The ORE of Politics will be available for subscription in late September. Speak to your Oxford representative or contact us to find out more.

Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, POLITICS ( (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 24 September 2017

The Two-Good Theory in Practice: From Abstract Generalization to Specific Inference

Summary and Keywords

The “two-good theory” is a theory of foreign policy that is meant to apply to all states in all situations; that is, it is general. The theory is simple and assumes that states pursue two things in theory with respect to foreign policies: change (altering aspects of the status quo that they do not like) and maintenance (protecting aspects of the status quo that they do like). It also assumes that states have finite resources. In making these assumptions, the theory focuses on the trade-offs that states face in constructing their most desired foreign policy portfolios. Further, the theory assumes that protecting realized outcomes is easier than bringing about desired changes in the status quo.

The theory assumes that states pursue two goods instead of the more traditional one good; for realism, that good is “power,” and for neorealism, it is “security.” This small step in theoretical development is very fruitful and leads to more interesting hypotheses, many of which enjoy empirical support. The theory captures more of the dynamics of international relations and of foreign policy choices than more traditional approaches do.

A number of empirical tests of the implications of the two-good theory have been conducted and support the theory. As the theory can speak to a variety of foreign policy behaviors, these tests appropriately cover a wide range of activities, including conflict initiation and foreign aid allocation. The theory enjoys support from the results of these tests.

If the research relaxes some of the parameters of the theory, the investigator can derive a series of corollaries to it. For example, the initial variant of the theory keeps a number of parameters constant to determine the effect of changes in capability. If, however, the investigator allows preferences to vary in a systematic and justifiable manner (consistent with the theory but not established by the theory), she can see how leaders in a range of situations can be expected to behave. The research strategy proposed, in other words, is to utilize the general nature of the two-good theory to investigate a number of interesting and surprising implications. For example, what may one expect to see if the United States supplies a recipient state with military aid to counter a rebellion? Under reasonable circumstances, the two-good theory can predict that the recipient would increase its change-seeking behavior by, for instance, engaging in negotiations to lower trade barriers.

Keywords: two-good theory, substitutability, empirical international relations theory, power, security, neorealism, foreign policy


The two-good theory of foreign policy was developed with the goal of providing an abstract, general theory of comparative foreign policy. That is, the hope was that it could explain all types of foreign policy decisions, for all states, in all historical state systems. It was specifically designed to help us understand how the use of various foreign policy tools (e.g., foreign aid, alliance formation, dispute initiation) interrelate and to uncover “substitutability” relationships among these tools. Our purpose is to evaluate the development and testing of this theory, with a particular eye toward discovering where further theorizing and further empirical analysis may usefully proceed.

The Two-Good Theory: Key Assumptions

Although a formal, mathematical model is at the core of the two-good theory (see especially Palmer & Morgan, 2006, chapter 5), it can be summarized in nontechnical terms. The theory is based first and foremost on the assumption that all foreign policy actions require resources and that the resources available to any state are finite. Even actions that are viewed as being relatively costless, such as issuing a diplomatic protest, require the time of diplomats and the (perhaps only primitive) technology to communicate. Other acts, such as the initiation of military conflict, entail significant costs in terms of lives possibly lost, money spent, and the unattractive possibility of defeat in combat. This simple assumption has a major implication for how we view what is to be explained. While most theories of foreign policy focus on decisions to use particular policy tools, such as force or economic sanctions, the two-good theory requires that we view a state’s primary foreign policy decision as focused on how to allocate limited resources across all available foreign policy instruments. The two-good theory views the use of all policy instruments as being interrelated and interdependent, at least in the sense that any resource devoted to the use of one instrument cannot be devoted to the use of another. The theory is focused not on explaining specific uses of specific tools in specific circumstances; instead, its goal is to understand the construction of foreign policy portfolios—the bundle of policies a state utilizes—given its resource constraints and preferences.

Second, the two-good theory is based on the assumption that all states at all times seek to produce two goods (things of value) through their foreign policy portfolios. These goods are labeled “maintenance” and “change.” Obviously, states actually pursue myriad more specific and precisely identifiable goals, from disease prevention, cultural exchanges, and increased trade to self-preservation, yet predominant explanations of foreign policy behavior are built on the assumption that states pursue a single thing of value—security. This generalization quite clearly characterizes neorealism. The influence of this assumption is illustrated by a brief perusal of the U.S. Department of State’s website, which shows that nearly every policy or foreign agreement is justified on the ground that it “enhances the security interests” of the United States. This is a useful abstraction, as it would be impossible—and not terribly instructive—to identify every interest affecting every foreign policy decision. One of our central concerns in developing the two-good theory is that assuming that all foreign policy actions are intended to produce a single good—security—makes it impossible to acknowledge, let alone consider, that trade-offs must be made in any policy decision. By incorporating the assumption that states seek to produce two composite foreign policy goods, the two-good theory complicates the traditional abstraction as little as possible (in that two is the smallest whole number greater than one) while allowing for considerations of the trade-offs that are inevitably involved when foreign policy decisions are taken.

The “two goods” of the theory, maintenance and change, are defined in terms of whether the goal of a policy is to alter some aspect of the status quo or to preserve some aspect of the status quo. The intuition behind this follows from the assumption that all actors at all times are satisfied with some things in the world and are dissatisfied with other things. Any policy that is intended to prevent an undesired alteration in the status quo is said to be producing maintenance, while any policy that is intended to bring about a desired alteration is said to be producing change. While all states seek to produce some of both goods, it is assumed that states vary in their preferences regarding the trade-offs they are willing to make across the two goods. Some are relatively maintenance seeking, and others are relatively change seeking. This can also vary across time for a given state. So, for example, the isolationist United States of 1930 was willing to sacrifice a great deal of change for a small amount of maintenance, while the globalist United States of 1950 was willing to sacrifice maintenance for change (see Palmer & Morgan, 2006, chapter 3).

Unlike in other theories that presume that some states are status quo states and others are revisionist, no moral judgment is assigned to the pursuit of either good, that is, no assumption is made suggesting that some types of states inherently seek change while others inherently seek maintenance; likewise no assumption is made suggesting that states are concerned only or even primarily with the structure of the interstate system. Desires to change or maintain the status quo can be over anything. For example, a state that devotes some of its resources to disaster relief can be motivated completely or in part by an altruistic desire to change the status quo by easing the suffering of disaster victims or by a desire to undermine the legitimacy of the local government authorities. This saves us from having to engage in the often-tortured exercise of arguing that such actions somehow enhance the security of the donor. Assuming that all states, at all times, seek to produce some quantities of both goods also frees us from the urge to assign blame for conflict.

Third, the two-good theory assumes that it is easier to maintain the status quo than it is to change it. In part, this captures the fact that efforts to defend the status quo can provide protection against multiple potential challengers simultaneously, whereas efforts to change the status quo must be targeted and directed. This also reflects a widely shared conventional wisdom that the status quo is sticky. This simple assumption has a critical implication: as state capabilities increase, a state will be able to produce more of both maintenance and change, but its ability to produce maintenance will increase at a decreasing rate, and its ability to produce change will increase at an increasing rate. Stronger states are better set to alter that world to their liking than are weaker states. Rather than seeing the strong as status quo oriented, the two-good theory’s assumption has it that weaker states lack the capabilities to change the status quo. Essentially, the theory sees change as a luxury good and maintenance as a necessity.

Obviously, states must address domestic policy concerns and allow for consumption and investment. It is assumed, however, that the budget for foreign policy is exogenously determined and that it is used entirely. This simplification permits trade-offs between foreign and domestic policy as well as those between current and future consumption to be ignored.

Finally, the two-good theory views the foreign policy activities that we observe, (e.g., the provision of foreign aid, the use of military force, alliance formation, diplomatic protests) as the instruments of policy that translate resources into maintenance and change. While almost any policy tool can be used to produce either good, it is assumed that some are better suited for the production of maintenance (e.g., defense spending, forming alliances with stronger powers, reciprocating the use of military force by another country) whereas others are better suited for the production of change (e.g., foreign aid, forming alliances with weaker states, initiating international military conflict). Moreover, it is assumed that the relative efficiency with which a given instrument translates resources into a particular good can vary with technology and with circumstances. For example, extensive trade relationships increase the relative efficiency of economic sanctions, while distance decreases the relative efficiency of the use of military force. The key foreign policy decision facing any state, then, involves determining how to allocate its resources across the various instruments of policy in order to optimize the balance of maintenance and change produced.

What We Learn from the Two-Good Theory

The assumptions outlined in the previous section suggest that foreign policy decisions are driven by three factors. First, the amount of resources available is a key factor affecting a state’s foreign behavior. States with greater capabilities can engage in more of all types of behaviors than can states with fewer resources. Second, states’ preferences regarding the trade-offs they are willing to make across maintenance and change will influence the proportion of their resources they devote to each of the two goods. Third, the relative efficiency with which various policies translate resources into the two goods will influence how resources are allocated across specific policy instruments. The formal, mathematical model provides much greater precision regarding how these factors interrelate to produce empirical explanations and hypotheses (Palmer & Morgan, 2006, chapter 5). Here a general flavor of the conclusions is provided.

First, the two-good theory leads us to expect that the greater a state’s resources, the more it will engage in all types of foreign policy behaviors. That is, more capable states will, on average, provide more foreign aid, impose more sanctions, join more alliances, have higher levels of defense spending, and use military force more often than will less capable states. This conclusion may seem intuitively obvious, but note that it flies directly in the face of one widely held piece of conventional wisdom—that a country realizes peace through strength. The two-good theory leads us to expect that the more capable (stronger) a state is, the more military conflict it will initiate. Moreover, the theory leads us to expect (actually by assumption) that as capabilities increase, the use of some instruments of policy better at producing maintenance (e.g., alliance with stronger states, defense spending, dispute reciprocation) will increase at a decreasing rate while the use of other behaviors better for producing change (e.g., alliance with weaker states, foreign aid, dispute initiation, economic sanctions) will increase at an increasing rate.

Following the logic of this expectation does lead to some other less obvious and non- and counter-intuitive hypotheses. One example involves how states respond to military provocations. Intuition suggests that when one state initiates a militarized dispute against another, the likelihood that the latter responds in kind increases as that targeted state increases in strength. The two-good theory suggests a more complicated relationship. Because reciprocating the use of force is viewed as a maintenance-seeking action and because the theory suggests that the use of maintenance-seeking policies increases with capabilities but at a decreasing rate, we are led to hypothesize that a strong state that is getting stronger is actually less likely to reciprocate the use of military force than is a weak state that is getting stronger. The basis for this expectation is that, on one hand, a strong state on the rise can ignore provocations and will see its resources better directed toward altering other aspects of the status quo; a weak state, on the other hand, is better served by devoting any new resources to fending off challenges. This is certainly not an intuitive hypothesis, nor is it one that to our knowledge has been investigated, yet the empirical evidence supports it (see Morgan & Palmer, 1997 and Palmer & Morgan, 2006).

It is more interesting that the two-good theory explains how the uses of various policy instruments are interrelated. The most important contribution in this regard rests with the theory’s clarification of foreign policy substitutability relationships. The notion of foreign policy substitutability was introduced by Most and Starr (1984, 1987). They noted that multiple instruments of foreign policy can be used to address a particular problem and that any given policy tool can be used for multiple purposes. One example they provide is to note that a state facing a hostile enemy can enhance its security either by forming alliances or by building its own arms. Most and Starr argue that this can explain why many of our empirical studies produce weak results. If we seek to associate one cause with one effect, we are doomed to fail. Following Most and Starr, a number of studies have sought to find evidence that substitutability relationships exist, but they generally have found that the use of policies believed to be substitutes (like arms and alliances) actually increase and decrease in tandem (see, e.g., Diehl, 1994).

These empirical results are consistent with the two-good theory, because the theory leads to the expectation that the frequency of use of all tools is closely associated with resource endowments. The two-good theory goes farther, however, and suggests that the use of any two policies that primarily produce the same good should be inversely related only when we control for resources and preferences. Thus, for example, defense expenditures and forming alliances with stronger countries—two maintenance-producing policies—can be positively associated in general but, more critically, should be inversely related when we hold resources and preferences constant. Palmer and Morgan (2006) provide evidence supporting expected substitutability relationships for a number of pairs of policies while controlling for a country’s level of resources. This suggests that substitutability relationships are more complex than is implied by Most and Starr. Investigating the implications of these complex substitutability relationships has led to a number of other non- and counter-intuitive results.

One such example addresses the issue of burden sharing in alliances. It is well known that many in the United States criticize NATO allies and Japan for failing to bear their fair share of the economic burden for the common defense. The basis of this criticism rests with the observation that the United States devotes a larger portion of its GDP to defense expenditures than do its allies. At the same time, U.S. allies frequently complain about the fact that the United States does not devote as high a percentage of its GDP to foreign aid as do they. The two-good theory suggests that these observations should not just be expected as a result of the alliances; they, in fact, represent a significant portion of the benefit of the alliances. The argument is as follows: When we observe the formation of an alliance, it must be the case that the alliance has become more efficient at translating resources into foreign policy goods than whatever the signatories were doing previously. Weaker allies gain maintenance from the alliance, since having a stronger ally significantly increases their ability to protect their interests. The strongest state in an alliance gains change because the weaker states add little to its ability to provide defense and because most such alliances contain provisions that give the strong allies something in return (like trade concessions, access to ports, or voting support in an international organization). Therefore, to stay at the preferred mix of maintenance and change, after joining an alliance with a more powerful state, weak allies move to shift resources away from maintenance-producing policies, like allocating high levels of resources to military expenditures and toward change-producing policies, increasing expenditures on foreign aid. Strong allies must do the opposite: having acquired change through the alliance, they reallocate resources in their foreign policy portfolios toward the production of maintenance This is how the alliance makes all members better off—if the alliance were not to do that, it would not form or remain in existence. This is exactly what we observe the NATO allies doing. Note, too, that this leads us to expect that NATO would continue to exist and even expand after the end of the cold war. Most traditional theories, which presume that an alliance needs a specific threat to exist, have a hard time accounting for the fact that NATO has persisted. Evidence suggesting that the pattern of alliance behavior predicted by the two-good theory is common can be found in Morgan and Palmer (2003) and Palmer and Morgan (2006), as well as in Leeds, Long, and Mitchell (2000).

Finally, the two-good theory explains a number of relationships between the actions one state directs at another and how that target state reacts. From the perspective of this theory, one state can influence another’s behavior only by affecting one of the factors that affect policy choices—preferences, resources, or the relative efficiency of policy tools. For example, the form in which foreign aid is provided affects how the recipient state alters its behavior as a result of the aid. Cash transfers, on one hand, simply increase the resources available to the recipient and will result in an increase in all types of behavior. Technical assistance, on the other hand, increase the relative efficiency of specifically targeted policies and will lead the recipient to devote more of its resources to those policies while reducing resources spent on some other policy. This finding led Morgan and Kobayashi (2013) to hypothesize that development aid would produce an increase in dispute initiations by recipients while military technical assistance aimed at enhancing defensive capabilities would lead recipients to reduce dispute initiations. They found evidence supporting both hypotheses. In this regard, the two-good theory provides a means for identifying unexpected consequences of foreign policy actions.

In another such study, Martinez Machain and Morgan (2013) examined how mutually agreed-upon troop deployments by one state to another affect the policy decisions of the host state. Assuming that troop deployments essentially serve to provide maintenance for the host country, the theory leads us to expect that host countries will shift resources away from other maintenance-seeking policies and put them toward change-seeking policies. Thus, host countries are expected to decrease their own troop levels and increase the frequency with which they initiate disputes. Furthermore, the extent to which the host makes these changes should be associated with the number of troops deployed from the sending state. Again, the evidence supports these hypotheses. Note that each of these expectations, taken singly, is intuitive. It makes sense, for example, that because of U.S. troops stationed in South Korea, the South Koreans need fewer of their own troops to provide for their defense. It also seems reasonable that a state hosting troops from an ally could use those troops as a shield behind which to pursue its own aggressive policies. Taken together, however, these expectations are counter-intuitive—it makes little sense for a state that is about to initiate a militarized dispute to precede that with a reduction in its troop levels, yet that is the general pattern we observe. The two-good theory reconciles, explains, and predicts those behaviors.

Evaluating the Two-Good Theory

How one evaluates such a theory depends on the criteria one adopts. One common approach to evaluating theory is to debate whether the assumptions on which it is based are realistic. One may argue, for example, that because it essentially treats all foreign policy as involving a single decision regarding how to allocate resources over various policy instruments, the two-good theory falls short. Its assumption on this point is not descriptively accurate. The criticism of the two-good theory, or any theory, based on that approach, however, is wrong-headed. If we know how some aspect of the world works, we do not need theory. If we need theory to understand some phenomena, we have to accept that theory involves abstraction, so all theory requires assumptions that could be deemed unrealistic. Another common approach to evaluating theory is to take that theory on its own terms and ask whether (a) it is logically consistent and (b) whether the hypotheses it produces are supported by empirical evidence. By these criteria, the two-good theory does quite well, as we have seen in this discussion. These criteria are not sufficient, however. A theory could meet these criteria well and yet be almost totally useless in the sense of producing much usable knowledge or in the sense of helping us understand the world.

The approach to evaluation taken here is somewhat different. It must be accepted that in developing or accepting a theory, one must always pay a price. To gain greater simplicity, tractability, or generality, one must usually abstract farther from reality. To answer one set of questions, one must often forgo even asking others. To accept some perspectives and insights, one must often abandon others, even if those others are long held and cherished. We must evaluate the balance between what is gained and what is lost when adopting the two-good theory.

One important point to emphasize in this regard is that the two-good theory is a deductive theory. It begins with the specification of underlying assumptions and derives, rigorously and systematically, empirical hypotheses that follow from these assumptions. It does not begin by presupposing a set of empirical relationships, trying to reconcile and explain those findings. There are advantages and disadvantages to this approach. First, a deductive approach can lead to a novel way of looking at the world. In this case, for example, foreign policy decisions involve decisions about how to allocate resources across instruments of policy, not whether or not to use a particular instrument in a particular case. Whether that is an advantage or a disadvantage depends entirely on whether that new way of looking at the world helps us gain insight or just confuses us. One of the primary ways of settling this question is to consider whether the theory leads us to look for empirical relationships that we never would have considered without it. On this score, the two-good theory does remarkably well. A major advantage of the deductive approach is that it provides some assurance that our explanations for a variety of observed relationships are not ad hoc or internally contradictory. The price one pays for this is that there may be observed empirical regularities that the theory just cannot explain. We must always accept that some (perhaps about 5%) observed relationships might be just statistical artifacts. A deductive theory may lead us to relegate some findings that actually deserve explanation to this category. Similarly, a deductive theory tells us exactly what empirical relationships to look for. In other words, a deductive theory does not require the use of any control variables in statistical tests of its implications; indeed, many deductive theorists reject the notion that one should include control variables in a statistical equation that are not included in the theory (Achen, 2005; Ray, 2003). That makes explaining empirical results relatively straightforward, but it means we may miss discovering interesting patterns serendipitously.

Another key aspect of the two-good theory is that it aims explicitly to unify explanations for a variety of phenomena. That is, one theory explains uses of force, defense expenditures, troop deployments, foreign aid allocations, alliance decisions, uses of economic sanctions, and so forth. Possibly, one could devise a better theory of, say, foreign aid allocation if that is all one wants to explain. It is possible that every type of behavior the two-good theory purports to explain is better understood if studied in isolation. If that is true, then there is certainly a trade-off between one theory that explains a lot of things—n in number—moderately well and a set of n theories, each of which explains one thing better. Two important points must be made in this regard. First, where one comes down on this argument depends on what one means by “understand and explain.” If, on one hand, one views explanation as involving something akin to accounting for as much of the statistical variance as possible in a dependent variable, then one is likely to do better focusing on one type of behavior at a time and utilizing as many independent variables as is plausible. On the other hand, if explanation is viewed as developing an understanding of the causal processes through which things are related, then the more inclusive theory will be better. Second, if Most and Starr are right regarding foreign policy substitutability, then it is unlikely that we can actually understand or explain foreign policy decisions by focusing on one type of behavior at a time. Most and Starr’s argument is persuasive, and their conclusion is compelling, so a more fully integrated theory is preferred, even if it means a somewhat weaker statistical fit.

Third, one of the costs that must be paid when a new theory is adopted comes from the risk that insights provided by existing theory may be abandoned. For decades, realism and neorealism have provided the basis from which analysts have explained foreign policy behavior. Even though many have challenged the realist paradigm, the field has continued to rely upon the insights it provides. Perhaps the most significant contribution of realism is the notion that in an anarchic international system, the single most important determinant of a state’s behavior is its power relative to other states. The two-good theory retains that insight, though it makes it more precise by focusing on measurable resources rather than an ambiguous notion of power, and it makes it more rigorous by specifying exactly how resources are translated into action. Moreover, Morgan and Palmer (2000) demonstrate that the mathematical structure of the two-good theory could be generalized to any number of goods and that, if we assume a single good, the model could be viewed as capturing at least one version of realism. They demonstrate that hypotheses following from the one-good version of the model fail to receive empirical support, significantly diminishing realism’s appeal and utility. A second major failing of realism comes from the fact that it cannot address trade-offs across goods that states must make. Along these same lines, the two-good theory draws on insight provided by perspectives that assume that much of global politics are driven by the fact that some states are status quo oriented while others are revisionist, such as power transition theory and long-cycle theory (see Palmer & Morgan, 2006, chapter 2), though the two-good theory refines this insight in two ways. First, the two-good theory is based on the assumption that all states want to change some aspects of the status quo and want to maintain others. Second, it does not assume that the extent to which a state holds these desires is related to state power. In this way, the two-good theory allows for the possibility that the most capable state in the system desires the most change. This avoids the tautology that is inherent in the other theories. Consider power transition theory, for example (Organski & Kugler, 1980). To account for the observation that some transitions result in major wars while others occur peacefully, the theory holds that it depends on whether the rising challenger was satisfied or not, but we have no way of identifying a priori whether a challenger is satisfied until we observe whether a war occurred.

Fourth, one cannot provide explanations for specific events from the perspective of the two-good theory. That is, the theory cannot explain why President John F. Kennedy chose to blockade rather than to attack Cuba during the Cuban Missile Crisis. There is no defense of the theory on this point, and if one’s goal is to be able to address this type of question, one should adopt a different theory. It is not clear that the two-good theory is worse in this regard than is any other theory that attempts to explain general patterns of behavior, but explaining choices to carry out specific actions is beyond the scope of the theory, and, to be sure, these are important questions. This is the price one must pay for any theory that is intended to provide general explanations for a wide range of behaviors conducted by all states in all historical contexts. To gain the ability to make comparisons across a wide variety of contexts with a relatively simple theory, it is necessary to sacrifice some ability to explain the specific choices made in particular cases.

Fifth, and somewhat related to the previous point, the two-good theory has no way of addressing strategic interaction. The model that is at the heart of the two-good theory is decision theoretic. That is, it assumes that state actors solve a maximization problem with no consideration of how other actors are going to behave. Intuition suggests that actors understand that their behaviors interact to determine outcomes and that they should consider and try to adapt to one another’s strategies in both conflictual and cooperative situations. Game theory, the main tool for studying situations of strategic interaction, has certainly made significant contributions to our understanding of international processes. It would seem that it should be possible to integrate the two-good theory and a game theoretic model. The two-good theory explains changes in states’ foreign policy portfolios, which encompass all their behaviors and interactions with all other states, not a specific opponent or partner in isolation. This creates two significant problems for a game theoretic approach. First, since it is based on the assumption that all actions are interrelated, it is unlikely that we could ever find well-behaved and stable game-theoretic equilibria. The fact that the two-good theory assumes that policies are substitutable means that multiple combinations of behaviors produce the same distribution of change and maintenance. It is not at all clear how this could be interpreted in a game-theoretic sense; the model is silent about how other states interpret and respond to one state’s specific foreign policy portfolio. Second, foreign policy portfolios include behavior directed at all other states. It is unlikely that we could make much sense of strategic behavior in such an environment. While some game-theoretic models have allowed for n-person situations, these have focused on coalition-forming behavior directed at a single problem. Two-good theory does sacrifice the ability to consider strategic interaction.

Sixth, the two-good theory largely ignores the role of domestic politics in determining foreign policy. Some work has been done showing how domestic institutions, very abstractly defined, affect the aggregation of individual preferences into the trade-offs a state makes between change and maintenance (Morgan & Palmer, 1998), but otherwise the two-good theory is based on the assumption that states are unitary actors. This ignores the role that competing domestic actors have in selecting among substitutable policies and in implementing policy decisions. It also forces us to ignore the role that factors like audience costs and principal-agent relationships play in influencing foreign policy decisions. State leaders’ personal ambitions and office-holding goals play no role in the theory. Again, abstracting away these factors is a price one pays for the ability to examine how policies interrelate. One could easily imagine that these types of factors would be quite important in selecting among substitutable policies, so this may be key to addressing some important questions that the theory cannot, at present, handle.

To Where from Here?

Given the current development of the two-good theory, what direction should future research take? One obvious answer is to conduct many more empirical tests of hypotheses regarding foreign policy behaviors. The theory is extremely rich and leads to many expectations regarding how foreign policy portfolios change and, more important, how various policy instruments are related. Empirical expectations may be refined along a number of dimensions, however. One area for such refinement would be if empirical studies could test much more precise hypotheses. For example, the theory leads us to expect that as foreign aid allocations increase, expenditures on developing advances in military technology (R & D) decrease, as resources allocated to each of these policies are designed, essentially, to produce change. Evidence exists to support this contention (Palmer & Morgan, 2006), but the theory could actually lead to expectations regarding the precise dollar amounts. Currently, data of sufficient quality to test such precise predictions do not exist. Another refinement would be if the hypotheses tested involved variables that have, to date, been excluded from any analysis. One example is that the two-good theory suggests that one important policy resource is time—for instance, the amount of time a state leader or foreign secretary has to devote to myriad policy problems is finite. The theory tells us that certain factors should be associated with how this time is allocated; for example, as a state increases in its capabilities, chief foreign policy actors should spend an increasing amount of their time on change-seeking behaviors. Again, data appropriate for testing such hypotheses are not presently available.

The two-good theory was developed in a fashion that permits complications and extensions. Two possible extensions may prove particularly useful. First, while the theory is based on the assumption that there are two foreign policy goods, the formal model is fully general with the number of possible goals. “Two” is the smallest number of goods that allows for the consideration of trade-offs, but it would be possible to consider more complex relationships. Care must be taken, however, so that the theory does not become nonfalsifiable. There is a danger that one could explain away bad empirical results by asserting there must be another good unaccounted for. To avoid that, it is necessary to be very specific about what other goods are being included and why. One prominent possibility would be to allow for saving/investment of foreign policy resources. This especially makes sense for instruments like alliances and defense spending, which may be made less to produce an immediate good than to make it possible to meet goals at a time in the future.

The second theoretical extension that may prove useful would be to do more to incorporate the role of domestic politics. Currently, the theory explains very broad patterns in the allocation of foreign policy resources. It is likely that multiple ways of allocating resources across available policies would serve to maximize utility gained from the production of change and maintenance. The precise allocation is likely a function of domestic politics. One area in which the empirical results for the two-good theory’s expectations have been poor has been in considering substitutability between economic sanctions and the use of force. One may conjecture that the problem rests in the fact that a state’s leadership can order the military into action but that it can impose economic sanctions only by inducing firms to comply with sanctions law. This allows nongovernmental actors to affect directly the implementation of policy. The specific design of other policies may be affected by lobbying efforts (e.g., foreign aid that stipulates what goods or suppliers can be purchased). Consideration of these factors may provide a means of refining expectations from the two-good theory and of using it to provide insight into other aspects of foreign policy decisions.

To close, the two-good theory has provided a novel way of examining and understanding foreign policy decisions. It is a rigorous, systematic, deductive model that leads to a large number of hypotheses, many of which have found empirical support. It has produced non- and counter-intuitive findings and offered explanations for a number of things consistent with conventional wisdoms. There are shortcomings with the theory and its application, only some of which may be correctable. On balance, however, it has proved useful and should provide the basis for a great deal of additional research.


Achen, C. (2005). Let’s put garbage-can regressions and garbage-can probits where they belong. Conflict Management and Peace Science, 22, 327–339.Find this resource:

Diehl, P. F. (1994). Substitutes or complements? The effects of alliances on military spending in major power rivalries. International Interactions, 19, 159–176.Find this resource:

Leeds, B. A., Long, A. G., & Mitchell, S. (2000). Reevaluating alliance reliability: Specific threats, specific problems. Journal of Conflict Resolution, 44, 686–699.Find this resource:

Martinez Machain, C., & Morgan, T. C. (2013). The effect of US troop deployment on host states’ foreign policy. Armed Forces and Society, 39(10), 102–123.Find this resource:

Morgan, T. C., & Kobayashi, Y. (2013, June). Just what did you think would happen? Modeling the effects of foreign policy actions. Paper presented at the Jan Tinbergen European Peace Science Conference, Milan, Italy.Find this resource:

Morgan, T. C., & Palmer, G. (1997). A two-good theory of foreign policy: An application to dispute initiation and reciprocation. International Interactions, 22(3), 225–244.Find this resource:

Morgan, T. C., & Palmer, G. (1998). Room to move: Security, proaction and institutions in foreign-policy decisionmaking. In R. Siverson (Ed.), Strategic politicians, institutions, and foreign policy (pp. 193–220). Ann Arbor: University of Michigan Press.Find this resource:

Morgan, T. C., & Palmer, G. (2000). A model of foreign policy substitutability. Journal of Conflict Resolution, 44(1), 11–32.Find this resource:

Morgan, T. C., & Palmer, G. (2003). To protect and to serve: Alliances and foreign policy portfolios. Journal of Conflict Resolution, 47(2), 180–203.Find this resource:

Most, B., & Starr, H. (1984). International relations theory, foreign policy substitutability and “nice” laws. World Politics, 36, 383–406.Find this resource:

Most, B., & Starr, H. (1987). Inquiry, logic, and international politics. Columbia: University of South Carolina Press.Find this resource:

Organski, A. F. K., & Kugler, J. (1980). The war ledger. Chicago: University of Chicago Press.Find this resource:

Palmer, G., & Morgan, T. C. (2006). A theory of foreign policy. Princeton, NJ: Princeton University Press.Find this resource:

Ray, J. L. (2003). Explaining interstate conflict and war: What should we control for? Conflict Management and Peace Science, 20, 1–31.Find this resource: