In this post, I want to try to pull together a couple of puzzles I’ve been thinking about, and suggest that these puzzles have a sociological explanation which often gets left out of economic theorizing—and that this explanation, in turn, has some startling implications for a theory of the firm, particularly the question of why independent startups seem to be more successful at radical innovation and growth than are more established firms.
The puzzles I am interested in are: (1) The Peter Principal: Why employees tend to ‘rise to their level of incompetence’ within organizations; (2) The relative infrequency in real-life organizations of ‘compensating differentials,’ and other mutually agreeable trades between employees and employers, which seem so logical and figure so strongly in economic theory; and (3) What is meaningfully different about ‘startups’ that allows them to do activities that cannot be done inside of more established firms—or, in other words, why startups per se should exist, rather than start-up like activities taking place inside the ownership structure of, and with all the supports of, established firms.
My proposal for answering these questions is that, contrary to workhorse models of economic theory, employee compensation and benefits do not function solely as incentives and rewards for our contributions. Rather, most humans tend to see compensation and benefits as recognition of status and social standing within the organization. And humans are highly social animals, who are very anxiously attentive to these symbols of our social standing, particularly as they compare to those of our relevant reference group, such as other members of our organization. Altogether, this means that humans tend to demand what I will call ‘status concordance’ in compensation, authority, benefits, perks, and rights within a firm—we want all of them to increase in lockstep as we move up in the firm in tenure and responsibilities—and we become very uncomfortable with status dissonance, when one person has more authority but less pay, etc., (unless that dissonance is between people who are members of different perceived peer groups, due to having a different educational background, skillset, or career track).
Thus, a human social constraint stops organizations from making certain kind contracts and agreements with employees which, from an economic perspective, would be extremely efficient. A particularly interesting application of my theory is the question of why so many extreme innovation and growth activities take place in startups rather than inside more established firms. Startup innovation activities are often driven by very aggressive, ambitious, and creative young people, who must, in order to innovate, be given the leeway and authority to make risky decisions, who often work extremely long hours, and who have very high-powered equity incentives that could make them extremely, extremely wealthy if they succeed. If such a startup were owned in-house as a division of a larger firm, it would generate status discordance and angst among the firm’s employees. In one direction, the firm would not be able expect one division’s employees to work longer hours than other employees in the firm; on the other it would not be able to give aggressive upstart young people high-powered incentives, scope to make risky decisions, and the potential to get very rich;… without inviting status discordance and outrage within the organization.
The Peter Principal
My thinking about this question was prompted by the job-market paper of a doctoral colleague and friend of mine. Her paper focuses on an employee-evaluation system that a company implemented in order to mitigate the well-known ‘Peter principal.’ (Since my friend has not chosen to publicly post a working paper online, I won’t share many more details about the author or paper.)
The economic explanation for the ‘Peter principal’ is simple. Organizations use employee promotion for two distinct purposes: (#1) To match people to jobs to which they and their skills are well-suited, and (#2) to incentivize and reward employees for good performance in their current job. If #1 factors in substantially, this means that if I perform well in Job Level 1, I get promoted to Job Level 2; and if I perform well in Job Level 2, I get promoted to Job Level 3, and so on. But once I reach Job Level N, and do not perform well in that position, I do not get promoted, and I stay in that position until I retire. So, people get promoted to the level of their incompetence, and stay there, and as an organization ages, it tends toward a state in which everybody is in the position in which they are incompetent.
From an economic perspective, the Peter Principal should be pretty simple to solve. We’ve been aware of this problem since 1969, so organizations should just recognize that it exists and stop using promotions for the purpose of #2, for incentives and rewards. Instead, they should just reward employees in the most efficient way, by giving them more money, while keeping them in the roles to which their capabilities and talents are most well-matched. Organizations should communicate to their employees that they should not think about the level of pay of a particular job, but instead think of the level of pay as a function of performance, and think of job ‘level’ as a simple matter of matching and sorting of skills and roles.
But, of course, organizations have not solved this problem yet, and they continue to suffer from the Peter Principle. Think, for example, how natural it is—how much we take it for granted—that we talk about the pay level of a particular job, assuming that the variation of pay within each job level should be trivially small, and that the variation in pay between job types is the most relevant determinant of pay. We have all largely internalized the idea that pay is a function of position, not a function of performance. The Peter Principle has three large costs: (1) organizations get people making decisions in roles at which they are incompetent; (2) organizations lose the gains they once had from employees who were highly competent in their positions; and (3) organizations end up committed to paying out large yearly dividends to people who are not doing well in their roles (assuming a reluctance to fire employees).
Why is the Peter Principle so hard to overcome, when it is so very costly and when in theory it should be solvable? One obvious factor is that promotion is not just a blunt tool to give employees higher pay; rather, employees value promotions for status and recognition. So, you might think that the explanation is that the sorting-and-matching role of promotions can’t be completely, cleanly separated from incentives and rewards—in short, because people value status and recognition, you can’t motivate people without promotions.
Now, I think that this is definitely part of the explanation for why the Peter Principle still plays a role—for why firms don’t just match employees to jobs on skills and reward them with payment—but it’s not totally satisfactory, just yet. After all, economic theory models people with rational preferences as willing to make trade-offs among the goods that they desire. So, in this vein, perhaps employees long for more status and recognition, but they have a price: for the right salary or bonus, they’ll be just as happy to stay in their current position, where they are most competent. Given the costs of the Peter Principle, my intuition is that this price would be lower than the price of the alternative. Moreover, it would seem logical that organizations would evolve other ways to give employees status and recognition, less stark and blunt than permanent hierarchical promotion.
So this is, I think, a very interesting puzzle and a challenge for conventional economic approaches to modeling employee incentives—and this puzzle is only partially, not completely, resolved by the claim that people value status and recognition within their organizations.
Compensating differentials and other nice trades:
As discussed, economic theory assumes that rational people are willing to trade off among the various goods that they desire. In labor economics, these tradeoffs are often called ‘compensating differentials.’ For example, it is often supposed that jobs that are dangerous or unpleasant (say, being a garbage man) will have to pay higher wages than other jobs that require similar skillsets but are less dangerous or unpleasant, in order to attract workers. As a matter of economic theory, the idea that such compensating differentials should exist follows axiomatically from the simple assumptions that people have preferences (things they like, and things they dislike) and labor markets are competitive. It would be simple to model an equilibrium in which workers had roughly equal skillsets and sorted into jobs with differing levels of pay vs. other benefits and detriments, based on their diverse preferences for those goods.
Now, it turns out that, empirically, there is some evidence for compensating differentials in some settings—but compensating differentials are surprisingly rare and small, in my reading of the literature. This is, I hasten to emphasize, a surprising fact, given that the existence of compensating differentials should follow axiomatically from three simple assumptions (a.) people are willing to trade off the goods that they desire, (b.) people have diverse preferences, and (c.) labor markets are competitive. From this perspective, compensating differentials should not be occasional curiosities that we can just detect with very clever statistical methods—they should be everywhere. Not just between jobs, but within jobs as well. But they don’t seem to be. There are no law firms that I am aware of that offer two possible career tracks: one with lots of document review, but enormously high pay, and one with less document-review, but lower pay. All corporate law firms seem to impose both a decent portion of document review, and high pay packages, on all of their prospective associates. There is one entry level. There is one track. There is one job description. There is one pay package. (For people entering the same role with the same educational credentials.)
The near non-existence of compensating differentials is actually just a subset of a bigger phenomenon, which is the general reluctance of firms to negotiate various kinds of mutually-agreeable trades with their employees. Consider the inefficient, but ubiquitous, practice of giving business travelers unconstrained expensing, or travel budgets, or doing in-house travel procurement: Employees get to spend as much as they want on first-class flights, hotels, and meals on the road, up to a usually quite high ceiling, but do not get to take home any savings. How can firms which seem to be obsessed with cost-cutting in many areas continue to do things this way? Why not just pay the employees more, and then let them choose how much of that money they want to spend on eating out vs. saving? From the perspective of economic theory, giving people money to make their own choices among these tradeoffs is clearly superior. In practice, there are several easy or clever ways the firm could arrange for employees to make their own choices around the costs and benefits of hotels and flights, while allowing them to take home savings from cheaper options.
One common justification for allowing unconstrained or budgeted business expensing or is that business expenses are tax-favored. But I’m skeptical that this justification suffices: Can the percentage difference in marginal costs arising from taxes account for the order of magnitude difference in the cost of first-class vs. coach-class flights? Would all the business travelers in first-class really prefer flying first class and taking home no extra cash to the alternative of flying coach and taking home an extra 60% of the $1000+ cost savings? It seems hard to believe. It seems equally hard to believe, as well, in the case of hotels—it is hard, I think, to suck $300+ of utility of from a room in which we spend most of our time unconscious.
A unifying theory: status concordance
I suggest that all of these puzzles can be explained by a single theory of compensation, which runs contrary to the canonical economic approach. (The canonical economic approach sees compensation as functioning to provide incentives ex ante by giving rewards ex post.) I propose that, instead, people interpret compensation—in all of its forms—as a recognition of their standing within an organization.
Compensation as status symbol has many implications, given that human beings are hierarchical, social animals who are extremely attentive to these symbols. (Usually unconsciously: We consider it unseemly to be explicitly attentive to status, but we also consider it unseemly to do millions of other things that have been implicitly arbitrarily deemed low status, so I will not believe most people who claim they are an exception to this rule.)
The first implications is that, rather than observing a world in which employee compensation packages trade off among many different goods (less pay in return for more freedom, etc.), we would expect to see people and organizations demanding what I call status concordance. When compensation is not an incentive, but a status symbol, people will not accept reasonable tradeoffs between the different dimensions of compensation, but will demand that the various dimensions of compensation move up in lock step—because now deficiency on a single dimension of compensation, relative to some peer reference group, is interpreted as a slight, and such slights are extremely, extremely pricey.
So, in a world of status concordance, the people with higher pay get more freedom get better health insurance (even though their higher wealth should already reduce their demand for insurance) get more prestigious titles, get better offices, get more authority, and so on. And the people in the same status levels are offered the same pay packages, same authority, same freedoms, same benefits, etc. and variations and tradeoffs among those goods are not discussed, considered, or negotiated by the many diverse people all being matched to one particular job description.
Now, of course, the fact that a senior corporate lawyer gets both more pay and a better office than an college intern is largely explained by differences in bargaining power, as is classically modeled in economics. The senior corporate lawyer has better outside options, and she can use this to negotiate better pay and a better office. So differences in bargaining power can surely explain why the correlations between the various components of compensation are greater than zero. But I do not think it can explain why they are so very close to one. Given that there are many people who are close together in terms of their outside options/bargaining position, and quite diverse in their preferences, you would still expect to see more examples of discordance, in the classical economic world of stable individual convex preferences and Nash bargaining.
So, I propose that status concordance is important to understanding firms’ compensation practices, and that my status concordance theory can explain many of the above mysteries and inefficiencies, and many more as well.
Given status concordance, the Peter Principle becomes highly intractable. If you have an extremely high performer in Job Level 1, that person will reasonably be able to demand recognition and status for that performance. Can you just offer that person gobs more money to stay in Job Level 1, where he is well-matched, rather than Job Level N, where he will be at his level of incompetence? In economic theory you can, but in practice, we humans just can’t deal with this: This employee would hardly be able to tolerate seeing his peers who performed less well than him but were better matched to higher roles promoted to authority over him. Paying him gobs of money while keeping him in Job Level 1 would make everyone in the organization uncomfortable: his superiors would feel less comfortable giving him instructions; his superiors would feel outraged at their own pay. Common knowledge of relative rank and status would be disrupted, which would make it more difficult to coordinate commands throughout the organization. Even if the organization and its consultants were to beg the employees to think of promotions as simple matching, people wouldn’t accept it—we are incorrigible social animals in this regard.
The status concordance theory of employee compensation can explain why professionals seem very committed to having ‘paid vacation’ and other forms ‘paid leave.’ ‘Paid leave’ is of course a fiction, as a job that pays $100,000 with two weeks of paid leave per year is identical to a job that pays $104,000 from which one takes two weeks of unpaid vacation per year. (Indeed, according to decision theory, the paid-leave job above is strictly inferior, since the unpaid leave job contract additionally gives the employee more options, beyond the option to get the exact same package as the paid-leave contract.) But to the extent that people see their compensation not as a reward/incentive for work, but as a recognition of their standing, then taking unpaid leave strikes them as an insult to their standing in the organization—a loss of rights and recognition—rather than as, say, a choice that any human being with a life outside of work should enjoy making from time to time. “To think, that they would reduce my salary, just for taking a vacation, which is a reasonable thing for me to do, my right after all I’ve given this organization!” So, we tend to prefer having our leave rights described as ‘paid leave’ and having the costs essentially amortized in our salaries over the remaining weeks—having money and choices taken away, a thing which conventional economic theory sees as self-contradiction—rather than face economic tradeoffs that could allow us to make choices that might cause status-concordance anxiety.
The status concordance theory can explain why we don’t see interesting job packages that reflect tradeoffs that meet the diverse preferences of different people, or, in other words, compensating differentials within job tracks. In my experience, most corporate lawyers in the world will report that they work too many hours and earn more money than they need. But corporate law firms mostly don’t offer career tracks with both lower pay and shorter hours, because doing so would introduce status discordance. If associates exit law school as equals, then, upon being split in career tracks, one track would resent the superior pay or the superior hours of the other.
The status concordance theory can explain business-class ticket prices. Most business travelers would not pay first-class ticket prices out of their own pockets, and the tax advantage of business expenses can’t account for a difference so large. So there should be a profitable trade to be made between business travelers and their firms. The status concordance theory provides an explanation: If businesses used a rational comping scheme for business travel, most of the business travelers would themselves choose to fly coach, but then, at that point, they would feel status discordance, feeling that their seat in coach did not properly recognize and reflect their status within their organization. So, we once again seem to prefer to have money, choices, and freedom taken away from us to avoid having to face status-discordance. (By the way, I thank business- and first-class travelers for their charity in subsidizing flights for everybody else, allowing this doctoral student to travel the world for amazingly cheap prices in coach. You and your firms are martyrs.)
A theory of startups:
For a long time, I’ve been looking for a theory of startups. We seem to talk about startups constantly these days—they are at times the object of uncritical worship and at other times the object of cheap ridicule. But whatever we think of the startups that are in the news, a question that has always bothered me is, why is it even meaningful to talk about startups per se? Whatever activities we associate with startups—good or bad—why can’t those same business activities—the actual things that actually happen, rather than the our verbal description of the firm doing them—happen in bigger, established firms? Why does it make sense for people to say ‘I want to work for a startup’ instead of saying ‘I want to do a job that involves doing X’? Strictly speaking, a startup is just a legal form at a point in time—a ‘startup’ is just a newish standalone business entity. There is no necessary tie between being newish and standalone and doing any particular activities, or not doing them. So, why is it meaningful to talk about the activities, prospects, or characteristics of startups per se?
I propose that status concordance in employee compensation can provide a theoretically-grounded explanation for why startups per se might be meaningfully different from other firms, and capable of different activities, in particular the extreme growth and innovation for which they are usually lauded and promoted. I make three assumptions: (1) status concordance matters, (2) people look to the members of their legally-defined organizations as their peer groups, and (3) organizations are reluctant to fire employees. With these three assumptions, it follows that startups can give high-powered incentives and scope of authority to young employees, that they would not be able to give them while part of a larger, older organization. And these high-powered incentives and scope to experiment and make mistakes are required for radical growth and innovation. So it is economically efficient for startups to be a thing—for radical growth and innovation activities to be housed inside newish standalone firms.
An important fact about startups, and the radical innovations and growth which they ideally aspire to achieve, is that they are usually largely made up of young people. Why is this? I see several candidate explanations. One factor might be that people seem to be at their most ambitious, aggressive, and creative in their youths—people mostly achieve most extraordinary genius in mathematics, music, and invention in their twenties. I think there are obvious evo-psych explanations for this, but I won’t dwell on them in this post. Another factor might be that young people are better at certain kinds of innovation because their minds are more adapted to the current state of technology. Another factor might just be that young people have more free time and energy, before they have kids. Whatever the reasons, the takeaway is simple: Startup activities are driven largely by young people, who, by virtue of their youth, would be ranked lower in the status hierarchy of a more established firm.
Another important fact about startups is that they are doing risky business: developing new products is an activity whose modal profits are negative, but whose top-quantile returns are extremely high. A final important fact is that startups are in fast-moving businesses: time is of greater essence when racing for share in a network economy, than it is when debating whether or not to change the packaging of an established CPG brand, for example. These two facts have some important implications: First, startups need to have control systems in which the employees have scope to quickly make decisions which could go wrong, which have potential downsides. Second, startups need to give their employees equity incentives to motivate them and make them willing to take risks, against the long-odds they are facing.
When you combine these facts with the status concordance theory, it’s seems that older established firms could run into trouble when trying to develop startup-like activities in-house. Basically, archetypal startups are doing activities where you need to give young twenty-somethings a lot of scope to make some potentially bad decisions, while offering them equity incentives that could potentially make them extremely wealthy. In an older, more established organization, the scope of control, mistakes (both potential and realized), and potential riches of cocky newcomer employees would be offensive and status-discordant to many of the more senior, established employees. Established employees would find it nearly impossible to have startup-like compensation packages and scope of control in one (younger) part of the organization, and established-firm-like compensation packages in the remainder. So this constraint on compensation packages within an organization stops more established firms from optimally incentivizing radical growth and innovation activities.This provides one reason why it might make sense that radical growth and innovation activities must often be housed in new, upstart firms.