A Guide to Choosing a Rebalancing System

Choosing the wrong system will make it impossible for you to serve your clients the way you want to...

How should you choose a rebalancing system? It’s not a small matter. Getting this wrong will make it impossible for you to serve your clients the way you want to; it will hinder your growth, your compliance and your profitability.     

This is a decision maker’s guide to choosing a rebalancing system. We’re not going to get into specific systems or even specific features. We’re going to look deeper, discussing two key architectural differences that fundamentally set rebalancing systems apart:

1. Sleeve vs. sleeveless
2. Rules vs. optimization

These are not the only considerations that matter. You’ll of course want to look at features, integrations, support, complementary products, security, price, etc. But if you get the basic architectural elements wrong, getting the rest right won't fix it. 

Sleeve vs. Sleeveless

With the sleeveless (aka “holistic” or “blended model”) approach, a single account is managed holistically. The case for the sleeveless approach is that it is operationally simpler and supports superior tax and risk management, which is better for investors.

With a sleeve-based (aka a “subaccount-based” or “partitioned”) approach, accounts are divided into two or more subaccounts, which are rebalanced separately. The case for the sleeve-based approach is that it supports sleeve-level performance reports, which are sometimes expected by investors.

We discussed the sleeve vs. sleeveless issue in more detail in two earlier posts: The Ultimate Guide to Sleeves Part I and The Ultimate Guide to Sleeves Part II.  We won’t repeat the entire discussion, but here was our conclusion: 

…{the} holistic approach is less complex, less expensive and better for investors, but may run counter to client expectations, shaped by past practices. In the long run, the holistic approach will dominate. The advantages are too great and the countervailing pressure of client expectations will recede as memories of past practices fade.  

We’re already seeing increased adoption of the holistic approach, with four specific drivers:

1.  General fee compression has put pressure on wealth management firms to reduce costs, and sleeves are expensive.
2.  As wealth managers de-emphasize “product,” they are de-emphasizing product-level performance reports, including sleeve-level reports. In their place, they are adopting a goals-based framework, as well as stressing risk and tax management, which are best managed in a holistic, non-sleeved approach.
3.  The rise of low-cost robo solutions has put pressure on firms to increase the level of customization provided. This lowers the value of sleeves, since customization and tax management compromise the integrity of sleeve-level reports.
4.  In response to the proposed DOL rule, firms are looking for ways to provide scalable, customized, compliant solutions that meet the fiduciary standard. This is only possible through increased automation. The complexity of sleeves makes sleeve-based approaches harder to automate.
   

Rules vs. Optimization 

The heart of a rebalancing system is its analytics — what trades does the rebalancer generate and why?

As the name suggests, rules-based analytics are a set of rules, like “sell any stock overweighted by more than 20%” or “don’t sell any position with short term gains.” The simplest and most common rule is “clone,” i.e., “create the trades that make the portfolio exactly match its target.”

The case for rules-based analytics is that they’re simple — and therefore easy to understand and use in conjunction with a manual rebalancing process. They also tend to be very low cost, sometimes even bundled in (at no cost) with other systems. The case against rules-based analytics is that they are not powerful enough to handle sophisticated risk and tax management. This means that if you use a rules-based rebalancer, you’re either not going to be offering high levels of tax and risk management, or doing so will remain a largely manual process.

Advanced rebalancers that can automate complex customization and tax management are optimization-based. An optimizer can be described as a trade-off engine. It’s like an old-fashioned balance scale. Imagine you’re trying to decide whether to sell something. You put all the reasons to sell on one side. You put all the reasons to hold on the other. Whichever side is heavier wins.

The case for optimization-based analytics is that it lowers the cost of offering high levels of customization, tax and risk management. When the cost of doing something goes down, you can offer more of it. Not only can you bring advanced services to smaller accounts, you can provide a higher level of service to large accounts. Tax-loss harvesting, for example, can go from being a year-end event to something you do year-round. You can add tax budgets and implement sophisticated substitution strategies. It pays off in terms of higher after-tax returns and lower dispersion. Our average client reduces both dispersion AND average tax burden by more than 60%.  

Complex Rules?

We noted earlier that rules-based systems are too simple to handle complex customization and tax management. You might think that all that would be needed would be more rules. The idea is appealing, but in practice, the rules would need to be so complex that they would become unwieldy. They’d be too complex for users to understand them. So complex, in fact, that they become difficult for software programmers themselves to understand.

To get a sense of why it’s not practical to use rules alone to handle complex trade-offs, consider the rules you’d need to simultaneously balance the following factors that need to be considered to avoid short-term gains:

• Time-until-long (it’s one thing to sell 300 days till long term; it’s another to sell 3 days before)
• The client’s tax rates
• How concentrated the position is (OK to hold onto a 0.5% weight position for 300 days, but not a 50% weight position)
• How large the gain is (OK to sell a 5% gain but not good to sell a 500% gain)
• The security’s ranking (OK to hold onto a neutral-ranked security, but not a weak-ranked position)
• Does the position contribute to violating a “max asset-class” constraint (i.e., it’s a large-cap position and large cap is overweighted)? Or would selling the position help satisfy some other unrelated constraint, like a cash withdrawal? If so, are there other positions that could be sold instead? What are their rankings? What would be the cost of selling these other positions? Would selling them violate some other constraint? Would it add to drift?

This list of considerations is not complete — we could add min-trade size, ESG constraints, lot rounding, etc. — but it gives an idea of some of the challenges of trying to handle complex portfolios with rules alone. Which is why a no rules-based rebalancer is sophisticated enough to automate advanced tax management and transition.  

In contrast, optimizers handle this complexity with ease. If you go back to the image of the old-fashioned balance scale, it’s not hard to keep adding factors to one side or the other. Mathematically, it’s more complicated, but conceptually it’s easier than a complex set of rules.

This doesn’t mean that rules have no place in rebalancing. It’s not possible for an advisor (or as we’ve noted, even programmers) to tell a rebalancer what they want done in every possible situation. However, it’s entirely feasible for advisors to give the rebalancer rules delineating behavior they don’t want. And within those fences, advisors can let the system do its thing. For example, client statements will often include a pie chart of both the recommended asset allocation and the client’s actual asset allocation. Because of constraints, taxes and trading costs, there can be good reasons to have these two be a bit different. But you probably don’t want them too different. So you restrict the rebalancer to look only at trades that keep portfolios within certain asset allocation bounds. This is how optimization-based analytics work in practice. It’s really “optimization within user-set, rules-based boundaries.”

Putting It All Together

In theory, advisors have four basic choices in choosing rebalancing systems. But as far as we know, there are no optimization-based sleeved systems, a combination that would be both expensive and ill-suited to risk and tax management. In practice, that leaves three basic types of rebalancers. Here’s a summary of their main advantages and disadvantages:

There isn’t a single best rebalancing system for everyone, but getting it right goes to the heart of your practice — your ability to serve your clients the way you want. And getting it wrong will keep you and your business, well, off balance.

For more on this topic, check out What is Rebalancing Automation?