Erik Snowberg

The Promise of Prediction Markets

The ability of groups of people to make predictions is a potent research tool that should be freed of unnecessary government restrictions.

Explaining the Favorite-Long Shot Bias: Is it Risk-Love or Misperceptions?

The favorite–long shot bias describes the long-standing empirical regularity that betting odds provide biased estimates of the probability of a horse winning: long shots are overbet whereas favorites are underbet. Neoclassical explanations of this phenomenon focus on rational gamblers who overbet long shots because of risk-love. The competing behavioral explanations emphasize the role of misperceptions of probabilities. We provide novel empirical tests that can discriminate between these competing theories by assessing whether the models that explain gamblers’ choices in one part of their choice set (betting to win) can also rationalize decisions over a wider choice set, including compound bets in the exacta, quinella, or trifecta pools. Using a new, large-scale data set ideally suited to implement these tests, we find evidence in favor of the view that misperceptions of probability drive the favorite–long shot bias, as suggested by prospect theory.

Television and the Incumbency Advantage in U.S. Elections

We use the structure of media markets within states and across state boundaries to study the relationship between television and electoral competition. In particular, we compare incumbent vote margins in media markets where content originates in the same state as media consumers versus vote margins where content originates out of state. This contrast provides a clear test of whether or not television coverage correlates with the incumbency advantage. We study U.S. Senate and state gubernatorial races from the 1950s through the 1990s and find that the effect of TV is small, directionally indeterminate, and statistically insignificant.

Selective Trials: A Principal-Agent Approach to Randomized Controlled Experiments

We study the design of randomized controlled experiments in environments where outcomes are significantly affected by unobserved effort decisions taken by the subjects (agents). While standard randomized controlled trials (RCTs) are internally consistent, the unobservability of effort provision compromises external validity. We approach trial design as a principal-agent problem and show that natural extensions of RCTs - which we call selective trials - can help improve the external validity of experiments. In particular, selective trials can disentangle the effects of treatment, effort, and the interaction of treatment and effort. Moreover, they can help experimenters identify when measured treatment effects are affected by erroneous beliefs and inappropriate effort provision.

How Prediction Markets Can Save Event Studies

This review paper articulates the relationship between prediction market data and event studies, with a special focus on applications in political economy. Event studies have been used to address a variety of political economy questions - from the economic effects of party control of government to the importance of complex rules in congressional committees. However, the results of event studies are notoriously sensitive to both choices made by researchers and external events. Specifically, event studies will generally produce different results depending on three interrelated things: which event window is chosen, the prior probability assigned to an event at the beginning of the event window, and the presence or absence of other events during the event window. In this paper we show how each of these may bias the results of event studies, and how prediction markets can mitigate these biases.

Prediction Markets for Economic Forecasting

Prediction markets - markets used to forecast future events - have been used to accurately forecast the outcome of political contests, sporting events, and, occasionally, economic outcomes. This chapter summarizes the latest research on prediction markets in order to further their utilization by economic forecasters. We show that prediction markets have a number of attractive features: they quickly incorporate new information, are largely efficient, and impervious to manipulation. Moreover, markets generally exhibit lower statistical errors than professional forecasters and polls. Finally, we show how markets can be used to both uncover the economic model behind forecasts, as well as test existing economic models.

Information (In)Efficiency in Prediction Markets

We analyze the extent to which simple markets can be used to aggregate dispersed information into efficient forecasts of unknown future events. From the examination of case studies in a variety of financial settings we enumerate and suggest solutions to various pitfalls of these simple markets. Despite the potential problems, we show that market-generated forecasts are typically fairly accurate in a variety of prediction contexts, and that they outperform most moderately sophisticated benchmarks. We also show how conditional contracts can be used to discover the markets belief about correlations between events, and how with further assumptions these correlations can be used to make decisions.

Examining Explanations of a Market Anomaly: Preferences or Perceptions?

Abstract This paper compiles and summarizes the theoretical literature on the favorite-longshot bias, an anomaly that has been found in sports betting markets for over half a century. Explanations of this anomaly can be broken down into two broad categories, those involving preferences and those involving perceptions. We propose a novel test of these two classes of models that allows us to discriminate between them without parametric assumptions. We execute these tests on a new dataset, which is an order of magnitude larger than any used in previous studies, and conclude that the perceptions model, in which bettors overestimate the chances of small probability events, provides a better fit to the data.

Accounting for Behavior in Treatment Effects: New Applications for Blind Trials

The double-blind randomized controlled trial (DBRCT) is the gold standard of medical research. We show that DBRCTs fail to fully account for the efficacy of treatment if there are interactions between treatment and behavior, for example, if a treatment is more effective when patients change their exercise or diet. Since behavioral or placebo effects depend on patients’ beliefs that they are receiving treatment, clinical trials with a single probability of treatment are poorly suited to estimate the additional treatment benefit that arises from such interactions. Here, we propose methods to identify interaction effects, and use those methods in a meta-analysis of data from blinded anti-depressant trials in which participant-level data was available. Out of six eligible studies, which included three for the selective serotonin re-uptake inhibitor paroxetine, and three for the tricyclic imipramine, three studies had a high (>65%) probability of treatment. We found strong evidence that treatment probability affected the behavior of trial participants, specifically the decision to drop out of a trial. In the case of paroxetine, but not imipramine, there was an interaction between treatment and behavioral changes that enhanced the effectiveness of the drug. These data show that standard blind trials can fail to account for the full value added when there are interactions between a treatment and behavior. We therefore suggest that a new trial design, two-by-two blind trials, will better account for treatment efficacy when interaction effects may be important.

Decision Theoretic Approaches to Experiment Design and External Validity

A modern, decision-theoretic framework can help clarify important practical questions of experimental design. Building on our recent work, this chapter begins by summarizing our framework for understanding the goals of experimenters, and applying this to re-randomization. We then use this framework to shed light on questions related to experimental registries, pre-analysis plans, and most importantly, external validity. Our framework implies that even when large samples can be collected, external decision-making remains inherently subjective. We embrace this conclusion, and argue that in order to improve external validity, experimental research needs to create a space for structured speculation.