There are diminishing returns on cyber escalation because stronger attacks are more likely to provoke disconnection.
Jon Lindsay and Erik Gartzke 14, Jon R. Lindsay is an assistant research scientist at the University of California Institute on Global Conflict and Cooperation and an assistant adjunct professor at the University of California, San Diego School of International Relations and Pacific Studies, AND Erik Gartzke is Professor and Director of cPASS at the Department of Political Science @ UCSD, Lindsay, Jon R., and Erik Gartzke. “Coercion through Cyberspace: The Stability-Instability Paradox Revisited.” The Power to Hurt: Coercion in the Modern World, Oct. 2014.
The combination of cross domain deterrence and voluntary connection to the internet gives rise to a variant of the classic stability-instability paradox. In Glenn Snyder’s original articulation of the paradox, mutually assured destruction could deter nuclear war. However, MAD was not credible for, and might even encourage, limited conventional war.56 Or as Robert Jervis puts it, “To the extent that the military balance is stable at the level of all-out nuclear war, it will become less stable at lower levels of violence.”57 Cyber capacity is a poor substitute for nuclear weapons, myths of paralysis notwithstanding, yet there is a similar logic constraining the distributions of harms which are possible via information technology. To extend this logic to the cyber domain, there are a variety of deterrent mechanisms contain the most disruptive types of cyber attacks yet fail to contain, and even enable, a wide variety of online espionage, subversion, symbolic protest, and criminal predation. In cyberspace we observe a rather stable damage contest (i.e., no “paralysis” and limited “disruption”) but a very unstable intelligence-counterintelligence contest (lots of “espionage” and “fraud” vying with efforts at “control” and “mobilization”). Thus the actors that have the ability to carry out highly destructive cyber attacks (mainly state actors for now) lack the motivation to attack; by contrast, these same actors as well as many different actors have both the ability and motivation to inflict irritant aggression with little fear of suffering consequences By and large, cyber options fill out the lower end of the conflict spectrum where deterrence is not as credible or reliable. The very few cases of physically disruptive cyber attack we do observe—mainly powerful states conducting covert action or battlefield support operations against militarily weaker opponents—have notably involved stronger actors who not only have the capacity to plan and conduct a sophisticated attack but also have the ability to deter retaliation against their use of cyber attack. This cyber variant of the stability-instability paradox has a slightly different logic, however. In the nuclear realm, actors cannot disconnect from the threatened harm, and this is what makes the threatened destruction both mutual and assured. When there are many missiles with many warheads, the chance of intercepting them on the ground through a disarming counterforce strike or in the air through ballistic missile defense with any confidence becomes vanishingly small. Not so in cyberspace, where connection to the internet or acceptance of connections through it is voluntary. There is “no forced entry in cyberspace” in Libicki’s phrase, and so the hundredth cyber attack against a closed vulnerability is as ineffective as the first. Attackers thus rely on deception to exploit vulnerabilities and ensure that they stay open. However, offensive deception can fail in the “fog of cyberwar” and defenders can be deceptive as well, both of which are more likely against high-reward targets (where cross domain deterrence also more credible). The need to preserve internet connections to facilitate ongoing and future deception as well as the need to preserve stealth to avoid the consequences of getting caught imposes discipline on attackers. Actors cannot enjoy the substantial benefits of interconnection without accepting some risk of exploitation (hacking to spy) and attack (hacking to disrupt). Thus the successful “lockout” of the internet, with advantage accruing exclusively to one political group or another, is not realistic. Moreover, because these harms share similar techniques, the observed abundance 39 of the former represents a latent potential for the later. The latent escalatory potential of even minor irritants leads to rampant fears of unrestrained catastrophe, to be sure. Yet this latent potential is difficult to harness for targeted coercion because the threat is self-effacing. Declared cyber threats that highlight the vulnerability to be exploited are readily mitigated. Instead, the ineradicable threat of cyber catastrophe (ineradicable as long as the internet continues to be useful) creates a general if diffuse deterrent effect among all parties who value their connection to the internet. No one who wants to make money on the internet really wants to have a cyberwar, and this includes states as well as criminals. Which types of actors are most able to benefit through internet coercion and which are most vulnerable to coercion? Large powers like the U.S. are highly dependent on the internet but also highly skilled at inflicting harm, both through cyber and traditional military force. Poor powers across the digital divide may have little vulnerability at all, while medium powers may have vulnerability but lack a range of forces to deter attacks. This might imply a “curvature” to the utility of cyber coercion. Big-capable countries are vulnerable to cyber harm but can deter through other military instruments. Poor states are not vulnerable. It may be the prosperous small or digitally developing who are in trouble, since they cannot credibly deter and have high dependence on the internet. The information revolution is often thought to be a boon to non-state actors, and indeed it is, but mainly in the irritant class of cyber operations. Moreover, the increasing ubiquity and sophistication of information technologies can be expected to have something of a democratizing effect on intelligence and counterintelligence techniques whereby firms and citizens will have access to and be concerned about the types of things that were historically the purview of obscure state intelligence agencies. However, it would be a mistake to use the increasing ferment of low-intensity information contests to infer the shape of higher 40 intensity activity. On the contrary, the traditional logic of war will continue to dominate the expression of cyber aggression. Because threatened internet harms depend on voluntary connections in the first place, and as many actors have alternative means to inflict (cross domain) harm in retaliation, the coercive utility of cyberspace is actually somewhat limited. At the same time an ever increasing variety irritants and more temperamental adjuncts becomes available for global political interaction. The “net” result is that opponents have strong incentives to impose costs via the internet but also to keep those costs low enough to preserve interconnection and avoid retaliation. Therefore, contests in damage will remain relatively stable while contests in intelligence will be increasingly unstable. The human-built world is becoming more complex, to be sure, but it is not necessarily more dangerous. As long as it is desirable to connect to the internet tomorrow, there will be only limited harm via the internet today.