To understand New Year resolutions, we need to understand time-inconsistent decisions, commitment devices, and why those devices fail.
Let's think like quants and build a simple model: to diet or not to diet, that is the question.
To answer that question, people weigh the value of having a good lazy time (eating, not exercising), call it $v_0$, versus the value of being fit, call it $v_1$, with the understanding that you only get $v_1$ after a delay $t_1$ to get into shape, no matter when you start.
Resolutioners follow a variation of the Kate Moss rule: nothing tastes as good as being fit feels, in other words, $v_1 > v_0$. The problem is the $t_1$ delay, because people discount value when it's delayed. People like instant gratification and delayed sacrifice, whereas exercise and diet are instant sacrifice and delayed gratification.
In fact, we know from many experiments that people are willing to delay gratification later, say starting at $t_L$, just not now:
a) Given a choice between $\langle v_0$ now $\rangle$ and $\langle v_1$ with a delay $t_1 \rangle$, they choose $v_0$: they eat pizza and binge-watch 'Dracula' on Netflix instead of exercising and eating high-protein, low energy (carbs + fat) foods.
b) But, given a choice between $\langle v_0$ at time $t_L \rangle$ and $\langle v_1$ at time $t_L + t_1 \rangle $ (i.e. the same choice, but with a delay of $t_L$), they choose $v_1$: if asked on Nov 15 whether they're willing to join a gym and start eating more healthy food on New Year's day, rather than spend the next year binge-watching Netflix and eating pizza, they choose the gym and healthy food.
If we use the standard exponential discounting of economics and normative decision-making (also finance, where it actually comes from), with some rate $r$, this behavior can't happen:
$ v_0 > v_1 \exp(-r \, t_1)$ (the first choice)
implies
$ v_0 \exp(-r \, t_L) > v_1 \exp(-r \, (t_L + t_1))$ (the opposite of the second choice),
since the second inequality is just the first multiplied on both sides by $\exp(-r \, t_L)$, which is the discounting equivalent to a delay of $t_L$.
There's a different type of discounting, hyperbolic discounting, that captures these effects, but by its own nature leads to temporally inconsistent-decisions, so it's a bad guide for decision-making.*
So, people don't follow the rationality of economists; anyone surprised? No?! Right. What does that have to do with New Year Day, an arbitrary date? Simple: it's all about commitment: a tool to manage one's own irrationality. It's an arbitrary date from a sidereal point of view, but not from a social point of view.** People celebrate, there's some talk of renewal, and therefore it becomes a focal point for the decision. It acts as a commitment device, especially if the resolution is made public to one's friends.
Why does it fail?
I believe three main reasons, based on occasional observation of others:
1 - Bad information leads to bad outcomes early on. People get bad information and hurt themselves in the gym or eat food that leads to significant hunger so, rationally (ironic, isn't it?), they stop exercise and diet. Note that this really is rational in the strict sense, because what they realize is that $v_1 < v_0$.
The problem is that their $v_1$ is low due to bad information about diet and exercise, which unfortunately is rampant. If they had good information they would get a high $v_1$ and stick with the program, but alas where it comes to fitness and diet the worst disinformation around tends to have the best marketing.
2 - The arbitrariness of the date and the fact that it's a psychological or social trick is known to the resolutioners themselves, so they eventually de-commit by rationalizing away any value the arbitrary date might bring. This is why gym people tell friends talking about their upcoming resolutions to start immediately (basically this is pointing out the time-inconsistency illustrated by the choices above and the obvious solution of sticking with one of the decisions, preferably the second one.)
3 - People revert to type. Sometimes people realize that their expectation of the value of being fit (the $v_1$) was based on other people's preferences and media narratives; that they really don't value health and fitness as much as they thought they did and that life is too short to give up pizza and ice cream.
So, what is to be done if one has friends who make these resolutions? Based on my totally anecdotal unquantified analysis in the three preceding points, there are two main interventions:
First, get them good information. For exercise I recommend John Little and Dr. Doug McGuff's book Body By Science as the foundation for understanding exercise and Dr. Brett Osborn's book Get Serious as an important complement for people over 35.
Diet is a minefield, so I'll just say what worked for me: intermittent fasting on a high protein-to-energy ratio eating. It worked because it didn't rely on self-control or discipline; it relied on never being hungry. I find that Mangan150 and TedNaiman on Twitter are good sources of information.
(A side note here on diet advice from athletes and personal trainers: a lot of people who are very fit passing their physique off as knowledge and some people with actual formal education but who never struggled with weight issues behave as never-smokers telling smokers who want to quit smoking to "just don't smoke!"
If only dealing with one's temporal inconsistencies were that simple... usually it's easy to identify these unhelpful people, because they focus on counting calories and "energy deficit." Getting an energy deficit is like not-smoking, an outcome, advice no more practical than "just don't smoke"; counting calories is terrible advice as I've shown before.)Second, encourage them by managing expectations. I agree with the human fountain of expletives, Alan Roberts: people out of shape are trying to improve themselves, they're often ill-at-ease in a gym, and they don't need fitter people making them feel bad about themselves. Also, make sure they moderate their expectations and enjoy their newfound fitness: they're not going to compete in Ninja Warrior by July, but they'll be able to walk the trails in Castle Rock Park without getting a cardiac event.
Compliance will be rewarded.
- - -
* Hyperbolic discounting is a good description of how people make decisions in reality, so it's a good tool to analyze other people's decisions; however, if you're trying to make the best decisions yourself, hyperbolic discounting leads to time-inconsistent choices (as the ones above); as the obesity and unfitness epidemic shows, time-inconsistent choices have bad consequences, so when possible one should use exponential discounting which by definition forces time-consistent choices.
** It's actually about being close to the Winter Solstice, the shortest day of the year, and a superstitious celebration/offering to ensure the days start getting longer, but let's not quibble.
