Our PhD Students Have Got Talent

Handling new tasks in a different function* is attractive because it brings broader knowledge and better understanding of the business. Cross-functional transitions force us to deal with unusual situations and new audiences and train us to see a problem from multiple perspectives, so that we can make better decisions in future. But at the same time, they pose a significant challenge because skills that make us successful in the current job might not be relevant in a new function. Broadly speaking, experience in multiple functions may make us well-rounded individuals but also leave us dilettantes in every field that we have tried.

Should we specialize or diversify to boost our career?

To answer this question, we need to understand what knowledge and skills one acquires and loses by engaging in cross-functional rotations and what knowledge and skills are needed to perform high level jobs. Importantly, managerial skills are multidimensional: some of them depend on the function, such as domain expertise, while others are universal, such as negotiation skills or strategic thinking. To disentangle those different components, we look at two dimensions of the managerial experience: in different functions and in different managerial roles.

Experience in multiple functions improves one’s chances to reach higher organizational levels but might be not as beneficial earlier in one’s career.

From functional perspective, we found that experience in multiple functions improves one’s chances to reach higher organizational levels but might be not as beneficial earlier in one’s career. Because job complexity increases with hierarchal level, domain expertise becomes less useful at more senior levels, while the need for broader knowledge and more general skills increases. Accumulating various work experiences stimulates integrative complexity, the ability to consider and combine multiple perspectives, and improves strategic thinking, a critical competence for senior managers. However, for lower-level managers, who mostly engage in operational rather than strategic tasks, it makes more sense to specialize to deepen domain expertise, which helps managers to faster identify task-specific problems and find better solutions within their area of expertise.

The question is even trickier for managerial roles

First, let’s define what is a managerial role. A managerial role is a predictable set of behaviors one can observe in daily activities and associated with a particular managerial position in an organization. In other words, it is a way of managing. A manager has three major ways of managing related to action, people, and information. Let me explain.

1)    Action:

In the first way, the necessary action is performed by the direct involvement of a manager. It represents a “doer” role, which includes supervising projects, solving problems, making decisions, and negotiating deals.

2)    People:

The second way implies managing indirectly by encouraging people to take necessary action. It is a “leader” role, which involves motivating and developing people, building teams, maintaining culture, and creating useful connections. 

3)    Information:

The third way is even less direct: a manager collects, processes, and shares information, but also controls information flows to influence people and ensure that they take necessary action. It is an “administrator” role, which might be exercised by creating systems, designing structures, implementing procedures, or imposing directives. 

Although managers might have a large repertoire of roles, they tend to gravitate towards one of these types. Notably, dominant managerial roles vary across functions and hierarchical levels. Therefore, by moving cross-functionally, one may acquire experiences in either similar or different managerial roles, and these experiences may be either beneficial or harmful for potential promotion. 

We found that at the lower managerial levels, dominant managerial roles are people focused. At the higher levels, the focus shifts from managing people to direct action, while information processing is useful at all levels. These findings suggest that managers with “doer” profiles are more likely to achieve higher organizational ranks than those whose experience is concentrated in interpersonal interactions.

Managers with “doer” profiles are more likely to achieve higher organizational ranks than those whose experience is concentrated in interpersonal interactions.

Methodology

We studied organizational career histories of 422 managers in a large pharmaceutical company. We collected data on all their job transitions within the company and studied job descriptions to better understand their roles in every appointment. We classify each job in three dimensions: the functional affiliation, the managerial role, and the hierarchical grade. While functional affiliation and hierarchical grade are clearly stated in the company’s HR records and aligned with its job catalogue, managerial roles are more subtle. So we had to carefully study the job descriptions to determine what types of activities a manager performs in a particular job and to infer what role(s) he or she plays. Based on this data we calculated the level of functional specialization of each manager and statistically tested how it influences one’s chances to be promoted from lower to middle management and from middle to senior management. We also looked at managers’ experience in different types of managerial roles and identified what roles are most valuable at different managerial levels.

Interview with Knowledge@HEC

Based on these results, what do you recommend to someone developing a managerial career and interested in achieving higher organizational ranks?

First, build a solid foundation in one function, but when you reach the middle management rank, think about expanding your experience outside of your areas of expertise. 

When planning a cross-functional move, think about what the dominant managerial role in a new function would be. It might sound counter-intuitive, but the more your experience is concentrated in “doer” and “administrator” roles and less in “leader” roles, the more skills necessary for higher-level jobs you will develop.

What do these results mean for organizations that think about developing their future leaders through cross-functional job rotations?

Not every cross-functional move is beneficial. The best leaders grow out of specializing generalists, those who develop broad knowledge and cognitive complexity though experience in multiple functions but keep concentrating in managerial roles most valuable at higher organizational levels. Therefore, paying close attention to managerial roles when developing a job rotation program is an essential condition for its success.

The best leaders grow out of specializing generalists.

*Function means functional unit in an organization: Sales, Marketing, Finance, IT, etc. Each function is associated with different tasks and requires different skillsets. Diverse experience means experience that spans multiple functional units.

Can you explain your field of research?

Recently, I work in the theoretical side of decision under uncertainty, focusing on building formal frameworks and proposing elicitation methods that I think are useful for people to gain a deeper understanding of the topic. The other side of the field tries to empirically uncover patterns of behaviors that people exhibit under uncertainty.

Who is more #risk averse? To understand, it’s required to study people with the same risk attitude. Yet #HECPhD Fan Wang brings an improved definition of #ambiguity attitude so experimental #researchers can now compare and cluster people with different #risk attitudes - #DTEAconf pic.twitter.com/x88pIlFmNa

— Knowledge@HEC (@HECKnowledge) June 3, 2019

How do you define ambiguity attitude and risk attitude?

Risk is the type of uncertainty that is precisely quantifiable, like the chance of heads when tossing a fair coin (which is 50%). In comparison, ambiguity is un-quantifiable uncertainty, like the chance of success of a truly novel business idea. Risk attitude and ambiguity attitude are usually measured in terms of how much the decision maker evaluates risky bets and ambiguity bets, respectively. The decision maker with lower evaluations is deemed more risk (and ambiguity) averse. 

Risk is the type of uncertainty that is precisely quantifiable, like the chance of heads when tossing a fair coin, while ambiguity is un-quantifiable uncertainty, like the chance of success of a truly novel business idea.

What do you bring to the field of decision sciences?

Let’s go back to the example of a novel business idea, and suppose that we want to compare the ambiguity attitudes of two venture capitalists (VC) who are interested in financing the idea. Again, the standard method suggests comparing their monetary evaluations of the idea. This is reasonable, but it has been shown that this makes sense only if the two VCs have the same risk attitude, that is, if the chance of success of the idea were precisely quantifiable, the two VCs must have the same evaluation. This strong requirement renders comparing ambiguity attitudes practically impossible.

I propose a method which in our example amounts to comparing the two VCs’ subjective estimates of the idea's chance of success.

In my research, I propose a method which in our example amounts to comparing the two VCs’ subjective estimates of the idea's chance of success. And I formally show that this method does not require a common risk attitude. Therefore, my paper establishes a separation between ambiguity attitude and risk attitude, which provides decision-theoretic support to many applications and empirical tests.

While many may think creativity will be a key human capacity as work processes get increasingly automated, AI-applications get more creative as well: for example, AI-painted portraits sell at exorbitant prices, pop hits feature AI-generated music, and mainstream media ‘hire’ AI-journalists to publish thousands of articles a week. How workplace AI-application will impact creative processes, crucial for innovation and organizational development, is an important question that needs rigorous scientific attention.

Portrait d'Edmond de Belamy by an Artificial intelligence software (Source: Christie's)

What can research bring to the field of creativity with AI?

Research can help find ways to foster creative work and innovation augmented by the numerous AI-provided capacities. We need research that explains how we can create and adapt artificial agents and arrange working processes such that employees do not have to be defensive against AI and be more creative.

By now, we have learned that many people believe that artificial agents – these are, for instance, AI-run apps like Siri or Alexa, or social robots you may see at an airport - are not suited for creative work or work that requires diplomacy and judgement skills. So, normally people would not care much about AI trying to compose a symphony or write a novel as long as they believe that humans are creatively superior. But we also know that when AI is involved in creative work – for example, generating image suggestions, or even creating logotypes at a design studio, people often reject its results and evaluate them lower even if the AI objectively performs just as well as a human. 

This is a problem because when employees believe that AI should not be creative and it turns out to be exactly that, they may feel threatened by it, fearing it might substitute them, and these feelings are normally not very conducive to creativity. In addition, when threatened, employees will try to protect themselves – after all, we want to feel that we, humans, the creative species, are unique. This hence may also hinder creative work.

How can your research help manage creative human-AI interactions?

Together with my advisor Professor Mathis Schulte, we investigate the differences in how people work on creative and non-creative tasks when they collaborate with AI, compete with it, or use results of its work as a benchmark for their own performance. We also investigate how these interactions are different from same situations in which a human counterpart is involved instead of an AI agent. 

Creative interactions with AI is a nascent area of managerial knowledge, and we are excited to be making first contributions to it. We show, for instance, that people use the same psychological mechanisms in comparing themselves to AI agents, even when they believe these agents can’t really compete with people creatively. We also show that what we believe about ourselves and about AI matters, and while some beliefs can make employees work harder on creative tasks, other may demotivate them. Knowing what these configurations are will help managers set up collaborations such that AI is used to its full capacity, and people blossom in creative work.

We show, for instance, that people use the same psychological mechanisms in comparing themselves to AI agents, even when they believe these agents can’t really compete with people creatively.

So, what main differences do you find between human-human and human-AI creative interactions?

First of all, we saw that people did not take as much time to work on a creative task – that is, they were not as effortful - when they collaborated with an AI as when they collaborated with a colleague. This was not the case for a non-creative task, in which people worked on average for the same time both with an AI and a human collaborator. To illustrate, in one of our experiments, the creative task was to come up with recycling ideas for a medical mask (this task could be, for example, a challenge in an engineering competition, or a topic for an entertaining article, which was the setting in our experiment). The non-creative task was to find recycling-related words in a word matrix that needed to be tested as a game for publishing online. 

People were not as effortful to work on a creative task when they collaborated with an AI as when they collaborated with a colleague.

We also found that an AI’s performance did not matter for such a decrease in effort. In a different experiment, we asked the participants to propose ideas on how to motivate their colleagues to move more or to eat better. We then showed them either a list of nonsense ideas (which really was a text generated by an algorithm!) or a list of sensible ideas. We told the participants that the list they see was either created by an AI or by someone else who also participated in the experiment. As a result, some people saw a nonsense list of ideas and thought it was proposed by an AI, and others thought that same list was suggested by a human. Regarding the list of sensible ideas, a third group saw a sensible list that was ostensibly AI-written, and the last group thought it was a human suggestion. Then, we again asked the participants to propose ideas on one of the topics. We found that no matter whether the ideas made sense or not, participants who looked at ideas that were introduced as suggested by other people took almost half a minute longer on the second task than those who thought the ideas were AI-generated. 

Interestingly, this was also the case in the non-creative task, in which participants looked for a specific character on an old-Russian birch bark, looked at the results of work of an AI or a human, and looked for another character again. Although we did not expect them to, participants who saw the results of AI’s work, both done well and failed, took less time to work on the task in the second round. 
 

The results of these two experiments suggest that in many cases, especially when we think AI has no place in creative work, we might be less motivated to exert effort on it than we do when working together with another human.

In many cases, we might be less motivated to exert effort on creative work when there is AI participation, than we do when working together with another human.

So what are the conditions in which AI can motivate us try harder on a creative task? 

This we tackled in a further experiment. In it, we again had participants think about creative ideas for the use of a medical mask or medical gloves. But, unlike in the first experiment, we first subtly made half of the participants remember that creativity is what makes humans unique, while the other half did not see such a reminder. We also told the participants that either an AI or a person came up either with many creative ideas, or with only a few. The results of these manipulations were astounding: while people who were reminded that creativity is a uniquely human trait and saw that AI came up with 20 ideas worked on average for almost full five minutes, those who were also reminded of creativity uniqueness but saw that it was a person that thought of 20 ideas took only 3.5 minutes to work on the task. We also saw that when high creative performance of AI was unexpected, people were much more threatened by that AI. 

What are the implications of these findings for practice?

These findings have very important implications for practice: what should we tell employees about AI-applications that can be used in creative work to not have them disengage as they have nothing to prove against supposedly non-creative AI? How do we motivate them to perform creatively without scaring them with AI? These are the questions that I am looking forward to solving in my future work.

You’re using some unusual online services to conduct your research. Can you describe how you recruit your participants and why these services are so attractive?

People who participated in these experiments were recruited either on Prolific, which is a rather popular British service for academic research, or on Yandex.Toloka, which is a Russian service for recruitment of individual workers on mini-tasks like data markup. 

Using both platforms is a great advantage. First, I can test whether the effects I hypothesize are universal across different cultures and nationalities. Second, as specialized platforms have recently been criticized for data quality, people that I recruit at Yandex.Toloka, which is primarily used by businesses for crowdsourcing, are naïve to experimental studies and, as the majority of these tasks are quite mechanical and dull, they are intrinsically motivated to participate in my experiments – I have received quite a few messages from participants saying that they enjoyed the “task” and would be willing to do it again. Although, unfortunately for these participants, that will not be possible to not compromise the integrity of the research, I think it says much about how people are unlikely to sabotage participation.

I can test whether the effects I hypothesize are universal across different cultures and nationalities.

Finally, Yandex.Toloka offers a great gender balance in the sample (51% of participants I randomly recruit there are women), people come from different walks of life and mostly have had quite a lot of working experience – their average age is almost 36 years. These factors make the use of both services a great strength of my research design, and I thank GREGHEC Laboratory for enabling me to use them by providing research support in the form of funding that got me almost 1200 participants.

Can you explain your recent research about the value of privacy in online bank loaning?

My recent research paper, “The Value of Privacy: Evidence from Online Borrowers”, studies the monetary value consumers attach to the privacy of their personal data. Whether and how much consumers value data privacy are not obvious questions. For example, previous studies based on lab experiments and survey data often find small to zero value for privacy. This is commonly referred to as the “privacy paradox”, i.e., the claimed intentions to protect privacy divert from how people actually behave online. 

I investigate this question in a real life context, where online borrowers have to relinquish personal data in exchange for consumer loans.

I investigate this question in a real life context, where online borrowers have to relinquish personal data in exchange for consumer loans. Because there is no official credit score in China, online lending platforms commonly require applicants to disclose extensive amount of information, in order to assess their creditworthiness. This implies that individuals borrowing from such platforms face a clear tradeoff between data privacy and credit access. In this context, their preferences for privacy can be truthfully revealed by their actual choices.

How do you quantify the monetary value of privacy?

To quantify the monetary value of privacy, I exploit large-scale field experiments, which involve around 320,000 participants, on a major Chinese peer-to-peer lending platform. These experiments randomly vary the amount of information required from applicants and the cost of borrowing. Applicants’ decision to disclose or not then truthfully reveals their preferences for privacy. In particular, the experiments allow me to measure applicants’ sensitivities to disclosure requirements and to cost of borrowing. By contrasting these two sensitivities, I back out the monetary value of privacy. The $33 value in the headline result can be interpreted as the fee reduction required to exactly offset the drop in loan demand cased by additional data questions in the application procedure.

What have you found? 

The headline result is that online borrowers in China are willing to share their social network ID and employer contact in exchange for a $33 reduction in loan origination fees. This is a non-trial amount considering that it represents 70% of the average daily salary in China. 

In addition, the platform makes 10% less in expected revenue by requiring applicants to disclose these additional data. This is because such data policy discourages a significant fraction of applicants from completing the applications.

These results suggest that maybe there is no “privacy paradox” after all. When we measure the value of privacy in a real-life context, the number can be far above zero. This paper also contributes to the ongoing debates about whether the recent privacy protection regulations, such as the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA), benefit consumers or not.

What are the implications for FinTech companies and for customer’s privacy? 

A key innovative aspect of the FinTech business model is to use customer data to generate revenue, either by using the data to improve the risk model, by profiling customers and sending them targeted advertising, or by selling the data to third parties. The overall impact of data collection on firm profit depends on how much consumers value data privacy and how useful the data is for generating revenue.

The overall impact of data collection on firm profit depends on how much consumers value data privacy and how useful the data is for generating revenue.

My paper shows that there are cases where collecting more data is socially inefficient. This is when the data items are valued to a large extent by consumers but have little economic benefit for the platforms. Therefore, it is a win-win situation if FinTech firms can identify such data items. The framework provided in this paper can be easily generalized and exploited by firms to conduct this type of analysis.
 

My research paper shows that there are cases where collecting more data is socially inefficient.
 

A word by Professor Johan Hombert, Huan Tang’s thesis supervisor

Huan’s research on privacy is making an impact because it provides a methodology to put a number on how much individuals value privacy. For example, Huan has been able to quantify that borrowers on a Chinese peer-to-peer lending platform needs to be given more advantageous loan terms worth $33 in exchange for revealing their ID on Tencent QQ, the Chinese equivalent of WhatsApp. 

Researchers previously calculated such numbers using data from surveys and lab experiments. However, such data suffer from well-known methodological problems such that results derived from these data can only be taken with a grain of salt. Instead, Huan’s methodology rely on “field data”, that is, data from actual borrowers taking actual loans, which are considerably more reliable, especially for quantitative analysis. 

Huan’s methodology can be used by FinTech platforms and more generally by B2C internet platforms to assess how much business they could gain by optimizing the personal information they require from customers. 
 
The HEC Paris Finance Department is proud to see the impact Huan’s research is making. Huan is now joining the London School of Economics as an assistant professor of finance, thereby continuing the exceptional track record of placement of HEC finance Ph.D. graduates (in recent years, HEC finance Ph.D. graduates were hired by universities such as MIT, Harvard Business School, Princeton and Wharton). 

Who are the Wall Street gurus?

Wall Street gurus are famous investors who are considered as experts because they were able to get large profits in the past and / or currently manage several billion dollars with their funds. The most famous gurus include for example Warren Buffet, Bill Ackman, Jim Chanos, or Carl Icahn. While they are often invited by different media to give their opinion on the economy and various stocks, they have fascinated individual investors for a long time and many books and movies feature their career and investments. For these reasons, individual investors listen to them and intend to mimic their supposedly superior investments. 

In my research, I focus on one of Wall Street gurus’ communication tools, the investment conferences (for example, the Sohn Investment Conferences or the Robin Hood Investors Conferences). During these conferences, Wall Street gurus are invited to share their opinions on the economy (such as the consequences of unemployment), discuss public policies (such as the impact of a change in tariffs) and, notably, pitch their favorite long or short investment ideas that they are supposed to have implemented in their own fund. They also regularly take advantage of their exposure to highlight their success and increase their reputation. For instance, some of them review their previous ideas if they were profitable.

The investment ideas and forecasts of the Wall Street gurus

I study about 100 investment conferences and more than 1,000 investment ideas between 2009 and 2018. I find that Wall Street gurus’ investment ideas are profitable and allow the investors to “beat the market” for about six months. In other words, the return of the investment ideas exceeds the performance of the Standard & Poor's 500 index during this period. While this is a good reason to listen to the Wall Street gurus, I also find that the profitability of their investment ideas decreased during the decade, implying that this is not necessarily a good strategy nowadays.

Wall Street gurus’ investment ideas are profitable and allow the investors to “beat the market” for about six months.

Additionally, I study the forecasts of future stock prices, also called target prices, disclosed by these gurus at investment conferences. When the gurus recommend the investors to buy a stock, they expect the stock price to increase by 85% on average. These forecasts are extreme as the Standard & Poor's 500 historical annual returns is only around 10%. After two years, I find the return of their investment ideas is only 25% on average. In other words, their forecasts are overwhelmingly biased. I find a similar extreme bias if I focus on their selling recommendations. In that case, they expect the stock price to decrease by around 60%.

When the gurus recommend that the investors buy a stock, they expect the stock price to increase by 85% on average.

Because they often share profitable investment ideas and can also provide guidance to analyze a stock, you should definitely listen to the Wall Street gurus. Nevertheless, their forecasts are extremely biased and should always be interpreted cautiously. Always keep in mind that they are supposed to pitch stocks that they own themselves. Therefore, if many investors buy the stock, the stock price will increase, and the gurus will directly benefit from this increase. They face an inherent conflict of interest.

At these investment conferences, the gurus also often claim that you should “do your own work”, that is, consider their ideas and then, decide by yourselves if the idea is worthwhile. 

“Do your own work.”

A compelling example is a discussion of David Einhorn - a hedge fund manager who predicted the subprime crisis - during one of his pitch. He notified the audience about his confusion when the stock price of a firm decreased instantaneously when people read the title of his presentation. The audience had not actually seen the slides or listened to his arguments but was already trading the stock. You should listen to the gurus but not blindly! This advice is not only worth for equity investments but apply to all assets, including the cryptocurrencies for which the manipulation through pump-and-dump schemes is frequent and substantial.

Why did both of you want to tackle this issue?

Claire Linares: I think it began when we saw the 2013 finding being published in the mainstream media, books and academic papers. We found the cellphone’s physical impact interesting enough to want to replicate this effect. It was actually to be a starting point of our research which we sought to extend into the field of creativity. But as we were conducting the studies, we were unable to find the initial effect. So this is what motivated our research.

Anne Laure Sellier: The claim that the mere presence of a phone would affect social interaction and would be deleterious to high quality relationships is formidable. Just imagine the consequences of smartphones today in negotiation rooms, at the U.N., for example, where you have leaders talking about going to war or not. We wanted to capture its impact. For theoretical reasons, as Claire said, we thought if it's true, it may hurt creativity beyond social interaction. But Claire conducted several studies and found nothing. However, don’t get me wrong: the effect is not necessarily bogus. It may have been there in 2013 in those two studies. But it certainly required replication. You cannot extrapolate about an effect based on only two studies.

As you said, the 2013 findings were widely accepted and reported on. Do you think your conclusions will enjoy the same attention?

ALS: I was looking at some articles recently showing just how bad it is to have a smartphone in business meetings. So, this perception is still very present. I imagine we’re not going to make waves with our results. It should certainly get as much media attention as the 2013 paper. But it probably won’t because intrinsically a failed replication paper rarely gets cited as recent research empirically shows. Let’s hope things will change going forward. We’re very pleased that the journal PLoS ONE chose to publish this work out of academic rigor.

You know, this is nothing new in terms of the relationship of mankind to technology. Whenever a new technology appears, there are fanciful theories that the world is coming to an end. If you look at the introduction of television, for instance, you can find articles from the 40s and 50s in the United States where men were warned to protect women and children from TV screens. Any new technology has this kind of vampire clout to it. That's kind of scary. And, rightly or wrongly, it's even more the case for cell- and smartphones because we know of cases of addiction which are made worse because smartphones are constantly at our fingertips.

The paper did involve lots of interesting elements, such as measuring pleasantness, appreciation, motivation, autonomy, control. All this research involved giving people tasks like creating a toy or telling an interesting story. How do you measure these, Claire? 

Across our two studies, we had a total of 356 volunteers split into 136 groups working on what we call a toy task. They had to discuss between themselves and then design a toy for a child from five to 11 years old. After the sessions, we had independent curators judging their originality and the other criteria pertaining to the creativity of these toys. We compared the creative measures of the toys between groups who had been exposed to the mere presence of a smartphone versus the group that had been exposed to the presence of a notebook. And we conclude by saying that there is too much concern over the impact of the mere presence of smartphones and how it can affect interaction, be it personal or professional, in terms of the professional context. Unfortunately, one result of the all-negative view of smartphones in meetings is that there have been widespread bans on smartphones in meetings. We should conduct more research on the impact of the active presence of the smartphone rather than its mere presence. For instance, what is the influence of notifications ringing during meetings, what happens to the interaction when someone checks his or her phone for messages, etc.?

You're saying that, in light of the results of your research, these policies go over-the-top?

CL: Well, certainly the media has insisted that the mere presence of a smartphone can be dangerous. We didn't find this impact. However, as we mentioned before, there are probably other negative effects caused by smartphones being in meeting rooms.

The media has insisted that the mere presence of a smartphone can be dangerous. We didn't find this impact.

Did you want to add something, Anne Laure? 

ALS: Yes. Of course, you don't want a smartphone inside a criminal court at the time of debates because you don't want anything to be recorded and get out of the closed room. This is a valid concern. But that's not what we're talking about. We're talking about those companies who require people to leave their phone outside of the room, the way cowboys left their colts outside of the saloon, just putting your phone on airplane mode is enough! And more generally, I think it is simply time for the business world to devise an etiquette for smartphone use.

In your research paper, there are a few caveats that you mention when you compare your own research to that of Andrew K. Przybylski and Netta Weinstein. For example, people have evolved in their attitudes towards smartphones and simply gotten used to their presence.

CL: Yes, we considered how we could reconcile our results with the findings of Przybylski and Weinstein. And one explanation is that time simply has passed. They published the article in 2013. So they conducted their studies in 2012 or earlier. We conducted our studies in 2018. Six years of technological progress is a lot. Furthermore, the participants in our studies were people in their 20s, this generation of people grew up with phones that were always there. Returning to your earlier point, the 2013 results may have been genuine, it’s may just be that time has passed and the mere presence of mobile phones no longer has the same impact.

***

Learn more soon in an upcoming podcast.

Crushing student debt is almost a fact of life in the U.S., where the cost of education routinely runs in the tens of thousands of dollars. While investing in a college education brings undeniable benefits, it also imposes a hefty burden on young (and even not-so-young) adults, potentially lessening their ability to buy a home, start a family or invest in a business. Not to mention that the staggering amount of student loan debt – an unprecedented $1.6 trillion in 2020 – possibly also holds back economic growth. 

And the figure is rising, as are calls to alleviate the burden of debt. 

At a time when many individuals, especially those with low to moderate income, are struggling during a pandemic-induced recession, student loan forgiveness is viewed as both fair to individuals and important for economic vitality. But do good intentions actually translate into financial benefits for borrowers?

Not all debt plans are created equal

A duo of researchers examined the financial situation of indebted households under various debt-relief policies – universal or capped forgiveness, and income-driven plans, targeted to specific borrowers. The researchers, Sylvain Catherine of Wharton University (formerly Ph.D. student at HEC Paris) and Constantine Yannelis of Chicago Booth, carried out their analysis by looking beyond student loan balances (what’s left to repay – the variable usually examined), because they felt it wasn’t truly representative of the weight of loans on individuals’ budgets. 

“Instead, we compute present values based on what people are actually repaying, which depends very much on their earnings,” Sylvain Catherine told Knowledge@Wharton.

Their conclusion? Not all plans are created equal. They have different distributional effects. In other terms, their costs and benefits are not evenly redistributed among various population groups – which may challenge some proponents’ view of fairness.

Universal debt forgiveness favors high earners, offers unequal benefits by race

Many progressives have been demanding universal debt forgiveness – i.e., (more or less) wiping the slate clean for everyone after a certain number of years, regardless of an individual’s financial situation. Such plans may be capped, discharging only a certain amount. In essence, they disproportionately favor those with higher incomes. 

“Any policy that is a universal loan forgiveness policy or a capped forgiveness policy – say forgiving debt up to $50,000 – is going to give most of the dollars in forgiveness to upper-income individuals,” Constantine Yannelis told Knowledge@Wharton. 

Specifically, under a universal loan forgiveness policy, in present value terms, individuals among the highest 10 percent of earners would receive $6,267 in forgiveness, while those in the bottom 10 percent would receive $1,276 in forgiveness. In a way, it is hardly surprising that plans are (almost by default) geared toward higher earners: Many holders of high loan balances completed graduate and professional degrees, and thus earn higher incomes.

The benefits of student loan forgiveness are unevenly distributed also by race and ethnicity, the researchers found. The average loan balances are highest among blacks, at $10,630, while those for whites are $6,157 and for Hispanics and others $3,996. After adjusting for the present value of those loans, universal loan forgiveness would lead to roughly equal average benefits for whites and blacks, but would yield significantly lower average benefits for Hispanics and other groups.

Better targeting middle and lower-income groups

In contrast, an existing loan forgiveness plan known as Income-Driven Repayment (IDR) links payment to income and forgives remaining debt after a fixed time, often 20 or 25 years. Under current IDR plans, borrowers pay 10 or 15 percent of their “discretionary income,” the part of their earnings above 150 percent of the poverty line. 

The researchers devised a scenario in which they expanded the existing IDR plan and found that it was more progressive than universal forgiveness plans. 

The researchers found that the Income-Driven Repayment was more progressive than universal forgiveness plans.

“Expanding the generosity of income-driven repayment plans, or enrolling more people in these plans, leads to the benefits of forgiveness going to the lower middle and the middle class, rather than the top percentiles of income distribution,” Yannelis told the Wharton site. 

Individuals in the bottom half of the earnings distribution would receive three-fifths of dollars forgiven, and borrowers in the top 30 percent would receive one-fifth. In terms of the racial and ethnic effects, forgiveness amounts are twice as high for blacks relative to whites and the general population. Hispanics and others see lower loan forgiveness amounts relative to other groups, however.

Designing better policies

For policymakers, these findings embody a timely contribution as they review plans to apply resources to the pressing question of debt. Indeed, the vast majority of student debt in the U.S. is directly disbursed or guaranteed by the federal government. 

Policymakers must be careful in shaping policies to deal with student loans “because they might sound progressive on paper, but they are very regressive,” Catherine noted. 

Of course, so-called distributional effects are not the only factor driving the debate, and such issues as how student loans impact career choices, hinder entrepreneurship or whether debt relief has macroeconomic consequences also must be considered.

You’ve been studying the adoption of AI by venture capitalists. Can you give us some background on your decision to investigate this topic?

In my research paper selected for the job market, I study how the adoption of artificial intelligence by venture capitalists (VCs) to screen startups affects the funding of early-stage innovative companies. The motivation relates to the fact that the past two decades have witnessed rapid growth in data availability and processing thanks to statistical techniques such as machine learning and AI. The adoption of these technologies by financial intermediaries has however raised concerns regarding their effects on investment decisions and, more broadly, on the allocation of capital. In my research paper, I focus on a key class of financial intermediaries— VCs, which are private equity investors in startups with high growth potential and play a crucial role in the financing of innovation.

Can you explain how venture capitalists use AI tools?

In recent years, dozens of VCs have adopted AI technologies for screening startups, i.e., sourcing, evaluating and selecting startups to fund. These VCs have developed their own proprietary platform which automatically tracks and scores startups in terms of future return prospects. Put simply these VCs employ algorithms to detect quantitative patterns in historical data from previous startups and extrapolate them to predict a new startup’s outcome. 

How do you identify the venture capitalists who use AI?

I develop a classification of VCs to determine whether and when they adopt AI. Using job and employee data, I identify VC firms that hire data scientists who develop machine learning algorithms for investment screening, and I call these employees AI-related employees. Using job starting dates, I classify a VC as becoming AI-empowered from the date it hires one AI-related employee. 

So, how is the use of AI by VCs changing their funding strategy?

In a nutshell, I show that VCs that adopt AI become better at identifying good quality startups, i.e., those that survive and receive follow-on funding, but only within the pool of startups whose business is similar to that developed by past companies. At the same time, VCs that adopt AI become less likely to fund breakthrough companies, i.e., startups that achieve an IPO or obtain highly cited patents. This finding is associated with an increase in the share of their investments being oriented toward startups developing businesses closer to those already tested. These results are consistent with AI exploiting past data informative about companies similar to past ones but not informative about breakthrough companies. Overall, my paper shows that AI adoption by investors might hinder the allocation of capital to breakthrough innovations.

VCs that adopt AI become better at identifying good quality startups, but only within the pool of startups whose business is similar to that developed by past companies.

How do you define and identify innovative startups vs. startups similar to past ones?

I construct a measure of “backward-similarity”. Specifically, I measure the similarity of the text we find in a startup’s business description and compare it to those of previous VC-funded startups in the same industry. So high backward-similarity startups run businesses similar to those that have been already tested by past startups. In contrast, low backward-similarity startups are more likely to be innovative and to develop novel products.

A few words on the method now: how can you be sure that your results are causation and not correlation?

To provide causal evidence that AI adoption leads to changes in VCs’ investments, I use a plausibly exogenous shock to one often cited determinant of a VC’s decision to adopt AI: the number of potential investment opportunities it faces. Indeed, given the large, fixed costs of evaluating investments and the limited scalability of VC firms, more investment opportunities make screening more onerous, creating incentives for VCs to adopt AI technology to automate screening with a view to saving time and costs. Specifically, my empirical strategy uses a quasi-natural experiment: the introduction of Amazon Web Services (AWS), i.e., cloud computing services by Amazon. This shock lowered the cost of starting new software- and web-related businesses, leading to more startup creations in specific industries and thus an increase in investment opportunities faced by VCs. 

What do you think are the big implications of AI adoption for the VC industry and the funding of innovations?

Taken together, my results show that AI adoption by VCs affects how they select their investments and, more broadly, how capital is allocated among young innovative companies. This suggests that AI adoption by investors can shape the nature of innovation and thus can have a significant impact on future growth trajectory.

Employee shareholders can represent a significant investment group in certain companies. In the US, employees own 8% of the stock in their companies. In France, around 51% of all employees have access to employee stock schemes and in 2018 their shareholdings represented approximately 3.5% of the total capital of French firms. But to invest, do employees need to love their company and approve its ethical performance?

Growing ethical concerns for investors – but what about employees?

Environmental, social, and governance (ESG) standards and the performance of companies in relation to those standards help socially conscious investors screen their investments. Such considerations are becoming increasingly important to investors. But we wanted to find out if this is true for all elements of the investor community, particularly for employee investors.

In theory, employee share ownership should be a motivational tool for workers as well as a useful savings instrument for the employee and a potential source of finance for the company. Such schemes are widely encouraged through tax breaks and other government incentives. 

Our quest was to better understand to what extent the ESG performance of a firm affects its employees’ loyalty. So, we decided to analyze the willingness of employees to invest in their business’s stock as a function of its ESG performance.

Our hypothesis was that an employee’s decision to buy the stock of their employer reflects their satisfaction with the firm’s policies. Increased employee satisfaction leads to increased employee loyalty and, this in turn, leads to an increased willingness to invest in the company through stock ownership rather than other investments.

To explore the link between employees’ investment decisions and the ESG practices of their employers, we were able to access an anonymized dataset on French employees enrolled in company-sponsored savings plans. The data was supplied from Amundi Asset Management, the leading asset manager for employee savings plans in France that manages some €66.8 billion in employee assets. It enabled us to analyze the monthly investment behavior of more than 380,000 employees.

To measure the ESG performance of the companies, we collated ESG incidents recorded in the RepRisk database. RepRisk screens media, stakeholders and third-party sources for news and information related to companies’ ESG practices providing daily counts of negative ESG news at the company level. The ESG data is also classified in distinct ESG issue categories. Essentially, we used the RepRisk negative news count as a proxy for negative ESG practices and then investigated how this relates to the investment decisions of employees.

Negative ESG news turns employee investors off

Our first finding was that employees are significantly less likely to invest in their company, or invest considerably less, following negative ESG incidents. Our analysis of the data, when controlling for other factors such as cyclic economic and company characteristics, indicates that the likelihood of an average employee investing in their company stock drops by eighteen percentage points (46% relative to the full sample mean) when the number of negative ESG incidents at the company doubles over the year.

Employees are significantly less likely to invest in their company, or invest considerably less, following negative ESG incidents.

In cash terms that is an average individual reduction in investment of €377 over the year. This is economically significant and large relative to the typical annual employee investment in their company’s stock of around €500.

But we also wanted to go a little deeper and see if we could better understand the motivation behind these investment decisions. Fortunately, the granular nature of the RepRisk ESG classifications enabled us to understand the factors more precisely behind investment behavior and to relate it to the three classes of ESG performance: Environment, Social and Governance. The results surprised us.

Me, me, me?

We found that these investment decisions did not respond to governance-related events, and, in some cases, even appeared to respond positively to negative environment-related incidents. With further analysis, this astonishing result was largely driven by employees working in more polluting sectors, whose loyalty with respect to their employers appears to increase following negative environmental incidents.

Overall, we found that the satisfaction and loyalty of employees is unrelated to the environmental performance of their companies. We also found that employees who tend to invest in Socially Responsible Investment funds and younger employees – both factors associated with greater environment consciousness – are only slightly more sensitive to negative ESG incidents involving their company.

Actually, the most important drivers of investment decisions were related to social-related reports such as overwork, low remuneration, spying on workers, harassment, employee suicides, or discrimination against trade union membership.

Most important drivers of investment decisions were related to social-related reports such as overwork, low remuneration.

Among these social incidents, those relating directly to the employee’s own working conditions were the most likely to influence investment decisions. We also found that French employees respond much more to such social events occurring in France than those happening in the same company but abroad. 

Our findings suggest that employees’ decisions are driven mostly by ESG practices that directly affect their everyday life. Personal benefits are the key determinants of employee satisfaction and loyalty with respect to their employer. Employees, it seems, are not altruistic, but focus on their own self-interest rather than wider ESG factors when considering their investment decisions.

Contrary to other studies and current views of investor decision-making, our empirical evidence suggests that overall investors tend to monitor the ESG performance of companies, and this is a significant factor in their investment decisions. However, a specific class of investors, the company employees, clearly have a tight focus only on their own well-being.

With employee activism emerging as an important phenomenon, it is important to understand whether employee interests are aligned with those of shareholders and other key stakeholders. These results can improve our understanding of the trade-off faced by firms between the benefits associated with employee shareholding and any potential tension this could create with other dimensions of the business.

We will be undertaking further research to explore this phenomenon more deeply.

Research shows that 98% of Facebook’s total revenue in 2019, amounting to $20 billion USD, was generated by the advertisements hosted on the platform. Yet social media platforms such as Facebook, Instagram and TikTok rely on content creators to attract users, and this is what attracts companies to purchase advertising on these platforms.  

Among these content creators, a sizable number are influencers, whose posts lead to increased user engagement with both influencer and platform. In 2019, 39% of all Instagram profile owners who had more than 15,000 followers were influencers. Their posts comprise the main source of platform traffic, increasing the likelihood of higher advertising revenues. More engaged users lead to a higher number of people viewing ads. Influencers are attracted to stay on social media platforms and use them more if the tools available help them increase their follower numbers. Followers equate to currency for influencers. The more followers they have, the more valuable they are.

In 2019, 39% of all Instagram profile owners who had more than 15,000 followers were influencers.

To increase user traffic, social media platforms work to launch innovations to improve services and compete with other entities, such as the launch of “story” concept. Yet, technology does not create value on its own. Those creating value are influencers on social media platforms. Technology introduced needs to be helpful for them in reaching and engaging with their communities. We wanted to investigate whether innovations do always achieve this.

The introduction of the “Story” feature: a step towards “normalizing” influencers

The “traditional” social media post is a piece of content that influencers put time and energy into creating. They are staged items that are dressed up well and designed to stay on the platform forever. They might include a feature of a new fashion item, a performance-inducing sports drink or a clever gadget. In posts, influencers show themselves at their best and are aspirational, attracting followers. To a certain degree, the follower idolizes the influencer. Influencers are not seen as “ordinary persons.” The relationship is one-way, influencer to follower.

In 2016, Instagram introduced the “Story”, aiming to increase engagement between content creators and users. Instagram believed this would in turn make the platform more compelling for advertisers and raise revenues. The Story presents the influencer with the opportunity to post impromptu content that provides a glimpse of their day, for example, “I ate this pizza,” or “I saw this animal.” This type of post would remain on the platform only for 24 hours and then disappear. The underlying idea was to create more possibilities for influencers to engage with their followers. The Story is designed to be a filler between more carefully crafted posts.

Because the influencer is seen as more of a “normal person” through self-disclosure at different level, the relationship with the followers changes.

As a less-curated item, the Story is a tool that allows the content creator to produce an item quickly that gives a deeper insight into them, and which may be more revealing than the typical post. Investigating this, we saw that this leads to self-disclosure at a level that normalizes the influencer. Because the influencer is seen as more of a “normal person” through self-disclosure at different level, the relationship with the followers changes. While prior to the Story, the influencer might have been somewhat revered by followers, the Story may lead to more of a two-way relationship between influencer and follower. So we wanted to understand if culture had any impact on this. 

We decided to research what happened as a result of the introduction of the Story feature, and if this change was having the effect that social media platforms expected it to across the board, i.e. higher engagement. 

Self-disclosure is not always culturally appropriate

What we discovered was that in some cultures the use of the Story leads to a higher level of user engagement with followers, but in other cultures it has a detrimental impact. This is due to the “Power-Distance” element of culture. Power-Distance is concerned with the relationship between people in authority and those that are subordinate to them. For example, cultures that have a high level of power-distance include China, Japan, India, Malaysia and Russia. Those with low power-distance include Denmark, New Zealand, Austria and Ireland.

In cultures that have a high level of power-distance there is a greater level of respect for hierarchy, and there is an understanding that power is unequally distributed between different elements in society. This is considered legitimate and even desirable in such societies. Evidence from the research suggested that people in such cultures may not appreciate self-disclosure from those in powerful positions, such as influencers, though this was not well researched in the past.

There might be less comments on "Stories" in high power-distance cultures because people want to hold their idols up higher than them.

What we found was that people in low power-distance cultures were more likely to both like and comment on Stories and engage with the influencer. Conversely, in those where there was higher power-distance, while followers did like influencer Stories, they were less likely to comment on them. We believe this is the case because in high power-distance cultures, people want to hold their idols up higher than them and do not want to see them as “normal people” in the same way that other low power-distance cultures will enjoy and accept.

Social media platforms must consider all stakeholders and cultures

From undertaking our research, we highlighted that social media platforms need to carefully consider all stakeholders when rolling out new features. Not all stakeholders in all regions may benefit equally from certain features, due to cultural differences. This could impact adversely on advertising revenues, so is well worth taking into consideration at the outset when devising new strategies.

Critically, the Story is just one novel development from social media platforms. There are now, and there will continue to be, more new features in the future as these platforms evolve. Social media platforms need to take into account cultural differences in different regions when rolling out these features, noting that they may not be of equal value in some areas. They may even be detrimental to the aims of content creators in some regions. This could negatively impact on advertising revenues. Analysis in advance of development and running tests in different regions could identify unforeseen issues. 

Social media platforms need to carefully consider all stakeholders as well as cultural differences in different regions when rolling out these features.

Equally, influencers should not adopt technology for the sake of it. They should assess its usefulness for them and the impact on the specific communities they are seeking to engage with. This will help them to determine if the new feature will bring any value to user engagement. Influencers on all social media platforms should also consider the cultural implications of releasing different types of posts. For example, on TikTok, for some regions influencers should think about the content of their videos and the level of self-disclosure within before posting. 

Influencers should not adopt technology for the sake of it.

Listen to the podcast:

HEC PhD Graduate Revolutionizes Health Practices in D.R.C.

The statistics make for stark reading: with only seven years to go, most African nations face a gargantuan uphill struggle to reach the UN SDG targets. As a result, almost 500 million Africans could be living in extreme poverty by 2030 - at least 70% of whom could still be there in 2050. Which is the main reason behind PhD graduate Anicet Fangwa devoting his academic studies to finding solutions bringing together his competences in engineering and strategy.

Fangwa's work on 1,000 health clinics in the Democratic Republic of Congo could have wide-ranging consequences on health management throughout the continent. In the Knowledge@HEC podcast, Breakthroughs, he explains why. The future Assistant Professor in Strategic Management at McGill University also shares his research on the positive impact women leaders are having in running these same health centers in the DRC.

Doctor Anicet Fangwa, you decided to focus your research and field work on the Democratic Republic of Congo. Why choose this country?

I was born in Cameroon and the DRC for me has always been this huge country full of untapped potential, but which is also characteristic of Africa in general: in the sense that it is full of resources but it is beleaguered by crisis, civil war and instability. The challenge is to help this country reach stability and wealth. For me, studying DRC was a way to study Africa in general, it represents African in miniature, and the idea is to find solutions to its problems, knowing that if it works there, then it should be able to work everywhere else in Africa. 

The healthcare in the DRC has a pyramid structure which is guided by the government and several NGOs and NPOs (non-profit organizations). They fund and control public healthcare and its systems. Your focus is on the non-profit organizational side and how management practices can impact the running of health centers and use these bundles of governance mechanisms. In your analysis, you say that management practices in this context are as relevant as economic logic to improve healthcare. What do you mean? 

I mean that, as economists, we tend to try to develop policy and financial aid that will increase the amount of resources available. But what I'm trying to say is that there's a behavioral component to the aid. This analyzes how the recipient, the beneficiary of this aid, can better grasp the funds that has been put at her or his disposal and generate much more value out of this fund. Each and every year billions and billions of dollars are invested in developing low income countries, but the improvement and the advance in terms of impact is quite limited. I really wanted to investigate why? Why do we still have such a incomplete impact? And so I look at the behavioral component in terms of management.

That’s to say, we study the routines that health centers put in place, the practices that they implement and how those practices can actually enable them to benefit from the fund they receive.

We found that funding is complementary to good governance.

One of the key takeaways that we have is that funding is complementary to good governance. If you only pour funds into the health centers, what we see is that they improve the volume of the activity, they are able to serve many more people and deliver much more drugs. But the quality of their service is not improved at all. However, when you provide those funds in parallel with good management practices (auditing, feedbacks and the commitment to actually change the behavior of the recipient of the fund), we see an improvement both in quality and quantity. We measure this quality of service by studying the infant mortality rate. In our study of 1,000 rural health centers, we find that we are able to reduce the number of deaths by almost 50% over a four-year period! Keep in mind that the DRC is one of the worst countries to be born in, with an infant mortality rate of 70 for every 1,000 births (Unicef, 2020). Compare that to the USA which has 6 deaths per 1,000 births and France which has 4 deaths per 1,000 births...

An important dimension of proper management is the need to have both staff and the pregnant women follow the process formalized by the WHO to arrive at the end of the pregnancy in good health. Infant mortality is often the result of inappropriate behavior and not following the WHO protocol in its entirety. Proper managerial practices can reduce this, there’s no need for additional medical training or changing employees since, with a proper managerial skill set, we are able to help the latter give the best of themselves and deliver high quality services.

Anicet, in the second of the three papers you presented at your PhD, you focused on women in the boardroom and the impact on the organizations they worked in. It was in the same setting in the DRC where gender inequality is a real issue. Your research suggests that there is a very positive correlation between the higher representation of women in leadership roles and issues such as more investment or greater sympathy for the stakeholders. You write that women bring a lot of positive behavioral evolutions, there's more benevolence, they are less traditional and less obsessed with security. Could you elaborate?

Well, like you mentioned, the DRC is a very patriarchal country. But we were lucky to find some health centers where women led the male employees. We found that women leaders do deliver better, higher quality services in those health centers. But they do it using interesting means that are counter to current literature. The latter suggests that whenever women run an organization or a group, they tend to be more collaborative, engage more in discussion and are able to generate much more interaction between the different groups.

But what I find in the DRC is, this being a highly patriarchal society, women leaders tend to actually reduce the interaction with the group, they call fewer meetings and decide alone on topic like finances. In other words, they rely more on top down management instead of collaborative management in this specific area They also tend to avoid internal conflict since they are aware that the surrounding communities will side with the man instead of the woman. One thing I would love to further explore is the impact of communal conflict on women leaders. My guess is that they will be less concerned about these ethnicity questions and treat all patients with equity. But I do not yet have the data to corroborate my hunch.