Buffett's 5 second intrinsic value calculation

[luutzu]

Thanks luutzu, great post and insight into your investment process.

When you work out forecast earnings for the next 12 months, to what extent do you generally find your figure differs from 1 year consensus forecast normalised earnings/EPS figures quoted on online broker websites?

Just in case you're using the IV = Earnings/r, where r = risk free rate, or interest rate. Not recommended to use r as that.

The r is your required rate of return, so if it's the risk free rate like bonds or term deposit, you might as well put your money there and take on no risk. Investing in businesses have risks. So if r is 4%, that's a P/E ratio of 25... roughly mean you'll get your money back in 25 years - see Switesh post on bird in the hand.

With my estimates... I don't do forecasts. I simply use the what-ifs analysis and after I get to understand the business well enough, see that it is of high quality, financially sound, does not show the need to raise new capital or take on debt to survive... Once I'm happy with it I just do a simply straight line scenario of sales remaining as is, goes down by x%... what price I'm happy with under such circumstances and see what's the offer is.

Though there are companies, as Phillip Fisher discussed in his very good book... that some companies are so exceptional, so able and well managed that they continue to grow year after year... that kind of company would deserves a premium on its "reasonable" price.

How much does the market tend to pay for that premium I don't know, but you can see what you're comfortable with knowing the business as you have come to know, knowing the market level. Though above 25 times earnings is really really pushing the bound of reasonable, high expectation.

Anyway, I'm just starting to sort out a standard approach I could use, never work in the finance industry so take what i learn for what it's worth.

But one thing to take away is Buffett's and Graham's advice to treat stock as a business, not as a financial instrument. As Buffett said, once you think like that, everything else will follow.

-----

With consensus and expert forecasts, with rating agencies... It's safer to ignore them completely. The expert forecast seem to be very good at forecasting next quarter or next year's earnings... but that could just be pure luck or pure manipulation by management - often I looks like simply straight line projection. No one knows until it's too late.

In my opinion, to accurately forecast a company's future earnings; to forecast it to a high degree of certainty - which is what you ought to else it's just guess work and is more harmful than useful... to get a forecast right to a degree acceptable by science, say +95% certainty the data and samples and figures aren't random guesses... that takes a lot of brain and a lot of work... most of which an analyst working for a fund manager who then give it out for free to newswires for free just can't afford to have.

With the few lawsuits against the major rating agencies in the US after the GFC... most say their work are no more than opinions, like newspaper columns, not advise or expert analysis. But yea, somehow their work affect companies and gov'ts ability and costs to raise fund and attract investors.


Let say you want to forecast a company's earning in two years. How to go about it?
Look at the industry, the structure, the players, the suppliers, the customers, the economy, changing consumer demand... Then once you know all these players and their relationship and influence on one another...
You don't need to go further to know how insane it is to think anyone could do it, and if they could, rating agencies could pay for it.

 

[TPI]

Thanks luutzu, what you're saying makes sense.

Personally though, in the interests of simplicity in valuation approach I am still leaning towards using 1 year forecast consensus normalised EPS for the E in the IV = E/r formula.

Reason being is that it is an easy figure to find and do so very quickly, a 1 year figure is more likely to be more accurate than a 5 year figure, these figures are based on company guidance which maybe more conservative to begin with, and small misses in forecasts are unlikely to materially affect the end valuation which is going to be an imprecise figure to be used for general guidance anyway.

Interesting article on this below:

http://www.fastgraphs.com/_blog/Res...-03-14-Carnevale-Forecasting-Future-Earnings/

For CBA as an example: using Google Finance/Thomson Reuters data, CBA's FY15 EPS consensus estimate is $5.4961 (the mean figure based on 16 analysts estimates).

If my required rate of return is 10%, CBA would be worth $54.9461, which could be calculated in less than 1 second once the EPS figure is obtained.

If I have a strongly positive long-term view on CBA and used a lower required rate of return of 8%, CBA would be worth $68.70.

With today's share price being $74.80.

Notwithstanding of course all the qualitative and personal factors (eg. the requirement for greater yield which maybe present despite apparent over-valuation) that go into the final purchase decision.

 

[galumay]

Personally though, in the interests of simplicity in valuation approach I am still leaning towards using 1 year forecast consensus normalised EPS for the E in the IV = E/r formula.

Personally I dont like using any forecasts, when I am using earnings in calculations I only use previous year earnings and I adjust that if its not in line with historical growth.

 

[McLovin]

Personally though, in the interests of simplicity in valuation approach I am still leaning towards using 1 year forecast consensus normalised EPS for the E in the IV = E/r formula.

This is just a terminal value formula that assumes zero growth. Have you played around seen what happens at various risk levels or if you add 1-2% growth? It kind of shows up the pointlessness of that formula.

Also, those one touch formulas are prone to extrapolating the good times in a boom and the bad times in a bust.

 

[luutzu]

This is just a terminal value formula that assumes zero growth. Have you played around seen what happens at various risk levels or if you add 1-2% growth? It kind of shows up the pointlessness of that formula.

Also, those one touch formulas are prone to extrapolating the good times in a boom and the bad times in a bust.

True that it assumes zero growth, but that's the point though.

Also true that if we apply this to companies doing very well because its industry is booming, we'd be poorer when that boom burst.

But...

That's why we shouldn't take last year's earnings, or the average previous earnings at face value.
That's why we look at the company's quality, its market position, its financial strength, its management.

That's easier said than done, and when it's done - I think I did it right on two occasions so far - it's still quite scary to see the market doesn't agree with you.

But when you study the company properly, happy with its quality and position, comfortable with its price... as Buffett said somewhere, something good might happen.

One of the advice Fisher gave was to ask whether the company has done well because it is lucky or it is able; or whether it is both able and fortunate. If the answer to a company doing well right now is simply that it was lucky, that its industry is booming and a rising tide lifts all boats; that it's lucky because one of its products does very well... then may not want to get in to it... But if it was well managed, if its management had made more intelligent decisions than poor ones, that there are R&D or other efforts to produce new products/services, to expand further beyond the current luck and boom, it's worth the risk.

So we can look at, say the mining companies or the industrial services... and see that it is obviously not booming, going downhill. If the investor know the business and the industry... I don't think it's correct to say that the entire industry and every business in it will drop by x% in revenue.

Who knows, a top quality operator in a contracting industry may even benefit from the slow down.. It might die too but it could manage to survive and buy off smaller and weaker rivals now struggling or going bankrupt; because of its position and reputation, it might win contracts that would go to someone else at lower prices in boom time.

Anyway, that's just my opinions and what I understand this approach to imply... to do otherwise is to try and predict the future, most likely from understanding the general industry or macro-economic trend... that and to pay at a price where that future never come - an approach which might be safe in the downturn but prove expensive in boom time.

 

[galumay]

t... But if it was well managed, if its management had made more intelligent decisions than poor ones,...

I have been thinking about this criteria a bit lately, I am not sure that its not better to look for bad management being able to make good profits - if you pick companies with really good management making good profits, there is always a risk that management will change to less good or just make some poor decisions and profitibility will suffer.

On the other hand a really badly managed business that is very profitible has more 'risk' of management changing for the better!

I look at companies like RIO, which is probably the most incompetently managed company I have ever had a close association with , and say TLS which has suffered massively from poor management - yet both have managed to make squillions of dollars and deliver real wealth to shareholders.

Its a bit counterintuitive but maybe I need to filter for poor management/high EPS

 

[luutzu]

I have been thinking about this criteria a bit lately, I am not sure that its not better to look for bad management being able to make good profits - if you pick companies with really good management making good profits, there is always a risk that management will change to less good or just make some poor decisions and profitibility will suffer.

On the other hand a really badly managed business that is very profitible has more 'risk' of management changing for the better!

I look at companies like RIO, which is probably the most incompetently managed company I have ever had a close association with , and say TLS which has suffered massively from poor management - yet both have managed to make squillions of dollars and deliver real wealth to shareholders.

Its a bit counterintuitive but maybe I need to filter for poor management/high EPS

Very true. haha

I think Peter Lynch or Buffett said similar things - it's better to buy a business any idiot can run, because one day an idiot will run it.


I read "Barbarians at the Gates" before and yea, how KKR analyse Nabisco for the takeover was along the line you're thinking. The RJR-Nabisco's CEO was known for the extravagance fleet of executive jets, suites, private hangar etc... so KKR deduct these expenses but still couldn't make the numbers... until some executive somewhere within the empire points to them that the CEO ordered him to spend x amount on needless advertising to keep the profit low as he prepared for the management buy out that led to all these.

I heard of this executive where I worked buying a $5000 chair. An office chair, gas lift... $5 grand... dam. Public company, not his money.

 

[McLovin]

True that it assumes zero growth, but that's the point though.

Well, I guess we disagree. I always thought a valuation should have some modicum of reality in it.

 

[luutzu]

Well, I guess we disagree. I always thought a valuation should have some modicum of reality in it.

Once you can estimate a company's current earning power, and being current does not mean it will report this earning, or achieve this earning right now - that earning power would show itself in one or two years time, after the current difficulties or the current boom for example...

Doing such estimates under various scenarios and there is no need to assume x% growth in the future. It is already expected to grow by at least the rate r that you required into eternity. The future might proved it exceed or decline from that requirement... if exceed, the business is worth more etc.

To assume growth at x% in the future you tend to project the current reported earning, if not projection but do analysis to get the earning power then estimate its growth - in one or two phases - you can do that, I just don't think I could do it, and if I could, my understanding of this approach make it unnecessary.

Just like Alice Schroder said in that YouTube, Buffett would say, after his calculations, that this company earns this much and I need this much return, can this company do it; If it can, it would be at this price, is the market offering this company at that price?

 

[McLovin]

Doing such estimates under various scenarios and there is no need to assume x% growth in the future. It is already expected to grow by at least the rate r that you required into eternity. The future might proved it exceed or decline from that requirement... if exceed, the business is worth more etc.

It isn't expected to grow by r. R is the required return.

 

[TPI]

Personally I dont like using any forecasts, when I am using earnings in calculations I only use previous year earnings and I adjust that if its not in line with historical growth.

Fair enough, I guess this is a more conservative approach which also makes sense.

In the end having some relative consistency in your approach is the main thing I think.

 

[TPI]

This is just a terminal value formula that assumes zero growth. Have you played around seen what happens at various risk levels or if you add 1-2% growth? It kind of shows up the pointlessness of that formula.

Also, those one touch formulas are prone to extrapolating the good times in a boom and the bad times in a bust.

When I looked at purchasing a small business I got my accountant to do a valuation.

To do so he looked at current year earnings, as well as the last 3 years history of earnings, made some adjustments for items that were unlikely to appear in the next year (but without extrapolating and forecasting the next year's earnings figure as I am by using forecast consensus EPS figures) to come up with a figure for future maintainable earnings, then divided this figure by the required rate of return (which was largely based on recent comparable industry sales).

If this sort of approach is often used to value small businesses, why can't it be used for listed companies?

 

[McLovin]

When I looked at purchasing a small business I got my accountant to do a valuation.

To do so he looked at current year earnings, as well as the last 3 years history of earnings, made some adjustments for items that were unlikely to appear in the next year (but without extrapolating and forecasting the next year's earnings figure as I am by using forecast consensus EPS figures) to come up with a figure for future maintainable earnings, then divided this figure by the required rate of return (which was largely based on recent comparable industry sales).

If this sort of approach is often used to value small businesses, why can't it be used for listed companies?

How realistic do you think it is that CBA will never, ever grow earnings? Inflation is 2-3% year, so using a no growth model assumes that over time CBA's earning will continue to fall in real terms.

As to why your accountant valued that way, most small business seem to be valued based on payback period with really low multiple p/e (2-4x). So that's a 25%-50% rr. As for the reasons, the earnings are usually less stable, the input of the owner can be critical to the success, the growth opportunities are usually pretty limited, there's no real competitive advantage etc. For that reason they are usually conservatively valued based on low multiple of current earnings.

 

[luutzu]

It isn't expected to grow by r. R is the required return.

Grow for me by at least r, into infinity. Grow or at least maintain its return on equity.

Growth I don't take to be growth in sales or revenue. A good business is expected to also grow its sales, but I'd be perfectly happy if its sales are just at or above inflation, as long as its margins are maintained, and maintained so with little or no debt, as Munger and Buffett say.

I approach it as though I own the entire business. It does me little or no good if my company grow its equity from retained earnings while profitability decline - that just mean business is doing tough and it needs more of my cash, cash it now return at a lower margin, cash that without the company might do much worse.

In the same vein, it does me the owner no good if the company grow its sales but its profitability declines and its balance sheet is stuffed with debts. I'd be fine with sales decline but margin is maintained or at least not deteriorated too much.


the DCF models is fine if apply to assets with predictable cash flows and interest rates/growth rates... As you said, if estimates are wrong by 1 or 2 percent, valuation is very different. Add to that, if inflation is more than expected, if growth is lumpy, or did not materialise in the timeframe as predicted... I'd be in a panic as to what causes what.

I run my own business and it could be just me but I could never estimate how much I'd be making next year or two. But I'm pretty good at predicting my margins though - I try to keep at the same level as previous; though with higher scales, I tend to get higher margins, economies of scale and stuff.

 

[McLovin]

Grow for me by at least r, into infinity. Grow or at least maintain its return on equity.

That's not growth. I think you need to understand how those formulas work, because it seems like you don't really understand them at all.

$x/r will deliver you $x and not a cent more for inifinity. You will still be getting your required "r", because if you buy something for $100 paying $10 then you'll get your 10% return as long as it earns $10. So in the case of CBA, CBA will be earning $5.50 in 10, 20, 30, 50, 100, ∞. With inflation running at 2-3% you are going backward in real terms. That's why I don't see the point of using such a method of valuation. But anyway, you do what you like.

 

[luutzu]

That's not growth. I think you need to understand how those formulas work, because it seems like you don't really understand them at all.

$x/r will deliver you $x and not a cent more for inifinity. You will still be getting your required "r", because if you buy something for $100 paying $10 then you'll get your 10% return as long as it earns $10. So in the case of CBA, CBA will be earning $5.50 in 10, 20, 30, 50, 100, ∞. With inflation running at 2-3% you are going backward in real terms. That's why I don't see the point of using such a method of valuation. But anyway, you do what you like.

OK, "grow" can be confusing... return.

I've reviewed this many times over the years. It's simplistic but that's why I understand and can do it

This formula is the present value of an ordinary perpetuity. So we have a series of cash flows (earnings) stretch into infinity. Yes, if the E and the r remain the same, then we'd have a problem... but we adjust them as we need to.

So I do the work, come to see that earning is E and at the moment my return is r... what is the value of a bond that returns E at r into infinity right now? P= E/r. Then next year or two... due to inflation, due to my greed... the required r is now r+i... What is the E now? If it's the same, then value is now less.

So while it seems like it's fixed, it's not fixed and static because it's reviewed and adjusted as required.

Since the E and the r are reviewed and adjusted all the time... and once you review and adjust these two variables with the changes you see in front of you, and do so when changes are clear and apparent, it results in more accurate estimates than estimating future growth in sales and future interest rates/inflations etc.

Don't know, it's what make sense to me and what I'm comfortable with.

Here's a shorter version of Schroder's speech on what she saw Buffett does:

 

[TPI]

How realistic do you think it is that CBA will never, ever grow earnings? Inflation is 2-3% year, so using a no growth model assumes that over time CBA's earning will continue to fall in real terms.

I see what you are saying McLovin, but my interpretation is that with this approach the valuation in practical terms is made for a particular point in time only, and not meant to include x number of years of growth into the future at y% pa, as luutzu also describes.

Though indirectly you can factor this in by using a slightly lower required rate of return to express a degree of confidence in the future growth prospects of the business.

Just out of curiosity, I would be interested in what intrinsic value you would place on CBA using your own valuation approach?

Also luutzu and galumay, if you are still reading this thread, what intrinsic value would you calculate for CBA using your own approaches?

Thanks!

 

[DeepState]

I see what you are saying McLovin, but my interpretation is that with this approach the valuation in practical terms is made for a particular point in time only, and not meant to include x number of years of growth into the future at y% pa, as luutzu also describes.

Though indirectly you can factor this in by using a slightly lower required rate of return to express a degree of confidence in the future growth prospects of the business.

All valuations are for a point in time. We wish to know what a company is worth right now in order to assess if it is trading cheap/expensive/fair right now. The discussion is about methodology for assessment.

Capitalising current earnings is, broadly, a reasonable thing to do. It's the assumptions beneath it that matter a great deal because so much hangs off the Earnings / FCF / Dividends that are capitalized.

If you are going to use a multiple of E/FCF/D approach, you will be using what is called the Gordon Growth model, ultimately. As McLovin and I have pointed out, the gap between the growth rate of the E/FCF/D and the required rate of return is the primary driver of valuation if you are going to do this.

If you want to hash earnings growth by saying there is no growth, it is an unrealistic portrayal of the vast bulk of companies. However, as you say, you can drop the required return figure such that, in total, the valuation doesn't change.

Is that a realistic thing to do? If CBA is expected to grow at a rate faster than inflation, but you assume zero growth, does the required return actually drop by the growth rate? Absolutely not. In general, you need a higher return to compensate you for risk when the growth rate is zero or negative unless it is just a cash box.

This is why adjustments to the discount rate which are made just to ape the lack of growth rate assumed is incorrect and are merely efforts to disguise an inaccuracy and force an artificial fit. This remains so if you have some notion of a fixed hurdle rate.

As McLovin said, valuations should try to reflect reality. Bending a capitalization formulation out of shape to allow for poor assumptions is not a move in this direction.

McLovin's approach is sound.

 

[herzy]

All valuations are for a point in time. We wish to know what a company is worth right now in order to assess if it is trading cheap/expensive/fair right now. The discussion is about methodology for assessment.

Capitalising current earnings is, broadly, a reasonable thing to do. It's the assumptions beneath it that matter a great deal because so much hangs off the Earnings / FCF / Dividends that are capitalized.

If you are going to use a multiple of E/FCF/D approach, you will be using what is called the Gordon Growth model, ultimately. As McLovin and I have pointed out, the gap between the growth rate of the E/FCF/D and the required rate of return is the primary driver of valuation if you are going to do this.

If you want to hash earnings growth by saying there is no growth, it is an unrealistic portrayal of the vast bulk of companies. However, as you say, you can drop the required return figure such that, in total, the valuation doesn't change.

Is that a realistic thing to do? If CBA is expected to grow at a rate faster than inflation, but you assume zero growth, does the required return actually drop by the growth rate? Absolutely not. In general, you need a higher return to compensate you for risk when the growth rate is zero or negative unless it is just a cash box.

This is why adjustments to the discount rate which are made just to ape the lack of growth rate assumed is incorrect and are merely efforts to disguise an inaccuracy and force an artificial fit. This remains so if you have some notion of a fixed hurdle rate.

As McLovin said, valuations should try to reflect reality. Bending a capitalization formulation out of shape to allow for poor assumptions is not a move in this direction.

McLovin's approach is sound.

Thank you RY. Not having any formal training in accounting, this really clarified the distinction between McLovin and Luutzu's approaches.

However... what is McLovin's (and your) approach? How do you attempt to factor in growth etc in your valuation?

 

[DeepState]

Thank you RY. Not having any formal training in accounting, this is really clarified the distinction between McLovin and Luutzu's approaches.

However... what is McLovin's (and your) approach? How do you attempt to factor in growth etc in your valuation?

No worries. The below is the bread and butter approach. McLovin obviously has a lot of skill and experience, and I'll leave it to him to describe how full bore M&A valuations and corporate-advisory/private-equity valuations work in case that's on your plate.

First, it is very tough and often not thought to be possible to capture the growth path of an equity by a simple, single, growth rate together with a discount rate. To the extent that you think growth rates will not be constant, you can see that a single figure is really an approximation. The more varying the periodic figures are, the worse the approximation. Even normalizing the earnings for capitalization can easily move valuations with similar information loads by +/-10% or more. As such, in many companies, it is worth going for little more detail. These are particularly the ones with fairly steady key elements in P&L, FCF and/or dividends that you can allow for.

The growth rates for different stocks are determined differently. It depends on what drives the stock. Again, everything is an approximation. In the case of CBA, one approach would be to take a look at NOMINAL GDP growth, the credit expansion related to this as a credit system and CBA's share of it. Unless you have amazing insight into macro, it is reasonable to grab estimates off forecaster surveys (incl RBA, Treasury, NAB) for GDP and inflation. You can look at historical relationships for credit and CBA's share. Nothing heroic needs to be done. Then you need to assess margins, capex, fixed costs, tax.... You can get an idea from examining historical accounts and comparables. Voila, you have the basis of a working model.

Don't get scared by this. It is actually a very small number of key assumptions. Once you knock out your first, it becomes pretty clear. Any fundamental analysis requires a look into these variables. Might as well use them to value the firm. You can get these for the most part. Exact is not really important. No-one has exact. Sales, gross margin, fixed costs, capex, tax. Those are the key bits that drive a FCF valuation. You do not need a 200 line model. By doing this you also figure out where key sensitivities are which will help you assess the required margin-of-safety/moat. This clearly differs for different situations. Doing calculations in this way helps you figure out the size of uncertainty you are facing if you cannot get a good handle on the input. This is very important.

You can then fade these in to some sort of steady state after a few years (depending on how long you figure these assumptions get to some sort of steady state - a company never achieves steady state in reality, just at some time, you fade a company into a state where you think it would be in equilibrium and then grows from there at a steady rate. This steady rate is usually below Nominal GDP growth. Not doing so is assuming the company will ultimately become World GDP) and then use the Gordon Growth model in the 'outer' years (this might be around 3-6 years, that is the practice in funds management and broking. That's half to one business cycle for a company's financials to settle down. Longer for resource companies whose mine lives are understood). Discount the whole lot back and you have a valuation that takes market implied major drivers and allows you to figure out how they translate to earnings/FCF/D and discount it at whatever your required rate of return is (normally these are after tax return on 10 year long bond rate + around 5% for equity risk premium although you can adjust this for beta if you like). You can add the value of franking to this as well.

Now, it can be argued that you have so many assumptions that everything is too wobbly to bother with. I think that is a possibility for wildly unpredictable companies. They have cash flows all over the place and no dominant steady driver. These companies are almost impossible to value anyway other than via scenario. A capitalised approach is of no help here either. However, as you get to more stable companies, that argument gradually fades in favour of something like the above or variants thereof.

Rather than using zero growth and some ad hoc, compensatory, adjustment to required rate of return, the above is an attempt to produce a realistic assessment of valuation. To the extent that each parameter brings uncertainty, it was already in the capitalized current E/FCF/D method already. Not looking at it is not exactly making the problem go away. However, the benefit of the slightly more expansive approach is that you can see what's going on and you can make a better effort to model the activities of a company and actually understand what is driving the valuation at a level that offers a viewpoint and offers a chance to get in the ballpark. Importantly, it will help you much more accurately judge what the margin of safety might be. When valuation ranges of a firm like CBA will easily be about 30% wide, this is very important. The width of the estimate uncertainty partly highlights why people who do this have such small numbers of stocks they are prepared to contemplate. Not many will actually look cheap enough to feel comfortable with. Also, the kinds of people who do this buy certain types of companies that lend themselves to this type of approach.

Using a capitalized approach, it is exceptionally hard to figure out what the margin of safety should be because the central estimate is already unknown. What does it mean to vary the 'gap' between zero growth and the adjusted discount rate (however adjusted) by 1%? An adjustment to an adjustment??? That grinds out a massive number of underlying changes to the actual valuation whose interactions are not even visible.

I hope you find the opportunity to take the effort if valuation is a part of your investment considerations.