You are on page 1of 14

CLAIMS VS.

REALITY IN
TURNOUT FIGURES IN
EASTERN EUROPEAN
ELECTIONS

Zachary Fieldhouse & Cecilia Cotero Torrecil


International Election Observation and Survey Methodolo
Proj
WHAT WE’RE LOOKING AT
• Turnout figures
• Why is this important?
• Best indicator of citizens’ participation
• Indicator of legitimacy of the government
• Health of the democracy
• How is it measured?
• Official government estimates
• Election observation mission (EOM) data
PROBLEM AND RESEARCH QUESTION
• Problem is that official turnout figures are sometimes highly
suspicious in some new democracies.
Goals:
• Development of a turnout fraud indicator (formula) based on
observations in polling stations.
• To check whether the turnout match with official figures.
RESULTS
1.Formula is promising but could do with some revision.
• The formula consistently underestimated the turnout levels as compared to
official figures.
2. Discrepancies were minimal in elections that were assessed as either good or
very good.
• Largest discrepancies found in those countries with poorly rated elections
(bad or very bad). Examples include Tajikistan 2015 and Kazakhstan 2011.
OSCE, ODIHR AND DEMOCRACY?
The Organization for Security and Co-operation in Europe
(OSCE):
• founded in 1972 as a ‘multilateral forum for dialogue and
negotiation between East and West’.
• addressing issues such as human rights, democratization, policy
strategies, counter-terrorism and economic concerns.
• 57 participating states from Western and Eastern Europe,
Central Asia and America
• The Office for Democratic Institutions and Human Rights
(ODIHR) is the OSCE’s human rights branch. It promotes
democracy and human rights through observing elections in
member countries.
DOCUMENT OF THE COPENHAGEN
MEETING (1990). . .
• Document which sets the principles for democratic elections in the
OSCE region.
• It includes wide-ranging commitments to hold democratic elections,
respect the rule of law and Human Rights and establish a
representative and accountable government.
• OSCE participant states are committed to it by law. Their obligations
include:
• To hold ‘free elections [. . .] at reasonable intervals by secret ballot.
• To ‘invite observers from any other [. . .] participating States and any
appropriate […]organizations who may wish to do so to observe the
course of their national election proceedings. . .’
WHAT DO OBSERVERS DO?
• Observers are drawn from all participating countries and are ‘deployed in
teams of two’ on election day to observe the ‘opening, voting, or counting
(tabulation) process’.
• Observers monitor and report on the conditions inside and outside the
polling station by filling in lengthy forms (one per polling station
visited).
• The team stays between 30 and 45 minutes at each polling station and
must use this time to introduce themselves to officials and to explain what
they are doing.
• At the end of their assessment, the observers will come to a consensus
opinion as to the overall quality of the election on a four-point scale (very
good, good, bad, and very bad).
BSERVATION FORMS AND LINKS TO TURNOUT

- Four to nine ranges to choose from

- Overall assessment
WHAT WE DID. . .
• OSCE missions since 2009
• In Albania, Tajikistan, Armenia, Georgia, Ukraine, Macedonia, Kazakhstan, Moldova,
Kyrgyzstan, Azerbaijan, and Bulgaria.
• Observation forms - data on turnout figures collected by observers in the 30 to 60 minute
allotted period.
• Extrapolate a turnout figure for each election following a formula.
• Compare our formula-produced turnout figure with the official turnout figure to see
whether there are discrepancies.
HE FORMULA - Median value for each range taken and
multiplied by the number of polling
ALBANIA 2011 stations.
- Sum of the products equals the total
oters per Polling Station (30 min.) Number of Polling Stations
number of observed people who voted in
23
the 30 minute period.
3 51
- (17,676 people)
10 171
-15 168 - Multiply this figure by 24 or 28 to
-20 169 extrapolate the amount for the whole day
-30 218 - (17,676 x 24 = 424,224).
-40 83
-50 27 - Divide by the number of polling stations
50 31 visited by the EOM.
tal 941 - (424,224 ÷ 941 = 450.8225)

Average # Registered Voters per Polling - Divide by average # of registered voters i


Station in EOM Polling Stations: EOM polling stations
- (450.8225 ÷ 890 = 50.65%)
890
DISCREPANCIES WITH OFFICIAL FIGURES
EOM Observer Turnout Official Turnout

Albania 2011 50.65% 50.90% The formula


Albania 2013 52.39% 53.50% consistently
Armenia 2012 46.42% 62.87% underestimated th
Azerbaijan 2010 39.64% 49.56% turnout levels as
Bulgaria 2013 43.14% 51.30%
compared to officia
Tajikistan 2015 29.84% 87.80%
Georgia 2010 47.15% 47.70%
figures.
Georgia 2012 46.74% 59.76%
Ukraine 2012 39.51% 57.50%

Ukraine 2014 (Parl) 34.29% 52.40%

Ukraine 2014 (Pres) 43.89% 60.80%

Macedonia 2011 57.76% 63.50%


Kazakhstan 2011 35.69% 89.98%
Moldova 2009 42.09% 57%
Moldova 2010 45.59% 63.35%
Kyrgyzstan 2009 48.79% 79.13%
REASONS FOR DISCREPANCIES
• Fraud and manipulation of turnout figures
• Requirements in national law
• Legitimacy and/or funding

• Amount of time that observers spend during their 30 minutes speaking to


officials and not counting voters.
• Another possible reason for the underestimation of turnout may be due to the
choice of polling stations by the OSCE.
• An overrepresentation of urban polling stations in EOMs would invariably skew the findings.
(Ukraine 2012 – 1,071 people per polling station vs. 1,598 people for observed polling
stations).

• Problems with limited number of ranges that observers can select


• Creates underrepresentation where large numbers of people vote and observers can only
select the >50 category.
RECOMMENDATIONS
• Add another range category to the form to allow better estimates for polling
stations with large numbers of voters (>50).
• Calculate the results for urban and rural stations separately. Take into account
that urban polling stations are over-represented. This assures that bias in
sample choice does not influence the results.
• Assign the task of counting voters to one of the two observers for the duration
of the 30 minute observation.
• More data cleaning is necessary. Change implausible data.

Further research into this area is necessary including revision of the formula.
THANK YOU FOR YOUR ATTENTION!

ANY QUESTIONS?

You might also like