M* recently reclassified VPMCX as large cap blend. It had been classified large cap growth even though its portfolio had been in the blend column since 2018.
Its performance looked very poor compared with its "peers": 50th percentile (2018), 84th percentile (2019), 94th percentile (2020). (Though in 2021, with value ascending, it looked great - somewhere in the top 5%).
As a result of its long term "poor" performance, it had been rated 2 or 3 stars earlier this year. But with the reclassification, it's suddenly a five star fund. Nothing has changed. Its
half brother POGRX is a bit more growthy, so M* has left it in the LCG category. As a result, that fund sports a 2 star rating.
An example how even unintentional drift affects star ratings is FSMAX. It tracks the S&P 500 completion index. M* writes: "This has been one of the strongest performers in the mid-cap blend category over the trailing 10 years through July 2020." But it has been on the blend/growth boundary since at least 2017, and M* recently reclassified it as growth. So this perennially strong performer is now rated 2 stars.
I'm not faulting M* here. Funds that do not sit near the center of a style "box" have a good chance of being over- or under-rated. This will happen regardless of what box they're dropped into.
Comments
This is just a remarkable, damning summary, which I repost for emphasis:
... reclassified VPMCX as large cap blend. It had been classified large cap growth even though its portfolio had been in the blend column since 2018.
Its performance looked very poor compared with its "peers": 50th percentile (2018), 84th percentile (2019), 94th percentile (2020). (Though in 2021, with value ascending, it looked great - somewhere in the top 5%).
As a result of its long term "poor" performance, it had been rated 2 or 3 stars earlier this year. But with the reclassification, it's suddenly a five star fund. Nothing has changed.
I've had a similar experience with MIEIX.
On 09/30/20, MIEIX was classified as a Foreign Large Growth fund with a 3 star rating.
The corresponding category rank was: 5 Yr - 66; 10 Yr - 49; 15 Yr - 24.
On 12/31/20, MIEIX was classified as a Foreign Large Blend fund with a 5 star rating.
The corresponding category rank was: 5 Yr - 6; 10 Yr - 5; 15 Yr - 3.
Although no material changes were made to the fund, it "improved" considerably!
If a fund is "miscategorized", this rating may be of little value.
Unfortunately, some investors overly emphasize star ratings when buying or selling mutual funds.
John Rekenthaler from M* discusses the star rating in the following article.
Link
https://www.morningstar.com/articles/754147/morningstar-categories-introduction-update-2016
M*'s style box was created in 1996 (according to most sources). But until 2002 funds were rated according to broad categories, as Rekenthaler stated.
All ways of benchmarking performance - using a large universe (e.g. US equity), using more precisely defined peers (style boxes), using an index blend that matches the fund's portfolio - have weaknesses. My intent was to highlight a weakness of comparing a fund with a "box" of peers: instability.
In addition, the quality of the comparison degrades as one moves toward a boundary even without crossing it. This is something I have posted on a few times in the past year prior to the market rotating to value.
Funds in an given style box, say blend, that tended toward value generally ranked more poorly than their peers that had a growth leaning. Normally this distortion effect isn't pronounced within a box. But because the gradient between growth and value was so large in the past couple of years one had to keep it in mind when looking at fund ratings.
If you want stability, you can go back to M*'s old way of rating funds. Compare all US equity funds together. A fund can't drift out of this "box" because there's only one box. But then the possibility of ratings distortions within this one box become much larger. This is what Rekenthaler is calling "style effect".
If you rate a fund against a hypothetical index benchmark blended to match the fund's portfolio (as I believe S&P is doing, or at least was doing years ago), you lose the sense of how well the manager is positioning the fund in its investment universe. What you get is just a measure of specific security selection, e.g. was the decision to invest solely in Coke rather than in the market mix of Coke and Pepsi a good one?
If you like, fault M*'s choice of methodology (style boxes). Though each methodology has its weaknesses. The discontinuity in a fund's star rating is inherent in the style box methodology. M*'s quality of execution doesn't create that discontinuity, but rather determines when discontinuity manifests.
In 2018, MIEIX returned -10.66%. M* says that placed the fund at the 19th percentile. That's the same percentile as ARTIX was ranked, with its -10.86% return.
In 2019, MIEIX returned 28.40%, only good enough for a 46th percentile ranking. That's the same percentile as TWEIX got with its 28.37% return.
http://performance.morningstar.com/fund/performance-return.action?t=MIEIX
http://performance.morningstar.com/fund/performance-return.action?t=ARTIX
http://performance.morningstar.com/fund/performance-return.action?t=TWIEX
However, on those same pages, annual performances are compared with the fund's current category, not the category the fund was in for each past year.
I have very different bonds in my tax exempt IRAs, in which I take more risk, assume more volatility, and seek to produce higher total return than in my taxable account. Again, I ignore stars and medal information, in selecting those differing funds, from a wider array of M* categories.