KEEPING TRACK OF YOUR INVESTMENT PERFORMANCE
Measuring your portfolio's performance is necessary to honestly evaluate
whether its excess returns (over the passive index approach) adequately pay
for your time and effort. Of course you may enjoy the 'game' for its own
sake. Regardless, you need to know how much your fun may be costing you. Research shows that investor's perception of their own performance has zero correlation to their actual results.
The following three steps (weekly, monthly, yearly) should keep you in charge of your portfolio,
without allowing it to take over your life.
Investors should make a conscious decision how frequently to monitor investments. The more frequently a portfolio is monitored,
the more investors are driven by market noise and emotions, rather than long-term
fundamentals. Frequent monitoring allows you to 'see' temporary losses. You react and sell, even while the asset would most likely continue to appreciate. Lack of attention prevents flip-flop decisions, bandwagon jumping, and frequent trading. Ignorance truly is bliss. If you must look at your portfolio frequently, look only at its total value, not the individual positions.
Monthly: The spreadsheet for Portfolio
Transactions gives you more information and history than any of the
pre-packaged software. The month-end results are frozen and saved, and the
YearToDate profit totaled. Your tax return can be checked against its distributions and sales columns. The spreadsheet reports in your home currency, so there is no delusion
about FX risks. Each January you start with a fresh slate and forget the past - the past is 'sunk'. You benchmark on a calendar year basis. You need this detailed history in order to improve your investing. You can see;
- the portion of your profits coming from income vs capital gains.
- the turnover speed of your portfolio.
- whether you got rid of losing positions fast, or hung on for large losses.
- whether your losses were from FX or the actual stocks.
- whether your gains were from a broad mix of stocks, or from big winners but only a few lucky picks.
- whether 'sales' were months after profits peaked or timely.
- whether you did a better job picking debt securities than common stock or foreign stock.
- whether the $$profits from that risky stock you sweated over, were really meaningful to your overall profits.
Click to enlarge.
Multi-Year: At year-end add your results to the Multi Year Returns
spreadsheet. It gives you positive (hopefully) feedback for renewed
savings and investing. It allows you to compare your long-term returns
to the index's compounded returns, and plan for the future. The following is an example.
Notice that each year's return is weighted equally. The multi-year rate of return generated is comparable to other people's. When asked what your long-term rate of return is, this un-weighted calculation is your answer. This is how every index is calculated. This is what every mutual fund does. That is what you would do with any benchmark you use.
Factor in the timing of your savings/draws by creating a fictional benchmark portfolio that increases with your actual Savings in the same way as your actual portfolio (the right column above). Compare your portfolio's ending value to the benchmark's ending value. Here you see that although your annualized returns (in the box at bottom) show you outperforming the index in recent years, that has not been sufficient to overcome the effects of poor results when younger --- in spite of the recent great % returns being earned on a larger $ portfolio.
PORTFOLIO YEARLY RETURNS
Visualize your portfolio as a list of securities that includes a line for cash and maybe a line for debt. Imagine a circle around the portfolio. You measure the value within the circle at the beginning and end of the year. You factor in additions and withdrawals that cross the boundary of the circle. The timing of those cash flows affects how much principal was at work earning profits.
Dividends and interest income do not cross the boundary. When received, the cash gets added to the line for cash. It never jumps outside the circle. It stays within what you are measuring. Same with purchases and sales of securities - the cash stays inside the circle. Debt should also be inside that mental circle around the portfolio, so that the measurement of the portfolio's return will include the effects of leverage. If borrowed from your broker, the negative cash balance will be included in the broker's statements, and interest payments will move cash around inside the circle, but not jump over the boundary.
If you borrow to invest using a bank's Line Of Credit, include that negative value inside the circle. Its interest charges would show up as a negative Distribution in the portfolio tracking spreadsheet above. If its interest costs are not paid and increase the debt's balance, then there are no cash flows across the circle's border. But if the interest is paid from some other account, the $ interest is treated like added savings crossing the boundary. It shows up in the spreadsheet in the Sales column of the Cash line.
The best way to measure your rate of return is the same way mutual funds do - by 'unitizing' your portfolio and its cash flows. For this you need the value of the portfolio just before each cash flow. The ratio of the cash added/subtracted as a percentage of the account's value at that time, determines the number of units added/subtracted. The Portfolio Transactions spreadsheet includes a box for this calculation.
The example following starts with an arbitrary 100 units to make the math easy. There is one withdrawal and one addition during the year. Each time, the value of the account was recorded and the unit count adjusted for the additions/withdrawals.
|No. of |
|start ||$100,000||100||$1,000 || |
|withdraw ||$105,000 ||100||$1,050||($10,000)||10,000 / 1,050 = (9.5)|
|add ||$98,000 ||90.5 ||$1,083||$25,000 ||25,000 / 1,083 = 23.1|
|end ||$119,000 ||113.6 ||$1,048 |
|Return = ||(1,048 / 1,000) -1 = 4.8%|
This math produces a % Return that can be compared to other reported returns, e.g. indexes, benchmarks, other people's returns, your own returns of other years. It measures the results of your investing decisions. It removes from the calculation the impact of cash flows that result in larger / smaller portfolios during period of the year when returns were high /low. It presumes that cash flows were not strategically timed: you put in cash because you had saved it, you pulled cash out because you needed to spend it.
This is the same methodology used to measure the performance of professional fund managers. These people have no control over when cash flows in or out of their portfolios, so their returns are calculated 'per unit'. Benchmark indexes also use this methodology. E.g. the S&P 500 index is quoted as the price of one unit. Its yearly return is the return of one unit held for the full year without additions or draws. This methodology is commonly called the Time-Weighted Rate of Return.
The competing methodology is called the Dollar-Weighted Rate of Return. It uses math referred to as IRR (Internal Rate of Return). It can also be called the "Personal Rate of Return". This is the metric used by Canadian brokers starting in 2017 to calculate your account's return. They are forced to use this metric by the regulatory CRM2 reforms. As a result the industry has sold this metric as 'more correct' and 'personal' than the Time-Weighted Return. Garbage. They chose this methodology because its calculation requires input data that brokers already collect.
To judge its value for yourself, consider a market ETF priced at $100 on Jan 1, that stays steady at $100 for the whole year. On Dec 31 it jumps to $120. The stock's return is 20%. What would your return be if you had invested $1,200 on Jan 1? Your portfolio would be worth $1,440 at year end for a 20% return - just like the stock itself. Your profit is 1,440 - 1,200 = $240. For both the T-W and D-W methods, the return would be 20% ($240 profit / $1,200 invested).
What would your return be if you trickled your savings in over the year, started with just 1 share worth $100 and added $100 savings (= 1 share) at the start of each month? By Dec 30 you have added $1,200 - still valued at $1,200 because the stock price is flat. But on the last day the price rises 20% making your portfolio worth $1,440. The Time-Weighted calculation would consider the first purchase to be 'one unit'. Each monthly purchase added another unit - totaling 12 units at year's end. The $1,440 ending value = $120 'per unit', with a $20 (= $240 / 12) profit 'per unit'. The T-W return would be $20 / $100 = 20%. This makes sense because you invested only in the one stock that itself earned 20%.
But the D-W Return would be 38.78%. Essentially IRR starts from your $240 profits and divides that by the average $ invested throughout the year. Since the $1,200 trickled in evenly over the year, it is reasonable to say that is equivalent to $600 invested all through the year. The $240 profits / $600 invested = 40% return. This higher metric congratulates you on your wisdom to delay funding the account.
The D-W Return implicitly presumes that the 'other $600' - the savings added while NOT yet in the account - also earned a 40% return. This is the metric's problem. It is not a valid presumption for most all investors. Historically this methodology was created for businesses to choose between competing projects. In that situation the implicit presumption would be correct. The metric's use by business for this purpose is widely accepted. But for measuring the returns of a retail investor it is meaningless.
Be very clear that even the industry body requiring use of this metric explicitly warns that ... "benchmarks are not relevant comparisons to an individual investor’s Personal Rate of Return". The D-W return cannot be compared to other investors' returns or even your own in other years because it ... "is based on the individual investor’s specific deposits into and withdrawals. Since each investor has a different combination of deposits and withdrawals, each investor could have a different Personal Rate of Return".
Should you pat yourself on the back for almost doubling the return of your benchmark? No. Were you really smart in choosing assets that returned more than the market? No, you just bought the market ETF. Did you made a strategic timing decision? No, you just added savings as you were paid. Is that 38.78% comparable to anything quoted in the markets? No.
INDIVIDUAL SECURITY RETURNS
The easiest calculation of returns from a single position is if all dividends received are used to purchase more shares of the same security (as in a DRIP), and there are no subsequent purchases or sales. In this case only, you can be accurate and include the compounded profits from the reinvested income.
- Record the market value of all the shares owned at the beginning ( or at purchase). E.g. 500 shares at $2.00 each = $1,000
- Do the same with all the shares at the end of the period. This will include the shares bought with reinvested dividends. E.g. 520 shares at $2.35 each = $1,222
- Your profit is the difference between the beginning and end. E.g. 1,222 - 1,000 = $222
- Ignore the $$ received as dividends during the measurement period because their value is now reflected in (2) above.
- Divide the $profit by the $invested. E.g. 222/1,000 = 22% profit from both income and capital gains.
When dividends are not DRIP'd the return calculated for an individual security will never be perfectly accurate. To be accurate you would have to include the profits from where ever the dividend was reinvested. But most people would attribute those gains to the second security bought with the dividend dollars. So it is necessary to assume $0 profits from reinvested dividends. Still assuming only one purchase (or beginning $) and no sales, the calculation of the holding-period return is the ratio of the $ Profits divided by the $ Invested.
- Calculate the increase in market value. E.g. you bought for $100 and it is now valued at $150 for a capital gain of $50.
- Add the $ dividends or interest received. E.g. you bought at a 4% yield and received 2 of the 4 yearly distributions. 4% of $100 divided by a half year = $2 income
- Divide the $ Profit by the $ Invested. E.g. 52/100 = 52%.
Multiple purchases (or sales) make things a lot more complicated. There is no 'correct' way to handle this. The Rate of Return is the ratio of the $ Profit divided by the $ Invested.
- The $ Profit in the numerator is easy. Starting $ + Purchases - Sales - Dividends + Profits = Ending $. Solve for the Profits.
- The $ Invested in the denominator depends on the timing of the subsequent purchases and sales. How much money was at work earning that profit on average throughout the period? Simple assumptions will give you a 'good enough' result.
- For transactions just before the end of the year ... divide by the Starting value only because that is all that was at work for most all the year.
- For transactions right after the start of the year ... divide by Starting value +/- all of the purchases / sales.
- For transactions in the middle of the year, or in multiple flows through the year, assume half the amount was available through the whole year. Divide by the Starting value +/- half the purchases / sales.
Those adjustments for purchases and sales within a calendar year will produce a calculated rate of return the is 'good enough'. But when you try to measure a security's return over many years with many cash flows in between, the adjustments won't work. You would have to use an IRR Calculator.
The rate of return metric you should ignore is the one published by your broker. This cumulative % return covers the change in stock price over the full period of ownership, ignoring dividends. It is meaningless because each position you own will have been held for different time spans. The only value of your broker's metric is to warn you of income tax on sale.
We always compare our returns to some metric - usually a large index. It is human nature. It is important feedback you need for deciding whether you should be stock-picking or passive indexing. So how do you pick the correct index?
Use the 'total return' variant of any index (as opposed to the most common 'price return'). Dividends and interest must be considered reinvested to earn an additional return, just like the cash in your account is used for new purchases. It should make no difference if your holdings pay a higher/lower yield than the benchmark index. It would only be appropriate to use the price-return index if your portfolio's yield equals the index's yield, and you are withdrawing all dividends when received (e.g. in retirement), and nothing more.
You may consider choosing the 'equal-weighted' variant of your index if it exists (as opposed to 'cap-weighted'). There are arguments pro and con. If your own portfolio is a rough basket of equally weighted stocks, with sector weights more equal than the economy, an equal-weight index may seem more comparable. But your portfolio may be a blend of index ETFs, with high-conviction stock picks making up only 50%. Then a cap-weighted index is more comparable. Also you are probably more likely to include large-cap stocks in your stock picks just because they are so big.
The most commonly heard advice is to calculate a personal benchmark from a blend of indexes based on your asset allocation policy. Eg. for a 60/40 portfolio use the S&P Index (times 0.6) plus the government long-bond index (times 0.4). But when you create benchmarks to look exactly like your portfolio, you essentially guarantee your portfolio returns will match the benchmark. Which destroys the whole point of benchmarking.
Better to admit you would really like to earn the higher returns from 100% equity, but recognizing your intolerance for volatility, you feel it worthwhile holding some bonds as long as they only reduce your returns by (say) 1.5%. So you benchmark against 'the S&P 500 index less 1.5%'. This benchmark gives you feed back on your asset allocation choice.
This benchmark allows you to see more clearly the ongoing cost of holding low-return assets - a cost you only recover during years of market crisis. It prevents you from fudging results by changing the benchmark debt's maturity. You would have seen clearly the windfall benefits from owning debt since 1985, as yields have fallen - even though the debt was bought for safety without the expectation of capital gains.
The measure of the '% income you are willing to lose' gives you perspective when buying put options instead of asset allocating. A 60/40 portfolio can easily cost you more than 1.5% each and every year. Multiply that 1.5% by the benchmark index's price. Then look up the 1-year put option you can buy for that same price.
Think twice before prorating your benchmark with more indexes. Mostly the addition of multiple asset classes is because you feel they will somehow goose your portfolio's return. Your decision to allocate to foreign stocks, junk bonds, REITS, or small caps is a strategic choice. You benchmark to find out whether that choice was correct. So the benchmark itself cannot include those asset classes.
Passive indexers often dismiss the proof that individuals can outperform the market, by arguing that the 'wrong' index was used for the benchmark. E.g when Warren Buffett outperforms the S&P Index, they claim "He did not outperform the Value Index and that is what he owns". This argument is wrong because it uses hindsight to determine the benchmark. Buffett is free to own stock, options, debt, futures contracts, etc. He was never constrained to value stocks. He chose what he owns freely. He made the choice to buy value stocks, and beat the market, so he deserves the kudos. Slightly off-topic, here is an excellent analysis of all the factors going into Buffett's good returns (from luck, aptitude, personal contacts, family, business model, business structure, investing choices).
Passive indexers also dismiss stock-pickers' returns by arguing that stock-picker's returns should be risk-adjusted. The main metric used for risk-adjusting is Beta, but Beta was proven useless decades ago. The Fama-French 'risk-factors' have also been proven to have nothing to do with risk. See the discussion on the Beauty Pageant page. Even ignoring the problem with effecting any risk adjustment, you don't NEED to risk adjust. You are probably just fine with whatever risk you end up with. If you can tolerate volatility why should you not recognize any out-performance that results? See the long list of reasons to not risk-adjust on the Active vs Passive page
Just as important as picking the correct benchmark, is correctly stating your own returns. No cheating.
- Include the interest costs of money borrowed to invest, even if it is secured by your home.
- Include the foreign exchange gains/losses in converting to the currency which you will eventually be needing. That may be your home currency if you plan to stay put in retirement. It may be a foreign currency if you are planning to move.
- Remove the savings added to and cash withdrawals from the portfolio, from your calculation of profits .
- Include all your accounts in one calculation. Chances are you will have tried to optimize taxes by keeping different types of securities in different accounts. But dollars are dollars no matter what the account label. In retirement or in an emergency it won't matter where the cash comes from.
- Don't benchmark against your friends because it is highly likely they don't know how to correctly measure their returns and conveniently fudge the numbers.
- Don't use arithmetic averages for long-term results because that overstates the true geometric average.
- Don't use Yield-On-Cost YOC calculations that measure nothing meaningful and are meant to make you feel your returns are larger than they really are.
- Don't ignore capital gains if you have decided you are a 'dividend investor'. Ignoring feedback is putting your head in the sand. If your strategy is not performing you should know it.
METRICS FOR DIVIDEND-GROWTH INVESTORS
All the discussion above uses Total Return for measuring performance. In the 2000s, the cult of dividend true-believers invented the myth that their objectives and strategies require different performance metrics. All refuse to use Total Return as their primary metric. Many refuse to use Total Return at all. For them, everything is analyzed ONLY with Yield On Cost (YOC), Dividend $, and Dividend $ Growth.
But there is a disconnect between what the gurus preach and what really matters to them, and probably to you. When given a variety of portfolio results from the choices below, and asked which performed the best, their common sense prompted them to choose the largest $50,000 Portfolio D ... even when the performance metrics they promote indicate quite different choices, and were displayed for them to see.
Failing to understand the point of the exercise, they felt free to ridicule the question because the answer was so obvious. But that ridicule only validates the conclusion that what really matters at the end of the day, is the Total Return - even for Dividend-Growth investors. The gurus preach the use of other metrics, but what they really value is Total Return.
It is not necessary for the point being made here, but if you like, here are the 'stories' behind the portfolios.
* The only difference between Portfolios A and B is that, on the day before measurement, A switched stocks to buy higher yielding ones. This reset its YOC back to the current yield. But now he can brag that his portfolio is throwing off the largest $$dividend income.
* B was happy to brag how his decision to buy and hold high yield stocks resulted in his portfolio having the highest YOC.
* C chose to start with lower yielding growth stocks with a smaller payout ratio, whose management undertook to both grow the company and then increase the payout ratio. This allowed C to brag how his dividend$ grew at the fastest rate.
* The only difference between Portfolios C and D is that company management of the stock owned by D decided they could make better use of profits by reinvesting for growth, instead of paying out dividends. The trade-off worked. Although D's dividend income growth was lower. the increase in stock price more than compensated - creating the highest Total Return.
Which Performance Metric You Use Determines Which You Think Performed Best.
- Dividend Size
- You would pick Portfolio A with the largest $2,000 dividend. But why not Portfolio B? The difference was determined by one day's switch of stocks. Switching stock does not affect performance. Performance is created by stocks DURING THE PERIOD they are owned. Portfolio A simply 'put lipstick on the pig'. This metric is superficial and does not measure performance.
- Yield On Cost
- You would pick Portfolio B's 12%. But investors don't get bonus points for holding stocks without ever selling. The objective of investing is NOT to see who can hold a stock the longest. You may think that long holding periods are a good strategy for maximizing your returns, but they are never the objective. Many true-believers make a 10% YOC their objective. They base their retirement date on the assumption that a 10% YOC will cover their cash needs. But is Portfolio B really in a better position to retire? No. There is a discussion of YOC on the Valuation Metrics page.
- Dividend Growth
- You would pick Portfolio C's 12% dividend growth. But it is easy to grow fast when you start small. A company can double a small dividend easily, because growth can come from both an increase in earnings AND an increase in the payout ratio. But a company paying out 50% of earnings cannot double that to 100% of earnings without destroying all growth. This metric reflects a choice of strategy. It does not measure the resulting performance. By starting with low yielding stocks you give yourself a handicap to make the results look better.
Dividend true-believers feel that a 5% growth in their portfolio's $dividends is a better measure of success than a 5% increase in their portfolio's' $value. They are distorting dividend growth into a measure of rate-of-return. But a $dividend increase does not provide any 'return'. An increase in $dividends declared by a company can be replicated by simply switching stocks into one that pays a larger yield. Yes, an increase in $dividends may indicate that management thinks long-term profits will be higher. But if the stock price does not change, then the market believes the increased dividends will come at the cost of lower growth. It is only when the stock price increases as a result of the dividend increase, that the investor realizes a 'return' - a capital gain measured by the Total Return metric.
- Total Return
- You would pick Portfolio C with the 8% growth. This is the way mainstream finance measures portfolio performance - with rates of return that are comparable to other data. Total Return measures growth in value - value that can be used in the real world to buy more things today - or used to buy a larger income stream for the future.
It makes no difference what strategy you use. A larger ending portfolio will always give you more spending power then, or a larger income stream for the future. Only the Total Return metric measures this.
P.S. If you are thinking "Oh, this does not apply to me because I am not accumulating wealth. I am retired. What matters to me are withdrawals." Think again. Add an assumption to the story of all the portfolios, that (e.g.) $500 is withdrawn each year. What would change? The metrics for Dividend $$, YOC, and Dividend Growth would stay the same. The ending Portfolio Values cannot be measured without the year-by-year timing of returns and draws, but regardless .... Portfolio D would still have the largest ending $value. Portfolio D would still have the largest Total Return metric. The largest Ending Value allows for larger $draws today AND also provides the greatest promise for future income. Only Total Return measures this performance.
TRACKING the ACB for TAXES
Canadians must keep track of the ACB (adjusted cost base) of stocks and mutual
funds owned in Taxable accounts.
include all costs means you will eventually pay more tax than you
should. Only securities held outside tax shelters are of concern here because profits within RRSPs, RRIFs, RESPs, RDSPs and TFSAs are tax free. All transactions done in foreign currencies are translated into the Loonie equivalent at the exchange rate on the transactions date (not the settlement date).
You increase the ACB of your stocks when ...
- shares are purchased (either originally or additionally later). It makes no difference if they are in another broker's account. All are considered lumped together. The additional purchase is added exactly like the original purchase.
- distributions are automatically reinvested to buy more
shares. This can happen three ways.
1. DRIPs are set up to automatically reinvest all the distributions received, hence the name Dividend ReInvestment Plan.
2. ETFs and mutual funds distribute shares instead of cash in order to pass through taxable capital gains - usually at year end. With ETFs the number of shares you own may not
increase because the fund makes adjustments inside the fund. You know it has happened because your tax slip will include a capital gain amount for which you never received a distribution. These are called 'phantom distributions'.
Regardless, the distribution reinvested increases your ACB.
3. Sometimes when a corporation sells a division, it realizes capital gains that are better
realized in the hands of the shareholders. Because the company may not have cash available, they pay a capital gains dividend with shares instead of cash. Because there is no cash flow some investors ignore what is happening
'behind the scenes'. They think their original asset is simply growing in
value. In fact there are two transactions lumped together. The company distributes
dividends, then the investor uses that cash to buy more shares. Each purchase
of more shares increases the ACB by the amount of the dividend. Some brokers' do not include the $value of the transactions, only the number of shares issued. If so you should complain. At any rate the $total of the
year's distributions will be taxable and T5'd. That taxable amount will be the cost of the
additional shares purchased. The common errors resulting from not doing this correctly are discussed on the Valuation Metrics page.
- A tax loss is denied (superficial loss in Canadian tax) because
replacement shares are purchased too soon afterwards (within 30 days). In this case the
loss denied increases the ACB of the replacement shares so that it can
be claimed when they are ultimately sold. In the example below the Loss on Sale equaled $980 : = Proceeds ($3,000) less ACB (200*19.90=$3,980).
You decrease the ACB ...
- when all or some of the shares are sold. The same percent of the ACB, as the percent of shares sold, is removed. This amount goes into the calculation of capital gain for those shares sold - proceeds less ACB.
- when income trusts distribute what they call Returns of Capital (ROC). This
information is noted on your tax slips now.
- when flow-through Oil&Gas shares distribute deductible operating costs.
Staple a sheet of paper to the back of the broker's slip for the original
| ||Date ||CashFlow||#sh||ACB||ACB/sh
||b.) ||Div Reinvested||Dec||0||10||150
| ||Subtotal B4 Sale||510||10,150||19.90
|i.) ||Partial Sale||Mch||3,000||(200)||(3,980)||19.90
|c.) ||Loss Denied||(repurchased too soon)||980|
|ii.) ||ROC Dist'n Rec'd||Jun||200||(200)
| ||3:1 Stock Split||1830||10,550||5.77
Canadian Personal Income Tax returns require you to list all the securities sold during the year on Schedule 3. It is MUCH easier to update this schedule at the time of each sale, rather than at year end when you have forgotten or lost your records. The math is simply subtracting the ACB of the shares sold, from the total proceeds, to generate the gain. A calculator and piece of paper work fine. Or you can keep a simple spreadsheet.