If at First You Don t Succeed Try, Trilemma
Think about the last tough choice you had to make. Should I do this or that? Which way is faster, this way or that way? Should I spend my time and money here or there? Often we simplify difficult optimizations problems into dilemmas. Dilemmas are hard, but they are not that hard. Juxtaposition is one of the human mind’s greatest tools. We humans are adept at comparing two things, even on different scales. Kahneman and Tversky of “Thinking Fast and Slow” fame call this intensity matching. Two choices is relatively easy.
Two party models are extremely useful in trade, negotiations, client-server, principal agent and many other situations. The prisoner’s dilemma is a beautiful execution of game theory when two parties are faced with a classic dilemma, cooperate or defect. In many of these situations the trade off is readily apparent, some compromise may be created or in the case of the prisoner’s dilemma, can be mathematically derived. The problem with dilemmas is not that they pose a choice too difficult to answer, it is that that a simple static cost/benefit analysis can solve problems. I once overheard a great Jedi say:
No! Try not. Do… or do not. There is no try.
Yoda was right, there is no try in dilemma, but there is a tri in trilemma. I love trilemmas. Part of the reason trilemmas are so much cooler than dilemmas is that a third variable introduces syllogism. Deductive reasoning with two variables is boring. If not one, than the other. Three variables starts to get interesting, while introducing minimal complexity. The first formal trilemma I encountered was the Impossible Trinity, in an econ class on international finance.
Then in the corporate world we had a saying that is sometimes called the Project Management Triangle, but I prefer to call it the Client’s Trilemma. Trilemmas are also found in data management, for example the RAID (redundant array of inexpensive disks) trilemma that is bound by the same three constraints: speed, price and quality. I suppose as computers have gotten cheaper the questions have moved from problems of physical space to distributed systems problems like the CAP (consistency, availability or partition tolerance) trilemma. Regardless, trilemmas are the perfect “lemma” to me. There is something beautiful about a framing that is the minimum viable description of a dynamic system. To my ethnocentric view, three is the perfect number. Eastern philosophies have long held out the tetralemma as another way to frame problems. Perhaps it is the myopia of the Christian Trinity versus the complexity of Bhuddism. Regardless, trilemma seems right to me. Yesterday, I saw Dan Geer], chief scientist at In-Q-Tel, speak about cyber security, society and risk. He brought up what I will call Geer’s trilemma; choose two: security, convenience, freedom.
This trilemma is tricky because, although it is a cyber security paradigm, it applies more generally to a set of security policy choices. The trilemma of international finance or the project management triangle may elicit some preferences, but there are fundamental rules to guide your decisions. A country’s decision to impede the free flow of capital (as China has chosen) may retain control of its monetary policy and exchange rate, and in some ways this is a societal choice, but it is also a result of the economic position China is in. The United States would find it extremely difficult to restrict financial flows. You may link this to the USA’s social and policy choices around freedom, a strong dollar and, after a long political debate, a powerful and independent central bank. These decisions, however, are the result, not the cause of societal preferences. Below, is a toy model of the foundations policy decisions rest on. Perhaps this is too constructivist) a view of politics, but please bear with me. Realists) and liberals) may not see much of a difference between Geer’s Trilemma and the others. Perhaps the choices are simple optimization of power, as a realist would argue, or liberals might suggest that we can break some of these tradeoffs with international cooperation.
Geer’s Trilemma cannot be derived from our societal preferences because they are our societal preferences. Whether we have a functioning system to sort out these preferences is another issue. Politicians are not likely to ask voters to make unpopular choices. The bureaucracy will hedge their actions without a mandate from politicians to set clear priorities. Businesses and the market are apt to prefer convenience. Frictions are the enemies of transactions. Whether business and markets come down on the side of security or freedom, is probably a function of whether they are an incumbent or challenger. So far, I have been talking about national priorities, but with national boundaries weakening despite the efforts of various nationalist movements, this question may not be a national one. It is difficult to imagine a world with policy choices not tied to national governments and for the foreseeable future we will muddle along with our nation-state system, but really these choices are for you to choose. What are your preferences?