A few months ago I visited Houston to attend the 2013 meeting of the Society for Industrial and Organizational Psychology (SIOP). During that conference I presented an empirical study on ethical leadership. My co-authors and I found that ethical leadership is negatively related to deviance among work teams. Conversely, it is positively related to voluntary learning activities. Speaking very loosely, the evidence is consistent with the notion that ethical leadership increases effective work behavior, while decreasing problematic deviance. This is a classic “win-win” solution to performance problems. Supervisors benefit from superior performance (very broadly defined), while subordinates are treated justly and decently. Everyone is better off, other things being equal.

If you like that storyline, then I can improve on it. During my presentation, I observed that these findings are in line with other research on ethical leadership. They are also consistent, I would have added if time had not expired, with related work on servant leadership, authentic leadership, and transformational leadership. I know this is so because I have given similar talks on related work at other conferences. The package is so neat, clean, and superficially happy that I could (almost) write a popular business book about it.

As I sit here now, typing while my SIOP talk parades through my memory, I am wondering what it was that I missed. My uneasiness is captured in the famous 1892 short story that gave rise to a journalistic cliché. In Silver Blaze, Sir Arthur Conan Doyle’s Sherlock Holmes investigates the disappearance of a race horse and the possible murder of its trainer. Holmes has this famous exchange with Inspector Gregory of Scotland Yard:

Inspector Gregory: “Is there any point to which you would wish to draw my attention?”

Holmes: “To the curious incident of the dog in the night-time.”

Inspector Gregory: “The dog did nothing in the night-time.”

Holmes: “That was the curious incident.”

Later in the story, Holmes elaborates his reasoning: “a dog was kept in the stables, and yet, though some one had been in and had fetched out a horse, he had not barked enough to arouse the two lads in the loft. Obviously the midnight visitor was someone whom the dog knew well.”

The conversation over the “dog that did not bark” remains a masterpiece of detective fiction, and Doyle’s story juts into the memory of my ethical leadership talk. More to the point, in all of my presentations on this and related topics, there was an obvious question that no one (including myself) had thought to ask. Like the dog that didn’t bark, we automatically closed off the topic because we were too close to it.

Somewhere on the flight between Houston and Denver, it came to me. No one had raised the following issue: “If ethical leadership brings all the performance benefits that you say it does, then why does unethical leadership exist at all? That is, if ethical leadership boosts effective work behavior, as well as being socially desirable, then why would any leader choose to be unethical?” Even the most ardent sociopath, perhaps especially the most ardent sociopath, should behave ethically if there is a personal payoff, right?

Well, that all depends on what we mean by “payoff.” I wonder sometimes if we management scientists are just a tad naïve about the motives of business leaders. Tacitly, the aforementioned research assumes that mangers’ primary goal is to maximize their subordinates’ job performance (broadly defined, so as to also include more citizenship behaviors and fewer counterproductive ones). If we begin with this assumption, then the fact that ethical leadership seems to engender more effective subordinate work behavior is a decisive finding. People in leadership roles should attempt to be as ethical as possible, as they will reap these performance benefits in an agreeable way.

This logic is sound, but there is an issue with the starting assumption. Managers are not always as earnest as our research sometimes seems to imply. For many, their primary goal is to advance their careers, which means moving up the corporate ladder. To translate this into social scientific jargon, they are motivated to gain and hold social power. Higher subordinate performance might be one way to accomplish this objective, but this is not always so. Managers are likely to behave ethically if doing so serves their primary goal (power acquisition and retention), but they are equally likely to behave unethically if so doing would advance their careers. In sum, managers are not necessarily motivated by the things we behavioral scientists (or their subordinates!) wish they were.

If we want to understand why leaders behave unethically, we need to understand what they are trying to accomplish when they go to work. Their behavior should “make sense” within the context of their own goals, but not necessarily for the objectives that scholars might attribute to them. Businesspeople could have motives that, at least at first, appear unusual to the professors who are trying to understand their actions. In particular, businesspeople want status and power, as it serves their career goals. In some settings, the way to gain and hold power is through actions that most people would view as unethical.

It is on this point, more than any other, that many management theories of business leadership fall short. Managers are not (only) economic actors who attempt to maximize performance and profit. They are also political animals who attempt to gain standing within valued social groups. This quest for social advantage often leads managers to engage in conduct of dubious morality. In this regard, if in no other, I have found political science theories to contain a good deal of the hard-nosed realism that is often lacking in the organizational sciences. By beginning with different assumptions, political scientists are able to work out very different theories of leader behavior, and these anticipate that unethical action is often the result of leader’s career goals.

The Dictator’s Handbook: Why Bad Behavior is Almost Always Good The Dictator's HandbookPolitics, by Bruce Bueno de Mesquita and Alastair Smith, is a good example of what I mean. The Dictator’s Handbook presents an innovative leadership model, which the authors name selectorate theory. As Professors de Mesquita and Smith are political scientists, their theory is largely based on government. I will mostly stick with their emphasis, but I’ll also argue that selectorate theory is applicable to business settings as well.

Selectorate theory begins with the assumption that all leaders have roughly the same fundamental goal — to gain and maintain power. From this follows a negative corollary – leaders are not seeking to do good for their constituents, except to the extent that doing so allows leaders to maintain or enhance their power position. To return to the earlier example that promoted this discussion, business managers are not concerned with raising work unit performance, unless raising performance brings some benefits to them.

In The Dictator’s Handbook, Drs. de Mesquita and Smith emphasize one key factor that determines whether or not the leader is public spirited, selfish, or narrowly political. Indeed, managers are predicted to be downright unethical if this will advance their career goals.  Surprisingly, this key driver in the model is the manner in which the leader is selected for the power position. Essentially, the more people involved in the selection process, the better the leader is likely to behave once he or she has the job. Working through why this so is the subject of selectorate theory.

This simple causal mechanism may strike the reader as counterintuitive, at least it did for me, so let’s take a closer look at Drs. de Mesquita and Smith’s reasoning. We can begin descriptively. First, it bears mentioning that no one has the wherewithal to run a large polity entirely alone. Any social system as complex as a modern business organization, a government bureaucracy, a municipality, or a national government will require some governing coalition. That is, the leader will require the help of some body of people in order to gain and retain power.

A social unit, such as a nation or a corporation, has a roughly identifiable set of members (e.g., citizens and employees, respectively).  Selectorate theory divides this total set into at least three parts – the disenfranchised, the selectorate, and the winning coalition. Let’s take a look at each.

Many members of a social unit are disenfranchised, according to the theory, because they are not allowed to select leaders. For instance, in the United States people under the age of 18 are not allowed to vote. Likewise, in many organizations most employees have no say in hiring decisions or in managerial performance evaluations.

Conversely, the “selectorate” (S) is the subset of those members who can play an active role in choosing the leader. The selectorate may be further subdivided into two sets. The “nominal” selectorate is everyone with a supposed right to be involved. These people are sometimes called the “interchangeables.” The “real” or “actual” selectorate are those who participate in a meaningful way. These individuals are often called the “influentials.” For instance, in many democracies all adults are legally allowed to vote. This is the nominal selectorate (or interchangeables). However, many people choose not to exercise their franchise. Only those who vote are part of the actual selectorate (the influentials). For our purposes here, though, we will deemphasize this distinction, though I’ll refer back to it briefly.

The third group, the “winning coalition” (W), is those individuals whose support is necessary for the leader to grasp and hold power. For this reason, they are also called the “essentials.” The winning coalition is a subset of the selectorate (or, to be specific, a subset of the actual selectorate).

By now you may be seeing the problem. Things may be less democratic than they seem. For instance, a publically traded corporation is governed by a board of directors. These directors are elected by a vote of the shareholders. However, voting is not that simple. Some people might hold non-voting stock (disenfranchised), while other stockholders hold small amounts and so don’t bother participating (they are not influentials). In practice, the real selectorate tends to be the major shareholders, often institutional investors or descendants of the founders. And only a subset of these, perhaps 51%, will select the board members. Thus, the CEO is not beholden to the employees (disenfranchised), or the minor shareholders (not influentials), or even to a substantial percentage of the major shareholders (not essentials). Rather, the CEO is most concerned with the shareholders who voted or are likely to vote in his or her favor.  These are the people that the CEO needs to keep happy and doing so might not mean acting in the best interest of other groups. Often, the non-essentials will be mistreated if doing so preserves the winning coalition.

An example might make this clear. During the 1980s and 90s, the ethicality of workforce reductions were widely questioned because they had harsh consequences for workers, families, and communities. According to selectorate theory, it made perfect sense for CEOs to downsize their workforces, despite any moral qualms. This was because workers, families, and communities don’t get to pick the CEO and so they are of less concern to the leader. The major shareholders are part of the selectorate, and the major shareholders were the ones who reaped the biggest benefits when downsizing pushed up the stock price. When Milton Friedman tells us that the purpose of a business organization is to maximize shareholder value, he is putting a happy philosophical face on a cold political reality.

Adding precision to these observations, selectorate theory offers two numbers that help us quantify how a leader might behave. These are the size of the winning coalition (W) and the ratio of the winning coalition to the selectorate (W/S). We’ll take a look at each.

W, the number of essentials, is the most straightforward. To simplify the model quite a bit, a person in a leadership role can keep others happy by providing two types of goods – public goods, whose benefits are shared broadly, and private goods, whose benefits only accrue to members of the winning coalition. If W is small, and this is often preferred by leaders, a person can retain her power by pleasing only the small set of individual essentials. These can be accommodated through the use of private goods. Consider that dictators are chosen by a small winning coalition. Therefore, to retain power the leader needs to provide only a relatively few supporters with individual privileges, such as lucrative business contracts, sinecured “make-work,” bribe money, and so on.

The problem, from the point of view of some leaders, comes when W is large. When W gets big enough there are too many people to bribe. The leader must resort to offering good governance, especially the provision of public goods that can be consumed by all. For example, in a democracy leaders are elected by wide swaths of the populace. To keep their positions they need to please large groups of voters.

From this analysis, it is easy to see why systems with a large W, a broad winning coalition, tend to have better policy performance. Their leaders are forced to care about the sizable number of essentials that are important for retaining power. Drs. de Mesquita and Smith maintain, and empirical evidence supports them, that democracies have better policy performance than do dictatorships (e.g., their citizens are better educated and in better health). Notice the implications of this thinking for the organizational sciences. De Mesquita and Smith are suggesting that social responsibility is, in large measure, due to the nature of the winning coalition.

We gain additional insight if we also take into account the size of the selectorate (S). Professors de Mesquita and Smith consider the ratio W/S, the size of the winning coalition relative to the size of the selectorate. This is called the “loyalty norm.” When W/S is small (i.e., the essentials are a tiny portion of the selectorate), then each member of the winning coalition is easier to replace. There is a large pool of available interchangeables who would enjoy a slot on the winning coalition. If an essential (that is, a member of the winning coalition) does not perform as the leader’s likes, he or she can be booted out and a replacement brought in. Consequently, coalition members are less likely to defect when W/S is small, and this gives the leader a strong hand. They are more likely to engage in corruption, and they tend to maintain support by supplying private goods to the relatively few essentials in their winning coalition. Following from our earlier comments, the quality of governance tends to be poor when W/S is small. Here we see how unethical conduct (among other ills) results from the nature of the social system in which the leader is embedded.

When W/S is large (i.e., the winning coalition is an ample portion of the selectorate), then things are tougher for the leader but better for everyone else. In this situation the leader would find that each member of the winning coalition is more difficult to replace. The pool of available interchangeables is small, at least relative to the number of essentials. For their part, coalition members have more viable opportunities to switch their allegiance. For this reason, the leader needs to avoid personal corruption in order to maintain their loyalty. There is more to the matter. As the essentials become a large group, at least with respect to the size of the selectorate, it is usually more economical to provide benefits that can be widely enjoyed. Thus, in large W/S settings leaders tend toward public goods and away from private goods.

For an organizational scientist the findings for W/S are like being haunted by the ghost of Baron Acton. (“Power tends to corrupt, and absolute power corrupts absolutely.”) Consider the implications for employee empowerment. In the management discipline, we sometimes recommend involving workers in decision-making. Perhaps we might bring them into the selectorate. Regrettably, this sort of empowerment will sometimes have the opposite effect we intend, actually making matters worse. The risk is that empowerment could increase the size of denominator (S), thereby reducing the coefficient W/S. According to the theory, the only way that empowerment will improve policy performance is if it increases the number of essentials (W) faster than it increases the overall selectorate (S). Empowerment is probably not the answer, at least not as it is usually discussed in the business literature, unless employees are also brought into the winning coalition.

We are now in a position to answer our original question, and we can do so with reference to the evidence reviewed herein. To restate: “If ethical leadership brings all the performance benefits that we say it does, then why does unethical leadership exist at all?” According to many political scientists, that is a leading question because leaders are not necessarily interested in subordinate performance. They are interested in getting and holding positions of power. To accomplish this objective, they need a winning coalition of essentials. Hence, they sometimes behave unethically because doing so, in certain circumstances, helps them to maintain their group of essentials. This is especially true when the size of the winning coalition (W) is small in absolute terms and also when it is small relative to the entire selectorate (W/S). Under these conditions, poor conduct is an effective way to seize and maintain power, so leaders who do not engage in these behaviors (e.g., rewarding loyal essentials with private goods) are apt to be replaced by challengers that are less morally scrupulous. In short, leaders are acting realistically, though self-servingly, as a result of incentives within their work environment. If we want leadership that is moral (and also effective), then an effective strategy would be to increase the size of the winning coalitions, especially in relation to the size of the selectorate. For instance, we might involve teams of workers and managers in promotion decisions or we might have union participation in the performance reviews of senior managers.  The key would be to change the way that people obtain and hold power.

One does not need to endorse everything in selectorate theory or in The Dictator’s Handbook; I certainly would not. Nor does selectorate theory explain everything that goes well or wrong in the workplace. Even with these caveats, we should not underestimate the power of a conceptual model, which begins with a novel starting assumption about leader motivation. According to Professors de Mesquita and Smith, unethical conduct, and poor governance more generally, is not a random failing brought about by incompetence, evil intent, or ill-will. Rather, it is a rational means of pursuing and maintaining social power, at least when W and W/S are small. Tellingly, according to this view business leaders are neither monsters nor geniuses. Rather, they are people, trying to make the best of the situation in which they find themselves. Sometimes that “best” is less than we might wish it could be.