Why Philip Hans Franses had students look in wallets at a DIY store

Eureka! Every scientist contributes a small amount to the sea of knowledge that humanity already possesses. It always starts with wonder, followed by research, discovery and results. In this series, scientists talk about their eureka moments.

A 300-euro note does not exist. Yet in 2002, a German cashier was willing to give change for one. For econometrician Philip Hans Franses, the incident led to research that is still relevant today: why are our banknotes and coins divided into the denominations they are, and which ones do we really need?

The wonder

“I have been fascinated by predictions that are genuinely based on something ever since I completed my PhD. I think it is important to be able to show how you arrive at a prediction, so that others can check your calculations and repeat the research. I completed my PhD in 1991, among the first at the Tinbergen Institute, and began working on time-series analyses and predictions of variables with seasonal patterns. That fascination has only grown, especially now with generative AI. We econometricians have actually been studying what these models do – making predictions based on large datasets – for 20 years.”

The eureka moment

“It was 2002, and the euro had just been introduced. People were not yet entirely used to the new denominations – the mix of coins and banknotes. I read a report on NOS Teletekst about someone who wanted to pay with a 300-euro note at a petrol station in Germany. Of course, that note does not exist, but the cashier was willing to give 250 euros in change. I thought: how did we arrive at the denominations we use, and why did it nevertheless seem logical to someone that a note like that might exist? Having such a good question occur to me all of a sudden really felt like a eureka moment.

“The euro comes in denominations of one, two, five, 10, 20, 50, 100, 200 and 500 euros, in both coins and banknotes. It had already been calculated that this series was optimal, because it allows transactions to be carried out as efficiently as possible using as few different banknotes and coins as possible. In other words, as little money as possible changes hands in a payment. But whether people actually do this efficiently had not yet been studied.”

The research

“Together with a PhD candidate from De Nederlandsche Bank, I set up a study into how people use cash. It was fun. We sent student assistants to Gamma to simply ask people if they could take a look in their wallets. Then we observed how they used their wallets to make a payment.

“We compared that behaviour with the most efficient way to pay. For example, if you have to pay 15 euros and 2 cents, you do so with a 10-euro note, a 5-euro note and a 2-cent coin, because any other combination takes more than three actions. So for each amount, there is a combination of banknotes and coins that is the most efficient. That allowed us to count which denominations appeared most often in transactions.

“We saw that people do indeed want to pay as efficiently as possible. We also saw that we could manage without the note worth 100 euros. Even the 10-euro note turned out not to be that important, but the 20- and 50-euro notes proved crucial.

“The fun thing is that we backed up the theory with data by playing Monopoly in my office. In Monopoly, you can see exactly how much money someone has, which meant we could keep track of what they had in their wallet, what they had to pay and how they paid. Then we played games in which we removed the 20-euro note, followed by the 50- and 100-euro notes, to see what happened. What we found supported our theoretical model: people do not have specific preferences for any particular note, and try to pay as efficiently as possible.”

The aftermath

“It is useful to know which notes must not be missing when you restock cash machines, and the central bank can adjust the mix of denominations in circulation accordingly. Our research attracted a lot of media attention, in newspapers, on the radio and in the economics journal ESB.

“I have done various follow-up studies, for example on whether people found it easier to spend euros than guilders. It turned out they did, because the amounts seemed smaller. And the research is still relevant: people continue to make cash payments. My research also came in handy when a newspaper reported two months ago that cash machines would henceforth also dispense 100-euro notes. Apparently, they are now more often needed to make payments efficiently. Twenty years ago, we thought that note was unnecessary, but that is no longer the case.”