The Dividend Capture Strategy is an investment strategy aimed at realizing a profit from dividends, while holding the underlying stock as short as possible.
It might also be named “Dividend Hopping” as investors with this strategy may purchase several stock in a short sequence in order to obtain each respective dividend without holding on to the stock.
The Dividend Capture Strategy is questionable because the market tends to reflect the dividend in stock prices. The moment the stock owner becomes entitled to receiving the dividend, the underlying stock will lose value roughly in line with the dividend, or the after-tax value of the dividend associated with the stock. Loss in stock price and dividend income may offset each other, but may be taxed differently, leading to a potential after tax loss to the investor. And last not least, transaction cost may outweigh profits.
Real life performance of the stock however might not be a defined loss, but rather an unrelated random volatility, up or down, which can be greater than the dividend payout. This would make such a strategy highly speculative.
As Investopedia puts it: If this were profitable, computer-driven investment strategies would have already exploited this opportunity.
Nevertheless, to the extent that markets are not efficient, a profit is theoretically possible. Also a hedge against stock price decline might offer benefits where possible – while increasing transaction cost.
Exploring the Dividend Capture strategy.
- Declaration Date: Date on which the company publishes the amount of dividend paid out, can be weeks before Ex-Dividend
- Ex-Dividend Date: Date on which stock will start to trade without dividend entitlement
- Date of Record: Date on which ownership of stock will be recorded for dividend entitlement, usually the day after Ex-Dividend
- Pay Date: Date on which the dividend will be paid out, can be several weeks after the Date of Record
Minimum Holding Period for Dividend Capture
For how long do I need to hold stock in order to receive the dividend?
One night. Purchase before business close, the day before ex-dividend date. Sell after business open of ex-dividend date. See these articles:
- Dividend Capture Strategy (borntosell).
- Selling Shares before the Ex-Dividend date (Investopedia)
- How and When are Stock Dividends paid out (Investopedia)
- What’s the difference between the Record Date and Ex-Dividend Date (Investopedia)
- Confusingly, the image showing a two day period is shown here, too.
Two days, I conclude from these articles, but I think they are not clearly written, so it is actually one night nevertheless. The “confusing” texts are these:
- How to use the Dividend Capture Strategy (Investopedia)
- Critical comment: This article seems inconsistent or unprecise. The graphics in the article show a timeframe of two days. The text however says “The underlying stock could sometimes be held for only a single day.” Holding for the day seems irrelevant, holding for the night is most certainly decisive. Holding requirement is not a defined time period during the day but rather a point in time, e.g. business start, of either Ex-Dividend date or Record Date – the article doesn’t clarify which one. To the contrary it says “then selling it immediately after the dividend is paid. ” Payment date is certainly irrelevant, as it occors weeks after recording who is entitled.
- Ex-Dividend Date vs. Date of Record, what’s the difference? (Investopedia)
- Authored by “Investopedia Staff”. Defines that stock mus be purchased prior to Ex-Dividend date, and held on Record Date.
- “The ex-date or ex-dividend date is the trading date on (and after) which the dividend is not owed to a new buyer of the stock. The ex-date is one business day before the date of record.”
- “The date of record is the day on which the company checks its records to identify shareholders of the company. An investor must be listed on that date to be eligible for a dividend payout.” And: “The date of record is the date in which the company identifies all of its current stockholders, and therefore everyone who is eligible to receive the dividend. If you’re not on the list, you don’t get the dividend.” This sounds like stock must be held on that day, too. “Current stockholders” and “Listed on that date” are difficult to interpret differently.
- Afterwards it seemingly contradicts itself in the example: “If you want to sell the stock and still receive the dividend, you need to sell on or after [Ex-Dividend Date]”
- The only way to reconcile the statements would be that the on the Date of Record, the company will list those who owned stock on start of trading of Ex-Dividend Date – which would then mean that the minimum holding period is actually one night. The fact that the company checks the records a day later does not mean stock still has to be held while the company performs the verification.
- Authored by “Investopedia Staff”. Defines that stock mus be purchased prior to Ex-Dividend date, and held on Record Date.
If it is two days, how long is “two days” really? At what time do I have to buy stock and at what time am I allowed to sell stock on the respective days? Stock opening? Stock closing? Noon?
According to borntosell, link above, the investor must buy the stock by close of business before Ex-Dividend date. The stock may be sold right at business start on Ex-Dividend date, and the owner of the previous night will still be entitled to the dividend. It’s the most precise description of all the articles quoted. It’s also the oldest one, dated 2013.
According to Brandon Opre, from above cited link on Investopedia: “So if you want the dividend, you need to be an owner the day before the ex-dividend date.” This implies that the stock may be sold right on business start on Ex-Dividend date as well. Unfortunately it is not explicitly stated.
If the time is really was two days, then who gets the dividend when selling stock on the Ex-Dividend date, at any time during the day? Nobody? Does the company then save the dividend on anything traded on Ex-Dividend date? It would not make sense to me.
I conclude from the above that the real minimum holding period to be entitled to the dividend is one night. From the point of view of the stock market, it is a logical point in time, since between market close and market opening the following day, no trading time has passed.
The Ex-Dividend date for German Stock is the day after the shareholder conference, so stock must be purchased on the conference day or before to be eligible.
Past performance of stocks around Ex-Dividend date.
In an efficient market, the stock price should drop by the amount of after-tax dividend on the night before the Ex-Dividend date. An observation will be made difficult by general volatility of the stock. Markets react overnight to all kinds of news. In order to check the statement in real life, extensive statistical examination would be required. For a rough check, maybe some isolated examples of stock through the night before Ex-Dividend, compared to similar stock which does not have the same dividend date, could be insightful.
Let’s start with just one random example of a stock price change when it turns ex dividend.
Hormel Foods HRL had its last Ex-Dividend date on 10 July 2020. Dividend payout is quarterly. The amount of the dividend was 0.23 $. How did the stock perform during the night before Ex-Dividend date?
09.07.2020, 21:30 CET: 47,010 $ (CET is 6 hours before Eastern Daylight Time, so this would be 15:30 local time NYSE, 30 mins to close.)
10.07.2020, 15:30 CET: 47,530 $ (CET is 6 hours before Eastern Daylight Time, so this would be 9:30 local time NYSE, market open.)
This example shows no downward impact of Ex-Dividend date, any impact is likely superimposed by other greater factors.
Let’s pick an example of yearly dividend. Yearly dividend, being larger, could be more likely to have visible impact.
China Life Insurance Company LFC had its last Ex-Dividend date on 06 July 2020, yearly payout. The amount was 0,5134 $. How did this stsock perform during the night before Ex-Dividend date?
02.07.2020, 21:30 CET: 10,910 $ (CET is 6 hours before Eastern Daylight Time, so this would be 15:30 local time NYSE, 30 mins to close. July 3 market was closed for July 4 holiday, 4 and 5 weekend)
06.07.2020, 15:30 CET: 12,610 $ (CET is 6 hours before Eastern Daylight Time, so this would be 9:30 local time NYSE, market open.)
Once again, no visible downward impact of Ex-Dividen date. Several large upward jumps were observed around the time of dividend payout. It probably is a terrible example. Lets pick a period further in the past and a different player.
Quarterly dividend for Ford Motor Company, Ex-Dividend January 29, 2020. Dividend is 0,15 $.
28.01.2020, 21:30 CET: 8,970 $ (CET is 6 hours before Eastern Daylight Time, so this would be 15:30 local time NYSE, 30 mins to close.)
29.01.2020, 15:30 CET: 8,915 $ (CET is 6 hours before Eastern Daylight Time, so this would be 9:30 local time NYSE, market open.)
Now here we have downward movement. The downshift exceeds the dividend payout by a factor of 3,7 (0,55 $) so this is probably also superimposed by other factors.
I’d like to compare this to the performance of General Motors on the same night.
28.01.2020, 21:30 CET: 33,640 $ (CET is 6 hours before Eastern Daylight Time, so this would be 15:30 local time NYSE, 30 mins to close.)
29.01.2020, 15:30 CET: 33,755 $ (CET is 6 hours before Eastern Daylight Time, so this would be 9:30 local time NYSE, market open.)
Direct competitor General Motors went up during the same night. They had no dividend impact (dividends pad in December and March). So this example in principle shows downward movement for a stock that has dividend payout and upward for a direct competitor. However given that general volatility on both stock exceeds any dividend payout, and previous examples have shown no correlation, I don’t buy into the dividend payout being a driver for stock price change as observed. I believe the dividend impact in these few individual randomly picked cases was a minor part in what ever stock price change has occurred.
This leads to one important conclusion for me:
General volatiliy of underlying stock is a significant risk factor for the Dividend Capture Strategy.
How can general volatility be mitigated for a short term like one night?
Short term hedge for the stock market price
Hedging against such volatility could be achieved with a number of constructions.
One example would be the one put on investopedia, under the title of Dividend Arbitrage:
“To illustrate how dividend arbitrage works, imagine that stock XYZABC is currently trading at $50 per share and is paying a $2 dividend in one week’s time. A put option with an expiry of three weeks from now and a strike price of $60 is selling for $11. A trader wishing to structure a dividend arbitrage can purchase one contract for $1,100 and 100 shares for $5,000, for a total cost of $6,100. In one week’s time, the trader will collect the $200 in dividends and the put option to sell the stock for $6,000. The total earned from the dividend and stock sale is $6,200, for a profit of $100 before fees and taxes.”
The example consists in purchasing stock and a put option at the same time. The put option makes sure the sales price is known in advance, eliminating volatility by executing the option.
In this example, the investor found a hedge which cost im only 1$ per share to buy. (Put option costing 1 $ above internal value). Critics will be quick to point out that this is utterly unrealistic, in particular when the option is traded after the dividend is announced – the option will price in this change. So the theory is great, practice may not be so great.
But apart from that, it is good to learn that a hedge could be made by purchasing the stock and a put option, and excercising the put.
I’d like to learn whether a stock can be hedged using warrants. Other than a real option, a warrant does not entitle you to buy or sell stock, but it grants you cash settlement by the rules stipulated in the warrant.
Spoiler: The answer, as found below is: The ask/buy spread of warrants alone can be greater than the stock price change, rendering this excercise useless. Anyways, I’ve written it down, learned from it and will show the analysis below.
Warrants – Optionsscheine
The owner of a warrant receives the right, but not the obligation, to buy or sell a certain base value at a specified price, before the expiry date of the warrant. Usually, this right is executed by payment, not by delivery. This means, upon execution of the right, there is no sale or purchase of actual shares, but the difference between the current stock price and the face value of the warrant is being paid. This differentiates a warrant from an option. (Option vs. Optionsschein).
By executing a call warrant, the owner of the call will receive: Underlying share price – Strike price. If the call strike price is lower than the current share price, the owner receives money. The warrant is “in the money”. If the call strike price is above the current share price, it makes no sense to execute the call. It is considered “out of the money”.
By executing a put warrant, the owner of the put will receive: Strike price – underlying share price. If the put strike price is lower than the current share price, it makes not sense to execute the put. If the put price is higher than the current share price, the owner receives the difference. The warrant is “in the money”.
A warrant costs money to buy, and it can be sold, similar to the underlying stock itself. The price of the warrant will change in line with the change of its underlying base value.
The price of a call warrant is going to follow the price of the underlying stock. If the stock goes up, so goes the price of the call warrant. The price of a put warrant is inverted to the price of the underlying stock. If the stock goes up, the put warrant goes down.
Therefore, an investor cannot hedge volatility by buying stock and call warrants. However it is possible to hedge volatility by purchasing stock and its correlating put warrant in parallel.
How to define which warrant to buy and how many of them?
- Bid/Ask spread
Often warrants have a Ratio of 0.1. This means, to compensate 1 share, 10 warrants are required.
Delta indicates the correlation of warrant price and underlying stock price. Put warrants have a correlation between 0 and -1, call warrants have a correlation between 0 and 1. Delta indicates how much a change in stock price will be reflected in the change of the warrant price
The amount of put warrants required to compensate volatility of a stock to be purchased can be indicated as follows:
For example, in order to compensate 1 share of Apple with put warrants which have a Delta of -0,30 and a ratio of 0,1, the amount of warrants required is 33.33, so the investor may want to buy 1 share and 33 warrants, or 10 shares and 333 warrants and so on.
In order to compensate volatility, it is useful to choose a warrant deep in the money, and far in the future. They have the highest delta (or lowest, in case of put options where delta is negative.) High delta means less warrants required. Expiry dates far in the future also mean higher delta, and less loss of time value over time.
The formula also shows that in order to compensate the volatility of stock with call warrants, a negative number of warrants must be purchased, i.e. the warrants should be sold, not bought.
This compensation comes at a cost. No hedge is free. The cost of this type of compensation is the Bid/Ask spread, plus the transaction cost for buying and selling the warrant.
It must also be added that not every stock on the stock market has an easy to find and readily available call or put warrant available. Large cap stock has a higher probability for the availability of derivatives.
In order to add some more numbers to the academic formula above, lets get a real life quote and watch it over time.
Example: Deutsche Post, 27.07.2020, 14:45 CET, 34,85 € bid 34,86 € ask stock price
Highest negative delta put warrant: WKN TR2JES
Ratio 0,1; Strike price 36€ (in the money); expiry 16.12.2020, Delta -0,63, Purchase price 0,36 € (ASK), sales price 0,35 € (BID)
If I want to own 1000 EUR worth of Deutsche Post stock, I have to buy 29 shares.
To compensate 29 shares, I have to buy 460 put warrants.
Stock price: 29 * 34,85 = 1009,20 €
Warrant price: 460 * 0,36 = 165,60 €
Total cost: 1174,80 excluding transaction cost or taxes.
Tomorrow this time we could verify again. Or, even better, we now look up how this has performed vs. a change in the past.
01.07.2020, the same warrant came at a cost of 0,51 EUR. The underlying value of Deutsche Post has been trading at 29,20 EUR.
Stock price: 29 * 29,20 = 846,80 €
Warrant price: 460 * 0,51 = 234,60 €
Total cost: 1081,40 excluding transaction cost or taxes.
This means, this hedge would have been imperfect, even without transaction cost – however: It is difficult for me to verify the delta value which has been shown on the 1st of July. In order to compensate successfully, a higher number of warrants would have done the trick.
Warrant price: 643 * 0,51 = 328 € plus Stock price 846,80 € leads to a total cost of 1174,80. The number of warrants 643 would have been purchased if the delta on that day had been -0,451. Whether or not the delta has been there I don’t know. This is all hypothesis and assumptions.
Let’s add more examples where a put warrant is available on HSBC.
320,00 € bid 320,10 ask stock price 27.07.2020, 15:20h
Put warrant TT0HBH: Ratio 0,1, Delta -0,36, Bid 4,03, Ask 4,08
13,86 € bid 13,87 ask stock price 27.07.2020 15:24
Put warrant TR88NE: Ratio 0,1, Delta -0,98, Bid 1,47, Ask 1,49
Put warrant TR88NB: Ratio 0,1, Delta -0,99, Bid 1,14, Ask 1,16
Put warrant TR88NA: Ratio 0,1, Delta -0,97, Bid 0,84, Ask 0,86
141,14 € bid 141,18 ask stock price 27.07.2020, 15:28h
Put warrant TR9M7C: Ratio 0,1, Delta -0,83, Bid 5,49, Ask 5,54
Put warrant TR9T9L: Ratio 0,1, Delta -0,79, Buid 5,58 Ask 5,62
Put warrant TR9T9Q: Ratio 0,1, Delta -0,82, Bid 7,29, Ask 7,34
I shall continue posting tomorrow. A one day change is the relevant mechanism, so such a one-night study may hold a lot of insight.
New day new luck, this is how the same stock and warrants look right now:
35,08 € bid 35,09 ask stock price 28.07.2020, 10:07h
Put warrant WKN TR2JES: Ratio 0,1, Delta -0,62, Bid 0,33, Ask 0,34
141,78 € bid 141,82 ask stock price 28.07.2020, 10:12h
Put warrant TR9M7C: Ratio 0,1, Delta -0,83, Bid 5,45, Ask 5,49
Put warrant TR9T9L: Ratio 0,1, Delta -0,79, Buid 5,54 Ask 5,58
Put warrant TR9T9Q: Ratio 0,1, Delta -0,82, Bid 7,25, Ask 7,30
324,70 € bid 324,75 ask stock price 28.07.2020, 10:42h
Put warrant TT0HBH: Ratio 0,1, Delta -0,36, Bid 3,84, Ask 3,89
13,91 € bid stock price 28.07.2020 10:18
Put warrant TR88NE: Ratio 0,1, Delta -0,98, Bid 1,46, Ask 1,48
Put warrant TR88NB: Ratio 0,1, Delta -0,99, Bid 1,13, Ask 1,15
Put warrant TR88NA: Ratio 0,1, Delta -0,97, Bid 0,83, Ask 0,85
What strikes the eye first:
The absolute price change of warrants is often smaller than the guaranteed loss from ask/buy spread. This means, the hedge is nothing but a loss, guaranteed.
Ask/Bid spread of 0,02 € on a warrant which costs 0,83 € represents 2,4% loss before the trading even started. This is an order of magnitude which may well be greater than any after tax dividend payout.
Without further analysis, the intention of hedging a small stock price change with warrants can be discarded.
The numbers show for example: Purchasing 31 Apple shares on 27.07. would have cost 9.942,10€ (assuming a trade fee of 1€ as charged by Trade Republic). To try to hedge that, I asssumed a purchase of 861 put warrants. This should reflect the full invest based on ratio and delta. The warrants would cost 3.513,88€. Already on the day of purchase, the spread loss would be 1,28%. This is due to 0,04€ spread between bid and ask prices, multiplied by 861 warrants, and the assumed trade fee of 2€ (buy+sell). So purely due to this, the hedge will definitely cost 45,05 € without having delivered anything.
One day later, the shares have risen. By consequence, the options have fallen. This is expected. However there are two issues: a) as mentioned above, the hedge costs money due to the spread, b) the hedge has overcompensated the gain of the base value. It has fallen more than the apple stock has risen. Sales prices are, again reflecting a 1€ assumed trading fee: 10.064,70 for the shares and 3.305,24€ for the options. Overall result is a loss of 68,04€, even though the shares actually rose by 140,60€.
Components of the loss are the spread loss of options of about 1.3% in this case, plus the overcompensation of the hedge of about 12%.
I assume the Delta which is reported for the option is not a value with absolute truth in it, or I lack understanding for other implications which affect warrant price changes.
For the other examples: Deutsche Post, Volkswagen show between 37€ and 48€ guaranteed loss on the warrant position based on 10k€ invest in their base value. Shell even shows 150€ guaranteed loss on the warrant position. It may have to do with the lever. Small lever means higher invest in options and higher absolute loss on their spread.
Even disregarding the spread loss, the hedging attempt was not reliable. Deviation of the hedge was:
Deutsche Post 31% overcompensated
Volkswagen 6%, 27%, 31% undercompensated respectively for the three different warrants
Apple 12% overcompensated
Shell 50-52% overcompensated on all three different warrants.
The problem may be that the minimum increment for the change in value of warrants is 0.01€, one cent. One cent already causes a large change in value due to the high number of warrants required. The minimum increment leads to very poor accuracy of such a hedging attempt.
The case for Shell for example showed roughly double the change, the other possibility would have showed zero change. With an option value so low in absolute price, the minimum change can already swing the picture altogether. And let’s not forget that the spread loss is already twice as large at the desired hedge behaviour.
So the Shell case is not representative due to the change being the minimum increment.
Deutsche Post: Change was 2 cent. Deviation is smaller than minimum increment.
Volkswagen: Change was 4-5 cent. Actually, one of the options could have performed better: TR9T9Q. Minimum increment is slightly smaller than the deviation. Well, as a 1 cent increment would have brought this one close to perfection, it is plausible that the few minutes delay in recording were just those where the value switched.
Warrants with a high lever have a high impact of minimum increments, as the absolute price is low. Warrants with a low lever have a high impact of spread, as the required volume for the hedging attempt is high. The minimum increment may additionally impact effectivity.
Dividend capture strategy cannot be hedged with put warrants. Unhedged, it is speculative, as the base volatility may outweigh dividends. Also, in an efficient market, a profitable dividend capture should not be possible.
This does not mean that dividend capture should not be done. In a bullish market, it may be a nice “add-on” to grab dividend, additional upward potential for a trade. In that sense, it might be a bonus to other speculation, assuming that some inefficiency in the market can be skimmed, without being the profit driver in itself.
Search may continue for other hedging mechanisms, e.g. real put options or call writing, but there is little optimism. Cost may outweigh benefits, and call writing may be a professional option not available to retail broker accounts.