|
Edward Augenblick
Job Market Candidate
Stanford University
Department of Economics
579 Serra Mall
Stanford, CA 94305
650-804-5281
ned789@stanford.edu
|
|
|
|
|
Research
Pay-As-You-Go: Theoretical and Empirical Analysis of a New Auction Format (Job Market Paper)
This paper investigates a new auction mechanism called the "pay as you go" auction, which is similar to a war of attrition.
In this mechanism, (1) placing a bid costs a fixed (non-refundable) amount, (2) each bid raises the price of the good by a small fixed increment,
(3) a player wins if their bid is not followed by another bid within a short fixed amount of time.
Using a game theoretic analysis, I find a unique set of equilibrium hazard rates that satisfy a variety of desirable conditions.
I test these predictions using two datasets on 166,000 auctions and 14 million individual bids.
Significant overbidding occurs in these auctions, yielding average revenues of over 62% the value of the good
(for a total of 18 million Euro profit over the auctions).
While the empirical hazard rate is near the predicted hazard rate at the beginning of the auction,
it deviates more as the auction progresses, leading to extremely high (~250%) "instantaneous" profits in later periods of the auction.
I attempt to explain this result with a model of sunk costs. Using the bid-level data, I show that the expected value of
bidding increases as a player's experience increases, largely due to the increased strategic use of "aggressive" bidding.
Finally, I attempt to address the long-term prospects of the market for these auctions.
Using high frequency auction supply and user data, I estimate the current and optimal supply of auctions for a given number of users.
This analysis suggests that the structure of the auction creates barriers to quickly developing a large userbase,
allowing the most-established competitor to continue making large profits in the medium-term. This analysis is supported by auction-level data
from five competitors.
The Effects of Choice Context on Decision-Making: An Application to Voter Fatigue (with Scott Nicholson)
This paper exploits a natural experiment in which choice fatigue is
isolated as an explanation for the usage of heuristics in decision-making. The empirical
application provides evidence that voters who see a given contest relatively
further down the ballot are more likely to vote "no" and to abstain. Within-election
exogenous variation in ballot position is primarily due to differences in the set of
overlaying local political jurisdictions. Our central finding is that lowering a proposition
10 positions on the ballot increases precinct-level "no" votes and undervotes by
1.3 and 0.7 percentage points, respectively. Interestingly, 8 of 124 statewide propositions
in the dataset have winning margins within the range of the "no" estimate.
The empirical analysis employs a unique precinct-level panel dataset of votes cast
for the entire menu of federal, state & local ballot choices in primary and general
elections between 1992 and 2006 in San Diego County, California. Implications of
the results range from the dissemination of information by firms and policy makers
to the design of electoral institutions and the strategic use of ballot propositions.
Reference Points in Political Fundraising (with Jesse Cunha)
This paper presents evidence from a field experiment on the impact of different types of campaign solicitations on contribution behavior.
We sent one of three mailings to 10,000 potential donors of a Democratic congressional candidate in 2008.
The mailings contained either a reference to the average contribution to Republican candidates, the average contribution to Democratic candidates,
or no reference to contribution behavior (the control group). The group that received the Republican reference contributed at twice the rate of the
control group and the group that received the Democratic reference contributed one-and-a-half times the rate of the control group.
The distribution of donations was also affected by the treatment. Particularly, the group that received the Republican reference contributed more,
with a mode of twice the stated Republican average contribution. This suggests that social comparisons can have a significant impact on contribution behavior.
|
|