A voluntary standards body can compel nobody. It publishes a document fixing the pitch of a screw thread or the dimensions of a pallet, and any firm is free to ignore it; almost none do. The usual explanation of that obedience is what may be called the neutral-forum account. Firms need to converge on some common design; several designs would serve about equally well; and what the body supplies is a place to converge that none of the converging firms owns. On that account neutrality is the whole of the mechanism, since a body suspected of favouring one member could not coordinate the others, who would decline to be coordinated.
Two things must be granted to the account before it is set aside. It describes the greenfield case accurately: where a technology is new and nobody has yet built anything, committees do appear to choose on merit, and the design chosen is frequently not the one the largest participant brought. And neutrality is not decorative. Bodies seen to have delivered a result to a dominant member have lost their memberships, and a body without members publishes documents nobody reads.
The greenfield case, however, is the rare one. In a mature industry the committee convenes long after the investment has been made. Participants arrive having tooled their plants to a particular tolerance, trained their fitters on it, and sold equipment that assumes it. What is in dispute is then not which design is best but whose plant must be re-tooled, and what the committee settles is the distribution of that expense, which may fall on a single participant or be parcelled out among several. Two observations fit this reading and not the other: proceedings run longest where prior investment is deepest rather than where the engineering is most intricate, and bodies that abandoned consensus for majority voting did not thereby publish any faster.
It would be too much to say that the committee merely records a bargain struck elsewhere. The record does not support that, and the outcome is often a design no participant brought. What the two observations support is narrower: that the committee's business is the allocation of conversion costs, conducted in the vocabulary of engineering. Neutrality, which the first account made the centre of the mechanism, is then doing a different job from the one assigned to it. It is not what makes convergence possible. It is what makes the losing party's loss bearable, a cost imposed by a procedure one entered voluntarily being easier to absorb than a cost imposed by a rival.
The consequence is a change in what one should examine in assessing such a body. Impartiality of procedure is not the first test. The first is whether a participant who loses can be held to the outcome -- whether withdrawal is expensive, whether the standard finds its way into the contracts of firms that never sat on the committee. A body whose losers can simply walk away has allocated nothing.
