Moral Economics Concepts
On the shoulders of: Altruistic Arbitrage, Certificates of Impact, Cognitive Load and Effective Donation, Coordination Problems for Donations, Donation Swapping, International Donation Trade, Direct Funding for EAs and especially Moral Trade.
Special thanks to Ryan Carey, Ben Hoskin and Leo Arruda for ideas and corrections.
In these writings we propose the creation of a subfield of knowledge, Moral Economics, and provide in broad strokes its characteristics. We assemble previous writings that can be considered prospective subfields. We also discuss several concepts that have been developed in recent years that are instrumentally useful for Moral Economic thinking, and propose several new ones that could also bear fruit if the study of this area develops further in the future.
- Introducing Moral Economics
- Examples of Moral Economics Concepts
- Branches Within Moral Economics
- Moving Moral Economics Forward
- Direct Funding Between EAs
- Agential Identity in Moral Economics
Examples of Moral Economics Concepts
Altruistic Arbitrage: Analogous to the concept of Arbitrage, the concept serves to facilitate reasoning about opportunities for trade that are morally profitable.
Certificates of Impact: These consist of certificates that are bought after the fact for a past altruistic action or a fraction of one, which later can be traded or sold. An attempt to make the Moral Market for actors with altruistic inclinations be explicit and numerical, by creating a form of floating currency that permits exchange in space and time between the two markets – notice that inter-temporal backwards trade was not available for altruistic actions before this idea, you could not today cooperate and trade with past altruist action.
Intertemporal Coordination Problems for Donations: To donate maximally impactfully to a fundraiser with a deadline, it seems prima facie that one should wait until the last minute to guarantee that someone else donates instead of you, since your goal is that resources go to that event, not that your resources go to that event. This way you make sure your moral values are satisfied, and keep the resources to satisfy even more of your values. This is different from buying an apple for yourself, where you anticipate the only way of getting it is if your resources purchase it. In the altruistic case, as long as someone pays for the valuable thing, your value is satisfied.
This however poses a tragedy of the commons type incentive, where each individual will be better off (morally) by holding on to their resources for longer than would be desirable from the group’s perspective. It becomes more difficult to predict how much will be raised by the fundraising. This concept is related to apathy bias or the bystander effect in behavioral economics. This does not happen in classical economic theory because individuals are not indifferent over who gets the cake, the stock, or the wage as long as someone gets it.
Direct Funding for EAs: A direct linkage between an individual who wishes to have their values promoted and another who attempts to promote these values in exchange for money.
This has been attempted within the Effective Altruist community by creating Gratipay and Patreon as donation portals for EAs. Although it is much cheaper per worker hour than donating to institutions – back of the envelope calculation during the EA Think Thank says between 2x to 3x cheaper within a country and up to 5x depending on currency exchange – this form of trade doesn’t have a large fraction of EA donation resources yet. Direct funding also permits funding for causes that are unknown or neglected within the EA community. If one single donor would like to sponsor an EA to research how Emulation Economics would disrupt the economy, and a single researcher is interested in doing this, direct donations between the two could guarantee this is done, whereas institutions or certificates would not. Direct Funding will be further discussed in a later post in the series.
Donation Swapping: Frequently there are tax or legal reasons for two donors with different interests and values to swap their donation targets. Because the market does not have absolute liquidity and zero viscosity, it is frequently advantageous for individuals to A and B to reach an agreement on how much to donate to causes 1 and 2, and then A who values 1 donates to 2 and B who values 2 donates to 1, because some market inefficiency makes that morally cheaper for them. This is a form of regulatory arbitrage.
Moral Trade: …as described by Toby Ord: If two parties have different moral views, then there is another type of trade that is possible: they can exchange goods or services such that both parties feel that the world is a better place, or that their moral obligations are better satisfied. We can call this moral trade. For instance, a person concerned with animal welfare might offer to donate to Oxfam if their friend, who is concerned with global poverty, agrees to abstain from eating meat for a year.
Moral trade is the philosophical underpinning on top of which moral economics can be built. If there was no moral trade, there would be substantially less reason for a science of moral economics. Moral Trade can happen almost irrespective of moral stance, which we can dub the Moral Trade Orthogonality Thesis. For any ranking of actions according to their moral permissibility or desirability, whether measured in (a)virtue,(b)ranking seen from a veil of ignorance, (c)utilons, (d)hedons, (e)sacredness or (f)accordance with universal reason, there might be ways in which trade could facilitate the attainment of moral good from the perspective of an individual who endorses that ranking scheme. The divergence in moral views between individuals greatly facilitates this trend.
Tragedy of Commonsense Morality: In Moral Tribes, Joshua Greene distinguishes the types of problems that are partially solved by our moral instincts and agreements (Me versus Us) from those which our moral instincts and system 1 morality are not equipped to solve (Us versus Them). He dubs this latter type the Tragedy of Commonsense morality. Throughout the book he makes a mixed empirical and philosophical case for utilizing system 2, manual mode morality when making decisions of the Us versus Them type, and trusting our senses more for Me versus Us type problems. He further advocates the usage of utilitarianism as a currency when solving problems of the Us versus Them type, though for space reasons I will not describe the many reasons for which he considers this desirable. It is worth noting that the book ends with a link to the Givewell website, indicating Greene’s alignment with the use of Moral Trade as strategy to generate morally desirable outcomes.
Acausal Trade: Acausal trade is trade which doesn’t involve (relevant) causal contact between the trading parties. If I assume there is an altruist in India who would prefer that I don’t eat beef and she assumes there is an altruist in the US who would prefer that she alleviates the suffering of five poor people in her city, we can make a Moral Acausal Trade, creating a world that is best for both of us without ever interacting: I eat less beef and she helps those five people in a way I couldn’t.
Moral Economics Branches
Many other concepts are of service to reason about Moral Economics. To accelerate the development of the field, permit me to go ahead and propose the creation of subfields or branches of Moral Economics.
Economics of Counterfactuals
I propose the creation of a sub-branch of moral economics which investigates the differences between moral economics and regular economics that arise due to their different treatment of counterfactuals – also called replaceability effects within economics, also called opportunity costs in some cases. Differently from rival goods like apples in economics, where my having an apple precludes you having that apple, frequently in Moral Economics what matters is that someone gets an apple, and both of us accrue value out of that individual receiving it. Let me explore below some concepts that are specifically related to counterfactuals and how they behave differently in Moral Economics:
- Labor replaceability: See MacAskill and Todds’ theses. An amoral doctor may care about acquiring resources (money) and that she saves patients. A moral doctor frequently will instead care changes in use of resources from what would happen otherwise. They would prefer to save more lives than if someone else was a doctor, and may also divert a fraction of their resources (say 50% of their salaries) towards saving more lives indirectly through charity or astronomical waste interventions.
- Ripple effect risk analysis: Although companies and States have incentives to do risk analysis of ripple effects in the future, altruists have a much stronger reason due to a few factors:
- 1)Most of economics externalities are not externalities from an altruist perspective, since they generate, remove or preempt values from agents that are within the circle of altruistic concern
- 2)There are vastly more individuals who are not you than who are, and vastly many more potential individuals in the future than now (10^30 is a conservative estimate if silicon based moral agents are impossible, 10^50 if possible)
- 3)Power law distributions are more common in large scale economies than in small trades or partnerships.
Some of the work conducted by Sandberg, Hillebrand, Ord, Armstrong, Douglas, Bostrom and more recently the Open Philanthropy Project fall within this scope. Ripple effects are particularly important for altruists because altruists are more impartial about how far in the future the consequences of their actions generate benefit. Risk analysis is important because the same is true for negative consequences.
- Donation overdetermination: Many interesting game theoretic questions arise when donations are overdetermined. There may be a Schelling point of how much to donate when an individual anticipates other individuals donating, see also Non-Indexical Decision Theory, in the next post.
- Unilateralist Curse: A complex failure mode that arises out of several agents who can each take a unilateral decision, even when all the agents have value aligned collaborative intentions.
- Haste Consideration: To consider what would have happened if you had done otherwise, it is very important to take in consideration the causal impact of acting sooner or later. Matt Wage and I wrote two very similar texts advocating a 2 year haste consideration at different times.
Another valuable sub-branch of Moral Economics is the study of models of economic agents which are useful to reason about economic situations and states in which it matters to what extent the economic agents should anticipate their behavior being similar to that of other economic agents, or to what extent they should assume those agents are themselves or not. Two distinct forms of boundaries are relevant, the boundary for agents that behave like me in similar contexts, and the boundary for agents that are me. Since this is a complex sub-topic and there is much to be said, I’ll take Leo Arruda’s suggestion and save it for the next post.
Moral Economics as a Response to Coordination Problems
Some readers might be tempted to see moral economics as a solution to coordination problems. In some cases, knowing what a moral agent should do indeed solves or helps coordination problems to be solved. When instead of MaxSelf you are MaxSum, many games that were negative zero sum or zero sum become positive, that is, the structure of incentives which created the game is more favorable to moral players than to selfish players. This isn’t always the case, in very few cases, the invisible hand actually defeats morality and altruism even when judged by the total sum of good. These cases, however, are infrequent.
A more frequent and widely studied problem when interactions between more selfish and less selfish agents end up being less beneficial for the altruistic agent, who would be better off in a group that is more altruistic, that punishes cheaters, that is more moral etc… These cases are frequently studied in evolutionary biology (a nice summary is Novak 2013), and it is worth mentioning that the benefit accrued by a selfish agent is not considered benefit to the altruistic ones within biology, but frequently should in Moral Economics when consequentialist altruistic agents are involved.
For instance, today I stunned a bike shop owner when telling him I have no intention of making sure my bike seat isn’t stolen, and prefer to purchase new seats sometimes instead of locking that one, because I consider the benefit accrued by the thieve and the counterfactually would be stolen other person to also be accrued by me (at some 0.5<x<1.0 rate), making the case for seat protection too weak to matter. At our effective altruists house in Berkeley I have also observed other people acting in similar ways towards us, their housemates.