I may have asked this question some years earlier, but I hope that perhaps something has changed and I can find an answer to my question.
My question is that I have two Ameritrade accounts, regular and IRA, into which I made additions from time to time, never on a regular schedule. Besides that, I made various trades at various times, as well, received dividends and whatnot.
What I would like to know is a fair calculation of my rate of return on this account, year by year, properly taking into account things such as added money etc.
Excel's XIRR() function will. You only need to trap deposits and withdrawals from the account, as well as the final value, if you want IRR on an account level. The dividends and interest are happening internally and contribute to the end value and IRR.
If you want IRR for specific investments you'll need to track purchase/sale dates, dividends paid, and current values. But it's a lot easier to research the historical performance of the given security instead.
Tad, thanks a lot. Google Spreadsheets has two useful things: the XIRR function that you recommend, and GoogleFinance function that lets you access current and historical stock prices (and a lot more!).
I set up a spreadsheet, to count cash value only, as follows: On the left column there is a transaction date. On the right column there are cash flows; such that the starting amount is negative, all cash additions are negative, all disbursements are positive, and the final yesterday portfolio value is positive.
I started in May 2003, as, sadly, Ameritrade statements are not accessible beyond 6 years.
I fed this data into XIRR function and got a total XIRR return from May 2003 until yesterday to be 4.5% per annum.
For the same period, thanks to GoogleFinance function, the return of S&P 500 index, WITHOUT dividends, was -0.59% per annum. I do not know how to do a S&P total return (with dividends) in GoogleFinance. But the data for 2003-2009 suggests 1.75% average yield for the period. This makes S&P total return to be approximately 1.16% for the period.
Note that this fully ignores tax effects, as I did not pay taxes from my brokerage account. Going into tax effect would be very painful, but my first guess based on some money losing trades is that I did not pay a lot of capital gains taxes because I sold a couple of losers and did not sell a lot of winners. I did not get that much in the way of dividends. This is very unscientific, of course, but my hunch is that taxes did not amount to much.
All trading commissions are fully included in the calculation, since I went by account balance.
The good news is that my account, if I believe my own math, outperformed S&P. S&P index changed from 963 in May 2003 to 888 in May
2009, by 3.36%.
The bad news is that this outperformance amounted to not a lot of money, in terms of my monthly income, and took an inordinate amount of personal time and work (inordinate to my standards). I would be much richer, for the period, if I put all my trading money into spiders and spent tha time programming computers for profit. I also spent big money on finance related books, which again is not counted above.
The results are very revealing and very sobering.
I did the same for my Ameritrade IRA, which, fortunately, is insulated from tax effects. There, for the same period, the rate of return was
6.40%, which outperformed S&P by 5.2%. It sounds like an impressive number, but amounted only to a very small sum, because that IRA has not much money.
I think that I enjoy investing activity, thinking about this stuff etc, but there are other things that I enjoy not less. A lot of food for thought.
Thanks for recommending the spreadsheet approach Tad.
[I hope the moderators tolerate this thread, to me nothing is more fundamental to an individual investor than figuring out how they're actually doing with their investment choices!]
Glad that was helpful...in case you're interested in the 2.0 approach, here are some more ideas. There are a weaknesses in the IRR comparison you did, in addition to the ones you mentioned (taxes, dividends). The first may have an easy solution, the second is a lot more difficult.
If you had no deposits and withdrawals since your starting date, comparing your IRR to that of a benchmark index over the same period gives a meaningful result, and estimating the effects of dividends for the index isn't hard to do. But that's an unusual case, and IRR blows up when you factor in timing of your deposits/wdls. As an extreme, imagine the scenario was:
Deposited $1,000 March 3 2000, bought three stocks; this is "day 1" for the S&P500 benchmark Deposited $100,000 March 3 2009, bought a stock index fund Calculated returns May 2009
Obviously a market-beating IRR doesn't tell you much, because the timing and amounts of your investments would produce a much higher IRR for your portfolio than the index over that period (the start date was near the peak of the 2000 stock market bubble, but little was invested; the second date had a rapid ~40% runup right after it, with a lot of money going along for the ride). This is an extreme but at some level the issue crops up with any deposits/wdls.
How to address this? It isn't easy because a price index doesn't show the effect of dividends. One method might be to download data for one of the total return indices (e.g. the Russell 3000 series with all dividends reinvested) and turn each deposit or withdrawal into a purchase or sale of an "index unit" based on the price of the R3K-w/divs that day. Instead of using IRR, tally up your index units at the end, compare their value to what you ended up with. That will at least solve the timing aspect of the comparison. I've started this basic thing in an MS Access app but never got it quite there.
The second issue is risk-adjusting that IRR, which is a pretty open-ended thing. I think most people are good with "did I beat the market?" as a starting point. But if you owned 50% bonds, comparing to the S&P 500 isn't a fair comparison. This obviously can get a lot more complicated if you try to factor in that aspect because you might need to know your asset allocations throughout to do a fair comparison.
I am sure that the brokerage industry would hate what I would like to see proposed, which is to require them to calculate and compare investors' results with several benchmarks.
.... snipped good points ...
This depends on how IRR is calculated, which is not something that I know. I would imagine that a "value weighted IRR" would be indeed more useful, as it would compare apples to apples.
I did try, with some approximations, to take dividends into account when comparing with SPY.
That should work very well and is a good idea.
That brokerage account, was 100% stocks only at the time. My approach was to keep my stock trading at Ameritrade, bonds and money market at Vanguard, euros at a bank etc. However, in the future, due to rise in various asset allocation ETFs, I will use Ameritrade more for non-stock investments. It will, without a doubt, make it more complicated to compare returns, but it will make other things easier such as changing asset allocation.
Another aspect of risk, besides allocation to asset classes, was diversification. In this respect, my own investments were very poorly diversified in some sense, but some investees themselves were rather well diversified.
I am sorry for sounding gloomy in the previous post. The return on time invested was not much. However,
1) I am much more optimistic regarding future returns in general, compared to our past ten years results, given much more investor friendly price levels. Our family still has 100% of our retirement money in stocks, after a big but poorly timed move from money markets in November.
2) Since I have a more assets than I had 6 years ago, producing a similar outperformance for the next 6 years will yield a much better return on time invested. I did not realize it when I made my previous post. Of course, this outperformance is very far from assured.
i
The standard IRR calculation is defined thusly (my paraphrase):
"The internal rate of return is that discount rate in a net present value calculation which sets the net present value of a series of cash flows to zero."
In general, the calculation can only be done iteratively. There's no arbitrary closed-form solution.
By its nature it will be value-weighted (otherwise known as dollar-weighted).
The cash flows input to the calculation have deposits to the portfolio as positive cash flows, withdrawals as negative cash flows, reinvested interest/dividends not even appearing (since they remained inside the portfolio) and the portfolio balance on the calculation day as a "withdrawal" of the balance amount on the calculation day (e.g. a negative cash fflow).
It can be jiggered (by jiggering the cash flows used as the input to the calculation) to be so-called time-weighted.
Time-weighted returns (one might think of them as value-neutral) arguably give a better idea of how a fund/portfolio manager is doing. But value-weighted (also known as dollar-weighted) returns give the tell you the what an investor in the fund/portfolio actually experiences.
Rich, thanks for a great explanation. Now I understand it a little better and I think that IRR is a perfectly good measure for what I need to know.
Since I answered my question "what was my return compared to S&P", the next question is "how likely is that the outperformance in it is due to luck". This is a much harder question to answer, as I have very little statistical data to work with. This is a question that I probably will leave unanswered for now.
Just to clarify it, up until a few days ago, I had no idea whether I was outperforming or underperforming the index. I did not even try to outperform it, I simply did things that made sense and did not do things that did not make sense, and tried not to trade too much.
Furthermore, I will use my Ameritrade account to hold things other than American stocks, and will buy ETFs when appropriate (euros, bond ETF etc) , so the comparison against benchmark indexes will not be as directly applicable as it was to the prior 6 years.
I am not, by any means, setting a future goal to myself to outperform any stock index. I will try to not look at the performance comparison too often (it is set up to autoupdate with current stock prices), as I think that it is counterproductive.
But if that happens, so much for the better.
My future goal is to buy things that are priced attractively, and avoid fads and follies.
BeanSmart website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.