Dividends vs. Share Buybacks? A Halfords Case Study

I found an interesting alternative to buybacks today, thanks to TMF. I like CLS's novel way of not buying its own shares in the market, but buying them back from its shareholders. Sharescope does give the yield but does give the P/E, which is 4.9. The ticker is CLI. Price yesterday £6.23.

The original article was prepared for my share club, this edited version is for stockopedia because it offers a simple political and financial solution to the share buyback problem. A line of text in the budget perhaps, saying; "Share buybacks will be illegal unless shareholders interests are given priority. To use share buybacks, the company must first offer the buyback to its own shareholders before it makes the offer to the market.

This wording needs a lot of improvement, I will be happy to work on it if my idea has merit.

Here is the link to the detailed article;

http://www.fool.co.uk/news/investing/2012/02/02/40-discount-no-dividend-but-you-still-get-an-incom.aspx?source=ufwflwlnk0000001

MadDutch

There's an interesting article on efinancialnews today in which an EC-backed group of economists calls for an outright ban on stock buybacks.

http://www.efinancialnews.com/story/2012-02-03/eu-economists-call-for-ban-on-share-buybacks?ref=email_37191

Finnov, a research collaboration between seven European academic institutions including the UK’s University of Cambridge, said share buybacks should be subject to stronger governance control or banned outright.

It said the measure, used by publicly-listed companies to boost flagging share prices, was a symptom of “impatient capital” that threatened to undermine the stability of financial markets.

Terry Smith, CEO of Fundsmith also wades in:

“There should be a degree of disclosure and transparency. Companies should have to say why they’re buying back shares and why it’s better than paying a dividend.”He also said that share buybacks could contribute to financial instability. He said: “It’s a form of de-equitisation, to pay out in that manner reduces the capital base of a company and makes it more unstable under adverse conditions.

The Finnov report, titled Financing Innovation and Growth: Reforming a Dysfunctional System, was published on Thursday at a conference at the UK House of Commons.

It marks the end of a three-year project started in March 2009, with European Commision funding of €1.49m, as a response to the financial crisis.

Yes the FT article was a very distorted view of the situation. For example it said "Having established the extent of surplus cash companies then have a choice of handing it back through a one-off special dividend payment or in a buy-back". One important other option was omitted - namely to return it to shareholders via a tender offer. That is often the best option.

It also quoted Lord Wolfson as saying that "A special dividend.....it's a payment to one set of shareholders at one moment in time". Well it gets paid to all equity shareholders at the time it is declared. I simply do not understand the reference to "one set of shareholders" - was he arguing it should be paid to past or future shareholders?

Another oddity is that the article is focussed around the concept of "surplus cash" when many buy-backs take place when a company often has substantial debt, and sometimes even increases it to enable it to do buy-backs. The motivation is clearly therefore for other reasons than having "excess" or "surplus" cash.

Personally I always as a matter of routine vote now against share buy-back resolutions other than in investment trusts. The more you examine this issue, the less justified they appear. And there have been much better articles in the FT in the past on this subject which gives a better view.

Mark,

Rio Tinto latest;


29/12/12. Reply from Rio Tinto reception. She had a system failure, did not hang up and I accept that. Asked if my mail was resolved. I said no. My reply;

“The problem has not been resolved or even addressed. Rio Tinto’s Finance Director Mr Elliott was quoted in the Financial Times and FT.com, he was reported as justifying the world’s biggest miner’s policy of buying shares instead of returning the shareholders money to the owners of that money. Rio will come under increasing criticism if it does not address this matter.
Mr. Elliott’s name is mentioned, so he must see the FT article and the comments on FT.com. Please ensure that that his assistant, PA or secretary has my email and subsequent correspondence; and gives it to Mr. Elliott to read.
Michael Van Moppes.

Debate on ft.com this morning re Rio Tinto. My contribution;


Share buy backs usually destroy wealth, they only create value for shareholders if the balance sheet assets are undervalued. Buying £100 of assets for £70 is intelligent business, but buying £100 worth of assets for £150 is not agnostic, it is stupid.

In today's markets, the consensus advice is to buy high yielding FYSE-100 blue chips. Mr Elliot, the finance director of Rio Tinto mentioned above, should know that.

MoneyWeek published research which proved that increasing dividends lifts the share price by much more than buy backs, I am asking the editor Merryn Somerset Webb, for permission to send the article to anyone who wants it.

Rio Tinto's dividend is rubbish, only 2.13%. One Rio share earns 450p, but the hapless shareholders receive a 67.4p pittance. That is an unnecessary 6.7 times dividend cover.

I do not buy Rio Tinto shares because of this.

Dunno Mark.

Hadn't seen other trades - that just stood out for size and premium.

Not related to the thread but what's with HFD today?

Slowly ginding upwards, trending to the close around 289p then a 1.46m UT trade at 301p !!

That last trade compares to the 2.35m for the day ie a very big deal.

If management could tell which way the share price was going to move in advance they'd be in the wrong job I think! Personally I think Halfords is fairly cheap in the 340-390p range and an absolute steal at 300p. I bought in at 295p and am very happy with my 7.5% yield. For a company that's a market leader in it's field, with a great return on capital, low debt and reasonable long term growth prospects I'm very happy to be able to buy it this cheap.

Halfords has been buying up a lot of shares in open market over last several months (as per their announcement in April).
It's indicative that much of their buying has been in the 340p-390p range, destroying value. Yes, some recent buys have been at lower levels of less than 300p. 
But as a prospective shareholder this does not inspire confidence in the management's ability to allocate shareholder capital efficiently. EPS incentives do not align shareholders and managers equally!

To respond to all the points:

Really like Terry Smith's take on things - especially like the slide he has where he shows low P/B shares respond very well to buybacks. Really wish I could get the same data for P/E but I assume it's similar!

Agree with Lex in it's conclusions. As I said it's exactly the same as buying more shares so for an investor who is truly LTBH and doesn't sell if the share starts getting too highly priced then it can be value destroying. For me though, I don't hold a share unless I think it's significantly undervalued so buybacks complement my investing preferences.

True it's hard to accurately determine the intrinsic value of a trading business as earnings can be very hard to predict but in the case of say Halfords you can buy at such a wide margin of safety you can be reasonable assured of purchasing below intrinsic value.

Mark

Interesting take from Lex

http://www.ft.com/cms/s/3/b81c6984-cd92-11e0-b267-00144feabdc0.html#axzz1VhPmy9yS

Investors who do not match the buy-back with a proportionate sale are in effect buying more shares. That hurts whenever companies buy back at the wrong time.

Xig

The problem is that intrinsic value, for a trading business is unknowable

Indeed but perhaps more importantly this is likely to be wildly different dependant on your view of :

1) Macro picture
2) If you can combine the assets with your own to create synergies
3) Future prospects

This might be of interest - I just came across it - still digesting it - download the two files at the end of the piece

It's by terry smith of fundsmith fame

Hi Mark,

I don't believe we have exchanged emails off-board! Are you confusing me with someone else? I did sit opposite you at the last Mello event though if you remember me :)

Indeed you're correct in saying that the model is too 'perfect' in the sense that no company has earnings that grow so predictably forever and that this is a gross oversimplification. Also, some companies can grow with very little capital required to do so (Buffett's famous example of this is his purchase of See's Candy - the brand equity gives the business great pricing power). Other businesses are cyclical and do not follow such a pattern of steady growth.

Even setting the ROI to -5% however (I guess this would correspond to the business declining despite the reinvestment) and taking a 5 year period to compare to your model I still get buybacks as being slightly better (due to the reinvestment effect) for a P/E of even 15.

I agree that a lot of boards tend to like share buybacks due to bonus reasons and this encourages share buybacks regardless of the price relative to the intrinsic value of the business. Buying an overvalued asset isn't a good use of capital. If the share is overvalued though you probably shouldn't be owning it in the first place!

I guess the key point I'm trying to make is this: If one purchases a share because they believe the currect share price undervalues the company (be it due to assets or future cash flows) then share buybacks are simply the same as you buying more of the shares yourself - it amplifies your exposure to the underlying asset by giving you a greater % ownership of it.

It's exactly the same as the company paying you a dividend and you thinking "Great! Now, where shall I reinvest it? I know, the halfords share price is still awfully cheap...." apart from it's even better than that as it's slightly more tax efficient.

On a side note the more I look at Halfords the more I think the share is ridiculously undervalued... decent growth over a 10 year period, reasonably predictable earnings, incredible ROCE, low debt, huge yield and v low PE! What more could you want!? Definitely going to buy some come payday :) Thanks for the tip!

Mark

P.S. Please could you host your model somewhere so I can take a look? I'm curious as to what's different to mine.
P.P.S. Thanks for the AGM write up - a really useful resource!

fwiw I'm against buy backs as any beneficial effect is so small as to be negligible, compared to the effects of other factors on the price
However i see XOM has bought back a massive 28% of it s stock over the last 10y (is this a record?) as shown by this graph

Since the PE has ranged between 10 and 15 has value been destroyed?

Roughly 2/3 of spare cash has gone on buy backs - one wonders where the price would be if there were no buybacks! ...............  40$?

Oh .... and I almost forgot - non of the above has been adjusted for inflation!

schober

PS Does anyone know of a list of high yield stocks that don't buy back stock?

Marben,

I would like to respectfully disagree with the conclusion drawn in this article and explain to you my logic for doing so. I have constructed a spreadsheet to support my conclusions:

http://dl.dropbox.com/u/6961653/dividendsvsbuybacks.xlsx

I'm not sure how you've constructed your spreadsheet so am unsure as to how your method has deviated from mine but I shall explain how I did my analysis and we can compare methodology after.

I have three variables in my model - company P/E, ROI and Payout %.

In my two scenarios (Dividends vs Buybacks) I start with a company earning £100m with 100m shares in issue for a nice easy EPS of 1. Both scenarios keep the P/E of the company constant (a big assumption I will elaborate on later in my analysis). In the dividend model the defined amount of money (from the payout ratio) is paid out from the company and the remainder kept back and reinvested at the ROI specified. In the buyback scenario the model reduces the shares in issue by the appropriate amount given the capital spent on buybacks and the appropriate share price.

As each 'year' in the model ends the cumulative percentage return from an investment made in the shares at their initial price (£15 with the inputs in that file, as the P/E is set to 15) is calculated by the following formula:

(Sum of all dividends paid to date + (Share price now - Share price at purchase)) / Share price at purchase

As you can see from the model, regardless of the parameters entered in for P/E, ROI and payout ratio buybacks always come out ahead.

So, why is this? Does this mean that all share buybacks are beneficial? Are they better than dividends?

The discrepancy between the two methods of returning capital to shareholders in this model is due to one thing only, and that is the fact the dividends are not in any way reinvested in any asset. They cease to increase in value whereas under the buyback scenario they are reinvested - by increasing the % ownership of the underlying business per share.

However the difference between reinvesting this cash (buybacks) and not in this model (dividends) is exaggerated at low P/Es and for businesses that can earn a high return on capital. For a P/E 15 company, the % increase of the buyback scenario over the dividend scenario after 10 years is only 21%. As P/E comes down this difference rapidly compounds. If the company is permanently valued at a P/E of 5 (an obvious mispricing for a company which can earn a 10% return on capital and has stable earnings growth) the % increase of the buyback scenario over the dividend one is 91% after 10 years - a huge difference. Conversely, try setting the P/E to a bubble price, say in the 50's, and watch the difference come down to nothing.

However there is one big caveat to this model - the P/E of the business is held constant. In reality such a relationship is not realistic; P/Es swing around as sentiment and business performance does. If you initially have the business valued at a ridiculous bubble price of say P/E 40 and then slowly bring this down to a reasonable P/E of 10 then your share buybacks have been value destroying as you have simply been purchasing an overpriced asset the entire time (Microsoft might want to take note of this point).

Conversely, run the model with a low P/E then adjust the P/E back to a reasonable one in year 10. Now that's value creation!

So what does this imply for the intelligent investor? Share buybacks are equivalent to receiving the dividend and simply buying more of the share you currently own. This will simply amply the investor's exposure to the underlying asset and will be value-creating if the asset is cheap but value-destroying if the asset is overpriced. Because of this, I have no issue with the shares I own doing share buybacks because I own the shares precisely because I believe they are underpriced (I hope!!) and they are merely performing a practise I already agree with - buying an asset that's cheap.

I'd be interested to hear your thoughts on my interpretation of this.

Mark

The theory of share buy-backs is complex - there have been analyses published in the FT that questioned the original theory. In any case, even if a buy-back was justified it is best to do it via a tender offer than by market share purchases. Of course one of the problems (as one can see at Halfords) is that management is often incentivised by eps based bonus schemes to support buy-backs rather than do other things with the cash. At Halfords I can see lots of good opportunities to re-invest that cash in the business at good rates of return. I consistently vote against share buy-back resolutions as a matter of routine (except in investment trusts).

Hi marben

To be fair to M-M there is a caveat: that taking on debt finance potentially exposes a company to bankruptcy risk. Oilexco's management either ignored or failed to understand this. There's a recent article on the FT by John Plender on exactly this point: http://www.ft.com/cms/s/0/01e19a6e-759e-11e0-80d5-00144feabdc0.html#axzz1TPmlQrEW.

Of course calculating tail end risk isn't a science, so cash returns in whatever form should be conservative.  If a buyback involves taking on extra debt for tax efficiency purposes I think you'd want to be pretty comfortable about future cashflow projections (or possibly just not a shareholder).  But you've made that point already :)

Excellent thread, Btw.

timarr

....the one that's always appealed to me is that people tend to separate income from capital and are willing to spend the former rather than the latter. This would fit with a number of behavioural theories around mental accounting and self-control so that by separating dividends into a separate "mental account" which they can spend they avoid the temptation to sell shares to fund their lifestyle

Yes I think there is a very great deal of truth in this. I consider it to be completely irrational (and have been living happily for a decade on the proceeds of capital, having never had an income-producing share in my portfolio until the last month or two) but I cannot deny that plenty of people exhibit a strong preference for dividends.......

...which is a point I am entirely relaxed about, as it means that there is demand for income-producing shares that have gone ex-growth and, more importantly, the price of stocks with growth potential is lower than it would otherwise be (if there were more competition for them) - so I can buy them more cheaply.

ee

There are other factors that seem to drive the use of buybacks: there's a correlation between a company conducting buybacks and the size of a CEO's stock option plan, for instance. Of course, correlation isn't causality but still ... In addition there's also evidence that the introduction of accounting rules forcing the recognition of eps dilution through option plans increased the level of buybacks, presumably to counter the dilution effect: (written about in http://www.psyfitec.com/2009/11/buyback-brouhaha.html)

The classical economic theory that argues that whether value is returned to shareholders in the form of dividends or share buybacks is the Miller-Modigliani dividend irrelevance hypothesis. Unfortunately this assumes that shareholders are rational and ignores the evidence that they actually prefer a good dividend. There are multiple theories why this should be so, but the one that's always appealed to me is that people tend to separate income from capital and are willing to spend the former rather than the latter. This would fit with a number of behavioural theories around mental accounting and self-control so that by separating dividends into a separate "mental account" which they can spend they avoid the temptation to sell shares to fund their lifestyle (written about in http://www.psyfitec.com/2010/01/psychology-of-dividends.html)..

There's also the tendency of managements to want to smooth dividends because of the negative signals given by cutting them: so excess earnings are often returned in the form of special dividends or, more frequently, buybacks, rather than by increasing dividends, regardless of whether this is value enhancing or not. Because of this dividend forecasting tends to be more accurate than earnings forecasting, although that's not necessarily a good thing. Rationally we should want the largest dividend possible all the time, even if this means more frequent cuts, but if you're an investor assuming a gradually rising income stream and planning your spending on that basis you'll tend to have a different view, even if that means your overall income over a period of years is less than it otherwise would have been. Strange, but true..

It's a fascinating area, but one which proves once again that rationality depends on your perspective.

timarr

It shows that if the buybacks are made on a P/E of ~5 or less they are unequivocally beneficial to shareholders; between 6 and 8 they are not particularly damaging. At 10 and above they are much more questionable (but not as damaging as I'd thought previosuly).

This is the sort of conclusion that one would think of as being intuitively correct. It is simply that the future cashflows are being priced at a discount to the rest of the market.....and, if they aren't, then it probably isn't worth doing a buyback.

Of course in the real world, future earnings are never static. If the likelihood is for declining earnings, then a dividend now is likely to be more beneficial. If growth may reasonably be expected, then buybacks on a slightly higher multiple may be justified, as this will accelerate the rate of increase of dividend per share.

This is absolutely the key set of arguments. If there is a growing business, then continuing to reinvest in it would generally be shareholders' preference - so buybacks of shares tend to be favoured (for a related example, consider the fact that SOCO International (LON:SIA) fully covered all the share awards in their LTIP by buying shares in the market some years ago when the share price was a fraction of today's price).......

....but if, OTOH, the business was expected to struggle more...or was seeing a cyclical top in its revenue stream...then investors would prefer to see cash returned to them. It is for this reason why cash takeover bids are preferred to paper bids at market tops - but people are happier to take paper at market lows.

One of the reasons that buybacks get criticised is that they often seem to be carried out by very large, mature, fully-valued companies who are more interested in manipulating their EPS (presumably because EPS is one of the key metrics that affects executive compensation in the case of those companies).

ee

D'OHHHH!!!! I owe everyone a huge apology. I've just rechecked my spreadsheet and found a major error in my calculation of TSR: rather than adding the difference in opening and closing share prices to dividends, I was adding the difference in EPSs. Using the correct calculation I get the following values for TSR (with buybacks) vs P/E at which buybacks are performed:

This radically alters my conclusion. It shows that if the buybacks are made on a P/E of ~5 or less they are unequivocally beneficial to shareholders; between 6 and 8 they are not particularly damaging. At 10 and above they are much more questionable (but not as damaging as I'd thought previosuly). Of course in the real world, future earnings are never static. If the likelihood is for declining earnings, then a dividend now is likely to be more beneficial. If growth may reasonably be expected, then buybacks on a slightly higher multiple may be justified, as this will accelerate the rate of increase of dividend per share. With a 10% annual growth rate for earnings, I have verified that buybacks made on a P/E of 12 remain beneficial to TSR, whereas with a 10% annual decline even buybacks made on a P/E of 5 are not beneficial, compared to simply paying a cash dividend.

Regards,

Mark

Hi all,

I quite agree that buybacks in the case of an "asset play" are a different proposition to buybacks for an "earnings/income play", i.e. where your investment is valued based on its NAV, as opposed to where it is valued on the basis of a future earnings & dividend stream. As ee says, Encore's rigorous approach to its buyback programme was exemplary and could only add to value per share. Firms such as Polo Resources (LON:POL) and Regent Pacific (HK:0575) have similarly used buybacks to add to asset value and benefit shareholders.

However, my analysis shows (to my own surprise) that buybacks almost never seem to be justfied, on fundamental grounds, by companies valued on earnings/dividends: shareholders will be better off if the excess cash is simply repaid by means of an extra dividend. Tender offers are a special case and different considerations may apply.

FW Thorpe (LON:TFW) illustrates an interesting example of an approach that benefits shareholders nicely. It pays a reasonable (but not exciting) progressive dividend - and sits on a substantial cash pile. Every 3 years or so, the cash pile grows beyond what management deems reasonably necessary for their very prudent treasury approach and they reward shareholders with a nice special dividend. Thorpe only rarely buys back shares.

Regards,

Mark

FWIW, One of the best posts I've seen on Tax effects of buybacks was from Geng on TMF:

http://boards.fool.co.uk/obviously-this-will-improve-earnings-per-share-12246468.aspx

A very interesting topic Mark and I wish I could give it more time but I'm on holiday till 11 August and only typing this whilst having a quick fag....

The textbook example has to be Next where, IIRC, they have now bought back over 55% of the shares in issue. Using surplus cash, over capex and divi needs, to reduce the shares in issue is an obvious way to enhance value as opposed to wasting it on dubious acquisitions and gives a very clear demonstration of a management committed to its core business over the years.

Another obvious example would be Character Group, though, IIRC, some of the buybacks were from management - dubious (?)

Given a stable business, delivering the excess cashflow in a combination of rising divs and shrinking share base looks a winner. Unfortunately many co's buy back at the wrong price - look at the absence of buybacks during the Crunch....however, buying back when you've got a FCF yield in the double figures is much more acceptable.

Disc - long HFD from 360p.

Cheers


D - who may have had a bottle or two so don't rely on anything I've said above. ;-))

There's a very minor effect on stamp duty as well in that a company purchasing its own shares will pay stamp duty if the shares AREN'T cancelled and are held in treasury.