Portfolio Rebalancing - Present and Future

The second quarter was a period of heightened activity in many clients’ portfolios. As discussed in last quarter’s Commentary, we began restructuring portfolios in April by replacing seven actively managed mutual funds. Separately, during this past quarter, we also undertook our annual rebalancing of client portfolios.

Although we monitor client portfolios on an ongoing basis, and rebalance along the way as clients add cash to, or make withdrawals from, their portfolios, once a year we comb through every portfolio to make sure each one is sufficiently close to its target asset allocation. Portfolio theory and empirical evidence suggest that rebalancing is a very important practice, and one that many investors either overlook or recoil from, since it often means selling one’s winners and buying one’s laggards. The main benefit to rebalancing comes from the tendency for asset class performance to “revert to the mean” after periods of especially good or especially poor performance. In a perfect world, rebalancing would allow you to harvest gains in one portion of the portfolio at the top of the market and reinvest the proceeds in another segment of the portfolio that has done poorly by comparison with the first, but which is about to turn around. In the real world, where peaks and troughs are apparent only in hindsight, periodic rebalancing still adds considerable value

There is a band around the target allocation within which we’re willing to let the actual allocation wander without correction. While there are benefits to being “on target,” those benefits come at some cost in terms of taxes and transaction charges, which we seek to minimize. To balance those conflicting objectives, our current practice is to rebalance only if the actual allocation is more than 5% from the target (i.e., if the target for domestic large cap equity is 30%, we’ll only rebalance if it is below 25% or above 35%). Where the target allocation is less than 10%, our rebalancing range is plus or minus 50% of the target.

We also have targets for sub-asset classes (i.e. how much “value”; how much “growth”) and for specific funds (how much in fund x, or y, or z in a particular sub-asset class), but we allow increasingly greater tolerance at these portfolio levels. The correlation of returns between sub-asset classes within an asset class, and funds within a sub-asset class are relatively high, so the benefits of rebalancing are lower here than at the asset class level.

Advancing The State of The Art with “iRebal”

In the past, the annual rebalancing effort has been anticipated with about as much enthusiasm as one would muster for a tooth extraction. Looking at every portfolio and attempting to rebalance all the pieces while minimizing transaction costs and taxes, and while attempting to place tax-inefficient investments in tax-deferred accounts was not only tedious but so subject to error that it required multiple layers of review.

Over the past year, Kochis Fitz, in conjunction with two other wealth management firms, has developed an artificial intelligence tool to automate this process. Throughout the software design process, we compared notes on best practices with our peer firms and incorporated these into a program called iRebal, for “intelligent rebalancing.”

After being fed data on each client (tax rates, capital losses, etc.) and each portfolio (target asset allocation, special account limitations, idiosyncratic client preferences) over the prior 6 months, iRebal made its debut in our firm in April. Already, in release 1.0, it has produced several important benefits in more reliably consistent application of rebalancing criteria across portfolios, at much greater speed. Over time, as the software incorporates ever-more information, we expect that this will free up more of our firm’s resources for those client-specific portfolio issues and client service needs that we have no desire to automate.

The three firms that developed iRebal are also underwriting some groundbreaking research into the optimal frequency for rebalancing (monthly? annually?), the optimal bands around the asset class targets (bands that vary with the volatility of the asset class, for example, as opposed to a fixed, uniform band like our current 5%), and asset placement between taxable and tax-deferred accounts. We also hope to have this research be able to quantify the benefits (in terms of enhanced risk-adjusted rate of return) attributable to each of these constituent parts.

With the benefits of this research and an increasingly powerful iRebal, Kochis Fitz intends to remain at the forefront of the science and the art of managing portfolios for taxable individuals. However, no software tool we can presently envision will ever completely replace the insights of experienced wealth managers. Rest assured, while iRebal will aid our management efforts, your portfolio will continue to have our finger prints all over it.

Mike Fitzhugh

_