cancel
Showing results for 
Search instead for 
Did you mean: 
     
Highlighted
Beliavsky
Participant ○

Misclassification of Bond Mutual Funds

A bond fund that takes more credit risk than its peers may outperform most of the time but do much worse in a financial crisis. Investors should avoid paying high fees for credit beta. Table 9 of the paper lists fund families with a large fraction of misclassified funds.

Don’t Take Their Word For It: The Misclassification of Bond Mutual Funds
Huaizhi Chen, Lauren Cohen, and Umit Gurun #26423
November 2019
Abstract:
We provide evidence that mutual fund managers misclassify their
holdings, and that these misclassifications have a real and
significant impact on investor capital flows. In particular, we
provide the first systematic study of bond funds’ reported asset
profiles to Morningstar against their actual portfolios. Many funds
report more investment grade assets than are actually held in their
portfolios, making these funds appear significantly less risky. This
results in pervasive misclassifications across the universe of US
fixed income mutual funds by Morningstar, who relies on these reported
holdings. The problem is widespread- resulting in about 30% of funds
being misclassified with safer profiles, when compared against their
actual, publicly reported holdings. “Misclassified funds” – i.e.,
those that hold risky bonds, but claim to hold safer bonds– outperform
the actual low-risk funds in their peer groups. “Misclassified funds”
therefore receive higher Morn! ingstar Ratings (significantly more
Morningstar Stars) and higher investor flows due to this perceived
outperformance. However, when we correctly classify them based on
their actual risk, these funds are mediocre performers. Misreporting
is stronger following several quarters of large negative returns, and
it is strong at the fund family level. We report those families that
have the highest percentage of misreported funds in the sample.

12 Replies
yogibearbull
Valued Contributor

Re: Misclassification of Bond Mutual Funds

As this is a study related to M* bond fund classifications/misclassifications, may be @RyanM can find someone from M* statistical group to comment. Most likely, M* is aware of this study and may have responded to the authors.

From what I have seen from M* posts, M* fund credit analysis is poor [or N/A] for funds that use derivatives, leverage, nontraditional strategies. I don't know if the authors considered plain vanilla bond funds separately to see how M* and their classifications differed for those.

Authors have noted that M* changed its fund credit rating methodology in 2010 [Appendix B] that lowered most fund credit ratings, and they seem to have accounted for it.

YBB
0 Kudos
Community Manager RyanM Community Manager
Community Manager

Re: Misclassification of Bond Mutual Funds

Jeff Ptak @syouth1 might have some insight here. 

0 Kudos
syouth1 Employee
Employee

Re: Misclassification of Bond Mutual Funds

Hi.

Sorry for the delay. 

A few thoughts for what they're worth:

- Data integrity and accuracy is critically important to us. Given this, we take seriously any questions that are raised about potential discrepancies in our dataset. Accordingly, we have a team of senior researchers and data experts examining the claims in this paper.

- We have not yet completed our analysis of the paper's central claim--that asset managers routinely overstate the credit quality of their holdings when compared to aggregate measures of credit quality one derives by calculating it bottom-up using the credit ratings assigned to the instruments and the associated default rates. However, our preliminary analysis casts doubt on their conclusions, which we believe have been colored by the way they have incorporated 'Not Rated' bonds into their calculation of average credit quality. We will have more to share shortly.

- With respect to the assertions the paper makes regarding the reliability of Morningstar's category classifications and Star Ratings, we strongly disagree. Putting aside our doubts about the paper's central claim (i.e., that funds overstate credit quality, misleading us), the authors seem to conflate the Morningstar fixed-income style box with the Morningstar categories that house fixed income funds. The fixed-income style box classifications have no bearing on the Morningstar category classifications. We rely upon the Morningstar category classifications when ranking funds for purposes of assigning the Star Rating. We see no evidence that we've misclassified bond funds for the reasons the paper cites.

- The paper also asserts that with funds misclassified (again, an assertion we strongly disagree with), the Star Rating would be comparing funds in the same peer group that have markedly different risk/credit-quality profiles. This assertion is questionable at best, but seems to overlook the fact that the Star Rating compares funds' *risk-adjusted* returns. Thus, if a fund is taking on scads of risk compared to others in its category, the risk-adjustment should pick this up, harmonizing the comparison accordingly. 

I hope this is useful.

Kind regards,

Jeff Ptak

Morningstar Research Services

RickatMirage
Follower ○○○

Re: Misclassification of Bond Mutual Funds

It would be interesting to know how M* treats non-rated bonds in their credit analysis. I just found one of my bond funds, as reported by M*, has 19.63 % of their assets in non-rated bonds.

0 Kudos
yogibearbull
Valued Contributor

Re: Misclassification of Bond Mutual Funds


@RickatMirage wrote:

It would be interesting to know how M* treats non-rated bonds in their credit analysis. I just found one of my bond funds, as reported by M*, has 19.63 % of their assets in non-rated bonds.


See Appendix B of the paper linked in the OP. Funds use in-house ratings for bonds not-rated by rating agencies. On average, M* considers that not-rated have credit risks similar to BB.

YBB
0 Kudos
RickatMirage
Follower ○○○

Re: Misclassification of Bond Mutual Funds

Thanks. According to the appendix of the paper, after 2010, muni not-rated bonds were classified by M* as BB and others as B.

0 Kudos
myob
Explorer ○○

Re: Misclassification of Bond Mutual Funds

I would never rely on M* data as it relates to any fund or ETF.  There are countless examples of M* that do not correlate at all to actual fund holdings or ratings.  If you want accurate data, I would strongly suggest going to the issuing funds website to get an accurate assessment of any fund.  M* is more entertainment than facts.

0 Kudos
Community Manager RyanM Community Manager
Community Manager

Re: Misclassification of Bond Mutual Funds

FYI, here is the official Morningstar response.

0 Kudos
yogibearbull
Valued Contributor

Re: Misclassification of Bond Mutual Funds

M* rebuttal is based on (i) treatment of not-rated bonds and (ii) M* Style Box vs M* Categories.

As M* assigns BB rating by default to the non-rated bond group, this may generally be higher than what detailed analysis of non-rated group may show.

The NBER paper uses terms such as styles, style categories, style classifications almost interchangeably.

FWIW, M* Styles and M* Categories have confused me too. They match mostly but sometimes don't. The way I understand it, M* Style is portfolio manager's current investment style while a fund's asset mix would determine its category. Eventually, the style and category should converge but may diverge temporarily.

But if M* Styles are not as important as indicated in M* rebuttal, why are they displayed so prominently? And why there isn't more M* literature differentiating the two? Being around M* boards for almost 10 years, if I am confused, then I am sure others are too. And clearly, NBER authors are confused on this.

 

YBB
0 Kudos
syouth1 Employee
Employee

Re: Misclassification of Bond Mutual Funds

Hi yogibearbull.

I agree it can be confusing. Here's how I think of style-box vs. M* category classification. The style-box is a snapshot at a given point in time. The category classification is a portrait--I suppose you could say a time-lapse--of what the fund's style has looked like over a period of time, which we typically define as 36 months.

Unlike U.S. equity funds, where there's a one-to-one relationship between the nine regions of the style box and corresponding Morningstar categories (i.e., LV, LB, LG, MV, MB, MG, SV, SB, SG), there's no such one-to-one relationship in fixed income. That is, we don't have nine different categories corresponding to the 3x3 combinations of duration and credit quality in the fixed-income style box a la 'Morningstar Short-term High-Quality' or 'Morningstar Long-term Low-Quality', and so forth. Yet the paper is written in such a way that equates the fixed-income style-box classifications with our Morningstar category classifications, which is incorrect.

As you know, we use the Morningstar category classifications, not the style-box classifications, to compare and rank funds for purposes of assigning Star Ratings to them. The paper asserts that because we've mis-classified funds--again, we dispute this--we've therefore also incorrectly assigned Star Ratings to funds. We disagree.

I hope this is useful.

Kind regards,

Jeff Ptak

Morningstar Research Services

0 Kudos
hku
Explorer ○○

The Morningstar Mistake Not to Make

Bonds are supposed to be the boring part of most portfolios, but there is no shortage of controversy surrounding them lately. Now, a new academic paper says Morningstar is getting faulty data, a claim the investment research firm disputes.

The faulty data can cause Morningstar (ticker: MORN) to misclassify bond funds and misapply its widely known star ratings, which, in turn, can cause investors to make bad decisions, according to the paper, which is titled “Don’t Take Their Word for It: The Misclassification of Bond Mutual Funds.” It was written by Huaizhi Chen of the University of Notre Dame, Lauren Cohen of Harvard Business School, and Umit Gurun of the University of Texas at Dallas. The authors arrived at their conclusions by comparing summary fund

Morningstar rates each fixed-income fund by putting it into a style box “based on measures of risk, together with assigning to other forms of categorization,” according to the paper.

Morningstar does have a “style box” for bond funds, capturing the underlying holdings’ credit and duration risks, but the firm said in response to the paper that the box isn’t used to place funds in its category scheme.

Morningstar’s bond style box expresses a fund’s two main characteristics—overall credit quality (high, medium, low) and duration (long, moderate, short)—as opposed to its categorization system, which assesses a fund’s holdings more finely.

For example, funds in Morningstar’s high-yield, or junk-bond, category and multisector category will both check the low-credit-quality box. But the junk-bond fund will have its entire portfolio in junk bonds whereas the multisector might have, say, 35% of its portfolio in junk bonds. (The firm’s famous style box for stock funds measures capitalization—large, mid, or small—and investment style—value, blend, or growth.)

In addition to confusing the bond style box for bond fund categories, Morningstar says that the discrepancies the authors describe are eliminated when nonrated bonds are accounted for. Some entities have credit ratings, while the securities they issue don’t, Morningstar explains. In that case, a fund company may justifiably apply a rating to the holdings, but Morningstar will show the holdings as “not rated.” The result is higher levels of nonrated bonds than those self-reported by fund companies.

As debt issuance proliferates and boomers continue to retire, shifting their portfolios to fixed income and seeking yield, the issue of what’s in a bond fund is likely to become more, not less, contentious.

0 Kudos
Community Manager RyanM Community Manager
Community Manager

Re: The Morningstar Mistake Not to Make

Threads merged.

0 Kudos
Announcements

To learn more about the recent changes to Morningstar.com, please see the updated FAQ.

See recent posts and all our forums or access the old forums here.
 

You can read the community guidelines in