Thread by @Jess_Hoel

Hey #EconTwitter, THE (preliminary) RESULTS ARE IN!

Is cash fungible?

Do people exhibit the same risk and time preferences when offered money in cash or money in Venmo?

Do people exhibit the same risk and time preferences over round numbers and pennies?

No, No, and No.

Some background: In the second half of my experimental class, we designed, ran, (quickly) analyzed, and today we presented a new experiment. I pitched my idea 13 days ago. The students pitched theirs 11 days ago. We voted. One of their ideas won.

https://twitter.com/Jess_Hoel/status/1072638297242042369

$2688 and 18% of the student body later, here we are.

I withdrew 18.8 POUNDS of cash and coins from the bank.

(Forget all those “cookies get you better evals” studies, have you ever tried literally BUYING better evals? Apparently I’m doing that now.)

So, how much money ARE people willing to give up to get their payment in Venmo instead of cash?

We asked people: “Would you prefer $A in cash or $B in Venmo?” where A and B varied between $11 and $9.97. Details in the image below.

When cash and Venmo were offered at parity, 66% preferred payment in Venmo. When cash paid $0.01 more, 45% of people still preferred Venmo. When cash paid $0.19 more, 24% still preferred Venmo. When cash paid $1.03 more than Venmo, 14% still preferred Venmo.

How much does this affect measured risk preferences? We asked people: “Would you prefer a 50-50 gamble between $X and $Y, or would you like $Z for sure?”, with numbers chosen to bracket the same CRRA parameters across 5 tasks. See choices in image below.

When payments are made by Venmo, people are less willing to take the gamble and more likely to choose the certain payment. The difference is statistically significant when the certain amount is less than the expected value of the gamble.

This surprised me. I thought people would behave more risk neutrally when offered payments in Venmo. I thought the Rabin (2000) critique (unreasonable risk aversion over small stakes) would apply to cash and not Venmo. I was wrong.

The students were not surprised by this result. They said that when offered smaller certain amounts in cash, they didn’t really value it, so they were more willing to take a chance for bigger money.

But when smaller certain amounts were offered in Venmo, they were afraid they would be disappointed if they took the gamble and lost, and they valued the smaller amount in Venmo, so they preferred the certain payment.

A behavioral twist that surprised me!! Gosh research is fun.

Why did we offer 4 low stakes risk tasks? We suspected the difference in expressed preferences depends on whether choices are offered over whole numbers or pennies.

i.e. We’ll get different answers for “50-50 at $6 and $1 OR $2.50 for sure?” compared to “50-50 at $2 and $7 OR $3.79 for sure?” even though those choices bracket the same CRRA utility parameter.

The difference in expressed risk preferences between cash and Venmo is large when the gamble is offered over whole numbers and the certain payment is offered in pennies. The difference is much smaller when all payments are offered in whole numbers.

We haven’t had time to think through the time preference results, but for what it’s worth, here are the preliminary charts.

Obviously we need time to analyze these data properly, but those are our preliminary results!

Now I’m going to @ some people who have helped or expressed interest so far. What do you think, #EconTwitter?

@mikeanton13 @SMasschelein @etjernst @c_a_gravert @jt_kerwin @estebanjq3 @jhaushofer @KeithMMEricson @SylvainCF @paulgp @wsandholtz @KareninKenya @BerberKramer

Also, feel free to @ me with suggestions and commands to make my graphs look prettier. I know this has been a hot topic on Twitter these last 6 months (I hear there are t-shirts?), but my youngest baby was born 5.5 months ago so I haven’t been paying attention.

BIGGEST of shout outs to my class. This was the most fun I’ve had teaching a class, and I am so proud of all we accomplished in less than two weeks!

This is why it’s great to teach @ColoradoCollege.