## 4.4 Pandemics

## Source details

**Canonical URL:** [4.4 Pandemics](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024195-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024195-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024195-print-pdf.pdf.json)

---

### Role within the paper
- Section 4 surveys personal experiences of non-economic aggregate shocks, explicitly including pandemics alongside migration, wars, terrorist attacks, and natural disasters.
- Pandemics are examined as aggregate shocks that can leave long-term legacies on preferences and beliefs and are situated within a broader inquiry into how such shocks shape preferences and beliefs across generations.

### Key themes and framing
- Large aggregate shocks (e.g., the Global Financial Crisis, the COVID-19 pandemic) are rare within a lifetime but can leave long-term legacies on preferences and beliefs.
- Direct economic and social scarring from such shocks can be large and persistent; effects on preferences and beliefs may be even more long-lasting.
- The survey draws on interdisciplinary theory (sociology, social psychology, economics) and empirical work using large surveys and new datasets to analyze impacts, with attention to timing (impressionable years versus later life) and persistence.

### Generalized trust, trust in institutions, and trust in science
- Empirical findings on trust vary with context and government response; both eroded and strengthened trust outcomes are reported depending on setting and response quality.
- Key quantitative results:
  - Aassve et al. (2021): A 1 death per thousand increase in influenza mortality is associated with a 1.4 percentage point decrease in generalized trust.
  - For countries neutral in World War I, a 1 death per thousand increase in influenza mortality is associated with a 2.2 percentage point decrease in generalized trust.
  - Aksoy et al. (2024) / Aksoy et al. (2021):
    - Experiencing a pandemic during the impressionable years erodes trust in institutions and in scientists.
    - The individual with the highest exposure to a pandemic has, on average, 5.1 percentage points less confidence in the national government than those with no exposure (long-term effect).
    - The individual with the highest exposure to epidemics is, on average, 11 percentage points less likely than those with no exposure to trust scientists; results driven by people with little or no science education.
    - Changes in trust in scientists are associated with negative views toward vaccines and lower rates of child vaccination.
  - Fluckiger et al. (2019): An increase in average Ebola prevalence of one case per 100,000 people raises post-epidemic trust in parliament, the president, and the police by around 0.15 standard deviations compared to the pre-epidemic era (short-run estimate).
  - Bol et al. (2021): COVID-19–related lockdowns were associated with a three percent increase in trust in government in the short term; lockdown increased satisfaction with democracy and trust in government by about three percent and support for the party of the prime minister/president by about four percent.

### Political preferences and preferences for redistribution
- Fewer studies focus on political preferences and redistribution after pandemics; available evidence shows heterogeneous effects across contexts, election types, and short-run versus long-run horizons.
- Key quantitative results:
  - Campante et al. (2024) — Ebola scare in the United States (2014):
    - A one standard deviation increase in Ebola concerns decreased the Democratic vote share by four percentage points in House elections and by three and four percentage points in Senate and gubernatorial elections.
    - A one standard deviation increase in Ebola searches decreased voter turnout by about 1.4 percentage points.
    - Increased negative views about immigrants observed among voters.
  - Baccini et al. (2021) — COVID-19 and county-level voting change between 2016 and 2020:
    - A county with 100 more COVID cases per 10,000 people reduced its Trump vote share from 2016 to 2020 by an additional 0.12 percentage points on average.
    - Authors’ counterfactual claim reported: if cases of COVID had been lower than five percent, Trump would have likely won reelection (presented as in the source).
  - Rees-Jones et al. (2022) — U.S. attitudes toward safety nets (June 2020; 2,516 respondents):
    - Respondents living in a county with a one standard deviation higher number of COVID-19 deaths are 8.4 percentage points more likely to support long-term unemployment insurance expansion.
    - The same respondents are 9.2 percentage points more likely to support long-term expansions to government-provided healthcare.
- Cross-study summary:
  - Heterogeneity across contexts is substantial: effects depend on prevailing social and economic context, government response, and whether analysis captures short-term versus long-term impacts.
  - Few studies examine variation by life period of exposure; where examined, exposure during impressionable years produces stronger, longer-lasting effects.

### Mechanisms and life-period heterogeneity
- Mechanisms identified in the broader literature that apply to pandemics include:
  - Experience-based belief updating (impressionable-years effects): exposure during adolescence/early adulthood tends to have stronger and more persistent impacts.
  - Government competence and response framing: robust government responses can increase trust (Fluckiger et al., 2019; Bol et al., 2021), whereas weak responses are associated with long-term declines in institutional trust (Aksoy et al., 2024).
  - Science exposure and education: erosion of trust in scientists concentrated among those with little or no science education, with downstream effects on vaccine attitudes and child vaccination rates (Aksoy et al., 2021).
- Timing matters: several studies highlight stronger scarring when pandemics are experienced during the impressionable years.

### Policy-relevant implications
- Policymakers should account for generational heterogeneity: pandemics disproportionately affecting specific cohorts can generate long-term shifts in political alignment, trust in institutions, and trust in science.
- Government response quality matters:
  - Robust, credible responses can bolster short-run trust in government and satisfaction with democracy (Fluckiger et al., 2019; Bol et al., 2021).
  - Weak or poorly communicated responses can produce long-term erosion of trust in government and scientists (Aksoy et al., 2024; Aksoy et al., 2021).
- Public-health messaging and science education are critical: declines in trust in scientists are linked to vaccine hesitancy and lower child vaccination rates, suggesting investments in scientific literacy and trusted communication channels can mitigate persistent adverse effects.

*Source: 4.4 Pandemics, IMF Working Papers — Aggregate Shocks and the Formation of Preferences and Beliefs.*

### 4.4 Pandemics ..........................................................................................................

### 4.4 Pandemics

### Role of pandemics within the paper's scope
- Section 4 reviews the literature on personal experiences of non-economic aggregate shocks, explicitly including migration, wars, terrorist attacks, pandemics, and natural disasters.
- The survey situates pandemics within a broader investigation of how aggregate shocks affect the formation of preferences and beliefs across generations.

### Key themes from the introduction relevant to pandemics
- Large aggregate shocks (e.g., the Global Financial Crisis, the COVID-19 pandemic) are rare within a lifetime but can leave long-term legacies on preferences and beliefs.
- The paper emphasizes that the direct economic and social scarring from such shocks can be large and persistent, and that effects on preferences and beliefs may be even more long-lasting.
- Historical analogues are used to motivate the inquiry: the Great Depression’s long-term effects on beliefs and attitudes and more distant episodes such as hyperinflation shaping national aversions to inflation.

### Findings on shocks and preference/belief formation (general findings applicable to pandemics)
- Aggregate economic shocks permanently change various types of preferences and beliefs, including:
  - political preferences,
  - risk preferences,
  - trust in institutions.
- The effect of shocks experienced in young adulthood is stronger and longer lasting.
- Understanding the origins of preferences and beliefs is crucial for policy design because beliefs influence demand for government spending, taxation, and other policy responses.

### Theoretical and empirical frameworks informing pandemic analysis
- The survey draws on multiple disciplines to explain how shocks like pandemics may shape preferences and beliefs:
  - Sociology: theory of generations and formation of collective memories.
  - Social psychology: age periods relevant for formation of preferences and beliefs.
  - Economics: models of shocks as belief-twisting events and learning-from-experience frameworks.
- Empirical work uses large surveys and new datasets on individuals’ preferences to systematically analyze impacts.

### Policy relevance (implications for pandemic-related policy)
- Policymakers should recognize that shocks such as pandemics disproportionately affect specific generations, leading to potential long-term shifts in preferences and political alignment.
- Political actors and policymakers may need to tailor policies to address generational impacts from pandemics, given generational differences in values and beliefs formed by such shocks.

*IMF Working Papers — Aggregate Shocks and the Formation of Preferences and Beliefs (excerpt including introduction and section indexing; reference to Ronald Brownstein, “Is Gen Z Coming for the GOP?” The Atlantic, June 8, 2023).*

### conclusions and highlights avenues for future research. Table 1 provides a summary of the papers reviewed in

### wpiea2024195-print-pdf - conclusions and highlights avenues for future research. Table 1 provides a summary of the papers reviewed in

### II. Historical Experiences and the Formation of Preferences and Beliefs
- Standard models assume stable preferences (Stigler and Becker, 1977), but historical examples and behavioral/psychology evidence indicate experiences can alter preferences and beliefs (Lowenstein and Angner (2003); Malmendeier and Nagel (2011); Giuliano and Spilimbergo (2023b); Stevenson and Wolfers (2011); Algan et al. (2017); Guiso et al. (2024)).
- Four approaches reviewed to explain long-lasting effects of shocks:
  - Shocks as belief-twisting events
    - Friedman and Schwartz (1963: 673): Great Depression “shattered beliefs in a ‘new era’.”
    - Cogley and Sargent (2008): Great Depression altered confidence in common beliefs, generating pessimism; persistence arises from slow Bayesian updating from macro data.
    - Core idea applies beyond the Great Depression to other extreme events.
  - Experience-based learning (Malmendier and Nagel, 2011, 2016)
    - Individuals overweight lifetime-experienced data and underweight distant past.
    - Two implications: strength of updating depends on age not time; young people update beliefs more strongly than older ones.
    - Empirical links to stock-market participation, inflation expectations, consumption, home ownership, mortgage choice.
  - Formative-age hypothesis (three variants)
    - Impressionable-years hypothesis: core attitudes formed mostly in early adulthood; evidence of socialization between ages 18 and 25 (Krosnick and Alwin, 1989).
    - Increasing-persistence hypothesis: malleability decreases gradually with age (Glenn, 1980).
    - Lifelong-openness hypothesis: people remain highly malleable throughout life (Brim and Kagan, 1980).
    - Distinction: lifelong-openness vs. impressionable-years/increasing-persistence which emphasize young adulthood.
    - Neuroscience support: grey matter volume dynamics; “synaptic tagging and capture” (Frey and Morris, 1997) — emotionally powerful experiences induce stable synaptic connections; young brain may form stronger, more lasting connections.
  - Sociological theory of generations (Mannheim, 1952; Bourdieu, 1993)
    - Cohort experiences in youth create generational character; suggested age range around 17 to 25 for formation of generational differences.
    - Survey evidence (Schuman and Scott (1989)): peak recollection ages 15–27 for major events, consistent with impressionable-years.
- Synthesis: disciplines emphasize role of aggregate shocks in forming preferences and beliefs, differing on timing (formative age vs. recency) and persistence.

### III. What preferences and beliefs are altered by aggregate economic shocks?
- General note: hard to empirically distinguish preferences from beliefs; many studies emphasize one interpretation over the other; survey measures often conflate them (Malmendier (2021b)).

H3: 3.1 Recessions

- 3.1.1 Political preferences and support for redistribution
  - Krosnick and Alwin (1989): panels divided into age groups (18–25, 26–33, ...) show greater malleability of political preferences in early adulthood; sample sizes limited.
  - Giuliano and Spilimbergo (2023b)
    - Identification: regional variation in economic conditions where a person grew up; data from General Social Survey linking preferences to past regional macroconditions at age 16.
    - Findings for the US and World Value Survey:
      - Experiencing a recession when young → more likely to vote for a right-leaning party and less supportive of government redistribution.
      - Effect period includes adolescence before 18 and remains stable afterwards (supports impressionable-years).
      - Effect magnitudes: similar to effect of years of education and to effect of having annual family income rise from $1,000–$2,999 to $25,000 or more; 50 percent larger than the effect of being unemployed.
  - Carreri and Teso (2023)
    - Sample: US members of Congress born after 1911; exposure measured by state macroeconomic conditions where member grew up.
    - Finding: experiencing a recession during early adulthood → ~0.15 standard deviation increase in ideology score (more conservative).
    - Within-party comparison: 17 percent of a standard deviation change, arguing against selection into party as sole mechanism.
  - Algan et al. (2017)
    - European regions, 2009–2010 financial crisis; 2SLS using pre-crisis share of construction as instrument for regional unemployment.
    - Finding: a one-percentage-point increase in regional unemployment increases the populist vote by one percentage point; severity drives support for populists.
  - De Bromhead et al. (2013); Funke et al. (2016)
    - De Bromhead et al.: 171 elections across 28 countries, 1919–1939; one standard deviation decline in growth increases fascist vote share by 1.75 percent; in countries without prewar democracy and preexisting fascist parties, a one standard deviation increase in growth associated with 4.8 percent decline in extreme right vote share.
    - Funke et al.: 20 countries, 1870–2014; in five years following financial crises, far-right vote shares increased by about 30 percent.
  - Abramitzky et al. (2023)
    - Kibbutzim, Israeli electoral data 1977–2019; members experiencing severe financial crisis 1989–1996 moved politically right; lower support for socialist parties and increased support for liberalized labor markets.
    - Age heterogeneity: short-term effects similar across ages; persistent effects only for younger cohorts (consistent with impressionable-years).
  - Summary conclusion: economic shocks consistently move political preferences to the right across contexts, ages at exposure matter (supporting impressionable-years), and effects persist across time and regimes.

- 3.1.2 Prosocial preferences
  - Ambiguous theoretical direction: recession could increase altruism or increase selfishness and in-group protection.
  - Identification: follow Giuliano and Spilimbergo (2023b) strategy—regional exposure while growing up.
  - Cotofan et al. (2024)
    - US General Social Survey and regional income while growing up.
    - Finding: lower regional income during impressionable years → less openness to immigrants.
    - Magnitude: a log-point decrease in regional income increases anti-immigrant sentiment by 0.4 points on a five-point scale.
    - Effect larger than that of a log-point change in household income or of being unemployed; stronger for less educated respondents or those whose fathers are less educated.
  - Bietenbeck et al. (2023)
    - Global Preference Survey (approximately 65,000 respondents in 75 countries).
    - Prosocial index via principal-component analysis of trust, altruism, positive reciprocity, negative reciprocity.
    - Finding: experiencing a recession during impressionable years decreases prosocial preferences for the rest of life.
    - Trait-specific: trust and altruism not significantly affected; negative reciprocity not significantly affected; largest malleability in positive reciprocity.
    - Magnitude: effect comparable to being a man rather than a woman.
  - Summary: empirical literature consistently finds recession experience decreases prosocial preferences.

- 3.1.3 Generalized trust and trust in institutions
  - Stevenson and Wolfers (2011)
    - US time series (1972–2010), cross-country 2006–2010, and US state-level unemployment variation 1972–2010.
    - US finding: a 5-percentage point increase in unemployment correlated with a 13.5 percentage point decline in share expressing confidence in banks, and a 4.5 percentage point decline for trust in Congress, major companies, and business executives; newspapers and Supreme Court effects less robust.
    - Cross-country: confidence in financial institutions and national government is very cyclical; stronger results for OECD countries.
  - Algan et al. (2017)
    - European regions, 2009–2010 crisis; 2SLS with pre-crisis share of construction instrument.
    - Finding: one standard deviation increase in unemployment reduces trust in parliament by 3.6 percentage points; similar decline for trust in politicians; smaller effect for trust in legal system; trust in police moves opposite direction.
    - Generalized trust remains unaffected.
    - Heterogeneity: effects stronger for respondents without a college degree and for older cohorts (smaller confidence intervals).
  - Ananyen and Guriev (2019)
    - Russian regions during 2009 crisis; instrumented using industrial structure.
    - Finding: a 10 percent decrease in income associated with a 5 percentage point decrease in social trust.
  - Summary: negative shocks correlate with reduced trust in political institutions across contexts; stronger among less educated individuals—consistent with populism literature.

- 3.1.4 Risk preferences
  - Malmendier and Nagel (2011)
    - Learning-from-experience model using Survey of Consumer Finances (1960–2007); link risky behavior to lifetime-experienced stock and bond returns weighted by age-based scheme declining linearly to zero at birth year.
    - Findings:
      - Going from the 10th to the 90th percentile of experienced stock returns implies:
        - 10.3 percentage point higher probability of being in higher risk-tolerance categories.
        - 10.2 percentage point increase in probability of stock-market participation.
        - 7.9 percentage point increase in asset allocation to stocks.
      - For bond market: same change in experienced bond returns implies 11.4 percentage point increase in probability of bond-market participation.
      - Beliefs vs. preferences: evidence leans toward belief-based channel—experienced returns predict expected future returns (a 1 percentage point increase in experienced return associated with a 0.5–0.6 percentage point increase in expected return); differential effects by asset type argue against pure preference change.

H3: 3.2 Inflation experiences
- D’Acunto et al. (2021a)
  - Dataset: almost 100,000 households in Kilts Nielsen Consumer Panel; new survey data June 2015 and 2016 on experienced and expected inflation.
  - Finding: strong correlation between expected inflation and household CPI inflation, especially “frequency CPI.”
  - Frequency CPI: association with expected inflation 20–50 percent larger than household CPI; households pay more attention to price increases than decreases.
  - Explained variation: R2 from past observed price changes on individual inflation expectations = 10 percent; including geography, age, education increases R2 (split by age and education R2 increases to 25 percent).
  - Evidence holds back to 1953 (Malmendier and Nagel, 2016).
- D’Acunto et al. (2021b)
  - Gender differences in inflation expectations explained by grocery-shopping exposure: within household, main grocery shopper has higher inflation expectations; controlling for main grocery shopper makes gender coefficient insignificant.
- Summary: learning-from-experience applies strongly to inflation expectations; frequency of encountered price changes matters; evidence supports belief-updating channel.

H3: 3.3 Trade shocks

- Identification challenges: trade is endogenous to preferences and political systems; two main empirical strategies: cross-regional variation in endowments/specialization and exogenous trade shocks.

- 3.3.1 Political preferences
  - Autor et al. (2020)
    - Local exposure to Chinese imports instrumented with Chinese exports to other developed countries.
    - Findings:
      - 1 percentage point increase in trade exposure → 5.2 percentage point increase in Fox News market share (2004–2012).
      - Districts with more exposure: 13.2 percentage point higher likelihood of electing a conservative Republican (2002–2016).
      - Interquartile range increase in import penetration → Republican vote share increased by one percentage point between 2000 and 2008.
      - Campaign contributions: 35 percent increase in both left-leaning and right-leaning donors over 2002–2010; 54.8 percent higher campaign contributions from both liberal and right-leaning donors over 2002–2016.
    - Conclusion: greater trade exposure induces net shift toward right-wing candidates and increases polarization.
  - Colantone and Stanig (2018c) and country studies (Italy, France, UK, Germany)
    - Chinese import shock associated with increased votes for nationalist/far-right parties; similar Brexit and nationalistic outcomes in various country studies.
  - Feigenbaum and Hall (2015)
    - Trade competition per worker: $1,000 increase leads to 0.7 percentage point decrease in probability that district’s representative casts a “right-leaning” or “free-trade” vote in Congress (indicating more protectionist behavior).
  - Methodological note: most papers do not analyze life-cycle differential effects.

- 3.3.2 Preferences toward democracy and authoritarian preferences
  - Tabellini and Magistretti (forthcoming)
    - Instrument: improvements in air transportation relative to sea since 1960 (Feyrer (2019) approach); sample 74 countries, 1995–2020.
    - Finding: doubling exposure to economic integration with democracies increases individual attitudes towards democracy by .58 points on a 1 to 4 scale.
    - Impressionable-years evidence: peak effect between ages 20 and 28; effect persistent.
    - Trade with autocracies → no effect.
    - Mechanism: imports (higher-quality goods; trust-intensive goods) primarily drive transmission of pro-democracy attitudes.
    - Macroscopic effect: 80 percent increase in integration with democracies increases Polity2 score by roughly four points.
  - Ballard-Rosa et al. (2021)
    - UK survey 2017 (1,913 respondents); instrument Chinese imports to US for UK imports.
    - Finding: two standard deviation increase in import penetration increases authoritarian preferences by one third of a standard deviation; results driven by aggression component; leads to a four percentage point increase in voting for Brexit.
  - Colantone and Stanig (2018b)
    - European Values Survey 1988–2008; instrument Chinese imports to US for imports to Europe.
    - Finding: one standard deviation increase in income shock decreases preferences for liberal democracy/democracy by .058/.11 standard deviation; stronger effects for less educated people.
    - No robust result for “authoritarian personality” as proxied by values transmitted to children.
  - Summary: trade exposure can shape attitudes toward democracy and authoritarianism; effects depend on partner regime type (democracies vs autocracies), age at exposure (impressionable years), and education.

### Cross-cutting conclusions and research directions (high-level synthesis from reviewed literature)
- Timing matters: a consistent body of evidence supports the impressionable-years hypothesis—exposure during adolescence/early adulthood (roughly ages 16–28 in many studies) has persistent effects on political attitudes, prosocial preferences, trust, and democratic values.
- Channels: many findings point toward experience-based belief-updating rather than pure preference mutation, particularly for financial market participation and inflation expectations; disentangling beliefs versus preferences remains empirically challenging.
- Heterogeneity: effects are often stronger for less educated individuals and for younger cohorts; severity and type of shock matter (e.g., severity of unemployment increases populist vote; imports from democracies promote pro-democracy attitudes).
- Consistent empirical patterns:
  - Recessions and financial crises → political shift to the right, increased support for populist/extreme parties, reduced prosocial preferences, reduced trust in political institutions.
  - Inflation experiences → higher inflation expectations driven by frequently encountered price changes; grocery-price exposure explains household gender differences.
  - Trade shocks → increased political polarization and right-leaning outcomes in many contexts; trade with democracies transmits pro-democracy attitudes, while some import shocks are associated with increases in authoritarian preferences and anti-immigrant sentiment.
- Methodological note: robust identification strategies (regional exposure, instruments exploiting external supply shocks, transport-mode shifts to instrument integration) are central to isolating causal effects.

*IMF WORKING PAPERS — Aggregate Shocks and the Formation of Preferences and Beliefs (conclusions and highlights; selected avenues for future research summarized from the reviewed section)*

### conclusions are that trade shocks do affect preferences and beliefs. The effect on beliefs depends on the

### wpiea2024195-print-pdf - conclusions are that trade shocks do affect preferences and beliefs. The effect on beliefs depends on the

### Trade shocks, preferences, and beliefs
- Trade shocks do affect preferences and beliefs; the effect on beliefs depends on the context and on the interplay with identity.
- Evidence generally finds trade shocks associated with a movement towards the political right.
- Most papers indicate the effect of trade on preferences and beliefs is mediated by cultural rather than economic concerns.
- Most papers study short-term effects of trade exposure; exception: Tabellini and Magistretti (forthcoming) find strong and permanent effects of trade on preferences for democracy if the shock is experienced during the impressionable years.
- Margalit (2012) finds that people exposed to a cultural prime had a 14 percent more negative view of economic integration than those in the control group; there was no treatment effect among respondents exposed to the libertarian treatment.
- Ferrara (2023) and Strain and Veuger (2022) provide evidence that the political impact of trade is mediated by attitudes toward immigrants and racial/ethnic minorities.

### Aggregate economic shocks and the rightward shift in political preferences
- Recessions are associated with permanently lower earnings (Oreopoulos et al., 2012; Von Wachter, 2020), yet aggregate shocks often correlate with a rightward shift in political preferences.
- Standard income channel would predict a leftward shift because lower income correlates with stronger desire to redistribute (Meltzer and Richard, 1981); observed rightward shifts imply countervailing mechanisms.
- Proposed mechanisms explaining rightward shifts:
  - Early hardship strengthens materialism and desire for security (Inglehart, 1990); evidence: Cotofan et al. (2020) find lower regional income while growing up is associated with a weaker preference for a meaningful job and a stronger preference for a high-income job.
  - Recessions can make individuals more selfish and less concerned with social equality than with economic efficiency (Fisman et al., 2015).
  - Identity politics: shocks reinforce group identity, increase support for nationalist economic policies, reduce support for redistribution (Shayo, 2009; Bonomi et al., 2021; Grossman and Helpman, 2021).
  - Scapegoating: crises increase blaming of minorities/immigrants (Bauer et al., 2021; Cotofan et al., 2024); social sanctions against anti-minority behavior are lower during economic crises (Bursztyn et al., 2022, 2024).
  - Fiscal concerns: highly educated politicians may view government responses as wasteful if they suffered less from the recession (Carreri and Teso, 2023; Thal, 2017).
- Summary insight: adverse economic shocks move preferences to the right despite income-preferences correlations; factors include desire for stability, sense of community, identity politics, scapegoating, fear of fiscal burden, and the narrative/ideology framing shocks.

### Migration shocks: short-run backlash and long-run integration effects
- Review conclusions (Rodrik, 2021; Alesina and Tabellini, 2024):
  - Short run: immigration typically triggers natives’ backlash, increases support for anti-immigrant and populist parties, and lowers preferences for redistribution.
  - Backlash is higher when immigrant flows increase rapidly, immigrants are low-skilled, come from different racial/ethnic/religious backgrounds, and are not spatially integrated.
  - Mechanisms: economic backlash (labor-market competition, public goods) and cultural backlash (group identification).
- Historical and long-run evidence:
  - Tabellini (2020) finds European arrivals in U.S. cities (1910–1930) increased probability of voting for a conservative representative; effects attributed to cultural backlash.
  - Giuliano and Tabellini (2023) find the historical presence of immigrants increases support for the Democratic Party, welfare spending, and a higher minimum wage; magnitudes: a 5-percentage point increase in the historical immigrant share raises the probability of identifying with the Democratic Party by 5.7 percent and support for welfare spending by 2 percent of the sample mean.
  - Contact hypothesis and horizontal cultural transmission can reduce initial backlash over time (Allport, 1954; Bursztyn et al., 2024).
- Effects on sending communities:
  - Pfutze (2012): a one percentage point increase in the proportion of migrant households raises probability a municipality had a non-PRI government by at least 1.7% (1994) and more than 3.4% (1997).
  - Barsbai et al. (2017): departure of one percent of the community to the West is associated with approximately a 0.7 percentage point decrease in Communist vote share in 2009; a 1 percentage point increase in eastward migration increases Communist vote share by about 0.4 percentage points.
- Methodology note: many studies use regional variation and previous stock of migrants as instrumental variables (shift-share approaches) to address endogeneity.

### War: heterogeneous effects on cooperation, trust, and identity
- Overall finding: exposure to war influences social preferences, often increasing in-group cooperation while eroding generalized trust.
- Meta-analysis and review findings:
  - Bauer et al. (2016) meta-analysis (16 studies, over 40 countries) groups outcomes and finds a positive but moderate effect of war on preferences and beliefs: overall cooperation effect of 0.08 standard deviations.
    - Participation in social groups: 0.11 standard deviations.
    - Community leadership and participation: 0.16 standard deviations.
    - Prosocial behavior in experimental games: 0.17 standard deviations.
    - Voting and knowledge of politics: 0.02 and 0.06 standard deviations.
    - Cooperation with the in-group drives results: 0.25 standard deviations; no significant effect on the out-group.
  - Fiedler and Rohles (2021) (39 articles) find war generally associated with an increase in in-group cooperation and a decline in generalized and political trust.
- Evidence on specific outcomes:
  - Cooperation: Gneezy and Fessler (2012) show living under an ongoing external threat temporarily increases willingness of senior citizens to reward cooperation and punish non-cooperators.
  - In-group preferences and political participation: exposure to violence increases ethnic identity and in-group ties (Rohner et al., 2013); war increases ethnic vote share and reduces interethnic friendships (Hadzic et al., 2020); external threats can increase supranational identity and trust in institutions (Gehring, 2022) — effect magnitude: 17 percent of a standard deviation.
  - Generalized trust: evidence of erosion (Rohner, Thoenig, and Zilibotti, 2013; De Luca and Verpoorten, 2015; Alacevich and Zejcirovic, 2020); some studies (Gilligan et al., 2014) find increased altruistic giving post-conflict (Nepal).
- Long-term effects:
  - Conzo and Salustri (2019): exposure to WWII in childhood decreased trust by 4-5 percent and decreased probability of voluntary/charity work by 6.3 percentage points.
  - Adhvaryu and Fenske (2023): exposure to war up to age 14 across 17 African countries (1946-2005) does not influence political engagement or several measures of preferences and beliefs.
  - Besley and Reynal-Querol (2014): historical conflicts (1400–1700) associate with lower out-group trust and higher ethnic identity.
  - Tur-Prats and Valencia Caicedo (2020): Spanish Civil War effects — areas formerly Republican vote more center-left today; Nationalist-occupied areas vote more right; magnitude ~10 percent of vote share.
  - Fontana et al. (2023): north of the Gothic line (Italian Civil War/Nazi occupation) vote share for Communist party was about 9 percentage points higher in the 1946 election; effect persisted until late 1980s–early 1990s.
  - Berman et al. (2023): 1942 Ellwood bombardment (minimal economic damage) produced long-lasting conservative shifts in local political preferences and increased fear language in newspapers.
- Risk preferences:
  - Voors et al. (2012): Burundi civil war associated with increase in risk-seeking preferences, standardized beta coefficient of 0.26.
  - Callen et al. (2014): in Afghanistan (2002–2010), risk tolerance observed mostly under uncertainty; fear increases preference for certainty.

### Terrorist crises: immediate rightward shifts, short-lived effects, and heterogeneous trust responses
- Terrorism is unique because it is intended to influence preferences and beliefs via high visibility and unpredictability.
- Political preferences:
  - Akay et al. (2020): terrorism in Germany, the UK, and Switzerland (1994–2013) increases conservative vote shares by approximately 0.19–0.35 percentage point, a 0.7-1.3 percent increase over the average; effects stronger among older, richer, more risk-averse people, and residents in densely populated areas with high shares of migrants from high-terror countries.
  - Berrebi and Klor (2008): a terror fatality within three months of elections is associated with a 0.45 percentage point increase in locality’s relative electoral support for the right bloc in Israel.
  - Getmansky and Zeitzoff (2014): rocket range from Gaza since 2001 associated with right-wing vote share 2–6 percentage points higher in localities within range.
  - Kibris (2011): in Turkey, each additional casualty per hundred thousand voters in two years before election costs the governing party 0.5 percentage points; average casualties per hundred thousand voters ~7 implied ~3.5 percent loss in 1995.
  - Brouard et al. (2018): terrorist attacks in Paris (2015) and Nice (2016) move political attitudes to the right, with left-wing sympathizers moving right more than right-wing sympathizers; shift concentrated on security issues.
  - Finseraas and Listhaug (2013): 2008 Mumbai attack produced a conservative swing between -0.01 and 0.43 on the left-right scale; effect not long-lasting.
  - Schüller (2015): 9/11 in Germany—left-wing supporters reduce strength of support; right-wing supporters increase strength.
- Trust in institutions:
  - Perrin and Smolek (2009): 9/11 associated with change in trust in government (rally-round-the-flag effect) among sample of young adults (15,127 observations); heterogeneity: African Americans, women, and less-educated less inclined to rally.
- Overall: terrorism typically shifts political preferences to the right in the short run, with limited evidence of long-lasting effects; mechanisms remain multiple and not decisively settled in the literature.

*IMF WORKING PAPERS Aggregate Shocks and the Formation of Preferences and Beliefs*

### 4.4 Pandemics

### 4.4 Pandemics

### 4.4.1. Generalized trust, trust in institutions, and trust in science
- Evidence from past pandemics shows lasting effects on generalized trust and trust in institutions and science, with effects depending on context and government response.
- Aassve et al. (2021):
  - A 1 death per thousand increase in influenza mortality is associated with a 1.4 percentage point decrease in generalized trust.
  - In countries that were neutral in World War I, a 1 death per thousand increase in influenza mortality resulted in a 2.2 percentage point decrease in generalized trust.
- Aksoy et al. (2024) and Aksoy et al. (2021):
  - Experiencing a pandemic during the impressionable years erodes trust in institutions and in scientists.
  - The effect on trust in government is driven mostly by democracies and is stronger for weak governments.
  - An individual with the highest exposure to a pandemic has, on average, 5.1 percentage points less confidence in the national government than those with no exposure; the effect is long-term.
  - An individual with the highest exposure to epidemics is, on average, 11 percentage points less likely than those with no exposure to trust scientists; results are driven by people with little or no science education.
  - Changes in trust in scientists are associated with negative views toward vaccines and lower rates of child vaccination.
- Fluckiger et al. (2019):
  - Exposure to the Ebola epidemic in West Africa enhanced trust in government institutions when the government responded robustly.
  - An increase in average Ebola prevalence of one case per 100,000 people raises post-epidemic trust in parliament, the president, and the police by around 0.15 standard deviations compared to the pre-epidemic era (short-run estimates).
- Bol et al. (2021):
  - COVID-19–related lockdowns were associated with a three percent increase in trust in government in the short term.
  - The lockdown increased satisfaction with democracy and trust in government by about three percent and support for the party of the prime minister/president by about four percent.

### 4.4.2. Political preferences and preferences for redistribution
- Fewer studies focus on political preferences and preferences for redistribution; many use instrumental variables for identification.
- Campante et al. (2024) — Ebola scare in the United States (2014):
  - A one standard deviation increase in Ebola concerns decreased the Democratic vote share by four percentage points in House elections and by three and four percentage points in Senate and gubernatorial elections.
  - A one standard deviation increase in Ebola searches decreased voter turnout by about 1.4 percentage points.
  - Increased negative views about immigrants were observed among voters.
- Baccini et al. (2021) — COVID-19 and county-level voting change between 2016 and 2020:
  - COVID-19 cases negatively affected Trump’s vote shares in 2020, particularly in urban counties, in states without stay-at-home orders, in swing states, and in states Trump won in 2016.
  - A county with 100 more COVID cases per 10,000 people reduced its Trump vote share from 2016 to 2020 by an additional 0.12 percentage points on average.
  - Counterfactual analysis suggests that if cases of COVID had been lower than five percent, Trump would have likely won reelection (authors’ counterfactual claim reported as presented).
- Rees-Jones et al. (2022) — U.S. attitudes toward safety nets (Understanding American Study, June 2020; 2,516 respondents):
  - Respondents living in a county with a one standard deviation higher number of COVID-19 deaths are 8.4 percentage points more likely to support long-term unemployment insurance expansion.
  - The same respondents are 9.2 percentage points more likely to support long-term expansions to government-provided healthcare.
- Cross-study summary:
  - Heterogeneity across contexts is substantial: effects depend on prevailing social and economic context, government response, and whether analysis captures short-term versus long-term impacts.
  - Few studies examine variation by life period of exposure; short-term and long-term effects may differ.

*Source: 4.4 Pandemics, IMF Working Papers — Aggregate Shocks and the Formation of Preferences and Beliefs.*

### Conclusions

### Conclusions

### Overview
- The literature on shocks and beliefs has expanded over the last 30 years owing to theoretical and empirical advances, improved survey availability across countries and time, stronger identification and econometric techniques, and interdisciplinary collaboration with social psychology and sociology.
- This review focuses mostly on different forms of trust, political preferences and preferences for redistribution, and risk preferences, and examines: (a) insights from other disciplines, (b) theoretical reasons aggregate shocks could have permanent effects, (c) whether the life period of exposure matters, and (d) persistence of effects.
- Table 1 in the source summarizes main results across many studies.

### Major conclusions from the literature
- Shocks do affect beliefs and preferences; this finding is consistent across reviewed studies.
- Timing matters: evidence supports a “life cycle” or “impressionable years” effect—shocks experienced at certain life stages have stronger and more lasting impacts on beliefs.
- Effects are long-lasting: many studies document “long-term scarring” where past events—even very old ones—have permanent effects on individuals or populations.
- Negative aggregate economic shocks generally shift attitudes and beliefs to the political right, despite the theoretical possibility they could increase demand for redistribution and left-leaning policies.
  - Potential mechanisms include increased materialism; increased insecurity and individualism (Inglehart, 1990); scapegoating/blaming minorities or immigrants (Bursztyn et al., 2022, 2023); and identity politics reinforcing group identity at the cost of class identification (Shayo, 2009; Bonomi et al., 2021; Grossman and Helpman, 2021).
- Adverse non-economic aggregate shocks also tend to move preferences rightward, but results are heterogeneous and depend on context and government response, particularly for natural disasters where framing and prior beliefs shape outcomes.
- Experiential effects typically translate into actual behavior rather than remaining purely attitudinal.

### Quantified behavioral and political effects (key statistics preserved)
- Experiencing a recession, especially when young, is associated with a lower probability of voting for a Democratic president in the United States:
  - A one standard deviation increase in the measure of shock is associated with a 0.017 standard deviation decline in voting for a Democratic president (Giuliano and Spilimbergo, 2023b).
  - Comparable effect sizes: years of education standardized beta = 0.011; being unemployed standardized beta = 0.019 (Giuliano and Spilimbergo, 2023b).
  - Cross-country standardized beta = 0.021; effect of being unemployed = -0.012; effect of being a woman = 0.029; effect of being married = 0.024.
- Financial experience effects:
  - A 5 percentage point increase in experienced real stock returns is associated with an approximately 10 percentage point increase in the probability of stock market participation and, for participants, an approximately 7 percentage point increase in the percentage of assets allocated to stocks (Malmendier and Nagel, 2011).
  - CEOs who grew up during the Great Depression rely less on external financing and choose more conservative capital structure (Malmendier and Tate, 2005; Graham and Narasimhan, 2004).
- Inflation experience effects:
  - A 1 percentage point increase in experienced inflation increases a borrower’s willingness to pay for a fixed-rate mortgage, with a loss of $8,000–$16,000 for the year 2000 (Malmendier and Nagel, 2016; Botsch and Malmendier, 2020).
  - For policymakers/experts: a one standard deviation increase in experienced-based inflation forecast implies a hawkish dissent of about one-third and lowers the probability of a dovish dissent by about one-third (Malmendier et al., 2021).

### Heterogeneity, mechanisms, and life-period effects
- Life-cycle/impressionable-years effects are documented across many studies; younger cohorts often display stronger and more persistent responses to shocks.
- Mechanisms:
  - Learning from experience (e.g., formation of inflation expectations as a function of experienced inflation).
  - Changes in economic circumstances (temporary vs. long-lasting) that alter redistribution preferences (Meltzer and Richard, 1981).
  - Identity and group dynamics, scapegoating, and shifts in materialism and individualism.
- Individual-level shocks vs aggregate shocks:
  - Aggregate shocks are easier to study because they mitigate endogeneity concerns; personal shocks are nonrandom and may be confounded by omitted variables.
  - Yet individual-level quasi-random shocks (e.g., earnings or wealth shocks) can be informative about mechanisms and can complement aggregate-shock evidence.

### Evidence on political preferences, participation, and polarization
- Personal negative shocks (mostly job losses) increase support for redistribution and reduce trust in the political system in the short term (Margalit, 2019).
- Political responses to shocks can go both left and right; conditions determining direction and long-term persistence are not well understood.
  - Examples of documented political outcomes include:
    - Increase in voting for far-right populist parties and decline in trust in parliament/politicians/legal system following unemployment spikes (Algan et al., 2017).
    - Cross-country historical evidence linking recessions to increased support for right-wing anti-system parties (De Bromhead et al., 2013; Funke et al., 2016).
- Political polarization induced by shocks may make post-crisis macro-responses and reforms harder to implement (Mian et al., 2014).

### Consumption and “scarred consumption”
- Personal, local, or national experiences of unemployment can cause “scarred consumption”—permanently lower consumption spending, with the largest effect from personal experience (Malmendier and Shen, 2020).

### Research agenda and policy implications
- The field needs a formal life-cycle theory of belief formation that integrates multiple mechanisms and can be embedded in macro and political-economy models, including heterogeneous generations in voting and decision models.
- A broad theoretical framework should clarify when different types of responses to shocks should be expected and provide a roadmap synthesizing diverse findings.
- Greater collaboration with social psychology and neuroscience is recommended, along with the use of experiments and alternative methodologies to identify precise mechanisms.
- Studying individual-level quasi-random shocks could yield complementary evidence and help identify mechanisms behind aggregate-shock effects.
- Understanding how shocks alter policies via their effects on beliefs—particularly through political polarization and fractionalization—remains an important area for future work.

*Source: wpiea2024195-print-pdf — Conclusions.*

### 2016. Nielsen Local

### wpiea2024195-print-pdf - 2016. Nielsen Local

### Trade shocks and import competition
- Colantone and Stanig (2018c)
  - Shock: Trade shocks
  - Preferences/beliefs: Voting for nationalist far-right parties
  - Main finding: Increase in voting for nationalist, far-right parties
  - Setting: District level data for 15 Western European countries 1985-2006, European Social Survey 2002-2008
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous
- Feigenbaum and Hall (2015)
  - Shock: Chinese import competition
  - Preferences/beliefs: Voting for the right, protectionism or foreign trade bills
  - Main finding: Incumbents vote more to the left, toward more protectionism on foreign trade bills.
  - Setting: Roll=call behavior and electoral outcomes in the US House, 1990-2010
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous
- Rodrik (2021) (trade)
  - Shock: Trade shocks
  - Preferences/beliefs: Voting for nationalist far-right parties in various
  - Main finding: Increase in voting for right-wing parties
  - Setting: Review articles from different countries
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous
- Tabellini and Magistretti (forthcoming)
  - Shock: Trade shocks
  - Preferences/beliefs: Preferences toward democracy
  - Main finding: Economic integration with democracies increases preferences for democracy
  - Setting: International Values Survey, 1990-2020, 74 countries
  - Life period: Impressionable years
  - Persistent/contemporaneous effect: Persistent effect
- Ballard-Rosa et al. (2021)
  - Shock: Chinese import penetration
  - Preferences/beliefs: Authoritarian preferences
  - Main finding: Import shocks from China implies stronger authoritarian preferences
  - Setting: Newly collected survey on 1,193 UK respondents in 2017
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous
- Colantone and Stanig (2018b)
  - Shock: Chinese import penetration
  - Preferences/beliefs: Preferences for democracy, a liberal democracy, authoritarian values
  - Main finding: Import shocks from China implies lower support for democracy and a liberal democracy but have no effects on authoritarian preferences
  - Setting: European Values Survey 1998-2008
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous

### Migration shocks
- Rodrik (2021) (migration)
  - Shock: Migration shocks
  - Preferences/beliefs: Voting for nationalist far-right parties in various
  - Main finding: Increase in voting for right-wing parties
  - Setting: Review articles from different countries
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous
- Tabellini (2020)
  - Shock: Age of Mass Migration 1910-1930
  - Preferences/beliefs: Voting for conservative representative
  - Main finding: Increase in voting for conservative representative; Increase in conservative voting for 1910-1930
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous
- Alesina and Tabellini (2024)
  - Shock: Migration shocks
  - Preferences/beliefs: Natives backlash, support for anti-immigrant and populist parties, preferences for redistribution
  - Main finding: Immigration triggers natives backlash, increases support for anti-immigrant and populist parties, and lower preferences for redistribution
  - Setting: Review articles
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous and long-term effects
- Alesina et al. (2021)
  - Shock: Migration shocks
  - Preferences/beliefs: Preferences for redistribution
  - Main finding: Immigration lower preferences for redistribution
  - Setting: 16 European countries from 1990-2010
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous effect
- Giuliano and Tabellini (2023)
  - Shock: Age of Mass Migration
  - Preferences/beliefs: Preferences for redistribution and political preferences
  - Main finding: Immigration increases preferences for redistribution and voting for the democratic parties
  - Setting: Political ideology and preferences for redistribution, 2006-2020
  - Life period: No different life-time periods studied
  - Persistent/contemporaneous effect: Long-term effect
- Pfutze (2012)
  - Shock: International migration
  - Preferences/beliefs: Probability of a party in opposition to the former state party PRI in Mexico
  - Main finding: Increase in the probability of an opposition party victory
  - Setting: Municipal election for the 2000-2002 electoral cycle
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous-effect
- Barsbai et al. (2017)
  - Shock: Migration after the 1998 Russian crisis
  - Preferences/beliefs: Political support for democratic and liberal parties
  - Main finding: Decline in Communist vote in the elections of 2009
  - Setting: Change in the share of Communist votes in the parliamentary election between 1998 and 2009
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous-effect

### War, conflict, and political/social effects
- Bauer et al. (2016)
  - Shock: War
  - Preferences/beliefs: Social cooperation, 16 studies
  - Main finding: Conflicts in different countries — War leads to cooperation and pro-social behavior
  - Life period: No life-time periods studied, with few exceptions
  - Persistent/contemporaneous effect: Contemporaneous or medium term effect
- Fiedler and Rohles (2021)
  - Shock: War
  - Preferences/beliefs: Generalized and political trust, in-group preferences
  - Main finding: Decline in trust and political trust, increase in in-group cooperation
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous effect or medium term effect
- Gneezy and Fessler (2012)
  - Shock: War (2006 Israel-Hezbollah conflict)
  - Preferences/beliefs: Cooperation
  - Main finding: Living under an external threat increases the willingness to reward cooperation and punish non-cooperators; Stronger effects on senior citizens
  - Life period: Contemporaneous effect
- Bauer et al. (2014)
  - Shock: War
  - Preferences/beliefs: In-group preferences
  - Main finding: War promotes in-group preferences; Effects stronger between the age of 7 and 20
  - Settings: 2008 Georgia conflict with Russia over Ossetia, 1991-2002 civil war in Sierra Leone
  - Life period: Contemporaneous effect
- Rohner et al. (2013)
  - Shock: War (Lord’s Resistance Army insurgency (1986–2006) in Uganda)
  - Preferences/beliefs: In-group ties
  - Main finding: War promotes in-group ties
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous effect
- Hadzic et al. (2020)
  - Shock: War (1995 Bosnia War)
  - Preferences/beliefs: Ethnic voting
  - Main finding: War increases ethnic vote share, increase in-group trust
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous effect
- Gehring (2022)
  - Shock: War (Russian invasion of Ukraine)
  - Preferences/beliefs: European identity and trust in institutions
  - Main finding: War increases European identity and trust in European institutions; Results stronger for older people
  - Life period: Contemporaneous effect
- De Luca and Verpoorten (2015)
  - Shock: War (Lord’s Resistance Army insurgency (1986–2006) in Uganda)
  - Preferences/beliefs: Civic and political participation
  - Main finding: Increase in civic participation but no effect on electoral turnout
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Contemporaneous effect
- Cassar et al. (2013)
  - Shock: Civil war in Tajikistan (1992-1997)
  - Preferences/beliefs: Political attitudes, Membership in groups, willingness to engage in impersonal exchange, collective action and political participation
  - Main finding: War hampers trust within localities, decreases willingness to engage in impersonal exchange, reinforces kinship-based norms of morality
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Effect more than a decade after the end of the civil war
- Gilligan et al. (2014)
  - Shock: Nepalese civil war (1996-2006)
  - Preferences/beliefs: Altruistic giving, public good contributions, investment in trust-based transactions, and willingness to reciprocate trust-based investment
  - Main finding: Increase in altruistic giving, public good contributions, investment in trust-based transactions, and willingness to reciprocate trust-based investment
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Conzo and Salustri (2019)
  - Shock: Exposure to World War II during childhood
  - Preferences/beliefs: Generalized trust
  - Main finding: Decline in generalized trust in 13 European countries
  - Life period: Early childhood
  - Persistent/contemporaneous effect: Long term effects
- Adhvaryu and Fenske (2023)
  - Shock: War (17 African countries with conflict between 1946 and 2005)
  - Preferences/beliefs: Political preferences
  - Main finding: No effect on political engagement, Afrobarometer 2005
  - Life period: Age 0-14
  - Persistent/contemporaneous effect: Long-term effects
- Besley and Reynal-Querol (2014)
  - Shock: War (18 African countries with conflict between 1400 and 1700)
  - Preferences/beliefs: Outgroup trust
  - Main finding: Lower outgroup trust
  - Life period: No lifetime period studied
  - Persistent/contemporaneous effect: Long-term effects
- Tur-Prats and Valencia Caicedo (2020)
  - Shock: War (Spanish Civil War)
  - Preferences/beliefs: Trust and voting behavior
  - Main finding: Lower trust and trust in institutions associated with the Civil War if a person was exposed to the Spanish Civil War
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Long-term effects
- Fontana et al. (2023)
  - Shock: War (Italian Civil War and Nazi occupation of Italy, 1943-45)
  - Preferences/beliefs: Voting behavior
  - Main finding: In places where Nazi occupation was higher, the vote share for the Communist party was higher
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Long-term effects
- Berman et al. (2023)
  - Shock: War (Ellwood bombardment in California by a Japanese submarine in 1942)
  - Preferences/beliefs: Political preferences
  - Main finding: Conservative shift in presidential and gubernatorial elections as a result of the bombardment
  - Life period: No life-time period studied
  - Persistent/contemporaneous effect: Long-term effects
- Voors et al. (2012)
  - Shock: Civil war in Burundi, 1993-2003
  - Preferences/beliefs: Risk preferences
  - Main finding: Increase in risk seeking
  - Life period: No life-time period studied
  - Persistent/contemporaneous effect: Medium-term effects
- Callen et al. (2014)
  - Shock: War (Afghanistan)
  - Preferences/beliefs: Risk preferences
  - Main finding: Increase in risk seeking
  - Life period: No life-time period studied
  - Persistent/contemporaneous effect: Short-term effects

### Terrorist attacks
- Akay et al. (2020)
  - Shock: Terrorist attacks
  - Preferences/beliefs: Political preferences
  - Main finding: More support for conservative parties
  - Setting: Data on political preferences for Germany, the UK and Switzerland, 1994-2013. 70,000 terrorist events from the Global Terrorist Database
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Berrebi and Klor (2008)
  - Shock: Terrorist attacks
  - Preferences/beliefs: Political preferences
  - Main finding: Increase in support for right-wing parties
  - Setting: Polling station data on vote share for different political parties and terror attacks in Israel between 1984 and 2004
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Gould and Klor (2010)
  - Shock: Terrorist attacks
  - Preferences/beliefs: Political preferences
  - Main finding: Increase in support for the right wing block
  - Setting: Israel National Election Studies, 1988-2006
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Getmansky and Zeitzoff (2014)
  - Shock: Terrorist attacks
  - Preferences/beliefs: Political preferences (Right-wing bloc vote-share in Israel, 1999-2009)
  - Main finding: Increase in support for the right-wing block
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Kibris (2011)
  - Shock: Terrorist attacks
  - Preferences/beliefs: Political preferences
  - Main finding: Increase in support for the right-wing parties
  - Setting: Electoral district level data, 1987-1995
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Echebarria-Echabe and Fernández-Guede (2006)
  - Shock: Terrorist attacks (March 2004 terrorist bombing in Madrid)
  - Preferences/beliefs: Political preferences
  - Main finding: Increase in support for right-wing parties and increase in racial prejudice against Muslims but also Jews
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Montalvo (2011)
  - Shock: Terrorist bombing in Madrid before and after March 2004
  - Preferences/beliefs: Political preferences (voting behavior of people living abroad casting votes before the attack and voters in Spain casting votes after the attack)
  - Main finding: Lower consensus for the incumbent conservative party
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Brouard et al. (2018)
  - Shock: Terrorist attacks (Paris 2015 and Nice 2016)
  - Preferences/beliefs: Political preferences
  - Main finding: Terrorist attacks moved political preferences to the right
  - Setting: Survey from the “Policy Priority Barometer” and “Local Elections” projects, 2014-2016
  - Life period: No life-time period studied
  - Persistent/contemporaneous effect: Short-term effects
- Finseraas and Listhaug (2013)
  - Shock: Terrorist attacks
  - Preferences/beliefs: Political preferences
  - Main finding: Shift in conservative direction on the left–right wing scale, but not precisely estimated
  - Setting: European Social Survey 2008
  - Life period: Not life-time period studied
  - Persistent/contemporaneous effect: Short-term effects
- Schüller (2015)
  - Shock: Terrorist attacks
  - Preferences/beliefs: Political preferences
  - Main finding: Weaker support intensity among left-wing, higher strength of support among the right wing
  - Setting: German Socio-Economic Panel, 1999-2003
  - Life period: No life-time period studied
  - Persistent/contemporaneous effect: Short-term effects
- Perrin and Smolek (2009)
  - Shock: Terrorist attacks (9/11/2001)
  - Preferences/beliefs: Trust in government
  - Main finding: Trust in all level of government increases after 9/11/2001
  - Setting: Wave III of the National Longitudinal Survey of Adolescent Health
  - Life period: Sample of young adults
  - Persistent/contemporaneous effect: Short-term effects

### Pandemics and public trust/preferences
- Aasve et al. (2021)
  - Shock: Pandemics (Exposure to the Spanish Flu)
  - Preferences/beliefs: Generalized trust
  - Main finding: Exposure to the Spanish Flu reduces generalized trust
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Long-term effects
- Aksoy et al. (2024)
  - Shock: Pandemics
  - Preferences/beliefs: Confidence in institutions
  - Main finding: Lower trust in institutions
  - Setting: Gallup World Polls, 2006-2018, EM-DAT International Disasters Database
  - Life period: Impressionable years
  - Persistent/contemporaneous effect: Long-term effects
- Aksoy et al. (2021)
  - Shock: Pandemics
  - Preferences/beliefs: Confidence in scientists
  - Main finding: Lower trust in science
  - Setting: 2018 Welcome Global Monitor, EM-DAT International Disasters Database
  - Life period: Impressionable years
  - Persistent/contemporaneous effect: Long-term effects
- Fluckiger et al. (2019)
  - Shock: Pandemics
  - Preferences/beliefs: Trust in government
  - Main finding: Increase in trust in government
  - Setting: Afrobarometer, 2012-2015
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Bol et al. (2021)
  - Shock: Pandemics
  - Preferences/beliefs: Trust in government, satisfaction with democracy, vote for the party of the current Prime Minister/President
  - Main finding: Increased vote intentions for the party of the Prime Minister/President, increased trust and satisfaction with democracy
  - Setting: Web based survey for 15 European countries between March2-April 3 2020
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Campante et al. (2024)
  - Shock: Pandemics
  - Preferences/beliefs: Voting in congressional and gubernatorial elections
  - Main finding: Lower vote share for Democrats
  - Setting: US, county level electoral data, 2002-2014
  - Life period: No life-time period studied
  - Persistent/contemporaneous effect: Short-term effects
- Baccini et al. (2021)
  - Shock: Pandemics
  - Preferences/beliefs: Voting for Trump between 2016-2020
  - Main finding: Covid19 negatively affect Trump’s vote share
  - Setting: Country level data on electoral outcomes
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Rees-Jones et al. (2022)
  - Shock: Pandemics
  - Preferences/beliefs: Support for unemployment insurance and government provided healthcare
  - Main finding: Increased support for unemployment insurance and government provided healthcare
  - Setting: Individual panel level data, June 2020, January-February 2021
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term, medium-term effects

### Natural disasters and behavioral effects
- Achen and Bartels (2017)
  - Shock: Natural disasters (droughts and floods)
  - Preferences/beliefs: Voting for the incumbent
  - Main finding: Voters punished incumbent parties
  - Setting: State level data for the US on droughts and floods, 1896-2000
  - Life period: No life-time period studied
  - Persistent/contemporaneous effect: Short-term effects
- Lazarev et al. (2014)
  - Shock: Natural disasters
  - Preferences/beliefs: Support for authorities
  - Main finding: Exposure to the disaster increased support to the United Russia Party, the village head, the governor, Putin and Medvedev.
  - Setting: Survey questions collected by the authors
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Healy and Malhotra (2009)
  - Shock: Natural disasters
  - Preferences/beliefs: Voting for the incumbent
  - Main finding: Voters reward the incumbent presidential party for delivering disaster relief spending, but not for investing in disaster preparedness spending
  - Setting: US county level electoral data, 1988-2004
  - Life period: No life-time period studied
  - Persistent/contemporaneous effect: Short-term effects
- Albrecht (2017)
  - Shock: Natural disasters
  - Preferences/beliefs: Political trust and satisfaction with the government
  - Main finding: Political trust and satisfaction with the government are largely not affected
  - Setting: European Social Survey, 2001-2017
  - Life period: No life-time period studied
  - Persistent/contemporaneous effect: Short-term effects
- Whitt and Wilson (2007)
  - Shock: Natural disasters
  - Preferences/beliefs: Cooperation in a public goods game
  - Main finding: Evacuees from New Orleans relocated to Houston-area shelters — Increase in good cooperation
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Cassar et al. (2017)
  - Shock: Natural disasters (2004 tsunami)
  - Preferences/beliefs: Risk aversion, pro-social behavior and patience
  - Main finding: Increase in risk aversion, pro-social behavior, trust and impatience
  - Setting: Experiments in rural Thailand after the 2004 tsunami
  - Life period: No life-time periods studied; Four years after the disaster
  - Persistent/contemporaneous effect: Four years after the disaster
- Cameron and Shah (2015)
  - Shock: Natural disasters
  - Preferences/beliefs: Risk aversion
  - Main finding: Individuals who experienced a flood or earthquake exhibit more risk aversion
  - Setting: 1,550 individuals from 120 rural villages in East Java, 2008
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Said et al. (2015)
  - Shock: Natural disasters (2010 flood in Pakistan)
  - Preferences/beliefs: Risk aversion
  - Main finding: Individuals who live in areas affected by the 2010 flood in Pakistan are more risk averse, but personal experience of a flood mitigate the effects.
  - Setting: Experiment and survey data
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Chantarat et al. (2019)
  - Shock: Natural disasters
  - Preferences/beliefs: Risk aversion, altruism, impatience, trust of friend and local government
  - Main finding: Greater risk aversion and altruism, lower impatience and trust of friends and local government
  - Setting: Survey conducted in 2014
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Eckel et al. (2009)
  - Shock: Natural disasters (Katrina’s evacuees, 2005, 2006)
  - Preferences/beliefs: Risk aversion
  - Main finding: Women increase risk seeking behavior
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects
- Brown et al. (2018)
  - Shock: Natural disasters
  - Preferences/beliefs: Risk aversion
  - Main finding: Increased future expectations of losses, but results mediated by ethnicity
  - Setting: 259 Fijian households from 14 rural villages
  - Life period: No life-time periods studied
  - Persistent/contemporaneous effect: Short-term effects

*IMF WORKING PAPERS — Aggregate Shocks and the Formation of Preferences and Beliefs*

### References

### wpiea2024195-print-pdf - References

### Major research topics covered by the referenced literature
- Epidemics and pandemics: historical and contemporary political, trust, and behavioral consequences (e.g., “Epidemics and trust: the case of the Spanish Flu”; “The Political Scar of Epidemics”; “Revenge of the Experts: Will Covid-19 Renew or Diminish Public Trust in Science?”).
- Conflict, violence, and civil war: effects on trust, cooperation, ethnic voting, social cohesion, and political participation (e.g., “How exposure to violence affects ethnic voting”; “Legacies of Violence: Trust and Market Development”; “Violent Conflict and Behavior: A Field Experiment in Burundi”).
- Natural disasters and climate shocks: impacts on risk preferences, social protection, political support, and policy attitudes (e.g., “Risk-Taking Behavior in the Wake of Natural Disasters”; “Natural Disasters, social protection, and risk perceptions”).
- Trade, globalization, and import competition: local labor market effects, political backlash, electoral outcomes, and cultural responses (e.g., “The China Syndrome: Local Labor Market Effects of Import Competition”; “Global competition and Brexit”; “Low-wage import competition and populist backlash”).
- Immigration and migration: electoral effects, remittances and political beliefs, democratization, and cultural transmission (e.g., “The Immigrant Next Door”; “Do migrants remit democracy?”; “Immigration and voting for the far right”).
- Macroeconomic shocks and crises: financial crises, recessions, and their long-run effects on political behavior, redistribution preferences, and risk attitudes (e.g., “Going to extremes: Politics after financial crises, 1870–2014”; “Recession Experiences Shape Prosocial Attitudes”).
- Formation and persistence of beliefs, preferences, and values: generational effects, emotional tagging, experience effects, and political socialization (e.g., “Growing Up in a Recession”; “Emotional Tagging and Belief Formation-The Long-Lasting Effects of Experiencing Communism”; “The Problem of Generations”).
- Terrorism and external threats: impacts on voting, public opinion, and ideological orientation (e.g., “How terrorism affects political attitudes”; “Minding the effects of 9/11 and other attacks”).
- Trust, social capital, and populism: links between trust, institutional legitimacy, and the rise of populist politics (e.g., “Inherited Trust and Growth”; “The European Trust Crisis and the Rise of Populism”; “The Political Economy of Populism”).

### Common empirical approaches and data sources signaled by the references
- Historical and archival analyses spanning events such as the Spanish Flu, Great Depression, World War II, and the Great Inflation.
- Natural experiments and quasi-experimental designs exploiting localized shocks (e.g., trade shocks, terrorist attacks, natural disasters).
- Laboratory and field experiments measuring behavioral responses to violence, conflict, and migration.
- Cross-country and panel analyses of survey-based measures (trust, political attitudes, preferences) and administrative labor market outcomes.
- Interdisciplinary connections to psychology and neuroscience for mechanisms of memory, emotional tagging, and decision-making (e.g., “Emotional intensity predicts autobiographical memory experience”; “Synaptic tagging and long-term potentiation”).

### Thematic linkages and recurring emphases
- Persistent influence of formative experiences (war, recession, epidemic) on long-term political beliefs and economic preferences.
- Interaction between economic shocks (trade, recession, financial crises) and cultural or identity-based political responses.
- Role of trust—both interpersonal and institutional—in mediating political and economic outcomes after aggregate shocks.
- Evidence-generating focus on how shocks reshape voting behavior, support for redistribution, and demand for government intervention.

*Aggregate Shocks and the Formation of Preferences and Beliefs — Working Paper No. WP/2024/195*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024195-print-pdf.pdf_