The Uses of Diversity. David Ellerman
also extends to basic questions of handling risks. One often finds the almost knee-jerk response to any risk problem; the solution is the parallel-oriented strategy of diversification. “Don’t put all your eggs in one basket.” Use one’s resources to place many small bets rather than one or a few large bets. In terms of a tree search, this is the strategy of making an extensive parallel search of the first boxes on many different branches rather than an intensive search over the series of boxes along one branch.
The usual diversification recommendation to “spread your eggs between many baskets” makes the implicit assumption that the baskets have fixed characteristics (e.g., a fixed probability that the eggs in a basket will be broken by the basket tipping over or being dropped). But if the characteristics of some baskets can be significantly changed by a commitment of attention and control to one or a few baskets, then the dual risk reduction strategy emerges. Put all your eggs in one or a few baskets, watch them very carefully, and make sure the baskets are not tipped over or dropped.10
The logic of commitment enters through the door of flexible characteristics—in this case, the risk characteristics. The probabilities of a person having an accident might not be fixed independently of the person’s activities but might depend on the precautions taken by the person. When the logic of exit (diversification) works too well, for example complete insurance against the consequences of accidents, then the insured will neglect taking precautions and hazards will increase, that is, moral hazard. To combat moral hazard, one has to move back a little way toward the logic of commitment by exposing the insured to some of the costs of accidents, for example deductibles and co-pay in medical insurance.
This foreshadows a basic theme. When the choice environment is parsed into a set of options each having certain characteristics, the choice of a commitment-oriented strategy or an exit-oriented strategy will in part depend, respectively, on the extent to which the characteristics are flexible or are fixed.
Two Biological Reproductive Strategies
The choice of risk reduction strategy hinges on the question of whether or not a commitment of resources to one or a few options can significantly change the risk characteristics of those options. In biology, all organisms face the reproductive risk that their genes may not survive into the next generation. Since organisms only have limited reproductive resources to address this risk problem, they tend to “choose” one or the other of the “corner solutions” to solving this risk reduction problem.
If organisms have little or no control over the risk characteristics faced by their offspring, then the organisms will tend to favor a diversification strategy called r selection. They spread their reproductive resources over many offspring as with most insects and fish, and then the offspring face the odds with little or no parental assistance. The emphasis is on the quantity rather than quality of offspring. However, if the parents can significantly control the risk characteristics faced by their offspring and if the offspring can be produced with a “quality” to improve their odds, then that species will tend to favor the commitment-oriented strategy called K selection. They invest their reproductive resources in a few “quality” offspring as with mammals and then they provide intensive parental care to the offspring.
Two Ways to Increase Reliability
Another example is from reliability theory in engineering. The risk is system failure when a component breaks down. The parallel-oriented approach to reducing that risk (increasing reliability) is redundancy. In the general parallel (exit-oriented) case, there may be two ways to explore parallel options, one at a time or, if possible, simultaneously. One form of exit-oriented behavior is to try one option, then another, and so forth. Another form of exit-oriented behavior is to try many options simultaneously as when eggs are spread between many baskets or when biological resources are spread over many offspring. These two forms of the parallel-oriented approach come out clearly in reliability theory.
The one-at-a-time approach to redundancy (parallelism) is called standby redundancy. The system uses one option or subsystem until it breaks down and then it switches to the standby system. An airplane might have a standby radio or navigational system; if the primary system breaks down, it can switch to the standby system. The simultaneous approach to parallelism is called active-parallel redundancy. For instance, an airplane with several engines is using active-parallel redundancy; if one engine breaks down, hopefully the others will suffice to land the plane safely. Otherwise, they may have to switch to the standby way of getting to the ground safely, parachutes.
Scarce resources are used to build reliability into a system. In some cases, using redundancy to increase reliability may not be technologically possible. The dual way to increase reliability is to use resources to improve the reliability characteristics of the components of the one system. Thus, the trade-off between the two logics in using resources to reduce risks and increase reliability is “redundancy versus improved elements” (Von Alven 1964, 240).
These examples illustrate the second point that there is another way to reduce risks than just distributing one’s eggs between many baskets.
The Two Ways to Get Flexibility, Performance, and Efficiency
Flexible Options or Flexible Characteristics in Given Option?
In the options and characteristics model, the two system logics differ by what is fixed and what is flexible. With the parallel- or exit-oriented logic, the characteristics of an option are taken as fixed, so flexibility is obtained by exiting one option and taking another. If workers do not have the desired skills for a new technology, then relocate or fire those workers and hire others with the desired characteristics. With the series- or commitment-oriented logic, the option is taken as fixed, so flexibility is obtained by changing the characteristics of the option. Workers with some commitment from the company for employment security would be more willing to learn new skills and retrain for new technologies since they would not be “working themselves out of a job.” It is not a question of flexibility or not, but of achieving flexibility in one way in one system or in another way in the other system.
Exit versus commitment are the two logics that run through the design of institutions. The market is an institution that operates largely on the basis of the logic of exit. Economics developed first as the theory of markets and market behavior, so economics tends to see the world through an exit-oriented lens. “The economist tends naturally to think that his mechanism (exit) is far more efficient and is in fact the only one to be taken seriously” (Hirschman 1970, 16). For instance, under the exit-oriented logic all labor questions are “labor market” questions, while under the alternative commitment-oriented logic (e.g., in a Japanese-style firm), a labor question is a “human relations” or “human resources” question.11
Organizations (including political units or polities) would seem to be the natural setting for commitment-based strategies. If there are some costs or barriers to exit in the face of dissatisfaction, then the exercise of voice may be the better way to change things. Hence, one might expect to see organizations designed on the basis of the logic of commitment.
Allocative Efficiency or X-Efficiency?
More broadly, take an option to be a particular use of a resource and take the characteristic of the resource to be its productivity or effectiveness in that use. Then the two system logics give two ways to get improved performance and efficiency. The exit-oriented strategy is to move resources to higher-productivity uses (e.g., through the market) and the commitment-oriented strategy is to get higher productivity out of resources in the given use (e.g., in an organization). The exit-oriented notion of efficiency is allocative efficiency associated with markets. Resources in a certain use have fixed productivity so it is a question of the allocation of resources to the higher-valued uses. The commitment-oriented notion of efficiency is X-efficiency (Leibenstein 1966; 1984) where the principal variable productivity of a resource in a given use (e.g., in an organization) is human effort.12
If a skill is quite standardized and available on the market, then in that case the exit-oriented logic can work in a firm. The threat of dismissal calls forth more effort. The low-trust system is self-reinforcing in its system logic. Low trust leads to highly explicit contracts with competitive arms-length relationships