Researcher stories: Mathematical models can also describe human behaviour
2022-12-16 11:35:15Each week from September, we have featured a distinguished Corvinus researcher who has recently received recognition for their work. This time you read about Barna Bakó, researcher of the Institute of Economics who believes that everything can be described with mathematical models, provided there are sufficient data.
“I would be lying if I said that I planned a research career, because it was not the case. I was an active climber during my student years, so I was looking for a job where one could easily manage their free time. The researcher’s career gave me that, and I was able to travel to China, Nepal and India to climb mountains and do my work at the same time, because you can think anywhere,” he points out right at the start of the interview, so that no definite planning should be assumed for his progress to the 2021 Corvinus Research Excellence Award in the senior lecturer-researcher category.
Plenty of data, accurate prediction
He attended two universities simultaneously, and finished the mathematics course at ELTE before Corvinus.
I applied for the economics theory degree programme, but since there were only two applicants that year, it did not start. Instead, I studied economic policy and minored in operation research.
His studies gave him a strong mathematical foundation, while he was actively focused on people’s everyday behaviour and decisions. “I was full of questions that I liked to think about, whether I was climbing mountains or spending my time elsewhere. Why is Santa Claus? Why is it embarrassing to pay with a hot food voucher on a first date?” – Barna talks about what he still likes thinking about up to the present day.
As regards his ars poetica, he quotes the economist David Friedman: “it is the economist’s job to sail the ship of rationality and conquer the world”. He is convinced that every phenomenon in the world has a rational and explainable aspect.
Barna is most interested in the data available in a large sample and the models that can be created from them, because it makes people’s behaviour not only describable but also predictable. “Of course, there are always exceptions and individual cases, so there will never be 100 per cent accuracy in any model made by us, but we always strive to be as clear and accurate as possible in our predictions.”
In 2009, he spent a whole year in the Netherlands, which was a decisive period for him. He highlights his experience there as an example of how incentives influence human behaviour, and what he has seen in his fellow doctoral students due to the pressure resulting in the repayment of their study grants.
In the Netherlands, the rule says that students must complete their PhD in 3-4 years, otherwise they have to refund thousands of euros. I found that because of that my peers spent more time investigating a question than we would, which led to more in-depth research later on. I think the Dutch are not any smarter or more talented, yet we often think of them as they are more successful than us. I think the secret is simple: they are surrounded by the right incentives that push them to high levels of performance.
Bubbles and gamblers doomed to lose
He speaks with enthusiasm about some of his research projects; first in an area often explored by political science, and increasingly addressed by economics recently. Over the last 10-20 years, we have spent more and more time online and, as a result, we have started to exist in social and societal bubbles. E.g., the promotion system of Amazon matches our taste and offers a particular product exactly at the price the customer is willing to pay for it. While political sciences are actively exploring this area in relation to opinion bubbles, it is less so in economics.
“We looked at the dispersion of prices in the online space and found that instead of converging, prices are dispersing over a wider interval. We think the reason behind that is really the bubble that we get into as consumers, and we have no knowledge of the whole market, so there is no reason for companies to converge their prices. Rather, they should target a specific bubble with their products,” says Barna about the conclusions of the study.
Another study of his looks at the ‘hot hand’ terminology often used in sports in the case of gamblers. The term ‘hot hand’ refers to the hypothesis that if a basketball player makes a series of successful shots, a view will develop that they cannot make a mistake that day, and therefore will be more likely to score on their next attempt.
Based on data of a large online gambling provider, Barna, and his fellow researchers, investigated whether this trend really exists. “This is a very actively researched area, and, to the surprise of many people, a 2014 study found that there really are such things as hot-handed gamblers. Since then, several papers have questioned the related result, and our analysis, based on a large sample, i.e., millions of observations, refutes it”.
Don’t we cycle if we cannot use Uber?
Their paper, analysing the correlation between Uber’s withdrawal from Hungary and the use of MOL Bubi was published in 2020, and was an unexpected success. “With my co-authors, we never work on a subject just to make the world talking about it. Yet, this article was suddenly being shared on Twitter.”
When Uber withdrew from Hungary in 2015, it caused a kind of shock on the market and in society. With his fellows they examined how did the exit of Uber impacted the use of MOL Bubi and, according to Barna, the data clearly reflected its effect:
The interesting issue is always what is behind the data. And from the data we see that quite a few Bubi users used the bike-sharing service complemented with Uber. One of the consequences of that is that if Uber is not available, they will not use Bubi either, but, e.g., will switch to their own car.
Barna always elaborates the topic first, and only then starts looking for a place to appear. Certain topics require that the adoption process does not take too long, because there is a risk that the topic will become outdated: “A good friend of mine has already received feedback from a prestigious journal that the article is very good, but the examples are outdated. Well yes, because by the time the review took place, four or five years had passed,” he talks about one of his experiences. In his view, the better journal the researcher targets, the longer the review process takes.
Constant summit attack
He used to spend months in the mountains every summer during his student years, but lately he has had increasingly fewer opportunities to do so, but when he does find the time, he loves to climb.
As he explains, from a distance climbing is always an attractive activity until you reach the base camp. After that, you question every single step: “Shall I turn back now, or at the next one? Then somehow you still don’t turn back. And when you reach the summit, you are overcome by rare emotions. Even if not necessarily only good things happen to you in the process because of the adverse circumstances. And then, when you come down the mountain, you find yourself thinking of other summits”.
From a distance, there seems to be a lot of similarities between conquering mountains and the publishing process, but Barna rather smiles at this suggestion:
This is an interesting parallel. You would have to look at the data, but as I am the only sample, I would rather not draw any far-reaching conclusions from this amount of data.
Author: László Tucsni, Corvinus Communication