On July 5th 2015, Greeks voted convincingly (61% no) against a bailout proposal in the referendum whereas according to some polls, a substantial majority of the Greek people seemed to want their parliamentarians to pass the new agreement (the `agreekment’), which imposes, in many ways, even harsher measures than the conditions which were voted on in the last referendum. How is this possible? Are Greek voters irrational? Is this not (again) a proof of the large mood swings and unfoundedness of voters’ preferences?

**Choices under uncertainty**

What is very important to realize is that the issues on which one asks voters to choose in these circumstances are highly complex. They involve a complicated mixture of macroeconomic and monetary issues, international and domestic political organization, social sustainability etc. If one asks economists and other experts about the expected consequences of all the different scenarios (assuming that one even has a clear view of all the possible outcomes), it is clear that most don’t have any clear view on this. Precisely because there are quite limited (if not to say almost no) comparable precedents. Hence, some sort of uncertainty is the real challenge facing everyone that has to make a choice on voting `yes’ or `no’ in the referendum, or whether or not to support the attitude of the Greek PM Alexis Tsipras in accepting the current agreement conditions.

Economists and decision theorists have long made the distinction between choices under certainty (where outcomes are known and sure), risk (where all possible outcomes are known, but there exist some objective probability for each particular outcome to be realized), uncertainty (where possible outcomes/consequences are known, but the different likelihoods for each particular one to realize is not known) and finally choice under ambiguity. Under the last category, neither the set of possible outcomes or the set of probabilities one should attach to them are known by the decision maker (in this case, the voter).

From the description of the different environments of choice, it is very likely that the complex situation of Greece and the Eurozone crisis could be best analyzed within the framework of ambiguous choice. However, as this is one of the most difficult analytically tractable environments to perform some analysis, let’s consider choice under uncertainty to provide some rationalization for the Greek voters’ preferences. In the standard (mathematical) description of choice under uncertainty, it is important to make a distinction between preferences (tastes) and beliefs. Preferences describe what an individual would choose in case (s)he is completely certain about the final outcome of a given choice. In particular, it describes the ranking of alternatives under the easiest environment of certainty.

Economists use the mathematical tool of utility numbering to represent these preferences. In particular, let A and B be two different outcomes, then in case A is preferable to B for an individual, we write that u(A) > u(B), where u(A) (u(B)) is the `utility’ provided by outcome A (resp. B). No probabilities involved here. If I give different alternatives to a consumer (or a voter) (s)he will just pick one by comparing the resp. utilities of the different outcomes. However, now assume that an individual knows that the different outcomes can be A or B, but has no clue on what the likelihood is that each outcome will realize itself, given that he chooses some action. In this case, the individual has to form beliefs on the probabilities of the different outcomes. In particular, let p(A,x) (p(B,x)) indicate the belief that the individual attaches to A (resp. B) being realized given that he takes the action x (equivalently, p(A,y) and p(B,y) have the same interpretation, but for another action y). As we assume that it is known that A and B are all the possible outcomes, we must have that p(A,x)+p(B,x)=1 (for all actions x). Given the utilities of A and B (describing preferences over outcomes) and the beliefs concerning likelihoods of their realization, which action is the individual going to take? Well, decision theory suggests to use the (subjective) expected utility criterion, i.e. choose action x above y in case:

*p(A,x) u(A)+ p(B,x) u(B) > p(A,y)u(A)+ p(B,y) u(B)*

**The Greek situation**

Let’s now apply this reasoning to the situation in Greece. In this case, it is suggestive to consider three outcomes w.r.t. the negotiations (obviously this constitutes a massive simplification, but it helps to clarify the point). First, A=no Grexit, debt restructuring (or at least a better agreement in terms of what Syriza wants), B= continuing with the bailout conditions (austerity measures) and C= Grexit. Suppose that the preferences for the (average) Greek voter are represented by: u(A)> u(B)>u(C). This means that Grexit is the least desirable outcome for the Greek voter, whereas A is the most preferred. B (continuing along the lines of austerity) is the status quo condition and is considered to be the middle of the road alternative. For the referendum, x1=no in referendum, whereas y1=yes in referendum. By the outcome of the referendum, assuming rationality of the average Greek voter, we should have:

*p(A,x1)u(A)+p(B,x1)u(B)+ p(C,x1)u(C ) > p(A,y1)u(A)+p(B,y1)u(B)+ p(C,y1)u(C )*

Now, w.r.t. the current agreekment, it must be true that saying `no in parliament’ (x2) is preferred to `yes in parliament’ (y2). This means:

*p(A,x2)u(A)+p(B,x2)u(B)+ p(C,x2)u(C) > p(A,y2)u(A)+p(B,y2)u(B)+ p(C,y2)u(C)*

Many people find these two conditions incompatible, they point to mood swings or irrationality to describe these seemingly contradictory choices for actions. However, it should be clear that, by casting it into the framework of choice under uncertainty, the paradox disappears. Indeed, only when x1 and x2 (resp. y1 and y2) are considered the same. However, given the different context, it is quite plausible to assume that the voters have different beliefs pertaining to the outcomes. Indeed, they were likely to overestimate the likelihood of increased bargaining power by voting no in the referendum (i.e. assuming a rather high value p(A,x1)) and underestimate the probability of a Grexit. These beliefs were largely supported by the discourse of Greek PM Alexis Tsipras and the Syriza leadership. On the other hand, w.r.t. the vote in the Greek parliament, the beliefs changed as the same PM (on which the average Greek voter seems to put a lot of trust), now said that Grexit and chaos are very likely in case the parliament would not adopt the bailout plan.

**Leadership as basis of choice**

The conclusion of the description provided here above is that the important factor in situations of uncertainty are the beliefs of the people one trusts. In this case, it means the evaluation of the Greek PM Alexis Tsipras. This insight, the fact that democracy is not about rational voters expressing their will on outcomes and then merely selecting people to implement these outcomes, is quite old. The famous political economist Joseph Schumpeter outlined it in his classical book “Capitalism, Socialism and Democracy” . In particular, Schumpeter commented on the need for leadership in (democratic) politics, because of the intrinsic complexity. The dangers in neglecting this need are also clear:

“History however consists of a succession of short-run situations that may alter the course of events for good. If all the people can in the short run be ‘fooled’ step by step into something they do not really want, and if this is not an exceptional case which we could afford to neglect, then no amount of retrospective common sense will alter the fact that in reality they neither raise nor decide issues but that the issues that shape their fate are normally raised and decided for them. More than anyone else the lover of democracy has every reason to accept this fact and to clear his creed from the aspersion that it rests upon make-believe.”

To reiterate the argument, the average voter in an election is not necessarily irrational or merely responsive to emotional (mood) swings, however, given the complexity of (modern) politics and economics, (s)he is vulnerable to direction and information fed to them by leaders. This is a natural constitution of politics, as Schumpeter clarified. Denying this or hoping for complete rationality in a utopian framework is dangerous. Focussing on good, responsible and competent leaders is important and then induces (rational) choice under uncertainty with voters. Needless to say, I am not expressing myself (in this article) on my personal opinion on whether or not Tsipras has made the right decisions or given the proper advise. I do have such an opinion, but I leave the expression of these opinions to other media/articles.

Photo credit: Justin Grimes; http://bit.ly/1Mq0ZUC