Portfolio Rebalancing: 'Do You Want To Know A Secret?'

|
Includes: CRF, DDM, DIA, DOG, DXD, EEH, EPS, EQL, FEX, FWDD, HUSV, IVV, IWL, IWM, JHML, JKD, OTPIX, PPLC, PPSC, PSQ, QID-OLD, QLD, QQEW, QQQ, QQQE, QQXT, RSP, RWL, RWM, RYARX, RYRSX, SBUS, SCAP, SCHX, SDOW, SDS, SFLA, SH, SMLL, SPDN, SPLX, SPSM, SPUU, SPXE, SPXL, SPXN, SPXS, SPXT, SPXU-OLD, SPXV, SPY, SQQQ, SRTY, SSO, SYE, TNA, TQQQ, TWM, TZA, UDOW, UDPIX, UPRO, URTY, USSD, USWD, UWM, VFINX, VOO, VTWO, VV
by: Reel Ken

Summary

The theory behind portfolio rebalancing.

Why risk and return are two sides of the same coin.

The "secret"  behind portfolio rebalancing effectiveness.

Dynamic rebalancing is what it really should be about.

Portfolio Balancing is a technique that many investors utilize. It also seems to be a fundamental characteristic of most money managers. Some mutual funds, variable annuities, 401k plans and brokerages even offer automated versions to make it as simple as possible

It's very easy to develop a strategy and "back-test" it to see what worked and what didn't. Unfortunately, one can't "back-test" the future. I for one, certainly wouldn't pretend to say I know the future. "Past performance is no guarantee of the future" is much more than a caveat for performance.

So, let me state that I don't know whether or not Portfolio Rebalancing will prove to be effective in the future. I'm not even sure it was effective in the past. What I can do, is present the reader with an analytical look at Portfolio Rebalancing so they can be prepared and decide for themselves its efficacy to their own circumstances.

More importantly, I will provide some tactics that the reader can use to adjust their portfolios in a dynamic way. In so doing, this article will, by necessity, stretch the comfort level of many readers. That should be viewed as a good thing (I hope).

Theory

First: What is Portfolio Rebalancing?

The most popular definition looks something like this:

"Rebalancing is the process of realigning the weightings of a portfolio of assets. Rebalancing involves periodically buying or selling assets in a portfolio to maintain an original desired level of asset allocation.

In Chart Form, it looks like this:

Image result for portfolio rebalancing images

Portfolio Rebalancing is predicated on two main postulates.

1) Risk reduction: Once the investor establishes a risk profile, they should rebalance to maintain that risk profile. Allowing asset classes to grow, unabated, increases risk exposure.

2) Increased Returns: The systematic selling of asset classes that out-perform and reallocating to under-performers utilizes buy-low-sell-high and is accretive to returns.

Before we can truly understand Portfolio Rebalancing we need to put on our thinking cap.

These two concepts, Risk Reduction and Increased Returns really go hand in hand. They cannot be quantified without looking at them in tandem. Consider the following:

1) If returns are greater using Portfolio Rebalancing, then the extra returns, in and of themselves, reduce risk.

Let me explain by an example: Let's say one has an initial portfolio of $100,000 that is 50%/50% stocks and bonds. Without rebalancing it grows to $150,000. If Portfolio Rebalancing increases returns, then it would grow NOT to $150,000 but $160,000. That "extra" $10,000 acts as a cushion against a drop when compared to the portfolio that just grew to $150,000.

Portfolio Rebalancing clearly reduces risk by having fewer stocks and more bonds in a portfolio. If Portfolio Rebalancing is also accretive to returns, then it's a risk reduction "twofer" when compared to not rebalancing.

2) If Portfolio Rebalancing returns are NOT greater, but less, then it follows that risk is increased relative to the unbalanced portfolio. This may be a little hard to grasp, so let me expand on the previous example.

Example: If returns are sacrificed through Portfolio Rebalancing it's simply the other side of the coin. Reversing the previous example, not rebalancing grows to $160,000 and Portfolio Rebalancing grows to only $150,000. Well the increased growth provided by not rebalancing provides a $10,000 cushion that can absorb a sizeable drop before it equalizes with the rebalanced portfolio.

In this example, the $160,000 portfolio without rebalancing would have become stock "top heavy" and might consist of $100,000 stocks and $60,000 bonds. This looks "scary" to the investor that would have had a Rebalanced Portfolio with $75,000 each in stocks/bonds. But this is tunnel vision in viewing it simply in terms of asset allocation. The Rebalanced Portfolio does have less risk when compared to itself, but NOT when compared to the unbalanced portfolio. Let's see why ...

Assuming a drop so large that bonds go up 10% in value, stocks would have to drop about 45% before the two portfolios equalized. The extra $10,000 in growth goes a long way relative to the Balanced Portfolio. Not that a 45% or greater drop couldn't happen, but just to give an example of the extent that greater returns provide for greater risk protection.

Here's the tricky part ... The extent to which Portfolio Rebalancing actually reduces risk is intertwined with the extent to which it influences returns. The greater/lesser the returns, the greater/lesser the risk relative to not rebalancing.

Putting it another way: Accumulating more money provides a cushion against a drop.

This forces us to look at whether or not Portfolio Rebalancing provides greater returns

The theory is really an adaptation of buy-low-sell-high. So, it would seem to make sense, except for the fact that one is dealing with stocks and bonds which have different growth characteristics.

There are zillions of charts, articles and, research papers on the subject. Unfortunately, they are all subject to "back-testing" and data mining. So, I'm not going to take one side or the other. Instead, it's more important to understand than conclude. So let's look at one of the more typical charts:

Well this chart wants us to conclude that Portfolio Rebalancing is accretive to returns. In fact, the heading leads the reader to that conclusion. End of story? Not really, let's examine the chart more closely.

From 2002 to 2009 it seems to be pretty much a draw. The substantial "edge" comes in 2009-2014. But look closely ... in 2009 the portfolio was rebalanced and the "edge" was created between 2009-2010. Thereafter, the "edge" remained constant, just as it had in the years 2002-2009.

Now we all know that 2009 represented a major bottom and the start of one of the greatest bull runs in history. It cannot be surprising to anyone that moving money from bonds to stocks at this time would pay off ... and so it does in this chart.

Unfortunately, the chart only runs to 02/27/2014. The market is up almost 35% since then and rebalancing would have lost 1/2 that 35% and the "edge" would probably be wiped put. So, the advantage shown in the chart is transient.

Similarly, consider that the chart's data starts at 1998, but it truncates the 1998-2002 portion (data mining?). My estimate is that the advantage in favor of rebalancing from 2002-2008 was a result of rebalancing during the 2000 dot-com bust. So, the pattern seems to be no real advantage until a big drop, then an advantage that dissipates over time. Wash, rinse and repeat.

Next, add to this that the chart doesn't factor in trading costs and taxes on repositioning. So, what we can actually conclude is that Portfolio Rebalancing is likely "neutral" for returns provided one rebalances during a major drop.

So, what we have can be summarized as follows: Portfolio Rebalancing will NOT show enhanced returns unless there are major drops in the market. Even with these major drops, unless they are frequent enough, any advantage is transient. Portfolio Rebalancing is likely neutral over returns.

That actually is common sense. Stocks have traditionally outperformed bonds and stocks are generally up 2/3rds of the time. If one systematically transfers from stocks to bonds, then the result should be reflected in a lower overall return. The "swing factor" would be volatility causing enough large drops that enable Portfolio Rebalancing to buy-low-sell-high.

In any event, if Portfolio Rebalancing is neutral for returns, then the Risk Reduction is "free". This is what Portfolio Rebalancing is counting on.

Theory Meets Reality

The key to performance of Portfolio Rebalancing is dependent upon larger down moves. Routine volatility does not seem sufficient to overcome stocks historic over-performance relative to bonds.

But there's a hitch to this. If, and this is a very big "if", the investor doesn't take advantage of rebalancing during big drops, Portfolio Rebalancing will suffer from diminished returns relative to not rebalancing.

In a previous article of mine, I spent a great deal of time discussing the fact that most investors do the opposite of what they're supposed to do. During drops, especially big drops, they react emotionally and don't buy. I recommend readers look to that article or any number of other articles discussing negative returns associated with emotional investing.

Referring back to the chart above ... all gains are predicated on rebalancing in 2008-2009. But let's not forget that this was a time when investors were worried that money market funds would "break the buck", the world-wide economy would collapse and apocalypse was imminent. Investors were willing to accept negative interest rates on Treasuries just because they felt safety was uppermost. How realistic is it to just blindly assume the investor moved a large proportion of their "safe" bonds into stocks at that time?

If the investor just waited six months, till September 2009 instead of March 2009 to rebalance, they lost a 50% upward spurt. "He who hesitates has lost".

So, what we have is very simple. The investor must have the discipline to re-allocate towards stocks during major downturns. If they hesitate or delay until markets improve, Portfolio Rebalancing will NOT be neutral but show considerable drag. Possibly as high as 200 to 300 basis points per year. These lower returns mean less money for the investor. Less money means more vulnerability to true dollar risk.

Don't get me wrong. I'm not suggesting one shouldn't Portfolio Rebalance. I'm just pointing out that if one does, then it is incumbent upon them to follow through and fight any emotional reactions. If they can't do this, they are actually increasing their risk, not reducing it.

Dynamic Rebalancing 1.0

The discussion so far has been devoted to Portfolio Rebalancing between equities and bonds. That is how most investors see and practice Portfolio Rebalancing. However, there are many different asset classes: Large Cap; Small Cap; Overseas; Commodities; Gold, etc.. Many investors have at least a smattering of these.

As long as equities outperform debt, then rebalancing from equities to debt is a losing strategy. It's possible that bonds will outperform equities in the future, but historically the return on stocks is about twice the return on bonds.

There is another way to look at Portfolio Rebalancing. View it in terms of assets that are correlated and assets that are inversely correlated. Stocks and bonds are inversely correlated. Small Cap and Large Cap stocks are correlated. Instead of just balancing equities against bonds ... balance equities against equities. For instance, if one has Large Cap and Small Cap, rebalance those two equity positions. Over long periods of time, their performance is closer and they often switch being "in favor". Similarly, one may rebalance Gold and Bonds on the premise that they are more correlated than Gold and stocks.

The theory is simple ... it is version of mean reversion of returns. Stocks routinely "wobble" as they move. Correlated assets get to the same place but their zigs and zags are not precisely in tune. Rebalancing between correlated asset classes takes advantage of mean reversion by selling TEMPORARY over-achievers and buying TEMPORARY under-achievers. It is very different than balancing between two assets classes, such as stocks and bonds, when one is expected to outperform the other over time.

Dynamic Rebalancing 2.0

There is a way to have one's cake and eat it too. First, rebalance between correlated asset classes. Then rebalance uncorrelated asset classes based upon presumed performance.

Just for illustration, let's say the long term presumed return on equities is 10% and on debt it is 5%. Assuming a 50/50 asset allocation between equities and bonds and ignoring any volatility or deviation from the average ... 2/3rds (10%/%15%) of the total portfolio return is artificially attributed to equities.

An example might help. We start with a portfolio allocated 50/50 between stocks and bonds. Let's say the return in one year is stocks up $7,000 and bonds up $5,000. Traditional rebalancing would move $1,000 from stocks to bonds. Dynamic rebalancing would attribute 2/3rds of the combined $12,000 return, or $8,000 to stocks and only $4,000 to bonds. Therefore, stocks are under-represented and a resultant move of $1,000 out of bonds and into stocks would balance the portfolio.

In this example, the Dynamic Rebalance is opposite traditional rebalancing ... instead of moving from stocks to bonds, one moves from bonds to stocks. The rationale is simple: The stocks underperformed their presumed rate and will mean revert upwards. The bonds over-performed their presumed rate and will mean revert downward. So, one allocates according to what is most likely to happen tomorrow, not what happened today.

Another example: Stocks and bonds each go up $5,000. Traditional rebalancing would do nothing. Dynamic rebalancing takes the total return of $10,000 and allocates $6,666 to stocks and $3,333 to bonds. This would result in $1,666 moving from bonds to stocks.

Let's give yet another example: The FED unexpectedly raises rates. Stocks are up only $4,000 and bonds drop $2,000. Traditional portfolio rebalancing would move $3,000 from stocks to bonds (each class ends up +$1,000). Dynamic Rebalancing would move $2,666 from stocks to bonds. Stocks end up $1,333 and bonds $ 666.

In essence, rebalance on the comparison of returns to presumed returns. This will grow the presumed better asset class.

Moderating Balances

Let's assume that an investor starts 50/50 and engages in Dynamic Rebalancing. Let's also assume, that over time, stocks do outperform bonds as anticipated. This creates a slow increase in the proportion of the portfolio allocated to stocks. As stocks continue to outperform, the proportion will continue to increase .... but it will increase at an ever decreasing rate. If stocks continue to outperform bonds, eventually the allocation will reverse and a slow move from stocks to bonds will occur.

Let me give an example: The portfolio starts out 50/50 and grows (favoring stocks) to $100,000 bonds and $150,000 stocks. In the next year, bonds return 5%, or $5,000 and stocks return 10% or $15,000. On first glance, one would think no rebalance occurs as each precisely hit their benchmark.

But if we go through the math, the $20,000 total growth, when allocated 2/3rds to stocks, stocks should be allocated approximately $13,000 and bonds $7,000. So there is a $2,000 move out of stocks and into bonds.

In a more extreme example, the portfolio grows to $100,000 bonds and $300,000 stocks. Bonds return 5%, or $5,000 and stocks return 10% or $30,000.

Going through the math, the $35,000 total growth is allocated 2/3rds to stocks. That means stocks are allocated approximately $24,000 and bonds $12,000. So there is a $6,000 move out of stocks and into bonds.

This is very different in application than traditional portfolio rebalancing. Traditional rebalancing moves money towards bonds, BEFORE stocks get an opportunity to outperform. Dynamic Rebalancing moves money to bonds AFTER stocks have shown gains.

When to Rebalance

Whether one uses traditional or dynamic rebalancing, rebalancing small amounts is really not too wise. One gains very little and has to deal with trading costs and potential taxes. One can set a threshold imbalance that would trigger a rebalance. For instance, a minimum divergence of 3%, or 5% seems reasonable. It avoids the nuisance of small moves and lowers trading costs. One would simply keep track and wait until the imbalance hits the trigger point.

Next, when and how often to rebalance? To many this is a very difficult question. There are so many divergent studies it's astonishing. Should it be monthly? Quarterly? Annually? Beginning of year? Mid-year? The list goes on.

I find all these studies are nonsense. It is clear from all the data that one rebalances after major drops or major swings. It is "event" driven not "time" driven. All these studies fail to see the forest before the trees.

Though it runs contrary to conventional wisdom one should rebalance when the threshold is reached. This might occur after several years of low volatility. Or it might trigger randomly if stocks drop quickly. In some years, stocks might hit the trigger several times, resulting in multiple rebalancing in the same year.

Here's a sourced graph that seems to indicate the 5% threshold works well. Please note, it suffers from the same issues regarding 2009 as the previous chart.

Actually this chart wants to tell us several things. The theme it tries to portray is that Portfolio Rebalancing is accretive and a 5% threshold might work.

Careful examination reveals the "sub-story". We can easily see the 2009 theme because we know to look for it. But there's much more revealed. A 5% threshold rebalancing is triggered on any move more than 5%. This happened THREE times in 2010, TWICE in 2011 and ONCE in 2012. Look closely and one can see the "pop" after those drops as the bull marched on.

What we can learn is that the combination of thresholds and rebalancing can be an effective one-two punch. This suggests rebalance using "thresholds" and "events" and not arbitrary dates.

I can't leave this subject without what is, in my opinion, the poster child for "threshold" versus "time". Let's look at the typical rebalancer that rebalances on January 1st. Well, in 2014 we had a 10%+ drop on September that recovered by January. In 2015 we had a 10%+ drop that recovered by January. In 2016 we had a 10%+ drop in January, but after January 1st. In all three of these instances ... all in very recent memory ... the "time" rebalancer took no advantage, at all. The "threshold" rebalancer, would have rebalanced each year at the most opportune time. Assuming a 50/50 allocation, the "threshold" balancer would have gained almost 20% over the "time" balancer. All in less than 18 months.

Remember, the postulate for Portfolio Rebalancing is that one sells low and buys high. Utilizing thresholds takes advantage of the randomness of big drops. Provided one follows through ("human factor") threshold rebalancing tells us when and how to rebalance.

Summary

Portfolio Rebalancing can be a very effective tool to reduce risk and possibly enhance returns. Before one enters into such a strategy they should recognize when and why it works. In many cases (and certainly historically) one can reduce risk by NOT "robotically" rebalancing. Asset growth is ultimately the preferred risk reduction strategy.

Most investors don't achieve optimal results because they assume they are doing the right thing. Most end up with less cash and correspondingly more risk.

This result can be turned on its head. Greater returns and less risk ... provided ... the investor clearly understands that it is their actions on BIG DROPS that will tell the tale. It is essential they follow through or they will not achieve optimal results. Doing this requires an understanding of emotional constraints and acquiring the appropriate discipline.

Once the investor understands the fundamental factors governing success, they can move to Dynamic Rebalancing. This rebalances correlated asset classes amongst themselves and uncorrelated asset classes against assumed performance.

Now, no one can guarantee that any strategy will be the best strategy. The future may teach us that the best investment was money market funds, BitCoin or "Who Knows What"?

But if one wants to deal within the parameters of "most likely", they should at least be aware of what makes "most likely" tick.

Addendum

This article used Dynamic Rebalancing with an assumed 50/50 start point. Doing so made illustrations simple. Investors may fit the more traditional 60/40 mix. That being the case, the 2/3rds to stocks isn't properly reflective. It would be 3/4ths instead of 2/3rds. Also, if one changes the assumed growth rates, say bonds at 3% and stocks at 7%, the formula also changes. I thought the article was long enough to spare readers the math. But those that are curious, can work it out and I'm available for help.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.