Cost of equity

[Evan]

In the EVA thread the OP wrote this, "a cost of equity calculation does not require beta, it’s just something that’s historically been used a lot." I thought it was an interesting statement because it's something I haven't thought about a lot.

Cost of equtiy has always evoked the thought of "risk free rate plus beta times the risk premium" for me. Thinking about it there obviously SHOULD be other ways to go about it, but I can't think of any. I'll put some more effort into this later because I feel like I should be able to come up with something reasonable, but I wanted to post it before I forgot.

So, how would you calculate a cost of equity without using a beta? I'd like to stay away from dividend models due to popularity of not paying divdends these days.

WACC

[Evan]

[ QUOTE ]
WACC

[/ QUOTE ]
This is not an answer to my question. Do you know what WACC means?

[midas]

Evan -

WACC is Weighted Average Cost of Capital.

Also, cost of equity is very theoretical and seldomly used in the real world. What are you trying to accomplish?

[Evan]

[ QUOTE ]


WACC is Weighted Average Cost of Capital.

[/ QUOTE ]
I know. That's not an answer to my question

[ QUOTE ]

Also, cost of equity is very theoretical and seldomly used in the real world. What are you trying to accomplish?

[/ QUOTE ]
Nothing really, other than satisfying curiosoty. What's the difference?

WACC = only thing i remember from finance class..lol

[buffett]

[ QUOTE ]
WACC = only thing i remember from finance class

[/ QUOTE ]
Remember that WACC uses Ke as one of its inputs, so it can't possibly be an answer to 'how to calculate Ke w/o beta?'

PS
Last week I reviewed a securities analysis assignment for a friend of mine attending Georgetown. To value Colgate, they used a beta of 0.16 (????), giving them a Ke of around 5.6%, which (when combined with their outrageous EPS growth assumptions) gave an intrinsic value for CL of around $400 per share. Awesome.

[FatOtt]

Charlie Munger made a comment in a recent annual meeting (2003?) that was something like, "At it's core, economics all boils down to opportunity cost." That's what I think about when I think about cost of equity.

The cost of equity is going to be the rate of return that investors require from holding a particular equity instrument. That's unobservable, obviously. What you can do is look at economically similar instruments to get an idea for what that number should be.

Beta is a historically popular number (obviously not popular with Berkshire Hathaway). The most common beta calculation comes from a regression of firm excess returns on market excess returns. The assumption is that there is a common market risk factor and you should earn the risk premium to the extent that the individual stock shares that risk factor.

But academics define lots of different betas, rather than the market beta. A beta is basically just the exposure of a firm to any particular risk factor. In the CAPM case, the singular beta is the exposure to the market factor. You can open that up to multifactor models, where each beta represents exposure to a particular risk factor. (The Fama-French three factor model, perhaps).

So maybe you're not satisfied with betas. You could say, "Screw the math, I just want to figure out my cost of equity." Here are some starting thoughts:
1. It must be higher than the risk-free rate
2. It must be higher than any available instrument with less risk (however you define that)
3. It should be equal to your required equity rate for instruments with similar risk

So you might start by saying, "I need the historical market's rate of return." But that's not exactly true because what you really need is the historical market premium plus the current risk-free rate. But that's not necessarily true because you can invest in an index fund (that many people would describe as less risky) and get that rate. So you need the current real risk free rate, plus expected inflation, plus the risk premium, plus some premium for holding an individual stock rather than an index.

How do you do it? I don't know - it's obviously messy. Using the math-oriented beta is just one way. Maybe you just know that you need at least a 10% return when you invest in something - that's your hurdle rate and therefore your cost of equity. The best estimate you can get for a firm's cost of equity (in my view) is to look at the implied cost of equity in the current share price. Make your best estimate of future free cash flows (or dividends, if available) and figure out the IRR that sets the present value of those cash flows equal to the current stock price. Do that for a number of different firms and maybe that's the current equity cost of capital.

It's an unsatisfactory answer, right? There's no good way to calculate the cost of equity in a rigorous fashion, other than saying it's the opportunity cost across all possible investment opportunities with the same degree of risk.

Which, by the way, comes back to provide some support for the historical beta. Usually when I'm talking to someone about beta, they hate it - just like you do. They usually hate it a little less when I say the following:

When you're buying a stock, imagine that you're only choosing to buy the entire market index or that stock. You can't compare the cost of equity for the entire index to the cost of equity for the individual stock - the index is less risky. But you can adjust for this! If you want to buy the index in such a way to equal the risk inherent in an individual stock, you can lever up or lever down the amount purchased. You can roughly match the risk of a levered index position to the risk of an individual stock by levering up or down until the volatility of the levered position is equal to the volatility of the stock position. Therefore, you should require a return on the individual stock equal to at least the levered return of the index. What's that? Beta is the amount of the leverage, so your required return is beta times the market's excess return. If you can't get at least that from investing in the individual stock, you're better off investing in the levered market position.

I don't know if that's satisfying to you, but it makes me think there's at least some value in beta even though I don't buy into it completely.

[Evan]

[ QUOTE ]
The cost of equity is going to be the rate of return that investors require from holding a particular equity instrument. That's unobservable, obviously.

[/ QUOTE ]
I think you might be stretching a bit here. This may be true, but I don't think you can say it's obvious.

[ QUOTE ]
In the CAPM case, the singular beta is the exposure to the market factor.

[/ QUOTE ]
You can always deefine "the market" in a different way if you want. It doesn't have to be hte S&P 500.

[ QUOTE ]
So maybe you're not satisfied with betas. You could say, "Screw the math, I just want to figure out my cost of equity." Here are some starting thoughts:
1. It must be higher than the risk-free rate
2. It must be higher than any available instrument with less risk (however you define that)
3. It should be equal to your required equity rate for instruments with similar risk

[/ QUOTE ]
Points 2 and 3 are sort of repetative and essentially restatements of the security market line.

[ QUOTE ]
what you really need is the historical market premium plus the current risk-free rate. But that's not necessarily true because you can invest in an index fund (that many people would describe as less risky) and get that rate. So you need the current real risk free rate, plus expected inflation, plus the risk premium, plus some premium for holding an individual stock rather than an index.

[/ QUOTE ]
Why do you need a premium for holding an individual stock? Obviously if the stock is riskier than the market as a whole you need a premium for that, but it's got nothing to do with the number of securities. You don't get rewarded for systematic risk beyond the market risk premium. Determining the cost of equity is basically an exercise in evaluating unsystematic risk.

[ QUOTE ]
The best estimate you can get for a firm's cost of equity (in my view) is to look at the implied cost of equity in the current share price.

[/ QUOTE ]
Woah. You're implying a lot of assumptions about efficient markets in that statement.

[ QUOTE ]
Do that for a number of different firms and maybe that's the current equity cost of capital.


[/ QUOTE ]
Are you saying that all firms' cost of equity should be the same?

[ QUOTE ]
There's no good way to calculate the cost of equity in a rigorous fashion, other than saying it's the opportunity cost across all possible investment opportunities with the same degree of risk.

[/ QUOTE ]
Fair enough, what is another way that is AS GOOD AS a historical levered beta or a built up beta?

[ QUOTE ]
Which, by the way, comes back to provide some support for the historical beta.

[/ QUOTE ]
I think a bult up beta is much more "reliable" than a historical beta.

[ QUOTE ]
Usually when I'm talking to someone about beta, they hate it - just like you do.

[/ QUOTE ]
I don't really hate it. And to some extent I even like built up betas in that I think they're reasonable ways to estimate the opportunity cost we're looking for. I just think it would be interesting to find other ways of doing it.

[ QUOTE ]
You can't compare the cost of equity for the entire index to the cost of equity for the individual stock - the index is less risky.

[/ QUOTE ]
This is wrong. There is no way "the index" is always less risky than an individual stock.

[adios]

[ QUOTE ]
So, how would you calculate a cost of equity without using a beta?

[/ QUOTE ]

I'm assuming from your post that you're referring to a single measure of beta used in the CAPM. Actually some equity is notorious for being highly risky but low beta. Gold stocks in particular come to mind. MREITs are another example IMO. I think FatOtt alluded to an arbitrage pricing model where several risk factors are weighted to determine the cost of equity. So.......... it depends on the firm in question. Something like PG, GE, MSFT the CAPM becomes a decent way to arrive at the cost of equity. Other firms that are dependent on the value of a single commodity or highly interest rate sensitive, perhpas a different pricing model is better. Gold stocks and MREITs have relatively low R-Squared values so I would opine that in evaluating whether or not the cost of equity should be determined by the CAPM, one should look at the R-Squared.

[Sniper]

[ QUOTE ]
This is wrong. There is no way "the index" is always less risky than an individual stock.

[/ QUOTE ]

In general, the risk in owning an index is less than the risk of holding a single stock, due to diversification reducing the overall risk!

That said, some indexes are overdiversified, and you can achieve similar diversification risk reduction by holding a few select stocks.

[Evan]

[ QUOTE ]
[ QUOTE ]
This is wrong. There is no way "the index" is always less risky than an individual stock.

[/ QUOTE ]

In general, the risk in owning an index is less than the risk of holding a single stock, due to diversification reducing the overall risk!

That said, some indexes are overdiversified, and you can achieve similar diversification risk reduction by holding a few select stocks.

[/ QUOTE ]

Okay, in general it may be true, but not often enough to make the blanket statement that indices are less risky than stocks.

The beta of a market is 1, it is not true that no stock's beta is <1.

[adios]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
This is wrong. There is no way "the index" is always less risky than an individual stock.

[/ QUOTE ]

In general, the risk in owning an index is less than the risk of holding a single stock, due to diversification reducing the overall risk!

That said, some indexes are overdiversified, and you can achieve similar diversification risk reduction by holding a few select stocks.

[/ QUOTE ]

Okay, in general it may be true, but not often enough to make the blanket statement that indices are less risky than stocks.

The beta of a market is 1, it is not true that no stock's beta is <1.

[/ QUOTE ]

I made post awhile back about evaluating the risk in investing all of ones money in a particular stock. FatOtt and I had a good exchange of views on what I wrote. I know you know this already but for others that may be following along in this thread, since individual company risk can be "diversified away," there is no risk premium for assuming individual company risk. To use my MREITs and Gold stocks example again, these stocks generally speaking have low betas but there is no way in the world that these stocks are lower risk than something like SPY. This is due to the fact that these stocks have low R-Squared values which mean that they're not highly correlated to the overall market and thus the beta's are a poor measure of risk.

[Evan]

You already kind of said this, but I'll reitterate it; beta is not a metric of risk in the sense most people might think. It measures unsystematic risk that cannot be diversified away. Those are NOT the same thing. Companies with low betas being risky isn't really the oxymoron many people think it is.

[FatOtt]

Evan,
I think we agree on most everything, but I'm confused about this post.
[ QUOTE ]
[ QUOTE ]

The cost of equity is going to be the rate of return that investors require from holding a particular equity instrument. That's unobservable, obviously.

[/ QUOTE ]

I think you might be stretching a bit here. This may be true, but I don't think you can say it's obvious.

[/ QUOTE ]

It's not obvious that the required cost of equity is unobservable? What kind of situation are you talking about? What kind of observable measures of costs of equity are you thinking about? I still think it's pretty obvious that costs of equity are completely unobservable, short of taking a poll of investors (and even then you're in the world of distinguishing between stated preferences and revealed preferences).

[ QUOTE ]
[ QUOTE ]

In the CAPM case, the singular beta is the exposure to the market factor.

[/ QUOTE ]


You can always deefine "the market" in a different way if you want. It doesn't have to be hte S&P 500.

[/ QUOTE ]

I don't know what this refers to - I don't think I mentioned the S&P 500.

[ QUOTE ]
[ QUOTE ]
what you really need is the historical market premium plus the current risk-free rate. But that's not necessarily true because you can invest in an index fund (that many people would describe as less risky) and get that rate. So you need the current real risk free rate, plus expected inflation, plus the risk premium, plus some premium for holding an individual stock rather than an index.


[/ QUOTE ]

Why do you need a premium for holding an individual stock? Obviously if the stock is riskier than the market as a whole you need a premium for that, but it's got nothing to do with the number of securities. You don't get rewarded for systematic risk beyond the market risk premium. Determining the cost of equity is basically an exercise in evaluating unsystematic risk.

[/ QUOTE ]

I believe you need a premium for holding an individual security rather than an index (talking about your only holding here, not talking about a basket of individual securities) because you need to be compensated for the total risk, both systematic and unsystematic, whereas you can diversify away part of that risk by buying an index. I'm not sure what you're arguing here. Is it possible that a firm's total risk will be less than an index's total risk? Obviously, if the firm has a beta of less than 1, it's theoretically possible, but it doesn't seem very likely for a normal, publically-traded operating firm (as opposed to a firm that just owns some treasury notes).

[ QUOTE ]
[ QUOTE ]

The best estimate you can get for a firm's cost of equity (in my view) is to look at the implied cost of equity in the current share price.

[/ QUOTE ]


Woah. You're implying a lot of assumptions about efficient markets in that statement.

[/ QUOTE ]

Yeah, I have no idea what you're talking about here. I'm not making any assumptions about efficient markets. Empirically, the IRR that equates future cash flows to the current stock price is the firm's current cost of equity. Are you saying that if the stock prices drops by 50% without a corresponding drop in the expected future cash flows, the firm doesn't face a higher cost of equity than they did prior to the stock price?

[ QUOTE ]
[ QUOTE ]
Do that for a number of different firms and maybe that's the current equity cost of capital.

[/ QUOTE ]



Are you saying that all firms' cost of equity should be the same?

[/ QUOTE ]


Sorry, I meant to say different but similar firms. I don't at all believe that all firms' cost of equity should be the same.

[ QUOTE ]
Fair enough, what is another way that is AS GOOD AS a historical levered beta or a built up beta?

[/ QUOTE ]

This is the crux of the matter, right? Because costs of equity are ultimately unobservable (I'm still interested in why you think otherwise), there isn't a validation that you can do. What would you use to determine whether one method is better than another at arriving at a cost of equity? The whole discussion/problem comes about because there's no agreed-upon way of doing it. There's not even any after-the-fact verification because ex ante estimates do not translate directly to ex post results.

I'm surprised you had so much disagreement with what I wrote, considering I agree with most of what you're saying. Just not the observability of costs of equity.

[Evan]

[ QUOTE ]
It's not obvious that the required cost of equity is unobservable? What kind of situation are you talking about? What kind of observable measures of costs of equity are you thinking about? I still think it's pretty obvious that costs of equity are completely unobservable, short of taking a poll of investors (and even then you're in the world of distinguishing between stated preferences and revealed preferences).

[/ QUOTE ]
I guess I took "observable" to mean something more along the lines of "attainable". Even still, I think there's a case for a regression beta to be an observable cost of equity.

[ QUOTE ]
I don't think I mentioned the S&P 500.

[/ QUOTE ]
You didn't. The point was that if you don't feel that "the market" reflects a good benchmark for a stock you don't have to use the standard. I wasn't really correcting you, just expanding on your point.

[ QUOTE ]
I believe you need a premium for holding an individual security rather than an index (talking about your only holding here, not talking about a basket of individual securities) because you need to be compensated for the total risk, both systematic and unsystematic, whereas you can diversify away part of that risk by buying an index. I'm not sure what you're arguing here. Is it possible that a firm's total risk will be less than an index's total risk? Obviously, if the firm has a beta of less than 1, it's theoretically possible, but it doesn't seem very likely for a normal, publically-traded operating firm (as opposed to a firm that just owns some treasury notes).

[/ QUOTE ]
You don't get rewarded for diversifiable risk because you can diversify it away. Therefore your cost of equity does not increase due to diversifiable risk.

[ QUOTE ]
Yeah, I have no idea what you're talking about here. I'm not making any assumptions about efficient markets. Empirically, the IRR that equates future cash flows to the current stock price is the firm's current cost of equity. Are you saying that if the stock prices drops by 50% without a corresponding drop in the expected future cash flows, the firm doesn't face a higher cost of equity than they did prior to the stock price?

[/ QUOTE ]
Basically you're assuming that all sotcks are fairly valued at the moment, right? That assumes that markets are efficient. You have two inputs (discount rates, cash flows) and one output (value). You have to start with two known variable to calculate the third. If you start with cash flows and value as givens then you're assuming the stock is fairly valued in order to find the true discount rates. If you assume that all stocks are always fairly valued then you assume markets are efficient.

To asnwer your last question, no, the lower stock price does not necessarily mean the cost of equity is higher.

[FatOtt]

[ QUOTE ]
Even still, I think there's a case for a regression beta to be an observable cost of equity.


[/ QUOTE ]

Regression betas over many, many firms over long periods of time might give you average betas that are unbiased. That assumes a rational expectations environment where, on average, investors' expectations of future returns are true. That's a fairly strong assumption - even for a guy living in Famaland, Chicago, like I do.

[ QUOTE ]
You don't get rewarded for diversifiable risk because you can diversify it away. Therefore your cost of equity does not increase due to diversifiable risk.

[/ QUOTE ]

Edited to say that I have potentially screwed up the issue and it's probably lead to confusion. There are two things going on: What I require for holding a stock and what the market will give me. If I were to hold only one individual stock in my portfolio, I would require an expected return greater than I would require for holding the index. The market will not give me that return because, by diversifying, they can reduce the risk. I'm talking about what I would require for an individual stock return whereas you seem to be jumping the gun to point out that the market would not give me that return (a point that I agree with). If we look at it from the perspective of the firm, if the market doesn't reward the investor for holding diversifiable risk, the firm's cost of equity shouldn't reflect that. At this point, I can't remember if we're talking about:
- ways that an investor would calculate his required rate of return, or
- ways that a firm would calculate its cost of equity

[ QUOTE ]
Basically you're assuming that all sotcks are fairly valued at the moment, right? That assumes that markets are efficient. You have two inputs (discount rates, cash flows) and one output (value). You have to start with two known variable to calculate the third. If you start with cash flows and value as givens then you're assuming the stock is fairly valued in order to find the true discount rates. If you assume that all stocks are always fairly valued then you assume markets are efficient.

To asnwer your last question, no, the lower stock price does not necessarily mean the cost of equity is higher.

[/ QUOTE ]

No - I am assuming that all stocks are valued, period. If the market irrationally prices a firm's shares so that the price is 10% of the "Buffettesque Intrinsic Value", that firm does face a higher cost of equity. Why? Because when they go to issue shares, they have to issue shares into the environment that is underpricing the equity. By issuing shares when the stock is extremely undervalued, the firm is incurring a very high cost of equity.

It's easier to see in the case of debt because the debt rates are observable. If the firm issues a bond into an environment that prices the bond at par value, the (marginal) cost of debt will be equal to the coupon rate of that bond. If the market values that debt at 20% of par, the cost of debt is much higher - that's true regardless of whether the 80% haircut was rational or not.

By the way, have you ever been to a Berkshire Hathaway annual meeting? If not, you should go - it's a good time. You should also get a group of students (I think you're in school) to go visit Buffett in Omaha.

[Evan]

[ QUOTE ]
Regression betas over many, many firms over long periods of time might give you average betas that are unbiased. That assumes a rational expectations environment where, on average, investors' expectations of future returns are true. That's a fairly strong assumption - even for a guy living in Famaland, Chicago, like I do.

[/ QUOTE ]
This is pretty much what a built up beta is.

[ QUOTE ]
What I require for holding a stock and what the market will give me. If I were to hold only one individual stock in my portfolio, I would require an expected return greater than I would require for holding the index.

[/ QUOTE ]
Okay, but you wouldn't just be holding the one tock and you COULD diversify. The "true" cost of equity for that firm should assume diversification is possible. The one stock portfolio was just an example used to demonstrate what a cost of equity is, not a real life scenario.



I don't think your cost of debt analogy is correct. Cost of debt is an explicit rate (although not when there is no recently traded or issued debt) while cost of equity is largely implicit. The price:rate ratio doesn't really hold or equity. It doesn't really hold for debt either, which is why people often turn to things like interest coverage ratios to impute accurate discount rates for debt when no recent transactions are available. The quoted interest rates on debt can be VERY different from what the firm would have to pay if it were to issue debt in the present.

[ QUOTE ]
By the way, have you ever been to a Berkshire Hathaway annual meeting? If not, you should go - it's a good time. You should also get a group of students (I think you're in school) to go visit Buffett in Omaha.

[/ QUOTE ]
Never been. You have to own a share to attend, correct? I've thought about buying a class B share just to go to the meeting, but I haven't done it yet. I am in school at NYU, so the group idea is a good one.

[buffett]

[ QUOTE ]
By the way, have you ever been to a Berkshire Hathaway annual meeting? If not, you should go - it's a good time. You should also get a group of students (I think you're in school) to go visit Buffett in Omaha.

[/ QUOTE ]
Are you one of the people behind The Buffett Blog??? Were you part of the most recent group of Chicago kids that went to Omaha and did the Borsheims/NFM/BRK tour?

Evan, I totally agree....you need to organize a group of NYUers and take them to Omaha. (Plus, you also need to attend the BRK AGM at least once while they're both still alive. Forget about the fact that WEB has already said everything's he's going to say and you're not going to learn anything....go for the experience and the camaraderie.) I went in October 2003, and it was easily the highlight of my two years of school.

[Sniper]

[ QUOTE ]
Okay, in general it may be true, but not often enough to make the blanket statement that indices are less risky than stocks.

The beta of a market is 1, it is not true that no stock's beta is <1.

[/ QUOTE ]

In the way that you used it, Beta represents correlation to the market.

Thus, while it can be used to measure the level of market risk; it can not be used to measure individual stock risk.

I suggest you spend some time on riskgrades (http://www.riskgrades.com/)

[Evan]

[ QUOTE ]
[ QUOTE ]
Okay, in general it may be true, but not often enough to make the blanket statement that indices are less risky than stocks.

The beta of a market is 1, it is not true that no stock's beta is <1.

[/ QUOTE ]

In the way that you used it, Beta represents correlation to the market.

Thus, while it can be used to measure the level of market risk; it can not be used to measure individual stock risk.

I suggest you spend some time on riskgrades (http://www.riskgrades.com/)

[/ QUOTE ]
I know beta is not a measure of an individual stock's risk. I didn't mean to imply that. It was a seperate but somewhat related statement, just in case people reading this thread weren't aware that betas can be <1.

[FatOtt]

Evan,
Have you had Damodaran at NYU? I was just flipping through one of his books today. How is he as a teacher?

Anyway, I'll stick to the part that I think we disagree on: the idea that market pricing does not determine cost of equity (or debt). The cost of equity is going to be how much the firm has to pay to acquire equity funding. When that firm is faced with a marginal project, how much will they be charged to finance that project via equity funding?

Looking at the stock price and (assuming you can estimate them) the expected future cash flows to equity holders will tell you what discount rate is currently implied by the stock price. If a firm were to issue new shares to finance the project, that's the discount rate that would presumably be applied to those shares as well.

I'm really surprised to see you disagree with the cost of debt. Suppose a firm issues a bond that specifies 10 annual payments of $4 million, plus 1 payment of $50 million at the end of 10 years. Looking at the price of that bond (the proceeds) will tell you the firm's cost of debt. If the firm receives $76,947,755, their cost of debt was 2%. If the firm receives exactly $50,000,000, their cost of debt was 8%. If the firm receives exactly $38,699,554, their cost of debt was 12%. It doesn't matter if buyers of the bonds were perfectly rational economic agents or if they were calculating the value of the bonds by multiplying their shoe size by their hat size - the proceeds from the bond issuance determine the cost of debt. How could it be otherwise?

The implied cost of debt may not be observable for illiquid offerings, but the theory is still the same - the discount rate that equates future cash flows to the current price is that firm's cost of debt (at least for that issuance).

Getting to the Berkshire meeting isn't hard. Every shareholder gets 4 invitations, so you can grab one from someone else who owns it. Because people were starting to sell them online last year, Berkshire started selling tickets on ebay to non-shareholders. $5 apiece, I think.

[Evan]

[ QUOTE ]
Have you had Damodaran at NYU? I was just flipping through one of his books today. How is he as a teacher?

[/ QUOTE ]
He only teaches grad classes, so I haven't had him yet. I've sat in on some of his classes and he's given a few presentations to a club I'm in. He is exceptionally good at explaining difficult ideas in ways that make them easy to understand.

[ QUOTE ]
When that firm is faced with a marginal project, how much will they be charged to finance that project via equity funding?

[/ QUOTE ]
Okay, I can agree with that. But does that means that's the rate you should be using to value the company? If it is then every company you value will be excactly fairly valued because you're taking the market price as a given. I'm not saying that's not the current market cost of equity, so I guess we agree. However, that is not necessarily the rate you should discount future cash flows at. I think you have to agree with that, right?

[ QUOTE ]
I'm really surprised to see you disagree with the cost of debt.

[/ QUOTE ]
Your example is correct when examined in a vacuum. The problem is that bonds are issues/traded much less frequently than stocks. If there was a debt issue or trade very recently then it is fine to use market quotes. In many instances though, there may have been months or years since any market data was quoted.

A good example of when market interest rates have been useless is Salton. About a year ago they went through massive refinancing after taking huge (see HUGE) writedowns and restructuring charges that equated to losses of about $8.50/share on a stock trading for ~$6 at the time. Obviously this was more an example of "big bath accounting" than anything else, but the point is they were doing very, very poorly. During this time their market quoted cost of debt remained the same, even though they were obviously sucking wind pretty hard (they had a TON of debt at this time, so it's not like these rates were immaterial). Next, they defaulted. For those that aren't familiar with finance, this is sort of like bancruptcy, excpet that in this case it was a technical default. When you borrow huge amounts of money, as Salton had, you get convenants put on your debt like, "you have to make $x/year in operating income". Needless to say, Salton fell WELL below the required interest coverage ratio due to the writedowns and restructuring. Here's the twist, they lose a bunch of money, default on their debt, but their cost of debt doesn't move up at all. Why? Because the debt holders really had no leverage to do anything to the company. Management was already getting fired, there were new and promising products comign to market, and it looked like things could turn around. Salton didn't issue any new debt and no one wanted to buy their old debt (and the current debt holders didn't really want to sell since they'd get next to nothing for it anyway).

I know that got a little longwinded, but the point is that market rates, for debt or equity, really don't imply as much as you might think about future rates that should be applied to future cash flows.

[FatOtt]

[ QUOTE ]
Okay, I can agree with that. But does that means that's the rate you should be using to value the company?
If it is then every company you value will be excactly fairly valued because you're taking the market price as a given. I'm not saying that's not the current market cost of equity, so I guess we agree. However, that is not necessarily the rate you should discount future cash flows at. I think you have to agree with that, right?

[/ QUOTE ]

Ok, so now we're back at the distinction between:
1) The cost of equity that the firm faces: that will change with the market price of the firm's stock. When the firm's stock price goes down (assuming consistent expected future cash flows), the firm's cost of equity goes up.

2) The cost of equity that you, the investor, would use to value the firm to determine if you believe it's over or undervalued.

At this point, frankly, I'll just stop and say what I do. If I'm trying to value a firm, I estimate future cash flows and calculate the discount rate that equates those future cash flows to the current stock price. If that rate is satisfactory to me, then I may invest. So, for example, I might think that DLTR's future cash flows appear to be discounted at somewhere between 15-20%. That's an acceptable return for me, so I might be DLTR. (I do, in fact, own DLTR.) If that wasn't an acceptable way of doing it, then you could use a beta calculation to estimate the cost of equity you think is appropriate. I agree that using the implied cost of equity to value the firm would be circular - you'd always get the current market price.

-------
I don't understand your Salton example at all. You state that "During this time [of huge writedowns and bad performance] their market quoted cost of debt remained the same", but then you also say "Here's the twist, they lose a bunch of money, default on their debt, but their cost of debt doesn't move up at all. Why? Because the debt holders really had no leverage to do anything to the company. Management was already getting fired, there were new and promising products comign to market, and it looked like things could turn around. Salton didn't issue any new debt and no one wanted to buy their old debt (and the current debt holders didn't really want to sell since they'd get next to nothing for it anyway). "

That bolded part is exactly what I'm talking about. If the market value of the firm's bonds significantly decreases, that represents an increase in the firm's cost of debt. You can look at the quoted prices of the firm's bonds and calculate the YTM of those bonds, which represents the firm's cost of debt.

Here's some evidence:
In their 2004 10-K (for the year ended July 3, 2004), Salton disclosed that the fair value of their senior subordinated debt was $225.3 million, compared to $274.9 million a year earlier. That significant drop in the fair value of the debt represents a significant increase in the firm's cost of debt - the discount rate (or YTM) that equates the future cash flows to the price of the bonds is the firm's cost of debt.

What is your evidence of the market value of the bonds not changing?


Edited to add:
I just looked up some bond rating changes for Salton. It looks like they had to big issues outstanding recently, one maturing in December 2005 and one maturing in April 2008. The December 2005 bond had the following rating changes (from S&P):
9/25/2003: B-
2/11/2004: CCC+
5/11/2004: CCC-

The ratings for the 2008 maturity bond have the same downward pattern from late 2003 to mid-2004. I think this is more evidence that the firm's cost of debt did, in fact, increase over time as their performance declined. You will admit that a firm's cost of debt is increasing as their credit rating decreases, right?

[Evan]

[ QUOTE ]
What is your evidence of the market value of the bonds not changing?

[/ QUOTE ]
Quotes from Bloomberg at the time. I don't remember the exact numbers, they did change slightly, but not nearly as much as they "should" have. This company had negative earnings almost 1.5x their market cap! Let's say your numbers were the market prices, that's a change of 22% or about equivalent to a downgrade from BBB to BBB-. Do you really think that's all the true cost of debt changed?

[FatOtt]

Sorry - we crossed streams. Take a look at my edit. Their debt ratings changed substantially during that time.

[FatOtt]

[ QUOTE ]
Let's say your numbers were the market prices, that's a change of 22% or about equivalent to a downgrade from BBB to BBB-. Do you really think that's all the true cost of debt changed?

[/ QUOTE ]

All right, I just did some quick calculations. First, the company is claiming in their 10-K that the figures I gave were the market values of their structured debt (estimated based on dealer quotes). So I think it's reasonable to say they actually are the market values.

Let's take a look at the implied YTM of these items. We're looking at two notes, one issued on 12/16/98 (face value of $125 million, maturity 12/15/2005, coupon 10.75%), one issued on 4/23/2001 (face value of $150 million, maturity 4/15/2008, coupon 12.25%). I assumed that coupon payments were made annually on the anniversary date, so that coupon payments of $13.4375 million were made annually on 12/15 for the first issue and coupon payments were made annually on 4/15 for the second issue.

At 7/1/2003, when the fair value (market value) of these bonds totaled $274.9 million, I calculate a weighted yield to maturity of 11.86% for the two bonds.

At 7/1/2004, when the fair value was $225.3 million, I calculate a weighted yield to maturity of 21.79% for the two bonds.

That's what I'm talking about when I say that the firm's cost of debt increased significantly. Also, these are relatively short-term maturities. The first issue expires about 18 months after the 10-k, while the second one expires about 4 years after the 10-k. I suspect that change in cost of debt would be much more substantial for longer-duration instruments.

I don't mean to be argumentative, I just happen to be interested in this stuff.

[Evan]

[ QUOTE ]
Edited to add:
I just looked up some bond rating changes for Salton. It looks like they had to big issues outstanding recently, one maturing in December 2005 and one maturing in April 2008. The December 2005 bond had the following rating changes (from S&P):
9/25/2003: B-
2/11/2004: CCC+
5/11/2004: CCC-

[/ QUOTE ]
Okay, so this basically substantiates what I'm saying. The point is that they never paid these rates. Their public debt stayed at rates relative to the B- rating.

They may well be paying this rate on new debt. I haven't looked at the company is over a year.

[ QUOTE ]
You will admit that a firm's cost of debt is increasing as their credit rating decreases, right?

[/ QUOTE ]
Yes, but the rate they pay doesn't always change accordingly. Usually it does; I'm not trying to rework the entire economy, just saying that quoted interest rates do not always equal cost of debt.

[Evan]

Okay man, you're obviously doing your homewokr here. I think we're getting too cuaght up in an irrelevant example and losing track of the real discussion. Maybe a more direct example will be better

Let's say a company has ZERO on balance sheet debt. None, nothing, interest expense is zero. But, they have a ton of operating leases. This company will have no bond rating and no publicly traded debt to quote. This does NOT mean that their cost of debt is zero. A good example of such a company is Bed Bath and Beyond (although I don't think they have zero debt, it's very very low and their operating leases are MUCH bigger).

My only point is that market quotes are not a good way to attain your cost of debt. At times they will be right, when recetn trades/issues were long ago they will not be. When companies don't have public debt they will not be. That is all I'm trying to say.