We care more about things that are about to happen than things that will happen a long time in the future. This is pretty rational. To give a financial example, I should prefer to have £100 today than £100 in a year’s time, since in the meantime I might be able to invest that £100 to earn some extra money.
A perfectly rational being would discount future events exponentially with a constant rate. If I discount the future at a rate of 110% per year, say, I would be indifferent about receiving £100 now, £110 on 31/12/18, or £121 on 31/12/19, and so on.
However, in 1975, economist George Ainsle published an important paper, Specious reward: a behavioral theory of impulsiveness and impulse control, in which he argued that humans tend to discount the future not exponentially, but hyperbolically. This means that we give near events much much more attention than they deserve. To continue the financial example, in an early study subjects said they would be indifferent between receiving $15 immediately or $30 after 3 months, $60 after 1 year, or $100 after 3 years. These indifferences reflect annual discount rates that declined from 277% to 139% to 63% as delays got longer, in contrast to exponential discounting, where the rate would be constant.
This can easily be seen to be irrational. Would you prefer £50 today or £60 in three months’ time? I bet you would prefer £50 today. How about £50 in 24 months time, or £60 in 27 months time? Most people switch their preference (if you didn’t, perhaps you can see that some other numbers would persuade you to), but this is irrational, since you’re being offered the same choice just two years in advance.
How does this apply to real life? I think it’s a fairly common experience to resolve not to have a biscuit tomorrow morning, but when tomorrow morning comes along, you have the biscuit anyway. Is there any way that we can make resolutions stronger, to increase our chances of following through?
Consider a concert pianist who gives a concert every night after dinner. In advance, she would prefer not to drink alcohol at dinner, and give a good performance. But when dinner rolls along, she drinks, sabotaging her performance. She temporarily prefers drinking on Monday to playing well, but the events on Tuesday and after are far enough away that she still prefers to play well on those nights. So if the pianist could somehow choose at dinner on Monday night not to drink for the rest of the week, the prospect of 7 good performances might be enough to persuade her not to drink on Monday. By bundling all of the decisions into one, she achieves a greater reward than by making 7 individual decisions.
This suggests that the best way to make resolutions is to create personal rules, deciding things once and for all (or with a time limit), which you will follow through with. Indeed, my personal experiences with giving up meat and alcohol (Dry January, etc) seem to suggest that it is easier to say “I will never eat meat” or “I will never drink alcohol” than “I’m going to cut down on coffee”.
But when you make such a rule, how can you make sure that you follow through? If never eating a biscuit with your morning coffee ever again gives you a greater reward than always eating a biscuit with your morning coffee, eating a biscuit “just this once” and thereafter never eating a biscuit gives you a greater reward still (since in the short term you still irrationally prefer a biscuit this morning than abstaining). For fans of game theory, you are essentially playing the repeated prisoners’ dilemma against yourself every time you make this choice.
The trick is to view your decision this morning as the strongest predictor you have of following through with your rule in the future. In effect, you are staking the total future value of the rule against your decision this morning, every morning.
Consider a queue of people which is arbitrarily long. I will ask each person in turn whether they would like £1 for themselves (take), or 25p for everybody (share). I’ll stop at random, after I’ve asked a large number of people. What is the best strategy? You prefer to take for yourself, but you would like everybody after you to share, and would share if you thought that this was necessary and sufficient to persuade the people after you to share. You might reason that the next person to come will use the chain of previous choices to determine their own choice, and that your choice will be most recent and hence most salient. If the next person sees lots of taking, they might judge that it is unlikely for the people after them to take, and hence take themselves. If they see a lot of sharing, they might judge that the people after them are likely to share, and hence share themselves by the same recursive logic.
As such, an effective way to modify your behaviour away from impulses is to create strict rules that you will follow no matter what. Any deviation from following the rule (“sharing”) is likely to cause future versions of you not to follow the rule either, hence devaluing the entire rule. So you have a very strong incentive to follow the rule on every occasion. (Incidentally, this encourages you to think of exceptions in advance and write them into the rule, so that you can look back on your past selves and see yourself as having not broken the rule.) This is bordering on Kant’s categorical imperative: an absolute, unconditional requirement that must be obeyed in all circumstances.
With all this in mind, I have a rule that I would like to follow from now until the end of Lent term (until 16/03/18 inclusive). To prevent a fall-off at the end of the period, I resolve now to punish any breaking of the rule in the last days at the beginning of my next rule of this kind (if this technique is useful, as I believe it will be, I will be using it again!).
By explaining all of this publicly, on New Year’s day, I am leveraging two additional motivational factors. Firstly, if I do not follow the rule, I will lose face. Secondly, New Year’s Day is a focal point and I am likely to attach more seriousness to resolutions made now (people are more successful at giving up smoking on New Year’s Day than on the 23rd June).
I will attend every lecture for courses that I am having supervisions in, with the exception of an extreme family situation (death or hospitalisation of close family). Additionally, I will work for at least 4 hours every day, in two 2 hour sessions, arranged as 25-5-25-10-25-5-25 (on-off etc). The only thing that counts as “work” is actively trying to solve example sheet problems or past paper questions. When working, I will not listen to audio, with the exception of Philip Glass’ “Music with Changing Parts” or white noise. I must not use any device which connects to the internet while working. All prescribed work for the day must be completed before 18:00, although I may do additional work afterwards if I want. Exceptions: if I’m visiting, or being visited, or at a job interview, I only need to do an hour (25-10-25) of work.
If I fail any part of the rule, I will not attend lectures or do any work for the 5 following days. This gives me further incentive to choose the choice in my long term interests whenever I might consider procrastinating. This threat is credible, since the loss of 5 days work is worth less to me than the expected value of a tool this powerful (personal rules) over the course of my life, so I will have a strong incentive to follow through. As described in Schelling’s “Art of Conflict”, the more credible a threat is, the less likely it has to be enforced.
If you’re interested in reading more about this, try Ainsle’s book “Breakdown of Will” or his website.