The Problem of Making Policy Based on the Premise ‘If’

April 18, 2008 • Commentary
This article appeared in the Australian Financial Review on April 18, 2008.

The United States employs a version of the precautionary principle when it confronts threats to national security. The US spends vast amounts on defences against threats unlikely to affect Americans. Experts, defence officials and politicians justify those expenditures by saying they are necessary to protect the public from worst‐​case dangers. Those claims ignore what is probable and what defences cost. They exaggerate the danger our enemies pose and strip resources from more probable dangers, making us less safe. Surrounded by the demons of possibility, the American public perceives a menacing and chaotic world that is mostly fiction.

Consider war as a species of risk, danger or uncertainty. We are not accustomed to this perspective. The theories that inform the study of political violence are not those that guide regulation of health and safety. The US Defence Department is not considered to be in the same business as the Food and Drug Administration. But we can glean insights into our defences from debate about regulatory policy. We can reveal choices among dangers hidden by talk of uncertainty and consider their cause.

Students of regulatory policy know of the precautionary principle — an idea about risk favoured by advocates of various health and environmental regulations. The concept can be stated as follows: Whenever some activity poses a possible risk to health, safety or the environment, the government should take preventive action. Government intervention is warranted even if the evidence that the activity is harmful is uncertain and the cost of preventive action is high.

In Laws of Fear: Beyond the Precautionary Principle (2005), University of Chicago law professor Cass Sunstein demonstrates that the precautionary principle is incoherent. The principle fails to acknowledge that decisions about risk, whether they regulate health hazards or arm against a state, cannot deal with one risk alone. Because resources are always limited, efforts to head off a particular danger take resources away from other government programs and from private investment that also reduce risk. Also, because of unintended consequences, actions that prevent one danger can create new ones. If we took the precautionary principle seriously, we would have to be cautious about all the dangers a particular decision touches. That includes the danger of doing nothing. Taken literally, the principle prevents all action and inaction, making it useless.

States often ignore this logical failure and apply the precautionary principle to particular hazards. Sunstein argues that in many of those cases, precautionary action will be more harmful to society than running the risk. Those are cases where the danger is small and the cost of prevention is large.

The use of asbestos as building insulation is an example. When contained in walls, asbestos’s carcinogenic effect is tiny — estimated to kill less than one person for 1 million lifetimes, far less than lightning. Still, removing the asbestos may prevent a few cancer cases. The precautionary principle can be evoked by those demanding the material’s removal. But removal creates new cancer risks and its cost is enormous. Whoever bears it, that cost will take money away from other riskreducing uses, be it savings, health care or education. Removal harms society more than leaving the asbestos in place.

Another example is genetically modified foods. European regulators argue that the uncertain risks of genetically modified crops justify limiting trade flows and the resulting higher prices on consumers. They exchange an uncertain risk for a sure one.

The illogic of the precautionary principle does not mean that states should not regulate against uncertain dangers. The point is that dangers should be evaluated by cost‐​benefit analysis. This means that decisions about risk should consider the cost that preventive action would avert, the likelihood that preventive action will work, and the cost of the action. Decision makers should also consider, as Sunstein notes, not just total costs and benefits, but the equity of their distribution.

The problem with cost‐​benefit analysis is that it relies on unavailable information about the magnitude and likelihood of the harm. Everyone would agree to head off disaster at low cost and to avoid costly defences against tiny dangers. Everyone agrees that research is helpful to getting policy right. But some degree of uncertainty is hard to extinguish. You never know, some will say, what the true cost of asbestos as insulation is. If science is never complete, costbenefit analysis is impossible.

The problem with this critique of cost‐​benefit analysis is that its virtue does not depend on getting rid of uncertainty. Analysts use cost‐​benefit analysis to get all the potential costs into the debate and force recognition of choice. They show that the pursuit of perfect safety, of chasing a danger out of existence, creates other dangers.

This point shows why debate about the precautionary principle is often phoney. Inherent uncertainty means that the decisions about risk are likely to be made by some criteria other than a principle about risk. That criterion will be a prior political preference — in the case of genetically modified foods probably protection of domestic producers.

Critics of the precautionary principle charge that it is a justification for regulation, not its cause — that the principle’s defenders care more about the environment than other public goods. Defenders of the principle claim that cost‐​benefit analysis serves corporate bottom lines. They are both part right.

Fights about regulating risks are about which risks to confront and which to accept, not about how much risk to accept. All government policies ultimately reduce one risk or another. Politics is competition between risk preferences.

Societies are not consistent in their approach to dangers. They are precautionary about certain risks and accepting of others. Americans are less fearful — less precautionary — than Europeans about global warming and genetically modified foods. They are are more cautious about second hand smoke, drug approval and nuclear proliferation. The differences cannot be justified by objective appeals to science.

Scholars offer various explanations for the origins of those preferences. Political scientists Mary Douglas and Aaron Wildavsky argue that culture causes risk perception. They claim that groups are organised by preferences about what dangers ought to be confronted collectively and that the rise of new political coalitions brings new priorities about danger. Psychologist Paul Slovic points to people’s tendencies to react to certain risks — such as those that are novel or involve a perceived loss of control — and the way those perceptions spread by social interaction and media. Massachusetts Institute of Technology’s (MIT) Harvey Sapolsky argues that risk perception results from the balance of the various special interests that benefit from society either confronting or running the risk. Special interest groups compete to guide public opinion about danger.

What these scholars agree on is that whatever their origin, political preferences drive demand for regulation of risks. Statements about the certainty or uncertainty of science are often disguises for those preferences.

In the US this discussion about the precautionary principle applies to national security dangers in two ways. First, American national security policy is explicitly precautionary and is thus subject to the same problems as the application of the precautionary principle in other policy areas. Second, the precautionary reasoning advanced to defend US security policies hides political motives. As with the regulatory arena, cost‐​benefit analysis can help expose choices among risks that advocates of precaution shroud with claims of uncertainty.

According to American pundits, politicians, and various national security strategy documents, uncertain dangers stalk the US — “We know there are known unknowns …But there are also unknown unknowns”, as former defence secretary Donald Rumsfeld once put it. We are told that the world is awash in civil war and terrorism, which, according to the 2002 National Strategy for Homeland Security, could strike us “any time, any place with virtually any weapon”, and might, as former chairman of the Joint Chiefs of Staff Richard Myers put it “do away with our way of life”. Terrorists are said to hide in sleeper cells across the country and to plot destruction from proliferating outposts in failed states. It is said that US satellites are ripe for attack and the US faces an “electronic Pearl Harbor” because our computers are vulnerable. We are warned that Iran’s nuclear weapons portend disaster, that Iraq may pull the Middle East into chaos, that North Korea might attack and that China may soon push the US around.

The futures depicted in those arguments are possible, but they are not probable. It is mere possibility, no matter how small and unlikely, that justifies the defences that policy makers advocate. Because you cannot know for certain the odds of a potential danger, the logic says, you must prepare for it. History teaches nations extreme caution — better safe than sorry.

Precautionary reasoning pervades official writing about US national security. President George Bush’s speeches and national security planning documents, which are supposed to guide our defence budget, are rife with it. They depict a world of swirling uncertainty and rising danger. They claim that the simple Soviet threat has been replaced by more varied and irrational ones that require expensive preventive measures. These statements avoid dealing in probability and comparing the costs of defences with the risks they defend against. They imply that uncertainty has made that calculation impossible.

Laying out his pre‐​emption doctrine at West Point in 2002, the President said “if we wait for threats to fully materialise, we will have waited too long”.

The National Security Strategy of 2002 echoes this preventive logic and says that today’s dangers are more “complex and dangerous” than the dangers the US faced in the Cold War. The 2004 National Military Strategy sees “uncertain and complex” threats. The 2005 National Defence Strategy claims that the primary danger that the $US600 billion defence budget confronts is the unknown:

Uncertainty is the defining characteristic of today’s strategic environment. We can identify trends but cannot predict specific events with precision. While we work to avoid being surprised, we must posture ourselves to handle unanticipated problems — we must plan with surprise in mind.

The 2006 Quadrennial Defence Review (QDR) — the defence planning document drafted every four years to guide US defence spending — argues that the US faces a hostile mix of terrorists, failed states, insurgencies, rogue states and large militaries like China’s. Like the other documents, the QDR does not estimate the threats’ likelihoods and recommend focusing on one or another on that basis. It contends simply that “managing risks” compels the US to prepare for all of them. The QDR argues that the US must act to head‐​off those risks through direct military attack, support of proxies or stabilisation missions. It then recommends retention of the weapons and forces we have, with a few tweaks.

The clearest recent example of precautionary reasoning about security is Vice‐​President Dick Cheney’s so called 1 per cent doctrine, as reported by author Ron Suskind:

If there was even a 1 per cent chance of terrorists getting a weapon of mass destruction — and there has been a small probability of such an occurrence for some time — the US must now act as if it were a certainty.

Of course, a low‐​probability chance of a catastrophic event justifies expensive defensive actions even under cost‐​benefit reasoning. But pretending that the catastrophe is certain justifies overly risky and expensive preventive actions. If there is a 1 per cent chance of a $US10,000 problem, you should not want to pay insurance worth more than $US100 to prevent it. If the odds are certain, you’ll pay anything less than $US10,000. That is the kind of reasoning that makes the vast defence budget and the Iraq War seem sensible.

Precautionary reasoning extends to more specific defence policies. Why does the US keep 30,000 troops in South Korea? Originally the troops remained there after the Korean War to protect South Korea against another communist invasion. That may have made sense when South Korea was a relatively poor country, North Korea received strong backing from Beijing, and Washington believed in the domino theory. But today the Cold War is over and South Korea has over 20 times the GDP and 10 times the military spending of its northern counterpart. Would aUS pullout from South Korea increase the danger of North Korean aggression? It might, a bit, although the North Koreans would know that the US could still bring air power to bear quickly. Would an attack endanger Americans? It might harm the economy and it certainly would offend American values. But strategy should not focus only on preventing remote dangers while ignoring the risk of tying down troops that are needed elsewhere and expending considerable resources and money. This does not necessarily mean that Washington should abandon Korea, but that it should consider all the relevant risks.

The continuation of the Iraq War is often justified using similar logic. Advocates of the war imagine the consequences of a US withdrawal as regional Iranian hegemony, a Saudi‐ Iran war, or a terrorist state in Iraq. Those lurid possibilities are generally offered without attempts to judge their probability or how the possible costs of withdrawal stack up against the certain costs of staying — costs that include further inflaming anti‐​Americanism, $US2 billion to $US3 billion in military spending each week, an inability to deploy troops elsewhere, lost political capital, recruiting difficulties for the army, and many dead and wounded American servicemen.

Only a precautionary ethos explains why the federal government spends large sums to protect every American state from terrorism. Terrorists could strike Arkansas. Based on history, if they did so, they would probably kill several dozen people in an explosion. That possibility is so remote and the utility of the spending in addressing the risk is so unclear that the right amount of homeland security spending in Arkansas, and most areas of the US, is probably zero.

The dirty secret of American national security politics is that the country is relatively safe. Official rhetoric shrouds an increasingly stable and peaceful world. There is no basis for believing that the world is becoming more uncertain and dangerous. The Cold War was not predictable. Few predicted its end. Few agreed on Soviet intentions, on how much to spend on defence, and on which states were worth defending from communist aggression or insurgency. The giant clash did not come, but the world was not stable. Rogue states prospered. Civil wars raged. States failed.

Today, peace, liberalism, and order are spreading. According to a 2005 University of Maryland study, there is less war now than at any time since the 1950s. War between states is far less common than it was during the Cold War, and seems to be disappearing among the most powerful states. Civil war, the far more common type, has been declining since the early 1990s. A 2006 study from the University of British Columbia reports that the incidence of genocide has declined sharply since the end of the Cold War. Though terrorism is up lately, the number of incidents has not surpassed that of the 1980s. The report also shows that wars have become less deadly on average even as they become less numerous. Meanwhile, GDP rates are rising around the world, even in Africa, aiding the spread of law and order via taxes.

Americans are among the most secure people in history. On average they live 78 years, longer than ever before. The threats that have historically driven states’ military spending — civil war and invasion — are unthinkable in North America. The closest thing to state enemies, North Korea, Iran, and Syria, lack the capability to attack US shores directly. Together they spend about $US10 billion a year on their militaries, less than a sixtieth what the US does. None of those states would have good prospects for territorial expansion even if US non‐​intervention were assured. And it is not clear how Americans would be endangered if they did expand.

Our biggest worry, terrorists, kill less Americans than allergic to reactions peanuts in most years. Even in 2001, terrorists killed less than a tenth as many Americans as the flu. The minority of the jihadist movement that is actively trying to kill Americans has shown no sign of a presence in the US itself, as the FBI has grudgingly admitted.

Why do Americans never hear the good news about their security? Why do US leaders conjure up so many possible monsters to destroy, and then overspend to confront them? One explanation is the imbalance of interests and government’s near‐​monopoly on information in the national security realm.

In other policy arenas, like environmental politics, there are strong private interests on both sides. Environmental groups preach precaution. Business interests advocate regulatory restraint. The result is a fair political fight that creates debate. As marketplaces of ideas go, that is not bad. In US national security politics, there is debate, but all the interests are one side. Both parties see political reward in preaching danger. The massive US national security establishment relies on a sense of threat to stay in business. On the other side, as former secretary of defence Les Aspin once wrote, there is no other side. No one alarms us about alarmism. Everyone likes lower taxes, but not enough to organise interest groups against defence spending. A scattering of libertarians and anti‐​war liberals confront a bipartisan juggernaut. The information about national security threats comes to Americans principally from people driven by organisational or electoral incentives toward threat inflation.

In the US one source of precautionary messages about security is the American two‐​party system. There is no dove party. In recent decades, the Republicans have won elections by preaching national security vigilance. Few Democrats respond by making the case for a security policy that accepts more risk in exchange for more savings. That position would encourage them to downplay security dangers, as isolationist Republicans did in the first half the 20th century. Instead, Democrats — particularly in presidential elections — move to the political right to neutralise national security issues. They balance their relatively dovish stance on Iraq by supporting the enormous military budget and demanding more spending on homeland security, aid to failed states, and preventing weapons proliferation. The result is a debate where no party profits by helping Americans perceive their safety.

Information about national security also tends to be precautionary because it is provided by a massive security establishment with an interest in a sense of danger. Scientist William Clark, writing about the history of risk assessment, notes that medieval Europeans did not much fear witches until they created an inquisition to find them. The institutionalisation of the hunt heightened perceptions of the danger hunted. A similar problem haunts modern Americans. The large supply of defence creates a large demand for it.

Not all threats are exaggerated. Those that lack institutions designed to confront the dangers (and plead for government support) can be neglected. Because there was no large interest designated to fight terrorism in the 1990s — no natural bureaucratic champion — the threat was arguably given too little attention. Likewise, one explanation for the anemic American reaction to Nazism before World War II was the absence of a substantial military establishment that would have gained by promoting it. Some will argue that Nazi Germany proves that you can never have too much vigilance. But the solution to no vigilance is not over‐​vigilance. And examples of American underreaction are rare. The general American affliction is threat inflation, not threat denial.

In times of great danger, a large set of interests that profit by promoting danger and defence are necessary. Defence is a public good and, as University of Georgia economist Dwight Lee has noted, it will be underprovided unless public entities gain from its provision. But if we conclude that the dangers of defence are greater than the danger they confront, how to right the ship? Can we avoid precautionary security politics? Obviously, Americans should give greater consideration to the costs of defences. But political problems need political solutions. There is no analytic holy grail.

As with other hazards, the best strategy for dealing with true uncertainty is gathering information to assess the magnitude of the danger, as MIT’s Kenneth Oye has written. In national security that means intelligence. Empowering intelligence agencies at the expense of military services might help steer the political energy created by fear into more productive uses.

More generally, what we need are interests that profit by exposing precautionary reasoning in security, the counterparts of the industries that encourage scepticism about the extent of various health and safety hazards. Expanding the American political system to include a party that wins at the polls by attacking militarism might accomplish this, but that is a pipe dream.Amore realistic solution is to provoke more competition among government agencies. Today in the Pentagon, “jointness” is a nearreligious principle. Open competition between military branches is taboo.

But if war‐​fighters need unity, civilian managers need rivalries to exploit. Security strategies should cap spending and pit organisational budgets against each other, eliminating the fixed shares between the services and threatening to move less useful funds out of defence into intelligence, diplomacy or foreign aid and even to the non‐​defence side of the budget. Made to fight, different federal agencies and departments might publicly wield theories about risk in budgetary battles. For instance, the army might note (or encourage those it funds to note) that terrorists are rarely found at sea, making the navy’s claims to counter‐​terrorism spoils less credible. The State Department might argue that terrorism is primarily a political problem, not a military one. The more public these fights, the better. Public fights let the public learn what dangers are more exaggerated.

We could also use fights that cross risk categories. We need more mechanisms that pit a dollar spent on health care against a dollar spent on defending Taiwan. One way to do this is to strengthen the budgetary overseers in places like the Office of Management and Budget. Another way of provoking competition is a budgetary crisis. Big changes in government tend to come from rare events like wars and landslide elections, not sudden epiphanies by policymakers; increased entitlement costs combined with an economic slowdown might force a day of reckoning where the 20 per cent of the federal budget spent on defence becomes an attractive piggy bank. A competition of risks might ensue. A government decision to take on more health care costs could create similar pressure. Of course, these comparisons occur somewhat today. This helps explain why the defence budget does occasionally decrease. More of this competition is needed.

An obvious fix is less secrecy. Reforming the system of classification to make more information public would make it harder for officials to maximise alarm by cherry‐​picking information for release. It would also distribute expertise or credible claims to it and empower independent experts.

Debates about national security could also use more truly independent experts. More thinktanks that encourage a contrarian ethos would be helpful, as would more academic security specialists. Tenure’s purpose is to insure intellectual independence. Unfortunately, in the US the academy has largely abandoned the study of national security politics as distinct from international relations.

No formula tells us how to maximise safety. But scepticism — towards both what we are told to fear and the defences we are sold to confront it — is a good start.

About the Author