The paper by Narayana R. Kocherlakota explores the implications of efficient risk sharing in a symmetric information environment without commitment. It examines whether and when imperfect diversification of consumption risk can be optimal in such settings. The author shows that if individuals are sufficiently patient, imperfect diversification is always sub-optimal in the long run, but if they are not, it can be optimal. The paper also demonstrates that the way history matters in efficient allocations in a symmetric-information/no-commitment environment can distinguish lack of commitment from other rationalizations of imperfect risk sharing, such as efficiency in the presence of asymmetric information. The analysis reveals that in efficient allocations, individual consumption is positively correlated with current and lagged individual income, conditional on aggregate consumption. This correlation is a sufficient statistic for the evolution of efficient allocations, making it possible to empirically distinguish between efficient allocations in symmetric-information/no-commitment and asymmetric-information/full-commitment environments. The paper concludes by discussing the implications of these findings for understanding consumption and income data.The paper by Narayana R. Kocherlakota explores the implications of efficient risk sharing in a symmetric information environment without commitment. It examines whether and when imperfect diversification of consumption risk can be optimal in such settings. The author shows that if individuals are sufficiently patient, imperfect diversification is always sub-optimal in the long run, but if they are not, it can be optimal. The paper also demonstrates that the way history matters in efficient allocations in a symmetric-information/no-commitment environment can distinguish lack of commitment from other rationalizations of imperfect risk sharing, such as efficiency in the presence of asymmetric information. The analysis reveals that in efficient allocations, individual consumption is positively correlated with current and lagged individual income, conditional on aggregate consumption. This correlation is a sufficient statistic for the evolution of efficient allocations, making it possible to empirically distinguish between efficient allocations in symmetric-information/no-commitment and asymmetric-information/full-commitment environments. The paper concludes by discussing the implications of these findings for understanding consumption and income data.