Intrinsic value analysis works, kinda

After next week’s monthly review of the Thrifty 30, I really need get back to uncovering undervalued companies. The question is will I use the newfound residual income model, which compares market value to intrinsic value?

Maybe. According to SG analyst Dylan Grice, who adapted the version I’m using, It works better than the trailing PE when you simulate buying the cheapest largish international stocks each month and shorting the most expensive:

But here’s the reason he’s cautious about basing residual income forecasts on the historical return on equity:

Profitability generally reverts towards the mean (without actually getting there – apparently companies ‘asymptote’ towards the mean). That’s why Grice allows ROE to tend towards the industry average over time in his model.

And here’s what you miss if you make conservative assumptions like mean reversion; the next Microsoft:

Companies with a durable competitive advantage will produce excess returns for decades and consequently earn massive residual income, which the model cannot incorporate into the company’s intrinsic value unless instructed to.

As Grice says:

Companies on average don’t maintain their competitive advantages over time, which is why it’s probably a fair assumption to make when screening. But some companies do. For every hundred or so AOLs there’s a Coca-Cola.

Fiddling with the Diploma model, I can easily make the shares look cheap by treating it more like a Coca-Cola. All I have to do is assume Diploma’s ROE reverts upwards to the mean (the industry average is higher), extend its period of excess returns to 10 years, and assume it doesn’t go into decline thereafter but grows in-line with global GDP (it’s an exporter).

Here’s the spread sheet (you can compare it to my original one):

The new model produces an intrinsic value to price (IVP) ratio of 1.5, meaning the intrinsic value of Diploma is 50% higher than the market price. The shares are cheap, but notice how over half its value now is earned over a decade in the future (compare [Line 10] to [Line 11]).

Which scenario is more likely? Six years of continued growth followed by decline, or 10 years of accelerated growth followed by stagnation? Who knows? A computer can’t tell whether a competitive advantage is durable or not, only Warren Buffett can.

OK, perhaps that’s a slightly pessimistic statement, but although I’ve been thinking about what makes a company great, I wouldn’t like to wager my hard-earned savings on my ability to differentiate temporarily successful companies from genuine super stocks.

So, since Microsofts and Coca Colas are rare, it’s better to take the conservative view. To keep the value in our investing, we should be confident that a share is cheap.

As I said in my second post on Diploma, which seems a long time ago now, that doesn’t mean forgetting about the company. A very conservative estimation of its intrinsic value is 200p. For most of 2008 and 2009 it was trading well below that level so just because the shares are expensive now, they needn’t be forever.

A strategy for cautious investors might involve stalking Diploma, recalculating its intrinsic value from time to time, until a temporary dip in price presents a buying opportunity.

Here are some more imperfections identified by Grice:

• The model uses return on equity, which favours indebted companies and punishes those with a lot of cash. This is surely a factor in largely unindebted Diploma’s lowly valuation, for example.
• It uses book value, the value of the business according to accountants, which generally (though as Grice points out not always) is more prudent than realisable market value.

For a private investor, calculating the IVP ratio is a lot more work than a dividing the market price by earnings, say, and a little less ambiguous. I say do it at least once, because it really exposes the assumptions we make when deciding to buy or sell a share.

Calculating the IVP regularly, might lead us to question our assumptions more, and who knows where that could lead.

Calculating the intrinsic value of Diploma

I suppose this day had to come. This is what investors do when they’d really like to invest in a company but the trusted measures they use, in my case the ten year PE ratio, say it’s a bad idea; the company looks expensive.

We look for a new measure that makes the company look cheap.

But first, some caveats:

• This is the first time I’ve done this, so I might not have done it right, and if you can see a mistake I hope you’ll tell me.
• Yorkiem has been a great help as we’ve batted the model backwards and forwards deciphering it, but any mistakes are mine.
• I have a few doubts which I will not dwell on in this post as I just want to write it down. The criticism, positive and negative, will follow, soon.

You may have seen this post coming. First, the courtship of Diploma, which is the kind of company I’d like to put in the Thrifty 30, if only I could buy it cheaply enough. Second, last week’s rumination on intrinsic value offered the tantalising prospect of a method that might account for Diploma’s likely growth in a more convincing way than I can.

The method is derived from Stephen Penman’s Residual Income Model*, and pinched from a note by SG analyst Dylan Grice.

Here’s Grice’s model loaded with my data on Diploma (and here’s an Excel version to play with):

Now for the hard bit.

The model starts with the company’s book value per share in [Line 3], which is 136.1p. This is the accounting value of the assets on its balance sheet net of liabilities, i.e. net asset value or equity.

It uses the median return on equity over the last ten years (14.9%) in [Line 4] to forecast net income in [Line 1].

So, the first year of the forecast (2011) gives a net income (return) of 20.2p on 136.1p of equity (at the end of 2010) because:

136.1p x 0.149 = 20.2p.

From 2011’s net income we forecast how much will be paid to shareholders in [Line 2] using the median ten year payout ratio, which for Diploma is 0.44 times eps (44% of profit).

20.2 x 0.44 = 8.9p

Having forecast earnings for 2011 in [Line 1], and dividends in [Line 2] we can forecast the book value of Diploma in 2011. It’s the book value in 2010 plus net income, minus the dividend:

136.1 + 20.2 – 8.9 = 147.4p

And from that we can forecast the residual income in [Line 5], which is net income less the expected 10% return on equity. This is may require a little explanation:

The model rests on the notion that its not worth paying more than book value for a company unless it earns more than the annual rate of return you desire. Grice uses the analogy of a bank account:

The simplest such example would be that of a bank account containing $100, earning 5% per year interest… What’s it worth? Let’s assume my desired return is 5%. The bank account is worth only its book value of $100…. It may be liquid, stable and even growing, but since it’s not generating any value over and above my required return, it deserves no premium to book value.

Since we’re talking about shares, his model desires a higher rate of return (10%). Any ‘residual income’ over 10% earned in future will add to the value of the company now – it’s a bonus. The residual income of 6.6p per share forecast in 2011 is derived from net income forecast in [Line 1] less 10% of equity at the end of 2010 [Line 3]:

20.2 – 13.6 = 6.6p

However, income in the future isn’t as valuable as money in our hands now, so we need to reduce it by our desired return of 10% a year [Line 6] to give its present value (PV) in [Line 7].

Replicating the forecasts for subsequent periods, we have all the data to calculate Diploma’s intrinsic value. It’s the original book value per share in [Line 3] plus the sum of the residual income expressed in today’s money [Line 8] plus the continuing value (CV) of the business after our forecast period also expressed in today’s terms [Line 10].

136.1 + 38.1 + 24.5 = 198.7p

Grice follows the common practice of valuing the continuing business as a perpetuity in [Line 9] earning the same income every year, but uses a really harsh discount factor of 25% because companies generally succumb to competitive pressures in the long-term and go into decline.

It’s a rational response to the typical investor tendency to extrapolate past success into the distant future, but it seems draconian when I look at Diploma, which has been consistently profitable over the last decade and is financed for a consistently profitable future (I’m not saying forever!). The problem is if we are too optimistic about prospects in the distant future most of the businesses’ present value could be derived from the most uncertain of the three components of value, book value, residual income, and continuing value. That will be true of exceptional business, but they’re few and far between.

The calculations imply an intrinsic value of about 200p in [Line 11], which is more than the current book value of about 135p, but nowhere near the current price of about 325p (I’ve rounded the numbers because the scientific precision of the spread sheet implies a degree of accuracy that isn’t even remotely warranted!).

Dividing intrinsic value by price gives the intrinsic value to price (IVP) ratio in [Line 13], which at just over 0.6 suggests Diploma shares are expensive. The intrinsic value is just over 60% of the market value.

There are huge assumptions in a model like this, which Grice acknowledges. The words of Benjamin Graham written in The Intelligent Investor, and repeated in Penman’s textbook* are haunting me as I type:

The concept of future prospects and particularly of continued growth in the future invites the application of formulas out of higher mathematics to establish the present value of the favoured issue. But the combination of precise formulas with highly imprecise assumptions can be used to establish, or rather justify, practically any value on wishes,however high, for a really outstanding issue.

I suppose Graham was talking about the discounted cash flow model, a cousin of the residual income model but although, for better or worse, Grice’s very conservative assumptions have kept the Thrifty 30 out of Diploma, it’s easy to see how with a tweak, here, and another there, I could make it give me the result I want. As I said, I’m going to criticise, and perhaps experiment with the model in a future blog. For now, I’m happy just to have got it out there.

BTW, John Kingham, UK Value Investor, extrapolated future earnings from historic ROE to value Mears.

* See: Financial Statement Analysis and Security Valuation, a heavyweight textbook that is very well written. One day I’ll work my way through it from beginning to end.

Some notes that may help, if you plan to do this at home:

• The model for Diploma diverges from Grice’s in [Line 4] Grice gradually reduces the ROE of the business to the industry average over the forecast period to account for likely competitive pressures but Diploma’s ROE is actually lower than the median for the Industrial Supplies sector. That may be because Diploma is much less indebted than some of its peers, or it may be a fantastic sector! I don’t think it’s because Diploma is a bad company, so I’ve assumed its ROE remains at the same level as its historic median over the forecast period.
• To establish the median ROE for the Industrial Supplies sector I exported the ten-year data series for each individual company in the Industrial Suppliers sector from Sharelockholmes, calculated the medians and averaged them (see the second tab of the Excel sheet).
• To establish the payout ratio I exported Diploma’s ten-year earnings per share and dividend per share record to calculate its median payout ratio (dps/eps – see the third tab of my Excel sheet).
• Here’s Grice’s worksheet for Akzo Nobel, unadulterated (click for a larger version):

Intrinsic value: the holy grail of value investing

If you think market valuations of companies are sometimes wrong then implicitly there must be correct values for those companies. If only you could work them out out you would truly be in the money. You would, in the vernacular of value investors, be buying dollars for 40c.

But the intrinsic value of a company now, depends on what it does in the future. If it goes on to make a lot of money, it’s worth a lot now. If its future lies in the hands of the administrator it’s not worth much. Since businesses are complex, like the economy, or the weather, it’s very hard to make accurate predictions about them.

Many of us use rough-and-ready proxies for value, like a multiple of earnings, or asset values, and compare the market’s valuation to them. Because these proxies have a loose relationship to intrinsic value we are, on average, buying companies at lower prices, without the bother of working out what exactly their intrinsic values are.

A recent note from SG’s Dylan Grice has got me thinking about the holy grail again though. It explains how he calculates intrinsic value, and when I’ve worked it out I’ll write a note about it on this blog. To start with, he’s sent me scurrying back to my copies of Security Analysis and The Intelligent Investor to find out what Benjamin Graham meant by intrinsic value, since just like Grice today, Graham compared it to the market price to decide whether a company was undervalued.

The first chapter of Graham’s 1940 edition of Security Analysis defines the concept of intrinsic value:

In general terms it is understood to be that value which is justified by the facts, e.g., the assets, earnings, dividends, definite prospects, as distinct, let us say, from market quotations established by artificial manipulation or distorted by psychological excesses.

But unlike the price targets we see calculated today, Graham stressed the indefinite nature of intrinsic value:

The essential point is that security analysis does not seek to determine exactly what is the intrinsic value of a given security. It needs only to establish either that the value is adequate—e.g., to protect a bond or to justify a stock purchase… For such purposes an indefinite and approximate measure of the intrinsic value may be sufficient. To use a homely simile, it is quite possible to decide by inspection that a woman is old enough to vote without knowing her age or that a man is heavier than he should be without knowing his exact weight.

Or to mangle an aphorism, if it looks like a bargain, and smells like a bargain, it probably is a bargain.

At it’s simplest, intrinsic value analysis is like using the plain old PE ratio. Saying a company is cheap because it has a PE of, say, ten is the same as saying its cheap because it should be worth more than ten times its past earnings, a level at which it might pay an investor a reasonable return of 10% in a typical year (one tenth of your investment of ten times earnings). But that’s a generalisation. Some companies will be facing remorseless competition, some may succumb to their own debts, and others may be mismanaged into oblivion. Some will earn bigger profits in future.

Although, I think, very few investors decide which shares to buy based purely on the PE ratio, or price to book, some mechanical investors rely on a cocktail of two or more statistics to shield them against some of the risks and tilt their portfolios towards value. Other investors, like me, add a dollop of common sense to their algorithms in the expectation it will juice the returns.

I doubt a significant proportion of private investors actually calculate intrinsic value, and seek to pay less than a pre-ordained fraction of it (Graham favoured 2/3 after estimating intrinsic value very conservatively). It’s just too much work, there are countless very well resourced analysts doing it in the City, and as the complexity of forecasting future sales and profit margins builds up when you consider all the factors that could influence them, so do the uncertainties.

Even Graham had had enough by the end of his career, saying in 1976:

I am no longer an advocate of elaborate techniques of security analysis in order to find superior value opportunities. This was a rewarding activity, say, 40 years ago, when our textbook “Graham and Dodd” was first published; but the situation has changed a great deal since then. In the old days any well-trained security analyst could do a good professional job of selecting undervalued issues through detailed studies; but in the light of the enormous amount of research now being carried on, I doubt whether in most cases such extensive efforts will generate sufficiently superior selections to justify their cost.

Instead he favoured a semi-mechanical system using the venerable PE ratio as a proxy for value, and the ratio of equity to assets to guard against poorly financed companies.

But Grice’s Intrinsic Value to Price (IVP) method is a step back towards the kind of shorthand most of us use. Instead of basing his calculation on earnings forecasts, he extrapolates future dividends and increases in book value (which together are earnings) from the median Return on Equity over the last ten years.

I’m attracted by it because profitability, of which ROE is a measure, can, like value and financial strength, help us discriminate between winning shares and losing shares (Joel Greenblatt’s Magic Formula demonstrates it). A number of bloggers, myself, Yorkiem and UKVI, included, are experimenting with including ROE in our screens because we believe that in general high ROE is predictive of higher earnings in future.

In other words, we’re looking for a way to build earnings growth into our calculations, without having to make traditional forecasts.

As I said, when I can do a worked example, I’ll share it with you. To speed up the process, I’ve sent Grice’s note to some bloggers who may be able to help.