A long season of high–stakes elections in Georgia, Ukraine, and now the United States is finally over. Once the last campaign posters are taken down, we may as well start asking: now what? Whether we like to admit it or not, the success of democracy is probably ill measured by the show of competitive campaigns or the transparency of the voting system. Instead, the success largely depends on how informed and engaged in the political process the general public remains after the elections.
This is something Ukraine has discovered the hard way since 2004. As engaged and passionate as people were during the 2004 elections, as sidelined and isolated from the political process they found themselves to be after the hard-won Orange Revolution victory.
In part, one may say, this happened because people had taken a step back from active political life and just waited for the democratic tree to start bearing the fruit of prosperity.
But is democracy really a key input into the production of a well-functioning economic engine? The evidence on the relationship between economic growth and democracy is mixed at best.
Branko Milanovic in his paper “The Modern World: The Effect of Democracy, Colonialism and War on Economic Growth” points out that the growth outcomes of democratization can be very different for different countries. In fact, as the figure below (left panel) shows, poor countries are unlikely to grow faster than average, whether they are classified as democracies or dictatorships.
So, should we view democracy not as an input into production, but rather a luxury good enjoyed by the society only once it manages to achieve a certain level of prosperity? Perhaps, free speech and civil liberties are not facilitators of economic growth, but rather expensive indulgencies – like nice clothes or automobiles.
Looking at the right hand side panel of the figure above, this does not seem to be the case. Rich democracies appear more likely to grow faster than the rest of the world, while rich autocracies are drawing from a very mixed bag of growth outcomes.
How do we reconcile these differences in the experience with democracy in rich and poor countries?
Perhaps it is best to think of democracy as neither a universal catalyst of growth, nor a luxury consumption good. Rather, it may be likened to a sophisticated instrument that has the power to facilitate growth in the hands of those who know how to use it.
If at the core of mature democracy is the aspiration to actively participate in the process of decision making, the ability to take ownership of the political and economic agenda of the country, then it would not surprise me that democratization often produces best outcomes in societies that are already well-to-do.
In such societies people are more likely to already have a good idea how to organize their economic lives more efficiently. Consequently, more likely to know what they want to achieve through the democratic process. In light of this, the hard work of learning to live with freedom and use it wisely starts only now, after all the campaign banners have been folded.
Comments
Interesting. The right panel of the is indeed intriguing. I had a quick glance at the paper and I was wondering how much is the author (of the paper) manages to make the data on such a long span (1820-2000) comparable. Meaning, I was wondering whether the fat tails of the rich dictatorships distribution are not time dependent. I actually suspect that right tail collects lots of observations from nineteenth century, while left one collects much more recent performance indicators.
In this case rich dictatorships might not be drawing single-year growth rates from the distribution with larger variance (compared to the democracies). The phenomenon might have to do a structural transformation of the world economy that has taken place over the two century period covered by the data.
Zak, I think the time dependence problem is mitigated a bit by the fact that he is using relative growth rates (single-year growth rate of a country minus the average world growth in that year).
Of course, I have to agree with you, this doesn't preclude the possibility of a bias - the growth patterns in the 19th and in the 20th centuries were very different. It would be interesting to take a closer look at the underlying data.
I am not sure that taking such a long-term perspective is a good idea. First, the nature of democracy has changed over time. What was democracy in 1820 would not qualify as such in the 21st century. Most importantly, the economic policy implications of greater democracy from 1830s through the early 20th century where not increased redistribution in favor of the poor but rather greater economic freedoms for the new capitalist class. Thus, democracy was almost by definition good for economic growth (through capital accumulation in the hands of a few and exploitation of colonies in Africa, Asia, Latin America and everywhere else).
(средняя температура по больнице)
The gradual expansion of voting rights (enfranchisement) to the poorer strata of the population (and women), the rise of labor movements in late 19th and early 20th century, the Bolshevik revolution in Russia, etc. transformed the nature of democracy and its economic policy implications. Democracy has now become about keeping the masses happy, whether through greater redistribution, public works or any other means. Between the two world wars Europe lived through a huge disillusionment with democracy, with one country after another falling prey to more efficient autocratic and totalitarian dictatorships. During this period democracy became synonymous with decadence and decay, not growth.
After the second world war we see a different kind of development, with democracy getting a new lease of life and once again being associated with prosperity and growth. This period is associated with neo-colonialism, i.e. the formal exit of former colonial powers from Africa and Asia and development of new economic "exploitation" mechanisms.
Finally, the rise of China in the late 20th century, gives rise to (or, rather, reasserts) an alternative ("Asian") model of economic development. Development with markets but without democracy. This models was quickly followed elsewhere in Asia, leaving the European democracies far behind in terms of GDP growth.
To summarize, by "mixing" economic growth and democracy over four or five very distinct periods Milanovic does not produce any useful insights other than coming up with a measure of the average fever in a hospital
He controls for the Zeitgeist (his term), so this at first sight not an issue.
Eric, I see your point, but just a couple of comments.
nly if XRCOMP is Election or Transitional
1. I am not sure you are right in your analysis of the "periods of democracy". The October revolution in Russia certainly did not at any stage result in what could be qualified as a democracy, despite redistribution (at least not by the definition Milanovic is using). In fact, to make the redistribution and the subsequent state capital accumulation work, extraordinary repressive measures were taken against the people (e.g. famine in Ukraine)
(I am attaching the link to the relevant files. In the Polity IV index excel spreadsheet, check the polity2 index for USSR. By BM definition anything greater than 5 is democracy, less than 0 - autocracy http://www.systemicpeace.org/inscr/inscr.htm )
2. The exit of former colonial powers from Africa and Asia also did not automatically translate into the democratization of the former colonies. In fact, the reverse has taken place for many such countries, and these transitions did not result in high growth. I'm not sure exactly what changes you had in mind when saying that democracy was getting a "new lease on life".
3. The "Asian model" as impressive as it seems can at least in part be attributed to the fact China started out very poor. Not surprisingly the catching up growth rates are much higher than in the "democratic West", where the economies were closer to the steady state.
The point is not that the autocratic regimes are uniformly bad for growth. It's just that results seem more or less hit and miss, at least for "rich" countries. (to borrow from an old nursery rhyme: when it is good, it is very very good, but when it is bad it is horrid).
4. One can argue about the nature of democracy in the 19th and the 20th centuries, but the measure (Polity IV index) Milanovic is using tries to avoid relying on the modern definitions as much as possible, and tries to define representative government structure pretty consistently through time. Just an example from a coding book:
Authority Coding (Democracy)
Competitiveness of Executive Recruitment (XRCOMP):
Election
Transitional
Openness of Executive Recruitment (XROPEN)
Dual/election
Election
Constraint on Chief Executive (XCONST):
Executive parity or subordination
Intermediate category
Substantial limitations
Intermediate category
Competitiveness of Political Participation (PARCOMP):
Competitive
Transitional
Factional
For more detailed explanations and definitions, you can check out page 14 in the code book, section 2.1 and page 20
(http://www.systemicpeace.org/inscr/p4manualv2010.pdf).
There is a lot of stuff to digest here and I am surely not doing justice to Milanovic. It is very difficult to statistically tackle the issue of growth and democracy given that the data and definitions could be stretched in various directions...
Reading your reaction, however, I realized that I was not very clear in my initial comments. I mostly focused on Western European countries (not Russia, not Africa, and not even China) as for much of the period under discussion democracy was largely a Western European phenomenon (including British dominions and the US). I tried to argue that democracy was associated with growth during much of the 19th century and early 20th century. It was associated with decay and decadence, as I put it, between the two world wars. And, finally, it was again positively associated with growth after WWII, until China’s economic ascendancy.
It may be, as Michael suggests, that Milanovic uses year dummies or another statistical technique to control for the effect of time. Nevertheless, I remain to be convinced that -- intellectually – his method brings us closer to understanding the impact of democracy on economic growth today or in the past. My preferred method would involve constructing a set of time-variable political economy models.
Here is a somewhat more detailed restatement of the phases in democratic development, as I see them.
The process of democratization in the UK, with which I am familiar relatively well, involved several stages of gradual and cautious expansion of voting rights and no less gradual ascendance of the House of Commons over the House of Lords. This was clearly a pro-growth development given that the fight for democracy was essentially a fight against landed aristocracy and the restrictions it tried to pose on the free reign of the capitalist class (the driver of growth).
This period also happens to be a period of major technological breakthroughs, such as rail and steam, which facilitated a vast expansion of the British empire, the leading democracy at the time (France finally became “democratic” in 1870). Thus, one should observe a strong correlation between the extent of democracy and growth during this period.
Importantly, given the ongoing process of urbanization and the vast reserves of low quality labor it created, the spectacular growth achieved during this time by the UK did not trickle down, leading to extreme poverty and destitution documented by Dickens (1812-1870). [This is no chance that Dickens is a contemporary of Marx (1818-1883)].
When and where the fight against landed aristocracy was completed, the process of democratization started serving a different purpose, with new parties and classes joining the political arena. In late 19th century and certainly in the early 20th century we see organized labor beginning to claim its rights for a bigger share of the pie. This changed the nature of democratic politics and its implications for growth. Democratic politics in the age of mass participation forced greater concessions from the ruling classes, pushing up wages and other costs of running the country. Arguably, many of the new developments such as expansion of education, were good for economic growth. However, frequent strikes and higher wages put democratic European producers at a competitive disadvantage at the same time as steam, rail, telegraph and other technologies spread to other countries. Likewise, managing the empires in this period turned out to be a relatively costlier affair, reducing the benefits of colonialism and slowing down the growth rates of European democracies.
I had no intention of claiming that the Bolshevik revolution resulted in greater democracy as far as Russia as concerned. If anything, it was the February revolution that produced (for a short while) a kind of parliamentary republic. What the Bolshevik revolution did, however, is raise the economic price of democracy in Western Europe, forcing the ruling elites to make further concessions to the working classes in the form of higher wages, better working conditions, shorter working days and social insurance. Some of these concessions were probably paid for by printed money, resulting in inflation and hyperinflation (as in Weimar Germany).
As a result, between the two world wars, democracy in Europe evolved into in-fighting. Growth rates declined, the rifts within the ruling elites deepened, leading to moral bankruptcy, military and political defeats at the hands of foreign and domestic dictators.
The post WWII period, indeed, marks a new beginning. It is not about democratic development in Africa (although ridiculous attempts to plant the seeds of democracy in former African colonies were made by the retreating British and French forces). Rather it is about a renewed association between democracy and economic growth in Europe (and Japan, to the extent that Japan is a democracy). To be democratic in this period carried an additional benefit of being a US ally in the cold war it waged on the USSR. Thus, democracy brought significant material goodies in the form of e.g. the Marshall plan, technical and humanitarian assistance. The European democracies during this period engaged in building massive welfare states while learning (the hard way) how to manage money supply and avoid runaway inflation. They continued to enjoy from the supply of cheap raw materials from their former colonies without using (too much) military power. At least until the 1973 oil crisis.
Eric, very interesting! I now understand better what you mean, and I agree with much of what you say. The idea about the time-varying political economy models (and there need not be one dominant model for any given time period) is certainly something worth developing further.
However, there are important nuances to be considered. The 19th century growth model (which as you said was associated with significantly lower compensation for the working class) was already on the way out in the 20th century. Its doom it seems to me was not spelled by the Russian revolution, but rather by the changing nature of technology. The need for a more sophisticated, educated workforce. Also, the increased competition in international trade, which created the need to develop internal demand rather than rely on the outside markets.
In a sense, the Marxian capitalism like the economic model of slavery may have worked (for a while) until it didn't, driven by the changing nature of technology. I think it is no accident that the abolitionist movement in the US was coming from the more industrialized North.
Well - many more interesting aspects to explore and think about!
Indeed, and thanks for a great blog post!
This discussion could be easily resolved had we the access to Milanovic's data. The only thing we have to check is what is the share of observations from 19-s century below and above mean in rich dictatorship distribution. I still think that that share is disproportionately high above mean and disproportionately low below mean.
I think it would have been much more interesting if he could concentrate on shorter time-spans (perhaps 20 years) and do the similar analysis for 20-year successive periods. Of course the densities would not be this smooth but that should not be the aim of the research anyway. This would show us how the dependency has been evolving.
Zak, that should be relatively easy to check, this data is in the public domain. The GDP per capita figures and growth rates are from Angus Maddison historical statistics. The Democracy data is from Polity IV University of Maryland project.
Here are the links:
http://www.ggdc.net/MADDISON/oriindex.htm ( go down to Historical statistics and download one of the Statistics on World population... files)
Political indexes are available at
http://www.systemicpeace.org/inscr/inscr.htm
Download the Polity IV excel spreadsheet. The relevant variable you want to look at is "polity2"
I agree, Zak, focusing on periods of 20-25 years sounds about right: 1832-1848-1870-1889 (major benchmarks in the 19th century British and European history)
1889-1914 (scramble for Africa, end of westward expansion in the US, beginning of organized labor movements)
WWI...
1920-1939... (the lowest point of "democracy" in Europe, its defeat at the hands of totalitarian dictatorships)
WWII...
1947-1975 (democracy on the rise in Western Europe and Japan against the backdrop of a painful decolonization process from India's independence till the end of the Vietnam war; associated student protests)
1975-1991 (from the first oil crisis till the fall of the Berlin wall, the highest point of "democracy" worldwide))
1991-today (post-Tienanmen Chinese/Asian ascendancy, EU expansion/crisis, end of the Cold War and beginning of the global war on terror; ballooning debt in the US)
Captcha Image