Portfolio construction implemented by institutional and individual investors comprises two levels of decisions.

LEVEL 1: ASSET ALLOCATION = investment strategy across asset classes Selection of asset classes among equities, bonds, cash, real-estate, commodities, FX, private-equity and other alternative investments. Fixing an investment range for each selected asset class.

LEVEL 2: SECURITY SELECTION = investment strategy within selected asset classes The two approaches are index-tracking and in-house selection, also called passive and active management. Nevertheless, a portfolio elaborated with one or the other approach reflects an investment choice, and there is fairly no passive choice. The main difference between the two techniques comes from the ability for some investors to replicate a portfolio's performance with a low tracking error, which only works with indices.

Defined benefit (DB) plans drive investment policy in their core portfolio so as to maintain their fundig ratio above a minimum level. In DB's satellite portfolio, as well as in defined contribution (DC) plans, investors focus on achieving top level risk-adjusted returns over a medium to long-term investment horizon.

To help tackle this important topic, a factor often underestimated needs to be taken into account: liquidity.
For instance, small capitalization stock prices can experience fast movements. Short-term volatility may be profitable to investors who will be able to trade these specific stocks because they don't face liquidity issues, which would not be the case for large pension funds and heavy mutual funds over the same short period of time. Funds returns are not always comparable since all funds do not work with the same operational constraints such as size and governance. Higher risk-adjusted returns could be expected from closed funds, but a small proportion of investors have access to the best ones. Therefore a challenge would consist in providing competitive risk-adusted returns over the long term that can be profitable to a wide population of investors. Asset allocation makes that possible. In particular, dynamic asset allocation has a great role to play.

Partly for liquidity reasons, huge investors such as DB and DC plans tend to focus more on asset allocation - which includes capitalization, sectorial and regional allocation - than on security selection. In this investment framework, asset allocation turns out to be the primary determinant in performance attribution. Xiong, Ibbotson, Idzorek, and Chen (2010) clarified this subject by breaking asset allocation total return into two factors: market return and asset allocation policy return in excess of the market return.

Two interesting articles:

A long-only asset allocation policy that sticks to stable weightings, on one hand benefits from a low portfolio turnover, but on the other hand bears a directionnal bias (called beta). In that case, asset allocation returns evolve in tandem with market cycles. This is a major risk because nobody knows in advance neither direction nor duration of future market cycles. It could affect DB plans' funding ratio as well as long-term returns in DC plans. Global diversification may help smooth returns, but there is often a preference for domestic assets because investors know them better and it caps FX risk.

Consider 3 Target Risk portfolios holding index-like baskets of US Large Cap equities and US Treasury bonds (all maturities) in USD. The equity/bond weightings are 30/70 for the conservative portfolio, 70/30 for the growth portfolio and 100/0 for a pure equity exposure.

Figure 1 reveals that over a long-term period, balanced portfolios comprised of a stable asset allocation (monthly rebalance and index-like security selection) are characterized by a symmetry in their extreme short-term total returns. The upside bar illustrates the highest one year-rolling total return, the downside bar the lowest one year-rolling total return, the dash illustrates the annualized total return over the whole period.

Figure 1. Extreme short-term returns and annualized long-term returns


Though securities' specific risk can be offset through diversification, the rule should also be applied across asset allocation strategies because there is a specific systematic risk in holding a single multi-asset allocation. As an example, a Target Risk equity/bond allocation generates extreme one-year total returns that are almost symmetrical. Many investors seek asymmetry in the risk/return trade-off rather than symmetry, as shown in figure 2. In DB and DC plans, because of the liquidity constraint, asymmetry may be achieved by undertaking dynamic asset allocation, primarily with highly-liquid equity and bond baskets (physical or synthetic). Efficiency of the dynamic asset allocation process is a key issue to which LongTermFocus brings a solution.

Figure 2. Symmetry and asymmetry in the risk/return trade-off

Diversification within asset classes is a good practice because it helps reduce specific risk versus systematic risk. The paragraph above pointed out the specific asset allocation risk in holding stable asset class weigthings: symmetry in the risk/return trade-off.

A tactical overlay might be added to the strategic asset allocation. In our view, TAA and DAA are substantially distinct. Nevertheless, a balanced portfolio would improve its efficiency by implementing a Level 1 diversification among asset allocation strategies in addition to a Level 2 diversification in security selection. Stable and dynamic asset allocations are truly complementary.

Strategic asset allocation (SAA) establishes long-term investment guidelines across a series of asset classes in order to match a set of risk limits and return expectations. In a DB plan's satellite portfolio or in a DC plan, implementing the multi-asset benchmark with static weightings incurs a specific asset allocation risk. This is why tactical asset allocation (TAA) and dynamic asset allocation (DAA) help improve SAA's risk/return characteristics over the long term.

TAA can trigger market transactions at any moment in time to try to benefit from perceived temporary imbalances in the value of eligible asset classes or subclasses. The investment range dedicated to TAA in each asset class can be fully utilized at once if indicated by the process. TAA is designed to opportunistically tilt the portfolio’s overall performance.

DAA also embraces a variety of investment strategies, some of them would rather refer to TAA. In our view, DAA stands between SAA and TAA. One DAA investment process might be sensitive to market cycles so changes in asset allocation are little influenced by day-to-day volatility. Policy decisions might occur at prescheduled dates, for example once a month or once a quarter. This frequency allows smoothing out short-term market fluctuations. Each asset allocation shift accounts for a portion of the available investment range. In case market conditions make it necessary, DAA cumulated shifts may significantly differ from SAA weightings during an extended period of time in order to fulfill the investors’ objectives. Therefore DAA provides efficient diversification. It contributes to reducing long-term correlation between the overall portfolio's returns and market returns, particularly in durable bear markets.

LongTermFocus' proprietary dynamic asset allocation process seeks to provide superior risk-adjusted returns (asymmetry in the risk/return trade-off) over the long run. Risk (resp. return) is measured by the lowest (resp. highest) one-year rolling return. The annualized total return over a long-term investment period stands for the center of symmetry. Figure 2 illustrates asymmetry and symmetry in returns.

LongTermFocus' DAA policy does not target a specific excess return on a given benchmark. It aims at participating in durable upside market trends while dampening impact of durable market downturns on its performance. Over medium and long-term periods, it is meaningful to draw up a comparison between LongTermFocus' DAA benchmarks and target risk portfolios. Figure 4 plots cumulative total returns of 3 portfolios since December 1989: LTF US Long-Term Dynamic Asset Allocation benchmark, a conservative target risk portfolio (30/70: 30% US Large Cap equities/70% US Government bonds) and a growth target risk portfolio (70/30).

Figure 4. Cumulative total returns since December 1989 with US equities/bonds in USD


LongTermFocus' DAA process excludes short positions. Its dynamic asset allocations are long-only, they comprise highly liquid assets like Large Cap equities and Treasury bonds. The process builds a series of portfolios invested in domestic assets in local currency: US equities/US bonds in USD, Euro equities/Euro bonds in EUR, UK equities/UK bonds in GBP. An investor decides how much to allocate to each region according to his own guidelines. FX risk is clearly identified. An other interesting point is that LongTermFocus' regional balanced portfolios do not hold the same asset allocation at the same moment. It is efficient to run regional balanced portfolios in parallel. This provides broader insight into international diversification.

Investors feel naturally less confident in downturns than in bullish markets because negative returns imply waiting during an undefined period of time to recover the portfolio's highest historical valuation. Investors' psychology is on the portfolio construction frontline such as the risk to concentrate buy orders on the peaks and sell orders on the dips. Human behavior fuels market evolutions through the interpretation made by each investor of financial and economic data.

To take this key human factor into consideration, LongTermFocus' DAA process is based on behavioral finance. It makes no prediction about market orientation or asset valuation. It translates into a dynamic asset allocation the changes in risk aversion embedded in the path of the portfolio's cumulative return. Though investors do not act in the same way, in the core part of a long-term portfolio, most of them share similar expectations for asymmetry in the risk/return trade-off over the long run. LongTermFocus has converted investors' natural vigilance into a quantitative and systematic decision-making process. Vigilance allows to allocate a large portion of the portfolio to equities in appropriate conditions. Figure 5 illustrates LongTermFocus' US asset allocation since December 1989.

Figure 5. LongTermFocus' US asset allocation since December 1989


Risk management is the very central topic in LongTermFocus' dynamic asset allocation process. Stress-tests are systematically carried to assess in advance the impact of severe market shortfalls on the portfolio's valuation. The process immunizes against market consensus. Memory of the highest watermark also contributes to describe investors' behavior. This approach differs from CPPI. In DAA, one-year drawdown better reflects risk than volatility because asset allocation changes over time, volatility is not constant.

LongTermFocus' DAA process modifies regional asset allocations once a month on the last business day. The investment range allocated to Large Cap equities accounts for 0% to 80% of the portfolio, the other part being exposed to Treasury bonds (all maturities). The typical monthly shift is between 0% to 10%. The in-sample period to test the process extends from 1989 to 2003. Results are out-of-sample since early 2004.

LongTermFocus' DAA systematic process builds benchmarks that can be considered as indices. It updates regional asset allocations only at pre-scheduled dates, with no additional decision taking place in between. Portfolios are well diversified.

DAA can be implemented in ETF or non-ETF mutual funds, as well as in other investment vehicles. Liquid and transparent underlying assets perfectly match constraints that institutional investors are usually coping with. Leveraged strategies could also be put in place. Bullet bonds linked to LongTermFocus' DAA process could challenge long-term swaps in hedging some defined benefit plans’ liabilities.

The process builds a series of portfolios invested in domestic assets in local currency (e.g. USD, EUR, GBP) leaving all liberty to an investor to set up his own mix based on his investment guidelines. A cross strategy consists in applying a domestic signal to an other region such as a US portfolio driven by the UK signal, or a Canadian portfolio driven by the US signal. Cross strategies would have the ability to invest in emerging markets.

LongTermFocus provides, to our knowledge, a unique series of regional long-only dynamic asset allocation benchmarks driven by behavioral finance, without involving economic prospective: key investment tools for a wide range of long-term investors.

LTF Long-Term Dynamic Asset Allocation benchmarks' returns displayed in this website are calculated including fees and expenses. Transaction costs are set at 5 bp for each basket of equities and bonds. Annual management fees are set at an indicative level of 70 bp. All dividends are reinvested. The two comparative Target Risk portfolios bear no costs. They are rebalanced every month on the last business day.

• Retirement schemes: defined benefit (DB) and defined contribution (DC) plans.
• Insurance companies: increasing capital's return, e.g. in the Solvency 2 funding requirement framework.
• Sovereign wealth funds.
• Individual long-term savings.