Pegah Rahmani dreams of mathematical proofs. Testing them is her day job.
“For a period, I was working on the proof all day and night. At night, when I was asleep, I’d think of something and when I woke up, I had ideas to test,” Rahmani said. “When I finally had the proof and saw the new axiomatization, it was beautiful.”
Wanting to share the beauty of her axiomatization proof means the economic theorist usually can’t just explain her work. Rahmani has to teach people about the role of correlation sensitivity in the processes of decision making first.
To explain correlation sensitivity, the Rahmani refers to start up culture. Most people know that most start ups fail within a few years of beginning. They also know about the origins of Amazon and Microsoft and might be tempted to invest in what could be the next big tech boom to burst out of one parent’s garage or another’s basement.
“Suppose you can invest in a safer start up with a modest return and a 10% chance of success, or in a moonshot with a massive payoff but only a 5% chance of success,” Rahmani explained. “If the risks are independent, many people go for the moonshot. If you’re likely to lose anyway, why not aim high?”
If the potential for the failure of both start ups are connected, or correlated, the risk picture starts to look different to potential investors.
“Say both companies are in the AI sector and new regulations hit,” Rahmani said. “In that world, when the sector survives, the safer startup almost always succeeds, while the moonshot still fails half the time. Suddenly, the safer option looks better even if nothing about the odds of each startup making it changed. What changed was how their fates were linked. That’s what I call correlation sensitivity.”
Typically, behavioural economists who study decision making focus on the potential for regret and salience people exhibit when considering risk. The regret is the danger while the salience focuses on the details that are different or stand-out that sway decisions. In most models that are used to study decision making, people can rank any two options perfectly, but Rahmani’s work allows for incompleteness in considerations.
“Conceptually, think of multiple correlation-sensitive models. If they all agree, you choose. If they contradict, you remain indecisive,” Rahmani said.
Correlation sensitivity makes decision-making much more complex.
“To account for correlation sensitivity, you often need to compare counterfactuals, what would have happened if the subject was offered different choices,” she explained. “That complexity motivates the need to accommodate incomplete preferences.”
Rahmani’s job market paper, Incomplete correlation-sensitive preferences: An axiomatic framework for decision making under uncertainty, adds a new model for understanding how people choose between those risky options when their correlation structure affects choice.
“Pegah considers decision makers who may be indecisive and derives a tight characterization of their preferences,” said Professor Yoram Halevy, Rahmani’s thesis supervisor. This is a very natural and fundamental generalization of existing models, as correlation-sensitive decisions tend to be more complex, which can lead to incomparability.”
For Rahmani, teaching others about correlation sensitivity is not just about sharing the beauty of her mathematical proofs.
“Correlation sensitivity isn’t just abstract,” she said. “It reflects how people actually make decisions. When failures are independent, you aim high. When they’re correlated, you focus on the remaining possibilities, and that changes your choice. This matters in real-world decisions like investing, career choices, or even strategic planning.”
She learned these lessons about clarity and the application of theory herself, in part as a course instructor for ECO326, Advanced Economic Theory (Game Theory).
“At first, I thought the math made everything clear. The proofs show exactly what choices people should make, but I realized that’s not true for everyone,” Rahmani remembered. “I began using very simple, intuitive examples from daily life, like working on a group project, to explain strategic situations. Using these familiar examples helps students see that game theory isn’t just abstract math.”
Return to the Department of Economics website.
Scroll more news.