A Life-cycle Analysis of Retirement Savings and Portfolio Choices: Optimal Asset Allocation and Location with Taxable and Tax-Deferred Investment
MetadataShow full item record
The expansion of Defined Contribution (DC) pensions provides households with an opportunity to save and make investments in a tax-favored fashion. Participants in the DC plans can enjoy the tax benefit from their contributions and investment earnings, but also get affected by the pension designs, such as employer matching policies, investment options and illiquid constraint. In this study, I analyze theoretically and empirically the optimal saving and investment decisions of the households in those pension accounts, and show that the pension characteristics play important roles in explaining the households' behavior. There are three essays in the dissertation. The first essay attempts to explain the observed asset allocation and location decisions for households making taxable and tax-deferred investment. I incorporate employer matching policies and other pension account characteristics into a life-cycle model of optimal intertemporal consumption and portfolio choice, which includes a taxable saving account and a tax-deferred retirement saving account. The model is estimated using data from the Surveys of Consumer Finances (1992 to 2007), and the structural parameters are recovered by the Method of Simulated Moments. After modeling the features of U.S. pension system, the predicted policy rules are able to explain the observed portfolio patterns of American households, who are more likely to hold equities in the tax-deferred pension account. The estimates and results show that employer matching policy induces higher proportion of total wealth held in the pension account, so investors tend to reduce equity holdings in the taxable account for precautionary saving purpose and boost equity investments in the tax-deferred account to maintain an optimal portfolio mix. I find that a 10 percentage point increase in the estimated employer matching rate makes investors reduce the average equity proportion in the taxable account by 22 percentage point and boost those holdings in the pension account by 10 percentage point. In contrast, since the employer stock match exposes the households to a riskier situation in the pension account than the cash match, it causes households voluntarily to hold less equity in that account, resulting in an average decrease of 4 percentage point in the equity ownership and 3 percentage point in the conditional equity proportion. Moreover, the policy experiment reveals that a deletion of Social Security taxes and payments makes the pension account the only source of retirement income, so households tend to put a higher proportion of savings in the tax-deferred account, and they are likely to invest conservatively and hold about 25 percent more of pension wealth in relatively safe assets. The second essay is about company stock investment in 401(k) plans. Company stock investment in 401(k) pension plans has become an important but risky asset in retirement wealth. Previous studies on the determinants of company stock holdings focus on the past stock market performance of company stock, but ignore the characteristics of the retirement plans and individuals, such as company size, employer matches, other pension assets, and financial wealth information. In this study, using the data from Survey of Consumer Finances (SCF), I provide an empirical analysis of the factors that affect company stock holdings in 401(k) plans, by analyzing a broad list of company features, individual characteristics, and financial wealth information. My preferred estimates suggest that, different from general stocks which are sensitive to risk preference and total wealth, the decision of whether to hold company stock is more likely to be affected by the employer's characteristics and the availability of other investment opportunities. Individuals who work in larger companies and receive more employer matches in the retirement account are more likely to hold company stock in the retirement account, and they are less likely to hold company stocks when the wealth outside the pension account is large and they have other retirement accounts. In addition, I find that the company stock share in 401(k) account is decreasing with pension wealth and total net wealth, which indicates that less wealthy individuals are those who are more exposed to company stock risk. The third essay analyzes the impact of investment choice on savings in defined contribution pensions. The striking growth defined contribution (DC) pensions have vastly expanded the number of individuals with some discretion regarding their retirement savings. One of the factors that may affect saving decisions is investment choice: namely the ability of the participant to direct the investment of the assets in the pension account. In most studies, people who report that they have control over assets allocation in pension plans do not distinguish the assets between the participant contribution and the employer contribution, but it is common for the employer's contribution to be constrained--often to company stock. In this study, I use the Health and Retirement Study (HRS) to estimate the impacts of unconstrained and constrained investment choices on participant saving levels in DC Pensions. The estimates and results indicate that participants with investment choice contribute over 3 percentage points more of their earnings into the defined contribution plan than people without choice, and people constrained in their investment contribute about 3 percentage points less in their retirement saving account. In addition, I find that male and lower income participants are more likely to contribute in a self-directed saving account.