6. FURTHER COMPLICATIONS AND ADJUSTMENTS

We also had to recognize that besides the innovator, other firms or research organizations may have invested resources (prior to the introduction of the innovation in question) in R&D and related innovative activities aimed at innovations of essentially the same kind as this one. Clearly, it is not easy to obtain data on the extent of such investments, but fortunately the difficulties seemed to be less than they might appear, for two reasons. First, only a limited number of organizations could reasonably have been expected to have been doing R&D in the relevant area, and if they had been devoting any substantial amount of resources to such work in the relevant time frame, it is inconceivable that the current executives of these organizations (and all their competitors) would have been unaware of it. Second, since these innovations occurred some time before our study, firms generally were quite willing to discuss whether they were carrying out work of the relevant kind at that time. Moreover, and this is particularly fortunate, our results are - quite insensitive to errors in the estimated investment in R&D carried out by others. Even if the true social research expenditures were ten times our estimate, and the true social development expenditures were double our estimate, the results change remarkably little. (To a considerable extent, this is due to the fact that R&D costs are often much less than the total cost of launching an innovation.)

Interviews with executives of the innovating firm as well as of other firms that could reasonably be expected to do (and be aware of) R&D of the relevant kind, indicated that, in most cases in our sample, no other firm or research organization was doing work aimed at roughly the same kind of innovation. Thus, in these cases, the private investment seemed to be a good approximation of the social investment. In the remaining cases, other firms or individuals were engaged in R&D aimed at the same kind of innovation. In the cases where this R&D was unsuccessful, we obtained as accurate an estimate as possible of the cost of this unsuccessful R&D, and added this figure to the R&D costs described in the previous section to get an estimate of the total social cost of the relevant R&D carried out by the innovator and others.

In the single case (among product innovations used by firms) where other firms or individuals were engaged (prior to the introduction of the innovation in question) in R&D aimed at roughly the same kind of innovation, and where this R&D was successful, it was necessary to recognize that the innovator's investment only resulted in the innovation's availability at an earlier point in time, not in all of the social benefits from the innovation up to the relevant time horizon. In other words, the proper comparison was between what would have occurred if the innovator had not carried out the innovation (but other firms were free to do so) and what in fact occurred. In this case, we obtained as accurate an appraisal as we could of the date when the innovation would have appeared if the innovator had not carried out the innovation, and we calculated the social benefits only during the period between the date when the innovation occurred and the date when it would have appeared if the innovator had not carried it out. But in this particular case a realistic (if perhaps somewhat conservative) estimate is that the second firm to produce the innovation would have come up with it when in fact it did so, regardless of whether the innovator preceded it or not.

For most innovations, these calculations were carried out for each year from the beginning of work on the innovation until 1973. Thus, our estimates of the social benefits were conservative, since all benefits after 1973 were ignored. But in some cases, this would have introduced a serious distortion, since the innovation was relatively new. In these cases, forecasts were made of the consumer surplus and the innovator's profits (adjusted for ­imitators' profits and for profits on older products) in each year up to 1980. These forecasts were based on firms' expectations concerning (P₁ - P₂), Q₂ , and the relevant profits in the next few years. They were intentionally very conservative, so whatever bias there was in the resulting rates of return should be downward.

Having made the calculations described in this and the previous section for each year, we had an estimate of the net social benefits (which may be positive or negative) from the innovation for each year. Then we could compute the internal social rate of return, the interest rate that makes the present value of the net social benefits equal to zero. In other words, it is the interest rate, I, that results in the following equality:

(2)

where B(t) is the net social benefit in year t, t is the first year in which the net social benefit is nonzero, and (t + n) is the last year in which the net social benefit is nonzero.

We computed the private rate of return from the innovator's investment in each innovation. To do so, we calculated the cash flow to the innovator from the innovation during each year. This calculation involved the subtraction of all costs incurred by the innovator to carry out, produce, and sell the innovation (including the allowance described in the previous section for R&D on uncommercialized projects) from the innovator's revenues from the innovation. Also, profits that the innovator would have earned on products displaced by the innovation had to be subtracted. The time period over which these computations were made was generally up to 1973, but in some cases (as in the case of the social rate of return) it extended to 1980. Again, the forecasts in the latter cases were decidedly conservative. The net private benefits in each year, like the Bs in equation (2), were deflated. The Consumer Price Index, which generally was used, is not ideal, but it seemed very unlikely that the results would be affected in an important way by this choice of a deflator.

Finally, one misconception concerning this study should be dispelled. Contrary to some assertions, the sample of innovations included in this study was not confirmed to "winners". We went to considerable trouble to get as representative a sample as possible. The innovations were chosen at random from those carried out recently by the cooperating firms. A very substantial number turned out to have low or negative private returns. (One interesting finding was that the social rate of return tended to be very high for them, as well as for the others). One of the contributions of this study, in our opinion, was that it included a much broader and more representative sample than any in the past. However, it was not, strictly speaking, a random sample since some firms refused to cooperate because they did not want to disclose such detailed data regarding their innovative activities.

Return to Table of Contents or go to next section.

Date created: June 15, 2006
Last updated: June 16, 2006

ATP website comments: webmaster-atp@nist.gov  / Technical ATP inquiries: InfoCoord.ATP@nist.gov.

NIST is an agency of the U.S. Commerce Department
Privacy policy / Security Notice / Accessibility Statement / Disclaimer / Freedom of Information Act (FOIA) /
No Fear Act Policy / NIST Information Quallity Standards / ExpectMore.gov (performance of federal programs)