Essays in Economics

Abstract

This thesis examines various topics of individual choice and welfare in economics. In the first chapter, I examine the choices of workers and business in coal mining. Coal mining is a dangerous occupation where costly fatalities and disasters may increase future accident costs. We use occurrences of deaths as shocks that affect the tradeoff between mineral output and safety. We find that government inspections and penalties increase after fatalities, and less-severe accident rates decrease by 10%. For mines in a disaster-affected state, less-severe accident rates decrease by 23%, and fatalities by 68%, saving up to $2 per hour in accident costs, with limited evidence suggesting that mineral productivity falls by 7%, or $14 per worker hour, and that the number of managers employed increases by 11%. In the second chapter, I examine how the distribution of welfare from decades of technological improvement in electronics, have benefited consumers. To do this, I examine the welfare gains to different income cohorts from the development of multiple categories of electronic products. Income dependent preferences are estimated in a dynamic model of demand. Key utility parameters, unique to income cohort, are identified from moments created using micro-level data containing purchase and demographic information. My results suggest that the benefits from the rise of electronics products may be far more egalitarian than conventional measures of inequality. Welfare gains to consumers in the bottom third of the US income distribution average approximately $1,000, while gains to the top third benefit are about $2,500. This is twice as equal as related measures of consumption inequality. In the last chapter of this dissertation, I examine the influence of early random outcomes on the choices and later success of players in a major online multiplayer game. Players misattribute early random shocks to the value of an important choice. I find that players are significantly less successful in subsequent matches when making the same choice that benefited from positive shocks early on. This effect, which occurred in just a handful of initial matches, lasts over the year long career of a player, spanning hundreds of matches of the game. The setting of the game provides a rich dataset, and is also a rare empirical example of a multi-armed bandit problem.