Portfolio and Consumption Choice for Long-Term Investors

最优资产组合的选择是金融经济学中的一个经典问题。在单周期的情况下，这个问题已经被很好地理解了，在重要的特殊情况下，最优投资组合权重的解析解是可用的。例如，当均值-方差分析合适时，最优投资组合权重是已知的资产回报的第一和第二时刻的函数。在多周期的情况下，这个问题就不那么容易处理了。在长期投资者没有劳动收入且具有对数效用或面临持续投资机会的特殊情况下，投资组合权重的显式解可用；但这些情况是容易处理的，因为它们简化为熟悉的单周期问题。萨缪尔森和默顿在20世纪60年代末的重要工作表明，在更一般的情况下，投资组合选择受到长期投资者对冲劳动收入和投资机会随机冲击的愿望的影响，但很难超越这一一般见解，找到作为状态变量函数的投资组合权重的明确解决方案。文献中的这种差距尤其令人遗憾，因为很明显，可处理的特殊情况并不能充分描述现实。对于大多数尚未退休的投资者来说，劳动收入比财务收入更重要。投资者具有对数效用的假设与在股票和债券市场中观察到的承担风险的高额回报不一致。投资机会不变的假设与实际利率和风险溢价随时间变化并与经济状况有关的大量证据不一致。只有解决了长期投资组合和消费选择问题，许多重要的政策问题才能得到解答。例如，美国财政部最近发行的指数化债券，首次提供了一种实际无风险的长期投资工具。这对保守的长期投资者的福利有何影响？财政部长认为，指数化债券的发行应该会增加储蓄。对于哪些偏好参数，这是正确的？影响有多大？社会保障制度改革的势头日益增强。许多私有化系统涉及将退休储蓄投资于平均回报率较高的风险资产。在什么条件下这能提高系统参与者的福利？本项目旨在通过推导跨期投资组合选择问题的几种变体的近似解析解来增加我们对长期投资组合选择的理解。每一种变体都集中在问题的一个方面：时变的实际利率、时变的风险溢价或劳动收入中的不可保险风险。

9809545 Campbell The choice of an optimal portfolio of assets is a classic problem in financial economics. In a single-period setting the problem is well understood, and analytical solutions for optimal portfolio weights are available in important special cases. When mean-variance analysis is appropriate, for example, optimal portfolio weights are known functions of the first and second moments of asset returns. In a multi-period setting the problem is far less tractable. Explicit solutions for portfolio weights are available in the special cases where a long-term investor has no labor income and either has log utility or faces constant investment opportunities; but these cases are tractable precisely because they reduce to the familiar single-period problem. The seminal work of Samuelson and Merton in the late 1960's showed that in more general cases portfolio choice is influenced by long-term investors' desire to hedge against random shocks to labor income and investment opportunities, but it has been very hard to go beyond this general insight to find explicit solutions for portfolio weights as functions of state variables. This gap in the literature is particularly unfortunate because it is clear that the tractable special cases do not adequately describe reality. Labor income is more important than financial income for most investors who are not yet retired. The assumption that investors have log utility is inconsistent with the large rewards for bearing risk that are observed in equity and bond markets. The assumption that investment opportunities are constant is inconsistent with a substantial body of evidence that real interest rates and risk premia move over time and are related to the state of the economy. A number of important policy questions can only be answered if one has a solution to the long-run portfolio and consumption choice problem. For example, the recent issue of indexed bonds by the US Treasury makes available for the first time a long-term investment vehicle that is riskless in real terms. How does this affect the welfare of conservative long-term investors? The Secretary of the Treasury has argued that the availability of indexed bonds should increase savings. For what preference parameters is this true? How large is the effect? There is growing momentum for reform of the Social Security system. Many privatized systems involve investing retirement savings in risky assets which offer a higher average rate of return. Under what conditions does this improve the welfare of system participants? This project aims to increase our understanding of long-run portfolio choice by deriving approximate analytical solutions to several variants of the intertemporal portfolio choice problem. Each variant concentrates on a single aspect of the problem: time-varying real interest rates, time-varying risk premia, or uninsurable risk in labor income.