|
|
| The KleinPost |
|
|
| More Tips |
|
Track Record - Simple but Critical Simple Models Are Often The Best Pros and Cons of the Distinct Portfolio Filter
|
| Click here to see all tips. |
| Tips Archives |
Simple Models Are Often The Best
Most advisors have a feel for what they seek when evaluating mutual funds or managers. However, even the most seasoned advisors can have difficulty listing and quantifying these features. Some may actually be qualitative. Others may be like the Supreme Court’s opinion on pornography: You can’t easily define them, but you know them when you see them. As a user of K4 Fund Selection or K4 Manager Selection, your task is to quantify your style requirements into the tool’s factors.
The K4 products provide plenty of factors to choose from – in fact, the number may initially seem overwhelming. For example, K4 Fund Selection Plus includes about 170 different options. Some users have been so intimidated, they gave up. Don’t be one of them. Also resist the urge to act like a kid in a candy shop by selecting too many factors. It would be wonderful if your top-ranked funds had superior 1-, 3-, and 5-year returns versus the category, the category index, and the primary index, but if return is so important to you, you can achieve the same results by thoughtfully selecting only one or two of these factors; you don’t need them all.
Careful factor selection is truly the key. The best scenarios don’t necessarily depend on the number of factors used but rather the breadth of the information they measure. In fact, some of the most effective scenarios use only a handful of factors. There are a number of reasons to attempt to hold down the number, first and foremost being the fact that the K4 tools only allow a maximum of 15. That’s actually more than enough for an effective scenario.
A second product-related reason concerns the number of trade-off questions. One of the unique and powerful features of K4 is its ability to simplify the factor weighting process through “adaptive conjoint analysis.” Rather than requiring the user to assign relative weights to all factors, it relies on trade-off questions that compare no more than two factors at a time. This simplifies the user’s task, yet the more factors used, the more trade-off questions generated. For example, a 4-factor scenario will usually generate six tradeoff questions; the number grows to 20 for a 15-factor model. While answering 14 more questions may not be an insurmountable task, most users would feel more confident with fewer questions.
A more critical issue is that of correlation. Many factors, although appearing in different categories, are actually highly correlated. For example, 5-Year Return +/- Category Index, 5-Year Alpha, and 5-Year Sharpe Ratio are calculated differently, yet they’re still highly correlated. A fund or manager that fares well on one will also fare favorably on the others – perhaps even to a similar degree. In this case, you would add very little to the scenario by including all three. The more factors you use, the greater the likelihood you’ll encounter this problem.
The biggest problem is the fact that you lose a degree of control over your scenario. If you use the three factors in the example above along with three others, your results will be greatly determined by 5-year relative returns. It’s likely you intended this but probably not to the extent that you’ll get. Even if you simply equal weight all six factors, the three correlated ones will end up representing about 50% of the scenario. If they’re all measuring one characteristic, it will end up carrying a lot more weight than you probably intended. You didn’t mean for this to happen but it did because of the factors’ correlation.
The best factor models (which are what you’re creating in a K4 scenario) use factors that are highly correlated with the results but very little with one another. The trick is to translate your investing strategy into the most representative factors. In most instances, the fewer you use, the better. Our research has shown that effective scenarios use only 4-8 factors. You might need more, but your challenge is to define your strategy with as few factors as possible.
Begin by trying to define your strategy as precisely as possible. At this point, don’t necessarily limit yourself to the factors in K4 but rather imagine explaining it to a potential client. What aspects would you emphasize and which would you say are the most important? These are the features that need to be included in your scenario. For example, “consistent long-term large cap deep value” is much easier to translate into K4 than the more general “large cap stock funds.” Interestingly, many users have reported never having gone through this process before, and even aside from the benefits of using K4, the exercise itself was informative to them.
Once you’ve clearly defined your strategy, the next step is to list the features and find the appropriate factors in K4. A simple two-column table can help do this. List each of your strategy’s features in the left column and then the appropriate factor in the right. The table below shows how this might look. Once you have this, you’re ready to build the scenario in K4.
| Criteria | K4 Factors |
| Product Type | Funds |
| Asset Type | Stock |
| Track Record | 5 Years |
| Asset Class | Large Cap/Value |
| Long-Term Return | 5-Year Return vs. Category Index |
| Representative of Category | 5-Year R-Squared to Category Index |
| Volatility | 5-Year Standard Deviation |
| Value Added | 5-Year Information Ratio vs. Category Index |
| Consistency | Rolling Batting Average (1 in 5) |
| Deep Value | Price/Book Value |
The entire process isn’t difficult as long as you have a clear idea of your strategy and what you’re looking for. The hardest part is getting over the hump of building your first scenario. Once you’ve done that, you might find a few things you’d like to change, add, or delete, but you will be well on your way to effectively translating your style. Don’t be intimidated by the number of available factors; you don’t have to use them all. In fact, you only want to use a few.
