- Comments (16)
- Text Size: Down Up
moneysense.ca, 3/01/07
What is my Portfolio Return?
It seems like such a simple question: How much did our combined portfolios return in 2006? I reported yesterday that according to Microsoft Money, our portfolios returned 9.5% during the year. Turns out, the answer is not so simple.
Let’s say our portfolio is valued at $1,000 on January 1, 2006 and we added $280 at various times during the year. At the end of the year, our portfolio had a value of $1,437. What is our return?
Microsoft Money reports that the return is 9.5% which must be an annualized return that could be skewed by investments that slid sharply just after buying towards the end of the year (Conversely, investments that rise sharply just after buying will skew the returns in the opposite direction).
If I use a far simpler calculation [(end value - additions)/start value - 1], our portfolio returns are a respectable 15.7%. If I use a slightly more complicated calculation [(end value - 1/2 additions)/(start value + 1/2 additions) - 1], our returns are slightly less at 14%. I’ll investigate exactly what returns are reported by MS Money reports in a future post and this little investigation confirms my gut feeling that our returns are better than reported.
moneysense.ca, 3/01/07
If you are calculating a return for investments made thoughout the year, your method would be incorrect. If additions are made equally through the year, the return on those would be approximately the equivalent of having bought them all in the middle of the year. If investments are made in unequal increments, the average amount would have to be a weighted average with the weight based on the number of days for each incremental investment
I actually just covered the 3 different ways that I use to calculate returns on my site.
Part 1 here: http://www.intthree.com/index.php?itemid=281
Part 2 here:
http://www.intthree.com/index.php?itemid=282
Part 3 here:
http://www.intthree.com/index.php?itemid=283
All 3 methods have their advantages and disadvantages but I think tracking 3 rates of return gives me a fairly good understanding of how our portfolio is doing relative to the market and specific mutual funds.
Quicken is even more bizarre in calculating returns on investments. I have absolutely no idea how the program creates the numbers.
To give you an idea of how bizarre it is my year to date annual return after one trading day on the TSX is 29.85%, despite this massize percentage increase the value of my portfolio rose by about $13 or closer to a real gain of about 0.02%. But thats not the most bizarre part. According to Quicken Loblaws has a one day gain of 3771.67% (yes thats right three thousand percent) despite an actual increase of only $1.10 per share yesterday.
There is an “official” formula for calculating returns when additions are made during the period (see here: http://www.freedownloadscenter.com/Business/Financial_Calculators/Portfolio_Performance_Calculator_Download.html. But it’s way too complicated for me to bother with (you have to calculate the return from the start to the day the first addition was made, then the return in the second period, etc. and do some basic maths using a totally non-intuitive formula, or use something like the spreadsheet link above).
For this reason I’ve never got beyond making a rough estimate using the basic (and incorrect)
“(end value-additions)/start value”
formula myself. But this hasn’t stopped me meeting my goals so far.
Even if you did go to all the effort of working out your performance “correctly”, you would have to decide if you want to calculate overall portfolio performance (and compare it to what??), or calculate performance figures separately for each asset class in your portfolio so that you have can identify suitable benchmarks. (You may also have to work out risk-adjusted returns if you want to compare, for example, a personal portfolio of 20 stocks with an index benchmark).
I think if you made significant additions to your investments during the year you would also have to adjust the benchmark performance so that it was comparable. eg. if you started with 50K in a S&P500 index fund and the added 50K more on the 1st March, you couldn’t just compare you’re performance to the straight gain in the S&P500 for the entire year – you’d have to get index values for the dates you made the investments and work out a suitable adjusted “index” figure to compare to.
At the end of the day you only need to be able to identify if your portfolio performance was sufficiently different from what you expected (ie. the benchmark you had in mind) that it indicates when you have done something wrong.
I don’t recommend using Oxcc’s method #1 (same as Canadian Capitalist’s formula). Ignoring what you have contributed throughout the year is leaving out important information. The only method of calculating return that really shows the full picture is the IRR (specifically XIRR in a spreadsheet program, which is IRR for non-periodic payments). This is what programs like MS Money and Quicken will give you under “annualized return.” This is most likely what the commenter James is seeing and is Oxcc’s method #2 above.
Gummy Stuff has lots of links on how these calculations work:
XIRR (good note here about the fact that you don’t include dividends as positive cashflows.
Dietz return – the approximate-quick-and-dirty formula at the bottom is the same as the Canadian Capitalist’s second formula (just rearranged with some algebra. After a quick scan of this stuff I think Dietz is exactly the same as XIRR.
Lots more at Gummy Stuff, just search the page for “return”. Note: I am not affiliated with Gummy Stuff in any way.
I think there’s too much analysis put into it. I use a spreadsheet and I only measure what really matters. 1) What is my original cost? 2) What is it worth today if I sold it? 3) What amount of distributions did I get from holding it?
As I keep saying on this site, average annual returns don’t make any sense because it’s usually all over the map. Total returns in my portfolio in 2006 range from +50% to -5% and different investments are made at different times in the year – it’s not averaged out for the year. Average returns are rather irrelevant.
IRR (or XIRR)is it when evaluating how the INVESTOR did. It correctly calculates the weighted average return of every dollar for the exact amount of time it has been in the plan. It is also called a “dollar-weighted” rate of return.
IRR reports what the INVESTOR got and it is obviously a very important number, but it has drawbacks for evaluating whether an INVESTMENT is good.
The IRR may be low if the investor bought at a high or sold at a low. The timing of these cash flows affects the return a great deal and the INVESTMENT obviously can’t control this.
If you want to know an INVESTMENT did, you need to eliminate the impact of different cash flows. The INVESTMENT must be compared to other investments using the same end date. This is called a “time-weighted” rate of return. Examples of this are the numbers found in mutual fund tables which show 1,3,5, and 10 year RORs.
Many times the INVESTMENT yields great returns, although the INVESTORS don’t do as well. Buying high and selling low will always be the cause of this.
Smart investors will always use the specific method that takes the ‘data’ and provides the best ‘information’.
Phil,
You sound just like my dad when you say “different investments are made at different times in the year” and “total returns in my portfolio in 2006 range from +50% to -5%”. My dad invests a whack of cash at the same time every year, in June, so then he can easily see in the following June what his return for that previous 12 months were. (He sometimes buys more investments throughout the year but doesn’t bother calculating performance of those). When I tell him that you can invest at any time in the year and use XIRR to figure out your investments’ annualized returns, he still doesn’t get it, and says “but different investments are made at different times in the year?” Read about about XIRR, learn how to use, and don’t tell people that “average annual returns don’t make any sense because it’s usually all over the map” and “average returns are rather irrelevant,” unless you really know what you are talking about.
Careful with your terminology, average annual return is something different. Most of us here are talking about “annualized rate of return” over a given period.
You say that you care about “what is it worth today if I sold it” and its “original cost”. Well what if you bought $1000 of an investment 1 year ago. It gained 50% and is now worth $1500. Just last month you bought another $10000 of that same investment because you were amazed at how well it did. That $10000 has stayed still and is still worth $10000. So your total cost is $11000 and your investment is worth $11500. I think what you would have done is looked at the final divided by initial and said, hey, my investment has gone up 4.5%! (Or would you have taken into account the fact that these two purchases were made at different times? Or do you just give up and not worry about the actual percentage return?) But the IRR is actually 28%. Incidentally, the underlying investment went up 50% during that year (from $1500 to $1000). So we ruined our nice 50% return a bit by that $10,000 buy late in the year, but whatever. 4.5% is a return but it is not very interesting or meaningful. Finding out that our annualized return was 28% is meaningful, however.
Rob makes an important point that an investment should be compared to other investments using the same time and date. Taking your portfolio, finding the 10-year annualized return using XIRR and then comparing that to the 10-year return of benchmark indexes and such is a good idea. Taking a total return like 4.5% in the above example and comparing that to benchmark indexes makes no sense.
That’s it for my rant for now.
It is true that IRR is the true measure of performance but what I did in my earlier post was compare the IRR of my actual portfolios with the annual return of the benchmark (to which no money was added). That is really comparing apples to oranges. Like Enough Wealth notes, it is far too troublesome keeping track of the benchmark for every little bit added to the portfolio.
Phil: I do think that you should worry about and track relative returns, so that you know exactly how well (or badly) you are doing compared to a passive portfolio that is a snap to construct. If you find that you are beating the pants off your benchmark, you should continue to do your magic (and hope it persists in the future). I suspect too many investors would find that their stock-picking abilities (after all studies show that even the majority of professionals have difficulty beating the indices) really suck and they are better served with just a passive portfolio.
[...] There is a very interesting discussion at the Canadian Capitalist’s blog about how to determine performance and there was some talk of Internal Rates of Return (IRR): [...]
Sorry for not posting my comment here…I already had one long comment…didn’t want to hijack the entire thread and make it my personal soapbox.
if the formula used by mutual funds, when they advertise their performance statistics to attract more cusotmers, was used by private investors, we could compare our portfolios with the professionals performance.
can any ‘responder’ on this site provide the formula ?
CNN had a formula (now apparently discontinued) in which initial values, dates of sales, purchases, dividends etc. were recorded for the period, e.g. 1 year, and the annual percentage return was calculated.
Neville,
It’s XIRR, look it up on Google. You can do it on any spreadsheet program, just be careful not to include dividends as a cash in-flow. The cash in-flows and out-flows are only those that came from you to the investment or from the investment to you.
Okay…so using Excel…Do I want to put in the opening balance of my investment? And then all of my deposits, right? Then, my closing balance? It has to have negatives, what do I do about that? Thanks.
does anyone know of a decent/reputable online dietz / time-weighted return calculator?
Agree that if your portfolio had distribution and withdrawal of invested capital during the period, The right way to calculate, what Morningstar call “personal return”, is XIRR. This function allows calculates internal rate of return with non-interval cashflows. As a result, “Personal Return” can shows how well you manage your investment.
However to see how your portfolio allocation performs, comparing to benchmark return, “personal return” can not be used. Reason is that benchmark calculates with assumption that all capital are reinvested, no inflow no outflow. Therefore, to see how good is your allocation without interaction of market-timing skill, you need to use another one, called “total return”.
Hope this helps,
How to calculate => sorry I personally dont know, try to find out now