Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

In this Discussion

Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.

    Support MFO

  • Donate through PayPal

Where can I find annual mutual fund performance data for 25 years?

It appears that M*'s annual mutual fund performance data only goes back ten years. If I'm wrong, please tell me how to unlock "older" annual performance or where I might find annual data going back 25 years or longer. I want to see how each fund performed at least back 25 years and preferably even longer. (And yes, I realize that the older the data, the greater the likelihood of changes in management, etc. that could discount the value of some historical data.)


  • edited February 19
    You can deduce this type of info from charts.

    M* Charts go quite far back although they are acting up this morning (try later from the following link).

    Portfolio Visualizer data go back tp 1985. PV LINK
  • Thank you, Yogi. Exactly what I needed.
  • msf
    edited February 19
    You can also use M*'s new "Interactive Charts". You'll find them on the home (quote) page for each fund. You need to click on the "Show Interactive Chart" button at the upper left of the graph shown.

    While I prefer M*'s legacy pages for most purposes, the interactive charts have the advantage of showing you the cumulative gain in percentage between the dates you specify. So if you give dates of 12/31/97 and 12/31/98, it will tell you the annual gain for 1998 without your needing to do the long division.

    Yahoo's finance pages also provide figures that one can use to deduce annual returns. Select the historical range of data to cover the years of interest. Then look at the adjusted price on Dec 31 of successive years and divide to get the growth for the selected year. Adjusted prices incorporate the effect of dividends, splits, etc., so it represents total return. You can download the data and let Excel do the division for you.

    There are minor differences. For DODGX, the M* interactive chart reports a gain from the end of 2020 (12/31/2020) to the end of 2021 (12/31/2021) of 31.73%. This is also what M* reports in digital form on the fund's performance page, confirming that these are the right endpoints to use.
    M*'s performance page for DODGX

    Yahoo Finance reports adjusted closing prices of 186.20 and 245.26 at the end of 2020 and 2021 respectively, for a gain of 31.72%. Going to the horse's mouth, D&C reports a 2021 return of 31.68%.
    Yahoo Finance, DODGX data
    D&C performance page

    This illustrates why I try to go as far upstream as possible to the data source if accuracy is important. One can find 10 years of performance data in a fund's prospectus, so by looking at a prospectus that's 15 years old one can get the annual returns for years 1997-2006, and by looking at a prospectus that's 5 years old one can get the annual returns for years 2007-2016.

    SEC fund filing search page:

    For example, for DODGX, here are the bar charts for those 20 years and the prospectuses they come from:
    image image

    Dodge and Cox Prospectus, May 1, 2007        Dodge and Cox Prospectus, May 1, 2017
  • @msf, good points.

    Both Yahoo Finance and Stockcharts use adjusted-prices (ratio-adjusted for distributions). These provide good enough approximations for cumulative or annualized TRs for up to 10 years, in my experience. Beyond 10 yrs, the approximation errors become noticeable, but still OK for most purposes. Yahoo Finance often delays making adjustments, or, some skipped ones are never fixed. As internal details are not visible at Stockcharts, I keep my fingers crossed there.

    It bugs me that Yahoo Finance provides adjusted-prices in tabular form but charts only actual-prices. So, Yahoo Finance charts are mostly useless except for short-term. Why not provide charts for BOTH actual-prices AND adjusted-prices like Stockcharts does (with _TICKER and TICKER, respectively).

    The SEC/Edgar is a great resource. Each fund prospectus and semiannual/annual report has 5-10 years of TR data and one can go back years. This does require lot of patience and hard work that can be repeated manually only for a handful of funds.
  • Does the M* Interactive chart assume reinvestment of dividends for the fund being analyzed? And does it also assume reinvestment of dividends for the fund being compared, whether it is an open-end fund or an ETF?
  • edited February 19
    @Jim0445, now it works fine. Original and comparisons in Interactive-Charts now have reinvestments. Both mutual funds and ETFs also have reinvestments. The 1st version had the flaw that it did reinvestments for mutual funds but not for ETFs, but now that is fixed. What I still don't like is that these Interactive-Charts cannot be linked via direct website URLs but one must use screenshot, image-host, MFO Insert-Image icon (I did a post on steps for posting images

    SP500 ETF and OEF Max-view, Image URL at host

  • Yahoo Finance shows the yearly returns and returns by quarter on their performance tab. For example, FGMNX has had 2 losing years -2.17% in 2013 and -2.00% in 1994. 2Q 2013 and 1Q and 2Q 1994 were when the biggest losses occurred.
  • I hadn't looked at the performance tabs on Yahoo. That's a really nice feature.

    Now, if Yahoo would only report accurately. FGMNX had three losing calendar years: the two you mentioned and also 2021. Yahoo show 2021's return as N/A, though it knows better. Yahoo gives December 31 adjusted closing figures as 11.58 (2021) and 11.68 (2020) for a loss of 0.85%, matching Fidelity's official figure.

    FWIW, according to the Yahoo performance tab, the fund's worst calendar quarters were in 1987: 2Q (-2.48%) and 3Q (-3.40%). The worst three month drawdown, irrespective of month boundaries, was around 6½%, from the close on July 17, 1987 to the close on Oct 16 (a Friday) or Oct 19 (a Monday) 1987. Nearly double the calendar quarter max loss.

    I got this by downloading the daily adjusted close figures from Yahoo and playing with Excel to approximate quarterly returns day by day.
  • @msf, +1
    I also found other inconsistent data in the performance tab to deem unreliable.
  • Excellent discussion, everyone, once again demonstrating the value of this site. Thanks.
  • M* is the provider for Yahoo's performance tab data.
    Considering the data source, it does not surprise me that some posters have discovered inaccuracies.
  • Given that M* provides a figure for FGMNX's 2021 performance and Yahoo reports N/A, perhaps it is how Yahoo transcribes numbers and not M*'s data that is the problem? Or perhaps Yahoo calculates its own figures (e.g. adjusted closing prices) from data provided by M* and it just doesn't know how to program accurate financial calculations?

    Then there's FGMNX's YTD performance as of now (Feb 22, 2022 close). The authoritative figure from Fidelity is -2.42%. The M* figure on the fund's quote page is -2.42%. Yahoo's fund summary page reports -2.43%.

    One of these is not like the others.
  • I have noticed several things about Yahoo Finance (YF) data:

    1. Data-feed errors are generally not fixed. This I have noticed at other sites too. Reason probably is that it is pointless to correct data-feed errors manually as the next data-feed refresh may just restore those. Fund families whose data are involved don't care - this I KNOW from my prior emails to YF and fund families.

    2. Yahoo Finance does process some of its data. Adjusted-prices is something unique to Yahoo Finance (and Stockcharts, etc). Not everyone is sold on this concept (others do Growth-of-10K). Another area is Treasury rates ^TNX, ^TYX, etc where YF doesn't follow the typical 10x rate scale convention (CBOE, Stockcharts, etc) and that leads to some fantastic transient error in that data at YF on some afternoons.

    3. I also suspect that in adjusted-price calculations, YF rounds results to 2 decimals in EACH step. This rounding error then propagates and becomes noticeable for periods larger than 10 years. Conceptually, the results from adjusted-prices and Growth-of-10K should be closer to what they actually are.

    Under Data Disclaimer, Yahoo Finance provides multiple sources of its data.

    "US quotes are real-time for NASDAQ, NYSE, and NYSE American when available from Nasdaq Last Sale and if not available it will appear delayed from the consolidated tape. See delay times for other exchanges below. Quotes are updated automatically but will be turned off after 25 minutes of inactivity.
    Financial statements, valuation ratios, market cap and shares outstanding data provided by Morningstar.
    Company profile data provided by S&P Global Market Intelligence.
    US equities and global index historical data and daily updates provided by Commodity Systems, Inc.
    International historical chart data and daily updates provided by Morningstar.
    Analyst estimates, earnings, corporate, economic events, non-US IPO, and insider transactions data provided by Refinitiv*.
    Top institutional and mutual fund holders provided by
    SEC Filings and US IPO data is provided by EDGAR Online, a division of Donnelley Financial LLC.
    Sustainability data provided by Sustainalytics and Morningstar.
    Upgrades and downgrades provided by Benzinga.
    Corporate governance scores provided by Institutional Shareholder Services."
  • I agree with your suspicion in #3 (rounding at each step). It seems so apparent that I've never bothered to check this hypothesis.
  • edited February 23
    @msf, I had a professor once who insisted that if the results were claimed to be accurate to n-significant digits, then EACH step of the calculation must be rounded to n-significant digits. This may have been the limitation of the old slide-rules. I initially argued that intermediate steps of calculations could be carried out with much larger number of digits (that calculators or computers allowed) and only the final results can be presented to n-significant digits. We disagreed on this but I went along with that professor's approach for THAT course only, but never later. I think that the Yahoo programmer(s) may have been a student of that same professor.
Sign In or Register to comment.