Thursday, 8 October 2009

Portfolio Rebalancing - What, Why and How

What is Portfolio Rebalancing?
Portfolio rebalancing is the purchase and/or sale of investments in order to bring the various types of holdings (e.g. Canadian equities, Canadian bonds, REITs, foreign equities, commodities funds etc), known as asset classes, back to the intended target allocation (see Asset Allocation post for how to create a target allocation).

Why Rebalance?
Different classes of investments go in different directions over periods of weeks, months or years. Some go up while others go down or sideways for a time. As a result, the original target percentages allocated to each class can get far out of whack. For instance, if equities were initially set at 60% of a portfolio and bonds at 40% but equities grow a lot while bonds remain stable, equities could rise to 70% of the portfolio. Suddenly the investor has a riskier portfolio with more allocation in the volatile equity class. Rebalancing is thus a method of risk control.

Rebalancing can also be thought of as selling high and buying low. Renowned financial author William Bernstein has written about a rebalancing bonus, whereby a disciplined strategy of rebalancing in most cases will provide higher returns than a simple buy and hold portfolio with the same asset classes. Rebalancing is most worthwhile and effective when the portfolio includes volatile asset classes that behave very differently from each other (have low correlation).

What are the choices for a Disciplined Rebalancing Strategy?
Portfolio allocations are changing every second that markets are open, so some leeway for movement is necessary. There are two primary methods for rebalancing, both of which are quite mechanical and remove the need for an investor to make arbitrary decisions about how much and when to rebalance.
Various researchers have compared the two methods and tested the numerical trigger points, out of which the rules below summarize a more or less middle position. There is no definitive best method and both methods below will be effective since data sets used to test have varied considerably and most importantly, the future will not be exactly like the past - returns, volatility and co-movements of asset classes will differ.

1) Rebalance at fixed regular intervals, such as once per year - other intervals like per month, per quarter or bi-annually are possible. Or,
2) Rebalance when an asset class deviates from target by more than 5% - this works best in a portfolio with four or five asset classes with individual allocations of 20% or more.

When a portfolio grows and an investor subdivides it into fine-grained asset classes, such as allocations of 10% or less per class, the 5% rule might be too high, such that rebalancing never occurs, which would mean never taking advantage of the rebalancing bonus. Unfortunately no research seems to exist to suggest what works best with many small asset classes, so I've come up with what makes sense based on the trade-off factors below. Rebalance when a class is more than half its weight above or below target - e.g. a class with a 5% target allocation only gets rebalanced when it reaches 7.5% or 2.5% of the total portfolio.

Trade-off Factors to Consider in Choosing a Method:
  • Costs - trading commissions, fund redemption or switching fees and capital gains taxes (in taxable accounts) rise with more frequent rebalancing and reduce portfolio return
  • Effort - the percent deviation rule requires constant monitoring of the portfolio. The yearly interval may be combined with an annual financial review.
  • Volatility - portfolios containing more volatile investments like real estate, commodities and precious metals can exceed allocation boundaries much more quickly. When markets go haywire, a set interval may miss an opportunity to "sell high and buy low".
  • Trends - markets and asset classes often have streaks of winning or losing years, which means that frequent rebalancing will either sell out too soon and miss part of the upside, or buy in too early on the downside. Some researchers have concluded that rebalancing only every four years produced the best long-term return.
How to Rebalance
Few portfolios remain static nor do they magically appear wholly formed. Almost all portfolios are either being built up with savings or drawn down by withdrawals. Such changes are a sensible moment to rebalance. A previous post Generating Cash: Asset Allocation with the Global Couch Potato Portfolio worked through an example for a net withdrawal scenario. CanadianFinancialDIY's Thoughts on How to Start a Portfolio from Scratch shows in detail a build-up scenario.

No comments: