Deciding to Sell a Mutual Fund
(This is a background discussion. If you don't want to read it online,
you might want to print it out and review it later. It will run about 9 pages.)
[Home]
How do I decide to sell a mutual fund? We believe that the decision to sell a mutual fund should begin with performance, and the best way to evaluate the peformance of any mutual fund is to compare it to an appropriate benchmark. A fund that consistently underperforms its benchmark is a strong candidate for sale.

What is a benchmark? A benchmark is a standard of measurement for mutual fund performance. Benchmarks come in many different varieties, but an index or index mutual fund is the preferred benchmark for most professional money managers and financial advisors. For example, the Standard & Poor's 500 stock index or the Vanguard 500 Index mutual fund can be good benchmarks for evaluating the performance of a mutual fund that invests in large capitalization U.S. stocks.

What benchmarks should I use to evaluate my funds? There's no easy answer to this question. For example, if you classify mutual funds by objective (e.g., "growth," "growth and income"), you would probably want benchmarks that mirror those objectives. Similarly, if you classify mutual funds by the capitalization of stocks that they own (e.g., large-cap, small-cap), you would probably want benchmarks classifed by capitalization. Since we view benchmarking as only the first step in the potential sell decision, we believe you should seek a benchmark that is reasonably close to the fund you are evaluating; however, there is little point in seeking the "perfect" benchmark (you won't find it anyway!).

How many different benchmarks should I use? Again, there's no easy answer. At one extreme, you could conceivably use ten different benchmarks to evaluate ten funds; at the other extreme you could use only one benchmark. However, if you apply benchmarks consistently, and interpret the results carefully, you should be able to evaluate every mutual fund that you own with a universe of just five or six benchmarks.

Let's say I've identified my benchmarks. Over what period should I compare the performance of my funds to their benchmarks? As a general rule, we suggest that you avoid performance comparisons over very short or very long periods (for example, one month or ten years). If we had to recommend a single performance period, we would suggest that you compare the performance of each mutual fund to its benchmark for the past three years (36 months). Even better, we think, is to compare a fund's performance to its benchmark over several periods -- ideally 12 months, 36 months (three years), and 60 months (five years). When you examine multiple periods, you get more depth and perspective on the performance of your fund.

How does FundAlarm present benchmark information? From the FundAlarm Home page, you can access benchmark performance data by clicking on the link that says "View fund tables." (You can also go to this page directly: http://www.fundalarm.com/search1.htm.) FundAlarm presents a separate data table for each fund, a sample of which appears below (the benchmark section is outlined in black):



The negative numbers in cells 3a, 3b, and 3c tell us that this fund has underperformed its "best" benchmark for the past 12 months, three years, and five years (by 12.73%, 6.74%, and 7.26%, respectively). Because this fund has three periods of underperformance, FundAlarm designates it as a 3-ALARM fund (cell 3d). Looking down one row (cell 7), we note that the benchmark used to evaluate this fund's performance is the Vanguard 500 Index mutual fund ("Vang 500 Idx").

What benchmarks does FundAlarm use? FundAlarm draws from a universe of six potential benchmarks, as follows:

Benchmark	Used for:	Description of Benchmark
Vanguard Balanced Index Fund	Balanced funds	A mutual fund that invests about 60% of its assets in large stocks and about 40% in high-quality bonds. The stock portfolio attempts to match the performance and risk characteristics of the Wilshire 5000 index. The bond portfolio attempts to match the performance and risk characteristics of the Lehman Brothers Aggregate Bond Index.
Schwab International Index Fund	International funds	A mutual fund that seeks broad international equity diversification.
FundAlarm's Specialty Benchmarks	Specialty funds	Designed by FundAlarm for evaluating the performance of "specialty" mutual funds. There are eight specialty fund benchmarks, one for each of eight specialty fund categories (communications, financial, health, precious metals, natural resources, real estate, technology, and utilities).
Vanguard 500 Index Fund	Domestic large-cap funds	A mutual fund that attempts to match the performance and risk characteristics of the S&P 500 stock index
Dreyfus Mid Cap Index Fund	Domestic mid-cap funds	A mutual fund that seeks to match the performance of the S&P MidCap 400 Index
Vanguard Small Cap Index Fund	Domestic small-cap funds	A mutual fund that seeks to match the performance and risk characteristics of the Russell 2000 index

How do you select a "best" benchmark for each fund?

• FundAlarm first reviews each mutual fund in our database to determine if it is a "balanced," "international," or "specialty" fund.
  
  
• If a fund falls into one of these three categories, it is assigned the "balanced," "international," or "specialty" benchmark, as appropriate.
  
  
• Each remaining fund is then analyzed to determine its "median market capitalization."
  
  
• After "balanced," "international" and "specialty" benchmarks have been assigned (as described above), mutual funds with a median market capitalization up to $2.0 billion are assigned to the small-cap benchmark (i.e., the Vanguard Small Cap Index fund); funds with a market capitalization from $2.0 billion to $9 billion are assigned to the mid-cap benchmark (i.e., the Dreyfus MidCap Index Fund); and funds with a market capitalization in excess of $9 billion are assigned to the large-cap benchmark (the Vanguard 500 Index fund).

What else should I know about the FundAlarm benchmarks? Of necessity, the capitalization thresholds for assigning small-cap, mid-cap, and large-cap benchmarks are somewhat arbitrary and artificial. For example, we would assign the mid-cap benchmark to a fund with a median market cap of $8.9 billion, while we would assign the large-cap benchmark to a fund with a market cap of $9.1 billion. In reality, these two funds might be quite similar, but the slightly larger fund would be evaluated against the large-cap benchmark, which in recent years has been considerably more difficult to outperform than the mid-cap benchmark.

If the median market cap of your fund is close to the $2.0 billion or $9.0 billion thresholds, we suggest that you also evaluate the performance of your fund against the benchmark that is immediately above or below. In our example above, the owner of the fund with $8.9 billion market cap might want to give some consideration to the large-cap (i.e., higher) benchmark. Similarly, the owner of the fund with $9.1 billion market cap might want to give some consideration to the mid-cap (i.e., lower) benchmark.

What is the practical significance of a 3-ALARM fund? In general, a 3-ALARM fund is a strong candidate for sale, but a 3-ALARM fund is not necessarily an automatic sale. This is a very important point! In addition to the 3-ALARM designation, you should consider other information about your fund before making the "sell" decision, such as: How risky has your fund been? How has it performed in relation to its peer group? How long has your fund had the same manager? What is your fund's median market capitalization? How rapidly is your fund gaining or losing assets? All of this additional information can be obtained from the FundAlarm data table.

Should I be alarmed if I own a 3-ALARM fund? Not necessarily. For example, many index funds are 3-ALARM funds, but only by a slight margin. If you own a 3-ALARM index fund, or any other fund that is 3-ALARM by a slight margin, you can probably ignore the 3-ALARM designation.

Can I relax if my fund isn't 3-ALARM? Not necessarily. For example, your fund may have underperformed its benchmark for the past 12 months and three years, and only narrowly beaten its benchmark for five years. To us, this looks a lot like a fund about to go 3-ALARM, and you might not want to stick around while it does. Here's a good rule: If a mutual fund has two periods of underperformance, put it on your "watch list," and check FundAlarm every month. If performance continues to lag for 9 to 12 months after it goes on your watch list, consider selling, even if the fund isn't officially 3-ALARM.

How does "risk" enter into the mutual fund sell decision? If you want to start a lively discussion among financial professionals -- itself a rare occurrence -- ask each of them to say a few words about "risk" in relation to mutual funds. It's unlikely that any two professionals will agree on the definition of risk, let alone the measurement of risk or its importance to mutual fund investors when making the "sell" decision.

Cutting through all the academic arguments and statistics, we think most people perceive mutual fund risk as the chance of having a "down" year, as well as the magnitude of the potential loss. But mutual fund risk must be viewed in context. For example, let's say you own Fund A and you are comparing it to a benchmark index fund (Fund B). If Fund A has historically outperformed Fund B, you might want to keep Fund A even if there was a good chance that Fund A was going to have a worse "down" year than Fund B. On the other other hand, if Fund A has historically performed worse than Fund B, and there was also a good chance that Fund A was going to have a worse "down" year than Fund B, why would you continue to hold Fund A? In both cases, Fund A is "riskier" than Fund B, but in the first case the greater risk of Fund A has been rewarded with greater return. In the second case, both risk and return are working against Fund A, and Fund A is a compelling candidate for sale.

How does FundAlarm present risk information? The presentation of risk in FundAlarm is built around four numbers, which represent a risk spectrum. Every fund (except in the Specialty category) is assigned one of these four risk ratings: -2 (least risky), -1, +1, and +2 (most risky).

To assign risk ratings, FundAlarm first computes the potential "down year" for each fund, which is equal to the fund's three-year return minus its three-year standard deviation. For example, a fund with a three-year return of 30.00%, and a three-year standard deviation of 20.00%, would have a potential "down year" of 10.00%.

Within each benchmark category, funds are then arranged in descending order of their potential "down year." Recently, for example, there were approximately 1,000 large-cap funds in the FundAlarm database, and their "down years" ranged from approximately +30.00% to -32.00%, as shown in the left and middle columns of the following graphic:

The "down year" of the benchmark fund in each category is identified and that becomes, in effect, the pivot point for assigning risk ratings to other funds in the category. In the graphic above, the "down year" for the benchmark fund is +5.00%. Those funds with potential "down years" better than the benchmark are split into two equal-sized groups, and they are assigned either a -2 or -1 risk rating. Those funds with potential "down years" worse than the benchmark are also split into two equal-sized groups, and they are assigned either a +2 or +1 risk rating. The right-hand column of the graphic above demonstrates this part of the process.

It should be noted that risk ratings are only relevant within each benchmark category. Thus, it is correct to say that a large-cap fund with a risk rating of +2 is considerably riskier than another large-cap fund with a risk rating of -2. However, it is not meaningful to compare the risk of a +2 large-cap fund with (say) a -2 international fund.

Should I be concerned if I own a fund that has a risk rating of +1 or +2? Not necessarily. Some "risky" funds are also excellent performers -- perhaps even "NO-ALARM" funds -- and in those cases you might conclude that your fund's extra risk has been rewarded with extra return.

What if I own a fund that has a risk rating of +2 and is a 3-ALARM fund? Now you are getting into danger territory. A 3-ALARM fund that also has a risk rating of +2 is the least desirable combination in FundAlarm, and it becomes a very strong candidate for sale -- but still not automatic.

Why wouldn't I sell a fund that has a risk rating of +2 and is also 3-ALARM? One reason might be that your fund has outperformed its peer group, which you can tell from cells 9a, 9b, 9c of the data table:



Our sample fund has, in fact, underperformed the average of its peer group for the past 12 months, three years, and five years ("Lower Lower Lower"). An indication of "Higher" in any of these cells would indicate that the fund had performed better than the average of its peer group.

What other factors enter into the mutual fund "sell" decision? There are at least two other factors to consider: manager turnover and change in fund assets.

• Manager Turnover. By itself, manager turnover is not a reason to sell a mutual fund. However, information about manager turnover is an important component of the sell decision, since turnover information can help you interpret fund performance data. In fact, performance information is much less useful in the absence of turnover information, and vice versa. In some cases, a change in fund management can explain a pattern of deteriorating performance, and reinforce a tenative decision to sell that was based on performance information. In other cases, a recent change in fund management may encourage you to hold onto a fund that has performed poorly.
  
  FundAlarm contains two sources of information on manager tenure: a column in each fund data table that shows manager tenure, and a separate section of news items ("Mutual Fund Manager Changes") that tracks manager changes since July, 1996, which was the the inception of FundAlarm.
  
  In the FundAlarm data table, information on manager tenure looks like this:
  
  
  Thus, cell 4 tells us that the manager of this fund has a tenure of 19 years.
  
  If your fund has underperformed its benchmark for the past 12 months/three years/five years, and the tenure of your fund manager is five years or greater, you know that the current manager was entirely responsible for the poor performance record. In our opinion, this is another strong signal to sell your fund -- you've given one manager at least five years to beat the benchmark, and that manager has failed to do the job. This is especially true if your fund is "riskier" than its benchmark (i.e., a risk rating of +1 or +2).
  
  Of course, it's also possible to face a situation in which your fund has performed poorly for the past 12 months/three years/five years, but the manager has been on the job for only a short period of time. In such a case, you can't hold the current manager entirely accountable for the fund's poor performance, but that doesn't necessarily mean you should continue to hold the fund. If the new fund manager has a solid track record at another, similar fund, you might want to continue to hold. If the new manager has never managed a fund before, or has a relatively poor record at another fund, you should probably go ahead and sell. How do you find out about a new manager's previous history? Our page of "Mutual Fund Manager Changes" has some of this information, and you can usually get additional information directly from your fund.
  
  Finally, you are likely to face some situations in which it is difficult to relate performance information to manager tenure. For example, consider a large-cap fund where the manager has a tenure of 2 years and the fund has the following performance information relative to the Vanguard 500 Index:
  

  5 Years		-4.5
  3 Years		-3.0
  12 Mos		2.0

  
  In this case, it is reasonable to credit the good 12-month performance to the new manager, and to attribute most of the poor longer-term performance to the previous manager. But is 12 months enough time to form a valid opinion of the new manager? Probably not, if you were buying this fund for the first time. But if you already own the fund, you might want to continue to hold it, and keep a close eye on the new manager's performance as he or she reaches the three-year and five-year performance milestones.
  
  
• Fund Size/Change in Net Assets Under Management. This information is important because some mutual fund managers are better at managing relatively small sums of money, and the performance of these managers may suffer as their asset base grows. Other fund managers may find that they can't buy their favorite stocks when they have large sums of cash to invest. Either way, rapid growth of assets under management can help explain a mutual fund's poor recent performance. Rapid loss of fund assets can also help account for poor performance, although not as directly.
  
  For example, consider ABC Small Cap Fund, a hypothetical fund that grew from $50 million to $1 billion of assets under management. Let's assume that, at any given time, the manger of ABC Fund has 50 good investment ideas -- in other words, a maximum of 50 stocks on his "A List." When ABC Fund had $50 million in assets, the fund manager could take an average position of $1 million in each of the 50 stocks on his A-List. Now that the Fund has $1 billion to invest, the average position must grow to $20 million. In fact, the manager of ABC Fund is faced with several possible choices, none of them particularly pleasant or desirable from the standpoint of the Fund shareholders:
  
  
  • He can take $20 million (or larger) positions in several of his best choices. However, since many small companies have total market capitalization of only $200 million or so, ABC Fund would probably end up owning 10% or more of the stock in several companies. Large positions like this are undesirable for liquidity reasons, and there are legal restrictions on how concentrated a mutual fund may become.
    
    
  • The ABC Fund manager can expand his stock list beyond 50, by considering less attractive investment opportunities -- in effect, move to his "B-List."
    
    
  • He can raise the median capitalization of his holdings, even though he is really only expert in selecting small cap stocks.
    
    
  No matter what the ABC manager does, fund performance is likely to suffer.
  
  In the FundAlarm data table, information about change in fund assets appears as follows:
  
  
  According to cell 10a, this fund most recently had $989 million ($MM) of assets. Three months ago ("-3 Mo."), this fund had $1,043 million of assets (cell 10b), which means that the asset base of this fund has decreased 5.2% (cell 10c) in three months.
  
  When evaluating information on change in assets, you should keep your fund's market cap in mind. As a general rule, small-cap and specialty funds have the hardest time handling large increases or decreases in assets, and large-cap funds have a relatively easy time handling asset swings of 10%, 20%, or more (which don't happen very often, anyway). If you own a small-cap or specialty fund with a changing asset base and deteriorating recent performance, you might be concerned. With a large-cap fund, the changing asset base probably wouldn't be much of a factor in any potential sell decision.

Please walk me through a comprehensive example, showing how to use the complete FundAlarm data table. Here's the entire data table for our sample fund, which we call the XYZ Fund, but which is taken from an actual fund data table. Let's review this table step-by-step, and see how you might analyze your own funds:

• We suggest that you always start by reviewing cell 3d for 3-ALARM status. In this case, the XYZ Fund is 3-ALARM, and by quite a wide margin (cells 3a, 3b, 3c). We also note that the XYZ benchmark is the large-cap Vanguard 500 Index (cell 7), which we will return to in a moment.
  
  
• 3-ALARM status makes us nervous, but we want to know more. We look to the right of cell 3d, to cell 5, and we note that the XYZ Fund has a risk rating of +2. This is a bad combination -- 3-ALARM and +2 -- but we're not ready to say "sell."
  
  
• We look down to cells 9a, 9b, 9c and note that XYZ has failed to match the return of its peer group over the past 12 months, three years, and five years ("Lower Lower Lower").
  
  
• Prospects do not look good for the XYZ Fund, but we still want more information. Perhaps XYZ is on the threshold of large-cap and mid-cap, and fund performance would look better if we could compare XYZ to the mid-cap benchmark? Our eye goes to cell 6, and we note that XYZ holds stocks with a median market cap of $68 billion. Conclusion: XYZ is indeed a large-cap fund, and the large-cap benchmark is appropriate.
  
  
• Perhaps XYZ has a relatively new manager, and the old manager was responsible for its poor record? We look at cell 4, and note that XYZ has had the same manager (or team structure) for 19 years. Unfortunately for XYZ, we conclude that the current manager must take all the blame for its track record.
  
  
• We're down to our last piece of data: change in fund net assets (cells 10a,10b,10c). If our fund were a relatively small one, and it was growing or shrinking rapidly, poor performance (especially poor recent performance) might be attributable to the rapid movement of cash. However, in the case of the XYZ Fund, we're dealing with a relatively large fund ($989 million in assets, from cell 10a), which has recently been shrinking at a rate of just 5.2% (cell 10c). It's unlikely that decrease in assets under management offers any explanation for the poor performance of the XYZ Fund.
  

We have now reviewed the entire data table for the XYZ Fund, so let's briefly recap what we have found:

• This is a true large-cap fund, and it has underperformed the large-cap benchmark by a wide margin (3-ALARM).
  
  
• The fund is considerably riskier than its benchmark (+2).
  
  
• It has underperformed its peer group (large-cap "blend" funds) for the past 12 months, three years, and five years ("Lower Lower Lower").
  
  
• The management team currently in place has been responsible for at least five years worth of poor performance.
  
  
• The fund's performance cannot reasonably be explained (even in part) by a rapid change in fund assets.
  
  

Based on the above, we would conclude -- very strongly -- that the XYZ Fund should be sold.

---

[Top | Home]

FundAlarm © Roy Weitz, 1997-2003