Tuesday 24 May 2011

Cross-Border ETFs - Here's a Free Tool to Compare Costs

Our last post ended by suggesting that investors should compare the total costs of investing in US-listed ETFs vs Canadian-listed versions. We've taken up our own challenge and offer a spreadsheet (click here for download) to help readers assess their candidate foreign ETFs. Here is a screenshot of the spreadsheet.


What the spreadsheet does
Our approach is to start with assumptions about growth of an index that two ETFs aim to track and adjust returns taking into account the various factors explained below. The spreadsheet takes into account these factors by allowing you to enter the data for pairs of ETFs and applying costs, taxes, rates and fees that influence net returns.
  • US holdings and International holdings - Each type has its own sheet since a) each would be a separate asset class within a portfolio - thus we want to compare within an asset class - and b) the tax consequences are different.
  • RRSP (& other registered retirement accounts like RRIF, LIRA, LIF), TFSA, Taxable account choice - Tax consequences differ for each type of account.
  • Direct or Indirect ETF holdings - Instead of holding foreign stocks directly, some Canadian ETFs hold US ETFs inside, which we call Indirect ETFs. Again, there are different tax effects.
  • Capital gains and Distribution/Dividend yield - The amount and proportion of each source of return affects the net return due to the interplay of the other factors. A usual starting point for picking numbers to enter is historical data. For example, the current distribution yield for the SPDR S&P500 ETF (symbol: SPY) is 1.74%. Estimating Capital gains can be a lot more problematic - are we looking forward to years of low 2-3% growth or higher 5-6% gains? The usefulness of the spreadsheet is that we can quickly plug in different numbers to see if the best choice of ETF and account varies and by how much.
  • Tracking error - This is the amount by which the ETF performance differs from its index. Although a very few ETFs may actually outperform the index in the occasional year, the typical and expected result is under-performance over the long term, which is what interests us. A key element of Tracking error is the ETF's Management Expense Ratio (MER) and that figure should be the minimum value to enter. However, it is not the only one and especially with International ETFs or currency-hedged ETFs, not necessarily the biggest component. Other Tracking error sources include: Sampling/optimization or Replication strategy, Rebalancing and Reconstitution efficiency, Cash dividend reinvestment drag, Trading costs, HST (Canada only), Currency hedging costs, Concentration limits on holdings and Securities lending (a source of revenue for many ETFs). Larry MacDonald's Investopedia article ETF Tracking Errors: Is Your Fund Falling Short? explains this topic well. As Tracking error rises to the 1% or more range it starts to heavily influence net results, as playing with our spreadsheet will quickly show. Canadian Couch Potato's tally of Tracking Errors on US and International ETFs listed in Canada shows widely varying errors amongst funds for 2010. Keep in mind that errors can vary, sometimes a lot, year to year. Some ETFs seem to be quite stable, others not. Tracking error very often turns out to be the key factor in differentiating ETF performance.
  • International withholding tax - The actual amount paid to foreign non-USA governments varies within various ETFs - often 8 or 9%, but up to the 15% maximum of international tax treaties. In only one combination of circumstances, a Canadian listed ETF that directly holds international securities in a taxable account e.g. Claymore's International Fundamental Index (CIE), can a Canadian investor recover that tax through a tax credit.
  • US withholding tax - Though not a variable you need to estimate or change in the spreadsheet as it is always applied at the 15% rate specified in the Canada-USA tax treaty, US withholding tax can have a big influence on performance at times, especially when the ETF distribution yield gets to 2% or more and account type is brought into play.
  • DRIP brokerage commission and auto-DRIP - Many brokers now charge $10 per trade so we have set that as the amount in the input cell. If you want to see how reinvesting more often e.g. twice a year or quarterly when the ETFs make distributions, simply enter the total brokerage trading commissions you will incur per year. The same input adjustment can be made if you unfortunately are paying higher brokerage commission rates. Two ETF providers in Canada - BMO and Claymore - will automatically reinvest distributions for free, so our spreadsheet will adjust for that savings too. Some discount brokers even offer a so-called synthetic DRIP for ETFs (call your broker to find out if it offers that service; see also this article from CanadianFinancialDIY), absorbing the cost of distribution reinvestment.
  • Transaction fees on foreign exchange - The amount charged by discount brokers to convert Canadian into US dollars or vice versa varies from broker to broker, ranging from 0.5% up to 1.5% - see MillionDollarJourney's broker comparison. Forex can have a significant influence, mainly when Tracking errors of ETFs are close and where Withholding taxes do not differ.
  • Annual contribution - See how rich you might get! Just for the fun and as a reference point, since it does not affect which ETF is a better choice, the annual contribution you invest in the ETFs can be varied.
What the spreadsheet does NOT do
  • Compare different indices - It is assumed the ETFs compared track the same index. Different ETFs, even within the same asset class, such as US equity, can produce markedly different returns and portfolio growth. Look at the variability of fund returns in 2010 in Couch Potato's Tracking error article we cited above. If you compare ETFs with different indices, remember that you are only comparing costs, not total returns.
  • Exchange rate variability - There is no attempt to incorporate effects of changes over time in the value of the Canadian dollar against the US dollar or international currencies. Year to year and over multiple years, the evolution of currencies strongly affects foreign investment net returns. Whether to hedge or not is a fundamental choice for the investor and comes before the choice of a specific ETF.
Examples (numbers in the download spreadsheet)

US Holdings
  • US-listed = PowerShares FTSE RAFI US 1000 Portfolio (PRF)
  • Canadian-listed = Claymore US Fundamental Index (non-hedged) (CLU.C)
  • Forex fee = 1.0%, rate applied at big bank brokers
  • Distribution yield = 1.5%, current rate per PowerShares; could also try 1.99% cited on Claymore (rate as of end of March)
  • Account type = 1 RRSP
  • ETF type = 4 Direct i.e. CLU.C directly holds shares of companies, not a US ETF
  • Auto DRIP = 1 since CLU.C is a Claymore fund
  • Tracking error US ETF (PRF) = -0.48%, estimated using historical results on page 11 of the 2010 Annual Report here on the PowerShares website
  • Tracking error Canadian ETF (CLU.C) = -1.5%, estimated future error. The 2010 error was a horrible -2.3% from page 12 of the Annual Report here on Claymore's website, but in future it should be much lower as the ETF switched to full replication in September 2010, holding all 1000 stocks in the index, instead of the partial replication by 400 or so stocks it had previously carried.
  • Bottom Line = Under the above assumptions, PRF gets ever further ahead till it is up more than 20% over its competitor after 30 years due to the harmful effects of CLU.C's high tracking error.
International Holdings
  • US-listed = PowerShares FTSE RAFI Developed Markets ex-US 1000 Portfolio (PXF)
  • Canadian-listed = Claymore International Fundamental Index (CIE)
  • Forex fee = 1.0%, rate applied at big bank brokers
  • Distribution yield = 3.2%, current index rate per Claymore; could also try 2.3% cited on PowerShares though that is net of international non-USA withholding taxes already deducted
  • Account type = 2 Taxable
  • ETF type = 4 Direct i.e. CIE directly holds shares of companies, not a US ETF
  • Auto DRIP = 1 since CIE is a Claymore fund
  • Tracking error US ETF (PXF) = -1.2%, estimated using historical results on page 11 of the 2010 Annual Report here on the PowerShares website
  • Tracking error Canadian ETF (CIE) = -1.5%, estimated future error. The 2010 error was a terrible -2.1% from page 9 of the Annual Report here on Claymore's website, but in future it should be much lower as the ETF switched to full replication in September 2010, holding all 1000 stocks in the index, instead of the partial replication by 300 or so stocks it had previously carried.
  • International withholding tax = 9%, estimate from Foreign Tax Paid actuals on page 123 of the Annual Report
  • Bottom Line = Under the above assumptions, CIE comes out a constant 3% or so better than PXF. Put the CIE into a TFSA, however, and its advantage grows over the years, reaching over 10% after 25 years. Beware of withholding taxes!
There you have it. Use the spreadsheet to see which costs matter and which don't under reasonable assumptions, or to check how far things would have to change to make a difference in choosing ETFs and which account to put them in.

Disclaimer: This post is my opinion only and should not be construed as investment advice. Calculations and formulas are believed to be accurate but are not guaranteed to be so. Use at your own risk. Readers should be aware that the above comparisons are not an investment recommendation. They rest on other sources, whose accuracy is not guaranteed and the article may not interpret such results correctly. Do your homework before making any decisions and consider consulting a professional advisor.

2 comments:

bubulac said...

When I tried to access the spreadsheet I got a message Error 404 Not Found.
Any suggestion?

CanadianInvestor said...

Google had made some changes that apparently affected the visibility of the spreadsheet. Hopefully it should work ok now.