Does the elderly’s private pension ownership intensify aggregate equity demand? Empirical evidence in the US

ABSTRACT

In this paper, we investigate how the old generation income structure affects aggregate equity purchases, using Flows of Funds Accounts and Survey of Consumer Finances. Our results suggest that the risk aversion that increases with age could be modified to incorporate the old’s pension ownership. In particular, private pension income to elder households are related to increased aggregate equity purchases, even considering other pension and all other income. In this sense, private pensions are a ‘stepping-stone’ to increased equity investment in US households.

KEYWORDS:

• equity purchase
• private pension
• old generation
• income structure
• life expectancy

JEL CLASSIFICATIONS:

• C12
• C13
• G2
• G14

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

14 Generally, financial institutions, like banks and security firms, do not provide loans for young investors’ purchases of equity. We have estimated the younger generation’s growth variable as $growt{h}_{35}{t}^{+}$ in Equation (1)(1) $stoc{k}_t^{retired}={\text{β}}_0+{\text{β}}_1{r}_t+{\text{β}}_{2}vol{a}_t+{\text{β}}_{3}R{A}_t^{retired}+{\text{β}}_{3}incom{e}_t^{retired}+{\text{β}}_{3}savin{g}_t^{retired}+{ϵ}_t$(1) ′′. However, we do not find significant effects from the young generation’s growth: (1′′) $\begin{array}{rl}& stoc{k}_t={\text{β}}_0+{\text{β}}_1{r}_t+{\text{β}}_{2}vol{a}_t+{\text{β}}_{3}mutua{l}_t+{\text{β}}_{4}bon{d}_t+{\text{β}}_{5}growth\mathrm{\_}{35}_t^{+}+{\text{β}}_{6}growth\mathrm{\_}{45}_t^{+}+\\ & {\text{β}}_{7}growth\mathrm{\_}{55}_t^{+}+{\text{β}}_{8}growth\mathrm{\_}{65}_t^{+}+{\text{β}}_{9}SSe{c}_t+{\text{β}}_{10}Priv{P}_t+{\text{β}}_{11}Pub{P}_t+{ϵ}_t\end{array}$(1′′) where $growth\mathrm{\_}{35}_t^{+}$ is the growth rate of the 35 and older against the total population.

1 Elders indicate people 65 years and older.

2 Several papers have employed data sets from the Flows of Funds Accounts, such as Pesando (1974), Christiano et al. (1996), Baker and Wurgler (2000) and Boyer and Zheng (2009).

3 Corporate equities include common and preferred shares issued by domestic corporations and US purchases of shares issued by foreign corporations (including ADRs). Shares traded on the New York and American Stock Exchanges, and on the NASDAQ Stock Market account for most of the total. Investor groups include households, mutual funds, foreign investors, insurance companies, pension funds, and others (Federal Reserve, 2015). Meanwhile, since a bond reaches maturity, it repays its par, or face, value to an owner. Therefore, bond deals with non-matured bonds.

4 It should be noted that in estimation Equation (1)(1) $stoc{k}_t^{retired}={\text{β}}_0+{\text{β}}_1{r}_t+{\text{β}}_{2}vol{a}_t+{\text{β}}_{3}R{A}_t^{retired}+{\text{β}}_{3}incom{e}_t^{retired}+{\text{β}}_{3}savin{g}_t^{retired}+{ϵ}_t$(1) of the appendix, the over-55 and over-45 age cohorts’ effects on aggregate equity purchases are negative and not significant, respectively. Therefore, the 65 age cut-off is the most appropriate from the empirical perspective. For more details, please see the appendix.

5 Full social security benefits for retired workers are available at age 66 (or later), while reduced early-retirement social security benefits are available at age 62. Disabled workers are eligible for Social Security Disability Insurance at any age. Pension income includes employer-sponsored pensions (including military retired pay), 401(k) accounts and Keogh plans, individual retirement accounts, veterans’ pensions, periodic payments from annuities, and insurance policies.

6 We use the adjective ‘direct’ with households’ equity demand in order to distinguish this from indirect equity purchases via mutual funds or retirement plans.

7 Household net equity purchase averages almost zero (or −0.01%) if we exclude outlier years 1985 and 2007.

8 Retirement account growth indicates the asset’s growth of an individual retirement account (IRA), an account-type job pension (defined contribution, DC) including 401(k)s, or a defined benefit (DB) pension.

9 The life cycle hypothesis would link increased risk aversion of the old generation to higher bond purchases.

10 We use cohort growth rates instead of levels because the level variables are nonstationary.

11 We also include household’s savings growth in Equation (4)(4) $\begin{array}{rl}& stoc{k}_t={\text{β}}_0+{\text{β}}_1{r}_t+{\text{β}}_{2}vol{a}_t+{\text{β}}_{3}mutua{l}_t+{\text{β}}_{4}bon{d}_t+{\text{β}}_{5}growth\mathrm{\_}{55}_t^{+}+{\text{β}}_{6}growth\mathrm{\_}{65}_t^{+}\\ & +{\text{β}}_{7}SSe{c}_t+{\text{β}}_{8}Priv{P}_t+{\text{β}}_{9}Pub{P}_t+{\text{β}}_{10}lif{e}_t+{ϵ}_t\end{array}$(4) to control the wealth effect, as below. We confirm that private pension positively affects households’ equity investment. $\begin{array}{rl}stoc{k}_t=& -0.5425+0.0112r_t+{0.0264}^{\ast }vol{a}_t-{0.8533}^{\ast }mutua{l}_t-0.2705bon{d}_t-4.2611growt{h}_{{55}_t}^{+}\\ & \left(-0.8609\right)\left(0.7024\right)\left(1.9331\right)\left(-1.8490\right)\left(-1.3270\right)\left(-0.5676\right)\\ & +7.6681growt{h}_{{65}_t}^{+}+0.0582SSe{c}_t+{1.6644}^{\ast \ast \ast }Priv{P}_t+0.5010Pub{P}_t+0.0040lif{e}_t\\ & \left(1.6779\right)\left(0.2436\right)\left(3.3944\right)\left(1.2793\right)\left(0.4700\right)\\ & -0.1364saving\mathrm{\_}growt{h}_t+{ϵ}_t\\ & \left(-1.8050\right)\end{array}$where $stoc{k}_t$ is the normalised households’ net purchase of stocks, $r_t$ is the return on the S&P 500 Index with dividend reinvested, and $vol{a}_t$ is the logged difference between the highest and lowest S&P 500 Index for year, $mutua{l}_t$ is the normalised households’ net purchase of equity mutual funds, $bon{d}_t$ is the normalised households’ net purchase of bonds, $growth\mathrm{\_}{55}_t^{+}$ is the growth rate of those 55 and older against the total population, $growth\mathrm{\_}{65}_{\mathrm{t}}^{+}$ is the growth rate of the 65 and over against the total population, $SSe{c}_t$ is the ratio of social security income compared to the total income of those 65 and over, $Priv{P}_t$ is the ratio of private pension income relative to total income of those 65 and over, and $Pub{P}_t$ is the ratio of their public pension income relative to total income, $lif{e}_t$ is the life expectancy, and $savings\mathrm{\_}growt{h}_t$ is the savings growth of the households.

***^,* indicates statistical significance at the 1% and 10% levels, respectively (t-statistics are reported in parenthesis).

12 A good example of private pensions would be 401(k)s. For public pensions, the California Public Employees Retirement System (CalPERS) is a well-known example.

13 The part that mutual fund purchase (mutual) is orthogonal with net equity purchase (stock) would be the residuals (${\widehat{ϵ}}_t$) from the following estimation. $\begin{array}{rl}stoc{k}_t=& -{0.0207}^{\ast \ast \ast }-0.1370mutua{l}_t+{\widehat{ϵ}}_t\\ & \left(-5.8838\right)\left(-0.3536\right)\end{array}$The residuals from the above are regressed on life expectancy (life) in the following equation: $\begin{array}{rl}{\widehat{ϵ}}_t& =-0.0046+{1.5557}^{\ast }lif{e}_t+u_t\\ & \left(-0.9882\right)\left(1.8014\right)\end{array}$***^,* indicates statistical significance at the 1% and 10% levels, respectively (t-statistics are reported in parenthesis).