author: niplav, created: 2021-01-13, modified: 2022-02-23, language: english, status: on hold, importance: 2, confidence: remote
After a discussion with a friend of mine, I thought it would be useful to structure my current thoughts about politics and economics in order to speed up my progress on that topic. Many of these points are only rough ideas, or me collecting unrelated ideas I find good and useful. Therefore, this text might already be out of date.
Dieser Gang der Ereignisse zwingt die Revolution, “alle ihre Kräfte der Zerstörung zu konzentrieren” gegen die Staatsgewalt, zwingt sie, sich nicht die Verbesserung der Staatsmaschinerie, sondern ihre Zerstörung, ihre Vernichtung zur Aufgabe zu machen.
— Vladimir Lenin, “Staat und Revolution” S. 36, 1918
Note: Writing this page is on hold until I have read a textbook on economics.
Before I get to my beliefs about possible object-level improvements, I think that there are some broad strategies that many people would agree on at least in the abstract.
Making politicians resonsible if they don't fulfill their promises.
This is known as “getting to Denmark”, and is clearly not trivial.
Or smth idk
Maximize aggregate preferences of humans currently living in the country (as citizens).
Not (although that would be nice) The One True Good Consequentialism, or Our Best Guess of Morality With Moral Uncertainty, since you can't coordinate around that.
I know. People don't really have stable, coherent preferences.
However, it's the best model I know of.
How deep does this instability & incoherence go?
Arrhenius 2000 gives one answer.
Ask people. Maybe quadratic voting on different metrics.
More just noise.
Has been so far.
Although caveats with regulation/taxation/x-risk
Having no growth kicks people back into zero-sum dynamics
Disaster-preparedness is easier & better (humans are risk-averse)
Though there's the effect of making disasters more costly Shulman 2012, and civilization more complex & therefore fragile?
Not valuable hedonically (Easterlin paradox) But makes paradise engineering more likely & easier
Alignment of markets is >90%.
E.g. is animal suffering in factory farming a public good?
The definition seems kind of hard to pin down exactly.
Aggregate their preference vectors linearly.
Harsanyi's theorem!
Many economists believe diminishing marginal returns, what reasons do they have?
Problem: Hard to compute.
Stifling pareto improvements.
Subsidizing the purchase of new technologies has been a frequent strategy of national governments in relationship to fisheries, with results that have at times been disastrous. The effort to finance the acquisition of a new technology presumes that local fishers will not adopt efficient new technologies without external aid. The conservatism of fishers in regard to the use of new technologies may reflect an awareness that the management of complex resource systems depends on a delicate balance between the technologies in use and the entry and authority rules used to control access and use. If the adaptation of new technologies is accelerated, the relationship between the rules and technologies in use may become seriously unbalanced. This is particularly the case when the rules have come about through long processes of trial and error and fishers do not possess legal powers to devise new rules and get them enforced. A focus on "production costs" alone, rather than on the total of production costs, transaction costs, and enforcement costs, leads to a narrow interpretation of efficiency (North 1986a,b). The rapid introduction of a “more efficient” technology by an outside authority can trigger the very “tragedy of the commons” that the same public officials presume will occur if they do not regulate the use of these fisheries.
— Elinor Ostrom, “Governing the Commons” pp. 259-260, 1990
Market based decision are quite different to those made by governments. If a firm offers a product that nobody wants, it soon leaves business. If government enacts a bad policy, rarely is it repealed. Instead, lip service is given to reform, and when the reform fails, more reform is pursued. Take, for example, farm support policies, which were enacte d to provide financial assistance to struggling farmers during and after the Great Depression. Henry Wallace, the Secretary of Agriculture from 1930 to 1940, argued that the farm subsidies were, “a temporary solution to deal with an emergency.”² But even though the average farm household income now exceeds that of the average US household, and even tough there are now many fewer farmers, farm policies persist.
— F. Bailey Norwood/Jayson L. Lusk, “Compassion, by the Pound” p. 323, 2011
Heuristic against: Disincentivizes production, since more redistribution makes production more like a public good, and public goods are hard to fund.
Heuristic against: People do seem motivated by large amounts of money.
Heuristic in favour: There are strong social incentives for people to want to work.
Heuristic in favour: Shlegeris says so.
Test your damn laws.
At least the regulation is not forever.
Strength of regulation scales with organisational size. (Is this actually good?)
VCG all the things!
Spectrum from "Politicians sometimes look at prediction markets" to "What do you mean by “politicians”? We quadratically vote our values, then the markets optimize for us"/Raikoth.
Possible solution: create a stable independent currency (maybe with crypto, isn't there a Wei currency that tries to achieve this?)
Mild optimization by reducing size/funding for prediction markets.
Resulting problem: this makes manipulation easier.
Don't know how to approach this one, maybe secret cabal of data collectors? Data collectors are not allowed to trade?
Select n random people (something like 10k, or 0.01% of the population), ask them how much they approve of the governing in the last year (or whatever timespan the prediction markets operate on). Try to optimize that measure.
Discovering values, not policies.
Use the same cartel laws for unions as for companies. Also the current implementation for unions is really weird.
See johnswentworth 2018. The whole method of analysing the way in which markets are information processing systems is fascinating. Related: Towards a New Socialism by Cockshott, Hayek's essay on the economic calculation problem, that one paper that claims markets are efficient iff P=NP.
Related question: How much are current economies limited by very hard computational problems, compared to mundane everyday "nobody bothered looking at this yet" problems?
Opportunity cost, huge: around $100b per year.
This is not to say that I believe climate change to be harmless, I don't, and currently believe the recommendations in Marblestone 2019a, Marblestone 2019b and Marblestone 2019c to be reasonable (although geoengineering seems quite dangerous, and should probably only be used as a last measure, if at all).
Regulating companies propagates down higher prices to customers, since what you want to regulate is more ineffective than current practices (i.e. costs more).
There are other reasons why regulating companies might be better (more concentrated leverage points, more fine-grained control).
But there's no free lunch here, people will probably suffer because of this.