Experience Open to Interpretation

US Presidents and Economic Coercion

Dawid Walentek

Ghent University

Puzzle

  • What does Joe Biden make of the CNN news?
  • Does it affect his behaviour in respect to sanctions?

Literature

  • Current research on  the effects of sanctions on approval ratings is at odds (Whang 2011, Webb 2018)
  • We know that high approval ratings are a 'desirable commodity' (Donovan et al., 2020)
  • We also know that many factors drive approval (Wlezien, 2015, 2017)
  • Yet, research assume that US presidents are not biased and able to interpret movement in approval ratings

Theory

  • Formal model of beliefs updating (Fryer et al., 2019)
  • Building blocks of the model:
    • signals open to interpretation
    • beliefs about audience costs for empty threats
    • beliefs about audience benefits for imposition
  • Convergence on preferences for economic coercion
  • Motivated by research on leaders beliefs about public opinion (Walgrave et al., 2023) and susceptibility to confirmation bias (Donovan et al., 2020)

Hypothesis

A political leaders is likely to impose economic sanctions if experiencing a period of lower approval ratings

Research desgin

  • Data
    • TIES
    • Gallup
    • US Bureau of Labor Statistics
  • Complex unit of analysis
    • 731 presidency-months between 1945 and 2005
    • 694 decisions taken in  282 months
    • red indicate at least one imposed sanction
    • 3 types of observations: lags, leads and treatments

Research design

  • Event-study design
    • timing-based (Miller, 2023; Clarke and Schythe, 2020)
  • Baseline one month before imposition
  • Anticipation effect: a change prior to treatment

Results

  • Dots to the left are leads
  • Dots to the right are treatment effects
  • We observe an anticipation effect
    • joint significance test for leads p<.05
  • No audience effect  

Results

  • Placebo test
  • Random assignment of comparable number of treatment months
  • Replicable (using my ORCiD)
  • Random walk

Results

  • Heterogeneity 1: majority imposed vs at least one imposed sanction
    • no difference
  • Heterogeneity 2: majority security vs majority trade
    • anticipation effect present for security & absent for trade
    • US President play a pronounced role in respect to security issues
    • low power of the test
  • Main findings robust to dropping FE, controls & presidents

Conclusion

  • Suggestive evidence for an anticipation effect
    • US presidents act as if domestic audience benefits or costs (or both) are present
  • No evidence for domestic audience effects
    • signal is too weak
    • costs and benefits balance out

Thank you