Get started today!

Applying Is As Easy As 1-2-3!



Two different sorts of constraints are potentially relevant. First, a household may be unable to borrow at the riskless interest rate to finance equity investments (Con-stantinides, Donaldson, and Mehra 1998). Second, a household may face fixed costs of equity market participation; if these fixed costs exceed the benefit of participation, the household may hold no equities (Abel 1998). These constraints may affect different households differently. The first constraint is particularly likely to bind on a household whose unconstrained optimal equity position is particularly large, while the second constraint is particularly likely to bind on a poor household with little total wealth. These different sorts of households may be differentially affected by a Social Security reform that alters portfolio constraints.
Clear understanding of these issues requires a well developed normative theory of optimal portfolio choice over the life cycle. Until very recently, however, theoretical work on this subject lagged far behind the familiar theory of single-period optimal portfolio choice. Samuelson (1969) and Merton (1969, 1971) showed that there are conditions under which long-lived investors choose the same portfolios as single-period investors, so that the investment horizon is irrelevant; unfortunately these conditions are highly restrictive in that they include power utility, returns on safe and risky investments that are independently and identically distributed over time, and, most disturbing of all, the absence of labor income.
In the last few years economists have returned to this topic and have begun to explore long-run portfolio choice when these restrictions are relaxed. Ross (1997) shows how horizon effects can emerge from more general models of preferences. Brennan and Xia (1998), Campbell and Viceira (1998), and Wachter (1998b) consider changes over time in the riskless real interest rate, while Balduzzi and Lynch (1997), Bar-beris (1999), Brandt (1999), Campbell and Viceira (1999), Kim and Omberg (1996), Samuelson (1991), and Wachter (1998a) consider changes over time in the equity premium, and Brennan, Schwartz, and Lagnado (1997) and Liu (1998) allow a more complex pattern of variation in both the real interest rate and the equity premium. The effect of labor income on portfolio choice has been explored by Bertaut and Haliassos (1997), Bodie, Merton, and Samuelson (1991), Cocco, Gomes, and Maenhout (henceforth CGM, 1998), Gakidis (1997), Heaton and Lucas (1997a), Storesletten, Telmer, and Yaron (1998a), and Viceira (1997), among others. In this paper we concentrate on the effect of labor income. The theoretical literature on this subject can be loosely summarized as follows. A household with labor income has an implicit holding of a nontradable asset, human capital, that represents a claim to the stream of future labor income. This nontradable asset can “crowd out” explicit asset holdings. If labor income is literally riskless, then riskless asset holdings are strongly crowded out and the household will tilt its portfolio strongly towards risky assets (Bodie, Merton, and Samuelson 1991).