The Rebalancing Dilemma: Why Letting Winners Run Might Not Be the Best Strategy

Examining the pros and cons of portfolio rebalancing, and why sticking with top performers might not always lead to optimal returns.

Photo by Calum MacAulay on Unsplash

In a recent interview, Guy Spier advocates for you to not rebalance your portfolio. He argues that normally most of the returns in your portfolio comes from a few selected winners. When you rebalance, you are taking dollars away from the winners and putting them into what may turn out to be losers. You are sub-optimizing.

I have the utmost respect for Guy. I was one of the first people to review his book 10 years ago way back when I was writing Value Stock Guide.

But on this topic, I wholeheartedly disagree with him.

What Exactly is Rebalancing?

We know that we rebalance to keep our portfolio allocation consistent. The traditional reason is that it keeps your risk profile in line with what you can tolerate.

Let’s say you are a typical investor and your financial planner has put you into a 60-40 portfolio. So you have 60% allocation to equities and 40% allocation to bonds. For simplicity sake, we will assume that these are broad-market index ETFs. Perhaps SPY 0.00%↑ and BND 0.00%↑ .

Over time, your equity portion may appreciate faster than your bond portion. You may end up with a 65-35 allocation or even 70-30. At this point, your portfolio may have started to become a little more volatile than you would like. So you rebalance by selling some of your equity positions and buying more of your bond positions to bring the portfolio back to the original 60-40 allocation.

Now you can sleep better.

As Guy Spier puts it, you are reducing your allocation to the asset that is better performing (and perhaps will continue to perform better in the future) and reinvesting in an asset that is a relative loser.

He wants you to let the winners run.

He gives an example of Buffett who has over time accumulated outsized positions in a few names and runs a concentrated portfolio.

His analogies are false. Next we will review the real reasons to rebalance, and why rebalancing does not necessarily mean that you are exiting winners.

We will start with a self-evident truth.

To Rebalance, You Need to have a Defined Allocation First

A defined allocation is what we call a model portfolio.

For example, a 60-40 portfolio is a model portfolio. There are many other model portfolios, such as Harry Browne’s Permanent Portfolio or Ben Felix’s 5 Factor Model Portfolio. There are several others and a quick internet search will reveal them.

There is a theoretical basis for why such portfolios exist.

For example, Permanent Portfolio is designed to weather all economic conditions. The goal is to protect your assets regardless of what is happening in the economy, and grow it in the process. Is this right for you? Perhaps, it depends on your risk tolerance.

Any of the Factor portfolios are designed to take advantage of counter-correlations between different factors. Size, Value, Quality, Momentum, etc are all factors that can independently outperform the market. However, they do not all move in lock-step. If you build a portfolio by combining these different factors, you can then even out your portfolio volatility.

These portfolios work because you periodically rebalance. You take the froth off the asset that is outperforming, and you build up positions in the assets that are still undervalued.

As the economy cycles through, your underperforming assets will get their time in the sun while the frothy assets will settle down. When this happens, you rebalance again and reallocate.

You are buying low and selling high without timing the market.

And this gives you a rebalancing bonus. Every little bit of this extra performance that you eke out when you incrementally buy low and sell high during the rebalancing process adds up and compounds. When you do this, your portfolio performance will be greater than the (weighted) average performance of your individual assets. This write-up by William Bernstein gives you the math behind this.

How do we Approach Capital Allocation?

We run a concentrated value portfolio. We do not own ETFs in this portfolio. Our factor allocations are through equities and are mainly restricted to Value and Small. We also happen to benefit from the quality factor as that is the by product our our selection criteria.

Our capital allocation process is not based on prescribed percentages. We do not equal weight, and neither do we pick a random number out of thin air based on our gut feeling. We calculate optimal allocation percentages for each stock in the portfolio based on its volatility, expected returns (which we derive based on our intrinsic value calculations during the research process), and its correlation with other stocks in the portfolio.

We use Kelly Criterion to determine optimal portfolio weights. You can read more about the process here.

I have also laid out the step-by-step process to do this yourself in Excel or other spreadsheet programs that you may use.

This gives us a portfolio that maximizes long-term returns and minimizes volatility.

As you can imagine, the portfolio drifts over time from the Kelly Criterion allocation. This can happen due to many reasons:

1. Some stocks appreciate while others may lose value

2. There is new information such as earnings reports, acquisitions, etc, that change our estimate of the expected returns (by raising or lowering the fair value or target price)

3. New stocks are added to the portfolio or we sell out of existing positions as they hit the target

Frequent rebalancing keeps our allocation to the optimal weights.

Now here is the important part that you should keep in mind, and that Guy Spier glossed over.

When a stock appreciates, it becomes over-allocated in the portfolio because the expected returns from the appreciated price are now lower, if there have been no changes in the fundamentals.

The portfolio at this moment is no longer growth-optimal. Not only we need to rebalance, but we actually need to re-calculate our weights to figure out a new growth-optimal portfolio.

Do We Ever Add to the Winners with this Process?

If there is no fundamental change in the company business, this process will reduce its weight if the stock has appreciated. This means we sell as the stock goes up, and will eventually exit the position altogether when the expected returns become zero or negative (the stock price hits our target price).

There is a possibility though that the company reports improving fundamentals and we raise our estimate of the intrinsic value. In such a case, the expected returns will continue to grow even if the stock price appreciates. When we recalculate the weights, we may keep or even increase our allocation to this stock. So yes, we may let the winners run, as long as there is good basis to do so.

Coming back to staying with his long-term positions, Buffett does this with quality companies that have enduring moat. These are the type of companies that keep growing their intrinsic value. While Buffett does not speak about this publicly, he either intuitively or explicitly uses the Kelly optimizations in the background. So did Bill Gross and many other super investors you may have heard of.