Accountability and Information 1

Final Project Essentials: Fixed Effects

Jeremy Springman

University of Pennsylvania

Global Development: Intermediate Topics in Politics, Policy, and Data

PSCI 3200 - Spring 2024

Logistics

Assignments

  • Today
  • Thursday (3/28)
    • Data Assignment 1 Due
    • Dunning et al. (2019)

Agenda


  1. Jablonski and Seim (2023)
  2. Data Assignment 1 Questions
  3. Final Project Essentials: Fixed Effects

Jablonski and Seim (2023)

Models of political accountability

What is political accountability?

  • Accountability is meant to produce responsiveness
  • Responsiveness means that a government “adopts policies that are signaled as preferred by citizens” (Manin, 1999)
  • If governments fail to adopt citizens’ preferred policies, they will be sanctioned

Assumptions

  • Politicians have information about what citizens want
  • Citizens have information about which policies a politician is responsible for

Models of political accountability

Accountability requires that Voters:

  • Observe credible signal of performance
    • Attribution
    • Benchmarking
  • Update their beliefs
  • Have credible alternatives

Models of political accountability


Accountability requires that politicians:

  • Seek to maintain office
  • Observe citizen preferences
  • Update their beliefs

Models of political accountability

Distributional politics

  • Target public spending to maximize electoral advantage

Responsiveness

  • Maximizing welfare across constituents

Decentalization

  • Both sides have better information at a more local level

Background

Background

Theory

  • Politicians aim to allocate to the community with the highest return on votes
  • This means that allocation is conditional on information about citizen preferences across eligible communities

Citizen utility: \(x_i = c_i(a) - p_i - d_i\)

  • \(c_i(a)\): Utility from investment \(a\)
  • \(p_i\): Utility of voting for a challenger
  • \(d_i\): Cost of voting

Theory

Politician utility:

\[ v_j(a) = \sum_{i=1}^{N} 1(x_i > 0, p_i + d_i \geq 0) \] To allocate on electoral rewards, they need to know:

  • \(c_i(a)\): Utility from investment \(a\)
  • \(p_i\): Utility of voting for a challenger
  • \(d_i\): Cost of voting
  • \(n\): The number of individuals

Theory

Gathering information is costly

  • Time
  • Coordination
  • Varies across constituencies (geography, social proximity)

Alternatives

  • Rely on heuristics
  • Collect information from more accessible places

Theory


By providing information to politicians, we can see how the lack of information affects allocation decisions

Research Design

Information about schools

  • Need: Class size, teacher housing, etc.
    • Target communities with high consumption utilities (\(c_i(a)\))
  • Aid: Existing aid projects at each school
    • Marginal returns to overlapping investments
  • Voting: Vote share at nearest polling station
    • Update on \(\bar{p_j}\)
    • \(d\) determines whether to mobilize swing or base

Research Design


Outcome

  • Allocation of NGO supplies

Heterogeneous effects

  • Transparency
  • Information costs

Research Design

Research Design

Questions

  • What do they mean by ‘factorial assignment’?
    • Each respondent has 1-3 pieces of information
  • In the estimation, how are they using interaction terms?
    • Estimate the impact of information based on provision
    • Estimate heterogeneous effects based on information costs and transparency

\[ P(Y_{nsi} = 1) = \phi(\beta_1z_i + \beta_2t_sz_i + \gamma X_is + e_{nsi}) \]

Description of Information

Findings

Findings

Findings

Findings

Findings

Policy Implications


  • Information gaps may explain disparities in public spending
  • Information can reduce disparities

Data Assignment Questions

Final Project Essentials

jrspringman.github.io/psci3200-globaldev/