

4. How can you send money to and communicate with people who do not have a mailing address (ie. the homeless)?
Technology is a big help in this area. It's trivial to give someone a "withdraw-only" bank account using electronic payments such as Interac with very little infrastructure. It is also quite cheap to communicate with literate persons by email. When mobiles phones become the mainstream payment medium (this is clearly only a few years away), payments could be made to a mobile phone account, and everyone could be expected to have one (the same way that everyone needs a health insurance card). This would solve the communication issue for homeless people permanently. Obviously mobile phones need to be cheap for this to work (~30$), and enough public electric outlets to charge them would be needed, but I think these two requirements are already met. Obviously airtime and SMS to communicate with the government should be free.

5. How do you prevent people front-loading all their spending as soon as they get their subsistence income cheque, requiring hunger alleviating begging later on during the payment interval?
Electronic payments make this really easy. Just send 7$ at 6h00 for breakfast, 8$ at 12h00 for lunch, and 9$ at 18h00 for diner, et voila! Electronic micro-payments work much better than monthly paper cheques, and they are economical. Better yet, it is difficult to convert such small amounts into paper cash, preventing the sums from being spent on the black market. So buying illegal drugs would require one to accumulate a day's worth of payments before withdrawing the large 20$ lump sum, which is extremely difficult for normal humans to do—hyperbolic discounting at its best!

6. How do you prevent people front-loading all their spending by borrowing against future subsistence income?
Making the subsistence income legally unseizable makes formal lending impossible. Informal lending (like buying 100$ of drugs against the next four days of revenue) seems very difficult to enforce: the amounts are too small to warrant much violence, and the threat of violence would also fail since collection would require four days without spending anything. The account's maximum value would be around 9$ for most people!

7. How do you prevent people from buying cigarettes, alcohol and drugs rather than food, requiring hunger alleviating begging?
This is a tough one as well. We could rely on hunger being a stronger incentive than the drug addiction. Or we could rely on hyperbolic discounting again: as long as the minimum purchased quantity of alcohol, tobacco and drugs cost more than one meal's worth of cash (9$ in this case), it is unlikely that people would go hungry six hours without eating in order to accumulate enough cash for a fix.

8. Is the amount low enough to keep minimum wage jobs attractive?
Probably. 9125$ of yearly subsidy + 16000$ of minimum wage * 0.7 income tax = 20325$. You can more than double your earnings by working: not a bad deal.

Sunday, January 20, 2008

In defense of destiny

Here is a great quote from a scene in Quebec-Montreal, in which two friends are chatting about how to strike-up a conversation with an attractive opposite sex stranger.

Female: "This jerk was hitting on me last night at bar. A total moron, half-drunk. He was going to Cuba, and invited me there with him. What a dweeb! Jeez! I'd never go with some stranger."

Male
: "It's not always easy to find the right words. It's as if talking to a woman you don't know isn't politically correct. Really, you always seem annoyed. It's like the only way to approach a woman is by chance; otherwise, it's impolite."

Female
: "That's not true!"

Male
: "That guy that was leaving for Cuba, had you met him there, after he tripped over your long chair on the beach, wouldn't his odds have been better?"

Female
: "But that's the ideal situation."

Male
: "Exactly, you are always waiting for the ideal situation, in which the guys is not so explicit about his intentions. It's the situation that wins you over, not the conversation."

Female
: "Chance is a question of destiny, and that's what women like."

Single men will usually strike-up a conversation with a female stranger in the hope of eventually having sex, that part is clear. Often the attraction is reciprocal, yet sex remains a rare outcome if the feeling of destiny fails to materialize. Why is this? Two forms of adverse selection, in favor of the man, are to blame.

First, it would be wise to assume that the initiator of the conversion has a certain ease in chatting-up strangers. This implies that his ability of finding other mating partners is quite high: his market for mates is much more liquid than that of the average person. It could also mean that the initiator has some inside knowledge that his odds are good: perhaps he has tried this before and it worked for him. So the odds of this guy wanting a long term relationship with her are much lower than in the population at large.

Second, the initiator chooses whom he chats to, in such a way to maximize his pleasure by choosing the most attractive, wealthy and intelligent person that he has access to. Whereas the chosen person can't expect to get any better than average (or below average) for any of her criteria.

When a rare uncontrollable event occurs in the proximity of two people of opposite sex, the social barrier to start a conversion basically disappears for both of them. It could be that something supernatural wants them to hook-up (unlikely); or it could be that the guy finally has something to say that could not have been planned ahead of time, which he can only say to a few people who are physically close to him. The female realizes that the male's advantage in choice has been materially eliminated. He quickly assesses that his odds with this person are exceptionally good right now, so he says: "You don't see that every day. Hi, I'm Jim..."

I suspect that if roles were reversed (that is men waited around for women to buy them drinks and chat with them in bars) destiny would start playing a key role for males in the process of long-term mate selection, and would eventually be disregarded as a fitness indicator by females. Putting yourself in a situation where "coincidences" happen often improves your odds of having destiny on your side. If you are still not convinced, just watch the movie Serendipity.

Though not very romantic, it would be quite easy to emulate destiny using a random process and matching roughly compatible singles at a time interval which follows a Poission distribution with an average of about two months. Perhaps Loto-Quebec could make this work if it was presented in the form of a lottery, in which the winning pair wins a date at a fancy restaurant. Hum... not a bad idea!

Saturday, January 5, 2008

Hyperbolic discounting and bubble gum

An interesting experiment can be conducted involving children and chocolate: offer a child the choice between one piece of Caramilk consumable immediately, or the entire bar consumable after one minute (expressed in an understandable time unit). Which one does he choose? Preschoolers overwhelmingly choose immediate gratification, whereas ten-year-olds choose to wait for the greater reward. Though this is not the most thorough example of hyperbolic discounting, it illustrates that the time horizon we consider when making decisions gets longer as we mature.

While having suppressed the most counter productive forms of hyperbolic discounting, older humans will likely proceed to eat the entire stash of chocolate there and then, rather than eating a piece a day, which would definitely give them more total pleasure.

The spending pattern of many adults is strikingly similar to the aforementioned glutton's food stock dilemma. During the two week period between paychecks, most people front-load their spending. When spending is on capital goods (those used for many weeks at a time) this makes complete sense. But spending is also skewed on goods that give only immediate satisfaction (like expensive wine, eating at a fancy restaurant, going skiing), forcing one to consume sub-par beer and frozen meals at the tail end of the pay cycle. Choosing slightly cheaper wine would have made more sense in retrospect, yet the cycle repeats itself identically the next time around: it seems that adults can't learn to distribute income evenly over a two week period, regardless of experience and incentives.

So either humans innately can't break the two week hyperbolic discounting barrier, or they can't learn how to do so as adults. There are certain abilities which humans can only properly acquire at a young age (accent free speech and honed motor skills come to mind), so perhaps taming impulsive spending is one of those abilities. It's a theory worth testing.

The concept of discretionary spending provided by constant income is far fetched for anyone who has never had steady job. Such a job is only available in our late teens or early twenties. Predictable cyclic income is necessary to experience the woes of hyperbolic discounting, which in turn provide the only path to overcome it. Parallels with the medical world abound: shielding a child from all diseases results in severe health risks later on in life, like chickenpox. So why don't we vaccinate our children against poor economic behavior? Not literally of course. Although...

Since a child's irresponsible spending habits are inconsequential for a parent, why not give them a low fixed income (say 30$ a month) from the age of six? Start with a dollar a day for the first month, then 2$ every two days during the next month, and so on, gradually weaning them from impulsive spending. I'm sure they'll find something to do with the money: I sure did. In the odd chance that I found a quarter on the ground, I would immediately buy candy; bubble gum was my favorite—it lasted longer. Unfortunately, I was born too soon to benefit from my future insight.

Some parents already do something similar at a later age (calling it an allowance), but such schemes have several shortcomings. For one, it's possibly already too late. In practice, the income is rarely steady: both cutting it off on bad behavior or giving extra income on good behavior—especially when your child is broke— weakens predictability which makes budgeting impossible, thus defeating the entire purpose.