Incentives for Innovators: Contracts Within and Between Firms - The problem of designing appropriate innovation

Leads to muted incentives for both tasks (with a few additional assumptions) 14 Limitations of the Multitask Model Most for-profit firms do not look like the RMV! What about relational contracts to motivate attention to the second task (e.g., Azoulay 2004)? – “do the ‘right thing’ and you’ll be promoted” Static model – Appears to ignore the long term, multi-stage feature of innovation proj. [...] c2) – The value of learning about the exploratory path in order to enhance opportunities for success in the second period An exploration incentive scheme requires the principal to be able to commit to the worker across both periods (and not renege when there is a failure in the first period). [...] If there is only short-term contracting, a wage of 0 for success in the first period will induce shirking, and so it will be impossible to generate the learning required for efficient research path choice in the second period Monitoring and Feedback: If the principal observes first-period outcomes (but the agent does not), it is useful to provide such information in the case of an exploration. [...] 27 Measuring Creativity Creativity := Outcomes “in the Tail” – Number of papers in top quantiles of the citation distribution – Prizes (Nobel, Lasker, election to NAS, IoM…) – Grad Student/Postdoc Placement Creativity := Frequency of “Strike-outs” – Number of papers in the bottom quartile of the citation distribution Creativity as Process: Extent of “Branching Out” or Recombination (Weitzman. [...] in Top 5% of the Citation Distribution 28 Problems with the AGZM evidence Poor man’s identification strategy – selection on observables – combined with differencing Can’t distinguish the effect of $$ from the effect of longer time horizons/rich feedback/freedom to experiment Can’t distinguish between the incentives and sorting effects of HHMI appointments (best postdocs/grad students seek ou.