Economic Games Behind Nuclear Deterrence

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

Practical [military] men, who believe themselves to be quite exempt from any intellectual influences, are usually slaves of some defunct economist1 DO NUCLEAR WEAPONS MAKE US SAFER? DOES MAD WORK? The U.S. President wants a world free of nuclear weapons. The U. N. Security Council, on September 24, 2009 adopted a resolution, unanimously approved by heads of all countries present, that said working toward a nuclear free world was a worthy goal. The Russians most recently published National Defense strategy has all but eliminated nuclear deterrence from its strategic posture. Yet, in the U.S., the Pentagon has recently completed a draft Nuclear Posture Review that claims thousands of nuclear weapons are still necessary for extended deterrence. Nuclear hawks and nuclear optimists abound, proclaiming that nukes are still needed. They claim that nuclear deterrence makes the world a safer place. Who is right? Who is wrong in this debate? Is nuclear deterrence still necessary in the world today? Does nuclear deterrence really work? In the brief that follows, we discuss nuclear deterrence as a major component of national defense strategy from its foundational roots in economic game theory and rational choice theory. From this vantage, nuclear deterrence as a major component of U.S. national defense strategy leaves a lot to be desired. THE UTILITY OF NUCLEAR DETERRENCE During 2008, the nations of the world spent nearly $1,500 billion (U.S.) on their military forces for the purported purpose of national defense.2 In the past 64-years, since the end of WWII, the total spent on national defense globally is around $60,000 billion. One consequence of this massive, ongoing diversion of global resources (human, economic, scientific and technological capital) from meeting basic human needs is the continued immiseration of many billions of the earth’s human population and the the neglect of pressing critical global environmental, social, and economic issues. As President Dwight Eisenhower stated in 1953: Every gun that is made, every warship launched, every rocket fired signifies, in the final sense, a theft from those who hunger and are not fed, those who are cold and not clothed. This world in arms is not spending money alone. It is spending the sweat of its laborers, the genius of its scientists, the hopes of its children. The most obvious weakness to allocating so much of the world’s capital to national military preparedness is that it often fails to protect nations from the war and destruction it is supposed to prevent. Mutual Assured Destruction (MAD) and its variants such as Massive Retaliation and Minimal Deterrence 3 are all based on economic game theory developed by John von

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 1 of 10

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

Neumann (1903-1957) and Oskar Morgenstern (1902-1977) in their 1944 work on Subjective Expected Utility theory (SEU) in Theory of Games and Economic Behavior. Sometime in the early 1950s after the first Teller-Ulam design thermonuclear weapon was invented (“the H-bomb”), von Neuman proposed a First Strike on the Soviet Union using thermonuclear weapons. He advocated this because his economic game theory calculations indicated that this move was the best strategy. Although many in the military agreed with von Neumann that a First Strike was necessary, President Dwight Eisenhower (1890-1969), intimately familiar with the uncertainties of war, rejected this strategy and asked for an alternative strategy. Herman Kahn (1922-1983), a RAND analyst at the time, came up with an alternative strategy - the game of nuclear deterrence. 4 Kahn’s version of nuclear deterrence was called Mutual Assured Destruction (MAD). Today, nuclear deterrence comes in many flavors. MAD, Massive Retaliation and Minimal Deterrence are all economic game strategies of deterrence through counter-force.5 As the force being used is nuclear weapons, these games are collectively described as a game of nuclear deterrence. It was originally designed as a two-party game, to be played by the U.S. and the U.S.S.R. The assumption was that if both sides in a two-party game have adequate deterrence, then neither side will strike first if by launching an attack, the attacker would be destroyed in a counterattack. This game has three assumptions: (1) approximate parity so that neither side has an advantage for First Use, (2) both sides have enough information that they can make informed, rational decisions as the game progresses, and (3) the players are rational. Nuclear optimists add another assumption: (4) the game can be played indefinitely through time. Provided these assumptions hold, the MAD version of nuclear deterrence was thought to place the parties in a Nash Equilibrium where neither party would rationally choose First Use of nuclear weapons. 6 The world would be safe from nuclear war. All that was required is that both sides possessed nuclear weapons, could survive an initial attack, and could counterattack. Nukes make the world safer. Defense Secretary Robert McNamara (1916-2009) first described ‘Assured Destruction’ to the American public in the early 1960s. A foundational premise of von Neumann and Morgenstern’s neoclassical approach to economic game theory is the assumption that the decision maker either uses or acts ‘as if’ he was able to use, a specific, unique probability distribution. This is the rational choice theory.7 In rational choice theory 'rationality' simply means that a person reasons before taking an action. The framework for ‘reason’ that a person uses is typically utility as measured by costs against benefits before taking any action.

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 2 of 10

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

ASSUMPTIONS OF THE RATIONAL CHOICE THEORY The rational choice theory assumes a number of things about reality in order for the theory to hold true and for the calculations of probabilistic decision states (IF that occurs, THEN this will happen, WITH this probability). The first assumption of the rational choice theory is that the system being analyzed is linear. That is, the system is not complex and does not exhibit emergence. 8 This simplifying assumption to real life enables the analyst to calculate a probabilistic forecast (normal distribution) of a optimal decision state in all instances. The second assumption of the rational choice theory is that the system being analyzed is subject to stationarity. That is, past history of the system can be used as data input to develop a model for how the system may behave in the future. Again, this greatly simplifies the process for calculating forecasts of future system states. The third assumption of the rational choice theory is a philosophical premise about reality that stems from Enlightenment thinking. There are two parts to this assumption: (a) reality can be adequately modeled rationally with the tools of mathematics and analysis; and (b) since humans will behave rationally in any economic situation, this rational behavior can therefore be modeled. Thus, there are built-in assumptions about the knowability of human behavior, how humans will make decisions in times that are both ‘normal’ and in times of crisis, and that limits of the Real are ‘defined’ by what is ‘rational’ (i.e. knowable through the tools of mathematics and science). For example, the use of backward induction in game theory requires that all future play will be rational (i.e. subject to economically calculated utilities). 9 The fourth assumption of rational choice theory is a belief in certius paribus. That is the system being analyzed can be ‘walled-off’ from other, interacting systems and the decision-space adequately described, all things being equal when , of course, everyone knows that this is merely a simplifying assumption to make the analysis easier. RATIONAL CHOICE THEORY ASSUMPTIONS NO LONGER VALID Today, we are aware that all four of these assumptions of the rational choice theory are incomplete or wrong. Economic and strategic systems are nonlinear, complex systems that exhibit emergence. The strategic situation being analyzed is just as likely to display nonergodicity and non-stationarity as to be ergodic10 and exhibit stationarity. Recent work in psychology suggests that humans do not always behave with rationality. Thus, new insights from behavioral economics question the premise of rational choice theory. Maybe most importantly, recent work in physics and mathematics has proven that all decision-states may not be modeled and that the Real may not be fully described or known via analysis and mathematics. Certius paribus is frequently used in arguments and typically is of little concern when the consequences of the decision

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 3 of 10

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

are small. But, when the consequences of a decision involve the future of life on the planet, the use of this simplifying analytical technique is mere habituated intellectual deceit; hubris of the most pernicious nature. UNCERTAINT Y, PROBABILITIES & A CALCULUS FOR DECIS IONS Actually, it was John Maynard Keynes (1883-1946) in his work, A Treatise on Probability (1921) that offers maybe the first non-additive, non-linear decision theory. He defines uncertainty as U=f(w) {0,1}. U is the value of uncertainty. w is the weight of the evidence. As w increases U decreases. When w=o are situations of ignorance or complete uncertainty. When w=1, a point estimate value for U probability may be calculated. Otherwise for any values between 0 and 1, an interval value of imprecise probability represents U. The interval {0,1} is sometimes referred to as p (rho) which measures the degree of confidence (ambiguity) in the decision maker’s information base. Keynes argued that uncertainty must refer to a state of the world where 0<w<1, not that the world is adequately described by w=1 (certainty) or w=0 (uncertainty). The offshoot of this way of thinking about uncertainty and the calculation of probabilities is that gaussian normal distribution curves rarely accurately represent probabilities in real life. Neoclassical economics and game theory is based on the assumption that a normal distribution fits the time series data best. Normality prevails. Instead, Keynes and Benoit Mandelbrot (b.1924) argued that there are many situations where w=0 and a Cauchy-Lorenz distribution more accurately describes reality where the mean (expected value) and variance (standard deviation) are infinite (undefinable). 11 No one can calculate the probabilities of these outcomes with any degree of certainty.12 Neoclassical economists and decision theorists often assume w=1 (a point value for U may be derived) and that Gaussian distribution curves apply. Only if w=1 can one specific probability distribution apply to the decision-space being modeled; i.e the rational actor theory. Keynes, along with Mandelbrot, and more recently Nassim Nicholas Taleb (b. 1960), argue that in real life w~0; the weight of the evidence in real situations is often 0 or very close to 0 in many areas of decision making. That is, many real life systems do not exhibit finite distributions and no expected precise estimate of probability may be attached to a decision about this system with any degree of confidence. This does not indicate that certain outcomes will not occur (are uncertain), only that their occurrence is unpredictable. It also suggests that a set of possible outcomes for a system may never be fully complete. That is, the system may exhibit results that are unknowable to some degree with any amount of analysis, no matter how much money is spent to gather evidence w. Whether for engineering or budgetary purposes, that we fudge our models to enable decision-making does not belie the fact that the future behavior of many of these systems (in post-modern science, the decision maker is

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 4 of 10

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

always part of the system being analyzed) is unknowable with any degree of certainty. THE INCOMPUTABILITY OF PROBABILITIES FOR RARE EVENTS Taleb makes many of these same points in his Fooled By Randomness (2001) and The Black Swan (2007) where he outlines Mendelbroit’s proofs of the incomputability of the probability for consequential rare events from empirical observations ("Black Swans"). Also, from empirical studies of risk in a time series, it appears that in the process of achieving generalizations from this data and/or deriving general rules from particular observations, hidden properties in the data are routinely missed. Decision-makers end-up overestimating the value of rational explanations of past data, and underestimate the prevalence of unexplainable randomness in the data. The result is managing systems as if a black swan will never occur, even though they almost always do. Taleb believes decision-makers ignore black swans because humans are more comfortable seeing reality as something structured, ordinary, and comprehensible (i.e. rationally explainable). Taleb calls this blindness to the Real the Platonic fallacy and argues that it leads to three explanatory distortions when developing models of reality in time: (a) narrative fallacy: using retrospective historicity to ‘explain’ the past. That is, the past occurred this way because of x, y, and z; (b) ludic fallacy: modern probability theory, utility theory, rational choice theory, and game theory that assumes a normal Gaussian distribution probability curve mistakes simple models of reality with the Real; (c) statistical regress fallacy: believing that the structure of the probability of x occurring that actually exists in reality can be fully developed and described from a set of data. Taleb also believes that people are subject to the triplet of opacity, through which historical descriptions of reality is distilled even as current events are incomprehensible. The triplet of opacity consists of (a) an hubristic illusion of understanding of current events; (b) a retrospective distortion of historical events to ‘fit’ current socially acceptable ways of describing reality; (c) an overestimation of what constitutes ‘factual information,’ combined with an overvaluing of the value of ‘expert knowledge’ (typically rendered by ‘experts’ possessing certain credentials, experience, or notoriety)) concerning the subject being discussed. These philosophical musings, based on solid mathematical and recent empirical studies, describe some common conceptual hicups for game theorists where theory wanders down alleys of infinite regress or traps decision-makers in an intractable loop that folds back on itself with self-referentiality.

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 5 of 10

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

STATIONARITY FAILS FOR UTILITY OF NUCLEAR DETERRENCE “Stationarity in a random process implies that its statistical characteristics do not change with time. Put another way, if one were to observe a stationary random process at some time (t) it would be impossible to distinguish the statistical characteristics at that time from those at some other time (t′).” Generally, the assumption of stationarity is helpful in modeling real-world problems as it enables the use of linear mathematical techniques to approximate the potential behavior of complex systems. However, all complex systems exhibit nonlinear processes and emergence (results may not be predicted from the historical, chaotic systems processes). Thus, in the real world, models based on stationarity (which almost all models are) may not necessarily be used with a high degree of confidence to predict future system states, especially if such system states are projected to occur over significant timeframes (from the perspective of the system being modeled). Where the rubber hits the road with stationarity is the growing realization that due to physical changes in the environment being modeled, the assumption of stationarity itself is invalid. That is, for certain decision-sates we are living in a non-stationarity world today. "This is something completely new -- to make decisions not on facts or statistics about the past, but on the probabilities for the future."13 BEHAVIORAL ECONOMICS QUESTIONS BELIEF IN RATIONALITY Behavioral economics advances the discussion concerning rational choice theory by discussing the theory’s limitations when psychological principles of individual behavior are taken into account. For example, behavioral economic theory looks at decision-making in light of: prospect theory (generalized expected utility of decisionmaking under uncertainty), bounded rationality (satisfaction vs. maximized utility), overconfidence, projection bias, effects of limited attention, time-inconsistent choice (behavior not based on expected utility, but on previous historical reinforcement experiences), hyperbolic discounting (changing discount rates based on length of forecasting period), fairness, reciprocal altruism, etc. UNCERTAINT Y PRINCIPLE VS. GAME INFORMATION SETS Whereas modernity assumes that rational human decisions can always reveal the route to solving problems, post-modernity has doubts concerning rationality as a means for solving all problems. Under post-modernity human rationality is suspect based on work by Freud in psychology (about the strength of unconscious motivations); neurobiology, in our ability to even apprehend reality (neurobiologists suggest that as much as 90% of stimuli reaching the human body remains unconscious [below conscious detection]); and work in quantum physics that calls-to-question some of humanity’s most cherished beliefs regarding descriptions of space and time and causality, especially work on nonlocality and other aspects of the quantum world indicate that science “is far from yielding an assured access to” all of reality. 14 What this

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 6 of 10

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

new understanding of physical reality suggests is that for certain complex systems, game information sets may be lacking crucial data to accurately represent probability distributions of expected outcomes. IS NUCLEAR DETERRENCE AN UNWINNABLE ECONOMIC GAME? What we have seen is that the assumption of rational choice theory being operational very near to the decision to employ nuclear weapons and during the conflict within which nuclear weapons are used is highly unlikely. Instead, a more likely scenario its that moves and counter-moves are counter-productive in that are more likely to intensify reasons for First Use and increate the tempo, rather than slow the game down or reduce First Use nuclear attack. This spiral-effect means that more people on both sides will loose their lives in each successive move and more property damage will occur. Using nuclear weapons adds potential the longer-term costs of ecological services loss. However, it is highly unlikely that through moves of violence and counter-violence that enough individuals on each opposing side can be killed or enough property damage can be accomplished even with the use of limited nuclear strikes to cause the spiral of violence to cease. Some opponents will continue to fight back well beyond the point where all infrastructure is destroyed. The game of nuclear deterrence feeds on violence. The strength of a game intensifies. Violent moves and counter-moves tend to produce a self-perpetuating game. A game that is unwinnable, if nuclear weapons are used. IF THE DETERRENCE GAME IS UNWINNABLE, WHY PLAY? We play because it is the only game we know. We play because the game is familiar. We play because our institutions are set-up to play this game and not another. We play because there are short-term economic drivers for continuing to play the game as originally envisioned. We play because nuclear deterrence is a foundational component of a much larger game - the game of national defense - a very large strategic game. 15 This game primarily prepares for war and engages in fighting wars through the application of deadly force. This game is founded on a premise that nuclear and conventional weapons provide deterrence to war. This game is a non zero sum game. 16 The game as presently played, limited to a capacity for counter-violence, an unwinnable game. 17 Nuclear deterrence is a chimera that results in the escalation of the manufacture of ever more deadly conventional weaponry to avoid the use of nuclear weapons. Instead of limiting the risk for global war or the use of nuclear weapons, this strategic game enables war to be used for ever more marginal conflictual situations, intensify the trade in armaments, and create large opportunity costs as scarce capital is diverted from solving real economic problems to fighting limited wars. 18 The overall

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 7 of 10

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

riskiness of this strategic approach is that, over time, the prospects of global war are heightened, not lessened. If that is the case, the continued care and feeding of doomsday machines a form of potential assisted suicide. Playing this unwinnable game must stop. THE ONLY WAR TO WIN AN UNWINABLE GAME IS TO NOT PLAY Some recommendations to get out of this perpetual war-making machine, which today, has become a Doomsday Machine,19 will require: deciding not to play the game of nuclear deterrence any longer; inventing a new, less dangerous game to play (hopefully one that does not require nuclear weapons). For example, is deterrence even a valid notion today; expanding the strategy of national defense beyond counter-force options; reallocating capital away from counter-force toward an expanded concept of national defense; and accomplishing all of this without crashing the economy as military Keynsianism is dismantled and jobs and technological innovation is siphoned from military hardware and adventurism toward some other national defense or national security requirement. ENDNOTES

John Maynard Keynes, quoted in Justin Fox, The Myth of the Rational Market: A History of Risk, Reward, and Delusion on Wall Street (New York: HarperCollinsPublishers, 2009), xvi. 1

Defense budgets in 2008 were: United States - $607 billion, China - $85 billion, France - $66 billion, Britain - $65 billion and Russia - $59 billion. The world total represents an increase of 45% in military-related budgets (in constant dollars) over the past 10-years (Stockholm International Peace Research Institute). 2

Massive Retaliation assumes a nuclear war can be fought that is not a Total War and that one side could ‘win’ before both sides were totally destroyed as anticipated by MAD. Limited Deterrence assumes nuclear weapons can be used surgically to destroy just enough of an opponent that they give up their aggression. MAD is still the pure and final form of nuclear deterrence. 3

Deterrence as a game strategy for national defense, invented after WWII, is based on the premise that by possessing weapons and a force structure capable of inflicting unacceptable damage on an aggressor and making certain that all opponents are aware of this strength, it is possible to achieve peace as opponents will surely make the rational decision not to attack a superior force. Nuclear deterrence is seen as a limit case for deterrence strategy. Mutual Assured Destruction is the extreme case of nuclear deterrence. However, the failure of deterrence for national defense may not be adequately addressed. For example, in von Clausewitz’s notion of Total War, if nuclear weapons exist, it is most likely these weapons will be used, and when used, it is most likely their use will be for mutual assured destruction, not limited use. 4

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 8 of 10

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

These flavors of nuclear deterrence are problematical in that current scientific studies suggest that even limited nuclear exchanges of 100 Hiroshima-sized bombs, would have global implications with significant cooling of the earth's surface and decreased precipitation in many parts of the world that potentially could lead to global famine. See Ira Helfand, MD, “An Assessment of the Extent of Projected Global Famine Resulting From Limited, Regional Nuclear War,” Royal Society of Medicine (2007). 5

This follows from von Neumann’s minimax theorem that as long as the two rational players’ interests are completely opposed, they can settle on a rational course of action going forward in a zero sum game. An equilibrium is forced by an interplay between self interest and mistrust and a strategy can be devised for playing the game where there are no regrets, no matter what each player ultimately choses for game moves. See William Poundstone, Prisoner’s Dilemma (New York: Doubleday, 1992), 97. 6

Rational Choice theory also underlies the myth of rational markets and the efficient market hypothesis, formulated first at the University of Chicago in the 1960s. The hypothesis is that capital markets are ‘rational’ in that they properly price the value of financial instruments at any point of time. Of course, the recent meltdown of financial markets in 2008 has exposed the fallacy of this hypothesis and its lack of empirical backing (Fox, xv). 7

Emergence “refers to ‘the arising of novel and coherent structures, patterns and properties during the process of self-organization in complex systems.’ The common characteristics are: (a) radical novelty (features not previously observed in systems); (b) coherence (meaning integrated wholes that maintain themselves over some period of time); (c) A global or macro ‘level’ (i.e. there is some property of ‘wholeness’); (d) it is the product of a dynamical process (it evolves); and (e) it is ostensive - it can be perceived” (Wikipedia). 8

Backward induction simply means reasoning backwards in time at t=n, from the end of a problem or situation, to determine a sequence of optimal actions at an earlier point in the game e.g. t=n-1. In game theory, the idea is to carry this process backwards in time from the end game until one builds a strategy that encompasses the best action for every possible situation (the information set), at every point of time in the future. 9

The end state of a system is not sensitive to initial conditions i.e. the system does not exhibit chaotic properties. Thus a natural distribution curve is a reasonable way to describe the potential states of the system over time with given probabilities. 10

Cauchy distributions have fat tails that exhibit kurtosis (infrequent but extreme deviations from the normal case). Also, the law of large numbers does not apply (measuring one instance or a million instances will not improve a probability forecast), nor does the central limit theorem (for large samples, the mean is normally distributed i.e. can be expressed by a gaussian distribution). 11

There really is uncertainty in the world concerning the future – uncertainty that cannot be reduced to statistical probabilities. Maybe this is what Secretary of Defense, Donald Rumsfeld, meant by "unknown unknowns." For Keynes, this irreducible uncertainty and uncalculable risk accounted for the instability of market economies and surprise outcomes of some game strategies. Today, surprises are expected in some games due to work in behavioral economics. 12

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 9 of 10

THE

ECONOMIC

GAMES

BEHIND

NUCLEAR

DETERRENCE

See mathematical expression of stationarity at http://cnx.org/content/m11102/latest/; how stationarity affects real-world models P.C.D. Milly, et.al., “Stationarity is Dead: Wither Water Management?,” Science (1 February 2008): Vol. 319. no. 5863, pp. 573 - 574 at: http://www.sciencemag.org/cgi/content/ short/319/5863/573; “History Can No Longer Guide Investors, Farmers: UN at http://www.reuters.com/article/GCA-GreenBusiness /idUSTRE57P22D20090826?rpc=60. 13

Bell’s Theorem proves nonlocality (correlation of our knowledge of the physical characteristics of particles at a distance) in a quantum field. Essentially, two particles, at a distance, can become entangled. Entanglement means that the state of one particle influences the sate of another particle. This phenomena is described by Schrodinger’s equations. Locality is the principle where influences ( “a set of knowable or unknowable parameters that are real numbers”) are not propagated at infinite speed. In Nature, locality is violated. See Bernard d’Espagnat, On Physics and Philosophy (Princeton & Oxford: Princeton University Press, 2006), 57, 64, 175, 436. 14

From the branch of social science and mathematics that describes game theory - the study of strategic decision making in situations involving uncertainty. 15

A zero sum game always has winners and losers as the size of the pie is fixed; one person’s gain is another’s loss. A non zero sum game is a game where either party’s interests are not completely opposed as one player’s optimal strategy may also benefit the opposing player. In this game all may loose and none gain. (Poundstone, 51-2, 97-99). 16

A game state where it is impossible for the player to win the game. The only options are restarting the game or stopping and deciding to play another game with different rules. 17

Even fighting limited wars with conventional weaponry is expensive. E.g. the total projected economic cost of the 2003-present Iraq War is estimated at $3,000 billion. 18

Doomsday Machine, first coined by Herman Kahn in the 1950s was a hypothetical $10 billion “device whose only function is to destroy all of human life.” See Herman Kahn, On Thermonuclear War (Transactional Publishers, 2007), 144. Assuming that the enemy knew you possessed a doomsday machine, under the game of MAD, this should serve as a deterrence to attack. To the best of our knowledge, the U.S. never set out to built a doomsday machine, and Kahn suggested it was a dumb idea to do so. However, in the 1980s, the Soviets built a doomsday machine, Perimeter (Mertvaya Ruka, Dead Hand), in response to President Reagan’s proposed Strategic Defense Initiative (SDI, Star Wars) missile defense shield. 19

The Soviets interpreted SDI as a First Strike system. Perimeter was designed to launch a nationkilling attack on the U.S. automatically in response to sensor readings of radiation and impacts on Soviet soil from U.S. warheads. First operational in 1985, the system is still live according to some knowledgeable military sources. The Soviets kept Perimeter a secret from the U.S. and did not consider it a deterrent to a First Strike by the U.S. Instead, Perimeter was thought of as a fail-safe system for their own military to relax the need to launch-on-warning as even if Central Command was wiped out, Perimeter would take over. Thus, survivability of command structures were no longer important to retaliate. See Nicholas Thompson, “Inside the Apocalyptic Soviet Doomsday Machine,” Wired Magazine 17.10 at http://www.wired.com/print/politics/security/ magazine/17-10/m..

LYLE A. BRECHT 410.963.8680 DRAFT 4.1 CAPITAL MARKETS RESEARCH --- Friday, October 9, 2009

Page 10 of 10