Inside Morningstar’s New Strategy Indexes

On Monday, I wrote a post about a new Canadian dividend fund from XTF Capital. That fund is one of five new products from the newest ETF provider in the country, all of which are based on indexes provided by Morningstar Canada. Two more began trading on Wednesday: the XTF Morningstar Canada Value Index ETF (FXM) and the XTF Morningstar Canada Momentum Index ETF (WXM).

In my original post, I criticized XTF Capital and Morningstar for not making the index methodologies public. It turns out that wasn’t quite fair:  XTF president Barry Gordon contacted me and explained that the index construction rules are indeed fully transparent, and although they were not on the XTF website at the time, they are now.

I ended up speaking with Mr. Gordon about his company’s partnership with Morningstar, and it’s an interesting story. Although XTF Capital is a small firm compared with its competitors, it has partnered with one of the most well-known research firms in the investment industry, and the result is some unique and interesting ETFs. They are all passively managed, but they use rules-based strategies designed to deliver better risk-adjusted returns than the overall market.

A new family of strategy indexes

XTF Capital approached Morningstar and asked them to create indexes based on their Computerized Portfolio Management Services. CPMS is a service that Morningstar sells to advisors and portfolio managers who are interested in executing particular investment strategies—such as those based on dividends, value factors, or momentum. Rather than doing their own security analysis, subscribers pay Morningstar a fee for comprehensive research and recommendations.

The long-term performance of the CPMS strategies are impressive—albeit hypothetical. Their Earnings Value strategy, which is the basis for the new value ETF, returned 19.35% over the last 10 years, compared with 7.54% for the S&P/TSX Composite. The CPMS Momentum strategy claims an annualized return of more than 20% since 1985.

Gordon readily admits that investors should not expect returns like that: rules-based strategies don’t always live up to their promise in the real world, where they face the headwinds of management fees, transaction costs, market impacts and human behaviour. The ETFs will have management fees of 0.60%—not including HST—and will incur trading costs and other expenses that will lower their returns compared with the CPMS numbers.

Moreover, Morningstar had to make sure that its new indexes were truly investable, which meant eliminating small, illiquid stocks and rebalancing quarterly rather than in real time. Gordon says the company will not announce the exact rebalancing dates to discourage front-running.

As for human behaviour, that’s one area where the new indexes might prove to be superior to your average portfolio manager. Gordon says he knows many people in the industry who subscribe to CPMS, “but I rarely find anyone who does exactly what the methodology tells them.” Like most human beings, fund managers tend to ignore data or recommendations that go against their intuition—and usually they’re wrong. “So it will be interesting to see what happens in the index context, where there will be no discretion,” Gordon says. “We can’t override it and say, ‘No, that’s too heavy a weighting to X.’ The index will generate whatever it generates, and we will replicate that.”

That’s music to the ears of a passive investor.

35 Responses to Inside Morningstar’s New Strategy Indexes

  1. Kiyo February 16, 2012 at 9:00 am #

    Tracking errors and management fees are small factors, a few percent at most.

    I would argue that the real reason you can’t expect the index to get the full 20% is because the computer algorithm was trained with the same data it was tested with. The analysts at Morningstar had all the market data from 1985 at their disposal and used it to train a computer to make money in that market aver that period. They then tested the algorithm using the SAME DATA. Guess what? The computer made lots of money. Hindsight is 20/20. Lets test the algorithm on the Tokyo Stock Exchange from 1949 to 1985 and see if it still beats the market!

    Another factor is that information makes the market more efficient. The only way to beat the market is to through market inefficiencies. The analysis found some inefficiencies in the market that could have been used to make money. However, the cat is out of the bag now. By finding the loopholes, people will start using those strategies, and they cease to be winning strategies.

    Still, I’ll pick a biased computer over an emotional human any day.

  2. Barry Gordon February 16, 2012 at 9:23 am #

    @Kiyo. I respect your skepticism about backtested results. The key difference here is that the track records that Dan is referring to are actual historical portfolios generated by CPMS – in other word those track records reflect the performance of the portfolios they have recommended over time. The survivor bias doesn’t exist because the track record is the actual stocks picked during the time frame. CPMS has long operated as a reliable data scrubber, specializing in compiling value oriented data and earnings reports, earnings surprises and price momentum, in particular. The indexes are not trying to exploit inefficiencies, just pick stocks that have strong fundamentals and technicals that CPMS/Morningstar expects to perform better, with less risk. Simply put, we’re trying to take the best of ETFs – transparency, low cost and liquid – and combine them with some disciplined traditional asset management tools (the CPMS screens), to create ETFs which strive to deliver a better risk adjusted return. We obviously cannot promise anything and I understand your reservations. We and Morningstar can only stand behind the integrity of the process. Time will tell how successful we are.

  3. Kiyo February 16, 2012 at 10:09 am #

    @Barry. So your saying that CPMS has existed since 1985, and that quoted returns are the ones you would have gotten if you had of followed CPMS’s advise–at that time–with the algorithm that existed–at that time (up to fees, tracking error, etc.)? In that case I misunderstood the original post and completely take back most of what I said in my last post as it is irrelevant.

    I would still like clarification on one point. You say the new ETFs “strive to deliver a better risk adjusted return.” In a perfectly efficient market, the ONLY way to get better returns is to take on more risk. I guess this is a very theoretical question about fundamental indexing (which I admit I know almost nothing about).

    You say “Time will tell how successful we are.” That is a fairly conservative statement, given that the index has a long real-time track record. Assuming you aren’t changing the algorithm, the new ETFs are just new packaging.

  4. Barry Gordon February 16, 2012 at 10:22 am #

    @Kiyo. I guess in theory you are correct – in a perfectly efficient market, everyone would have instant knowledge of all material information affecting the value of securities at any given time, and you could never beat a cap weighted index. The efficient capital markets hypothesis has been debated at great length.

    One point I want to be perfectly clear on – these indexes are not identical to the CPMS track records. Morningstar had to make changes to them to make them investable as indexes – in other words there are liquidity thresholds, and some of the factor weightings were changed to, among other things, generally tilt the strategies towards liquidity. If you go to the website — http://www.xtf.ca — you will see the index historical returns are generally less than their CPMS counterparts because of these changes. Also check out the methodologies. But the process and the screens are the ones that generated those track records. We believe in them and hope investors will too.

  5. Kiyo February 16, 2012 at 10:31 am #

    @Barry. Thanks for the clarification. To your credit, backtesting liquidity cuts should not be nearly as biased as backtesting the selection algorithm.

  6. Chris February 16, 2012 at 10:42 am #

    @Kiyo: “In a perfectly efficient market, the ONLY way to get better returns is to take on more risk.”

    Theory does not always match reality. Long bonds have outperformed stocks over the last 40 years at significantly lower beta.

  7. Andrew F February 16, 2012 at 1:01 pm #

    Barry, is it possible to get calendar year returns for the backtests?

  8. Barry Gordon February 16, 2012 at 1:16 pm #

    Yes. They are up on the Morningstar.com website. Go to http://corporate.morningstar.com/US/asp/subject.aspx?xmlfile=11.xml (or google morningstar indexes) and click through “Equity Indexes”, and then depending on which one you want, you have to search through their indexes – for example both Canada Target 30 and US Target 50 are under “Equity Income Indexes”. You choose the index you want and a fact sheet like this: http://corporate.morningstar.com/US/asp/popup.aspx?xmlfile=2649.xml&filter=CANDIV30.xml&page=Equity will show you the calendar year returns. Those sheets also highlight other metrics that may be relevant to you. Cheers

  9. Andrew F February 16, 2012 at 2:49 pm #

    Thanks, Barry. I almost have a hard time believing the alpha for the momentum and value indexes, even though I am convinced by the evidence of both the momentum and value effects. I’m more used to seeing it be a ppt or two.

  10. Charles Chase February 16, 2012 at 2:58 pm #

    Interesting. The criteria for this index used for XFM seems to be generally similar to the RAFI fundamental index used in CRQ. I would have expected the two to be similar, especially since there aren’t that many large Canadian companies, but the two ETFs seem to have very different holdings.

  11. Nick February 16, 2012 at 5:40 pm #

    How do we know CPMS didn’t try numerous strategies, some of which worked over the past 25 years, some of which didn’t, and Morningstar is just cherry picking the ones that worked and dropping the rest. This would be akin to any financial service company that plays up its actively managed mutual funds that had the best past returns up to the present. Of course someone will always “beat the index”, but it’s been shown long ago that, overall, investors’ average returns will equal the index’s minus costs. Either you put your money in an index fund, and get the average return, or you take your chances with an active strategy, human or computer controlled, and maybe you beat the index and maybe you don’t.

  12. Dan M February 16, 2012 at 6:03 pm #

    I just want to say that I really appreciate Mr. Gordon’s personal participation in these recent conversations. It’s refreshing to have a Senior Representative of a company get out there and enthusiastically discuss their products with the public. Furthermore, he heard what people wanted and gave it to them (posting the index construction rules).

    Too often these questions remain unanswered or you get a useles generic response from an entry-level employee or a computer. “Thanks for your comment. We value your busines. We will look into it…” blah blah blah.

    Thanks, Barry!

  13. Lake Superior John February 17, 2012 at 10:16 am #

    @Nick.

    Great post. From what I have read your statements are validated, time and time again!
    As for Mr. Gordon’s participation this is a breath of fresh air. Will others follow?
    Thanks Barry, Nick, and Dan.

  14. Canadian Couch Potato February 17, 2012 at 10:28 am #

    Thanks to everyone for your comments, and I echo Dan M’s sentiment. Mr. Gordon was very forthcoming with information about these new indexes, and I’m impressed with his willingness to engage directly with investors. Believe me, even as a journalist with the national media it is often difficult to get straight answers from people in the financial industry, so his openness is much appreciated.

    As a general comment about CPMS, I accept Mr. Gordon’s assurance that there is no data mining here. The value and momentum premiums are real. As always, however, the problem is whether the costs of implementing the strategies will overwhelm the benefits. The same is true with the RAFI fundamental indexes: yes, they may provide 2% a year in excess returns, but if they cost 50 bps more in management fees, incur 1% trading costs and lose another 50 bps to tracking error, so much for the advantage.

    Only time will tell, and I look forward to watching the performance of XTF’s products in the coming year.

  15. Barry Gordon February 17, 2012 at 10:29 am #

    @Nick. Again, while I share your skepticism of back testing and cherry picking, it really isn’t the case here. All of these indexes are based off real long term actual stock portfolios recommended by CPMS that they have been publishing (and charging people for) for years. Morningstar really has too much reputational risk at stake to simply data mine trying to find strategies that worked historically. As do we. We stand behind the methodologies. Thats all I can promise.

    I will endeavor to keep up with what everyone here is discussing. If you ever have a question of me, ask me by name in a post and my google alerts should pick it up. It just may take me a day or two to respond.

  16. Seldom February 17, 2012 at 2:10 pm #

    Ladies and gentlemen,
    That is a company president that’s worth listening to!

  17. Andrew February 17, 2012 at 2:33 pm #

    I disagree that the strategies will be arbitraged away, especially since they are now in an ETF product.

    Value and momentum can offer premiums because they are a disciplined approach to putting capital at risk, while many, if not most, of the rest of the market players operate using behavioural/emotional biases that are exploited by the strategy. The fact that they continue to work, that they may be exploited, attests to the prevalence of these biases and the difficulties to overcome them.

  18. Ogopogo February 18, 2012 at 6:00 pm #

    Hi, long-time reader, first-time poster. I’m thinking of placing a limit order of 50 shares of FXM and 50 of DXM this weekend to avoid a market spike on Tuesday. I’m a fairly new investor and definitely a couch potato, but I have $5,000 to play with in my TFSA, so I thought this would be a good start with an ETF that’s just getting off the ground.

    Any thoughts? Thanks.
    P.S. I’m also wondering why there were only 60 trades of WXM? Is this a bad sign for this ETF?

  19. Canadian Couch Potato February 18, 2012 at 6:09 pm #

    @Ogopogo: Thanks for posting. I can’t comment on whether either of these ETFs would be appropriate for your situation. But to answer your question about the low trading volume, this is not unusual for new and small ETFs. There are many Claymore and Horizons ETFs with very low trading volumes, too. Barry Gordon of XTF actually wrote a piece in the Financial Post that addresses this specifically:
    http://business.financialpost.com/2012/01/30/xtf-insight-understanding-the-basics-of-etf-liquidity/

    In theory, low trading volumes should not have a dramatic effect on liquidity, but I know from experience that it can indeed cause fairly wide bid-ask spreads. I would definitely recommend trading them with limit orders.

  20. Barry Gordon February 18, 2012 at 6:30 pm #

    @Ogopogo. Like Dan, I can’t comment on your specific investment needs, but I am flattered you would consider our ETFs. In terms of trading, I always recommend limit orders to everyone. You mention putting in a limit order this weekend. Be aware that you risk getting filled at prices that may be above market if you place your order like that prior to market open on Tues. personally I would wait to see how the market shapes up on Tues morning and then place your order, for whatever ETFs you end up choosing. I generally never trade during the first 15 minutes until the market settles in.

  21. Neil February 23, 2012 at 9:23 pm #

    I hope this isn’t a ridiculous question but I have been looking to invest in a dividend focused ETF. XTF Morningstar Canada Dividend Target 30 and Claymore Canadian Financial Monthly Income have made my shortlist. I love the way the Claymore website lays out the distribution schedule and history. I understand that XTF fund is new but I can’t find any real information on what to expect in the way of distributions besides being paid quarterly. The div yield (4.21%) even comes with a disclaimer (in the fact sheet) that this yield shouldn’t be expected. My question is, if this is fund focuses on dividends, where can I get information on the expected dividend?

  22. Canadian Couch Potato February 23, 2012 at 10:32 pm #

    @Neil: Not a ridiculous question at all. The Claymore ETF (FIE) has a fixed distribution, which means they pay you that 4 cents per share every month whether or not the dividend yield can sustain that. What this mean is that a lot of that is return of capital, not dividends. You can see the breakdown here:
    http://claymoreinvestments.ca/etf/fund/fie/distributions

    The Morningstar ETF will pay out dividends as they are received, without any return of capital. You should probably expect the dividend to be right around the 4.21% quoted. But remember, as the price the of the fund moves up or down, the percentage yield changes, so they can’t guarantee that. Hope this helps.

  23. Neil February 26, 2012 at 4:38 pm #

    Thanks. Really enjoy the blog. Learning lots!

  24. Kiyo March 18, 2012 at 9:00 pm #

    Sorry to revive an old thread, but I had a question related to limit orders and ETFs and this comment thread seems to be the most on-topic.

    Ideally, I would like to pay the NAV for the ETF. However, that’s not calculated during trading hours and, if the the volume is low, the bid/ask prices might not reflect the NAV well. So how do I choose the limit price?

  25. Canadian Couch Potato March 18, 2012 at 9:18 pm #

    @Kiyo: If an ETF is trading at a premium to the NAV, there really isn’t anything you can do about it. In theory, this should not happen, but in practice it definitely does. This is one place where mutual funds are superior to ETFs: the always trade at NAV with no spread.

    If the bid-ask spread is wide, the best you can do is put in a limit order a few cents above the ask price (assuming you’re buying) and hope you get a taker. Some brokerages offer Level II quotes, which allows to see the actual bids and asks of other traders to give you a better idea of the ETF’s trading activity. But in general, if your are making long-term buys and the spreads are not excessive (and with most ETFs they are not), then I wouldn’t worry too much about this. Spreads are really only a big issue if you’re doing a lot of trading.

  26. Barry Gordon March 23, 2012 at 5:30 pm #

    @Kiyo. The bid/ask spread on an ETF, regardless of what the historical volume has been, should always reflect the bid/ask spreads of the underlying securities. In other words, if the ETF holds, say, 30 stocks and they are all very liquid, the average bid/ask spread on each of those underlying stocks might be a penny, or two. So, the bid/ask spread made by the market maker should reflect that if they are doing their job. They typically price it a penny, or maybe two pennies, wider to make a profit for themselves. In sum, for an ETF that holds 30 stocks with an average bid/ask spread of one cent, the spread shown by the market maker should be two or three cents.

    Now, in relation to NAV, the spread shown will reflect very closely to actual NAV at any time. The Market Maker receives files daily (which are usually available on ETF co. websites by the way) which shows what is called an “iNAV” (no relation to Apple 🙂 ) – meaning indicative NAV. Essentially, it will reflect any expense accruals, expected dividends, interest, etc… So, when a Market Maker is showing the bid/ask spread, it incorporates all the things that actually go into calculating a NAV, and it should very closely reflect a penny or two most on either side of NAV.

    All of this adds up to why ETFs are good investment vehicles. Now, all that being said, you do have to be careful, because if you buy or sell in sufficient volume in a particular trade, you may “use up” the posted bid or offer, and then get your remaining fill farther away from actual NAV. Also, I recommend never trading in the first 10-15 minutes of market activity, or near the close, because wonky things happen as securities are opening up for the day and dealers adjust inventory at the end of the day/in preparation for the MOC (market on close).

    Soooo…, very long winded, you should be able to place a limit order either at the posted bid or offer, or a cent inside it, and be confident that you are getting filled at or very close to NAV. If you are going to buy in big volume (5000 shares or more) you can always call the ETF company, who should have a specialist that can help you hook up with the trading desk at the market maker to ensure you get the full fill at the level you are posting. Also, the ETF companies (like mine) try to keep a very close eye to make sure their ETFs are trading properly. Hope this helps

  27. Canadian Couch Potato March 23, 2012 at 5:52 pm #

    @Barry: Many thanks for the detailed answer. The inner workings of market making are a real mystery to most DIY investors, so this is some helpful insight.

    In the past I have traded some Claymore ETFs that had very wide spreads (both were “ETFs or ETFs”), and I placed market orders that were filled at surprising prices. Now I know better and always use limit orders.

  28. Barry Gordon March 23, 2012 at 6:59 pm #

    thx Dan. Under market making best practices the spread should reflect the true liquidity of the underlying securities. The less liquid the underlying, the wider the spread. However, as you say, strange stuff happens. Most market making is done through computer driven algorithms and sometimes the inputs go screwy intraday. Also, sometimes you can see market makers trading something rich for reasons that are unknown to me. It truly is the etf company’s job to monitor it, and most firms have at least one person who keeps a close eye on it. People who experience bad fills should call the company and complain.

  29. Andrew F March 23, 2012 at 7:02 pm #

    Barry, any thoughts on applying the same value and momentum criteria to the US market? Do those strategies generate the same kind of out-performance, or is there something particular to Canada (market inefficiencies) that allow these strategies to outperform?

  30. Barry Gordon March 23, 2012 at 7:44 pm #

    @Andrew thanks for the question. CPMS Morningstar has excellent track records for the US value and momentum strategies. It is safe to assume that we will look at all of these for future ETFs. Cheers

  31. John April 4, 2012 at 1:10 pm #

    Hi Dan,
    My question is how to correctly set limit orders. I have used them but, I’m mostly guessing on the limit order price based on current bid/ask spreads. As a long term investor of ETF’s inside my RSP is it a good idea to set my limit order below the bid price to expire over a few days to hopefully get a better price.

  32. Canadian Couch Potato April 4, 2012 at 1:23 pm #

    @John: If you’re buying, it often makes sense to make the limit order a few cents above the ask price. If the current price is lower than your bid, you will get the lower price, so there’s no risk you will overpay. But this protects you from surprises in case the price shoots up quickly: you are setting a maximum that you’re prepared to pay.

    You can always put in a buy order lower than the ask price (or a sell order higher than the bid price), but you run the risk of never having that order filled.

  33. gl October 29, 2013 at 9:29 pm #

    Just to provide an update on this ETF.

    We are now 1 yr 8 months into the launch and this is how return stack up:

    6 month:
    WXM +15.14%
    TSXComp +8.10%
    TSX60 +8.35%

    1 year:
    WXM +17.90%
    TSXComp +9.16%
    TSX60 +9.58%

  34. Andrew F October 29, 2013 at 10:55 pm #

    Actually Dan, are you considering covering the US versions of these strategies? They were just launched a few days ago.

  35. Canadian Couch Potato October 30, 2013 at 12:28 am #

    @Andrew: No plans to look at these new funds just yet. But stay tuned for a future post that includes an analysis of FXM.

Leave a Reply