II. Corruption in “Bad” Systems: Functional Corruption

We generally disapprove of corrupt acts. In fact, it appears that we have innate, and in fact evolved, brain architectures that nearly force us to react angrily to seeing others violate social norms.^{Footnote 15} However, there may be several moral considerations operating at once, and things that are objectively “corrupt” if no one does them may be perfectly socially acceptable in a context where, as the saying goes, “everyone does it.”^{Footnote 16} That doesn’t mean that there is no such thing as morality, but it may imply that our reaction to corruption is deeply contextual.

Further, it may be easy—because humans are clever, and in fact this cleverness is likely also evolutionarily adaptive—at coming up with facile justifications for why, though an action might generally be unacceptable, in this case it is pragmatically allowed, and perhaps even conditionally morally permissible.^{Footnote 17}

There are many acts that we might think of as corrupt, or almost corrupt. When I was in college I had a job as a night manager of a “Burger Chef,” a downscale fast food restaurant in Winter Garden, Florida.^{Footnote 18} A city cop would sometimes come by at about 10:30, when we were counting up receipts from the night in preparation to close. The cop could have been all sorts of places, but he came by Burger Chef. I had the front staff make sure to give him a big greeting, and to have a fresh pot of coffee ready to give him a cup, along with a fried apple pie. At first we didn’t let him pay when he pulled out his wallet; later, he stopped even offering.

The policeman would show up apparently randomly, different nights of the week, but always just before closing time when we were counting money in the back. Once when the owner of the store was in late and saw our little ceremony he waited until the cop had left and then called me into his office. I could see he was furious.

But not the way you might think. The owner said, barely avoiding yelling: “If a cop comes in just before closing time, you give him a g*****n fish sandwich and fries, plus a coffee and a pie!” The owner had several stores in not very nice areas, and having a police car parked out front near closing time on random nights, but fairly frequently, was cheap insurance.

Was it corrupt? I think most people not in that situation would say yes.^{Footnote 19} The policeman should have paid, and we should not have offered. As DeLeon put it:^{Footnote 20}

Robert K. Merton blamed corruption on the culturally corrosive influence of wealth:^{Footnote 21}

Thus, there are individual goals that can be constrained by formal rules, but “society” can also construct and perpetuate what are essentially soft, moral constraints that people still perceive as binding.^{Footnote 22} If these moral norms are transgressed, society has considerable resources—shame imposed by peers, guilt felt by violators, shunning, and so on—to enforce these “other” rules.

But what if society’s other rules sharply constrain economic activity? What if it is perceived that the government and its officials are themselves acting in ways that raise obstacles to business and obtaining the necessities of everyday life. Merton argues that in this circumstance corruption can be “functional,” meaning that society may simply look the other way rather than sanctioning transgressions, because everybody knows, citizen and official alike, that corruption makes the system work better?

Johnston also notes the benefits of functional corruption in dysfunctional systems:^{Footnote 23}

Thus, in a socialist or primitive command system, corruption actually creates a space where the price mechanism can function. The size of this “grey” or “black” market can be allowed to grow or shrink as the needs of the society change, creating wealth for many participants and allowing citizens to obtain food, clothing and other necessities in a way the forestalls actual revolution while preserving the formal powers of the state apparatus.

Johnston goes so far as to spell out a set of sufficient (though not necessary) conditions under which Merton’s “functional corruption” will occur.^{Footnote 24}

• • First, the “fruits” of government action, or the costs of inaction, must be sufficiently large that there is excess demand for these services. Thus, there must be some reason to want new telephones, to build new factories or apartment buildings, or to need permits to engage in some economic activity. This means that the state needs considerable “capacity,” in the sense that it can enforce sanctions that would otherwise prevent economic activity, and that it can deliver services (roads, electricity, telephone connections) when mobilized to do so.

• • Second, these benefits or escape from these sanctions can only be obtained by “dealing” directly with the government. Organized mafias that act as intermediaries can either replace, or thwart, the development of functional corruption. Thus, the state is obliged to secure its Weberian role as the monopoly provider of security and the sole legitimate user of violence and force. This is another kind of state capacity, and it represents a stable alternative to the development path of seeking to negotiate an “open access order,” as we will later see.

• • The “routine process” by which needed benefits are created, recognized, or conferred is time-consuming, inconvenient, expensive, and possibly uncertain. Alternatively, sanctions are arbitrary and devastating. These conditions may be endogenous, of course, to create a maximal difference between those who “pay to play” and those who try to avoid paying bribes or speed money.

Calling these sufficient conditions means that corruption will (not can) emerge. In such systems, functional corruption is a necessary lubricant for social intercourse and economic interaction.^{Footnote 25}

This calls to mind Adam Smith’s famous distinction about “good” and “bad” people, as opposed to good or bad systems. Smith was describing the British East India Company, but he could as easily be talking about citizens with corrupt governments, or even individual corrupt officials themselves:

So is “functional corruption” acceptable? As I will argue in the next section, the answer is no. But at the same time we can understand why people might excuse others, and certainly themselves, from the requirements of acting in ways that are not corrupt. Corruption is “good,” or at least acceptable, in a Pareto sense insofar as it benefits those who engage in the corrupt acts and harms no one else. In fact, if “the system” depends on corruption to get anything done it may appear that everyone benefits from the increased economic activity. Corruption can facilitate exchange in ways that are Pareto improving.

The problem is that the expectations of corrupt actions become part of the architecture of the social order itself. As in the case of the “junior engineer” in India, the right to collect extra-legal fees outside the system is capitalized as part of the “value” of the job to its incumbent. Any attempt to change the system will be thwarted by the need to compensate the incumbent for the rents that led to the efforts, and probably side payments, that allowed that person to acquire the job in the first place. Thus corruption is a problem, and is morally bad, insofar as it creates and fortifies an institutional order that prevents institutional changes that would imply Pareto improvements.

The distinction that I am making between short run and long will be unfamiliar to anyone not steeped in economics, so let me try to clarify it. The “short run,” in terms of my analysis, is a situation where institutions are fixed and the only cooperative activities are exchange of items already determined. The “long run” allows for larger and deeper innovations, both in terms of items being exchanged and the institutional setting in which cooperative activity can take place.

For example, as we will see in the next section, in a system where the only form of electronic communication between individuals is analog telephones, it may actually be economically beneficial to almost everyone involved to allow black market “bidding” in the forms of bribes to speed delivery of service to those persons or businesses with the highest values for this access. Compared to not allowing such bribery, in fact, a system with bribes will allocate telephone access more efficiently and will create higher levels of economic activity.

On the other hand, the fact that the business of allocating access to landlines is so profitable, in terms of bribes offered, will mean that potentially Pareto-improving changes to a system of cell phones will be fiercely opposed by those who profit from the existing locally optimal but globally inefficient system. Cell phones completely cut out the bureaucrat who must allocate physical landlines, and sharply reduce the ability of the authorities to hold up service.

As Merton put it:

In nations with “bad” institutions, corruption may be a rational response to dysfunctional or badly structured state capacity when there are substantial economic benefits to finding a “work around.” That likely means that the system of socially transgressive and furtive delivery of benefits and protection from sanctions itself becomes formally institutionalized, or nearly. Both the citizens who depend on cozy relations, and the government functionaries who depend on income, are likely to resist efforts at reform. And we are left to compare what is “seen”—a system that works, and in which government bureaucrats make a comfortable, reliable living, and phones and other services are delivered in an understandable and predictable way—with what is “unseen”—a system where technological advances disrupt old ways of doing business, and in which rapid economic growth delivers only uncertain benefits to any particular citizen. There is no way to guarantee that the current winners, at the inferior local optimum, will have anything like the same level of relative status, or even absolute income, at the new, superior global optimum.

Merton’s rather breezy conclusion—“alternative structures may fulfill [the services now delivered through functional corruption] without necessarily entailing its social dysfunctions”—is thus too optimistic. No one is confused, and having some development aid specialist come over and lecture about “social dysfunctions” will just bring laughter; compared to what? What can the nation that depends on functional corruption actually do about it?

III. Corruption as an Extension of Limited Access Orders

North, Wallis, and Weingast (2009; hereafter NWW) is modestly subtitled “A conceptual framework for interpreting recorded human history.” A short summary of their thesis is useful at the outset of this section.

The major question of development, including the creation and maintenance of institutions in which corruption is effectively controlled, might be asked in either of two ways. First, why is it that many nations fail to develop “good” institutions and thriving economies? Second, why is it that some nations succeed in developing “good” institutions and thriving economies? I understand that this posing of “two” questions where one is just the negative of the other seems facile, but they are in fact completely different questions.

The approach to answering is going to depend on whether the analyst believes that the “natural” state of human society is development and progress, or stasis and entropy. The account of much of modern economics is that the development of institutions to facilitate exchange is Pareto-improving. Consequently, barring some cultural pathology or social barrier, functional markets will always develop and prosperity will always ensue.

But this does not seem to be a very good empirical description of the world we live in. Many, and in fact most, nations seem to bump along at or near the bottom of the range of growth paths implied by their endowments of population and natural resources. Perhaps the phenomenon needing explanation is not failure, but success. What is it that the relatively few successful nations had that the failed states lack?

NWW claim that the surprise is development; the “steady state” for most of humanity is failed institutions and corruption. The central fact of human society is the “state,” or the centralized system of organization that solves the most fundamental problem, providing social order.^{Footnote 29} In particular, the state must prevent or contain the human capacity for violence, while itself monopolizing and focusing the means of exerting violence. Most states fail in both core functions, failing to prevent violence internally and failing to limit and safeguard its own capacity to exert violence. Citizens hurt each other, in myriad ways, and agents of the state itself harm or threaten to harm citizens in ways that violate the formal rules and constitutional provisions on which state power is ostensibly founded.

The problematic transition, then, lies in the ability of a state to move from a system where order is imposed top-down to a system where order emerges from permissionless innovation by individuals or private groups as groups of citizens and other people outside the nation order and reorder themselves into temporary, highly decentralized structures such as firms, voluntary associations like non-profits, and other kinds of weakly connected voluntary communities. As has now become part of the jargon, such systems of emergent order must facilitate permissionless innovation without extracting rents or blocking the development of the order by requiring permissions in the first place.^{Footnote 30} The problem, of course, is that the state’s capacity to exert violence gives it, or its agents, the formal and also the informal power to exact payments in exchange for permission, permits, or simply the withholding of threats of force.

NWW argue that both historically and for good theoretical reasons the “types” of state organization fall into one of three basins of attraction. Any system, once established, has considerable inertia and internal mechanisms for continuity, and can only be changed through wrenching transformations that often, at least in historical experience, involve violence and destruction of the essential institutions of the old society.

• • Primitive social orders, or hunter-gatherer societies with traditional charismatic or hereditary rulers.

• • Limited access orders, which have dominated the last 10,000 years.

• • Open access orders, which have arisen only in the past 300 years.

Limited access orders contain violence by political manipulation of the economic system to generate rents. These rents are created by using rules and decrees to limit entry by competitors, and by using the capacity for state violence to defend those rents, while collecting the tax revenues or tribute created by economic agents competing for those rents. The very survival of limited access orders depends on the ability of the state to create rents and then reliably to generate a stream of future rents by making a credible commitment to use violence to protect the rents that it is selling. If any other agent has access to the means of focusing violence, those rents are jeopardized, so the individuals who receive rents have large incentives to suppress organizations that might commit violence or foment social change. Rent “owners” have no choice but to support the current regime.

Open access orders, by contrast, foster the creation of emergent orders, and constrain violence and sustain social order through political and economic competition rather than rent-creation. Open access orders have developed in about two dozen countries, and all are both economically and politically developed.

To be sure that the distinction is clear, these latter two types of orders—limited access orders and open access orders—are entirely separate and generally internally stable equilibria. Limited access orders depend on restricting competition to survive; open access orders depend on human generated competition to grow.

To see how this approach illuminates the problem of corruption, it is useful to understand North’s distinction between “institutions” (the human devised rules of the game that shape and direct human activity) and “organizations” (the optimizing responses the emerge spontaneously in response to the incentives and constraints presented by the institutional setting). Each of the three “institutions”—primitive, limited access, and open access orders—evoke a different set of organizational responses. As NWW put it:

The reason that this insight is important for present purposes is that limited access orders are locked into the preservation of rents that prop up the power structures in the state and create rents—artificial value, but extremely profitable to the “owners”—in the “economy.” The “junior engineer” in the Indian system who saved up to make a payment of 10 million rupees expects to get a return on that “investment,” and that return takes the form of being able to collect fees and speed money for the years it takes that engineer to save enough to be able to buy a position as a senior engineer. Again, NWW note the difference; limited access orders actually “preclude” the development of markets, permissionless innovation, and long-term development. This is not simply a difference at the margin, but fundamental game-changer for development:

The theory advanced in this section relies on the observation NWW make about limited access orders. The ability of the state to control and focus violence enables it to create and protect rents. These rents might take the form, in early times, of a monopoly on the “right” to sell salt, or tea, or spices, within the area controlled by the state. This right is valuable, and since it is excludable the state can auction off the right in exchange for revenue.

These formal rent-protected revenue sources may sustain a state for long periods, but in every case they preclude innovations and creative destruction of providing these services in new, better, or cheaper ways. This feature of limited access orders alone might sharply constrain economic development, because it stultifies innovation. But there is an additional implication I am trying to draw out, in the way that limited access orders also foster and then preserve corruption.

This implication has to do with the signals provided by a properly functioning price system. Even a monopolized price system will reflect accurately differences in demand and the opportunity costs of resources, though it will incur the costs of deadweight loss and limited innovation that are well-known for monopolies generally.^{Footnote 33} But limited access orders cannot be expected even to generate “efficient” monopoly prices! Schleifer and Vishny argue that the real problem with using bottlenecks, permitting, and other artificial scarcities to create rents is that the price mechanism is almost completely suppressed. Most transactions will be based on an exchange of favors, or the pursuit of political advantage, rather than being based on the scarcity signals contained in prices.^{Footnote 34}

A limited access order where elites charged monopoly prices would be much more prosperous and dynamic than a society where many rent-holders obtained their rents through corruption. In a corrupt system, goods and services are allocated based on political advantage. Each “price” charged is actually the sum of the value being paid, the waiting costs of the purchaser, and the rent represented by the difference between these values and the shadow price represented by the “consumer surplus” of the purchaser or supplicant.

Elites and rent-controllers will generally charge less—possibly much less—than market-clearing prices to secure political goals and to satisfy prior promises of quid pro quo arrangements. The result is that prices cannot carry out their function, which is the impersonal coordination of the behavior of individuals.^{Footnote 35} “Natural” states (as NWW call primitive and limited access orders) cripple the ability of prices to coordinate impersonal activity, to enable permissionless innovation, and to convey information about the appropriate margins where cost, demand, and scarcity indicate activity should be focused.

The problem, then, connects to a general argument about the efficiency of cooperation and competition in different settings. Economists have often argued that competition is always beneficial, and that prices always “work” to allocate resources efficiently. But competition can be diverted from the dynamic creation of products and services resulting in effective creative destruction, and toward rent-seeking competitions to protect inefficient economic organizations against creative destruction.^{Footnote 36} If the form that competition takes is to seek the attention of powerful state actors in protecting artificial rents against new entrants, and if the form that price signals take is to measure out the cost of obtaining such protection, the actual functions of the markets are co-opted and corrupted beyond recognition.^{Footnote 37} What looks like competition is diverted into the particular means—bribes, favors, and the strategic erection of tollbooths near bottlenecks—that characterize limited access orders. Corrupting the market function of prices corrupts the society.

IV. Lock-In Limited Access Orders

There are two threads of argument, made separately, that in this section will be tied together.

The first thread is the claim that a nation with dysfunctional or malformed institutions can be improved by corruption. If access to basic services, licenses, and permits is controlled by a grossly inefficient bureaucratic system, it is best to resort to speed money and make the system more efficient. In fact, if the value created by the activity is sufficient, a bureaucratic bottleneck can be overcome. We might think of this as an application of the “Coase Theorem”: the value created being large guarantees that the activity will take place. The details of how bureaucratic obstacles are overcome are essentially epiphenomenal; entrepreneurs simply treat those obstacles as one more problem to be worked around. Institutions of grey or black markets will emerge if Merton’s three conditions of “functional corruption” are satisfied.

The other thread might be labeled the “transitional gains trap” or “lock-in.” This is two problems in one: problem A is that attempts to give money to particular groups will create rents or special benefits attached to certain features. For example, farmland for the very poor might be subsidized. But then the value of the asset associated with the rent will rise, capitalizing the rent. That means that the subsidy has no net effect, because the poor person now earns a normal return on the now more expensive asset that capitalizes the rent. But if the government then tries to eliminate the program, policy, or informal practice there is a substantial and possibly devastating loss to the “owner,” who suffers a capital loss. Having paid more for the asset in the first place, there is no lasting benefit but there is a cost if the rent is removed.

Problem B is that corrupt systems have exactly this “transitional gains trap” feature, such that any successful attempt to eliminate corruption will impoverish an entire class of bureaucrats, experts, and other “mandarins” who worked for years to acquire their current positions. The reason they worked or bribed their way into these positions is that they produce substantial revenues in the form of fees and bribes. The fact that the state itself does not formally sanction or approve these “assets” does not reduce their value. Eliminating corruption, even if in principle it is potentially Pareto-improving, becomes nearly impossible unless the state can buy out the holders of asset titles in the informal, corrupt economic system.

To summarize, then, corruption (1) makes bad institutions more tolerable, and may go a long way through a Coasian logic toward rendering them efficient; but (2) locks in bad institutions because of the transitional gains trap. This means that we might expect countries with problems with corruption to find that any kind of effective reform is very difficult. Corruption is not as growth-inhibiting as you might expect, under some circumstances. But efforts at reform will usually be diverted, exhausted, or even co-opted.

Corruption may facilitate exchange and efficiency given a set of rules, but it will make changing the rules or institutions of the society much more difficult, even if that change dramatically expands the overall size of the economic pie. Paradoxically, at least to an outsider, attempts at reform may engender moral outrage of insiders, because the attempt to change “the rules” is both an economic threat and “just not how we do things.” As Brennan and Buchanan put it: “The rules to which they tacitly agreed when they decided to participate have been changed. They have not implicitly promised to obey the new rule, so they cannot be accused of unjust behavior if they violate it.”^{Footnote 38} The corrupt bureaucrat was violating the formal law, but following the informal rules, in charging a “fair” level of speed money in exchange for the accelerated delivery. The attempt to end corruption amounts to an attempt to force a return to the law at the expense of the rules everyone has become accustomed to, and around which they formed their expectations.

As Douglass North put it:^{Footnote 39}

This conclusion is what I have tried to support here. The original form of the “lock-in” of institutions is economic, because the corruption of the system creates economic rents. But as time passes the rules are locked-in by expectations. We follow “the rules” simply because they are the rules.^{Footnote 40} But this takes a considerable amount of time to internalize, meaning that in the “short run” corruption is impervious to efforts at reform. The problem posed by North, Wallis and Weingast is elegantly simple, but perplexing: while in economics the “long run” is simply the shortest amount of time when all inputs are variable, in political economy the “long run” is the period required to break, modify, and lock in new institutions. The “short run” can therefore last a very, very long time.

A. An empirical comparison

To get a (simple) view on the problem I put together data on growth rates over the period of 2010-2015 from World Bank sources. I divided parts of the data (the sample is only 40 countries, for reasons I will explain) into four categories, and presented them in Table 1.

Table 1. Fraser Institute Freedom Index with GDP/Capita Growth Rates

Panel A: Good property rights AND low corruption

To form the categories, I used two different indices in the Fraser Institute’s Economic Freedom of the World.^{Footnote 41} The information presented is in no way a “test” of any hypothesis, but is rather intended to illustrate differences in corruption and other institutions.

The first category (Panel A) is the “good” countries. These nations have the highest sum of the “good property rights” and “absence of corruption” indices. They may have (as in the case of Qatar) other problematic institutions, in the sense of political repression, but they have good economic institutions and noncorrupt state apparatus. The mean growth rate in this category is 3.2, and the level of GDP per capita is well above the world average.

The second category (Panel B) is failed states, the nations with the lowest sum of the property rights and corruption indices. These are nations with both terrible property rights protection and appallingly corrupt institutions. Perhaps unsurprisingly, these countries are a hot, awful mess. The average growth rate is substantially negative, with only Argentina and Angola showing any kind of movement. The uncertainty that comes from lacking property rights and constant abuse by corrupt officials is the closest thing in the modern world to Hobbes’ “state of nature.”^{Footnote 42}

The third category (Panel C) represents the nations with the largest difference between property rights and corruption, measured by the Fraser Institute indices. That is, these nations have (on paper) pretty good property rights rules, but are infested with corruption, particularly (as in the case of South Africa) in some regions or some sectors of the economy. The result is a mean growth rate of 2.2 percent, with corruption sucking the vitality out of the creative impulses of entrepreneurs who are unable to capitalize on “good” institutions because of “bad” government.

The final category (Panel D) represents those nations with the largest negative difference between property rights and corruption. That is, these nations have pretty terrible property rights rules, but moderate corruption.^{Footnote 43} If corruption were always bad, we would expect the nations in Panel D to perform worse than those in Panel C. But that is not the case. The absence of protections in the sense of “bad” property rights rules is mitigated by access to corrupt work-arounds. The growth rate for the nations in Panel D is a healthy 4 percent.

If we were to classify China, it would belong in Panel D (though measuring averages of either property rights or corruption as an average across such a vast and diverse country would be nonsensical). The Chinese system is moderately corrupt, with old corrupt systems (nepotism, local norms of privilege for some families) remaining as a stubborn residue under new forms (cadres of the Communist Party who look the other way, or actively invest, when new and essentially private business enterprises are formed using “public” property and capital). Still, given how dysfunctional the formal institutions are in many areas it is hard to begrudge the positive effects that functional corruption have had. Half a billion Chinese citizens have escaped poverty in less than forty years, even though property rights are vague and contested. The problem, of course, will be escaping from the locked-in commitments made to corrupt officials, who can easily block reforms to any kind of open access order. In other words, China has progressed rapidly, and has done it in perhaps the only way that was feasible. But, what now?

V. Conclusion

This essay has argued that corruption, while morally questionable as a first-best “institution,” may be indispensable for mitigating the influence of bad institutions. Functional corruption, using Merton’s formulation^{Footnote 44}, can always be expected if three conditions are met:

• • There is enough economic opportunity to attract entrepreneurs, and there is enough state capacity to provide the services entrepreneurs require. In particular, the state can enforce sanctions that would otherwise prevent economic activity, and it can deliver services (roads, electricity, telephone connections) when mobilized to do so.

• • The state is strong enough, and organized enough, to control access to these services or solutions to hold-up problems, meaning that the state can secure its Weberian role as the monopoly provider of security and the sole legitimate user of violence and force, keeping mafias and organized crime from threatening property rights or extorting protection money.

• • The “routine process” by which needed benefits are created, recognized, or conferred is time-consuming, inconvenient, expensive, and possibly uncertain. Alternatively, sanctions are arbitrary and devastating.

However, while participation in corrupt practices may be morally permissible in a system where such practices are widespread and accepted as “part of doing business,” there is a pernicious consequence. Accepting corruption as conventional erodes respect for rule of law, and creating the informal state capacity to make functional corruption work in the first place creates a set of rent-producing entitlements that lock in corrupt institutions because of the transitional gains trap.

The result is that the transition from limited access orders to open access orders, the key transition for economic and political development, can be thwarted, possibly indefinitely. Even if open access orders are (potentially) Pareto superior, in the sense that they create dynamic economies that foster permissionless innovation that dramatically increase “the size of the pie,” there is no way to compensate the stakeholders who depend on particular rents in the limited access order.^{Footnote 45}

Perhaps more importantly, there is also a deeper problem. As Gaus argues, “It can’t be rational choice all the way down.”^{Footnote 46} It is asking a great deal for everyone to know and obey the law (as opposed, possibly, in the Hayekian distinction, to statutes or formal rules).^{Footnote 47} Asking them to know at the same time whether the law is good is too much. In the “short run,” as I have used the term, the acceptability of corrupt acts may yet be in flux, and the expectations of those in authority are not locked-in. But once everyone knows “that’s the system,” then those practices become both part of what is allowable for me and what I expect of you. As North argued^{Footnote 48}, the rationality assumption does not allow us to impose rational changes in institutions or moral norms.

Thus, while it is tempting to see corruption in a static sense as only contingently vicious—everybody does it, and it actually enhances growth in nations with bad institutions—the very Coasian nature of corruption as a work-around prevents growth in the longer term, dynamic setting. The transaction costs of transition for limited access orders infested with corruption to open access orders are prohibitive. Corruption is vicious.