In the summer of 411BC, a group of citizens launched an oligarchic coup d’état against Athenian democracy (Kagan 1987). Following political turmoil and recent defeats to the Sicilians, the conspirators had decided that Athens would be better ruled by a small group of elite leaders rather than a large democratic assembly. This short-lived oligarchic regime was not followed by a return to classic Athenian democracy, but by a hybrid regime – the Constitution of the Five Thousand – that continued to exclude the poorest citizens from power.
By then, the tension between democracy and oligarchy had been eloquently articulated in a pamphlet by the anonymous ‘Old Oligarch’, who argued that democracy gave too many men of low social background access to power, creating less talented and virtuous rule (Kagan 1969). The Old Oligarch’s complaint was not against elections, but the allocation of public office, which was done by drawing lots. The most distinctive piece of democratic hardware in Athens was not the ballot box, but the kleroterion – a randomisation device, made of clay, that drew names and matched them to offices. The Old Oligarch was an even-handed critic. He recognised that allocating offices by lot ensured the representation of the lower classes even more than elections did. But he argued that the lottery system was overused, and that the selection system did not put enough weight on competence, even though the Athenian constitution used a different method for positions such as military command for which competence was of great importance.
Selection of public officials by sortition (lottery), and the perception of a trade-off between competence and fair representation, did not disappear with Athenian democracy. For example, the medieval republic of Florence used a lottery coupled with explicit limits on the ability of powerful families to access office. In his study of Florentine politics, Brucker (1962) writes that “[t]he method of selecting officials by lot placed little premium upon political ability or statesmanship, but tended instead to exalt mediocrity.”
Last year’s Brexit referendum and US presidential election have revived these debates on the advantages and limits of democracy. Many see the two votes as having produced incompetent choices. Others argue that they were part of a political radicalisation driven by social groups that do not feel adequately represented. This debate has revived sortition too. In a recent book, Van Reybrouck (2016) argues that this method could bridge the gap between elites and voters.
If one accepts that broad representation and competent political leadership are both valuable, it seems worth asking what electoral democracy with political parties achieves. Is it possible to escape the tension between broad representation (which sortition would help to achieve) and competent rule (which sortition would not help to achieve)? Finding an empirical answer to this question, however, is difficult. If we wanted to evaluate how a society selects politicians, and how these politicians compared to the population they represent, we would need measures of competence and representation for citizens as well as for elected politicians. Few countries systematically collect information like this.
In our recent research, we analyse administrative data for the entire Swedish population, and for all politicians elected in Sweden, between 1998 and 2010 (Dal Bó et al. forthcoming). These data include commonly used – but potentially misleading – measures of competence, such as education and income. We instead focus on three measures. One is an ‘Earnings score’, based on a methodology developed by Besley et al. (forthcoming). For Swedish males, we also measure a ‘Cognitive score’ (IQ) and ‘Leadership score’ (personality type), based on data from military enlistment.
To measure representation, we rely on another remarkable feature of the data – its ability to link individuals in families and through generations. This gives us information about the social class of the parents of politicians. Using it, we can test Old Oligarch’s assertion that social representation goes hand-in-hand with a less competent political class.
Are politicians more competent?
First, we ask whether elected politicians in Sweden, on average, have been as competent as if they had been selected by lot, better or worse.
The possibility of ‘negative selection’ leading to incompetence would not surprise some economists. Able individuals face a higher opportunity cost if they enter politics, so the incompetent may dominate public life (Caselli and Morelli 2004).
The answer is that politicians, starting with municipal council members, and continuing with mayors and national legislators, are strongly positively selected. In Figure 1, we show the distribution of the Leadership, Cognitive and Earnings measures of competence, as well as Education. The distribution for the population is shown with thick hollow bars. The lightest shaded bars show the individuals nominated, but not elected, to a municipal council seat. The next shade of grey corresponds to elected municipal council candidates. The rightmost two shades of grey show mayors and national legislators, respectively.
Politicians are more qualified than the average citizen, since the distribution of traits for each type of politician shifts to the right relative to that of the population. We also see similar shifts in the distribution from municipal council candidates to members, to mayors, and to national legislators. In our research we compared these patterns to other elite occupations, and found that mayors have the same competence level as CEOs of mid-size companies (those with 25 to 250 employees).
Figure 1 Competence of politicians relative to the population
Source: Authors’ calculations from Swedish population data.
The pattern that competence monotonically increases with political power could, however, have been caused by three different regimes:
- Elitism: Elite membership (belonging to wealthy and influential families) grants access to power. If politicians look able, it is because elite membership correlates with ability, and positive selection is an accidental outcome of oligarchic politics.1
- Exclusive meritocracy: Politics selects for competence. But if competence correlates with higher socioeconomic status, meritocratic recruitment has the side-effect of excluding those from lower social groups.
- Inclusive meritocracy: Politics selects on competence, but manages to represent broad social sectors even if the elites have on average a competence advantage.
Does positive selection reflect elitism?
If elite status drove political success, then conditional on family (and hence social) background, individual competence (merit) would not matter much for selection.
But conditional on family background, individual competence traits matter a lot. The upper row of Figure 2 shows the distribution of Cognitive, Leadership, and Earnings scores for municipal politicians and their siblings. The distribution for politicians in all three traits is clearly shifted to the right. Selection within the family explains between 70% and 80% of the overall difference between politicians and the average citizen for all three measures. The middle and lower rows of the figure show the analogous distributions for mayors and parliamentarians compared to their siblings. If anything, intra-family selection pattern is stronger for more powerful politicians.
Figure 2 Competence of politicians relative to their siblings
Source: Authors’ calculations from Swedish population data
Figure 2 shows that selection is not purely elitist, but it would still be compatible with an exclusive meritocracy. Sweden might select able and meritorious politicians, but these individuals may still belong to the elite, because (as it happens to be the case, even in Sweden) ability is positively correlated with affluent social background.
Is Sweden an inclusive meritocracy?
If this were true, Sweden would promote competent people while attaining broad representation. To answer, we investigated the social origin of politicians by looking at the income distribution of their fathers in 1979. Note that an income distribution expressed in percentiles that corresponds to the whole national population should be uniform at the 0.01 level.
The left panel of Figure 3 shows the income distribution of politicians active in 2011. (They were elected in 2010, and these incomes predate their entry into politics.) Clearly, the skewed distribution shows that politicians tend to be high earners, reflecting their higher-than-average ability. It also suggests that they enter politics despite facing a higher opportunity cost than the average person. The right panel is striking. It shows that the fathers of the 2011 political class – measured in 1979, so at a similar point in their lives – were not high earners on average. In fact, the distribution of their income percentiles is very close to the uniform distribution of the full population. In other words, the social origin of politicians has the same income distribution as the entire population. Politicians in Sweden are not only competent, but also representative of all walks of life. Sweden looks like an inclusive meritocracy.
Figure 3 Income distribution of politicians and their fathers
Source: Authors’ calculations from Swedish population data.
Democracy and virtue
After the coup in 411BC, one of the first acts of the oligarchs was to take away the wages of Athenian public officials. A modern economist might think that such a policy would discourage competent people from going into politics, as it does not compensate for opportunity costs. The Athenian oligarchs had a different theory. They judged that only the wealthy would be able to devote time to public affairs, and this would increase competence. Which theory is right would depend on whether Athens had the conditions to create an inclusive meritocracy like Sweden’s.
In Sweden, competent people devote time to municipal politics even though – as in Athens after the coup – council positions are (largely) unpaid. But people who serve are not disproportionately from the elite. This suggests that electoral democracy can attain the representative leadership we would expect from sortition, while significantly improving on it by selecting highly competent leaders.
This certainly does not tell us that democracy always leads to a virtuous outcome. But it suggests that electoral democracy can help us avoid the tension between representation and competence that has troubled observers since ancient Greece. Our results invite more investigation into what characteristics of society may help bring about these outcomes.
Editors’ note: This column summarises research and contains figures from the paper “Who Becomes a Politician?,” forthcoming in the Quarterly Journal of Economics, published by Oxford University Press. For permissions to reuse the figures presented herein, please contact [email protected] and reference the original paper.
Besley, T, O Folke, T Persson, and J Rickne (forthcoming), “Gender Quotas and the Crisis of the Mediocre Man: Theory and Evidence from Sweden,” American Economic Review.
Brucker, G (1962), Florentine Politics and Society, 1343-1378, Princeton, NJ: Princeton University Press.
Caselli, F and M Morelli (2004), “Bad Politicians,” Journal of Public Economics 88: 759–782.
Dal Bó, E, F Finan, O Folke, T Persson, and J Rickne (forthcoming), “Who Becomes a Politician?” Quarterly Journal of Economics.
Kagan, D (1969), The Outbreak of the Peloponnesian War, Ithaca, NY: Cornell University Press.
Kagan, D (1987), The Fall of the Athenian Empire, Ithaca, NY: Cornell University Press.
Van Reybrouck, D (2016), Against Elections: The Case for Democracy, London: Bodley Head.
 The possibility of elitism is not an idle consideration in electoral democracy. Even Athenians thought that oligarchic elements would have an advantage at courting popular opinion and attaining influence; Pericles was often accused of demagoguery. Similar tactics by patricians were repeatedly observed in the Roman republic, where fame and money could buy support.