Discussion of Michelacci and Schivardi Francesco Caselli Four

Discussion of Michelacci and Schivardi Francesco Caselli

Four Questions on Intepretation

Theoretical Motivation Productivity (Endogenous) Volatility of returns in high return projects (exogenous)

Theoretical Motivation Productivity (Endogenous) With good risk-sharing Opportunities With poor risk-sharing Opportunities Volatility of returns in high return projects (exogenous)

Empirical Result • Negative coefficent on (Risk-Sharing Opportunity) X (Volatitlity) • Consistent with theoretical framework …

… but also consistent with Productivity (Endogenous) With good risk-sharing Opportunities With poor risk-sharing Opportunities Volatility of returns in high return projects (exogenous)

or with Productivity (Endogenous) With good risk-sharing Opportunities With poor risk-sharing Opportunities Volatility of returns in high return projects (exogenous)

or even with Productivity (Endogenous) With good risk-sharing Opportunities With poor risk-sharing Opportunities Volatility of returns in high return projects (exogenous)

Therefore • I am not yet convinced that Idiosyncratic risk discourages entrepreneurial activity and hinders growth (from abstract) • • • Finding seems to concern the role of riskdiversification opportunities, not risk per se Does this call for a change of title? Role of risk remains an open question

From levels to growth rates • Theoretical argument framed in temrs of levels • Empirics in terms of growth rates • Transition is not entirely trivial

From project returns to size of investments • Theory: investment size constant, but different returns • Empirics: results driven partially by amounts invested • Reconciliation: high-return projects more capital intensive? • That would be interesting

Family firms as (inverse) measure of risk-sharing opportunities • Idea: concentrated ownership results from lack of diversification opportunities • But: many other interpretations possible • E. g. suppose family firms are intrinsically more risk averse. Then (family X risk) interaction would be negative • Not sure that instrumenting with demographic shocks helps: it isolates variation in family ownership not due to lack of risk-sharing opportunities. Seems to be the opposite of what you want

Blonde-Blue-Eye Dummy • Identification assumption: Dnk, Nor, Swe, Fin have same coefficients in risk regression • Next year: paper by a Scandinavian economist assuming that Portugal, Spain, Italy and Greece have same coefficients in some regression

Conclusion • Should paper be recast in terms of effects of risk sharing opportunities, rather than risk per se? • Tighter link between theoretical arguments and empirical estimates (levels v. growth, productivity v. investment) • Can we think of a more convincing measure of risk-sharing opportunities?