A very good point, Zak! The version of HH-PMI that I calculated for this post reflects one thing: political pluralism within a parliament (the Uzbek example shows that even this is not always true). Given how imperfectly any elections reflect the underlying political preferences, my HH-PMI index does not do justice to the strength of a country's civil society. The latter could be assessed in the following two ways:

(1) If we are interested in political pluralism as a social phenomenon (strength of "civil society") we should be calculating HH-PMI using data from opinion polls measuring political preferences, not election results.

(2) Another alternative is to calculate HH-PMI using the actual proportion of votes cast for each party (rather than proportion of seats). This method is better than the one based on seat allocation because it removes distortions associated with the minimum threshold requirement, gerrymandering (redrawing of electoral districts), and the formula used to reallocate "surplus" votes (an important problem that is not very well known to the general public). This method also has one potential advantage over method (1): much larger sample size. However, if self-selection into voting is not random (as one would suspect to be the case in "managed democracies"), the results of elections and the index will be highly biased. The second problem is loss of information concerning parties and candidates that were barred from participation in the election (recent examples abound).

There is also a lot of criticism of those type of indices, so what exactly would lack of correlation be indicative of?

What is better is to find a counter-example, a country where such an index has a high value (much higher than 0.5, the case of two equal-size parties), but where the level of democracy is clearly at an appropriate level.

First, with the exception of Uzbekistan, HH-PMI is quite consistent (or correlated, if Lasha prefers this term) with The Economist Democracy Index that I referred to in the post. Please think of a counter example, as suggested by RT (e.g. a country with a high value of HH-PMI that we would think of as a great democracy).

Second, I am not saying that HH-PMI is a perfect measure of democratic deepening. For one thing, it is subject to manipulation through the PPP mechanism described in the post (see also my comment on Uzbekistan). I would rather say that a low degree of political monopolization (or, conversely, high degree of political pluralism), as measured by HH-PMI, is a good starting point for a democratic polity to emerge, a necessary but not sufficient condition for democratic deepening.

Third, I like HH-PMI for its simplicity and robustness. It certainly does not depend on anybody's subjective judgement.

Fourth, and related to the above point, any other index I am familiar with is calculated as a weighted average of scores a country gets on a certain set of characteristics that are that are deemed important for democracy. Such indices are subject to criticism for three obvious reasons:

a) They are based on "expert" surveys. The results therefore critically depend on sample size and choice of experts (who are surely not selected at random);

b) There is no universal agreement on which characteristics define democracy. For instance, are access to education and income inequality important attributes? Some would say that having access to healthcare, job and food security are more important attributes of democracy than having the right to vote every four or five year;

c) A priori there is no correct way to assign weights to particular characteristics of democratic governance (freedom of media, freedom of political association, freedom of elections, etc.)

The true value of the HH-PMI index for Uzbekistan should be 1. Uzbekistan is a great example of the PPP manipulation mechanism at work. All four major parties that combine for 135 seats were approved (sic!) for participation in the 27 December 2010 election support president Kerimov, there is no real opposition in the country. Moreover, according to Wikipedia http://en.wikipedia.org/wiki/Uzbekistani_parliamentary_election,_2009%E2%80%932010, "several opposition politicians have alleged that all candidates also had to be approved by the government before they would be placed on the ballot". A fifth party ("ecological movement") was granted fifteen seats in the parliament, one seat per region, without participating in the election.

You are welcome to try, Yasya

Eyeballing the data, it appears that you are right with the exception of Belarus and Georgia. Both are quite diversified (not having any natural resources to speak of) and yet their parliaments are dominated by a single party. Other than that it does seem that having a large natural resource extraction sector helps establish a political monopoly of sorts.

Without having done the regression I already have a theory about it :-)

Economic interests necessitate political representation. (e.g. the rising merchants and industrialists in Great Britain pushed for the Reform Bill of 1832; the newly emerging "managerial class" in Ukraine helped finance the Orange Revolution). Therefore, the more diversified the country's economy and the more competitive the industrial structure, the more diverse are economic interests, and the more plurality and competitiveness one may see in the political system.

(Of course a diversified economy does not immediately translate into a diversified parliament. For example, in Georgia the "middle class" may still be relatively week economically as well as politically, but this will change as the country develops).

There is a potential feedback effect from political competitiveness to the industrial structure as well: tightly controlled parliaments would tend to pass the laws which help concentrate economic power in the hands of the select few.

The countries with large natural resource endowments seem to be an interesting special case. One may argue that in these countries the political systems are "natural monopolies", as the economy consists of only one naturally monopolistic industry.

That said, however, I myself do not much buy into the argument of "naturally autocratic" states. The feedback effect mentioned above can seriously entrench the power of the few, making the economic (and political) diversification all but impossible without a revolution of sorts.

An excellent theory. Here is a possible refinement (to account for the Georgian case).

To generate political pluralism you need economic diversity AND organization/cohesion/leadership attached to each sector.

Small traders and restaurateurs (the bulk of Georgia's emerging middle class) are inherently unorganized, atomized and lacking prominent leadership (small traders are not made of the right material to provide leadership, that's why they stay small). Small traders are very good social material to support the emergence of dictatorships.

Economic diversity can translate into political diversity if the various economic interests are sufficiently organized and endowed with sufficient resources to compete for power.

Absolutely!

It seems that taking advantage of the economies of scale is beneficial not only economically. It is instrumental in acquiring a political voice.

Borrowing from Industrial Organization theory we may talk about the “optimal firm size” and the optimal level of economic diversification to create a system of political pluralism with effective checks on the executive power...

Lasha, please read my other comments. Uzbekistan is a special case. Its true index is 1. Kyrgyzstan parliament is quite fragmented, reflecting a true plurality of political preferences. It is quite democratic in this sense (there are no restrictions on any freedoms, and people are free to speak, form associations, parties, etc.). It surely has a long way to go in terms of institution building, poverty reduction, etc.

In any case, Kyrgyzstan, Moldova, Estonia, Poland and Hungary have HH-PMI values below 0.5, which allows for ideas to be contested. In countries with HH-PMI values below 0.5 the largest party would most likely have to form a coalition. This in itself forces a political bargaining and compromises. The lower the index, the harder it becomes to build a coalition, pass laws, accept budgets, etc.

Lasha, please read my other comments. Uzbekistan is a special case. Its true index is 1. Kyrgyzstan parliament is quite fragmented, reflecting a true plurality of political preferences. It is quite democratic in this sense (there are no restrictions on any freedoms, and people are free to speak, form associations, parties, etc.). It surely has a long way to go in terms of institution building, poverty reduction, etc.

In any case, Kyrgyzstan, Moldova, Estonia, Poland and Hungary have HH-PMI values below 0.5, which allows for ideas to be contested. In countries with HH-PMI values below 0.5 the largest party would most likely have to form a coalition. This in itself forces a political bargaining and compromises. The lower the index, the harder it becomes to build a coalition, pass laws, accept budgets, etc.

Just to reiterate: an index of political pluralism such as HH-PMI, does not capture "the true sense of democratic transformation". It is a necessary condition for a true democratic transformation to occur. I find it very hard to see how "equality before the law; civil liberties; human rights; judicial independence; freedom of political expression; freedom of speech; freedom of the press" can be promoted and protected in a single-party "democracy".

Also, as someone who reads Ludwig von Mises, you should appreciate the value of competition in any sphere, including politics. Political competition, as we all know, is particularly vulnerable to government intervention through meddling with constitutional and electoral law, licensing and registration of parties, legal prosecution and expulsion of political opponents, abuse of administrative power and resources (money, access to media, etc.). When HH-PMI is above 0.6 (which is equivalent to one party having more than 70% of the seats) you can be sure that some if not all of the above tactics are being employed.

A side remark: the lower the index, ceteris paribus, the harder it becomes to build a coalition, pass laws, accept budgets, etc., in any environment. Whether democratic institutions are developed or not. I believe this is a theorem in game theory (the "core" is smaller).

Your point on rent-seeking suggests an interesting topic for research: how does political competition affect rent-seeking behavior and welfare?