Experimental Design
We survey a representative sample of 3,000 people from Hong Kong. The survey includes five categories of questions. First, we ask about their perception of liquidity assets. Second, we ask about their perception of payment methods. Third, we ask for their allocation of liquidity across liquidity assets. Fourth, we ask for their choice of funding method for each payment method. Last, we ask for their choice of payment method in various transaction situations.
We randomly assign the sample into four groups of size 750: control, plain treatment, security-perception-enhanced treatment, and ecosystem-enhanced treatment. The control group and treatment groups differ in the choice sets in the questions; e-HKD is included in the choice set for the treatment groups but not for the control group when we ask for their allocation of liquidity assets, choice of liquidity assets, and choice of payment methods. The control group and plain treatment group have the same information provided. Three treatment groups differ in the information provided at the beginning of the survey.
In the plain treatment and control group, e-HKD is explained as a digital asset that is free from default risk.
In the security-enhanced treatment group, in addition to this, it is stressed that e-HKD accounts and transaction records are managed in a regulated secure system operated by commercial and central banks like a bank deposit in the savings/current account of a bank. It is intended to give the impression that e-HKD is more secure and regulated than other non-bank commercial money.
In the ecosystem-perception-enhanced treatment group, in addition to this, it is emphasized that e-HKD will have an ecosystem like that of a bank deposit that enables payment in any situation like salary receipt, tax payment, and securities settlement. It is intended to give the impression that e-HKD is not merely money for retail payment, but will be an infrastructure of the entire payment system as well as bank deposit in the existing payment system.
We have two random parameters for respondents. This variation is aimed to identify the price elasticity of liquidity asset holding and payment method choice. In questions 17 to 20, we have bank deposit interest rates XX%. XX is sampled from the uniform distribution [0, 0.3] (Over the past five years, the average annual interest rate for bank deposits has hovered at approximately 0.15%.). The interest rates are the same for all mentioned questions within each respondent's answers, meaning that only one XX needs to be sampled for one respondent. In questions 21 to 29, we have credit card cashback rates ZZ%. ZZ is sampled from the uniform distribution [0, 4] (The average credit card cashback rate for major banks is 2%.) The cashback rates are the same for all mentioned questions within each respondent's answers, meaning that only one ZZ needs to be sampled for one respondent.
Before conducting the main experiment, we carry out a pilot study with a sample size of 100 participants. After running the pilot study, we observe the results and modify the questionnaire. Then, we submit the revised design and pre-analysis plan to the AEA Randomized Controlled Trials (RCT) Registry.