Research Papers
Imperfect OPEC: An Oil Market Model (Job Market Paper)
OPEC is designated to be a cartel. Yet over the past decades, researchers have puzzled over whether its behavior is consistent with one. One size of the puzzle is that OPEC is the low-cost producer, yet it has lower market share than the competition. The other side being that despite this relatively restraint, the cartel is not consistently able to keep prices from falling below its stated targets. Theoretical models have struggled to model and explain the group’s behavior. I construct a model based on imperfect cartelization and bounded rationality: I find that OPEC faces multiple challenges in stabilizing the market due to the heterogeneity of producers and the long lead-times and lifecycles of oil projects. When those are taken into account, the model replicates closely the production paths of OPEC/Non-OPEC agents in the historical period from 1973-2017. The model shows that endogenous cyclicality in prices can arise naturally in the oil market due to over (under) investment following positive (negative) demand shocks. I use the model to construct long-term scenarios which exhibit a stationary but cyclical price pattern under various demand assumptions.
Interpretable Economics: Using Fuzzy Logic to Model Agent Behavior in Energy Markets
I construct a model of energy markets with linguistic rules to describe agent behavior. I prototype a stylized model of the oil market to simulate the production and investment decisions of producers and the price negotiation process between producers and consumers. I show that the interaction between agents can lead to a realistic pattern of endogenous cyclicality due to the presence of field development lags and that the price negotiation mechanism can successfully clear the markets.
Why Scenarios Fail to Model the Instability of the Oil Market
Scenario models of the oil market often generate linear outlooks that fail to foresee the possibility of significant “tipping point” dynamics in prices, even though those “boom-and-bust” have been a regular market characteristic since OPEC formed in 1973. We argue that the oil market is a continuously evolving system with changing structural features and behaviors and that an understanding of how the micro-foundations of the markets change is essential to modeling the energy transition. We use a micro-level dataset to discuss the empirical features that give rise to market complexity and ways to model them. Finally, we demonstrate through a model of adaptive price expectations that myopic investment behavior based on net present value and/or breakeven price calculations is not sufficient to capture the dynamics of producers' behavior.
Work in Progress
“Evolutionary Methods for Generating Trading Strategies in The Crypto Market.”
I use evolutionary methods to construct heuristics-based trading strategies for portfolio optimization in the crypto market. I use the evolutionary search algorithm to select from a large number of input variables (more than 500) for inclusion in the trading rules and show that the algorithm can successfully identify relevant variables. I use the strategies to optimize various measures of risk and return
“Systemic Racism or a Few Bad Apples: One and The Same?” (With Raffi Garcia and Merima Tricic)
Much of the political discourse on racism in the United States centers around the argument of "systemic racism" vs "a few bad apples". While academic research has documented the presence of systemic racism in multiple social domains, in this paper we show that the "few bad apples" defense -- that is, that most racism is conducted by a minority of bad actors, does not necessarily negate the presence of systemic racism at all. We use a stylized agent-based model: agents go through various 'stages' analogous to education/career phases where they exert effort and their performance is evaluated. The move to the next stage is determined by this evaluation so that the highest subjectively performing agents are placed in higher tiers in the next stage (higher ranking universities, more prestigious jobs etc..). we demonstrate that even in the presence of only a small percentage of 'discriminatory agents' when evaluating, outcomes for minority agents can be severely skewed as they face an increasing probability of being held back at each successive stage and as those effects are compounded. The end result being that minority agents are under-represented at higher tiers. This can also lead to a feedback effect as there will be less minority agents that make the decisions to evaluate, making it more likely for discrimination to persist.