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

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

The United States employs aversion of the precautionaryprinciple when it confrontsthreats to national security.The US spends vast amountson defences against threats unlikely toaffect Americans. Experts, defenceofficials and politicians justify thoseexpenditures by saying they arenecessary to protect the public fromworst-case dangers. Those claimsignore what is probable and whatdefences cost. They exaggerate thedanger our enemies pose and stripresources from more probabledangers, making us less safe.Surrounded by the demons ofpossibility, the American publicperceives a menacing and chaoticworld that is mostly fiction.

Consider war as a species of risk,danger or uncertainty. We are notaccustomed to this perspective. Thetheories that inform the study ofpolitical violence are not those thatguide regulation of health and safety.The US Defence Department is notconsidered to be in the same businessas the Food and Drug Administration.But we can glean insights into ourdefences from debate aboutregulatory policy. We can revealchoices among dangers hidden by talkof uncertainty and considertheir cause.

Students of regulatory policy knowof the precautionary principle — anidea about risk favoured by advocatesof various health and environmentalregulations. The concept can be statedas follows: Whenever some activityposes a possible risk to health, safetyor the environment, the governmentshould take preventive action.Government intervention iswarranted even if the evidence thatthe activity is harmful is uncertain andthe cost of preventive action is high.

In Laws of Fear: Beyond thePrecautionary Principle (2005),University of Chicago law professorCass Sunstein demonstrates that theprecautionary principle is incoherent.The principle fails to acknowledgethat decisions about risk, whether theyregulate health hazards or arm againsta state, cannot deal with one riskalone. Because resources are alwayslimited, efforts to head off a particulardanger take resources away fromother government programs and fromprivate investment that also reducerisk. Also, because of unintendedconsequences, actions that preventone danger can create new ones. If wetook the precautionary principleseriously, we would have to becautious about all the dangers aparticular decision touches. Thatincludes the danger of doing nothing.Taken literally, the principle preventsall action and inaction, makingit useless.

States often ignore this logicalfailure and apply the precautionaryprinciple to particular hazards.Sunstein argues that in many of thosecases, precautionary action will bemore harmful to society than runningthe risk. Those are cases where thedanger is small and the cost ofprevention is large.

The use of asbestos as buildinginsulation is an example. Whencontained in walls, asbestos'scarcinogenic effect is tiny — estimated to kill less than one personfor 1 million lifetimes, far less thanlightning. Still, removing the asbestosmay prevent a few cancer cases.The precautionary principle can beevoked by those demanding thematerial's removal. But removalcreates new cancer risks and its cost isenormous. Whoever bears it, that costwill take money away from other riskreducinguses, be it savings, healthcare or education. Removal harmssociety more than leaving the asbestosin place.

Another example is geneticallymodified foods. European regulatorsargue that the uncertain risks ofgenetically modified crops justifylimiting trade flows and the resultinghigher prices on consumers. Theyexchange an uncertain risk fora sure one.

The illogic of the precautionaryprinciple does not mean that statesshould not regulate against uncertaindangers. The point is that dangersshould be evaluated by cost-benefitanalysis. This means that decisionsabout risk should consider the costthat preventive action would avert, thelikelihood that preventive action willwork, and the cost of the action.Decision makers should also consider,as Sunstein notes, not just total costsand benefits, but the equity oftheir distribution.

The problem with cost-benefitanalysis is that it relies on unavailableinformation about the magnitude andlikelihood of the harm. Everyonewould agree to head off disaster at lowcost and to avoid costly defencesagainst tiny dangers. Everyone agreesthat research is helpful to gettingpolicy right. But some degree ofuncertainty is hard to extinguish. Younever know, some will say, what thetrue cost of asbestos as insulation is.If science is never complete, costbenefitanalysis is impossible.

The problem with this critique ofcost-benefit analysis is that its virtuedoes not depend on getting rid ofuncertainty. Analysts use cost-benefitanalysis to get all the potential costsinto the debate and force recognitionof choice. They show that the pursuitof perfect safety, of chasing a dangerout of existence, creates other dangers.

This point shows why debate aboutthe precautionary principle is oftenphoney. Inherent uncertainty meansthat the decisions about risk are likelyto be made by some criteria other thana principle about risk. That criterionwill be a prior political preference — in the case of genetically modifiedfoods probably protection ofdomestic producers.

Critics of the precautionaryprinciple charge that it isa justification for regulation, not itscause — that the principle's defenderscare more about the environment thanother public goods. Defenders of theprinciple claim that cost-benefitanalysis serves corporate bottom lines.They are both part right.

Fights about regulating risks areabout which risks to confront andwhich to accept, not about how muchrisk to accept. All government policiesultimately reduce one risk or another.Politics is competition betweenrisk preferences.

Societies are not consistent in theirapproach to dangers. They areprecautionary about certain risks andaccepting of others. Americans areless fearful — less precautionary — than Europeans about globalwarming and genetically modifiedfoods. They are are more cautiousabout second hand smoke, drugapproval and nuclear proliferation.The differences cannot be justified byobjective appeals to science.

Scholars offer various explanationsfor the origins of those preferences.Political scientists Mary Douglas andAaron Wildavsky argue that culturecauses risk perception. They claimthat groups are organised bypreferences about what dangers oughtto be confronted collectively and thatthe rise of new political coalitionsbrings new priorities about danger.Psychologist Paul Slovic points topeople's tendencies to react to certainrisks — such as those that are novel orinvolve a perceived loss of control — and the way those perceptions spreadby social interaction and media.Massachusetts Institute ofTechnology's (MIT) Harvey Sapolskyargues that risk perception resultsfrom the balance of the variousspecial interests that benefit fromsociety either confronting or runningthe risk. Special interest groupscompete to guide public opinionabout danger.

What these scholars agree on is thatwhatever their origin, politicalpreferences drive demand forregulation of risks. Statements aboutthe certainty or uncertainty of scienceare often disguises for thosepreferences.

In the US this discussion about theprecautionary principle applies tonational security dangers in twoways. First, American nationalsecurity policy is explicitlyprecautionary and is thus subject tothe same problems as the applicationof the precautionary principle inother policy areas. Second, theprecautionary reasoning advancedto defend US security policies hidespolitical motives. As with theregulatory arena, cost-benefitanalysis can help expose choicesamong risks that advocates ofprecaution shroud with claimsof uncertainty.

According to American pundits,politicians, and various nationalsecurity strategy documents, uncertaindangers stalk the US — "We knowthere are known unknowns . . .Butthere are also unknown unknowns",as former defence secretary DonaldRumsfeld once put it. We are told thatthe world is awash in civil war andterrorism, which, according to the2002 National Strategy for HomelandSecurity, could strike us "any time,any place with virtually any weapon",and might, as former chairman of theJoint Chiefs of Staff Richard Myersput it "do away with our way of life".Terrorists are said to hide in sleepercells across the country and to plotdestruction from proliferatingoutposts in failed states. It is said thatUS satellites are ripe for attack and theUS faces an "electronic PearlHarbor" because our computers arevulnerable. We are warned that Iran'snuclear weapons portend disaster, thatIraq may pull the Middle East intochaos, that North Korea might attackand that China may soon push theUS around.

The futures depicted in thosearguments are possible, but they arenot probable. It is mere possibility, nomatter how small and unlikely, thatjustifies the defences that policymakers advocate. Because you cannotknow for certain the odds of apotential danger, the logic says, youmust prepare for it. History teachesnations extreme caution — better safethan sorry.

Precautionary reasoning pervadesofficial writing about US nationalsecurity. President George Bush'sspeeches and national securityplanning documents, which aresupposed to guide our defencebudget, are rife with it. They depicta world of swirling uncertainty andrising danger. They claim that thesimple Soviet threat has beenreplaced by more varied andirrational ones that require expensivepreventive measures. Thesestatements avoid dealing inprobability and comparing the costsof defences with the risks they defendagainst. They imply that uncertaintyhas made that calculation impossible.

Laying out his pre-emptiondoctrine at West Point in 2002, thePresident said "if we wait for threatsto fully materialise, we will havewaited too long".

The National Security Strategyof 2002 echoes this preventive logicand says that today's dangers aremore "complex and dangerous" thanthe dangers the US faced in the ColdWar. The 2004 National MilitaryStrategy sees "uncertain andcomplex" threats. The 2005 NationalDefence Strategy claims that theprimary danger that the$US600 billion defence budgetconfronts is the unknown:

Uncertainty is the definingcharacteristic of today's strategicenvironment. We can identifytrends but cannot predict specificevents with precision. While wework to avoid being surprised, wemust posture ourselves to handleunanticipated problems — wemust plan with surprise in mind.

The 2006 Quadrennial DefenceReview (QDR) — the defenceplanning document drafted every fouryears to guide US defence spending — argues that the US faces a hostile mixof terrorists, failed states,insurgencies, rogue states and largemilitaries like China's. Like the otherdocuments, the QDR does notestimate the threats' likelihoods andrecommend focusing on one oranother on that basis. It contendssimply that "managing risks"compels the US to prepare for all ofthem. The QDR argues that the USmust act to head-off those risksthrough direct military attack, supportof proxies or stabilisation missions. Itthen recommends retention of theweapons and forces we have, witha few tweaks.

The clearest recent example ofprecautionary reasoning aboutsecurity is Vice-President DickCheney's so called 1 per centdoctrine, as reported by authorRon Suskind:

If there was even a 1 per centchance of terrorists gettinga weapon of mass destruction — and there has been a smallprobability of such an occurrencefor some time — the US must nowact as if it were a certainty.

Of course, a low-probabilitychance of a catastrophic eventjustifies expensive defensive actionseven under cost-benefit reasoning.But pretending that the catastropheis certain justifies overly risky andexpensive preventive actions. If thereis a 1 per cent chance of a $US10,000problem, you should not want to payinsurance worth more than $US100to prevent it. If the odds are certain,you'll pay anything less than$US10,000. That is the kind ofreasoning that makes the vastdefence budget and the Iraq Warseem sensible.

Precautionary reasoning extends tomore specific defence policies. Whydoes the US keep 30,000 troops inSouth Korea? Originally the troopsremained there after the Korean Warto protect South Korea againstanother communist invasion. Thatmay have made sense when SouthKorea was a relatively poor country,North Korea received strong backingfrom Beijing, and Washingtonbelieved in the domino theory. Buttoday the Cold War is over and SouthKorea has over 20 times the GDP and10 times the military spending of itsnorthern counterpart. Would aUSpullout from South Korea increasethe danger of North Koreanaggression? It might, a bit, althoughthe North Koreans would know thatthe US could still bring air power tobear quickly. Would an attackendanger Americans? It might harmthe economy and it certainly wouldoffend American values. But strategyshould not focus only on preventingremote dangers while ignoring the riskof tying down troops that are neededelsewhere and expendingconsiderable resources and money.This does not necessarily mean thatWashington should abandon Korea,but that it should consider all therelevant risks.

The continuation of the Iraq Waris often justified using similar logic.Advocates of the war imagine theconsequences of a US withdrawal asregional Iranian hegemony, a Saudi-Iran war, or a terrorist state in Iraq.Those lurid possibilities aregenerally offered without attempts tojudge their probability or how thepossible costs of withdrawal stack upagainst the certain costs of staying — costs that include further inflaminganti-Americanism, $US2 billion to$US3 billion in military spendingeach week, an inability to deploytroops elsewhere, lost politicalcapital, recruiting difficulties for thearmy, and many dead and woundedAmerican servicemen.

Only a precautionary ethos explainswhy the federal government spendslarge sums to protect every Americanstate from terrorism. Terrorists couldstrike Arkansas. Based on history, ifthey did so, they would probably killseveral dozen people in an explosion.That possibility is so remote and theutility of the spending in addressingthe risk is so unclear that the rightamount of homeland securityspending in Arkansas, and most areasof the US, is probably zero.

The dirty secret of Americannational security politics is that thecountry is relatively safe. Officialrhetoric shrouds an increasingly stableand peaceful world. There is no basisfor believing that the world isbecoming more uncertain anddangerous. The Cold War was notpredictable. Few predicted its end.Few agreed on Soviet intentions, onhow much to spend on defence, andon which states were worth defendingfrom communist aggression orinsurgency. The giant clash did notcome, but the world was not stable.Rogue states prospered. Civil warsraged. States failed.

Today, peace, liberalism, and orderare spreading. According to a 2005University of Maryland study, there isless war now than at any time sincethe 1950s. War between states is farless common than it was during theCold War, and seems to bedisappearing among the mostpowerful states. Civil war, the farmore common type, has beendeclining since the early 1990s. A 2006study from the University of BritishColumbia reports that the incidenceof genocide has declined sharply sincethe end of the Cold War. Thoughterrorism is up lately, the number ofincidents has not surpassed that ofthe 1980s. The report also shows thatwars have become less deadly onaverage even as they become lessnumerous. Meanwhile, GDP rates arerising around the world, even inAfrica, aiding the spread of law andorder via taxes.

Americans are among the mostsecure people in history. On averagethey live 78 years, longer than everbefore. The threats that havehistorically driven states' militaryspending — civil war and invasion — are unthinkable in North America.The closest thing to state enemies,North Korea, Iran, and Syria, lack thecapability to attack US shores directly.Together they spend about $US10billion a year on their militaries, lessthan a sixtieth what the US does.None of those states would have goodprospects for territorial expansioneven if US non-intervention wereassured. And it is not clear howAmericans would be endangered ifthey did expand.

Our biggest worry, terrorists, killless Americans than allergic toreactions peanuts in most years. Evenin 2001, terrorists killed less thana tenth as many Americans as the flu.The minority of the jihadistmovement that is actively trying tokill Americans has shown no sign ofa presence in the US itself, as the FBIhas grudgingly admitted.

Why do Americans never hearthe good news about theirsecurity? Why do US leadersconjure up so many possible monstersto destroy, and then overspend toconfront them? One explanation is theimbalance of interests andgovernment's near-monopoly oninformation in the nationalsecurity realm.

In other policy arenas, likeenvironmental politics, there arestrong private interests on both sides.Environmental groups preachprecaution. Business interestsadvocate regulatory restraint. Theresult is a fair political fight thatcreates debate. As marketplaces ofideas go, that is not bad. In USnational security politics, there isdebate, but all the interests are oneside. Both parties see political rewardin preaching danger. The massive USnational security establishment relieson a sense of threat to stay in business.On the other side, as former secretaryof defence Les Aspin once wrote,there is no other side. No one alarmsus about alarmism. Everyone likeslower taxes, but not enough toorganise interest groups againstdefence spending. A scattering oflibertarians and anti-war liberalsconfront a bipartisan juggernaut. Theinformation about national securitythreats comes to Americansprincipally from people driven byorganisational or electoral incentivestoward threat inflation.

In the US one source ofprecautionary messages aboutsecurity is the American two-partysystem. There is no dove party. Inrecent decades, the Republicans havewon elections by preaching nationalsecurity vigilance. Few Democratsrespond by making the case fora security policy that accepts morerisk in exchange for more savings.That position would encourage themto downplay security dangers, asisolationist Republicans did in thefirst half the 20th century. Instead,Democrats — particularly inpresidential elections — move to thepolitical right to neutralise nationalsecurity issues. They balance theirrelatively dovish stance on Iraq bysupporting the enormous militarybudget and demanding morespending on homeland security, aid tofailed states, and preventing weaponsproliferation. The result is a debatewhere no party profits by helpingAmericans perceive their safety.

Information about national securityalso tends to be precautionary becauseit is provided by a massive securityestablishment with an interest in asense of danger. Scientist WilliamClark, writing about the history of riskassessment, notes that medievalEuropeans did not much fear witchesuntil they created an inquisition tofind them. The institutionalisation ofthe hunt heightened perceptions of thedanger hunted. A similar problemhaunts modern Americans. The largesupply of defence creates a largedemand for it.

Not all threats are exaggerated.Those that lack institutions designedto confront the dangers (and pleadfor government support) can beneglected. Because there was nolarge interest designated to fightterrorism in the 1990s — no naturalbureaucratic champion — the threatwas arguably given too littleattention. Likewise, one explanationfor the anemic American reaction toNazism before World War II was theabsence of a substantial militaryestablishment that would havegained by promoting it. Some willargue that Nazi Germany proves thatyou can never have too muchvigilance. But the solution to novigilance is not over-vigilance. Andexamples of American underreactionare rare. The generalAmerican affliction is threatinflation, not threat denial.

In times of great danger, a large setof interests that profit by promotingdanger and defence are necessary.Defence is a public good and, asUniversity of Georgia economistDwight Lee has noted, it will beunderprovided unless public entitiesgain from its provision. But if weconclude that the dangers of defenceare greater than the danger theyconfront, how to right the ship? Canwe avoid precautionary securitypolitics? Obviously, Americansshould give greater consideration tothe costs of defences. But politicalproblems need political solutions.There is no analytic holy grail.

As with other hazards, the beststrategy for dealing with trueuncertainty is gathering informationto assess the magnitude of the danger,as MIT's Kenneth Oye has written. Innational security that meansintelligence. Empowering intelligenceagencies at the expense of militaryservices might help steer the politicalenergy created by fear into moreproductive uses.

More generally, what we need areinterests that profit by exposingprecautionary reasoning in security, thecounterparts of the industries thatencourage scepticism about the extentof various health and safety hazards.Expanding the American politicalsystem to include a party that wins atthe polls by attacking militarism mightaccomplish this, but that is a pipedream.Amore realistic solution is toprovoke more competition amonggovernment agencies. Today in thePentagon, "jointness" is a nearreligiousprinciple. Open competitionbetween military branches is taboo.

But if war-fighters need unity, civilianmanagers need rivalries to exploit.Security strategies should cap spendingand pit organisational budgets againsteach other, eliminating the fixed sharesbetween the services and threatening tomove less useful funds out of defenceinto intelligence, diplomacy or foreignaid and even to the non-defence side ofthe budget. Made to fight, differentfederal agencies and departmentsmight publicly wield theories about riskin budgetary battles. For instance, thearmy might note (or encourage those itfunds to note) that terrorists are rarelyfound at sea, making the navy's claimsto counter-terrorism spoils less credible.The State Department might argue thatterrorism is primarily a politicalproblem, not a military one. The morepublic these fights, the better. Publicfights let the public learn what dangersare more exaggerated.

We could also use fights that crossrisk categories. We need moremechanisms that pit a dollar spent onhealth care against a dollar spent ondefending Taiwan. One way to dothis is to strengthen the budgetaryoverseers in places like the Office ofManagement and Budget. Anotherway of provoking competition is abudgetary crisis. Big changes ingovernment tend to come from rareevents like wars and landslideelections, not sudden epiphanies bypolicymakers; increased entitlementcosts combined with an economicslowdown might force a day ofreckoning where the 20 per cent ofthe federal budget spent on defencebecomes an attractive piggy bank.A competition of risks might ensue.A government decision to take onmore health care costs could createsimilar pressure. Of course, thesecomparisons occur somewhat today.This helps explain why the defencebudget does occasionally decrease.More of this competition is needed.

An obvious fix is less secrecy.Reforming the system ofclassification to make moreinformation public would make itharder for officials to maximisealarm by cherry-picking informationfor release. It would also distributeexpertise or credible claims to it andempower independent experts.

Debates about national securitycould also use more trulyindependent experts. Morethinktanks that encourage acontrarian ethos would be helpful, aswould more academic securityspecialists. Tenure's purpose is toinsure intellectual independence.Unfortunately, in the US the academyhas largely abandoned the study ofnational security politics as distinctfrom international relations.

No formula tells us how tomaximise safety. But scepticism — towards both what we are told to fearand the defences we are sold toconfront it — is a good start.

Benjamin H. Friedman

Benjamin H. Friedman is a research fellow in defense and homeland security studies at the Cato Institute and a PhD candidate in political science at the Massachusetts Institute of Technology.