war on the rocks

The Futures Problem: Why Big Organizations Have Problems Making Long-Term Forecasts and What to Do About It

November 21, 2017

Editor’s Note: This is the lastest installment in our “Next War” series. 

“Hard to see, the Future is.”

–Yoda

It is 1921. Leading military professionals see a new world on the horizon. Italian theorist Giulio Douhet contends:

Would not the sight of a single enemy airplane be enough to induce a formidable panic? Normal life would be unable to continue under constant threat of death and destruction.

Airplanes would rain destruction upon industrial cities reducing the relevance of armies and navies. Yet, airpower rarely proved decisive alone. Even the nuclear strikes on Japan were launched from island staging areas seized by amphibious forces.

Today, advocates of the “third offset” search for new game changers that will assure the United States maintains a military position of advantage in an uncertain future. Large investment decisions loom in Washington despite ballooning budget deficits about where to spend the marginal defense dollar. What capabilities will deter Russian aggression in Europe while shaping long-term competition with China as well as limiting North Korean and Iranian regional threats?

Making big bets about the future against multiple competitors is a problem that confronts both military organizations and large, capital-intensive businesses. Firms like Exxon Mobil have to determine how to allocate resources between renewal energy and lowering the costs of deep water drilling or inland shale gas extraction. In other words, they all face a multiple futures problem. They have to prepare for a range of outcomes to maintain their competitive advantage.

This futures problem compounds overtime. The larger the bet and the deeper into the future you go, the higher the number of black swans and known unknowns ready to rain on your parade. Just as an energy firm must decide on the balance between hydrocarbon and renewal energy investments for 2040 in 2017, the U.S. military must make a prediction about the future character of war today. The question is how can large bureaucracies organize around this futures problem and make big bets about an uncertain future.

The answer lies in constructing an institutional process that connects evolving estimates of alternative futures to defense strategy. For military organizations, the goal is to think more like George Marshall and Albert Wedemeyer than George Patton. They must foresee the formations required to win future campaigns, not just prevail over the enemy in the present. To this end, large organizations should create incubators that connect with diverse external actors and model alternative future horizons. More important than the size is the right mix of personnel, a testament to big changes cultivated by small teams historically in the Office of Net Assessment and Google X. Equally important is the mandate: Connect forecasts to strategy. These teams generate estimates of an uncertain future that assess large trends, prevailing opinions, and possible interaction vectors through techniques ranging from scenario analysis to complex systems modeling. They collect diverse data inputs and make forecasts. As new information emerges or military experiments reveal new theories of victory, the teams update their forecasts. The net result of the analytical process is a frame of reference for crafting a modernization strategy. This strategy articulates investment priorities and trade-offs in relation to likely risks and opportunities. In this manner, a modernization strategy is the blueprint for acquisition and future force generation.

Escaping the Folly of Alchemy

Much of the futures literature and techniques used to produce it is reminiscent of alchemy. The search for certainty, like the quest for gold, crowds out better judgment when thinking about events 20 years away. Bold claims about fourth industrial revolutions, cyber doom, and rising hegemons provide practitioners with a false sense of certainty about still unfolding events. Worse still, these forecasts tend to intersect with inherent individual bias and group dynamics in large organizations under pressure to make bets about the deep future. Practitioners confuse rampant speculation with empirical fact and create a picture of the future prone to euphoria or damnation rather than the mundane middling distribution of events more likely to occur. It is easier in the board room or E-Ring to play on dreams of unproven disruptive technologies, “risk-free” swarm warfare, super tanks, or the fear of artificial intelligence run amuck. This quest for certainty, often biased towards a single technology, limits the ability of large organizations to update their estimates of the future. They forget the beauty of Bayes theorem and power of feedback loops and miss how groups adapt to the introduction of new technologies.

More structured approaches to forecasting the deep future tend to emphasize trends overtime, prevailing opinions, and estimating interactions. Importantly, none of these methods is used in isolation or should overly weight any one variable such as a breakthrough in military technology. Small teams synthesize and evaluate different forecasts of the changing character of conflict to create a running estimate of the future. This set of research hypotheses is updated over time as new information becomes available and actions taken to shape the present change the course of future events.

Approximating Trends

The first input to the running estimate of the future is an inventory of the big structures and large processes that can shape the future. In the military, planners use operational variables to approximate these trends. In economics, researchers use macroeconomic variables. These are used to construct time-series forecasts that look back to look ahead. These regression models estimate major trends overtime and try to use that historical pattern to predict future change. Another technique is to develop scenarios based on predominant trends, a method pioneered by Shell Oil in the 1970s. All of these approaches assume objective, material trends beyond the reach of any one individual shape the future. For example, studies integrate multiple economic and environmental trends, as drivers, to predict future migration patterns and the resulting risk of conflict.

Capturing Prevailing Opinions

Yet, the further you look into the still unfolding future, the more uncertainty and chance for abrupt change. In 1990, few would have imagined cell phones or social media in 2015 as key vectors for recruiting a global network of extremists through slick propaganda. Therefore, forecasters should reject certainty and embrace that there is usually little consensus about trends. Where one study will find a link between climate change, migration, and conflict, another investigation will question the relationship. Therefore, studying trends requires aggregating multiple studies and looking at the distribution of estimates. There are two techniques: Delphi and crowdsourcing. The Delphi method, originally developed by RAND to explore how technology changes war, assumes that experts have a disproportionate share of knowledge about the future. By surveying these sages, forecasters can predict likely futures. Alternatively, crowdsourcing assumes that disparate observations about the future are widely distributed. The larger the crowd, the more likely you are to find converging estimates that will better predict the future.

Mapping Interactions

The next step is to take the range of variables and make explicit bets about how these factors will collide to create alternative futures. From a systems perspective, complex interactions are often interdependent and produce emergent results. Therefore, any trends and estimates of their impact should be modeled as competitive interactions that have the potential to alter the future course of events. This simple, profound truth animates the enduring utility of wargames and processes like net assessment and competitive strategies. Peacetime bets on tanks produce incentives for adversaries to purchase cheaper anti-tank guided missiles. As adversaries allocate more resources to these missiles with their marginal defense dollar they forgo other opportunities and produce incentives to limit the number of tanks produced as costly active protection measures are integrated. Historically, wargames give defense officials the best vehicle to capture this interaction. For example, the Office of Net Assessment has historically funded wargames to explore competitive strategies including working on the maturing revolution in military affairs such as the 20XX study and the dominant maneuver series.

A Culture of Experimentation

The next step involves adapting your running estimate of the future into a series of experiments that help clarify the key components of a modernization strategy. You have to create a culture of experimentation that encourages risk taking and failing early and often as a means of clarifying forecasts about future concepts and capabilities that emerge from the running estimate of the future. Your experiments test your hypotheses about the next war without incurring the risk of fighting it. In a 2003 Center for Naval Analyses study, Brian McCue captures this need for military experiments, stating:

Judgment-based military decision-making works best when it has a strong basis in experience. Almost by definition, there can be no strong basis in real-world experience if the question at hand regards major innovation. Today’s standard military equipment was yesterday’s innovation, and last week’s hare-brained scheme. The tank, the airplane, the radio, and the machinegun were each, in their infancy, decried as useless, and yet today they are deemed essential. The rigid airship, the battlecruiser, and the tank destroyer were supposed to be great ideas, and yet they are now remembered for their disappointing results.

Experiments clarify risk and opportunities over time. For example, interwar naval experimentation, the refinement of War Plan Orange, connected the schoolhouse to the fleet, testing concepts, and capabilities required to achieve a position of advantage in future maritime campaigns. Without these experiments, it is doubtful defense strategists would have seen the importance of carrier aviation.

Too often modern defense experiments are too big to fail, a dynamic captured in Malcolm Gladwell’s portrayal of Millennium Challenge. The larger the wargame or field experiment, the more uncomfortable failure is for the bureaucracy. Therefore, encouraging small, distributed experiments in the operating forces limits risk while encouraging a culture of innovation and prudent risk taking at echelon. Imagine an Army where instead of a massive annual futures exercise like Unified Quest, a mix of junior and senior officers played a series of distributed wargames that let them test new concepts and capabilities. This diverse data provides an opportunity for controlled experiments while making the rank and file soldier part of the process.

Updating Estimates

Because the future is uncertain, any insights from a running estimate of the future and military experiments need to be updated as new events unfold or the enemy adapts. Similar to how major energy companies think about deep futures, you make a forecast, hedge your bets, and update your forecast over time as new information becomes available.

The importance of updating forecasts is a key feature of what Philip Tetlock and Dan Gardner call superforecasters. The idea of using new information to update forecasts dates back to Thomas Bayes and his 1763 treatise, “An Essay Towards Solving a Problem in the Doctrine of Chances.” Since then, multiple thinkers have built on the original idea and placed Bayesian inference at the center of a wide range of forecasting efforts. The defense strategist that does not update their probabilistic forecasts of the future risks blindly proceeding to irrelevance. After the Democratic National Committee hack, assuming all future cyber conflict looks like Stuxnet risks missing how rival states are targeting domestic audiences and political attitudes.

The Strategy Bridge

The running estimate of the future, updated through analysis and experimentation, produces the forecast required to develop a modernization strategy. According to Colin Gray, a strategy “seeks control over an enemy’s political behavior” through bridging “means and ends.” You achieve a position of continuous advantage not just by imposing your will, but shaping the decision horizon of your competitors as well as your partners. Strategy is about shaping the future through prioritizing the mix of instruments of power, to include military concepts and capabilities, required to achieve possible political objectives. It therefore requires estimating national interests and how they will diverge and converge between great powers, as well as emergent fault lines in the international system linked to events as varied as state breakdown, corruption, and online political warfare.

With this hypothetical range of political objectives alongside the running estimate of the future, the defense planner bridges ‘ends’ and ‘means’ to create a modernization strategy. This strategy explicitly states developmental priorities over different time horizons.

In the short-term, there is less uncertainty. Forecasts about the next four years are less uncertain, but never perfect, than the next 40. Modernization investments consist of buying off-the-shelve and already prototyped material. Large investments are not high risk in the short-term even though they can crowd out future investment opportunities. For example, why try to invent military artificial intelligence, when you can leverage significant investments made by firms like Alphabet, Amazon, and Microsoft. When you live in a world where commercial technology is evolving more rapidly than defense research labs, embrace this fact and buy it, don’t build it. Cold War skunk works-type labs certainly gave us stealth aircraft and other marvels, but collections of programmers in Silicon Valley funded by venture capitalists firms are more likely to give us the next breakthrough. Therefore, short-term modernization is as much about understanding the market for innovation as it is about making investments.

In the long-term, there is more uncertainty both about the future political landscape as well as the collection of concepts and capabilities most likely to provide a continuous position of advantage. Modernization investments consist of basic research and making multiple small bets. You are essential buying long derivative contracts that bet on alternative strategic futures. This time horizon requires the defense strategist to think more like a venture capitalist than a Cold Warrior. Investments that fail, and there will be many, should be quickly dropped while successful candidates are scaled up. Long-term modernization requires managing a diverse portfolio that cuts across different types of bets, say long-term precision and next-generation sensors, to prioritize your investment strategy.

You are looking for the connectors, those systems required for multiple technologies to bear fruit. For example, the radio that connected the tank to the aircraft and dismounted infantry and artillery enabled the translation of German kesselschlacht into blitzkrieg. If you were making long-term bets in 1918, in theory you would have to choose between aircraft, mechanized vehicles, artillery improvements, and communications technology. As advocates of tanks and aircraft made apocryphal predictions about future war, investing in the radio would not be immediately obvious. Yet, choosing the radio enables multiple small bets about improving the tank, the airplane, and artillery. Small bets in connected systems can create better returns on investment. The question is what technologies today, like artificial intelligence or autonomy, have that same connective payoff for 2040?

Towards a Flexible Modernization Strategy

Large organizations have to make big bets about an uncertain future. For military organizations, unlike firms, poor choices mean deterrence failures, battlefield defeat, and a loss of strategic position and standing. The creative destruction that animates competitive markets is too risky for great powers. Therefore, large military organizations need structured processes for translating their future forecasts into a clear modernization strategy. The modern military has to articulate risk and opportunity 20 years into the future and a clear concept of how to seize those opportunities and mitigate the risks. Formalizing a process for developing running estimates of the future based on analysis and experimentation, and updating those estimates, provides the insights required to craft a modernization strategy. This strategy articulates what means should be prioritized to achieve a range of strategic ends. Above all, a clear institutional process that connects forecasts with defense strategy provides strategic flexibility. It helps practitioners balance seizing opportunities and mitigating risks during long-term competition.

 

Benjamin Jensen, Ph.D holds a dual appointment at Marine Corps University and American University, School of International Service. He is the author of Forging the Sword: Doctrinal Change in the U.S. Army, 1975-2010 and the Next War series at War on the Rocks.

Neil Hollenbeck is a Fellow with the Army Future Studies Group and an infantry officer with previous assignments in the 82nd Airborne and 3rd Infantry Divisions and on the faculty at West Point. He holds an MBA from Duke University.

The opinions, conclusions and recommendations expressed or implied above are those of the authors and do not reflect the views of their organizations or any entity of the U.S. government.

Image: Africom