Incomplete Contracts: An Empirical Approach

The strategic ambiguity hypothesis posits that when some aspects of performance are observable but not verifiable, the optimal contract is deliberately incomplete. I test this result for the first time. Because a direct test is infeasible, I derive an equivalent result: incompleteness is optimal when some terms are legally void. I test this using executive contracts from S&P 500 firms. I find that firms pay severance in discretionary installments to induce their executives to comply with noncompete agreements-but only in California, where noncompetes are void. Outside California, noncompetes are valid and these same firms pay non-discretionary severance upfront. I conclude that firms use strategic ambiguity to circumvent legal constraints.