Does the US nuclear policy still target cities?

(This is a cross­post from LessWrong; here’s the origi­nal)

The his­tory of nu­clear strate­gic bombing

Daniel Ells­berg’s The Dooms­day Ma­chine brought my at­ten­tion to a hor­rify­ing fact about early US nu­clear tar­get­ing policy. In 1961, the US had only one nu­clear war plan, and it called for the de­struc­tion of ev­ery ma­jor Soviet city and mil­i­tary tar­get. That is not sur­pris­ing. How­ever, the plan also called for the de­struc­tion of ev­ery ma­jor Chi­nese city and mil­i­tary tar­get, even if China had not pro­voked the United States. In other words, the US nu­clear war plan called for the de­struc­tion of the ma­jor pop­u­la­tion cen­ters of the most pop­u­lous coun­try in the world, even in cir­cum­stances where that coun­try had not at­tacked the United States or its al­lies. Ells­berg points out that at the time, peo­ple at RAND and pre­sum­ably other parts of the US defense es­tab­lish­ment un­der­stood that the Chi­nese and the Soviets were be­gin­ning to di­verge in strate­gic in­ter­ests and thus should not be treated as one bloc. Nev­er­the­less, the top lev­els of the US com­mand, in­clud­ing Pres­i­dent Eisen­hower, were com­mit­ted to the ut­ter de­struc­tion of both Chi­nese and Soviet tar­gets in the event of a war with ei­ther coun­try.

The policy of de­stroy­ing cities is a legacy left over from strate­gic bomb­ing in World War II. The de­struc­tion of Hiroshima and Na­gasaki are the most fa­mous, but the fire bomb­ings of Ja­panese and Ger­many cities de­stroyed far more in­fras­truc­ture and kil­led far more peo­ple than the two atomic bombs. The given ra­tio­nal for strate­gic bomb­ing was to de­stroy the abil­ity of the en­emy states to con­tinue to make war. If a state can no longer pro­duce air­planes and tanks, ei­ther be­cause the fac­to­ries have been de­stroyed or be­cause there are no longer peo­ple to work in the fac­to­ries, then its abil­ity to re­sist is diminished.

Given the level of tech­nol­ogy and de­vel­op­ment in WWII, strate­gic bomb­ing had a chance at achiev­ing mil­i­tary ob­jec­tives, be­cause the con­flict was to carry on for mul­ti­ple years. On the timescale of years, a coun­try’s ca­pac­ity to build ar­ma­ments and re­sup­ply armies in the field can be cru­cial to vic­tory.

Nu­clear war changes this calcu­lus. In a mod­ern nu­clear war in­volv­ing SLBMs (Sub­marine launched bal­lis­tic mis­siles), ICBMs (In­ter­con­ti­nen­tal bal­lis­tic mis­siles), strate­gic bombers, and other weapon sys­tems, the ma­jor­ity of an ad­ver­sary’s mil­i­tary, in­dus­trial and pop­u­la­tion cen­ters could be de­stroyed in a mat­ter of days or hours. It is hard to imag­ine a nu­clear war last­ing years or even months. Without a pro­longed war, the origi­nal ra­tio­nale for strate­gic bomb­ing dis­ap­pears, or is at least much re­duced. A state may still wish to re­duce the ca­pac­ity of its en­emy to fight fu­ture wars, but it can no longer claim that the whole­sale de­struc­tion of cities is nec­es­sary to achieve mil­i­tary ob­jec­tives in the cur­rent war.

Why then, did early US nu­clear poli­cies call for the de­struc­tion of cities?

Nu­clear game the­ory in the 1960s

The de­struc­tion of cities was pri­mar­ily a threat of in­flict­ing harm rather than an at­tempt to de­stroy the ca­pac­ity of the en­emy to wage war. The idea, for­mal­ized by RAND game the­o­rist Thomas Schel­ling, was that both the United States and the Soviet Union would threaten mas­sive re­tal­i­a­tion against each other’s civilian pop­u­la­tions and in­dus­try to de­ter the other from start­ing a war.

Schel­ling de­vel­oped a cat­e­gory of game the­ory that in­volved what he termed “mixed mo­tive games”. Games where both sides sought ad­van­tage, but where the pay­off to one side did not strictly cor­re­late to the loss to the other side. In these types of games, both play­ers may wish to avoid out­comes that are mu­tu­ally un­fa­vor­able (Strat­egy of Con­flict, pg. 89). In the case of nu­clear de­ter­rence, both sides strongly preferred to avoid nu­clear war, and thus were both were de­terred from tak­ing ac­tions that would di­rectly lead to nu­clear war.

Much of Schel­ling’s work con­cerns it­self with how states in a nu­clear stale­mate can pur­sue their own ad­van­tage while avoid­ing es­ca­la­tion to nu­clear war. In this type of game, states try to ma­neu­ver each other into po­si­tions where the only pos­si­ble ac­tions are 1) es­ca­late and risk nu­clear war or 2) de-es­ca­late and con­cede some­thing to the other side.

Dur­ing the Cuban Mis­sile Cri­sis, Kennedy or­dered a block­ade of Cuba, be­liev­ing that such an ac­tion would not be suffi­cient for the Soviet Union to ini­ti­ate a war. The United States be­lieved that the Soviet Union would not try to break the block­ade, be­cause such an ac­tion would be rec­og­nized by both sides as start­ing a (nu­clear) war. Be­cause Kennedy proved cor­rect in his be­lief that the Soviet Union would not go to war over the block­ade nor risk ini­ti­at­ing war by break­ing the block­ade, the United States used to its ad­van­tage both coun­tries un­will­ing­ness to go to war.

What does this have to do with the tar­get­ing cities? To an­swer this ques­tion it’s nec­es­sary to con­sider how a nu­clear war might start. Although nu­clear pow­ers would al­most always pre­fer to avoid a nu­clear war, each has an in­cen­tive to strike first if they be­lieve nu­clear war to be in­evitable. By strik­ing first they may de­stroy their ad­ver­sary’s nu­clear forces be­fore they can be used. At this point, it’s use­ful to define a cou­ple of terms. Coun­terforce tar­get­ing refers to the tar­get­ing of en­emy mil­i­tary in­stal­la­tions, es­pe­cially other nu­clear forces. Coun­ter­value tar­get­ing refers to the tar­get­ing of en­emy in­fras­truc­ture and pop­u­la­tion cen­ters.

Con­sider the pri­mary goal of a first strike. Un­der the most plau­si­ble nu­clear war sce­nar­ios, it is to elimi­nate the nu­clear forces of the ri­val state; its ob­jec­tive is pri­mar­ily coun­terforce in na­ture. This is markedly differ­ent than the goal of a sec­ond strike. The pri­mary goal of a sec­ond strike, un­der nor­mal as­sump­tions of de­ter­rence, is ac­tu­ally to provide fulfil­ment of the pre-com­mit­ment made to re­tal­i­ate if ever at­tacked. That is, it is nec­es­sary to ac­tu­ally be com­mit­ted to at­tack­ing sec­ond so as to avoid be­ing at­tacked in the first place.

Schel­ling effec­tively ar­gued that the more pun­ish­ing the sec­ond strike threat­ened to be, the more effec­tive the de­ter­rent would be also. If true, then in the event of a nu­clear war, a state fol­low­ing the op­ti­mal strat­egy of de­ter­rence would tar­get cities as well as nu­clear tar­gets to make their nu­clear re­sponse as pun­ish­ing as pos­si­ble; that is, it would de­stroy both coun­terforce and coun­ter­value tar­gets. This would seem to lead to a policy for states to at­tack cities, with­out a sec­ond thought. In­deed, this was the policy of the US and the Soviet Union in the 1950s and early 1960s. How­ever, just be­cause tar­get­ing cities promised to be a more effec­tive de­ter­rent did not mean it promised to be the best policy.

Given some prob­a­bil­ity of nu­clear war, the effec­tive­ness of a de­ter­rent strat­egy ought to be weighed against the sever­ity of the re­sult­ing war were that strat­egy em­ployed. In other words, it might make sense for a state to com­mit to not tar­get­ing cities in a sec­ond strike if they them­selves do not have their cities de­stroyed in a first strike. While this may re­duce the effec­tive­ness of their de­ter­rent (and per­haps only marginally—nu­clear war is plenty dam­ag­ing with­out cities be­ing de­stroyed—the fal­lout alone will kill many mil­lions), it may also greatly re­duce the severe­ness of a nu­clear war.

Her­man Kahn, a promi­nent and con­tro­ver­sial RAND re­searcher, ar­gued that states would be ra­tio­nal to re­frain from de­stroy­ing cities in a first strike, to re­tain some bar­ter­ing power that might al­low them to save more of their own cities. The ar­gu­ment is that the defend­ing force might re­frain from de­stroy­ing many en­emy cities if do­ing so pre­vented their own cities from be­ing de­stroyed. Kahn be­lieved that the US should study and pre­pare for ne­go­ti­at­ing for the avoidance of US cities in a nu­clear war and that in or­der to do this the coun­try should:

  1. Develop the abil­ity to have suffi­ciently pro­tected or hid­den nu­clear forces to be able to both sur­vive a first strike and carry out coun­terforce and coun­ter­value at­tacks.

  2. Have “backup pres­i­dents”, or peo­ple with au­thor­ity to both or­der at­tacks and ne­go­ti­ate with the Soviet Union in the midst of a war, and that the US should have mul­ti­ple se­cure lo­ca­tions which are staffed 247 by these lead­ers.

Both Her­man Kahn and Thomas Schel­ling agreed that ne­go­ti­at­ing the end of a nu­clear war would be difficult, but both be­lieved it was crit­i­cal that nu­clear states re­main ca­pa­ble of ne­go­ti­a­tion. Schel­ling writes about this in his 1966 work, Arms and In­fluence:

The clos­ing stage, fur­ther­more, might have to be­gin quickly, pos­si­bly be­fore the first vol­ley had reached its tar­gets; and even the most con­fi­dent vic­tor would need to in­duce his en­emy to avoid a fi­nal, fu­tile orgy of hope­less re­venge. In ear­lier times, one could plan the open­ing moves of war in de­tail and hope to im­pro­vise plans for its clo­sure; for ther­monu­clear war, any prepa­ra­tions for clo­sure would have to be made be­fore the war starts.
...A crit­i­cal choice in the pro­cess of bring­ing a war to a suc­cess­ful close—or to the least dis­as­trous close—is whether to de­stroy or to pre­serve the op­pos­ing gov­ern­ment and its prin­ci­pal chan­nels of com­mand and com­mu­ni­ca­tion. If we man­age to de­stroy the op­pos­ing gov­ern­ment’s con­trol over its own armed forces, we may re­duce their mil­i­tary effec­tive­ness. At the same time, if we de­stroy the en­emy gov­ern­ment’s au­thor­ity over its armed forces, we may pre­clude any­one’s abil­ity to stop the war, to sur­ren­der, to ne­go­ti­ate an armistice, or to dis­man­tle the en­emy’s weapons.

His­tor­i­cal de­vel­op­ments in US nu­clear tar­get­ing policy

The United State’s nu­clear tar­get­ing policy has evolved from one of in­dis­crim­i­nate de­struc­tion of mil­i­tary and civilian tar­gets, in­clud­ing cities, to one that promises pro­por­tional re­tal­i­a­tion. While the pub­lic doc­u­ments, per­haps in­ten­tion­ally, do not make the US’s po­si­tion clear, their im­pli­ca­tion is that the United States would only tar­get cities in the event that their own cities were de­stroyed.The first nu­clear tar­get­ing plans ex­isted in the form of SIOP (Sin­gle In­te­grated Oper­a­tional Plan). This clas­sified doc­u­ment out­lined our nu­clear policy start­ing in 1961 un­til 2004, and now ex­ists in the form of the Oper­a­tions Plan (OPLAN). The first SIOP speci­fied all out tar­get­ing of both mil­i­tary tar­gets and pop­u­la­tion cen­ters, that is both coun­terforce and coun­ter­value tar­get­ing, in both first strike and sec­ond strike sce­nar­ios. Later SIOPs con­tained mul­ti­ple op­tions, in­clud­ing the op­tion to hold the bomb­ing of cities in re­serve.

This pa­per: “The Trump Ad­minis­tra­tion’s Nu­clear Pos­ture Re­view (NPR): In His­tor­i­cal Per­spec­tive” sum­ma­rizes how the Kennedy ad­minis­tra­tion be­gan to ad­vo­cate for a limited war sce­nario that spared cities:

Pres­i­dent Kennedy went so far as to en­dorse Sec­re­tary of Defense McNa­mara’s effort to get the Soviets to agree to a “no cities” nu­clear tar­get­ing rule, which McNa­mara and the Pres­i­dent soon aban­doned in the face of ob­jec­tions from NATO and the US Congress as well as the Krem­lin that the idea was to­tally un­re­al­is­tic. McNa­mara there­upon did a 180°turn to cham­pion a MAD arms limi­ta­tion (and re­ten­tion) pact with the Soviets – to pre­vent nu­clear war by guaran­tee­ing it will be mu­tu­ally suici­dal. The John­son ad­minis­tra­tion’s effort to ne­go­ti­ate such a treaty with Kosy­gin was aborted in 1968 by the Soviet Union’s bru­tal re­pres­sion of the re­formist Dubcek regime in Cze­choslo­vakia. McNa­mara con­tinued to work se­cretly with the mil­i­tary, how­ever, to en­large the menu in the SIOP (Sin­gle In­te­grated Oper­a­tional Plan) from which the pres­i­dent could se­lect limited and con­trol­led nu­clear re­sponses to a nu­clear at­tack – pre­serv­ing some pos­si­bil­ity of a nu­clear cease fire prior to Ar­maged­don.

Even though McNa­mara’s efforts to change nu­clear war plans to spare cities failed, his in­fluence led to changes in the SIOP that for the first time speci­fied a flex­ible re­sponse in nu­clear war plan­ning. Nixon would later make ad­di­tional changes to the SIOP, giv­ing the United States even more flex­i­bil­ity in nu­clear tar­get­ing sce­nar­ios. It is not clear whether the United States ever de­vel­oped a se­ri­ous “no cities” strat­egy dur­ing the Cold War, but it did at least lay the foun­da­tions for one.

For the first time in US his­tory, Pres­i­dent Obama’s ad­minis­tra­tion stated the US would not tar­get cities with nu­clear weapons. How­ever, this state­ment did not rule out es­ca­la­tion to coun­ter­value tar­get­ing in the midst of a nu­clear war, and is best in­ter­preted to mean that the US would only tar­get cities as a re­tal­i­a­tory mea­sure. From the same His­tor­i­cal Per­spec­tive pa­per:

Yet Obama, while con­ced­ing to this pre­sumed need to be pre­pared to ac­tu­ally use nu­clear weapons in ex­treme situ­a­tions, was not about to to­tally de­volve the plan­ning for such use onto the Pen­tagon…. he was adamant in his guidance to the mil­i­tary that if that cru­cial thresh­old ever had to be crossed, all op­er­a­tions had to be “con­sis­tent the fun­da­men­tal prin­ci­ples of the Law of Armed Con­flict. Ac­cord­ingly, plans will … ap­ply the prin­ci­ples of dis­tinc­tion and pro­por­tion­al­ity and seek to min­i­mize col­lat­eral dam­age to civilian pop­u­la­tions and civilian ob­jects. The United States will not in­ten­tion­ally tar­get civilian pop­u­la­tions or civilian ob­jects”(US Depart­ment of Defense, 2013 US Depart­ment of Defense. 2013. Re­port on Nu­clear Em­ploy­ment Strat­egy of the United States Speci­fied in Sec­tion 491 of 10 U.S.C. June 12.
The re­stric­tive rules of nu­clear en­gage­ment were trans­lated into the mil­i­tary’s doc­tri­nal lan­guage: “The new guidance,” elab­o­rated the Pen­tagon’s June 2013 Re­port on Nu­clear Em­ploy­ment Strat­egy, re­quires the United States to main­tain sig­nifi­cant coun­terforce ca­pa­bil­ities [jar­gon for di­rected at strate­gic weapon sys­tems] against po­ten­tial ad­ver­saries. The new guidance does not rely on a ‘counter-value’ or ‘min­i­mum de­ter­rence’ strat­egy [jar­gon for di­rected at cen­ters of pop­u­la­tion]. (US Depart­ment of Defense, 2013 US Depart­ment of Defense. 2013. Re­port on Nu­clear Em­ploy­ment Strat­egy of the United States Speci­fied in Sec­tion 491 of 10 U.S.C. June 12.
Did this mean that the United States was dis­card­ing its ul­ti­mate as­sured de­struc­tion threat for de­ter­ring nu­clear war? Clearly not. The guidance was care­fully drafted. Does not rely on is differ­ent from will not re­sort to. But more ex­plic­itly and openly than pre­vi­ously, the lan­guage in­di­cates that as­sured mas­sive de­struc­tion of the en­emy coun­try would be the very last re­sort in an already mas­sively es­ca­lat­ing nu­clear war, in which all the lesser op­tions had been ex­hausted and had failed to con­trol the vi­o­lence.

Pres­i­dent Trump’s nu­clear policy, as con­tained in the 2018 Nu­clear Pos­ture Re­view, differs in a num­ber of ways from Pres­i­dent Obama’s poli­cies, but doesn’t sub­stan­tially change the doc­trine of hold­ing the tar­get­ing of cities in re­serve.

If de­ter­rence fails, the ini­ti­a­tion and con­duct of nu­clear op­er­a­tions would ad­here to the law of armed con­flict and the Uniform Code of Mili­tary Jus­tice. The United States will strive to end any con­flict and re­store de­ter­rence at the low­est level of dam­age pos­si­ble for the United States, al­lies, and part­ners, and min­i­mize civilian dam­age to the ex­tent pos­si­ble con­sis­tent with achiev­ing ob­jec­tives.
Every U.S. ad­minis­tra­tion over the past six decades has called for flex­ible and limited U.S. nu­clear re­sponse op­tions, in part to sup­port the goal of reestab­lish­ing de­ter­rence fol­low­ing its pos­si­ble failure. This is not be­cause reestab­lish­ing de­ter­rence is cer­tain, but be­cause it may be achiev­able in some cases and con­tribute to limit­ing dam­age, to the ex­tent fea­si­ble, to the United States, al­lies, and part­ners.

Con­clu­sion and take­aways:

The US nu­clear tar­get­ing policy, in so much as pub­lic state­ments and doc­u­ments re­veal, has shifted sub­stan­tially from a policy of tar­get­ing cities by de­fault to a policy that leaves cities as re­serve tar­gets for full es­ca­la­tion sce­nar­ios. The US policy has never ruled out the pos­si­bil­ity of es­ca­la­tion to full coun­ter­value tar­get­ing and is un­likely to do so.

The maxim “no plan sur­vives con­tact with the en­emy” is es­pe­cially wor­ry­ing from the per­spec­tive of nu­clear war plan­ning. Dur­ing the early cold war years de­scribed in Daniel Ells­berg’s book, the mil­i­tary cul­ture pro­moted a ded­i­ca­tion to nu­clear readi­ness—so much so that officers vi­o­lated their own pro­to­cols to en­sure they could launch nu­clear weapons in a crisis. Readi­ness for re­tal­i­a­tion, es­pe­cially full coun­ter­value re­tal­i­a­tion, nat­u­rally trades off against risk of full es­ca­la­tion.

As both Her­man Kahn and Thomas Schel­ling made clear, com­mu­ni­ca­tion be­tween le­gi­t­i­mate au­thor­i­ties is es­sen­tial to the abil­ity to ne­go­ti­ate the end of a nu­clear con­flict. Yet the mil­i­tary value of dis­abling an en­emy’s nu­clear com­mand, con­trol, and com­mu­ni­ca­tions (NC3) ca­pa­bil­ities is large. This may be the biggest risk to cities; if Moscow and Wash­ing­ton are both de­stroyed in the early stages of nu­clear con­flict, then this could eas­ily es­ca­late to all out coun­ter­value tar­get­ing. It is im­por­tant not only that some com­mand struc­ture with the au­thor­ity to ne­go­ti­ate re­main in­tact on each side, but also that both par­ties can com­mu­ni­cate with each other and trust that the ad­ver­sary’s com­mand struc­ture is ac­tu­ally in­tact and ca­pa­ble of ne­go­ti­a­tion.

Fi­nally, all of this means very lit­tle if ei­ther of two po­ten­tial ad­ver­saries fail to make plans for a) re­frain­ing from ini­tial tar­get­ing of cities b) main­tain­ing NC3 ca­pa­bil­ities through an ini­tial nu­clear strike, and c) have the au­thor­ity and in­ten­tion to ne­go­ti­ate a peace & de-es­ca­la­tion. In­deed, pub­lic state­ments by Soviet lead­er­ship dur­ing the cold war sug­gested they had no in­ten­tion of spar­ing cities in a re­tal­i­a­tory strike, mak­ing any pos­si­ble US policy of grad­ual es­ca­la­tion po­ten­tially use­less. Ul­ti­mately, it’s im­por­tant to rec­og­nize here that not all nu­clear war sce­nar­ios have equal out­comes, and that both sides in a nu­clear con­flict could benefit greatly from en­gag­ing in strate­gic re­straint.

No comments.