if they die early.
6 They are relatively nonexistent in institutional DB
portfolios, with some exceptions, even though they offer a seemingly
ideal cash-flow profile for retirement planning.
7 The risk for the
insurance company, in offering these instruments, is that it needs to
diversify its risks (e.g., either cash flow or longevity) through signing
up a large diverse population and also in engaging in liability-hedging financial decisions much like those of an immunized institutional DB plan. However, unlike the DB plan that is back-stopped
by a corporate or governmental entity, an insurance company has to
support these annuity claims with its own capital (or hedge some
risk through re-insurance companies, which then raises the cost and
transfers the risk to the re-insurance company). Because the insurance company needs to be compensated for the use of its capital and
hedging incurs costs, annuities typically are expensive.
Reverse mortgages: Homes are not normally considered in typical
theoretical models of investments, but typically they are one of
individuals’ largest asset holdings and where savings often are
diverted during working life. However, given that one’s house is an
illiquid investment and does not generate positive cash flows but
instead requires negative cash flows, many are recommending the
use of more-intelligent reverse mortgages than the current offering
to finance retirement income (Merton 2015). Very simply, a typical
reverse mortgage requires the individual to turn over the ownership stake in a house to a counterparty at death. In return, the individual stays in the house and receives an income stream—much
like an annuity—until death, from the counterparty to the transaction. Depending on survivor clauses, etc., at death, the ownership
of the house is turned over to the counterparty. As with annuities,
the problem with reverse mortgages is that they are opaque and
expensive (Lucas 2015). In some instances, reverse mortgages pay
out lump-sum payments, and if investors spend this windfall carelessly, then they are at risk of not having the income stream they
require through retirement. In South Korea, these instruments are
referred to as “home pensions,” which removes the stigma associated with these instruments.
None of the public assets described above were designed to help
achieve retirement security; they serve the purpose of the issuer
(i.e., companies, government).
8 As a result, they do not provide a
hedge against retirement income. Annuities are costly, and as a
result, rarely found in retirement portfolios. Housing, which may
be a major store of asset value, is illiquid and difficult to convert
into the desired retirement income stream. This lays the first piece
of the foundation in the case for a new bond; namely, creating a
simple, liquid, publicly traded, low-cost hedge for retirement
income that effectively embeds the accumulation during working
life and the annuity payout at retirement.
How to Invest
Modern portfolio theory or MPT, which comprises Harry Markowitz’s
mean-variance optimization (MVO) and the Sharpe-Linter-Tobin-
Mossin capital asset pricing model (CAPM), is the backbone of
9 Retail and institutional investors globally use
In our LCH example, an investor using MVO would make assump-
tions about expected returns (using CAPM), volatilities, and cor-
relations of all investable assets, and try to generate 3.2-percent real
return annualized over the investment horizon for the lowest pos-
sible risk. The output of the model would be the investor’s strategic
asset allocation with target allocation levels to each asset, to which
the investor would rebalance periodically.
MPT and Challenges for Retail Investors
MPT and Liabilities
The first issue with using traditional MPT for retirement planning
is that in its simplest application, it ignores the uses of funds.
Rather than being held to earn the highest risk-adjusted return,
retirement assets are held to service some future liability. This
aspect of investing to service a future liability was recognized by
Merton (1973), focusing on a representative individual’s decision,
and also in Sharpe and Tint (1990), focusing on a pension fund’s
10 However, these versions of the model are not taught in
typical CFA or MBA classes, and even the liability-centric models
of asset allocation are used sparingly. This is most easily seen in
the large proportion of assets in portfolios that, rather than being
positively correlated to liabilities (which is what a liability-centric
approach would require), were negatively correlated to liabilities.
The two major crises in the 2000–2010 decade revealed the problems with such an approach and the funded status of all pension
funds declined dramatically.
The importance of liabilities in asset allocation decisions, or liability-driven investing (LDI), which makes the liability the reference point
in investment decisions, took on new relevance after the 2000–2002
tech bubble crash. However, the asset pricing implications of LDI
were addressed much later in a model called the relative asset pricing model (RAPM), first in Waring and Whitney (2009), but then
more robustly in Muralidhar et al. (2014).
11 , 12 The RAPM approach
demonstrates that investors should engage in three-fund separation;
namely, hold a liability-hedging portfolio, a cash portfolio, and a
market or return-seeking portfolio. Interestingly, this type of portfolio structuring is common among institutional DB investors governed by strict solvency regulations (e.g., Dutch pension funds,
U.S. corporate pension funds) because they are able to hold a
duration-matched liability-hedging portfolio. However, this same
approach cannot be adopted by an individual investor (say, a
twenty-five-year-old) or even an institutional DB portfolio with a
duration much greater than the current thirty-year bond. The liability-hedging portfolio cannot be created by even the most complex
financial engineering of market instruments (without adding