Firm Governance Advice: Why Math and Machine Learning Disagree

It has been a long time since agency theory started dictating management theory thinking about firm governance. According to this theory, management are primarily selfish and secondarily interested in firm profits and shareholder value creation. This makes them unreliable agents of the ultimate owners, the shareholders, so it is necessary to prescribe many medicines in the form of better surveillance of their actions and better alignment of their pay with firm outcomes. And so a small industry of governance research and advice has grown, resulting in many practices that are supposed to improve things.

Everything is backed by the mathematics of game theory and proofs showing that managers (well, CEOs primarily) will behave better if they are controlled. That makes the advice an undisputable fact except for one thing. The equations are not exactly the same as the practices, and there are many ways actual human beings will either fail to act according to predictions, as when boards do not implement the practices well enough, or will have counter-measures against the predictions, as when CEOs manipulate the board. So in a world of actual humans acting as directors on boards and CEOs of firms, does the advice hold true?

This is what we (Andrew Shipilov, Yeonsin Ahn, Timothy Rowley, and me) examined in research published in Journal of Organization Design. Our approach was simple. We had data on firm adoptions of 11 different governance practices and a series of firm outcomes, and we focused on Return on Assets, Debt, and Dividends distributed to shareholders. These are outcomes that shareholders care about because they concern profitability, risk, and money returned to the owners. We used modern machine-learning techniques to find out which of the practices predicted these outcomes best.

So, what did we find? The best way to summarize our findings is that what works in equations does not work in practice. Hardly any of the 11 practices had any effect on the three outcomes. Two that had effects – and independent audit committee improved profitability and director evaluations increased dividends – suggest that a major mechanism contributing to impact is whether directors have reason to pay close attention to the firm and counter-act selfish CEO actions. That is at least some encouraging news, though overall our findings suggest that much governance advice is hot air.

Shipilov, Andrew V., Yeonsin Ahn, Henrich R. Greve, and Timothy J. Rowley. 2024. The Impact of Governance Practices on Firm Outcomes: A Machine-Learning Exploration. Journal of Organization Design, forthcoming.

Life of The Disenfranchised Entrepreneur: Discrimination, Success, and Stigma

We love the stories of entrepreneurs who advanced from poverty to success and riches. How much do these stories reflect reality, and not wishful thinking? What are these stories missing? Recent research by Leandro S. Pongeluppe published in Administrative Science Quarterly examines a more modest – and so more realistic – story of social advance through entrepreneurship, and it offers a piece of realism and some important lessons.

The story of entrepreneurship being the path to success is quite realistic because poverty is often a result of labor market discrimination, so there is no way out except through entrepreneurship. Labor market discrimination is a powerful exclusion because employers discriminate against those who are visibly different: often minorities or women. In this research, such discrimination was against people living in the Brazil slums (“favelas”), who are easy to distinguish by their language dialect and address.
 

Entrepreneurship by the disenfranchised is not easy, though, and the whole foundation for the research was a set of programs teaching favela residents skills for forming and operating businesses. The skills were useful, because those who received the training were able to start businesses more often than their peers. Importantly, this happened even though those who received the training were no more likely to get a job after the training than those who did not—despite the fact that training someone for running a business also makes them more capable as an employee of a business. Labor market discrimination is a powerful exclusion.

So, with more entrepreneurship and higher income, after training we have a nice story of success, right? That’s where the traditional success story is incomplete. We are forgetting that discrimination against groups happens because they are not supposed to be successful, so when they succeed against the odds that’s wrong too in the eyes of others. They carry the stigma of their disenfranchised background in the favelas, and this stigma is imposed more strongly by others the more successful they are. More income means more prejudice and more stigma from those who are fortunate enough to be born to a middle-class life.

What to do? Obviously, training the disenfranchised for entrepreneurship is still right, and equally obvious it is hard, or impossible, to control the irrational responses of others. Even the old stories of entrepreneurs who advance from poverty are not enough. But we know the reason, of course. In the novels and the movies, those entrepreneurs looked just like the audiences. The stigma will not fade until we tell more stories of favela dwellers and minorities who succeed through entrepreneurship, and we learn to celebrate them too.

Pongeluppe, Leandro S. 2024. The Allegory of the Favela: The Multifaceted Effects of Socioeconomic Mobility. Administrative Science Quarterly, forthcoming.

Hire Those You Trust! But Actually It Is More Complicated

What is the relationship between trust and hiring? We all know the simple answer. Employers hire those who seem trustworthy, so trust and hiring are pretty much the same thing. But there is also a more complicated answer, and that one involves looking at how national cultures differ in the general trust levels. Suppose that two cultures differ in the level of trust – will employers in the high-trust culture hire more people than those in the low-trust culture? No, of course not, employers hire as many people as they need. But social trust levels still matter.

How they matter is the topic of research by Letian Zhang and Shinan Wang published in Administrative Science Quarterly. It involves a novel idea and some nifty analysis, and fortunately it is easy to summarize. Trust does not mean hiring more people, but it does mean hiring different people. The reason is that low social trust is associated with hiring for a specific job, with less expectation that the employee can develop new skills. High social trust means hiring for the firm, with an expectation that the employee can develop new skills and fill other jobs. High trust, then, means hiring for foundational skills rather than advanced skills. It means hiring an analyst for general math ability more than for skills in Laplace transformations.

This idea raises two questions. First, is it true? Using data on job postings from the European Union countries, Zhang and Wang found that it was indeed true. Employers in nations with high social trust hire based on more foundational skills than nations with low social trust. Moreover, the same multinational firm would hire more based on foundational skills in nations with high social trust than in nations with low social trust, so the same relation holds within employers as well. Job characteristics such as university education or work experience requirements reduced this effect but did not make it go away.

Second question, is it consequential? Well, look at the figure above. Nations in Europe differ quite a bit in social trust levels, as the horizontal scale shows (the range is from zero to one). The vertical scale is not so easy to understand, but perhaps it helps to know that a difference of 0.6 is less than the difference between attentiveness and mathematics (foundational skills), and electricity principles and Java (advanced skills). The figure shows that the average hire in each nation differs significantly by the trust level.

There are many possible consequences of these differences. We don’t yet know whether they all happen, but it is valuable to check each one. Hiring in high-trust nations means hiring for the long term and for multiple roles, giving greater room for personal growth and firm flexibility. Hiring in high-trust nations means less emphasis on specific expertise and credentials, so symbolic collection of certificates to get hired is unnecessary. Hiring in high-trust nations allows more diversity in teams doing a single task and better communication within teams, increasing creativity and productivity. Employers in low-trust nations may have lower access to all these benefits.

We do not know whether all these differences result from different levels of societal trust. Now that we know how societal trust changes hiring practices, we should be aware that they might exist, and both employers and employees might think of employment practices and careers differently.

Zhang, Letian and Shinan Wang. 2024. Trusting Talent: Cross-Country Differences inHiring. Administrative Science Quarterly, forthcoming.

Networks and Discrimination: Why Female Artists are Disadvantaged, and What They Can Do About It

It is not easy being an artist. Recognition of talent and creativity can be slow, sales only happen in small galleries, and initial sales are domestic and even local. The last thing artists need is discrimination in addition, but that is exactly what female artists get: a recent investigation showed that comparable paintings sell at a 42 percent discount if the artist is female.

Is there anything that can be done about such discrimination? This was the question that we (JungYun Han, Henrich R. Greve, and Andrew Shipilov) wanted to address with data on Korean artists and their exhibitions abroad. We found that female artists were less successful in exhibiting abroad, as expected, but that difference was not our main interest. Instead, we wanted to know whether we could identify anything in their careers that reduced or eliminated their disadvantage. We could.

An important step in the careers of many artists is a residency stay in which they share workspace in studios provided by the residency and also get to meet other junior and senior artists to gain inspiration and advice. Residency programs help artists succeed, which is exactly their purpose, but unexpectedly this was only true for female artists. Education in an elite art school provides top-notch technical training and artistic appreciation. Elite education helps artists succeed, which is exactly its purpose, but again there was a surprise: it benefited female artists more.

What is going on here? The best explanation for these two effects is not training, but social networks. Art residency programs and elite schools connect artists with others who can provide advice on how to approach galleries and even direct contacts to them. The best explanation for the male and female difference is that female artists have more to prove, so the benefit from a network tie is greater for them. In network effects we often see such effects – those who are accepted purely by who they are gain some benefit from a good social network, but not nearly as much as those who are discriminated against and need a social network to be introduced to the right people and become recognized for their achievements.

These effects offer clear advice for how to help women succeed in art, and probably also in other kinds of entrepreneurship and work. They also offer a warning to society because such differences can only exist because of discrimination.

Han J, Greve HR, Shipilov A (2024) The liability of gender? Constraints and enablers of foreign market entry for female artists. Journal of International Business Studies.

If Women Can’t Network and Women Can’t Move, Why is it Better for Women to do Both?

Among the many disadvantages that women have at work, here is one that is often overlooked: they have fewer opportunities to form beneficial networks, and even if they succeed, they gain less benefit than men. This matters greatly for their careers because network ties to coworkers help employees gain skills,learn about opportunities, and execute plans. A particular disadvantage is women’s problems in getting brokerage positions in network. A network broker is connected to people who are not directly connected to each other. Brokers gain separate pieces of information quickly and can quickly assemble them to form opportunities.

Why is it hard for women to become brokers? To begin with, it is hard for anyone because it requires reaching beyond the immediate work group. It is also hard because people are suspicious of brokers and may be reluctant to share information with them. In fact, the most effective brokers are those who are not known to be brokers. For women, these suspicions are especially strong because of the gendered belief that women maintain closer relations with proximate friends and coworkers. As a result, they gain less access to brokerage and less benefit from brokerage.

Changing jobs has many of the same disadvantages, even if the job change is just a reassignment ordered by the employer. But here is the interesting part: when women move, the brokerage disadvantages disappear. Both disadvantages. Women who move gain brokerage positions just as easily as men who move, and women who move obtain the same performance benefits as men who move. This is a new discovery from a paper by Evelyn Zhang, Brandy Aven, and Adam Kleinbaum published in Administrative Science Quarterly. Their idea, which turned out to be true, is that moving gives “license to broker” because network ties in the new workplace are necessary, and maintaining contacts with the prior workplace is expected – especially for women, who are supposed to be more stable network partners than men (again a gendered belief). So in this case two wrongs make a right.

Interesting? Let us not see this as encouraging information though. Even when workers benefit from gendered beliefs like this, the beliefs still create a warped workplace where opportunities and rewards are unfairly distributed.

 

Zhang,Evelyn Y., Brandy L. Aven, and Adam M. Kleinbaum. License to Broker: How Mobility Eliminates Gender Gaps in Network Advantage. Administrative Science Quarterly, forthcoming.

Failing Once, Failing Twice: What Makes Firms Search for Radical Improvements?

Cynics would say that firms don’t look for opportunities as much as they should. Instead, it is problems that generate search for improvements. The cynics would be right – what we call problemistic search, triggered by disappointing profits, is a real thing and it is more frequent than search for opportunities. That is bad enough, but actually things are worse.

Research by Thomas Keil, Evangelos Syrigos, Konstantinos Kostopoulos, Felix Meissner, and PinoAudia published in Journal of Management shows that multiple goals complicate things even further. This is because problemistic search can be replaced by self-enhancement. Executives and organizations engaged in self-enhancement do not solve problems, but instead they look for reasons to claim that there is no problem to solve. Chief among these reasons, I mean excuses, is finding a secondary goal that shows higher performance.

Does this happen? There is ample experimental evidence that individuals self-enhance when given the opportunity. This research is novel in showing that organizations can self-enhance in response to very important goals, and self-enhancement has important consequences.

Pharmaceutical companies rely on drug approvals for their profits, so having drugs pass the late stages of the approval process is a primary goal. They also need a good research pipeline, so drugs moving through early-stage approval is a secondary goal. How to get many drugs and novel drugs? A key decision is whether to search in the proximity of their current expertise, or whether to move into new disease areas and acquiring candidate drugs from other firms. Proximate search is safe but is a questionable strategy for a firm with low performance. Distant search is riskier but is the way to renew a firm with low performance.

What the firms should do in response to low performance is trivially simple. If the internal research is good, stay with it and do a proximate search. Otherwise do distant search. The good news from the research by Keil and coauthors is that the pharma firms behave exactly like that. They turn to distant search when the performance from the current search is low.

There is also bad news. They do this only when seeing disappointing performance on both the primary goal and secondary goal. Disappointing performance on the primary goal – the most essential one – is not enough to trigger distant search. Even worse, doing poorly on the primary goal and well on the secondary goal produces less distant search than doing well on both goals. For the sake of firm profitability, and for society getting necessary medicines, this is very problematic.

Self-enhancement is something we can understand and accept when we see it in individuals. It is a slightly childish thing to do, but people want to preserve their self esteem and want to look good in their own view, and that of others. Better than they deserve, even.

It is much harder to understand and accept self-enhancement by firms. Firms exist for practical reasons. They produce products and services, they develop improvements in products and services, and being good in actuality is much more important than being adept at self-enhancement. Unfortunately, this research is a reminder that there is self-enhancement in firms too. No doubt this is because the managers and executives of firms are people too, and the firms are lacking processes that control their individual self-enhancement.

Keil, T., E. Syrigos, K.C. Kostopoulos, F. D. Meissner, P. G. Audia. 2023. (In)Consistent Performance Feedback and the Locus of Search. Journal of Management forthcoming.

Talk is Expensive: When a Competitor Has Financial Ties to Media

We understand that media ownership can be translated into power, especially when a media outlet has a dominant owner and the context is politics. Rupert Murdoch, Fox, and Donald J. Trump are keywords that come to mind. That’s just a rich guy playing around with the governance of a nation, with no connection to the world of business competition, right? Wrong. Media ownership also affects competition among firms, and the effects are seen also when the ownership structure is more dispersed.

This is the main discovery made in a paper published in Administrative Science Quarterly byMark R. DesJardine, Wei Shi, and, Xin Cheng. Their starting point is the remarkable concentration in firm ownership that has happened following the growth in institutional investments in the form of fund management firms. These investors want to (in fact, are obliged to) maximize the returns of their holdings, so they will do whatever it takes to increase the value of the firms they own.

What does “whatever it takes” mean? This is where media ownership comes into play. An interesting feature of owning media firms is that media firms are involved in news gathering and reporting, which can influence the competitive balance of an industry. Hurt one firm, and the other gains. Report selectively, and the value of firms owned by the fund that also owns media outlets will increase. As a result, media talk is expensive for the competitors of firms that have a media connection in their ownership.

Such media effects are a very big deal because they show an illicit use of media ownership that tilts valuations of firms, and corresponding access to resources and success in markets, away from the products and services they provide. They can only happen as a result of unethical actions by media executives and editors.

The research they present has plenty of evidence. Media coverage turns negative when a competitor firm has financial links with the media. This effect is stronger for competitors with more similar product lines, so relevance increases negativity. The effect is stronger for competitors nearby, so proximity increases negativity. And, most perniciously, if the media company CEO has equity-based compensation, so the CEO gets paid more when the media company value increases, the effect is also stronger. In sum, negative media coverage is a result of financial links, and it is particularly negative when the competitive relations between firms are close and when the media company CEO is for sale.

Should we worry about this? People arguing that “talk is cheap” would not be too concerned about these findings. But media coverage has significant consequences for firms, especially for their access to financial resources, so seeing it can be tilted so easily means that there is one more area of competition that requires regulatory attention. We cannot have an economy and society in which consequential, expensive talk is for sale.

DesJardine, Mark R. , Wei Shi, and, Xin Cheng. 2023. The New Invisible Hand: How Common Owners Use the Media as a Strategic Tool. Administrative Science Quarterly, forthcoming.