There is no disguising it. This column is going to be disconcerting to all readers. And for a number of reasons: most readers will not accept that economics has much to say on such grand issues as democracy and personal freedom; few will even consider the possibility that economic theory can arrive at any firm conclusions about the relationship between the two; and absolutely none will expect the truly dismal and yet inescapable outcome of my argument.
To make my challenge even more arduous, the kind of reasoning I will deploy will be unfamiliar and yet I will need each reader to follow every step of the argument. If you are sitting comfortably, I shall begin.
At the mention of the word theorem, all readers will have distant memories of Pythagoras', where it is proved that the sum of the squares of the two shorter sides of a right triangle is always equal to the square of the longest side.
But what about an impossibility theorem? Like Wagner's music, it is not as bad as it sounds. Suppose I started with three incontrovertible statements, i.e., statements that anyone, indeed everyone, would accept as true. An impossibility theorem would prove that even though the three statements taken individually are correct, when taken together they produce a contradiction – an impossibility.
You will see the scale of the task ahead: defining both democracy and personal freedom in a way that must be acceptable to everybody. Luckily, the hard work was done over 50 years ago by one of the most original and truly inspirational economists: Amartya Sen (1998 Nobel prize winner), who I had the privilege of having as a teacher at the LSE. In what follows I will summarise, rephrase, and somewhat cheapen one of Sen's most famous papers.
How can we define democracy in a way that would be universally acceptable? By identifying a minimal feature that every possible interpretation of democracy would have to incorporate. At its most basic, democracy must refer somehow to how the preference of individuals are translated into the preferences of the whole society.
What about the following minimal definition of democracy:
(1) If every member of society prefers outcome X to outcome Y, then X must be the democratic outcome.
Notice that the above definition is not meant to be realistic; it simply sets the absolutely minimal requirement that any democracy must satisfy. To drive the point home: if you were to disagree with point (1) you would be saying that in your 'democratic' society an outcome unanimously regarded as bad can be imposed instead of an alternative universally regarded as better.
Now that you are on board with minimal definitions, let's apply the same principle to 'personal freedom'. As a shorthand, I shall call a society that upholds some form of personal freedom as 'liberal'. What is the minimal definition of liberal society?
(2) Every individual has the right to have one of his/her preferences respected by everyone else.
Let's stress again that the above definition of liberal society is not meant to be realistic, but minimal. According to it, for example, North Korea would be a liberal society, provided each citizen had the right, say, to decide the colour of his/her socks. The power of the definition resides precisely in its minimality: denying its acceptability would mean that a society where everyone has absolutely no right to choose anything could be called liberal.
The stage is nearly set. We need one more, slightly technical but intuitive, definition. For our theorem to have universal applicability it cannot restrict the preferences of individuals; after all,
de gustibus disputandum non est ('In matters of taste, there can be no disputes'). Let's consider the tale of the Bishop, the Atheist, and Monty Python's
The Life of Brian.
This two-individual society wants to be both democratic and liberal (as defined above): if both the Bishop and the Atheist decide that one outcome is preferable to another, democracy demands that their unanimous decision be respected. Similarly, both the Bishop and the Atheist have one personal decision (and only one) that would have to be respected by society at large. The issue in question is who (if anyone) should be allowed to watch a controversial movie. There are three possible outcomes to be considered:
N: Nobody should be allowed to watch the movie.
A: The Atheist should be allowed to watch the movie.
B: The Bishop should be allowed to watch the movie.
To determine how this society is to decide, we need to know the preferences of its two citizens.
The Bishop's preferences are predictable: his most preferred outcome is N ('nobody should be corrupted by this filth!'); his second-best choice is B ('I will endure the degrading spectacle to save the Atheist's soul'); and the worst is A ('if the Atheist watches it, she will be beyond redemption').
The Atheist's preferences are interesting: her most preferred outcome is that the Bishop should watch it (B) – 'it may open his eyes'; her second-best choice is obviously A ('I enjoy an anti-religious chuckle'); and the worst is N ('a waste of a good movie').
In this minimally liberal society, the choice whether an individual watches a movie or not should be respected by society. This means that the Atheist's preference of A over N and the Bishop's preference of N over B cannot be overturned.
Notice how harmonious society's preferences are: if only one copy of the movie were available, respect for personal freedom of both the Bishop and the Atheist would result in the Atheist watching the movie and the Bishop being spared the scarring experience. Our minimal definition of liberal society is complied with: the Atheist should watch the movie.
Suppose now that the issue of who should watch the movie were put to a vote. Both the Atheist and the Bishop are unanimous: both prefer that the Bishop should watch it rather than the Atheist. In our minimally democratic society, the unanimous vote in favour of the Bishop watching
The Life of Brian cannot be overturned.
And here we have it: either we uphold personal freedom and let the Atheist watch what she likes or we respect the unanimous decision of all citizens and let the Bishop watch the movie. Personal freedom (even in its mildest form) and democracy (even in its mildest form) are incompatible. Our impossibility theorem is proven.
As one can have expected, Sen's impossibility theorem did not pass unnoticed in the academic world and to date it has been cited in over 2,000 articles, making it one of the most influential in the field of social choice theory (this is what economists call logic applied to political and moral problems).
Unfortunately, one cannot detect much influence of this in what passes for political debate these days. I am sure that readers of the
Scottish Review can find a multitude of examples that fit the impossibility described above, starting with the dilemma between preferences for national independence and respect for the laws of the land.
A particularly telling example was provided in 1991 when the Algerian military cancelled the second round of a democratic election because after the first round it was clear that the Islamic Salvation Party would have achieved a majority large enough to change the constitution and impose a theocracy. This was not a local North African aberration, but the result of a deep-seated problem: sometimes a choice must be made between democracy and personal freedom.
Or, indeed, between human rights, democracy and international trade: writing in
The Guardian on 9 December 2022, the ironically-named Foreign Secretary, James Cleverly, stated that: 'Britain is not a passive observer on the world stage. We want oligarchs and dictators to fear us'. Two days later,
The Guardian reported that the very same James Cleverly announced a change of policy 'mark[ing] a downgrade of a commitment to human rights' to facilitate trade with two dozen middle-level countries 'that are not necessarily democracies'.
Dr Manfredi La Manna is a Reader in Economics at the University of St Andrews
