In democratic societies, public support is essential for policy implementation. Yet public opinion is fickle: it is constantly shaped by competing stories about what drives social and economic outcomes.
Take the recent experience of higher inflation in many countries. Depending on whom you ask, it was caused by fiscal stimulus, loose monetary policy, supply shocks, corporate greed or simply the political party in government. There are similar clashes in debates about crime: are trends driven by technology, immigration, welfare policy or policing norms?
Different narratives point to different culprits and different solutions. By definition, they cannot all be right. But some clearly gain traction even when they are false.
This raises fundamental questions: why do certain false narratives catch on? What makes them appealing? And what consequences do they have for public policy?
As an economic theorist of the ‘pure’ variety, my modus operandi is to take a messy social phenomenon like ‘competing policy narratives’, and try to distil it into a mathematical model that can deliver valuable insights into the phenomenon in question. This is what Kfir Eliaz (King’s Business School) and I set out to do in our 2020 American Economic Review paper ‘A Model of Competing Narratives’ (Eliaz and Spiegler, 2020).
It is probably best to convey our ideas with a simple thought experiment, abstracted from debates about globalisation in Western economies. In this example, there is a true narrative: technological change drives both a country’s trade balance and employment in its manufacturing sector. Trade policy can shift the trade balance but not employment.
This narrative is represented by Diagram 1, which is an example of a ‘directed acyclic graph’ – a common representation of causal models in statistics and artificial intelligence. (For a lively popular introduction to this field of research, see Pearl and Mackenzie, 2018.)
Figure 1: True narrative
Source: Authors’ methodology
If voters accept this story, they will oppose trade restrictions, which only harm consumers without helping workers.
But now imagine that politicians and pundits start peddling a false narrative, represented by Diagram 2. This narrative ignores the role of technology and argues that the trade balance has a direct causal effect on employment.
Figure 2: False narrative
Source: Authors’ methodology
If voters buy into this false narrative, they entertain the possibility that trade policy might help workers. So, these two narratives imply conflicting policies; but which one will hold sway over the public?
Kfir and I addressed this question with a theoretical argument based on two premises.
- Premise 1: False narratives generate beliefs by misinterpreting correlations that actually exist in historical data. For example, if the historical record shows that trade surpluses are associated with higher employment, the false narrative concludes that: ‘trade surpluses create jobs’. In reality, both outcomes are driven by favourable technological shocks. The false narrative confuses a mere correlation with a causal effect.
- Premise 2: Voters are drawn to hopeful narratives, which promise good outcomes according to the beliefs that they generate and the policies that these beliefs support. This is an assumption about political psychology: people are prone to wishful thinking, especially when their individual vote has little effect on outcomes.
When we put these premises together, an intriguing political dynamic emerges:
- When the true narrative dominates public opinion, governments implement free trade. Technological shocks then drive both employment and the trade balance, creating a strong correlation between the two.
- This correlation gives oxygen to the false narrative, which promises an easy fix: restrict trade to boost employment. The false narrative becomes more hopeful than the true one, because it implies that the benefits from trade restrictions outweigh their costs.
- Once the false narrative prevails, trade is restricted. Technology still drives employment, but now has less impact on trade. The observed correlation between the trade balance and employment gradually becomes weaker, eroding the false narrative’s appeal.
- Eventually, support swings back to the true narrative – and the cycle begins again.
In short, a ‘populist’ narrative can thrive on opportunistic misreadings of empirical correlations. Yet because it feeds off the successes of sound policies, its ascendance is self-limiting. Over time, public opinion oscillates between rationality and wishful thinking, free trade and protectionism.
Our 2020 paper went beyond this thought experiment, and presented a theory of competing policy narratives that drew more systematically on the language directed acyclic graphs. This type of cross-pollination is one of the pleasures of doing economic theory: the surprising encounters it creates between distant scientific fields – from AI to political psychology.
Like much of abstract theory, the value of our exercise lies less in definitive answers than in providing a language for posing new questions. For example:
- How do scapegoating narratives enable governments to pursue bad policies with intrinsic popular appeal, such as ‘irresponsible’ fiscal spending (Eliaz et al, 2025)?
- What role do false narratives play in the phenomenon of media bias (Eliaz and Spiegler, 2024)?
Our theoretical framework has also inspired empirical research that extracts popular narratives from textual data and links them to physical events (for example, see Andre et al, 2025). Thus, from both theoretical and empirical perspectives, the promise of our approach is that it offers systematic ways of thinking about that elusive and ever-shifting object: public opinion.
