Asset Allocation in China: Crafting the Optimal Allocation Strategy, Episode IX

4.8 Rebalancing the Portfolio

One of the great conveniences of simply allocating using the weights of the world market is that rebalancing is basically unnecessary.  If a country's market performs especially well, the investor's allocation to that country will grow, but so will that country's weight in the world market, so the investor continues to maintain weights equal to, or at least very close to, those of the global market.

However, this is not the case for weights selected, partly or fully, by an optimizer based on opinions that differ from the market's.  Such a portfolio can move away from the optimum, and in more than one way.  First, the performance of assets changes investor weights, even if the optimizer's opinion does not change.  So, for example, if the optimizer suggests 15% in U.S. stocks, and the U.S.  stock market rises 20% while the overall portfolio is only up 10%, then the investor will find himself over-allocated to the U.S.  Alternatively, as the world changes the assumptions used in the optimization can change, so that rebalancing is wise even if the portfolio weights have not moved much -- the investor would be holding a 2016-optimal portfolio in a 2017 world.

Of course in actuality both these things occur, and indeed they occur every day.  But rebalancing every day would be too costly, in effort and transaction costs.  We recommend the following rules of thumb:

- Rebalance anytime there is a truly major market event -- say, if asset classes making up 20% or more of the portfolio move 20% or more in a short time.  Thus after a market crash or a rapid appreciation, it's worth taking a look to see if it's wise to move back toward the old weights, or whether new weights should be computed and implemented.

- Rebalance annually even if there hasn't been a major event.  Small shifts can add up, and often expected returns move opposite to prices.  For example., if interest rates rise substantially, the value of an investor's bond portfolio will likely have fallen quite a bit, but with bond yields higher, bond expected returns are probably higher and so the optimizer may conclude it's wise to hold more, not less, in bonds.  So even if bonds only fell modestly as a fraction of the portfolio, and the desired allocation only rose modestly, since the directions are opposite it may be that the portfolio is now fairly far from the optimal spot.

- As to the size of reallocations in the rebalance, there is no way to state a general principle.  Often no, or very modest, shifts will be necessary.  On occasion best practices will suggest major changes to the portfolio -- the world can change a lot in a year.  Of course the factors noted in the previous section -- taxes, transaction costs, etc. -- should be accounted for in the annual rebalance as well as in the initial.  The advice of a skilled financial advisor is likely to be invaluable in dealing with these complexities.

4.9 The Impact of Diversification Benefits is Large

The ability to earn a significantly higher rate of return with the same risk budget has an extremely large effect on household wealth over time.  This is because of the power of compound returns -- an extra one or two percent per year, cumulated over 30 years, makes an enormous difference in the level of assets that households will end up with. 

Exhibit 20 below demonstrates how large the benefits of creating a diversified portfolio can be. We take for this illustration the example of a typical Chinese investor with 10 million RMB of wealth and moderate risk tolerance.  For a standard set of model assumptions we find that the model selects a portfolio targeted at delivering 10 percent volatility can be expected to deliver approximately 5 percent real return on average.  Compare this optimized portfolio to the portfolio that a typical Chinese investor might be likely to select: such a portfolio might consist of a handful of Chinese stocks, some bonds, and a significant allocation to cash.  A portfolio of that type that had 10% volatility, because it would not benefit from global asset class diversification, would only be expected to return about 2.5% real per year.  Although the difference of 2.5% per year appears modest, it adds up to tremendous impact overtime. Even after adjusting for inflation, the diversified portfolio, with no additional risk, is expected by the optimizer to deliver an additional 22 million RMB or so of purchasing power. The investor in this hypothetical example would have almost double the wealth without having taken on any additional risk. This is the power of diversification.

EXHIBIT 20 - CUMULATIVE WEALTH GROWTH FOR 10 MILLION RMB INVESTOR

CONCLUSION

This paper makes three major points:

1) The theory of finance shows that investors can benefit from diversifying both across asset classes and globally. Most investors fail to take advantage of these opportunities, and therefore create portfolios that have higher risk and lower expected returns than what they could achieve were they to invest optimally. However, the most sophisticated individuals and institutions and develop markets take full of vantage of this opportunity, and best practices are to diversify in these ways.

2) Historically, Chinese investors have invested primarily domestically. In great part this was because of restrictions on international investing that left few opportunities to fully diversify. In recent years, these restrictions have become much less severe, as changes in government policy have created numerous mechanisms for achieving global diversification. Some sophisticated investors have taken advantage of these opportunities, but it is still typical for many investors to be undiversified, perhaps in part out of habit.

3) In the final section of the paper, we layout theoretically optimal portfolios for constraint investors of various types. We look at nine types of investors who have three different wealth levels and three different levels of risk aversion. Each of these nine investors faces different constraints on their ability to invest in different asset classes and countries. For each of these nine investor types, we show the portfolio that is optimal given the assumptions of the model, including assumed constraints. We then show the benefits that a crew to investors who moved from the typical investment allocation for investors of that type to the portfolio deemed optimal by the model.

It is important to recognize that each investor’s circumstances are different and therefore that each investor’s optimal investment policy will differ based on their personal circumstances. Moreover, the results of our model are highly contingent on the model assumptions, and by their nature such assumptions are debatable -- for example, we use for all investors HMC's assumed expected returns for each asset class, but HMC may have the ability to execute at lower transaction costs or with superior manager selection ability that few investors can replicate.  And these are just some among the many important simplifications a model like ours requires.  Such models, and results like those delivered in this paper, are, we believe, and important starting point for investors to attempt to maximize investment performance and minimize risk. But they are only a starting point – final investment decisions should be made by investors after careful consideration of all factors involved, and with the assistance of wise financial counsel.