Rebalancing

Rebalancing is the process of realigning the weightings of one's portfolio of assets. Rebalancing involves periodically buying or selling assets in your portfolio to maintain your original desired level of asset allocation. Although the investor may start with a given portfolio design, as time passes the portfolio will move away from the original asset allocation targets.

Why rebalance?
As time progresses, the performance of the asset classes in your portfolio will tend to move in different directions and your portfolio will drift from the original asset allocation. As part of portfolio maintenance, it is desirable to periodically readjust that allocation, in order to control the risk.

Improving returns?
During some periods, rebalancing may boost returns slightly, at least before costs. For example, Arnott and Lovell (1992) looked at 50-50% mix of US stocks and long bonds between 1968 and 1991. Both asset classes had high returns during that period, and monthly rebalancing improved annualized returns by 7 basis points (before costs) compared to no rebalancing. Bernstein has concluded that, in some circumstances, greater returns may be achieved by periodic rebalancing.

An interesting explanation of why rebalancing enhances returns in some periods but not others is offered by Perold and Sharpe (1988). They look at t-bills and US stocks. In strong bull or bear markets, rebalancing hurts returns. For example, in a strong bull market, the investor who rebalances is selling stocks, but stocks continue to rise in value, so in retrospect, she would have been better off not rebalancing. Similarly in strong bear market, the investor is buying stocks, but they keep loosing value. However, in a flat but oscillating market, rebalancing improves returns, because rebalancing trades exploit the volatility and reversals. Of course, it is impossible to know for sure which direction the market will go in the future, so the investor must adopt a rebalancing policy that works reasonably well in all markets. Also, portfolios typically contain more than two asset classes.

Controlling risk
Over the very long run, stocks have returned more than bonds, and rebalancing mostly leads to selling stocks (the best performing asset class), so it has historically hurt returns. Yet not rebalancing would have meant holding a portfolio with ever-increasing volatility, eventually a nearly all-stock portfolio, which very few investors can tolerate, especially as get older. Also Dichtl et al. (2014) show that "relabalancing strategies significantly outperform buy-and-hold in terms of average risk-adjusted performance".

The investor should regard rebalancing as primarily a procedure for controlling the risk, not boosting the return. The question is not whether to rebalance or not, but when and how to do so.

When to rebalance?
An investor's investment policy statement should include a rebalancing rule. This strategy should then be adhered to without second-guessing, otherwise it becomes market timing. Vanguard lists three main types of rebalancing policies:


 * Time-only: The portfolio is rebalanced according to a schedule — daily, monthly, quarterly, annually and so on — regardless of how far the portfolio’s asset allocation has drifted from its target.


 * Threshold-only: Investors rebalance when their portfolios drift from the target asset allocation by a certain amount (such as 1%, 5% or 10%). Daily monitoring is required when using a threshold-only strategy.


 * Time-and-threshold: Investors monitor their portfolios on a scheduled basis but rebalance only if their current asset allocation has drifted from its target by a predetermined amount.

Vanguard concludes that there is no universally optimal relabancing strategy. They state that "choosing a rebalancing strategy comes down to personal preference. An investor must decide how far they are willing to let the portfolio drift from their target asset allocation and how much they are willing to pay in costs." Other authors disagree and have tested different rebalancing rules, typically using historical data.

Calendar rules
Several studies have been conducted on the optimal rebalancing frequency using calendar ("time-only") rules. Dichtl et al. (2014) looked at portfolios of 60% stocks and 40% government bonds in the US, UK and Germany, between 1982 and 2011, using historical rolling windows (e.g., 26 five year periods or 21 ten year periods) and Bootstrap simulations. With the rolling window approach, and accounting only for institution-like transaction costs (and no taxes), annual rebalancing typically beats or equals the average annualized returns of quarterly or monthly rebalancing, although the differences are minor. An individual investor would likely have higher costs than an institution, which is a further argument for relatively infrequent rebalancing.

Rattray et al. (2019) note another issue with monthly rebalancing, beyond higher costs: it makes drawdowns worse during crises. Quarterly or annual rebalancing typically improves returns (relatively to monthly rebalancing) during crises (see their Table 2). Although these authors go on to advocate "strategic rebalancing", this is again akin to market timing and not further considered here.

Hallett (2019) notes that asset prices exhibit both momentum (over the short term, the trends persist) and mean reversion (over the longer term, the trends do not persist). He recommends rebalancing every 2 or 3 years.

Thresholds
Rebalancing thresholds may be set on either an absolute basis - that is, as a percentage of the entire portfolio - or on a relative basis, as a percentage of the particular holding. For example, a statement that "Equity weightings to remain in the range 45% to 55%, with a nominal target of 50%" is a threshold of 5% absolute. An example of a relative threshold could be a statement that "Small Cap Stocks to be set at a weighting of 10%, with a deviation of not more than 25% relative" (which would be a 2.5% absolute threshold). A "5/25" rebalancing policy (that is, rebalancing when asset allocations drift more than 5% absolute or 25% relative from targets) has been advocated by Swedroe.

Vanguard does not believe there is an optimal strategy but states that "annual or semiannual monitoring, with rebalancing at 5% thresholds, is likely to produce a reasonable balance between risk control and cost minimization for most investors". In other words, they advocate a time-and-threshold policy. A threshold-only policy would theoretically require daily monitoring, which may not be practical and likely to induce investor stress.

Using portfolio cashflows
Rebalancing can be most easily done in accumulation mode by adding new funds (or redirecting portfolio income such as distributions, dividends, etc.) to that section of the portfolio that is most out of balance on the downside. This limits transactions costs and has no tax consequences. See example 2 below.

Similarly, in withdrawal mode, the investor can sell (for withdrawal purposes) a portion of that section of the portfolio that is most out of balance on the upside.

Buy selling and buying
In registered accounts, where there is no immediate tax liability, it is also often possible to switch funds from one component to another at reasonable cost. A sufficient portion of an overweight asset class is sold, then the funds are used to purchase an underweight asset class (see example 1 below). Rebalancing in this way in a taxable portfolio, however, may incur significant tax, and should usually be avoided in that case.

Detailed example 1: complete rebalancing with index funds
Matthew C. from P.E.I. has a simple index portfolio with four index mutual funds. His target asset allocation is:
 * 50% Canadian bonds
 * 25% Canadian equities
 * 10% US equities
 * 15% International equities

Matthew has a registered account and there are no tax consequences to rebalancing. Matthew has a "5/25" policy but he hasn't looked at an investment statement for five years (being busy on his farm) and he decides that a full rebalancing exercise is needed. He gathers the following current values from his account:
 * Canadian bonds (TD Canadian Bond Index – e): $5800
 * Canadian equities (TD Canadian Index – e): $4000
 * US equities (TD US Index – e): $1000
 * International equities (TD International Index – e): $2200
 * Total: $13000

The first step is to calculate the current allocation by dividing the dollar amounts for each fund by the total dollar amount:
 * Canadian bonds: 45%
 * Canadian equities: 31%
 * US equities: 8%
 * International equities: 17%
 * Total: 100%

Matthew subtracts his current allocation from his target allocation and gets:
 * Canadian bonds: +5% (he needs 5% more)
 * Canadian equities: -6% (he has 6% too much)
 * US equities: +2%
 * International equities: -2%
 * Total: 0%

So Matthew needs to buy 5% bonds (positive number) to get him to 50%, sell 6% Canadian equities (negative number) to get back to 25%, and so on. These percentages are reconverted to dollar amounts by multiplying them by the total account value of $13k:
 * Canadian bonds: +$700 (buy $700 worth of bonds)
 * Canadian equities: -$750 (sell $700 worth of stocks)
 * US equities: +$300
 * International equities: -$250
 * Total: $0

To keep things simple, Matthew decides to sell $725 worth of the Canadian equity fund and buys bonds with that amount. He then sells $275 of the international equity fund to purchase some US equities. Close enough!

Detailed example 2: new contribution with ETFs
Marilla C. has a three ETF portfolio which sits entirely in a RRSP account at a discount brokerage. Her target asset allocation is:
 * 40% Canadian bonds
 * 30% Canadian equities
 * 30% Global equities

She’s just added $5000 of cash to her account, and there was already $143 sitting there from recent ETF distributions. She reviews her current ETF holdings:
 * VAB (Canadian bonds): $34 658
 * ZCN (Canadian stocks): $35 217
 * XAW (Global Stocks): $31 123

The total value of her account, including the cash, is $106 141. She divides the market value of each ETF by the account total to get her current allocation:


 * VAB (Canadian bonds): 32.7%
 * ZCN (Canadian stocks): 33.2%
 * XAW (Global Stocks): 29.3%
 * Cash: 4.8%
 * Total: 100%

Marilla subtracts her current allocation from her target allocation and gets:
 * Canadian bonds: +7.3% (she needs 7.3% more)
 * Canadian equities: -3.2% (she has 3.2% too much)
 * Global equities: +0.7%
 * Cash: -4.8%

Marilla decides to buy more VAB (Canadian bonds) with all the cash available, $5143. This will not bring back Canadian bonds all the way to her target of 40% but will get her most of the way there in one transaction. The rest of the imbalances, for example the surplus of Canadian equities, can be addressed during her yearly rebalancing scheduled in a few months.

During market hours, she obtains a quote for VAB: the bid is $26.43 and the ask is $26.45. Marilla subtracts the $10 commission from her cash, which leaves $5133. Dividing this by the current ask price for VAB yields 194.1 shares. So she enters a limit order for 194 shares of VAB, using the ask price as her limit. This order is promptly executed and adds $5131.30 to her existing VAB holdings. Taking the commission into account, there is less than $2 of cash left in her account.

The current values and percentages are now:
 * VAB (Canadian bonds): $39 789.30 or 37.5%
 * ZCN (Canadian stocks): $35 217 or 33.2%
 * XAW (Global Stocks): $31 123 or 29.3%
 * Cash: $1.70 or 0.0%
 * Total: $106 131 or 100%

These calculations can be automated in a rebalancing spreadsheet to save time (see external links below for an example).

Rebalancing spreadsheets

 * Rebalancing spreadsheet, free download from Google Docs. Supplies three example approaches to rebalance a portfolio by allocation percentage, transfer amount, or final value; by forum member LadyGeek. To download, select File --> Download As --> Excel or OpenOffice.

Other links

 * Dan Hallett, Rebalance for long-term risk control not to boost returns, 2011
 * Dan Hallett, Rebalance for long-term risk control not to boost returns, 2011
 * Dan Hallett, Rebalance for long-term risk control not to boost returns, 2011