Helping Defense Planners Make More Informed Forecasts

Recent debates over future trends in armed conflict and how involved the United States might be in such conflicts have important implications for U.S. defense planners. Will conflicts like those in Syria and in Ukraine become more common? Will we see a resurgence in interstate war? And how is the United States most likely to respond to the range of contingencies that may arise? What size and types of forces is it likely to commit? How will these demands vary under different states of the world? Decisions about posture, training, recruiting, investment, operational planning, and strategic guidance rely on estimating how frequent both armed conflicts and U.S. interventions in them are likely to be, but reliable estimation of these trends is difficult. To assist in this type of estimation, RAND developed a more sophisticated, dynamic forecasting model. The model utilizes historical trends in warfare and U.S. interventions alongside trends in structural factors associated with them to estimate the risk of future armed conflicts and the demands they may place on U.S. military forces.

This tool was completed in February 2019, followed by security review by the sponsor and the Office of the Chief of Public Affairs, with final sign-off in April 2022.

Interactive Tool

With the interactive tool, users can then visualize what affect these assumptions have on the outputs of the model in something other than the baseline scenario. For example, if users want to understand the implications for ground troops in military interventions in such circumstances, they can customize a scenario based on one of the alternative states of the world and see how this changes the baseline scenario after 2025 when those states of the work kick in.

A view from a RAND tool showing a prediction about future armed conflict — If the U.S. keeps current alliances in place, interstate wars are projected to remain stable.

Compare this to an isolationist United States. In this scenario, the United States enters a renewed period of isolationism starting in 2025, potentially in response to continued backlash against globalization, its unequal costs and benefits, and concerns over the growing costs of U.S. forward presence. Tensions with key European allies and with Korea and Japan over differing views of the future of multinational institutions and agreements (including trade agreements) create additional strains. As a result, the United States takes significant steps to reduce its involvement in these institutions. Most significantly, while current interventions continue, the United States abandons participation in key multilateral alliances, such as NATO; bilateral alliances, such as with Korea and Japan; and multilateral trade agreements and organizations, such as the World Trade Organization, effectively eliminating all entanglements with the exception of existing bilateral trade agreements. U.S. policymakers also limit their involvement in and security concerns over areas of the world to those most core to U.S. interests: Europe, East Asia, and the Americas.

This sharp reduction in the U.S. engagement in the international arena has serious implications for international trade, norms, and regional balance of power across the globe. This is reflected in the model's prediction of an increase in new interstate conflicts.

When defense planners consider future armed conflict, they rely on trends. Will conflicts like those in Syria and Ukraine become more common? Will we see a resurgence in interstate war? And what will the most likely response be to the range of contingencies that may arise? U.S. defense planners use these trends to make a number of critical decisions—the size and types of forces the United States is likely to commit; how these demands will vary under different states of the world; and the effect these trends will have on posture, training, recruiting, investment, operational planning, and strategic guidance.

But reliably determining these trends and estimating the implications of them for defense planners entails accurate and credible forecasting (or predicting) of future events—something that countries have had a poor track record of doing. Simple extrapolation from recent trends may widely miss the mark or fail to capture the potential for a more radically different future than what is currently anticipated. In a project for the U.S. Army, RAND conducted an effort to identify potential future demands for U.S. Army forces, including size, location, and capabilities, to inform decisions on planning, posture, and investments. A key part of that project centered on developing a sophisticated, dynamic forecasting model that uses historical trends in warfare and U.S. interventions—alongside trends in structural factors associated with them—to estimate the risk of future armed conflicts and the demands they may place on U.S. military forces out to the year 2040.

Step 1: Model incidence and forecast future conflicts
Step 2: Model incidence and forecast future interventions using future conflicts as input
Step 3: Define force demands using insights from average characteristics of historical interventions
Step 4: Model conflicts, interventions, and force demands under several future states of the world using an iterative forecasting process

Steps 1-3 are iterative. Step 4 follows on after the iterations are from 1-3 are complete.

The model has four steps. It is iterative, going through steps 1–3 to estimate each additional year of data and incorporating each new forecast datapoint into the calculation.

We then introduce the interactive tool we created to allow users to interact directly with the model and explore the implications for trends in conflict and demand for U.S. forces of alternative future scenarios.

Step 1: Future Conflicts

The forecasting model starts by projecting future conflicts for U.S. intervention. This includes deterrence prior to conflict and post-stability operations after one, but the focus in on armed conflict itself—particularly the future incidence of interstate wars and intrastate conflicts. To do so, the model uses historical data to identify the factors most significantly associated with the occurrence and duration of these types of armed conflict. At the nation level, this include states’ political and economic institutions, their level of development, and their demographic factors. The model also includes factors at the international level, such as the relative military capabilities between regional and major powers, economic interdependence, the presence of territorial disputes, and the prevalence of democratic regimes. In short, the model learns from historical patterns of conflict to forecast future conflicts, using projected values of these key factors domestic and international factors. The intrastate conflicts graph shows an example of the output from the model in this step. It shows both the historical intrastate conflict data in the dark bars on the left and the model’s projections of future conflicts on the right. The darker line shows an average projection, while the shaded area indicates a range of plausible projections.

Step 2: Future U.S. Military Interventions

Given the set of conflicts identified in Step 1, the model then projects likely future U.S. military interventions, assessing the potential for three different types of U.S. ground interventions: armed conflict interventions (that occur in ongoing conflicts), stabilization interventions (in post-conflict environments), or deterrence interventions (that occur in states currently at war, but where the United States may be concerned about the risk of conflict). To produce these projections, the model identifies key factors historically associated with U.S. interventions in these different contexts, ranging from the nature of the U.S. relationship with the host country, to characteristics of the conflict, to structural political and economic conditions. The model then evaluates the historical relationships between these factors and incidences of U.S. intervention, projects future trends in these factors, and then uses the combination of the two to project the likelihood of future U.S. interventions. As with the previous graph, the one shows the historical incidence of U.S. military interventions on the left and a projection of future interventions (and a range of plausible projections) on the right.

A view from a RAND tool showing a prediction about future U.S. military interventions — U.S. military interventions

Step 3: Future Force Demands

Using output from the second step on each projected intervention, the model then estimates force demands—the size and type of U.S. ground forces involved. The model does this by first identifying how the size and type of U.S. forces have historically varied across interventions with different characteristics, including the location where they occur, the conflict that prompted the intervention, and the relationship between the United States and the host country. Based on this model output, the RAND research team then identified different intervention categories with “typical” sizes and force mixes based on these characteristics, placed projected future interventions into these categories based on projections of their characteristics, and assigned these typical size and force mixes to the projected interventions. Like the previous two graphs, this one shows historical numbers on the left and the projected numbers on the right.

A view from a RAND tool showing a prediction about the number of ground troops required for future U.S. military interventions — U.S. ground troops involved in military interventions

Step 4: Different States of the World

In the last step, the model conducts each of the prior steps based on specified scenarios for different states of the world. These scenarios include a baseline—"no surprises"—future in which factor trends proceed as currently anticipated out to 2040. However, surprises do occur, and the strategic environment used in the model can be changed, allowing users to explore the implications of such different strategic environments, where certain trends or events may diverge more radically from current expectations.

These different strategic environments can be defined in a number of different ways, including divergent political, economic, and military trends and events; that is, they can be defined by any variable that affects the model's projections of conflicts or interventions, as described above. The report that summarizes the model results includes four illustrative worst-case scenarios, ranging from a global pandemic scenario that sharply affects economic activity and populations to a sharply revisionist China scenario that involves a fracturing of the international order into U.S. and China-led blocs.

Beyond the illustrative scenarios in our report described above, the forecasting model is intended to be used to explore the implications of a wide and flexible range of future strategic environments. To facilitate this use, RAND has developed an interactive tool with an easy-to-use interface that allows users direct access to some aspects of the forecasting model. As such, users can design their own strategic environments and then explore the projected trends in conflict, U.S. intervention, and demand for U.S. troops that result—at both the global and regional levels. The tool allows users to vary assumptions about future U.S. policy choices and assumptions about the future trajectory of the international order (customize them).

Variable	Baseline	Alternative Future Options
U.S. Military Size and Spending	A slight decline in both U.S. military size and spending by 2040	Range from substantial increases, particularly in spending, to large decreases (20–25%) by 2040
U.S. Military Force Posture	U.S. force posture in 2018 remains in place	Include moderate increases in troops deployed to Europe, Asia, or the Middle East, or combinations thereof, or a complete withdrawal of all U.S. forces from other countries.
Extent of U.S. Alliances	U.S. alliances in 2018 continue to 2040	Range from large expansion of U.S. security commitments to encompass countries such as Sweden and Georgia to sharp reduction in U.S. alliances, including withdrawal from NATO.
Importance of International Trade	International trade continues to increase modestly	Include stark changes, such as U.S. entry into multilateral free trade agreements with European and Asian states or breakup of the current WTO-centered system into four competing regional trading blocs
Prevalence of Democracy	Continued modest increase in number of states becoming established democracies	Include potential for substantial increases in democratization in sub-Saharan Africa and Southeast Asia or rise in authoritarian regimes in key countries such as Brazil, Nigeria, and Poland

The ability to select options for these variables—and to create combinations of these options—allows users to clearly specify their assumptions and examine how projections for conflict and demands on U.S. forces change in response. Users can change assumptions about, for example, U.S. force posture, trends in global trade, or transitions in regime type and then compare multiple, alternative scenarios with substantive differences.

A view from a RAND tool predicting a rise in future U.S. military interventions if the U.S. pulls back from current security commitments — U.S. ground troops involved in military interventions with custom scenario