The Way People Lie in Markets

How to cite

Tergiman, Chloe; Villeval, Marie Claire. 2019. "The Way People Lie in Markets." AEA RCT Registry. AEARCTR-0004732. https://doi.org/10.1257/rct.4732-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Chloe Tergiman* , Marie Claire Villeval

* Primary investigator

Registry ID
AEARCTR-0004732
DOI
https://doi.org/10.1257/rct.4732-1.0
First registered
Last updated
Status
Completed
Trial dates
2018-09-01 – 2019-06-27
Countries
France ()
JEL codes
C91, D01, G41, M21
Registry record
http://www.socialscienceregistry.org/trials/4732

Abstract

The goal of this paper is to investigate how people lie in the context of markets with asymmetric information and when lying has varrying intensities in terms of probability of getting caught. Some lies are detectable ex post with certainty, but others guarantee plausible deniability. We look at which market mechanisms affect the amount and types of lies.

Intervention

We have 4 treatments, where competition and reputation are turned on and off independently.

Experimental design

Fixed Pair treatment (baseline): In each period, a project manager (PM, hereafter) is matched with an investor and pairs are randomly rematched at the beginning of each
new period. We now describe the timing of each period.
(i) First, Nature randomly draws a set of three cards for each PM. The cards convey information on the quality of the PM's projects. Each card has a (independent) 0.5 probability of displaying a star, which indicates a successful project. If the card does not display a star, it is
blank. Only the PM can observe his three cards and see how many of display a star. The total number of stars  is then in the (0; 1; 2; 3) set.
(ii) Then, the PM sends a cheap-talk message m to the investor regarding the number of cards that display a star, where m can be 0; 1; 2; 3.
(iii) The investor receives an endowment of 100 tokens. After observing m, the investor decides on an action a in (Invest;Not Invest). If a= Invest, the entire endowment has to be invested.
(iv) Nature randomly draws one of the PM's three cards, which determines the quality of the selected project. If the selected card displays a star, the investment is a success; if the card is blank, it fails.
(v) Finally, both the PM and the investor learn whether a star was on the card nature drew, regardless of whether the investor invested
or not. A history box is also displayed on the subjects' screen. This history box lists (m, a, whether a star was drawn by nature)
for each past period.

We have 3 other treatments:
Fixed Pairs: PM and investor are in a fixed match for all 27 periods
Random Triplets: 2PMs and 1 investor are randomly rematched in each period.
Fixed Triplets: 2PMs and 1 investor are in a fixed match throughout the 27 periods.

Design details

Randomization method
subjects randomly show up for a given treatment.
Randomization unit
experimental sessions.
Clusters
about 35 observations for each type of subject.
Observations
400.
Arms
at least 35 observations for each type of subject.
Primary outcomes
The frequency of extreme lies (detected ex post with certainty), high risk lies (detected ex post with 66% chance), low risk lies (detected ex post with 33% chance) and deniable lies (undetectable).
Secondary outcomes
We will also use individual characterisitics (Machiavellian scores as well as gender and how much people value efficiency) to show the robustness of our results.

Cited by

No citation data available for this registration yet.

Keywords