PESTEL framework

(Importing a draft post by Vipul Naik, with his permission; this can be edited freely.)

A few months ago, while researching scenario analysis, I came across an acronym called PESTEL. PESTEL stands for political, economic, social, technological, environmental, and legal. Broadly,these are the different dimensions we need to consider when evaluating different scenarios (another source I read talked of the STEEP framework, that omitted “legal” from the list, presumably because of its overlap with the political dimension).

Causes typically considered in EA contexts fare quite differently with respect to the dimensions where they involve the most change, and to effectively prioritize between causes, we need to have a general understanding of just how tractable each dimension of change is.

The easy case: the causes where the limiting factor is money

The causes that are simplest to understand are those where a proven approach simply needs to be scaled up, and the main constraint is money. We have a relatively clear idea of how additional money would translate to results on the ground. No significant changes are necessary on any of the PESTEL dimensions (though even these simple interventions may have ripple effects on these other dimensions over the longer term). GiveWell’s current and former top-rated charities, such as GiveDirectly, Against Malaria Foundation, Schistosomiasis Control Initiative, fit this description, at least as they are typically described: there is a proven (or semi-proven) intervention with a roughly determined bang for the buck, so if more money flows to the intervention, more stuff will be accomplished.

In this post, I will ignore the “easy cases”. Of course, to obtain a satisfactory answer to the cause prioritization question on the whole, we would need to compare the easy cases with the harder cases. But I want to begin by simply evaluating the harder cases. A lot has been written (particularly by GiveWell) on the evaluation of easy cases, so it makes sense to focus on the harder cases.

The Importance, Neglectedness, Tractability (INT) framework

In comments on the EA open thread, people have talked about three criteria for cause evaluation:

  • Importance
  • Neglectedness
  • Tractability

While importance, neglectedness, and tractability vary by cause, this post talks more about how the INT framework interacts in general with the PESTEL framework. Can we say anything in general about whether causes that require mostly political change are more likely to be important or neglected than causes that require mostly technological change? Can we say about about whether causes that focus on economic change are more tractable than those that focus on legal change? A general INT X PESTEL analysis can offer good Bayesian priors for our investigations of particular causes.

The political and legal dimensions

I’m clubbing the political and legal together, even though there are important differences: many legal changes can be accomplished through non-political channels (case law and judicial precedent can occur independently of “politics” as it’s usually conceived), and many “political” changes may not operate through any formal changes to the legal framework. Nonetheless, the main reason we care about political change is that politics is the mechanism through which laws are decided.

Some examples of big policy changes that have been considered high-potential from the EA perspective, grouped by category:

  • Libertarian/globalization: Migration policy (the extreme being open borders), trade policy (the extreme being free trade)
  • Libertarian/basic economic freedom: Getting governments around the world to the level of economic freedom seen in relatively economically free countries, such as the US and Canada.
  • Libertarian/libertine: Allowing markets in products that might be high-value to trade, such as human organs and tissues. Also, allowing trade that may not be intrinsically high-value, but where the crackdown is (believed to be) very socially and economically destructive. Examples here might be drug legalization and allowing prostitution.
  • Intellectual property: Various directions of IP reform have been considered, generally in the direction of making copyright, patent, and trademark law less stringent. The main social value justification is with respect to patent law in drug discovery, though patents might also have an impact on innovation in the software world.

There are “spending”-type policy changes, that mainly involve influencing how governments allocate money:

  • Foreign aid: The idea is to lobby governments of better-off countries to spend more on foreign aid, specifically on the most high-value components of foreign aid.
  • Technological investment: Lobbying governments to spend more (or less) on particular forms of technology.
  • Intranational stuff: Most of this isn’t promising from the EA perspective directly, but it could be based on flow-through effects. For instance, perhaps socialized universal health care and/or a radical free market in health care could deliver huge social value, both directly and through the increased wealth and technological progress it enables. Perhaps a more robust social safety net could also have strong flow-through effects.