This work search experiment was conducted in Tacoma, Washington. It tested four work search approaches that ranged in philosophy from an ";exception-reporting"; approach with no specific work search directives or monitoring to an intensive reemployment assistance early in the unemployment spell. Approximately 10,000 new unemployment insurance (UI) claimants were randomly assigned to one of the four treatment groups during the July 1986 to August 1987 enrollment period. The results indicated that various work search policies have different and important consequences for the UI Trust Fund. For example, relative to the standard work search policy, the authors found that more intensive reemployment services treatment reduces UI payments on average of one-half a week, or, $70 per claimant. This reduction is considerably larger than the increased administrative costs associated with this treatment. It appears that the impact, reducing UI duration, of this treatment is primarily due to raising the cost of remaining on UI rather than enhanced job search abilities. However, there is no evidence that the relatively rapid reemployment of claimants in this group occurs at the cost of lower earnings or hourly wage rates. The authors also found that the exception-reporting approach significantly increases UI outlays relative to the standard work search approach by approximately 3.3 weeks and $265 per claimant. Given that the costs of monitoring work search activities are relatively modest these results taken together indicate that it would be prudent for states to maintain an active work search policy.