By Adam Elkus
In Part I last month, I discussed the origins of the “failure is not an option” mode of strategic theory. Key to my conception was the idea of “failure is not an option” as a specific algorithm for ramming through a risky, controversial idea under highly complex, difficult, and uncertain circumstances.
If the possibility exists that anything less than a rigid formulation could lead to failure, the algorithm accepts the danger of a rigid design because its very rigidity ensures survival in a hostile or otherwise difficult environment. The logic of political survival famously holds that policy stability is strongest in a tighter coalition. The concept of “strategic essentialism” holds that subaltern groups should minimize individual differences at strategic points to present a united front to outsiders, despite the fact that such debate might serve an optimizing function over the long run.
The idea of implementing a massive and complex venture rapidly and decisively (with little room for error) is essentially just a rephrasing of the familiar the pre-World War I fear of losing a mobilization race. Under some circumstances, a nuclear balance could also degenerate into a “use them or lose them” dilemma in which a state risks the entire annihilation of its strategic forces and decision nodes in one murderous enemy salvo. There also seems to be — from Niccolo Machiavelli to Nathan Bedford Forrest – a general competitive heuristic that if you are to crush your enemies, you must strike as powerfully as you can and as quickly as you can. The heuristic is even repeated in the animal kingdom: queen bees famously kill their rivals upon emergence. But as the Germans discovered after the Schlieffen Plan and The Wire‘s Omar taunted, rapid execution and massive risk only pays off when it pays off. Fail and you run the risk of embroiling yourself in a quagmire that might have been avoided with more gradual and less rigidly planned execution.
The last aspect of the “failure is not an option” algorithm, “guarantee automatic consequences for failure” is perhaps the most interesting and complex. Whereas “failure is not an option” is to designed to optimize a wide variety of potential instances of the same general problem, the idea of automatic punishment is more ambiguous. Generally, the idea of automatic and unavoidable consequences for failure is intended to incentivize a “sink or swim” mentality due to the inevitably harsh punishment upon failure. But the distribution of the consequences for failure is not inherently specified by the “guarantee automatic consequences” instruction.
In a pure instance of the “guarantee automatic consequences for failure” instruction, defection from the plan is literally impossible. Cortes the conquistador scuttling his ships is the penultimate example. Either everyone succeeds together or they all die together. However, it is difficult to engineer such a circumstance due to the fact that one must close off any real possibility of escape. That said, a leader can also engineer this by forcing his subordinates to collectively cross a metaphorical Rubicon comprised of political, ethical, or sectarian norms of appropriateness. The classic heist movie cliche of the bank robbers being forced to kill the security guard or innocent witness is a cinematic example of this. All of your hands are dirty, therefore the group must succeed together or fail together.
The problem, however, is that the actual distribution of consequence in a high-risk endeavor is extremely variable. Consider a hypothetical (amalgamated) dictatorship at war that uses the threat of summary execution to optimize military performance. There are three possible implementations of the”failure is not an option” algorithm’s final component, each corresponding to a different distribution of lethal consequence.
The first implementation attaches a commissar unit to the back of each tactical formation. When it is time for the general offensive to commence, the tactical commander cries “death or glory, boys” and signals for the junior officers and NCOs to lead their men over the top. Anyone who falters is shot in the back by a special team of politically reliable riflemen and machine gun crews. The second implementation punishes only senior leaders. A general who fails to defend a critical city named after the Grand Sultan is visited by political officers that take him outside his improvised winter HQ to be shot in the head. A premier who oversees a losing war commits seppuku in his office with one hand while saluting his statue of V.I. Lenin with the other.
The third implementation is known as the “skin in the game” variant of the “failure is not an option” algorithm. Here, automatic punishment is equitably distributed. The war has been lost, and the dictatorship is forced to submit to what it considers to be humiliating peace terms. The political elites determine that no one party bears responsibility for the failure – a collective societal sickness has made the dictatorship weak and vulnerable. In order to better incentivize the decadent society to fight stronger when the dictatorship inevitably re-arms, it draws up a list of those to be executed that includes representative samples of every rank responsible. Corporals, junior officers, generals, cabinet ministers, and the Supreme Leader himself are all sent to the guillotine while cheering mobs chant “liberty, fraternity, equality!”
When considering American public policy, many analysts seem to believe that “skin in the game” is the best way to ensure optimal public policy outcomes. I will use the “skin in the game–conscription” variant to illustrate a sample argument:
“Skin in the game — conscription ” relies on the following assumptions:
- Imbalance in the distribution of potential consequences for failure is a major societal problem.
- Politicians feel free to wage indecisive, quagmire-like wars of convenience with ill-defined objectives.
- The burden on a few soldiers instead of the many is morally unfair and threatens collective cohesion in the larger society.
- Distributing potential consequence will deter politicians from waging unnecessary wars, rectify a moral error, and restrict wars fought to those of necessity and those with well-formulated political objectives.
However, as I will explain in Part III, the problem with these assumptions are that they all seem to raise the larger societal stakes. And that paradoxically seems to lead back to conditions when “failure is not an option” becomes an ideal forcing mechanism — which seems to create the very lopsided disasters that “skin in the game” at least partially is designed to prevent……