By now, no one should be under the illusion that spontaneous order is synonymous with good outcomes. Perverse spontaneous orders abound. Charles Johnson has written about The Invisible Fist with regard to norms of discrimination. Scott Alexander called perverse spontaneous orders Moloch phenomena, which I follow in the title. Even Hayek himself, in The Road to Serfdom, explained the political process as a perverse spontaneous order, anticipating work by Buchanan and Wagner.
There are two distinctions, therefore, to be made. First there is the difference between cosmos and taxis; between spontaneous orders and planned orders.1 In addition to this, we must distinguish between benign orders and perverse orders. Whatever criterion we use to distinguish between benign and perverse, it is certain that there will be examples in all four quadrants.
The key difference between beneficial and perverse spontaneous orders will be the relevant feedback mechanisms. Whatever our criterion, we will want good results to be reinforced, and bad results to be selected against. There are cases, such as institutional evolution, where (as Doug North emphasizes, contra Hayek) feedback is extremely weak, so spontaneous order cannot be depended upon to produce beneficial results. There are other cases, to return to the example of the democratic political process, where in the absence of strong exogenous (constitutional) constraints, the feedback process is such that political success is most easily won by policies (such as targeted spending and certain types of regulation) which cannot practically be rolled back until they lead to a crisis.
To the extent that these feedback mechanisms are explicable and manipulable, intervention within a spontaneous order will be a straightforward exercise in mechanism design. The major caveat, of course, is that any intervention within one order is always taken from within another “supervening” order. Gadamer stressed that one can only question one piece of knowledge by taking other pieces for granted; one can never evaluate all one’s knowledge at once. Likewise, there is no “god’s eye” standpoint from which one can step outside all orders. In this sense, cosmos is more basic than taxis. There is no such thing as a “pure” act of will; there is only intervention of one order into another.
Buchanan’s model of constitutional political economy takes account of two – possibly three – layers of intervention: the rules of the market order are set by the political order, and the rules of the political order are set (ideally, anyway) by a real or hypothetical constitutional order. It would be a mistake, however, to regard the constitutional level as an act of pure will. When Buchanan or North emphasize that we “have no choice” but to consciously engineer the institutional order, they neglect the fact that such engineering takes place within a normological order. Social science is but one feedback mechanism influencing the norms that determine the constitutional process, and not necessarily a particularly strong one. The U.S. had the good fortune of crafting its political order at a time when the normological order was heavily influenced by liberal thought and thus conducive to a robust extended order. However much the institutional order has decayed since then, who can doubt that throwing open those questions again at this point in time with another constitutional convention would be utterly disastrous?
Care must be taken that any intervention into one order is done from within another order whose feedback mechanisms are at least as benign as those of the order being intervened in. And this is easier said than done. After hundreds of years, the feedback mechanisms of the market order are fairly well understood: the “system constraint” is tight, and the feedback mechanisms tend to reinforce socially beneficial outcomes. The political order is less well understood. After 60 or so years, some progress has been made, though the system constraint binds less tightly (hence the increased importance of individual wills within it), and perverse outcomes are frequently selected for. The normological order, on the other hand, is even still almost entirely obscure. There are no determinants of change which are both general and well-accepted.
This, then, must be the core of the argument against “rational constructivism”, the supposed supremacy of the will in crafting orders: one cannot explicate every order within which one stands, so it is hubristic to subject an order which we have good reason to think works reasonably well to the feedback mechanisms of another order which are ill-understood and more likely than not to be perverse.
Leave a Reply