UNDERSTANDING CHANGES TO OWNERS' EQUITY
Since the mid 1990's investors' ownership interest in companies has been impacted MORE by changes to the companies' equity than by the companies' earnings. This is not an exaggeration. None of the reporting systems measure the effects of financial engineering. They are all geared to report on the company as a whole - not the portion attributable to a stock. Investors must learn how to correctly interpret and measure changes to equity.
Rates of Return To Investors
The different uses of profits generate different rates of return for the investor. This should factor into your decision on which is preferable. This table shows the math calculations. It is automated in the Different Yields spreadsheet for different input assumptions.
|Use Of Profits ||Multiply ||Rate of Return ||Equals||Return |
|25% ||to Dividends||* || E / P ||5% ||= ||1% |
|40% ||on Buy Backs||* ||E / P ||5% ||= ||2% |
|35% ||Reinvested||* ||ROE ||33% ||= ||12% |
|100% || || || ||Total||15% |
- Dividends paid on common shares are not created out of thin air. They are a transfer of value from the company to the owner. Each dividend dollar paid out of a business, reduces the value of that business (and the owner's investment) by the same amount. This is completely different from dividends paid on preferred shares. In that case the dividends ARE income because the principle (face value of the preferred share) has not been affected.
- Dividend dollars earn the investor a rate of return usually far lower than the return earned by reinvesting in the business. This depends on the incremental ROE, but generally cash inside a business is more productive than cash in the secondary markets.
- Management's choice to pay dividends indicates a lack of better business opportunities, or a preference for a certain type of owner.
- Companies with large DRIPs know the cash for those dividends will never leave the company bank account. With (say) 25 percent of investors in DRIPs, the company can declare a dividend one third higher than it could otherwise afford.
- When some shareholders DRIP and others take cash dividends, neither class of owners ends up better or worse off than the other.
- Management compensation that depends on capital appreciation may contain clauses protecting management from large dividend effects. The number of stock options held can be grossed up by the capital lost to the distribution. Similarly, when convertible securities are issued to 'friends of the business' there can be a provision to lower the options' exercise price because of distributions. The resulting dilution is not captured by any reporting metrics.
- Dividends can be used to hide management compensation, because dividends are not a deduction before Net Income. Shares are issued to management for debt. Dividends paid on those shares are journal-entry'd to pay down the debt and should be considered management wages. The issue of shares is just an excuse to label the cash 'dividends' instead of 'wages'. The exact same thing could be accomplished by paying wages and requiring the executive to buy shares with it. It is impossible to determine from the financials what the dollar value of this is. While this strategy will always raise reported Net Income in total, it will only raise EPS when the dividends are greater than earnings, so you see this strategy in REITs and Income Trusts with high distributions. See separate page of explanation.
- Dividends can be used to hide management compensation a second way. When shares are repurchased and not cancelled (or issued without being sold to outsiders), they may be in a trust structure for the benefit of management. That means any dividends received by those shares do not accrue back to the benefit of the company as a whole. They increase the value of the trust - usually spent buying more shares for the trust. This eventually is distributed to management.
- Dividends can be used to hide interest expense. It is not unusual to see the AVERAGE number of shares outstanding in a period being greater than the number outstanding at both the BEGINNING and END of the period. This may be the result of a delay between the time of issuing new shares on the exercise of stock options and the time of repurchasing them from the market. But this may also indicates the existence of financing with company shares. There may be an agreement with a bank to sell them shares for cash, and a repurchase price agreed at the start. That price includes interest on the cash 'borrowed'. Any dividends issued during the period would reduce interest added to the repurchase price.
- Share buybacks are equivalent to the payment of dividends. The cash moves from the productive operating sector of the economy to the secondary markets, just like dividends do. If 5% of the outstanding shares are repurchased then the investor has a 5% gain.
- Unlike dividends, the cash for buybacks goes to people selling their shares, rather than to the continuing owners.
- Unlike dividends, no taxes on owners are triggered by share buy-backs.
- When offset by the issue of additional shares (say, from the exercise of options), there is no benefit to owners. This is ignored in the many reports claiming that buybacks have replaced (historically much higher) dividends.
- The timing of share buybacks should be opposite from the time chosen to exercise stock options. Options will be exercised when the owner believes the stock price is peaking. Companies should invest in their own shares when the stock price is bottoming ... for the same reason everyone buys shares (buy low - sell high). In practice, companies may buy back shares BECAUSE of options being exercised - at the exact worse time.
- When US domiciled multi-nationals earn profits in foreign jurisdictions, the cash profits would be taxed if repatriated for share buy-backs. So US companies borrow in the US to fund buy-backs. This has two effects. Even while the foreign cash is not accessible it offsets US debt for net Debt/Equity ratios. While interest rates are low this strategy works fine, but when interest rates rise the company must either continue paying the higher interest, or else pay taxes on repatriating the foreign cash. Not good options.
- Shares may be bought-back, but not cancelled, at the time options are granted, at the price equal to the exercise value of the options. This creates the superficial appearance that the options are 'fully funded'. When the option is eventually exercised, those shares in treasury are used, with no change in dollar value recorded, or change to the gross number of shares outstanding, in Shareholders' Equity. In reality this does not change the opportunity cost of options. The two transactions should be measured separately. See the details at Comprehensive Income
- How investors should interpret treasury shares (showing as negative Shareholder Equity) depends on whether the treasury shares are effectively owned by all the shareholders, or whether they sit in a trust for the benefit of management only. The 'number of shares outstanding' used in the calculation of 'book value per share' and 'earning per share' should be the greater (gross) number when purchased shares are owned by management. Otherwise, the calculations use the reduced (net) number of shares. This also means the calculation of ROE based on 'eps' and 'book value per share' will differ from the metric based on Net Income and Equity. See this example from Legg Mason.
- Since buybacks are financially equivalent to dividends it is reasonable to conclude that valuation calculations based on the present value of future dividends should include buybacks as quasi-dividends. But it has been argued that this results in double counting, when the analysis is done on a per-share basis. The cash used for the buyback is the first 'count'. But the number of shares outstanding has been reduced as a result, so in future, the same amount of cash (in total) can fund a higher dividend (on a per-share basis). This growth in the dividend is the second 'count'.
- When profits are reinvested to grow the business, the investor will earn a rate of return equal to the ROE of the incremental business. The company's expected rate of growth equals the multiple of this incremental ROE and the proportion of profits reinvested.
- If retained profits are used to pay down debt, the investor's return will equal the interest rate of the debt.
- Whether reinvesting profits is better than paying dividends depends on the use to which the cash is put - the incremental ROE.
- Dividend Reinvestment Plans recycle the nominal dividend's cash back to the company by paying with script (shares) instead.
- This is equivalent to reinvesting profits and not paying dividends at all ... but inefficient due to costs.
- DRIPs are not equivalent to the shareholder receiving cash dividends, to be re-deployed in the secondary market. The cash is retained within the business where it is hopefully more productive than when in the secondary market .
- Investors earns a rate of return according to the same criteria as seen above in "Reinvesting Profits".
- A company's issue of shares and then payment of a dividend with part of the proceeds, is equivalent to an investor selling a portion of their shares in the secondary market. See example below.
- If additional shares are sold for market value, the investor is indifferent.
- If sold for less than market value, the investor suffers a loss equal to the discount. That loss will equal the gain to the buyer of the discounted shares - usually employment compensation from exercising stock options. Investors must measure this loss and deduct it from published Net Income and EPS. See the details at Comprehensive Income.
- Issuing shares does not dilute existing owners' wealth because $proceeds are received in exchange. The question is "Will the $proceeds be put to work earning the profits anticipated by the market price of the stock?"
- If issued at twice book value the $proceeds need only be invested at half the ROE of the pre-existing assets in order for the EPS to remain the same. If issued at 1.5 times book value, the proceeds need only earn a return equal to 2/3 of the pre-existing ROE for the EPS to remain the same. Etc. Money inside a business is worth more than money outside because it is more productive.
- If sold for less than "book value per share", each resulting share will have fewer assets working for it. EPS will be expected to decline. The company would be in dire straights before doing this.
- If sold for more than "book value per share", the premium will be shared by all resulting shareholders. Each share will have more assets working for it. EPS will be expected to increase. The increase should not be interpreted as "due to good management". It results from forces in the secondary market for shares, not management actions.
||Stock Options Align Management's Interest With Shareholders'... FALSE
- defer management's personal income taxes.
- are not measured correctly by the financial statements and so are open to abuse.
- are dismissed as a "non cash transaction" (see Cash Truths That Aren't) to gullible Retail Investors.
- are not long-term incentives because all kinds of events will
trigger an immediate vesting - events that can be of management's own
- do not serve to retain management talent because the poacher will cover the cost of options left behind.
- can be monetized (cashout without selling or triggering tax) by derivatives.
- will prompt share buybacks instead of the harder task of
growing the business with reinvestment. No effort is needed to double
the share price, when you can just buy-back half the shares.
- are free to management, in the sense that no money of their own is used to buy them. Nor have they paid income tax on the deemed cost.
- are taxed at half the rate of wages. The company loses half the tax deduction that would have resulted from paying wages.
- are recognized as ineffective by private equity, who require
management to buy AND PAY FOR shares of a value equal to many years'
||Share BuyBacks are Good ... FALSE
The decision to buy back shares is a knee-jerk reaction by
management, today. Management and the media say it returns value to
shareholders; it is more tax efficient than dividends; it is an
unambiguous 'good'. But it isn't !!!
The decision to buy back shares should be the fallout from a
thorough analysis by management of all the possible re-investments for
its earnings. The correct choice is the one yielding the greatest
return. The possible investments, in order of highest probable rate of return, are:
- A) The Core Business
- Most businesses earn a return on equity in the 10-15% range.
Reinvesting the earnings to grow the business is
usually the best opportunity for the business and shareholders.
- B) A New Business
- The next best investment would be in another business with only an incrementally lower ROE.
- C) Reduce Debt
- Most companies use debt to leverage the
return on invested capital into a higher return on equity (ROE) for shareholders. Reducing the
debt will lower the company's risk, but will also lower the ROE. You
need to compare the cost of debt to the ROIC(%) to see if
debt is good or bad. (See discussion at Leverage).
- D) Dividends
- give the shareholders the decision where to reinvest. The company
foregoes its own growth and implicitly declares it has no opportunities
that can't be bested by the shareholder re-deploying into another
company. Since business (generally) earns a ROE of 15-20% and shares
trade at twice book value, the shareholder's long-term return is only
7.5-10%. The company has decided it has no opportunity that will yield
even half what its current operations earn.
- E) Share Buy-Backs... should be the decision of (next to) last resort.
- If the share price is falling, companies should never buy-back
shares - for the same reason you shouldn't buy them personally. Your
investment would be losing value.
- While it is true that buy-backs are more tax efficient than
dividends, the downside is that shareholders are given no choice to
re-deploy into another company with better opportunities.
- Theoretically, the share price will go up an equal amount,
so you can sell for an equal profit. But in practice, the share price
does not respond so predictably.
- The biggest problem with share buy-backs is that the media
and analysts don't know how to factor them into their decision models;
the accountants don't show them on any financial statement; the
media refuses to teach the public how to calculate comprehensive earnings; the analysts using discounted cash flow models don't incorporate them.
- The media further muddies the water by integrating the issue of
stock options. Stock options and share buy-backs are two completely
separate decisions and transactions. They should be valued and reported
independently. They aren't though. So do it yourself.
- Because management compensation is half paid in options, it is in their best interest to spend every penny on
stock buybacks. The value of their options will rise without any
- The return realized by the company on its investment in its own
shares is the same as an individual shareholder's (the Earnings Yield =
flip of P/E = ROE divided by the Price/Book). This is pretty poor. Over
time it degrades the company's overall ROE. The media's simplistic
argument that "since there are fewer shares, each shareholder's stake
becomes larger" ignores the opportunity cost of that increase in EPS.
- F) Purchase an Existing Business ... usually the worst option.
- Buying out your competitor most often ends in grief. This is
especially true if the business bought was publicly traded at
multiples of book value. Making it even worse is paying a premium to
market value for control. Chances are high that the incremental return on the cost will be
less than the purchaser's pre-existing ROE - which is why the cost of Goodwill is never expensed.
When the business purchased is private, and the purchase is paid
for with shares that trade at healthy multiples, the probabilities of
success rise. The incremental ROE will be much higher. This is why
consolidators of fragmented sectors are frequently successful, at the
||Share BuyBacks Offset the Cost of Stock Options... FALSE
This is the most common way 'experts' dismiss the cost of stock
options. Sure, the number of shares outstanding does not change when
the company buys them back just as fast as they are issued. And sure,
the excess cost to buy back shares is offset by the gain from selling
your proportionate interest in the company. But ....
The following diagram shows a typical company's price structure.
The middle two columns reflect the two parts of a stock option
exercise. The first reflects the existing shareholders' gain in book-value-per-share when
they sell a part of their ownership in the company to new owners. The
second reflects the cost of management compensation equal to their lower exercise price (vs market value). The right column reflects a share buy-back.
The cost of the shares in the market exceeds what the company
originally sold them for (the exercise price), so there is a loss.
What the 'experts' ignore is the column for Compensation. The
company does not receive full market value for the options exercised.
The opportunity cost lost is management's gain. This is not a
'victimless crime'. Selling shares of a company is no different from
selling assets of the company. The shares represent a proportionate
ownership interest in all the assets. Assets given away to employees
are an expense to the company.
Buying back shares will never offset the cost of stock options.
There will always be the discount to market value that is management's
See also how to integrate these numbers into Comprehensive Earnings Per Share.
||Share Issues Dilute Shareholder Value... FALSE
When the same earnings are shared between
a greater number of shareholders, yes, each will receive less of the earnings. But new
shares are issued for a price. The proceeds are put to use
generating additional earnings. As long as the proceeds equal the
market value of the shares, the original owner should be
indifferent. He will be in the same position as if he had sold that
same % of his shares in the market.
The following example follows what happens to an owner when his company issues new shares at a market value greater than book-value, compared to the same owner who sells his shares in the market for a capital gain. They end up in exactly the same position.
- You start a business with $10,000 capital and 1,000 shares.
- The $10,000 turns out to be very productive and the company is re-valued to twice as much: $20,000.
- A new partner pays $20,000 for 1,000 new shares = half interest.
- The company doesn't need the cash so it pays $20,000 out as a dividend ($10,000 to each partner).
|Company grows || || ||$10,000|
|Your position now||1,000||$10||$10,000||$20,000|
|Issue new shares||1,000||$20||$20,000||$20,000 |
|Pay dividend|| ||($10)||($20,000)||($20,000) |
So What Happened?
- The company is not changed by the issue of new shares. The book value
of its equity is still $10,000. Its market value is still $20,000.
- Except it now has twice the number of shares outstanding - which doesn't make any difference.
- The owner
- Seeded capital = $10,000
- Received $10,000 dividend which left him with a
- Cash cost = $0, and an investment of
- 50% share in a $20,000 investment = $10,000.
this to selling half his shares in the secondary market. For 500
shares a buyer would pay $10,000, leaving the original owner with the same cash cost = $0 as the first scenario, and he would still have a 50% equity interest in a company worth $20,000 =
Conclusion: As long as the company receives full
market value for new shares, you are fairly compensated for giving up
ownership in the company. You 'realized' a capital gain from the 'sale'
of your ownership. This is exactly what happens when you see "Dilution
Gains" on a company's Income Statement. It is not an oxymoron. Their
subsidiary has issued more shares, so their ownership is diluted. But
they realize a gain because the increased value of the subsidiary has now been realized by a transaction.
||Companies Now Record the Cost of Options Compensation ... FALSE
In 2004 there was a concerted effort to force companies to measure
the cost of options, and include the expense in the Income Statement.
Users of Financial Statements lost the battle. Companies and their
accountants won. Yes, options are now measured, but not correctly. Now
you must cancel out the incorrect accounting as well as measure the
options cost yourself.
First ask: "what is the cost to the company, of compensation paid as options?".
- Can the whole thing be ignored because it is a non-cash transaction?
- Is the cost to the company the same as the benefit received by the options holder?
- Is the full cost of options determined at the time of their grant?
1. The cost of options cannot be ignored just
because it is non-cash. There is no such thing as a non-cash expense.
Either there is a barter transaction that should be considered two
separate cash transactions, or there is a timing difference between the
cash transaction and the reporting period. This argument is expanded on
the Cash Truths page.
2. Companies take the position that the total
benefit realized by management from options is NOT a cost of the
company. They claim the cost to the company is measured by the
Black-Scholes formula at the grant date. They say that the increase in
an option's value as the stock price increases, comes from 'the
market', not the company. Therefore no additional expense need be
recognized after the date of issue. This logic fails on analysis.
The parties on opposite sides of an option contract are equal and offsetting. Options are a zero-sum game - one person's gain is another person's loss. When management benefits from the option someone must lose. That someone is the business. The company cannot get rid of its option liability without either paying someone to assume it, or settling it. At no time does 'the market' assume the liability. Yes, the increase in an option's value is DUE TO the market's pricing, but the increase in value for the option-holder does not COME FROM the market.
Another way to support the argument (that the benefit to management
is exactly the same as the cost to the company) is to consider a
business where the owners and managers are one in the same people. No one ever suggests that this
management should be compensated with options. Everyone realizes the
cost to themselves as owners would negate any gain to themselves as
3. All options contracts have two transactions - the original creation of the contract when the premium is exchanged, and the final settlement. It cannot be argued that the company is the counter-party to one but not the other. The full cost of options compensation cannot be
known at the time of their issue. The future is unknown until the contract is settled and closed. The two transactions together (the opening payment of a premium and the closing settlement) determine the total cost.
Financial Statements record as an expense only the opening transaction - the calculated value of the premium. This valuation never changes, and the closing settlement is completely ignored. This violates the basic concept of the
Balance Sheet which is supposed to measure the value of the
assets/liabilities AT EACH POINT IN TIME. Correct accounting would re-value the option according to its changing intrinsic value due to a changing stock price.
An analogy may make this more clear. A company purchases goods from
another country, payable in that other's currency. The cost booked at
the date of purchase uses the exchange rate at that date. But when the
company eventually receives the goods and pays the bill, the exchange
rate will be different. The final cost to the company is the exchange
rate on the date paid. An adjustment to the booked cost must be made.
One excuse for NOT adjusting the cost of the options is that a
decrease in the stock price would result in a negative compensation
cost. Some think this would be unacceptable, but do not say why. The
stated objective of this type of compensation is for management to
'participate' in the fortunes of the shareholders. When profits decline .... the stock price is expected to fall ....
which decreases the value of the options .... which decreases the
compensation expense .... which decreases the hit to profits for
outside shareholders. That is just what is wanted.
The same system of participation (both positive and negative) is
used to record income taxes. When the company loses money, the
government 'participates' in the loss. A negative tax expense is booked
that reduces the hit to profits for outside shareholders.
The Correct Measurement of Options Compensation
Any measure of the costs of options must measure both the final value of
the options exercised and the change in the market value of options outstanding in the interim. The process is analogous to measuring 'cost of good sold' (Opening Inventory + Purchases - Closing Inventory = Inventory Sold).
Closing Liability - Opening Liability + Exercised = Compensation Expense.
||Diluted EPS Measures Option Dilution... FALSE
You should think of the exercise of stock options as if the option-holders (not the company) force all existing shareholders to give up to the option-holders a percentage of their shares at a price below market value. Diluted EPS measures only a proforma reduction in EPS. It ignores the much larger loss of ownership value.
A better metric for measuring the impact of options dilution
answers the question "What % of future earnings growth (before options
expense) goes to the holders of stock options and not me?" Or put
another way. "What % of reported earnings growth (not earnings) would disappear if
options were measured correctly?" After all, no one invests for earnings. We invest for earnings growth. This metric is equal to
(% options dilution) × (P/E ratio).
E.g. if the options outstanding equal 5% of the issued shares and the P/E = 20, then (5/105×20 =) 95%
of any increase in earnings goes, not to the shareholders, but to the
options holders: a HUGE opportunity cost. If the income was correctly stated most
of the growth would vanish. Without an increase in EPS, the stock price
would not go up, and the options held would not have gained in value.
The emperor has no clothes, and investors must stop giving him their
- The number of options outstanding is 5% of the issued stock.
- The P/E is stable over time at 20.
- The published earnings to start are $1.00/share, so the stock trades at $20.00.
- The exercise price of the options is $20.00, so the liability (intrinsic value per share) for options is $0.
What happens if the earnings increase 10%?
- Published EPS increases $0.10 to $1.10.
- The stock price increases $2.00 from $20.00 to $22.0.
- The options value will now be (22-20 =) $2.00.
- That liability 'per share' is (5 / 105 × $2.00 =) $0.0952 / share.
- If that increase in liabilities were booked as an expense it would cancel out (.0952 / 0.10
=) 95% of the increase in earnings.
But earnings DO increase when companies use options !!! Yes, but not when there are share buybacks to cancel the increased number of shares outstanding. Share buybacks would crystallize that $.0952 un-booked options cost in a real transaction (although not on the Income Statement).
Earnings can increase in spite of options' costs when there are no offsetting share buybacks. This is because the company gains from a share-issue-premium. That gain offsets the cost of options. Remember the diagram from above:
The cost of the options here is the $5 reduction from the $15 market value. But the share-issue-premium of $10 is big enough to more than offset the cost of options. That $5 excess share-issue-premium will boost future earnings. Its existence is not an excuse though, for failing to report the true cost of options. The cost of options and the share-issue-premium are two completely separate and distinct economic events. Shareholders are only indifferent to the issue of additional shares when they sold for market value. Discussed in #4 above
||Dividends Are The Preferred Return... FALSE
(i) Many investors prefer shares that pay dividends because they think
the cold cash 'proves' a company's earnings. This thinking ignores the reality that it is easier to raise cash by selling assets, than it is to earn it. Distributions to shareholders can more easily be a 'return OF equity' than an 'allocation of profits'. During the hay-days of Canadian Income Trusts a large portion of distributions were funded by borrowing money. Balance sheets changed from zero debt at IPO to 50% debt five years later. Dividends do not 'prove' earnings - earnings 'prove' the sustainability of dividends.
(ii) That same belief that dividends are 'cold cash proof' falls down from another perspective. What must the investor do after receiving this cash? He must reinvest it back into the market. He is left in the same position he started from - holding paper assets.
(iii) Retail Investors think the dividend is a kind of 'extra', that their choice is between x percent growth PLUS a dividend, or just the x percent growth. They do not realize that the dividend comes AT THE EXPENSE OF growth. It is a transfer of wealth from the company to the shareholder. While the dividend puts cash in his pocket, the value of his stock holding has fallen equally. The company's ability to grow is impaired. The Fama-French data show how higher dividend yield stocks have lower capital gains (from lower growth).
Some traders believe that buying a stock just before the dividend is paid, and selling on the ex-dividend date generates excess risk-free returns. In theory this won't work because the stock price falls by an amount equal to the dividend paid. In practice it has not been shown to work for other reasons. The market in general can move in the interim. The company can be revalued by the market in the interim. Then consider any taxes that would be owing on the realized dividends, that could have been deferred by selling the stock the day before. Then consider the transaction costs and the bid-ask spreads of less liquid stocks.
A 2012 paper measured the returns for this strategy from the Blue Chip stocks in four countries. Only in Tokyo were there trading profits. In London there were losses.
(iv) It is common to hear the argument supporting a preference for dividends that says "dividends account for 97% of investor returns". This argument deserves a section all of its own below.
(v) Another argument stresses the contribution from dividend income by deconstructing the components of 'real' returns - with inflation removed. The game they play is to subtract inflation from only the capital gains, leaving the dividends intact. (See paper from Research Affiliates). This makes the dividend portion seem much bigger. But why? Certainly owners of debt explicitly consider inflation to eat into the value of interest payments. So why does inflation not eat into dividend income?
Inflation eats into all returns, whether paid as dividends or realized as capital gains. When the total return comes from more than one source it is meaningless to attribute all the effects of inflation to either one or the other. It is only meaningful to subtract it from the total return. But even the academics use this sophistry.
Examples of this belief - Table 4 of Dimson, Marsh and Staunton's paper from which they ... "reinforce the point that the dividend yield has been the dominant factor historically". See also the Market Oracle. See also Exhibit 3 of GMO Montier's article where he subtracts inflation from the growth component - essentially capital gains.
(vi) Investors are skeptical about management's use of retained profits. They think profits will be wasted on executive jets, etc. You can find this belief right back in Benjamin Graham's 1934 classic. But his statement of 'fact' is justified only by 'if it were studied it would be found ...'. In reality companies earn handsome rates of return on those reinvested dollars. The long-term US average is 11%. You check the data on Sheet 31 of the data spreadsheet. It is not hard to pick stocks with returns higher than the average.
Often the pundits claim that it has been proved that companies paying more of their earnings out as dividends have higher earnings growth. They claim this is because management is forced to make every penny count and cannot build empires. This comes from two papers. The papers are discussed fully on the Growth page.
The economy has its ups and downs. The turmoil of the 1930s explains Graham's belief, but times change. Going forward from 2010 you must form your own opinion. Investors' correct preference (reinvest or pay dividends) should be decided by the opportunities each company faces. If it can earn 15% in a world where the investor can only earn 10%, the earnings should be retained. Since investment opportunities vary the dividend paid should also vary.
(vii) Retail investors like dividends because they provide a simple and easy substitute for proper security analysis. They think of the dividend yield as an 'interest rate' equivalent. They are encouraged to ignore the ups/downs of the stock's price with the assurance that in the long run, the stock's price will always go up. But a full year's 4% dividend is easily swamped by a 20% drop in the stock price. The dividend yield says nothing about the certainty of its continued payment. A high yield may represent a good deal, or a disaster in the making. This kind of short-cut analysis is short sighted.
(viii) Some dividend die-hards have now moved the emphasis away from the size of the dividend toward a company's history of dividend growth. This is the metric used to generate lists of "Dividend Aristocrats". Their use of the term 'dividend growth' implies that it is different from what we normally call 'growth' - growth in assets, growth in earnings, capital gains,etc. In fact it is NOT different. Sustainable increases in dividends can only come from growing earnings. Over short terms the company can increase the pay-out ratio (portion of earnings paid as dividends) but it will soon run out of earnings. Over time market valuations normalize so that capital gains follow earnings growth. Growth is growth is growth.
As would be expected, the Dividend Aristocrats have not usually paid high dividend yields. The growth in dividends must come from growth in earnings ..... which must come from reinvested earning ..... which means low dividend yields and/or high returns-on-equity.
Excess returns from Dividend Growth?
(a) (Gwilym, et al.) found in the UK that excess returns were only found when the sorted portfolios were equally weighted. The opposite effect was found when weighted by market capitalization - lower returns from the dividend-growers.
(b) FactSet (not an academic paper) also found that US returns were lower from the top dividend-growers. This contrasts with what is published on the Aristocrats. The official Aristocrat portfolio is equally weighted - so it reflect the returns of small and micro-cap stocks. Investors looking for the stability of large-cap multi-nationals should not expect excess returns.
(c) Gerber ("Dividend-Growth" Vol14, No1,2013) worked with only S&P stocks, isolating those with 10 years of dividend growth, and the safety of payments less than both operating earnings and forward earnings estimates. He found the dividend growers out-performed from 1981 to 2012 but they have under-performed zero-dividend stocks in the time frames since 2003 and since 2008.
(ix) As profits collapsed in 2008 it was common to hear dividends-paying stocks recommended because "you get paid to wait" for the price to eventually recover. But the dividend comes at a price. The company loses liquidity in an illiquid world. The company may be forced to issue additional equity that will permanently dilute your prospects of recovery. The company may miss opportunities to buy-up floundering competitors in an once-in-a-lifetime opportunity. All because investors themselves have locked management into a policy of sustaining the dividend at all costs.
(x) Retired investors claim they need products producing cash flow because they are now living off their investment returns. They claim that growth-investing results in the need to sell stock at in-opportune times - crystallizing losses. They claim that those asset sales permanently impair the portfolio's value.
In reality no investor's portfolio is completely static. It is regularly traded or re-balanced. At each time the investor can choose to NOT reinvest a portion for withdrawal. Most people asset allocate with asset classes that rise in value when stocks decline. Most hold cash balances. To presume that investments must be 'sold' to fund a cost of living is wrong.
The probability is that retired people will keep a much higher balance in their bank account because they do not want to be always worrying about overdrafts, and the regular top-up from the pay cheque is now missing. Withdrawals from the investment account will not be frequent and will be fairly large.
(xi) An argument is made that, when you will be wanting to own dividend stocks in retirement, you must also own them during the accumulation phase. Supposedly the switch at retirement from growth to dividend stocks will trigger capital gains taxes and reduce your portfolio. There are three counter arguments.
1) Except for the rich and those receiving inheritances, the vast majority of Canadians' retirement savings are in tax-sheltered accounts. Taxes are irrelevant.
2) Assuming they are not in tax-shelters, any dividends paid during your working years will be taxed yearly and at your top tax bracket. Capital gains can be deferred.
3) Almost nobody holds individual growth stocks for the 10, 20 or 30 years of accumulation. The normal turnover of securities will trigger capital gain tax gradually over the period, long before the date of retirement. The normal turnover of securities will happen in the dividend portfolio as well, with the same effect.
(xii) Dividends are frequently promoted with the claim that they are taxed at the lowest rates. This may be true for an individual, but is not true as advice without qualifications. When the portfolio is inside a Canadian RRSP, or RESP, or a TFSA taxes have no impact because there is no tax paid. When the investor is caught by Canadian Minimum Tax all income is treated the same. Also, at the top marginal tax bracket dividends are taxed at higher rates than capital gains. (See current marginal tax rates on different types of income.) Also, the effective tax rate paid on capital gains depends on the holding period of the asset. The longer the delay in being taxed, the lower the effective tax rate. See the sheet on the spreadsheet link above.
(xiii) Stock price volatility is definitely higher in stocks paying no dividends at all. When comparing between stocks paying different yields, evidence from Fama-French data shows that higher yields do NOT result in lower price volatility. And no, this is not because the highest excess returns come from the NEXT to highest dividend-yielders so you would expect the NEXT to highest to have the lowest volatility. The data shows the lowest volatility in the NEXT to lowest-dividend-yielders (Clemens, 2012, Exhibit 7). So the volatility difference may be more to do with the type of company than dividend policy.
(xiv) Many believe that a steady $$dividend in a period of stock price volatility, allows the reinvested dividend to purchase more shares when the stock is down, and less shares when the stock is high, producing extra returns from a dollar-cost-averaging effect. There are dividend investors who gloat about the 'lost decade' of 2000-2010, convinced they gained from the low prices.
The box below compares the outcomes when dividends are received from a stock with a steady price, to one whose market price drops in half before the dividend, only to recover 100% after. You can see that yes, more shares can be purchased because of the drop in market price. But the dividend payment has double the proportional impact on that smaller share price. The two effects equal and offset each other. Both situations end up with the same value.
|Steady Price||Variable Price|
|Total Position Value|| $ 100|| $ 100|
|# Shares Owned||1||1|
|Price Drops in Half|
|Stock Price B4 Dividend||$ 100||$ 50|
|Dividend Paid||$ 2||$ 2|
|Stock Price After Div||$ 98 (= 100-2)||$ 48 (= 50-2) |
|Drop in Price ||2%||4% |
|# Shares Owned||1.0204||1.0417|
|Total Position Value|| $ 100 (= 1.0204*98)|| $ 50 (= 1.0417*48)|
|Total Position Value||$ 100|| $ 100 (= 1.0417*96)|
(xv) Back-testing supposedly shows higher returns from higher yield stocks (except for the very highest decile yield) (Clemens, 2012, Exhibit 8). The Fama-French data set generates the following graph of returns broken down by decades. There is no generalization that holds true. If you open the Dividend Returns spreadsheet you can rotate the image to see through to all the decades yourself.
In the eight decades analyzed, the highest returns were found in the lowest-yield-deciles (0-1-2) in 4 of the decades. The highest returns were found in the highest-yield-deciles (8-9-10) in only 2 decades. The middle deciles (3-4-5-6-7) had the highest returns for the remaining 2 decades.
Before accepting back-testing at face value read the more nuanced understanding on the Screening page. You can deconstruct the excess returns supposedly from dividends into a combination of excess returns to low P/E stocks and excess returns to LOW dividend stocks when P/Es are in the average range.
(xvi) Some believe that dividends allow 'compounding to work faster'. They think that capital gains do not compound until the position is sold. They think that the math of compounding does not work with unrealized paper profits.
This idea must come from the mistaken thought that your portfolio (and the stocks in it) are valued at their purchase price. IF that were true then the portfolio's reported value would not change until a sale transaction triggers a re-valuation. But all that is not true.
Portfolios, the stocks within them, and even stock indexes, are constantly marked-to-market value. The stocks' value changes while the position is held. There is no change in value when the asset is sold. The stock is exchanged for equal cash.
The measurement of returns treats realized and unrealized profits exactly the same. Any price increase in Yr1 is counted as the profit of Yr1. That profit 'compounds' by being added to the original cost to form the new base for the calculation of Yr2's profits. E.g. A stock is bought for $100. It increases in value to $110 by the end of Yr1. That provides a 10% rate of return ((110 - 100) / 100). The $10 profit is considered to be 'reinvested' or 'compound' into the new $110 base for Yr2. To earn another 10% return the stock must rise in value by $11 to $121 ((121 - 110) / 110).
You can see the result of this compounding of unrealized price gains in the long-term charts of stock indexes. They look like hockey sticks because the yearly $gains become larger and larger, even though the %return is steady, because each year's gains 'compound'.
Inside the company, retained profits compound perpetually. E.g. each day 100 widgets costing $5 are sold for $8. The $300 profit is reinvested to buy 160 replacement widgets. The next day's profits are 160 widgets * $3 = $480. Etc.
A company has a choice of four ways to use earnings. A simple
example (following) compares the investor's returns under each option. For the two
options where investment remains in the business, the investor realizes
twice the return (20%, equal to the company's ROE), compared to 10% when profits are used to pay
dividends or buy back shares (20% ROE divided by P/Bk 2).
You many be tempted to argue that (B) Receiving a dividend and then using it to purchase more shares would be the same as (C) DRIPS. The distinction is that DRIPs increase the company's equity by reinvesting inside the business. Whereas purchasing more shares from the secondary market has no effect on the company itself. A dollar inside a business earning 20% ROE is worth twice as much as a dollar held by investors earning only 10%.
DRIPs put the investor into the same position as if the company had kept the cash and reinvested earnings (D) without ever paying a dividend. The cash does a circuitous route out to the investor (triggering his personal tax) and then back into the business. Of course there were transaction costs along the way. Reinvesting profits without the pretense is more efficient. Pundits promoting dividends as well as DRIPs are talking through their hat. They are the exact opposite of each other.
||Dividends Do Have Valid Benefits
Of course there are some good points in favour of dividends.
- For those who Asset Allocate and re-balance faithfully, the cash provided by dividends prevents you having to sell assets for this purpose. The transaction costs and taxes triggered by sales are saved.
- Many Canadian testamentary trusts are set up to distribute income to a surviving spouse during his/her lifetime. On their death the next generation is given the residual principal of the trust. If the trust earns no income (dividends or interest) there is nothing to distribute. So trusts are correctly heavily weighted with high dividend paying stocks.
- Canadians not in the top tax brackets, get taxed on dividend income at a lower rate than capital gains. See the spreadsheet comparing the tax brackets.
- The requirement to pay dividends keeps management's attitude correct. They would love to simply take the cash from common share issues and then see the back of those 'owners'. Dividends keep it clear that management works for those owners, that those owners demand a return on their investment. The replacement of dividends with share buy-backs removes that attitude correction. Although technically equivalent, they are not transparent, or considered an ongoing commitment.
- Dividends have a great emotional wallop for the investor. Getting the cash just feels good. When in retirement and withdrawing cash, it is emotionally easier to withdraw cash from dividends than from any purposeful sale of shares. Many people emotionally 'feel' that asset sales deplete capital while dividends leave capital intact.
- Healthy dividends prevent some of the games management play using equity for compensation. Option holders do not receive any dividends. And dividends reduce the capital gains on options. (Of course the compensation can be structured to come from dividends instead, or to increase in offset to any dividends.)
- There is no question that hefty dividends support share prices when the economy tanks and profits fall. When cash is being withdrawn in retirement the 'sequence of returns' affect is more deadly the farther a stock falls, so a non-volatile stock price reduces the damage. As well the dividend $$ itself helps fund the withdrawal - reducing the need to liquidate securities at their lows. The second sheet of the Dividend Vs Growth Portfolio spreadsheet where cash draws are modeled seems to show this protection from sequence of returns risk.
- It is during market corrections that stronger businesses use the opportunity to purchase weaker competitors. They must pay with cash or debt or shares. The dividend-paying business will not have the cash, but its share price may be relatively more strong because it was propped up by those dividends. As shown throughout this website, shareholders benefit from new share issues for good ROE opportunities when the stock is trading at multiples of book value.
- Dividends to common shareholders provides a cushion for preferred shareholders. Most often the common's dividends must be cut before management can cut the preferred's dividends. This ensures public notice of the cut along with the ensuing loss of management credibility. It also ensures that the people with votes are truly pissed off before the preferred owners get their dividend cut.
- Short-sellers of the stock must make good on all dividends issued by the shares they sold. This cost adds considerably to the cost and risk of short-selling. Keeping short-sellers at bay may be considered either a good thing or a bad thing.
||Yield On Cost (YOC) is a Useful Metric ...FALSE
This section has been moved to the page on Valuation Metrics. Sorry for the inconvenience.
||Reinvested Dividends Account For 97% of Your Returns ...FALSE
Advising clients to buy high dividend stocks may be good advice, but
advisors are lying when they quote the 97% number (or even the 60% number). They should find
honest reasons to support their advice.
What Is Relevant
If the current dividend yield on your portfolio is (say) 3% and you demand a 10% return for investing in risky stocks, then 30% of your expected return will come from dividends - 3% as a portion of 10%. This does NOT change the longer you hold your portfolio. Each year is calculated anew with the current dividend yield. And history is just the sum of all individual years.
If you choose to own high dividend stocks, the portion of your expected return from dividends will be higher than for the person who chooses low dividend stocks. Neither choice necessarily influences their total returns. During a market pull-back the small return means a greater portion of the return will come from dividends than during a fast growing market with high returns.
If that common sense argument does not convince you, consider the implications of the claim for retired investors. If it is true (that reinvested dividends account for 97% of your return) then retired people who are spending their dividends (and not reinvesting) must be earning only 3% of their potential. Garbage. Investors earn the same rate of return regardless what they do with their dividends, regardless what they do with their capital gains for that matter.
What Is Irrelevant
Historical dividend yields are irrelevant - along with their ratio to total returns. There are two long-term trends that will not revert to any mean.
1) The use of stock options for compensation has permanently changed management's motivation to pay dividends.
2) Stock buy-backs have replaced dividends because of their flexibility and because analysts now treat them as equivalent to dividends.
3) Multi-national companies now generate a large portion of their profits in foreign countries. Before dividends can be paid these profits must be repatriated. There is a dis-incentive to do so when there are additional taxes (US) due on that transaction.
Times have changed. What happened in the past is in the past.
The following graph compares yields vs capital gains as they have changed over time. It uses 5-year averaged returns in order to smooth the inherent volatility of capital gains and better show the relationship to dividends. From 1957 to 2009 the TSX return from dividends has equaled 37% of the total return. There is no cyclicality showing here to cause you to conclude that yields are mean-reverting. Instead what is shown is that during periods of high price appreciation (1980s and late 1990s) yields fall. This is basic math. When the denominator increases Div/P decreases, unless dividend $payments increase as the same rate .... which they have not.
Even these statistics overstate the impact of dividends. The measurement of dividend returns used includes both 1) the actual dividend payments from the index, plus 2) the capital gains earned from reinvesting those dividends that were paid earlier in the year. See the math at this spreadsheet.
The advisors who quote the statistic ("dividends provide x% of total returns") always present it as a reason to prefer dividend-paying stocks, implying higher returns will result. But there is no logical link between the two issues - no matter what correct percentage is used. Some of the investors in that historical period owned growth stocks and earned the benchmark return without any dividends. Other investors in that historical period owned high dividend stocks and earned the benchmark return with a much larger percentage coming from dividends.
The experts who have changed their argument from "97% of total returns" to "40% of total returns" still present the fact as a reason to prefer owning high dividend stocks. See this example from James Montier. But if dividends earn only 40% then capital gains must earn 60%. Since 60% is bigger than 40% why are the experts not promoting growth stocks instead? None of this argument makes sense. Note that Montier did not change to the 40% number because he realized Siegel's graph was wrong. He used Siegel's exact logic but shortened the time-frame so that his 40% did not have sufficient time to become Siegel's 97%.
What Is Flat Out Wrong
The 97% claim comes from a book by Jeremy Siegel ("Future for Investors" pg 126). The quote is "From 1871 through 2003, 97 percent of the total after-inflation accumulation from stocks comes from reinvested dividends. Only 3 percent comes from capital gains." He compared the portfolio value resulting from $1 invested in 1871 in a 'total return' index to the same $1 invested in a portfolio whose dividends were removed (what we are familiar with as a 'normal' index).
Over 133 years the spread between the two widens until by 2003 the portfolio without dividends is worth only 3% of the value of the 'total return' portfolio. He presented a graph looking like the one following and concluded that the area between the two lines was the return from dividends.
1: The original claim is being mis-quoted by the media and advisors. See again how Montier talks about "43 percent of S&P returns". Siegel did NOT say "97% of your returns". He said "97% of accumulation". He was NOT talking about 97% of the rate of return. He was talking about 97% of the resulting portfolio value.
2: Accepting what Siegel actually said, was his conclusion correct? No.
3: The correct interpretation of his graph is ..... given three portfolios invested in exactly the same assets, earning the exact same rate of return and type of return, the portfolio with principal added each year (the spread funded with yearly dividends) will grow faster than the portfolio without (the Total Return). A third portfolio with principal removed each year (the lower line with dividends removed yearly) will grow slower than both of those.
You can chart the exact opposite logic used by Siegel. The value of the Total Return Index would remain the same. This is the value of a normal growing portfolio. But instead of the lower line measuring that portfolio with dividends removed, now the lower line measures the portfolio with capital gains removed each year.
This graph shows reinvested capital gains accounted for 99% of the accumulated portfolio. Is it logical to claim that 96% of Total Returns come from dividends, AND 99% of Total Returns come from capital gains? Of course not. Neither graph measures the income split attributed to dividends or capital gains. The only thing both graphs show is the importance of reinvesting profits - from whatever source.
Look at Box (C) in the spreadsheet (Excel). What Siegel labeled as the value attributed to dividends is just the mathematical difference between the values of the two indexes (column E). But the same value can be obtained by considering it to be a stand alone portfolio (column F). Put the cursor on the spreadsheet's cells to see that the calculation used increases the portfolio each year by the total return earned by its own investments (mostly capital gains plus any dividends), PLUS the capital infusion equal to the dividends of the S&P index.
Another Way To Explain It
Start by simplifying the issue. First, reverse Siegel's decision to measure portfolios 'after-inflation'. Inflation does not change the essence of any conclusions. Second, present the graph with a log scale so the hockey-stick effect is hidden and the portfolios' growth rates are represented by the lines' slopes. Third, measure 1926 to 2007 because that is the period for which there are public numbers.
Now interpret the resulting graph above. The top line measures the Total Return Index, reflecting a portfolio where no capital is added or removed. The profits compound untouched. The bottom line measures the normal Index, reflecting a portfolio invested in exactly the same stocks, and earning the exact same rate of return. Its growth rate (slope) is lower than its rate of return because capital is removed each year equal to the dividends earned. The difference in the portfolios' growth rates (slope) equals the 4.2% return from dividends (= 10.2% - 6%). Over this period dividends provided 41% of the total return (= 4.2 divided by 10.2).
Conclusion: The size of the Difference area on Siegel's graph has nothing to do with 'returns from reinvested dividends'. Nor has it anything to do with 'accumulations from reinvested dividends'. It grows mainly from its own profits (dividends and capital gains), because the math presumes the removed dividends are themselves invested to earn the same returns as the source portfolio. It makes no difference what dividend yield is paid by the Index portfolio. Even a paltry 0.5% dividend yield would result in a graph looking essentially the same. It makes no difference to the spread if funded with dividends or capital gains, or interest, or any other type of income. In all cases the spread will eventually become 99.99% of the Total Return portfolio's value because its slope is steeper from capital injections, not due to the magic of dividends.
You must be aware that even academics can be full of hot air. Retail investors should smell a rat when they read the nonsensical statements like the following from Dimson, Marsh and Staunton, based on the same idiotic interpretation of Siegel's graph. "Dividend income adds a relatively modest amount to each year's gain or loss. But while year-to-year performance is driven by capital appreciation, long-run returns are heavily influenced by reinvested dividends. .... The longer the investment horizon, the more important is dividend income. ... Capital appreciation dwindles greatly in significance [over time]".
Say what? How exactly do they think this magical transformation occurs? It is common sense that long-term results are only the sum total of each year's results. What is consistently true for each year will be true over long time periods.
This idea that any one year's capital gain turns into dividend income if you just look back at it from a long-enough time span, gets trotted out by many people. See Exhibit 2 of the GMO Montier article. Lies, damn lies and statistics."