Unravelling Voters Perceptions of the Economy

1
Unravelling Voters Perceptions of the Economy

• Dr. Orla Doyle
• UCD Geary Institute

2
Format

• Motivation for analysis
• Stage 1 Economic Perception Formation
• Are economic perceptions objectively or
  subjectively determined?
• EEA Survey Dataset
• Ordered Logit Models of Economic Perceptions and
  results
• Stage 2 Determinants of Vote Choice
• Do economic perceptions or real economic
  conditions influence vote intentions?
• Multinomial Logit Models of Czech Voting
  Intentions and results.
• Conclude

3
Economic Voting Macro Approach Aggregate
Analysis
Economy Macro Variables (UE, Inflation, Growth)
Responsibility Hypothesis
Government Support
But Economies do not vote, people
do (Nannested and Paldam, 2000)
4
Economic VotingMicro Approach Survey Data
Economy
Obser- vation
Voters Perceptions
Objective Perceptions
Subjective Perceptions
Sup- port
Vote Choice
5
Macro versus Micro Divide

• Micro studies are less conclusive than macro
  findings and often find only weak relationships
  between the economy and voting behaviour.
• e.g. Kinder Kiewiet, 1981 Lewis-Beck, 1988
  Brug Franklin, 2000.
• Macro studies use objective economic conditions
• Micro studies use subjective economic perceptions
• ?macro and micro studies which examine the same
  phenomena find diverging results!
• To understand the dichotomy between the 2
  approaches it is necessary to determine the
  relationship between objective economic
  conditions and subjective perceptions of the
  economy.
• ? Must examine the formation of economic
  perceptions

6
Stage 1 Economic Perception Formation

• Economic Perception Formation has received little
  attention in the literature.
• Mainly focuses on United States economic
  perceptions. (Conover, Feldman and Knight, 1986
  1987, Duch, Palmer Anderson, 2000).
• How much do voters really know about the
  economy???
• Voters have very little information about real
  macroeconomic variables (Aidt, 2000 Paldam and
  Nannestad, 2000 Sanders, 2000).
• As the cost of acquiring information about the
  economy is very high given the benefit derived
  when casting a vote, a rational agent will try to
  minmise the effort they exert when formed
  perceptions of the economy (paradox of voting).
• Voters react to subjective, personal evaluations
  of the economy.

7
Aggregate versus Individual Perceptions?

• Most economic voting studies use aggregate
  measures of economic perceptions (Page and
  Shapiro, 1992 Erikson, 2000 Paldam and
  Nannestad, 2000)
• Rational Aggregation purges individual noise
  or random variation associated with survey data.
• An individuals perception is a stochastic
  variable Y Y0 Ys ?
• Y0 is the latent objective perception
• Ys captures differences due to subjective
  factors
• ? is the stochastic component.
• Recent studies (Duch, Palmer, Anderson, 2000)
  (Bartels, 1996) show that individual errors in
  measures of public opinion are systematic not
  random, hence aggregation will not produce
  unbiased aggregate measures. Ys ? 0

8
Are Economic Perceptions Subjectively or
Objectively determined?

• Subjective Factors
• According to Duch, Palmer and Anderson (2000),
  there are 4 sources of subjective heterogeneity
  1. Information, 2. Political Attitudes, 3.
  Socio-economic characteristics, 4. Personal
  economic experiences.
• Objective Factors
• Very few individual studies can include objective
  measures of the macroeconomy as these variables
  are constant across all individuals in the
  surveys.
• This study offers an alternative method- use
  regional economic data! (real economic variable
  yet varies across individuals and time).
• Test whether perceptions are influenced by
  subjective or objective factors

9
Dataset

• 11 Economic Expectations and Attitudes Surveys
  (EEA)
• Time-span 1990-1998
• 9 of the 11 EEA surveys were pooled to create
  three new groups
• E1 comprises of EEA03, EEA04 and EEA05
  (1991-1992)
• E2 comprises of EEA06, EEA07 and EEA 08
  (1993-1994)
• E3 comprises of EEA09, EEA10 and EEA11
  (1996-1998)
• Focuses on
• 1. Attitudes towards economic transformation
• 2. Political attitudes
• 3. Socio-economic characteristics
• Between 1113 and 2084 observations and
  approximately 130 variables in each survey.
• Repetition of Qs enables analysis of changing
  attitudes overtime.
• Untapped source of data.

10
Ordered Logit Models of Economic Perceptions

• Y1 Retrospective Economic Perception Question
• Would you say you are generally satisfied with
  the recent progress of the present economic
  reforms?
• Y2 Prospective Economic Perception Question
• Would you characterise the present period as
  being the beginning of a substantial improvement
  of our economic situation?
• Measured on an ordered 4-category scale ranging
  from
• Definitely Yes, Rather Yes, Rather Not,
  Definitely Not.

11
Independent VariablesX SEi, Ii , PESi,
UEi, WGi
12
Table 5 Ordered Logit Models of Retrospective
Economic Perception Would you say you are
generally satisfied with the recent progress of
the present economic reforms?
13
Table 5 Continued
14
Table 6 Ordered Logit Models of Prospective
Economic PerceptionWould you characterise the
present period as being the beginning of a
substantial improvement of our economic
situation?
15
Table 6 Continued.
16
Summary of Results

• On the basis of their statistical significance
  ideology and perceptions of one personal economic
  situation and education influence both
  retrospective and prospective perceptions
  formation across all 3 time periods.
• Provides evidence for the theory of cognitive
  consistency.
• Overall the socioeconomic factors contribute
  least to the systematic variation, only
  education, female, age and income proving
  significant in specific time periods.
• Income only becomes significant in the later
  surveys.
• Regional Unemployment and Average Wage does not
  reach any level of statistical significance
  (except in E1), hence suggesting that objective
  economic variables do not influence perceptions
  of the economy.

17
Stage 2Determinants of Vote Intentions

• Do economic perceptions or real economic
  conditions influence vote choice?
• Micro economic voting studies cannot determine
  the direct relationship between real economic
  conditions and voting decisions.
• Therefore they rely on perceptions of the economy
  alone and are mainly concerned with determining
  what type of perceptions influence vote choice
  (i.e. Retro vs Prospect, Egotropic vs
  Sociotropic).
• Stage 2 merges the micro and macro approach to
  economic voting by including both economic
  perceptions and regional economic conditions in
  the voting equations.
• Most post-communist economic voting studies are
  Macro analysis.

18
Multinomial Logit Model of Vote Intentions

• Dependent Variable
• If an election to the Chamber of Deputies of the
  Czech Parliament were organised now, which
  political party would you vote for?
• -Civic Democratic Party (ODS)- Base Category
• -Civic Democratic Alliance (ODA)
• -Christian Democratic Party (KDU)
• -Czech Social Democratic Party (CSSD)
• -Communist Party of Bohemia and Moravia (KSCM)
• -Others, includes vote intentions for all other
  small parties.
• Independent Variables
• -Retrospective and Prospective Economic
  Perception
• -Macroeconomic Factors (Regional UE and Wages)
• -Socio-economic Factors (same as before)

19
E1 Multinomial Logit Model of the Determinants
of Czech Voting Patterns
20
E2 Multinomial Logit Model of the Determinants
of Czech Voting Patterns
21
E3 Multinomial Logit Model of the Determinants
of Czech Voting Patterns
22
Summary of Results

• Both retrospective and prospective economic
  perceptions questions influence voting
  intentions.
• The more pessimistic you are about the economy,
  the higher your probability of voting for either
  CSSD or KSCM compared to voting for the
  incumbents.
• Socio-economic factors do not play a major role
  in influencing voting intentions, apart from
  education.
• The inclusion of ideology does not change the
  effect of economic perceptions on voting
  intentions.
• Real economic conditions do not influence voting
  intentions

23
Concluding Remarks

• Previous studies show that voters do not utilise
  objective information about the economy when
  forming their perceptions of the economy. These
  results confirm these findings even in a
  transitional setting.
• Economic perceptions are systematically distorted
  by ideology, personal economic situations and
  various socio-economic factors.
• These findings refute Evans and Andersen's
  assertion that ideology should not bias
  perceptions in transition economies.
• Confirms Duch et al. 2000 findings that micro
  economic voting studies should not rely on
  aggregate economic perceptions are a means to
  avoiding measurement error associated with survey
  data.

24
Concluding Remarks Cont.

• Vote choice is influenced by subjective economic
  perceptions rather than real economic conditions.
  
• While the economy does matter, voters are
  ultimately swayed by their perceptions of the
  economy, which are frequently contaminated by
  ideological preferences and personal experiences.
• May explain why, that even in times of economic
  fortune, incumbent government often fail to be
  re-elected.