Executive Summary

• Donors often navigate difficult choices among projects that have different outcomes and promote different values. Standard cost-effectiveness analyses allow us to see how changing assumptions (e.g., regarding moral weights or probabilities of success) can change our evaluations of individual interventions. However, they are less apt for dealing with uncertainty about these assumptions or navigating disagreements among stakeholders.
• Our Moral Parliament Tool allows users to navigate uncertainty about key assumptions that affect our assessment of interventions. It does so by representing diverse perspectives as delegates in a democratic decision-making process about how to distribute resources to various causes or projects.
• What’s new: We have added functionality to the original Parliament Tool, which allows it to be adapted to novel resource allocation problems simply by editing a spreadsheet.
• The tool is not limited to evaluating broad moral philosophies and cause areas. It can be applied to many kinds of worldviews (e.g., ‘EA should be a mass movement’, ‘insects matter’) and more concrete allocation decisions. It now handles questions like:
  • "I've decided to give to animal welfare—but how much to chickens versus fish versus shrimp versus insects?"
  • "I want to support global health, but I’m unsure about how much to weigh saving lives vs. increasing income. Which charities should I prioritize?"
  • "My family disagrees about healthcare priorities—how do we find an allocation we can all accept?"
• We have built four ready-to-use parliaments that demonstrate the versatility and utility of the tool:
  • Bioethics Parliament: Allocate healthcare resources in light of diverse bioethical views (e.g., Rule of Rescue, Utilitarian, etc.). [post / parliament]
  • GiveWell Parliament: Evaluate global health charities using different moral weights, including animal welfare considerations that GiveWell does not currently include [post / parliament]
  • Animal Parliament: Set funding priorities across different kinds of farmed animals (chickens, fish, shrimp, and insects) given uncertainty about sentience, welfare ranges, and project success rates [post / parliament]
  • Movement Building Parliament: navigate strategic decisions about growing effective altruism (though this is more speculative and serves primarily as a template for users to modify) [post / parliament]

• We provide instructions for how users can adapt the Parliament for their own resource allocation problems.

   

Introduction

Our Moral Parliament Tool helps donors make better decisions when facing uncertainty—whether navigating conflicting intuitions or trying to reach an agreement with others who have different values. It models viewpoints as delegates in a parliament who are deliberating about how to allocate a set of resources.^[1] It offers an efficient way to:

• See the implications of different perspectives
• Gain insight into how different deliberative procedures might resolve disagreements
• Make principled choices in the face of genuine uncertainty

We originally developed the Tool to model the deliberations of delegates with diverse moral worldviews (e.g., Kantianism, utilitarianism) regarding the distribution of money across different philanthropic cause areas (e.g., malaria charities, AI safety organizations). In this series, we show how the tool can be adapted to other resource aggregation problems that differ in terms of both project types (e.g., real or hypothetical, general cause area or specific project) and the values that are used to assess relevant options. We provide instructions for how users can configure the Moral Parliament Tool for their own purposes.

Why should you consider using the Moral Parliament Tool to assist in your decision-making? Many resource allocation problems involve navigating different kinds of uncertainty simultaneously:

• Empirical uncertainty: What are the chances that the projects we’re considering will succeed in bringing about their desired results? What is the expected magnitude of the project’s effects?
• Normative uncertainty: How much should we value the very different results achieved by different projects?
• Metanormative uncertainties: When delegates disagree, how should a collective decision be reached? When I partially believe in several different worldviews, what should I do, all things considered?

The Moral Parliament Tool provides a user-friendly interface for navigating all three types of uncertainty together.  While many of these functions can be performed by standard cost-effectiveness assessments, one of the Tool’s unique functions is in modeling a diverse set of worldviews and assumptions simultaneously. It allows you to easily test how changing the answers to any of these questions, or adding participants with different sets of answers, affects the value of projects and the overall results of deliberation. It offers multiple ways of aggregating across disagreements, revealing how the decision-making process itself shapes outcomes.

Making a Moral Parliament

Components

The Moral Parliament tool has three components:

1. Specifications of delegates’ worldviews
2. Worldviews’ judgments about projects
3. Methods for allocating resources in light of worldview judgments

A worldview encompasses a broad set of commitments about who matters, what matters, and how we should act. These commitments are represented in the Parliament Tool by numerical judgments on a set of normative dimensions, such as “how much consideration should we give to humans vs. chickens?” or “how much do you care about saving vs. improving lives?”. You can populate a parliament with delegates that represent the worldviews you’re considering.

Delegates in the moral parliament deliberate over a set of candidate projects. We use “project” as a label for any option that delegates may consider devoting resources to, where this may be a specific philanthropic investment (e.g., Against Malaria Foundation) or more general cause areas or strategies (e.g., Global Health and Development).

The Tool helps us derive how each delegate would assess each project by evaluating how well the project promotes the things that the worldview values. We take the proportion of the project’s effects that promote some dimension, multiplied by the importance the worldview places on that dimension, and by the overall magnitude of the project’s effects.

Finally, these delegates will choose a project (or allocation of resources across projects) by using an allocation method. The Moral Parliament Tool includes a variety of different procedures, including:

1. Voting: approval, Borda, ranked choice
2. Bargaining: Nash bargaining, proportional allocation (Moral Marketplace)
3. Social choice functions: maximin, Maximize Expected Choiceworthiness, My Favorite Theory

Specifying the problem space

To design a parliamentary deliberation, we need to characterize candidate projects and worldviews:

Step 1: Identify a set of normative dimensions that are relevant to the decision problem. A helpful heuristic is to ask: Which values serve as cruxes when evaluating these projects? What value differences could explain people’s different judgments in this area?^[2] 

Step 2: Specify a set of candidate projects. To set the normative dimension scores, ask what proportion of the project’s benefits goes to each category of that dimension. These should add up to 1. Then, set the Scale by asking: if you fully valued everything the project did, how much value would it bring about? To ignore differences in scale, set equal scales for all projects.

Step 3: Specify a set of worldviews. Idiosyncratic worldviews can be designed for each delegate. Alternatively, populate a list of worldviews that are common in the problem space you are modeling.

Step 4: Populate the Parliament with the correct proportions of delegates to represent the target population. The number of delegates representing a worldview may reflect:

• An individual’s credences in different perspectives
• Actual representation in a group decision

• How much weight should be given to different stakeholders

   

Getting started

Start with one of the five ready-to-use parliaments:

• Original Moral Parliament Tool: Resource allocation across philanthropic cause areas in light of common ethical worldviews
• Bioethics Parliament: Healthcare resource allocation with diverse bioethical values
• GiveWell Parliament: Global health charities with customizable moral weights
• Animal Parliament: Funding priorities across farmed animal species
• Movement Building Parliament: Strategic decisions about growing effective altruism

The Moral Parliament Tool is fully customizable. Minor edits — changing the scores that a project or worldview assigned to a normative dimension — can be made directly on the site. Larger edits, allowing entirely novel worldviews and projects, can be done via a spreadsheet method. Instructions can be found here.

If you would like our help in constructing a parliament tailored to your specific needs, please contact us.

Acknowledgments

The Moral Parliament Tool is a project of the Worldview Investigation Team at Rethink Priorities. Arvo Muñoz Morán and Derek Shiller developed the tool; Hayley Clatterbuck created the particular parliaments in this sequence. We’d like to thank David Moss and Urszula Zarosa for helpful feedback. If you like our work, please consider subscribing to our newsletter. You can explore our completed public work here.

1. ^{^}

    Delegates may represent different individuals or different perspectives within a morally uncertain individual.

2. ^{^}

    Many real-world disagreements stem from disagreements about empirical matters rather than, or in addition to, differences in values. The Moral Parliament Tool can capture some empirical disagreements about how likely projects are to be successful and the magnitude of their impacts. Empirical judgments about how well each project promotes each worldview’s values are made exogenously.