Osh’s Electoral Geography (Updated)

Map imagery (c) 2012 Google

by Nathan Hamm on 3/13/2012 · 6 comments

Results from last week’s municipal elections in Osh have not yet been certified. The Central Elections Committee has said it is still investigating reports of violations and that results in some precincts may yet be invalidated. Regardless, it says that possible invalidation of results will not change the outcome of vote that saw Uluttar Birimdigi, a party supporting current mayor Melisbek Myrzakmatov and formed out of a union between Ata Jurt and Butun Kyrgyzstan, two parties primarily supported in Kyrgyzstan’s southern provinces. The CEC has said that Uluttar Birimdigi will receive 22 of the 45 seats on Osh’s city council with one yet to be distributed pending final results of the vote counting. This means that Myrzakmatov is all but certain to continue as Osh’s mayor.

At the macro-level, Kyrgyzstan is often characterized as politically divided between north and south. As close watchers of the country know, however, this over-generalizes the social and economic divisions that manifest in Kyrgyzstan’s politics. Nevertheless, there is evidence that it is an accurate characterization of national politics. After last year’s presidential election, Fredrik Sjoberg mapped Almazbek Atambaev’s performance at the rayon level, showing an abundantly clear difference in his performance in northern and southern regions.

The same approach can be used at lower levels, highlighting the geography of political divisions within cities and rayons and providing a richer picture of political competition in Kyrgyzstan. The map below shows the city of Osh and surrounding villages in Kara-Suu rayon.1 Red areas are villages and electoral precincts within the city of Osh and villages that fall under the jurisdiction of city, which, like Bishkek, is essentially a province-level entity.2 The orange areas are villages in Kara-Suu rayon.

Osh city electoral precincts (red) and provincial Osh villages surrounding the city (orange). Map imagery (c) Google Imagery and TerraMetrics 2012.

The distinction between provincial Osh and the city of Osh is important, as the inclusion of rural villages southwest of the city and exclusion of other nearby villages generally puts Kyrgyz who live in villages near the city in the Osh city electorate and generally puts Uzbeks outside of it. The map below shows the ethnic composition of the city and surrounding villages.3

Ethnic composition of Osh and surrounding Kara-Suu Rayon. Darker reds indicate higher proportions of ethnic Kyrgyz while blues indicate ethnic minorities, almost exclusively Uzbeks. Territories under the jurisdiction of Osh city are bordered in yellow. Map imagery (c) Google Imagery and TerraMetrics 2012.

During the 2010 parliamentary and 2011 presidential elections, there was a close relationship between ethnicity, specifically Uzbeks, and results of voting in and around Osh.4 In 2010, the performance of Ar Namys looks very similar to a map of ethnic distribution.

Performance of Ar Namys in 2010 Parliamentary Voting in Osh. Reds represent lower performance while blues indicate higher performance. Map imagery (c) 2012 Google Imagery and TerraMetrics.

In 2011, Atambaev received some of his highest concentrations of votes in Uzbek villages and the heavily Uzbek Amir Timur and Turan sections of Osh.

Performance of Almazbek Atambaev during the 2011 Presidential voting in Osh. High performance is in blue and low performance is in red. Map imagery (c) 2012 Google Imagery and TerraMetrics

These last two maps show that, at least over the course of a couple years, the electoral geography of Osh and its vicinity is fairly stable. Uzbek neighborhoods in Osh — Turan, Amir Timur, and the areas surrounding Suleyman-Too — and Uzbek villages surrounding Osh tended to vote together in 2010 and 2011, first for Ar Namys, then for Almazbek Atambaev. In fact, these relationships are strong enough that the results of voting in Osh city precincts in Alymbek Datka and Manas Ata, the two districts extending to the north east along the M41, may be an indirect indication either of changes in ethnic composition of these neighborhoods since data was last collected or of the unreliability of the data.

Meanwhile, the neighborhoods of Ak-Tilek and Kerme-Too as well as the villages under municipal jurisdiction to the southwest of the city were consistently the least likely to support what might be called “northern” politicians. SDPK, Ata Meken, and Ar Namys all performed poorly in these areas in 2010, as did Atambaev in 2011. Ata Jurt and Kamchibek Tashiev performed fairly well in these areas in 2010 and 2011 respectively, but Adakhan Madumarov and his Butun Kyrgyzstan party were the strongest performers in those elections.

In 2012’s Osh city council election, there is a now familiar geography to the outcome of the vote. Areas where Ar Namys did well in 2010 and Atambaev did well in 2011 are the same areas where Uluttar Birimdigi, the political union of Tashiev’s Ata Jurt, Madumarov’s Butun Kyrgyzstan, and Osh Mayor Melisbek Myrzakmatov, performed most poorly.

Performace of Uluttar Birimdigi in the 2012 Osh Kenesh election. Red indicates higher performance. Map imagery (c) 2012 Google Imagery and TerraMetrics.

Similarly, those areas that had rejected “northern” politicians in the previous two elections, did so again in 2012, especially in the southwestern villages.

Combined performance of parties that are in government in Kyrgyzstan's parliament -- SDPK, Ata Meken, and Respublika -- in the 2012 Osh Kenesh election. Blue indicates higher performance. Map imagery (c) 2012 Google Imagery and TerraMetrics.

It’s worth nothing the ambiguous identity of Respublika, at least in the vicinity of Osh. While it never really dominates in any precinct or village in any election, there isn’t a very clear pattern to the geography of its support the way there is with support for more stereotypically northern or southern parties. It was competitive with Butun Kyrgyzstan and Ata Jurt in 2010 and with both Uluttar Birimdigi and SDPK in 2012. While not an overwhelmingly popular party, it appears to have a strong base of support in Osh that cuts across both the ethnic and urban-rural divide. Looking at its results in 2012 show that its inclusion in the above map as a member of government makes some areas appear to have softened in their opposition to “northern” parties.

Performance of the Respublika party in the 2012 Osh Kenesh elections. Blue indicates higher performance. Map imagery (c) 2012 Google Imagery and TerraMetrics.

While these results may not be entirely unexpected, mapping them clarifies the political and ethnic dynamics of Osh and its vicinity. It demonstrates that at least since the ethnic violence of 2010, there is a stable and distinct geography to the distribution of political constituencies in Osh and that ethnic minorities, and perhaps rural Kyrgyz as well, are fairly cohesive political units.

However, it highlights the importance of what is and is not Osh and who controls its destiny. The city, Kyrgyzstan’s “southern capital,” is incredibly important. Until the closure of Uzbekistan’s borders greatly curtailed trade, it was the economic heart of Southern Kyrgyzstan and an important source of consumer goods for Uzbekistan’s Ferghana Valley. It was an important staging point for the protests that eventually overthrew Askar Akaev, and bringing the city under control was one of the post-Bakiev transitional government’s major and ever-present headaches. Melisbek Myrzakmatov, who will continue on as mayor as a result of the March 4 election, is widely rumored to have played a role in planning and executing violence against Uzbeks in June 2010 and of profiting from the reconstruction of the city. He is, unfortunately, also fairly adept at playing political games. He has consistently been able to find and exert the limits of independence Bishkek will tolerate of him.

And so, the victory of Myrzakmatov’s party is extremely important to Kyrgyzstan’s immediate political future. While most of Kyrgyzstan’s political establishment seems to be buying into the idea of competing for resources through the parliament made more powerful by the 2010 constitution (even Tashiev has one foot in on playing by the new rules), Myrzakmatov’s victory is a potential blow to the parliamentary experiment. Both Madumarov and Myrzakmatov have every incentive to play politics by the old rules and invest in public constituencies to mobilize against the government and their rivals, and Tashiev, as their friend in parliament, can play things both ways. Whether they intended it or not, their electoral coalition bought them political space and have kept one corner of the post-Bakiev political order still unconsolidated.

They benefit, however, from the boundaries that define Osh’s electorate. Were the Uzbek villages on the immediate outskirts of Osh under municipal jurisdiction, the outcome would likely have been quite different, with an SDPK-Ata Meken-Respublika coalition government all but assured. It will be interesting to see how, if at all, these parties respond to this election. By banding together, they would likely be a far more potent political force, at least in Osh, though it is unknown whether or not past Respublika supporters would carry over to such a coalition. Or, if Osh is the prize most worth winning in the south, these parties, which are also in control of the national parliament, could always adopt a trick developed to its perhaps highest form in U.S. state legislatures, and just redraw Osh’s electoral boundaries.


Protesters in Osh are disputing the outcome of the vote saying that Uluttar Birimdigi won 23 of 45 seats on the city council. The Central Election Committee says that they are sorely mistaken and that the party only won 21 seats. Additionally, the CEC says that it invalidated the results of three precincts, and protesters are demanding voting be conducted again in these areas. Uluttar Birimdigi won more than 50% of the vote in two of these precincts and took about 40% in the third, so voting again likely would help the party reach the goal of a majority on the council.

In the comments, there is some discussion about being able to follow the maps. Hopefully the animated map below is a bit easier to follow. Click to view the animation.

Voting for Ar-Namys in 2010, Atambaev in 2011, and Uluttar Birimdigi in 2012 showing geographic stability in the electorate. Areas in which over 50% of the population is Uzbek are highlighted. Click to view animation.

Additionally, there is discussion about turnout levels. This map shows the changes 2010-2012.

Turnout in the 2010, 2011, and 2012 elections. Higher turnout is blue. Lower turnout is red. Click to view animation.

Image at top shows performance of Almazbek Atambaev during the 2011 presidential election in Osh city and Kara Suu rayon. Blue indicates higher performance. Map imagery (c) 2012 Google.

  1. Several villages northwest of the city near and beyond the old airport are not displayed. Because of name changes since independence and the differences between legal names and those used colloquially, especially by Uzbeks, I have been unable to identify which villages are which in this area, and they have been left off of the map at present.
  2. With a few exceptions in the immediate vicinity of Osh, it is difficult to map individual electoral precincts within or shared between villages with data available in print or online from outside the country. It is primarily for that reason that electoral results are aggregated at the village level in these maps. An additional benefit is that it allows investigation of relationships between electoral outcomes and other data, particularly that found in the 1999 and 2009 censuses, that is aggregated at the village level.
  3. Ethnic population figures are based on 1999 and 2004 data. While these figures are likely somewhat dated, the greatest changes have probably come inside the city of Osh, where the data available were least reliable.
  4. Though I and a colleague who is working on this project with me have not yet had an opportunity to confirm this observation, it appears that this relationship diminishes the farther a village is from Osh. It may be that the voting behavior near Osh is being influenced both by 2010’s ethnic violence, which primarily took place in Osh, as well as the ongoing tensions between Uzbeks in the city.

Subscribe to receive updates from Registan

This post was written by...

– author of 2991 posts on 17_PersonNotFound.

Nathan is the founder and Principal Analyst for Registan, which he launched in 2003. He was a Peace Corps Volunteer in Uzbekistan 2000-2001 and received his MA in Central Asian Studies from the University of Washington in 2007. Since 2007, he has worked full-time as an analyst, consulting with private and government clients on Central Asian affairs, specializing in how socio-cultural and political factors shape risks and opportunities and how organizations can adjust their strategic and operational plans to account for these variables. More information on Registan's services can be found here, and Nathan can be contacted via Twitter or email.

For information on reproducing this article, see our Terms of Use


Chris R March 14, 2012 at 2:44 am

Great, informative post Nathan. Also interesting to note that turnout for the Osh municipal elections was much higher than turnout in the same city for presidential elections, suggesting perhaps that the city feel there is more at stake. Karakol and Tokmok had higher turnouts for the presidential elections than for their respective municipals.

Respublika does have a southern nationalist underbelly if you look at the party list, but the general point that the N/S division can be over-hammed is one well made. Any vote can be bought of course, and the Respublika machine was surely the best funded during the parliamentaries.

Nathan Hamm March 14, 2012 at 11:48 am

Chris, thanks for commenting. You’ve reminded me that I wanted to look a bit more closely at turnout changes over the course of these three elections.

Your point on vote buying is a good one too. Even if voting were hopelessly corrupt, the value of this kind of project is that it shows the geography of different parties’ organizational capacities for winning elections.

Jay Ulfelder March 14, 2012 at 5:58 am

I learned a lot from your analysis in prose, but I found it very hard to get information out of the maps and am not convinced they’re the best way to present your data. There are so many comparisons we have to make in our minds within and across maps (e.g., scalar variation, strength of association), and then there are some comparisons the maps invite us to make that aren’t relevant to your analysis (e.g., spatial size of districts). I got lost quickly, and even when I tried hard to make specific comparisons, I wasn’t confident that I could do so accurately. FWIW, I think simple scatterplots would be much more useful here, alongside if not in place of most of the maps.

Nathan Hamm March 14, 2012 at 11:45 am

Jay, thanks for the comment.

As I like to put it, map lie. You’re right that they do a poor job of clearly illustrating, for example, the relationship between the Uzbek population of a village and voting for Ar Namys. And they are hard to compare. I am using a new-to-me GIS that required me to do way more data prep and manipulation by hand than I’m used to, and in the end, I cut corners with presentation.

One of the problems for analysis encountered in this project is that at the city level, our ethnic and voting data aren’t at the same scale. The way around this would be to aggregate votes up from the precinct level to the “territorial council” level at which we have some (less than entirely reliable) ethnic data. We lose the precinct level view of the results, though. It doesn’t come out here because, well, I didn’t take the time, but you can see some of the same relationships between ethnicity and voting at the sub-territorial council level.

Grace Zhou March 15, 2012 at 12:17 pm

This is a really interesting analysis and I found the maps useful; you do have to study them a bit but I think you can clearly visualize the overlaps in population and voting data. I’d been wondering about this after walking up Navoiy and being told it was now Kara-suu rayon even though we were still fairly close to city-center. Do you know when the rayons were drawn up and if/when there has been redrawing of electorate lines?

Nathan Hamm March 15, 2012 at 1:27 pm

I will hopefully be able to present these a bit more clearly in the future and in a way that shows the boundaries between the city and Kara-suu a bit more distinctly.

I don’t know when the rayons were originally drawn up, but the census has descriptions of all the changes made to them going back to about 1960. What’s relevant to Osh and its immediate vicinity is that little bits of neighboring villages were incorporated in 1960, 1966, and 1968. Almalyk, a small village in Nookat, was placed under municipal subordination, as were parts of Nariman (the long arm stretching to the northeast), Verkhniy Uvam (on the east side near the main highway leading out of town), and Tuleiken (the area south of the city between Turan and the Ak Buura).

In 1985, all the rural villages southwest of the city that are now under municipal subordination were transferred over from Tuleiken village council in Kara-suu to what is now called the Japalak village council (originally 60 Years of Kyrgyzstan) as an administratively subordinate area. Interestingly, they didn’t actually become part of the administrative and territorial border of city until 2004, when some land was also transfered.

I only have info on the electoral precincts back to 2009. There are very few changes, near as I can tell, but then again, in villages, the precincts have all been lumped together at the village level. Within Osh, some very small precincts including only hospitals, factories, etc. were eliminated after the 2010 election.

Previous post:

Next post: