State subsidies to R & D or innovative investments in firms are organized in many different ways. Examples from the plethora of extant subsidy instruments are tax incentives, grants to researchers, project grants, loans, conditional loans, and grants with royalty rights. Very little is currently known about the effectiveness of these subsidy forms. <p> In this paper we compare the effectiveness of eight forms of subsidy for R & D projects. The comparison is based on a survey of Swedish R & D managers, including detailed information about 214 research projects or project proposals. In a first set of results we report managers' general judgments about the effectiveness of different subsidy forms. Second, R & D managers were asked to judge how each subsidy instrument would affect the firm's decision about the size of each project and whether to conduct it. This allows an estimate of how such additional R & D each policy might induce. <p> There are two main conclusions. First, general subsidies do not seem to induce much additional R & D for a given amount of subsidy. Second, among specific subsidies so called "stock option grants" seem to induce most R & D per subsidy krona. These are grants that give the state a right to recoup some of its funding by exercising a stock option if the firm's value rises rapidly. The main reason that the stock option grant performs well is not that the state can recoup some of its costs but rather that firms do not accept this subsidy for much of the research that they would have conducted even without subsidy.
