It is hardly contentious to say that policy measures often fail to reach their intended goals. Frequently, they are not based on sufficient evidence, much less on a sound understanding of the underlying reaction of the economy to the proposed changes. Fostering such understanding and better policymaking is an important motivation driving Anna Rubinchik’s (THEOMET) research.
One of her areas of interest is the development of overlapping-generations models useful for policy evaluation. In a work-in-progress article with Alexander Gorokhovsky (UColorado), for instance, Anna works on a class of such models that captures key features of economic reality, like the hump-shaped life-cycle path of earnings, the midlife spike in the desire for immediate enjoyment, along with aggregate trends: a steady increase in consumption and a slight decrease in working hours. The next step, according to Anna, is “to use this model as a laborious yet exciting illustration of what the neoclassical economists were promising for a long time: a perfect foresight that can deliver predictions about a full response to a change in economic policy.”
Her approach contrasts with traditional analysis treating the response of an economy to interventions as a reaction to an impulse applied to an inanimate object, like the swinging of a pendulum suddenly pushed out of equilibrium. While some policy changes might surprise us, most are at least partially anticipated, triggering changes in our economic behavior far in advance. “I am looking forward to working jointly with faculty and graduate students. I hope that together we can calibrate the model and its extensions so as to quantify predictions about various policy changes, such as education programs or pension reforms.”
Meanwhile, Anna is venturing into neuroeconomics and related fields. Currently, she is working with a team of neurobiologists studying human neural firing patterns to uncover mechanisms behind boundedly rational behavior. But this is not her first foray into these fields. In a recent article with Flavia Aluisi (UHaifa) and Genela Morris (UHaifa), Anna proposes a learning model to analyze how rats process and respond to stimuli with multiple sensory dimensions. The team uncovers the effects of previous learning. "Using the model, we estimated the attention weights corresponding to different sensory dimensions and discovered a surprising similarity to biases in human behavior: having learnt to pay attention to a particular dimension of the problem may impede identifying the key to success in a new environment. While restricting attention to a few parameters speeds up the analysis, it might also blindfold a decision maker when the situation changes."
