This is a working list of areas at the internsection of economics and technology that I would like to see more work in. If you work in this space, or if you have ideas, or even if you just want to chat over a coffee (or a zoom call), hit me up!

  • Automated policy design: Policy makers often find that the policies they have design and legislated on, have unintended consequences, often because the people that the policies are designed to impact in a certain way, do not react the way that the policy designers expected them to react. There are numerous examples of well intended policies have unintended consequences. The feakonomics podcast recently had a great episode on just this topic. A prime examples of such a policy is the ADA or the Americans With Disabilities Act, a well intentioned piece of legislation passed during the George Bush I administration, that among other provisions, made it illegal for firms to discriminate on the basis of disability status while hiring or firing The unintended consequence, which was not anticipated by legislators, was that companies began to hire fewer people with disabilities because they are worried about getitng sued and dealing with litigation - the subject of a famous 2001 paper by Acemoglu and Angrist. At the core of this problem is predicting how human beings behave, and by extension, how organizations with certain incentives, would behave when exposued to a certain stimulus. Human beings do this instinctinvely - to some extent - when interacting with others, we instincively predict what the other person might be thinking or perceiving an interaction. What I’m proposing is doing this at large scale and thereby predict how a policy will play out. Do we have the technology to accurately predict what the effect of a particular policy would be, before the policy is implemented? This is the core question. I think we probably do. Solving this would be gamechanging. Not just for goverment policies, but for policies that are crated by any organization, including firms, NGOs, municipalities, etc.

  • Automated incentives detection: What is the incentive structure of any given economic actor: an individual, an organization or a system of interconnected set of entities? This is an important question for anyone looking to understand an economic system at any scale: a good example to use is the healthcare system in the US: the economic actors are: healthcare providers (hospitals, clinics, labs), doctors, nurses, patients, insurance companies, pharmacies, pharmacy benefits managers. Each of these economic actors interacts with more or more other actors in this list. Another example is the education system: school board, teachers, students, parents, board of education, etc. Understanding the economic incentives of each of these entities is non trivial. A person inside the system can describe it to you, but figuring it out on your own is very hard. One idea is to simply have a wikipedia-style crowd sourced database where people catalog the incentives of each economic.

  • Automated detection of externalities, positive and negative: An externality is simply the effect of an economic actor actions on other economic actors, either harmful or beneficial. Second hand smoke is a negative externality for people standing near a cigarette smoker – people who never chose to smoke are hamed by those who did chose to smoke. Companies that poluute the environment create a negative externality for the rest of the world via their pollution. Moderna created a positive externality for the travel industry when it created the Covid vaccine. Externalities, both positive and negative, are fascinating, for a few reasons. First, they are often hidden. And second, and much more importantly, both types of extenalities are indicative of often massive inefficiencies. Negative externalities often indicate that some economic actors are benefitting at the expense of others - sometimes this points to economic injustice against those affected by the extenality. Positive extenalities, like the Moderna example above, indicate that the entity adding value is not capturing it in the same proportion, which means that incentives are not well structured to add value (in my book, when an economic actor can add a lot of value by virtue of their unique skill, and can also capture most of that value, then such an actor has a strong incentive to add that value). For these reasons, I think it would be great to be able to detect both types of externalities – in a sort of automated manner, either using crowdsourcing, or perhaps using some type of AI.