This issue features two original research articles, three issues & policy articles and a book review section. The lead article is by Geoffrey Kingston and Susan Thorp (University of New South Wales, Australia) and addresses the issue of Annuitization and asset allocation with HARA utility. One of the puzzles in retirement economics is why individuals do not choose to purchase annuities and Kingston and Thorp explore in detail a real options model in which individual preferences obey the broad class of hyperbolic absolute risk aversion utility. The theory of Real Options argues that people might want to delay annuitisation at relatively younger ages because the price of life annuities might improve and annuitisation is irreversible. However, Kingston and Thorp show that the implications of a Real Options approach varies across individuals considerably. For example, when individuals have a desired consumption floor as opposed to CRRA preferences, they are more likely to want to purchase annuities earlier than later. It would be interesting to see empirical tests done in this area to test the relatively new Real Options theory as it applies to irreversible personal financial decisions, such as annuitization.

A new explanation for the well-known reluctance of retirees to buy life annuities is due to Milevsky and Young (2002, 2003): Since the decision to purchase longevity insurance is largely irreversible, in uncertain environments a real option to delay annuitization (RODA) generally has value. Milevsky and Young analytically identify and numerically estimate the RODA in a setting of constant relative risk aversion. This paper presents an extension to the case of HARA (or GLUM) preferences, the simplest representation of a consumption habit. The precise date of annuitization can no longer be ascertained with certainty in advance. This paper derives an approximation whereby the agent precommits. The effect of increasing the subsistence consumption rate on the timing of annuity purchase is similar to the effect of increasing the curvature parameter of the utility function. As in the CRRA case studied by Milevsky and Young, delayed annuitization is associated with optimistic predictions of the Sharpe ratio and divergence between annuity purchaser and provider predictions of mortality.

In this paper, we assess the welfare costs and gains of different scenarios of pension reforms in France, using a life-cycle model including various sources of heterogeneity and distinguishing between socio-occupational groups. The pension reforms considered combine features regarding the generosity of the pension system as well as features regarding the financing schemes: PAYG, the build-up of a temporary fund and that of a permanent one. We focus on both macro and distributional issues. It appears that (i) a considerable increase in savings is to be expected, even in the case where pensions remain generous, (ii) a considerable crowding-out effect would occur in the case of the constitution of a fund trust, (iii) reducing the generosity of pension seems relatively more beneficial to low-income low-life expectancy agents, while (iv) postponing the legal retirement age benefits relatively more high-income high-life expectancy agents.

The foreign property rule (FPR) requires that no more than 30% of the assets held in tax-deferred retirement savings accounts be foreign property. The FPR is supposed to increase the value of the dollar and reduce its volatility and decrease the cost of capital and promote investment in Canada as well as decrease the extent of inequality inherent in these plans. On the basis of evidence from the easing of this regulation from 20% to 30% over the period 2001/02 we find that it accomplishes none of these objectives. There was no measurable impact on the exchange rate predicted from the Bank of Canada's forecasting equation; the capital outflow from the change amounted to no more than two days trade in the foreign exchange market over the period 2000/01; Canada's equity markets did significantly better internationally when the FPR was eased than in the prior two-year period. Finally, closer inspection reveals that the rule exacerbates income inequality by imposing the largest costs on lower middle-income groups. We estimate that the increase in the FPR from 20% to 30% increased Canadians expected income by between 500 million and one billion dollars annually by permitting greater portfolio diversification. The complete removal of the FPR would increase income by an estimated additional 1.5 billion to 3 billion dollars annually.

Preparations are underway to revise national accounting to implement actuarial recording of pension liabilities for corporations and government as an employer. This paper extends this to unfunded public pensions with the help of ‘implicit tax’ in pension contributions. The clearest advantages of the revision appear in situations where pension liabilities are shifted from the corporate sector to government, and where part of the public pension system is privatized. The proposed revision raises public debt and deficit to new orders of magnitude. The paper provides a framework for setting the debt and deficit targets under both current and proposed definitions.

This paper takes a look at the modeling side of pension fund management. It is based on interviews with pension fund managers, regulators, consultants, and academics in four countries – the Netherlands, Switzerland, the United Kingdom, and the United States. The objective was to understand, through the experience of market participants, the role of modeling in managing defined-benefit pension funds. The 28 defined-benefit pension funds participating in the study have a total of €334 billion ($436 billion) assets under management. The findings of our study show that modeling is now considered an indispensable tool by many market participants. The need to manage the risk inherent in defined-benefit pension plans is the key motivation behind the growing use of modeling. In the Netherlands, for example, where private-sector plans did not experience serious underfunding problems after the 2000 market crash, the use of modeling is widespread and well-integrated in the decision-making process. Dutch regulators have recently mandated a risk-based approach and specified broad principles of sound modeling, including the marking to market of assets and liabilities.