higher price to use the road than those who are more willing to postpone trips. Our result
is similar to the delayer pays principle, in which drivers who value consumption most pay
other drivers who are willing to postpone their trips (Rafferty and Levinson (2009)).
This paper is divided into five sections. The next section explains the experimental
designs and set-ups, section 3 discusses the theoretical predictions of the experiment,
section 4 discusses the results of the experiment and section 5 concludes.
2. Experimental designs and procedures
We conducted a two-sided repeated double auction market with two-way traders. There
were total of twelve periods in the double auction market; the first two periods were trials
and are not reported. Each auction period lasted for 180 seconds. Thirty-three graduate
students from different faculties at the Universiti Sains Malaysia participated in the
experiment. Each period of the experiment consisted of three stages: (1) permit
allocation, (2) auction market and (3) decision to use the permit. The auction market was
introduced to correct for any initial misallocation of permits in the first stage. Banking of
permits was not allowed in the experiment.
In each period in the allocation stage, the “government” issued permits randomly to the
drivers for free. Each driver was initially endowed with a one-off $20 at the beginning of
the experiment and was told the “value” of using a road, which was drawn independently
for each period from a discrete uniform distribution over the range of [1,30]. The value
was a private value and traders did not know the values of other traders.
In the second stage, if the driver decided to make a trip, s/he needed a permit which s/he
had to buy from the auction market if s/he was not allocated a permit in the first stage.
If the driver decided to sell the permit to other drivers in the auction market, the price had
to cover the cost of not using the road. The introduction of the auction market corrected
the misallocation in the first stage by allowing transactions between high value buyers
and low value sellers. If these transactions occurred, the exchange surplus, i.e. V- cost,