@techreport{oai:grips.repo.nii.ac.jp:00001702,
 author = {MUNRO, Alistair and D’EXELLE, Ben and VERSCHOOR, Arjan},
 note = {https://www.grips.ac.jp/list/jp/facultyinfo/munro_alistair/, In the context of investment decisions, "contingency" refers to the influence agents may exert over the probability distribution of returns on investment. Often, contingency is difficult to detect and investment decisions are influenced by recent experience of (non-)contingency. To investigate the behavioural influence of prior (non-)contingency on investment decisions, we conduct an economic experiment in rural Uganda. Subjects are asked to invest any amount they wish of their endowment, with success dependent on whether they are correct in detecting the heavier of two objects. In one task, there is contingency: trying hard to detect the weight difference should influence the success probability. In another version, there is non-contingency: the weight difference is below the differential that humans are able to perceive. To investigate the effect of prior experience of (non-)contingency we experimentally vary the priming of (non-)contingency with a guessing game organised before the investment tasks. Our main  finding is that priming contingency raises investment in the contingency condition. We find in addition that stated perceptions of confidence are also affected by priming contingency. In both cases, the effect is mediated by individuals' risk aversion. Individuals who are less risk averse respond more positively to priming contingency. We conclude that alertness to contingency matters for investment decisions, the more so the less risk averse people are., JEL Classification Codes: C93, D03, D81, O13, The research documented in this paper was  financed by ESRC-DFID grant ES/J008893/1. Some of Alistair Munro's expenses were supported by JSPS KAKENHI Grant Number 25101002.},
 title = {Investment behaviour when agents’ influence on the success probability is hard to detect: An experiment}
}