Because basic research is so unpredictable a mainly merit-based reward system may disincentivize employees to try new things – in such a system employees would prefer to work in areas of the company where the chances of success are higher (e.g. development as opposed to basic research).
Risks need to be borne by the laboratory or company and promotions and pay raises need to be given to researchers that are not overtly successful.
Of course, outstanding results need to be rewarded but polarization should be avoided – rewards can be given in the way of budget or project allocations.
Example: From an investigation of Japanese production systems.
Source: Koike, K. (1990) Aspects of excellence in Japan’s production system. Economic Eye, Vol. 11, No. 4, pp. 27-31.