How to track your portfolio gains&losses correctly?

[smjsl]

People often talk about how many % their portfolio is making vs some benchmark, like S&P500. I wonder how you track it and more specifically, how do you account for
- additions
- withdrawals
- taxes
- dividends / distributions

For example, say you started Jan 1'2008 with $10,000 in your portfolio and you ended it with $12,000. During this time, you had:
- added $100 per month
- withdrew $1000 for some purchases
- paid $500 in portfolio-related taxes out of you liquid savings
- got $15 dividend 4 times a year
How much did your investment skills really give you for proper comparison with a benchmark?

I suppose one way would be to do as follows. Create a fake portfolio with your benchmark, and keep track of both portfolios:
- for every check you get from your job (if tracking your NW) or more generally, for every addition to your portfolio you are tracking, imagine you invested it in your benchmark (e.g. S&P500) with the closing price on the same date
- for every withdrawal, including taxes for the portfolio you are tracking, imagine you sold the benchmark at the closing price on the date before
- make sure your benchmark's dividends are reinvested and are paid on that number of shares you would have held at that ex-dividend date
- all dividends / distributions from your real portfolio have to stay "inside" it or you have to adjust for them as "withdrawals" otherwise.

I started doing something similar with my investment portfolio but quickly this became unmanageable for me even though I am not trading that often. In part, this was hard because figuring out taxes is somewhat of a pain. Also a simple S&P500 benchmark is not a good one since I have a high percentage of non large-cap non-US investments and tracking multiple indecies is harder with above approach. I suppose it's all doable in the end but quite time consuming.

What do you do?

 

[CyclingInvestor]

I keep all transactions, dividends, tax info etc. (since 1993) in an Excel file. Pretty easy to calculate IRRs, match shares sold to those bought, whatever is needed. Every once in a while I need to add a new column or page, like for calculating my 72t withdrawals a few years ago.

I use the SP500 as my benchmark, since I almost exclusively invest in large US companies.

 

[FIREd]

I use Quicken and double check with Excel. Both use the XIRR method to calculate returns.

 

[Alan]

I keep all transactions, dividends, tax info etc. (since 1993) in an Excel file. Pretty easy to calculate IRRs, match shares sold to those bought, whatever is needed. Every once in a while I need to add a new column or page, like for calculating my 72t withdrawals a few years ago.

I use the SP500 as my benchmark, since I almost exclusively invest in large US companies.

This is what I do to track ins/outs and to calculate IRR.

I don't compare against a benchmark as such. I compare against what the VG portfolio analysis says I can historically expect for my particular mix.

 

[smjsl]

Thank you for the replies and pointing me to the XIRR function. 2 follow up questions:

(1) To make sure I understand how to use XIRR, given my original example, is the following the correct set of values for XIRR inputs (I understand I need dates for this sequence too and the final numbers would be sorted chronologically):
-$10,000 (initial investment)
-$100 (additional money: listed 12 times for 12 months)
+$1000 (withdrawal should be positive money)
+$500 (taxes paid - should be treated as withdrawal at the date of tax payment)
+$15 (listed 4 times for the 4 quarterly non-reinvested dividends, but if they were reinvested they would not appear on the list)
+$12000 (final or current amount, which MUST NOT include non-reinvested dividends and other withdrawals listed above)

(2) I am still unclear on how you then compare with a benchmark (for those that do)? Do you manually compute the fake portfolio based on dollar numbers and dates of your real portfolio? For example, would not you need to calculate how many shares of your benchmark you are buying and selling each time and how many dividends you would get as a result, even if they are reinvested...?

 

[Alan]

Thank you for the replies and pointing me to the XIRR function. 2 follow up questions:

(1) To make sure I understand how to use XIRR, given my original example, is the following the correct set of values for XIRR inputs (I understand I need dates for this sequence too and the final numbers would be sorted chronologically):
-$10,000 (initial investment)
-$100 (additional money: listed 12 times for 12 months)
+$1000 (withdrawal should be positive money)
+$500 (taxes paid - should be treated as withdrawal at the date of tax payment)
+$15 (listed 4 times for the 4 quarterly non-reinvested dividends, but if they were reinvested they would not appear on the list)
+$12000 (final or current amount, which MUST NOT include non-reinvested dividends and other withdrawals listed above)

(2) I am still unclear on how you then compare with a benchmark (for those that do)? Do you manually compute the fake portfolio based on dollar numbers and dates of your real portfolio? For example, would not you need to calculate how many shares of your benchmark you are buying and selling each time and how many dividends you would get as a result, even if they are reinvested...?

For (1) - yes, that is the way to use the XIRR function.

for (2) - I personally don't feel I need that level of accuracy - way too much effort for me.

 

[FIREd]

I mostly invest in index funds, so I know that each one of my funds tracks its respective index fairly well. As far as my overall portfolio goes, the only benchmark I care about is my long term target return of 8% per year.

 

[Meadbh]

My statements include the appropriate benchmarks, which makes that relatively easy.

 

[CyclingInvestor]

(2) I am still unclear on how you then compare with a benchmark (for those that do)? Do you manually compute the fake portfolio based on dollar numbers and dates of your real portfolio? For example, would not you need to calculate how many shares of your benchmark you are buying and selling each time and how many dividends you would get as a result, even if they are reinvested...?

Much simpler. I actually use SPY as my benchmark. Each quarter I use Yahoo's historical data to provide me with the change in SPY's price, plus the dividend. For example, SPY closed at 91.95 on June 30 2009, 105.59 on Sept 30 2009, and paid a dividend of $0.508 on Sept 18 2009 when it closed at 106.72, so I consider my benchmark's return to be (105.59-91.95)/91.95 + 0.508/106.72 = 15.3%. Since I already have my own IRR, it is easy to compare.

 

[audreyh1]

I actually own shares in my benchmark. I use a balanced fund as a benchmark. When I was building my portfolio, I was adding to the benchmark at the same times. That makes it easy to track.

You could do the same with an index ETF.

I use Quicken to do most of my tracking and performance reporting.

Audrey

 

[smjsl]

I am sorry I am not being clear here (or maybe it's late ;-) )... I am trying to understand how to account for inflows/outflow into/out of the portfolio.

Meadbh, CyclingInvestor, please correct me if I am wrong but I don't think your methods are correct if you consider additions / withdrawal / etc.

Meadbh, I assume you mean that your brokerage sends you your benchmark changes in your statements. But if you deposited $1000 in the middle of the month, I asume they will not adjust their benchmark computation "portfolio" for that (see below).

CyclingInvestor, same thing. In your example, if you deposited $1000 on Sept 15, you'd have to make the adjustment for buying SPY at the same time when you compare to how your porfolio changed during same June 30 - today period. Also, you'd get more dividends then in your S&P500 portfolio than if you deposited $1000 on Sept 20.

Same applies to withdrawing a $1000 from your portfolio (including taxes).

So again, if you are trying to see if you are doing any better than SPY and by how much, I think you have to pretend you are really doing everything with SPY in terms of any incoming/outgoing funds into/from your portfolio. Otherwise, it's not apples-to-apples... Maybe there is a simple way, but I don't see it quite yet.

Audreyh1, your message actually was in part what inspired this thread ;-) You mentioned elsewhere that you are doing better than the benchmarks, which means you have funds outside of your benchmarks in your portfoio... but again owning a benchmark does not seem to account for above unless you distribute and withdraw all funds in exact same proportions. So, if your portfolio has 60% of funds in your benchmark and 40% outside of benchmark, any new $1000 has to go 60% into benchmark and 40% to other funds, and vice versa, any $1000 withdrawal has to come out as $600 out of your benchmark and $400 out of other part of the portfolio. Otherwise, I don't think you can base your comparison based on benchmarks changed vs overall portfolio.

Example:

T1 prices: myFund $10, myBenchmark $10.
Say you start with $1000 as 100 shares in myFund. (Or you could have had 100 shares of myBenchmark for comparison)

T2 prices: myFund $50, myBenchmark $5.
Say you added $500 to myFund with 10 shares $50 each; so now you have 110 shares of myFund. (Or you could have added 100 shares of myBenchmark and now had 200 shares of it.)

T3 prices (today): myFund $10, myBenchmark $10.
Now, your total is: 110*$10=$1100 (10% over initial amount and $400 less than overall $1500 invested). In any case however, what do you compare this myFund performance with? If you compare this with benchmark increase without accounting for T2 transaction, you would see 0% increase in myBenchmark, but if you actually invested in benchmark, you would have had $2000 now, or 33% increase over $1500 invested!

Final thought: perhaps for those of you already retired, your portfolio size is so large and your withdrawals (including taxes) and additions to it are so small that they do not make much difference... but for me, where both withdrawals and additions are large relative to the size of the portfolio, it makes a big difference...

 

[Brand New Day]

The question is: Does it really matter? Of course we all want to see it go up. But comparing it to a benchmark, WHY? It only matters that it is making more than I am spending, or/and that I am happy with the results. If neither of those are happening, then I need to worry about what it is doing and decide what to do about it.

I do track it, but rather simply in XL spreadsheet, start with jan 1 balance add and substract additions and withdrawls, then just compute % on current balance. Sure it may be off, but it is the BALANCE that I really keep track of, not the %.

 

[audreyh1]

Audreyh1, your message actually was in part what inspired this thread ;-) You mentioned elsewhere that you are doing better than the benchmarks, which means you have funds outside of your benchmarks in your portfoio... but again owning a benchmark does not seem to account for above unless you distribute and withdraw all funds in exact same proportions. So, if your portfolio has 60% of funds in your benchmark and 40% outside of benchmark, any new $1000 has to go 60% into benchmark and 40% to other funds, and vice versa, any $1000 withdrawal has to come out as $600 out of your benchmark and $400 out of other part of the portfolio. Otherwise, I don't think you can base your comparison based on benchmarks changed vs overall portfolio.

I was simply doing some spreadsheet "tests" to see if I could guess my portfolio performance from the S&P500. It turns out I really can't. Not surprising - my equity funds altogether are much more widely diversified than the S&P500. I was really amazed at the initial outperformance.

So - I don't use the S&P500 as my benchmark. I was just posting some info I had recently calculated out of curiosity on the S&P500 thread.

I had the spreadsheets for my calculations on the two different dates that I rebalanced my portfolio plus the current values. I simply took the total equity dollar amount on each of those dates, calculated the growth %, and compared against the S&P500 closing value on those dates. So it didn't take into account and S&P500 dividends paid out during those times, but since the yield is fairly low and time periods were not that long it probably doesn't make a huge difference.

Other than the rebalancing on those dates, I had not made any additions or withdrawals, so that made my calculations easy.

Audrey

 

[audreyh1]

I generally use the Quicken ROI calculation to calculate performance numbers for my portfolio. This does a reasonable job of handling withdrawals and additions IMO.

Audrey

 

[walkinwood]

smjsl,

Why not compare your individual fund performance to their benchmarks. I think that achieves the same thing that you're trying to do with a lot less complexity.

The only way I can think off to accomplish what you're doing is to create a second portfolio as you have already tried. To me, it doesn't seem worth the effort.

 

[Meadbh]

Meadbh, I assume you mean that your brokerage sends you your benchmark changes in your statements. But if you deposited $1000 in the middle of the month, I asume they will not adjust their benchmark computation "portfolio" for that (see below).

I get quarterly written statements. Online access does not show benchmarks. As far as I know, returns are calculated based on weighted asset allocation over time. As for checking on this every month, that's too much noise.

 

[smjsl]

Thanks for responses.

@Brand New Date, @walkinwood: the point of the exercise is to see whether I should stick everything into benchmark as soon as I get it without price considerations or try doing something outside of it by doing stock/fund picks and/or timing of purchases. It would be nice to prove to yourself that any "active" management you do (including both investment selection and timings of those selections) is better than the benchmark.

And since I see different posts talking about how people did compared to S&P500 or some other benchmark, I wanted to know how the comparison can be done... Of course, if you don't have any additions / withdrawals, it's easier - also probably means it does not include tax effects or that the person concluded that tax effects would be the same.. speaking of which, Audreyh1, even though you did not have withdrawals during the time periods you looked at, presumably you would still have to pay taxes for that period. Your rebalancing act may have triggered some taxes that would be smaller or greater than if you had everything in S&P500... then again, perhaps it's all in a tax-sheltered account.

 

[FIREd]

If I understand how the XIRR calculation works, I don't think you have to get very crafty to compare your results to a benchmark. The calculation returns an annual rate of return that can be interpreted as follows:

Let's say you started investing in 2000. During the years you made contributions and withdrawals to your accounts. Excel says your XIRR over this period is 8% a year. What it means is that if, instead of investing your money in stocks and bonds and whatever, you invested it in a savings accounts paying a fixed 8% a year in interests, you would have exactly the same amount of money today. So by calculating the XIRR, you already benchmark your returns so to speak. The only thing you have to figure out is whether or not that benchmark beats the annualized return for the S&P over that period of time (S&P today-S&P in 2000+dividends). You do not have to take cash flow in and out of the S&P into account in order to make that comparison. You should be able to compare your XIRR to any published S&P annualized return.

 

[audreyh1]

Audreyh1, even though you did not have withdrawals during the time periods you looked at, presumably you would still have to pay taxes for that period. Your rebalancing act may have triggered some taxes that would be smaller or greater than if you had everything in S&P500... then again, perhaps it's all in a tax-sheltered account.

No, as it happens, during those time periods it didn't trigger taxes even though it was all in a taxable account. I happened to have enough losses to counteract gains.

Audrey

 

[smjsl]

If I understand how the XIRR calculation works, I don't think you have to get very crafty to compare your results to a benchmark. The calculation returns an annual rate of return that can be interpreted as follows:

Let's say you started investing in 2000. During the years you made contributions and withdrawals to your accounts. Excel says your XIRR over this period is 8% a year. What it means is that if, instead of investing your money in stocks and bonds and whatever, you invested it in a savings accounts paying a fixed 8% a year in interests, you would have exactly the same amount of money today. So by calculating the XIRR, you already benchmark your returns so to speak. The only thing you have to figure out is whether or not that benchmark beats the annualized return for the S&P over that period of time (S&P today-S&P in 2000+dividends). You do not have to take cash flow in and out of the S&P into account in order to make that comparison. You should be able to compare your XIRR to any published S&P annualized return.

Not quite FIREdreamer. You are comparing your real 8% return from XIRR to "overall" return of S&P500 instead of what your particular return in S&P500 would have been. If you take a look at my example with T1, T2, T3 timestamps in earlier post, the benchmark's return is 0% overall but if you invested in it INSTEAD OF investing in your own fund, you would get a 33% return over invested amount (exact XIRR for it would be different of course depending on the relative T1/T2/T3 timings).

Put it another way, look at the past year - if you got 8% overall return during the year but you happened to invest a lot of money from incoming new funds in March, you really underperformed S&P500 benchmark (i.e. you would be much better off investing in SPY instead of doing your own thing); but if you invested most of your money a year ago, your results would be much closer to S&P500. The 8% XIRR return takes into account your actions and timing over the period and to compare it with S&P500 you have to really match it with what you would get from S&P500 using the same additional and withdrawn funds at the same times... (Note: I am talking about additional funds to your overall portfolio and withdrawals from your portfolio like taxes... funds that you sell but stay within your portfolio do not affect the benchmark portfolio since that is just your own market timing.)

 

[audreyh1]

smjsl - you sure are spending a lot of time figuring out how to model this!

But since you are trying to do something that I did, I will tell you what I did and the outcomes.

First of all - I never considered the S&P500 to be my benchmark, because it was very concentrated in large cap growth stocks in the late 90s when I started my portfolio. Plus it was market-cap weighted which I always thought was crazy anyway.

Since I was setting up a portfolio that had a given bond and stock allocation, I picked a well respected balanced fund with similar allocation as my "benchmark".

And the idea was for the same reason you did - if over several years managing my own asset allocation did not outperform the balanced fund, then what is the point of managing my own AA? Why not put in all in the balanced fund?

And to achieve a reasonable comparison, during the 2.5 years I was averaging into my target asset allocation, I put an additional chunk into the balanced fund. It was about 10% of my portfolio.

So that made comparisons relatively easy using the Quicken ROI functions which takes into account additions and withdrawals. For any time period, I could look at the ROI of my portfolio not including the benchmark, and compare it to the ROI for the benchmark fund. Straightforward and accurate comparison.

Now - some interesting stories about using a benchmark:

I started out using WEBAX. During the early 2000s, my AA portfolio did at least as well, but then I became aware of other balanced funds that were doing better - namely DODBX. That became my benchmark for several years outperforming my own AA, and by 2005 I was seriously considering just putting any new money that came my way into it and that is what I started doing.

But low and behold - 2006, 2007 - my more broadly diversified AA portfolio started to outperform DODBX - blowing it away in 2007, and really crushing it in 2008. Yep, new money I had added to DODBX in 2006/2007 was hurt badly whereas my original AA portfolio did way better. Of course the reason is simple - DODBX rode the wave of large cap value doing well much of 2000s, but by late 2000s large cap value did very poorly. My more widely diversified AA portfolio was behind for a while and then it screamed ahead.

So what happens? You guested it! Now I have yet another benchmark! OAKBX! But given the performance of my own AA portfolio over the past 5 years, I have renewed confidence in it and will stick with it. Where any new money goes - I don't really know. But I probably wont have that much new coming in anyway as I no longer have many assets outside of my retirement portfolio.

What is the moral of this story? Even if you pick a benchmark and then decide to abandon your own AA in favor it it, you might still encounter periods where your benchmark underperforms. Past performance does not guarantee future results!

It's a tough call!

But I still think it's useful to have some comparisons - it keeps you honest, it forces to you look at some of the details you might otherwise gloss over.

Audrey

 

[smjsl]

Thank you Audrey.

I think your benchmark picks clearly make more sense than s&p500. It's good to read about your experience and it's true - in the end underperformance in one period does not mean it in the next one (esp. if you learn something in the process).

I still think proper comparison would imply any additions and withdrawals from portfolio had to be added proportionally to your benchmark for proper comparison of your own picks vs what would have happened with benchmark (i.e. $X addition to portfolio must have been done with 10% of X going to your benchmark)... but like I said before, degree to which this matters depends on how large your contributions / withdrawals were relative to portfolio size.

In any case though, your experience gives me something to think about. Thanks!

 

[audreyh1]

The ROI function takes additions and withdrawals into account over any given time period, to comparing two portfolios works reasonably well.

While I was building the portfolio, I did do the consistent contributions to both portfolio and benchmark - 10% went to the benchmark.

Now, since I am retired, and any contributions/withdrawals are small compared to the size of each portfolio, I don't worry too much about what residual performance gain/loss might be attributed to contributions/withdrawals during any comparison time period.

Audrey