Response to: Response to: Curving for Justice

Dear friends,

The unparalleled (and unhinged) Wooden Womb wrote a response to my post about intergenerational justice in which he made four main points about where my framework for a way of determining a fair distribution of resources accross generations may be lacking. I want to thank zhim for his excellent piece which certainly merits a reply, congratulate zhim for zher excellent taste in clothes and commend zhim for zher fantastic use of Science. I will go through it point by point, but first I want to make a few clarifying statements.

WooWo suggested that the following characterised my position:
"Let’s graph expressed preference for income distribution across the course of their lives vs. what people actually get over the course of their lives."

Actually, though this is merely an elaboration not a contradiction, the model is designed not only to give us a picture of distributing resources over a person's lifetime, it is also supposed to give us an idea of how resources should be shared among different generations.

Obviously assuming either zero growth or that growth has limited effects on people's standard of living (a wildly innacurate assumption) these two things will be the same as each other. My reasons for excluding growth are firstly to make this simple enough to start from, and secondly that I think its importance is overemphasised a lot of the time. The number of houses per person hasn't changed that much, and x-BOXes aren't all that.

That was the clarifying staments. Now, let me address his main points:

1: People Don't Know What Is Best For Them

I must counter this argument by stating that here WooWo is absolutely right. People's lives are far too important to let them be the ones making decisions about them. However, most of the time terrible as they are they are still the best people for the job.
My argument was not supposed to be about giving people complete power over their lifetime income distributions, however. I wanted to make the point that if you did ask them what they wanted, they would give you an answer differing from what we have, but I don't think that's where we should end. I wanted to challenge people and policymakers to go to a different level of abstraction, to ask not what they do want, but what they would want, under some yet to be refined conditions. These conditions would mean that people try to consider what their distributions would be like from a neutral perspective so their current situation doesn't bias it.
When I first tried to think of conditions that would fulfil this role, I thought of making it so the person has to imagine they will transported into a random year in their life. This would not work though, as people would load all the resources into their final years because they might miss a lot of income if they put any into the first phase of their adult lives. So this first potential condition fails because it leads to perverse incentives.
A better condition which I have not found a problem with yet is to ask people to imagine that they will live through all of the phases of the life they designed but in a random order. This would hopefully remove biases people have towards resources now as opposed to resources later.
What the exact conditions would be are academic however. What is important is that we imagine such conditions could exist, and imagine what people would choose. That we we can arrive at distributions less influenced by people's current place in time.
Of course it doesn't completely work in practice. It is impossible to get people to accurately describe what they would do in hypothetical situations that don't, or possibly even can't, exist. That's why they're called thought experiments: if they were eminently practical they would just be experiments. Thought experiments are useful for framing a problem more clearly, and bringing out our intuitions, and pointing out inconsistencies, and other conceptual hi-jinks.
Maybe one day a mad scientist will come with a way of actually putting people in test conditions ans in a situation where they have to make these sorty of choices. It would almost certain be immoral, and fascinating.

I feel I have strayed a little off WooWo's main point here though. While all I have said above does help bring out the difficulties and also advantages of encouraging people to think critically about what they think the best distribution of resources over lifetimes would be for them individually, or for society more generally. As for the core objection about people not knowing what is best for them, this is something which in practice in society we work around. Most people have control over a lot of what they do with their lives. Social institutions nudge us towards things which are deemed good and vice versa, and there are blanket laws for and against some things, but a lot of important decisions are left to people themselves. That's why I don't think we can get a lot of mileage out of the people don't know sh*t objection. Besides, people vote and so collectively they exercise a great deal of power over their lives.
Voting is an interesting one to lookout. Voting gives us information about aggregated preferences, similar to my suggestion of how to use information from these curves to inform public policy. Collecting and aggregating these curves will give us helpful information about people's preferences which policymakers can use with discretion, informing but not dictating their decision. As an example of this works in practice, see Damian McBride's excellent piece about how budgets get written. Impact assessment is an important part of the process, so governments can see how their policies will distribute wealth, income, or other resources between different groups, including between generations: "what’s the distributional impact?".

So much for point 1. I have responded to WooWo by accepting the limitations of this approach while touting its usefulness as a policy starting point and pointing out that aggregated preference information (like voting) is already a large and more or less accepted part of the policymaking process, which this method is no departure from.

2: Temporal inconsistency.

I believe my response to point 1 also covers this. Abstraction provides a limited work-around for this problem, as does framing the question in such a say to try and provoke unbiased anwers. Impossible to perfect but still useful.

3: Satisfaction.

Actually how much people enjoy their jobs was not considered in my model, infact jobs are presented as a resource to be distributed as opposed to a burden. Other than the very worst jobs, by most physical and mental health measures working is pretty much always better than not working. As for the distribution of job satisfaction across ages, I imagine it is one of the main things that we imagine should by slowly going up, in order to have maximally worthwhile lives for people. A little struggle makes the reward all the sweeter. Indeed, Spinoza writes (and I think he's got something there) that happiness is not a state in itself, but a sensation we feel when we are transitioning from a worse to a better state.
When I say resources, we should usually think of resources considered individually. This includes both money and free time. Therefore we should imagine people as they get older getting a little more income and a little more free time. Some peope do love work, but with the volunteering sector there is always a chance for people to turn their free time into rewarding work at no cost.
I will write a blog post at a later point about trade-offs between work and leisure and how it is more complicated than you might think. The headline finding, as expressed brilliantly in Willets' The Pinch, is that earning more per hour often leads people to work more hours, not less.

4. Motivation.

In response to this last point I will draw what I think is a rather apt analogy between advancing in age and advancing in heirarchy, which often amount to the same thing.
In the '50s, bosses tended to earn something like like ten times what the average worker earned. Probably. The exact number is not really important. In the '10s it is more like one hundred times, among medium to large sized companies anyway. If that sounds unrealistic, lets remember that that just means bosses earning about £2 million, to the median person's £20,000. I think that is fairly representative, though the point will stand anyway. Surely people had sufficient motivation to try to advance their careers already in the '50s? Certainly capitalism did not fall apart due to lack of aspiration.
A similar story can be told with average wages for people of various ages. People are still going to work hard to increase their incomes to 150% of what it is now over the next few decades, they won't only budge for some figure like 400%, which is more what the differential between young and old workers is like at the moment. That's based on an estimate of £10,000 for 16-25 year olds (probably too high considering the 20% youth unemployment level, relative poverty endured by students, proliferation of unpaid internships, lower minimum wage for 16-21 year olds, prevalence of temporary and part time work, and non-participation in the labour force) and £40,000 for a 45-55 year old, which will probably be when most people's earnings peak. A range from £20,000 to £30,000 (to use these figures) or from £16,000 to £24,000 (to straddle my median income estimate) will still give people plenty of motivation to challenge themselves, while reducing the amount of lifetime suffering due to financial hardship for everyone.


That's it. I realise I have gone on much too long, but due to the high quality of WooWo's response I wanted to give an answer in some detail. Hopefully this has resolved more ambiguities than it has introduced and corrected more mistakes than it has made. I still think this is a helpful way to start thinking about distribution of resources between generations. I also commend WooWo's excellent conclusion about the robots and stuff. It's pretty cool (though it doesn't necessarily answer the question given the assumption that scarcity will always exist in some form) and I don't have anything nearly so clever to end on.

I tried to make a pun once about a dotty aristocrat who mistakenly thinks porpoise is an abbreviation of poor person, but I couldn't formulate it in a way which wasn't horrifically offensive to aristocrats, non-aristocrats and people with dots on them. It fell flat and I retreated to my room, feeling ashamed and like I had betrayed my moral convictions, wishing there was a God so there could be someone to ask for forgiveness for my slight which had wronged millions of people who I could never apologise to and who could never forgive me. I carry that mark on my soul still. This life is an endless procession of torments, failures and humiliations. The robots comment, that was better.