🏅 Top traders

I can't believe I only just discovered this! I'm worried that "Civil Liberties" might become an overly powerful fudge factor later down the line, but I also understand the need for a metric which prevents creating human breeding factories, putting the dying on perpetual life support, and enslaving all of humanity in the name of profit. Maybe it could be determined through a score or range vote by all council members? Or even the general public?

Well, the Model UN unfortunately seems to have stalled because people are too busy. Hopefully we can revive it at some later point. In the meantime, I will just score this based on how I imagine the Model UN game might go. Here is my current thinking:

5 points - Life Expectancy * Civil Liberties * log(GDP per capita) * log(population)
My initial suggestion, and still seems like a reasonable way to combine the metrics available, I think the life expectancy and log income per capita parts are well justified. The civil liberties really depends on what that number even means but I think linear scale makes sense given that the scale of the metric is 0-100. log(population) instead of linear seems reasonable given that I personally would much prefer having 1x population with 2x life expectancy, compared to 2x population with 1x life expectancy.

3 points - Life Expectancy * Civil Liberties * log(GDP per capita) * population
Same as the first one but linear instead of log population. As mentioned in earlier comments, I don't think treating population linearly is a good reflection of my personal values.

2 points - Life Expectancy * log(Civil Liberties) * log(GDP per capita) * population
I don't really know what the scale of civil liberties is, but it is apparently 0 to 100, so I think linear makes more sense than log.

0 point - Population * log(GDP per capita)
Worse than the above

0 points - (Population * Life Expectancy * Civil Liberties) / GDP per Capita
Discussed in the comments, I don't think dividing by GDP per Capita makes sense - higher GDP per Capita means more real goods that people have access to.

@jack - can we get a dedicated group just for your policy question markets? Model UN Game has everything. Model UN Futarchy is empty though?

@MattCWilson Adding them to the futarchy group. The game is still being set up. You can follow the discussion in the discord channel #model-un

@MattCWilson A society where everyone consumes what he produces and no one trades has a GDP of 0 and thus scores infinite points on that scale.

@Milli Precisely?

@MattCWilson People living in caves alone eating raw mushrooms is in my opinion not a society to strive for.

@Milli Neither is that mine?

@MattCWilson But it gets infinite points by the formula you suggested.

@Milli Taking your argument entirely seriously:

Sure, it would. Practically speaking, if things were trending towards cave inhabitation and mushroom consumption, we would likely see drops in population or life expectancy on the road to that extreme. Which would motivate us to steer things away from that endpoint. I’m not concerned that we’ll somehow fail to realize we’re on that path before we arrive there. Plus, no one is beholden to the math over common sense.

I’d love to discuss constructive ideas about to run our futarchy. I’m very not interested in pedantics. I hope you can understand, and if so I look forward to future discussion.

@MattCWilson I didn't construct the example to be pedantic. Any rating function that is broken in theory might also be broken in practice.
Conceptually your rating function punishes spending GDP on anything that is not Population, Life Expectancy or Civil Liberties. A lot of potentially useful things are in that category.
It even punishes increases to those three if they correlate with higher GDP growth (relatively).

@MattCWilson Real world example: USA vs Congo
USA:
Population: 331M, Life expectancy: 77 yr, GDP per capita (USD): 60K
Congo:
Population: 96M, Life expectancy: 60 yr, GDP per capita (USD): 577

Estimates for liberty values (multiplied values from https://worldpopulationreview.com/country-rankings/freedom-index-by-country by 10):
USA: 87, Congo: 56

USA = 331M x 77 x 87/60K = 37M
Congo = 96M x 60 x 56/577 = 559M

That means Congo would easily beat the US and probably all other western countries. Even if the liberty score (only the only non objective metric) is off by a factor of 10 Congo would still win.

@Milli Please refer to my comments elsewhere in the market. You’re narrowly focusing on something here that, imo, is precluding you from a fuller understanding of where I am coming from.

I completely understand the effect of putting GDP/C in the denominator, and have indicated it’s intentional. Pointing out to me the mathematical results of that choice isn’t doing anything to advance a conversation; you seem to be making a category error about what I do and don’t understand.

This is not a real world game; this is a toy model. We currently lack finer statistics that would get closer to the heart of the thing I am trying to measure, which is QALY/effort. Unlike the US and the Congo, the three countries in our game are starting from equivalent starting points, so (for now) sensitivity of the ratio to movement on the numerator or denominator isn’t too impacted due to scaling factors. But as I pointed out earlier elsewhere, some transform may be needed to account for scale.

The point you are making that’s most helpful to discuss, imo, are what things factor into GDP/C that don’t positively impact Pop/LE/CL. I agree (again, earlier and elsewhere) that my primary issue with GDP/C is that it is confounded by all sorts of monetary concerns that blunt its value as a measure. I also think everyone is very anchored in GDP/C as a valuable metric because you come from Earth where that’s a thing important country-running elites say is important. Personally, I care about having a high quality of life for lots of humans, and whether they spend their time painting, playing holodeck chess, or going mushroom caving is not my concern.

@MattCWilson I think the disagreement is broad-based, not narrow. I strongly believe GDP/C is highly correlated with quality of life, both in our game nation and in the real world, and you are saying you think it will be negatively correlated, right? I would be happy to bet on it. Feel free to make a market if we can find a reasonable operationalization (in game, we can look at correlation of e.g. life expectancy and civil liberties, or we can ask the game master to provide additional metrics).

I think the monetary concerns you're worried about are largely solved by using real GDP, which is a much better metric than nominal GDP, but even nominal GDP is a reasonable proxy for useful goods and services.

@jack Thanks. I appreciate the offer to bet. I prefer to find finer metrics and see if that moves us closer to a fruitful, tractable discussion. My concern with a bet is that we'd somehow need to both a) neutrally select policies w/o concern for GDP/C impact (positive or negative) and then b) evaluate whether any long-run gains in the stats we agree upon were more influenced by concurrent GDP/C growth or were orthogonal to it.

@Milli This is a multiple-choice market though? I’m not sure we’re here to pick one and max it, I’m reading this more about trying to mathematically formulate our country’s values?

Also I feel like most decisions we make will involve tradeoffs even when we aren't at the Pareto frontier?

@jack Not being at the pareto frontier means we can make a "decision" which moves us closer to the pareto frontier in all dimensions.
Examples for policies that probably don't have a negative impact on the stated metrics:

• Allowing same sex and unmarried couples to adopt kids

• Legalizing prostitution
  
  In many cases the "decisions" will be a more complex set of actions that make sure no metric is off worse.

@MattCWilson The formula is a way to pick a spot at the pareto frontier to aim for. When we maximize the formula we try to go on a straight line to the point we are aiming for.
It's highly improbable that this is the best way to reach the point we want to get at, if we even identified it correctly.
For example it might be easier to increase one variable first which makes it easier to increase the others.
If we focus on ways to increase all values we'll see which can be easily improved, which give us more resources, and which make the participants actually happy. We also reduce our distance to the pareto frontier more reliably.
If we hit it (which we won't) we can start thinking about evaluating the metrics against each other and doing tradeoffs.

@Milli These are good points, but I'm very skeptical that we can reliably formulate policies (or bundles of policies) that are expected improvements in all metrics, and that also have big enough effects to tackle big problems.

@RiverBellamy I was thinking of this idea too. Although I'm not sure about the linear in population part - empirically most countries would much prefer doubling the quality of life of the same population, vs doubling the population and keeping quality of life constant (cf repugnant conclusion).

Perhaps log(population) or something? Not sure.

@jack the so-called repugnant conclusion involves a decrease in quality of life, not holding it constant. I think most people would prefer higher populations if we don't have to give anything up for it. But ultimately I am voting my personal values, not my model of most people or countries.

@RiverBellamy I don't think the decrease in quality of life vs holding it constant thing is important, or if it is it's a loss aversion thing which isn't the key point I'm making. I'm just saying that if people are offered a choice between society A with 100m people and say 50 average QALYs per person, vs society B with 50m people and 100 average QALYs per person, most people wouldn't value those equally I expect.

The repugnant conclusion doesn't require decreases or changes at all, the original statement is just comparing static populations.

@jack According to Wikipedia, the original formulation of the so-called repugnant conclusion is "“For any perfectly equal population with very high positive welfare, there is a population with very low positive welfare which is better, other things being equal.” So it is very explicitly comparing populations with different levels of welfare.

Anyway, I think I do want to value those two societies the same. I don't find the so-called repugnant conclusion very repugnant.

@RiverBellamy Yes, it is comparing populations with different welfare - that's the same thing I'm saying too.

Many utilitarians accept the repugnant conclusion and I think that makes plenty of sense, but I personally don't take this view with respect to population ethics - all else equal I'd prefer more people living lives with positive welfare, but choosing between societies A and B above I'd much rather have society B (more QALYS and less people). I guess we should poll the members of the futarchy to see what they think.