Academic journal article Journal of Risk and Insurance

Explaining Low Annuity Demand: An Optimal Portfolio Application to Japan

Academic journal article Journal of Risk and Insurance

Explaining Low Annuity Demand: An Optimal Portfolio Application to Japan

Article excerpt


Using an optimizing financial planning model in the tradition of Merton and Richard we explore how individuals should determine their life insurance and annuity choices, given uncertainty about investment returns and mortality. Both consumption and bequests appear as arguments in the individual's preference function. The model explicitly recognizes the existence of social security in retirement, and of loadings on insurance premiums, due to administration costs in the life insurance and annuities markets. The model sheds light on the reasons for the thinness of voluntary life annuity markets worldwide. The relative importance of pre-existing annuitization through social security, the role of bequests, and premium loadings are quantitatively assessed within a single optimizing framework. Results are presented for a model specification calibrated to Japan.


The factors influencing decisions to purchase life annuities have received increasing attention in recent years. Renewed interest has been spurred by increasing reliance on defined-contribution (DC) plans for private retirement provision globally, and by population aging, which is leading to greater focus on the characteristics of retirement income products and markets.

Both these developments are manifested in Japan. Now the world's oldest economy, Japan's population peaked in 2006. (1) Among a large number of pension reforms enacted over the last decade, and motivated in part by the rapid demographic transition being experienced there, is legislation enabling DC plans to be offered by Japanese firms. DC plans have quickly become popular, although contribution levels and account values are, thus far, low. Defined-benefit (DB) occupational pension plans are seriously underfunded and benefits can be renegotiated, sometimes for pensions in payment. Japan's social security system typically delivers less than 50 percent replacement at retirement, and large stocks of wealth are held in cash management accounts, yet the life annuity market mediates only a few thousand life annuity contracts each year. In Japan, therefore, the question of why voluntary annuity demand is so low is of special concern.

This article seeks to shed new light on the puzzle of thin voluntary life annuity markets. In sharp contrast to previous studies, it takes as its point of departure Merton's (1969, 1971) model of optimal lifetime asset allocation, and exploits Richard's (1975) theoretical extension to longevity insurance markets. The Merton-Richard framework is consistent with that used by Yaari (1965) to conclude that in the absence of a bequest motive, and given actuarially fair annuity markets, an optimizing consumer will annuitize all wealth. (2) Within this framework, we use stochastic control techniques to calculate optimal levels of life insurance and annuity purchases for Japanese households. We then use the model to investigate numerically the relative importance of three traditional explanations for low annuity demand: the bequest motive, the proportionate value of loadings, and the level of social security benefits. Although the results reported here have been calibrated to the case of Japan, we believe the qualitative findings to have general applicability and interest.

Our results indicate that the most important factor inhibiting annuity purchase is the bequest motive, followed by social security provision. Loadings are relatively unimportant.

The article is organized as follows. We begin in the first section with a discussion of annuity markets in Japan and the literature on thin annuity markets. The second section details, interprets and extends the Richard model--the tool we use to analyze optimal demand for annuities. The parameterization of the model is described in the third section. The fourth section reports the results of our modeling and analyzes the findings while the final section concludes. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.