My brother and I are about the same age and when we were younger, we often got in trouble together. We weren’t delinquents. I’m talking about kid-trouble like running in the house when we were told not to and we bumped into something and broke it. When our parents asked what happened we would never tell on each other. We never agreed beforehand that we weren’t going to tell on each other; it’s just something we somehow agreed not to do. In retrospect, I think we subconsciously reasoned that if we were both at fault then the punishment we would receive would be far less severe. On a few occasions, we weren’t punished at all because our parents either didn’t know who to pin the blame on or they didn’t have the energy to deal with it (it was probably the latter). I’m not sure if this dynamic exists today amongst siblings, but this was our unwritten, but very real “kid-contract” for avoiding punishment. And it worked.
Unknowingly, my brother and I were engaged in a very common game of decision analysis called prisoner’s dilemma. By definition, “prisoner’s dilemma is a paradox in decision analysis in which two individuals acting in in their own self-interests do not produce the optimal outcome (1).” In more critical applications, the engaged individuals could be a group of people, businesses, or even countries. In the case of my brother and I, we knew in advance that we were not going to tell on each other. However, oftentimes each party involved does not know how the other will respond.
The most commonly told example to describe prisoner’s dilemma is the tale of the two bank robbers. There are two thieves who are apprehended by police after robbing a bank together. They are taken to the police station and separately interrogated. Neither thief trusts the other, so there is a level of uncertainty that each thief feels as each enters their interrogation room. Here’s what the thieves do know: If neither admit to the crime, they’ll each serve one year in prison. If one admits to the crime and the other doesn’t, then one goes free and other gets three years in prison. Lastly, if they both admit to the crime, they each get two years in prison. Clearly, the optimal situation for both of them is to not admit to the crime, but neither of them know for certain how the other will respond. In the absence of uncertainty, it is better for each of them to admit to the crime and serve two years in prison even though they both knew in advance that there was a better option.
The first time I learned about prisoner’s dilemma was in an economics course I took when I was getting my MBA. In the last class before the final the instructor forced us into a prisoner’s dilemma-like situation. She told us that if no one physically showed up for the final she would give everyone an “A” on the test. However, if any one person showed up, everyone who didn’t show up would get a zero. Given that this was a large class and we all didn’t know each other well, there was no real way we could collude to make sure no one went to the class on finals day. Supposing we did want to collectively conspire to not show up, the only way we could do so was to use an intra-class communication web portal, but the instructor had access to the portal as well and told us we would all fail the course if we attempted to use it for collusion. The only options we had were to either show up for the final or not to. Needless to say, on the day of the final everyone showed up and the first thing the instructor said to us before passing out the exam was, “So. I guess you all don’t really trust each other!” We all laughed, but we knew she was right. In hindsight, I think we all chose correctly.
Diving Deeper into Game Theory
John Nash was an American mathematician who won the Nobel prize for his contributions to the development of game theory, which is the underlying theory behind prisoner’s dilemma. According to the Corporate Finance Institute, “The Nash Equilibrium conceptualizes the behavior and interactions between game participants to determine the best outcomes. It also allows predicting the decisions of the players if they are making decisions at the same time and the decision of one player takes into account the decisions of other players. Take for example two companies, Company A and Company B. Both companies want to determine whether they should launch a new advertising campaign for their products. If both companies start advertising, each company will attract 100 new customers. If only one company decides to advertise, it will attract 200 new customers, while the other company will not attract any new customers. If both companies decide not to advertise, neither company will engage new customers. In the end, Company A should advertise its products because the strategy provides a better payoff than the option of not advertising. The same situation exists for Company B. Thus, the scenario when both companies advertise their products is a Nash equilibrium (2).”
In the business world, there are countless examples of prisoner’s dilemma and/or Nash Equilibrium scenarios that take place every day. For example, the price of non-renewable energy sources (e.g oil and gas) is (usually) based on supply and demand. When there is general consent between the rest of the world (ROW) and the Organization of Petroleum Exporting Countries (OPEC) in the Middle East regarding the volume of oil supplied to world markets, the industry is generally stable and everyone makes money. However, if OPEC decides to increase output when public demand hasn’t changed, the price of oil drops resulting in less profits for everyone.
Labor unions and corporations are in some sense engaged in prisoner’s dilemma. Unions are intended to protect its members from unfair or unethical practices of corporations. When employees are treated poorly, they can either continue to put up with it, quit, or go on strike. The power of going on strike lies in a majority of workers refusing to work, thereby putting the employer’s business at a standstill. However, if there are enough individuals who ignore the strike and go back to work, it weakens the position of the employees (and the labor union) and the company continues unabated.
To me, the concept of game theory/prisoner’s dilemma is one of the most intriguing decision analysis exercises out there. While I think it’s fun to think about how this concept works in the business world, it’s equally interesting to think about it in a social and political context. When the majority of people in a (democratic) society unite to effect change, there’s little that any government can do to stop them. This is one of my biggest complaints with people who complain about politics but refuse to vote. Unless you exercise your right to vote, stop complaining. Similarly, there are some (larger) retail stores I refuse to shop at because I think their wages and employment practices are unfair. When a business begins to lose enough customers due to some perceived wrongdoing, that business is likely to correct the problem. For example, in 2011 Netflix decided to bifurcate its DVD mailing service from its streaming service, angering millions of customers and causing a big drop in subscribers and the company’s stock price. Netflix quickly reversed the decision and returned things to normal. This is a classic example of game theory because even though Netflix knew that distancing itself from DVDs was a good long-term decision for the company, they did not know how their customers would react. Try to imagine performing a scenario analysis on this decision at that time:
Action – Separate DVD mailing service from streaming service. Great for the company long-term.
- Reaction 1 – Customers embrace the change. No issues.
- Reaction 2 – A few customers hate the change at first, but eventually embrace the separation.
- Reaction 3 – A large majority of customers hate the change and it hurts the company.
Action – Keep DVD mailing service and streaming service together. Not good for the company long-term.
- Reaction 1 – Customer sentiment remains the same.
- Reaction 2 – Customers slowly leave over time to Netflix competitors who offer more robust, standalone streaming services and Netflix ends up like Blockbuster.
In the end, things worked out for Netflix, but as you can see, making these types of decisions in real time is difficult because there is no certainty as to what the outcome will be.
Thank you for reading this (unusually) longer post from me. If you have any thoughts or examples of prisoner’s dilemma that you’ve experienced, I’d love to hear about it in the comments.
Cheers – KM