Session 12: Estimating Hurdle Rates - Debt & its Cost

Profit Playbook
12 followers
In session 12 of this 36-session corporate finance class,
I'd like to lay the foundations for the second way of raising financing,
which is to borrow money. I'll start off by defining what I mean by debt,
and then attach a cost to that debt as a precursor to coming up with the cost of capital.
These last few sessions, we've talked about estimating the cost of equity.
Started with the risk-free rate, we came up with an equity risk premium,
and then we went through the process of estimating betas for companies,
depending on the businesses they were in. All of that, though, was to give you the cost
of one component in your capital, which is equity. In this session, I'd like to move on
to the second component, which is debt. I want to talk about what goes into debt
and what the cost of debt should be. We need that to get to a cost of capital.
So let's set the table. To get from cost of equity to cost of capital,
we need to bring in how much money you borrow. So we have to define debt.
We have to attach a cost to that debt to convert into a cost of capital.
So the first of those questions is, what is debt? That sounds like a pretty simple question,
right? You should be able to look at a balance sheet and make that judgment.
Unfortunately, it's not that simple. There are three criteria I go looking
for to classify something as debt, and here are the three criteria, or three characteristics.
The first is you have to have a contractual commitment to make payments, and that's what
separates debt from equity. When you issue shares, people might assume that you will pay
dividends. They might even expect you to pay a dividend, but you're not contractually obligated
to pay a dividend. So if you don't pay a dividend, they can't sue you. If you borrow money,
you enter into a contractual commitment to make interest-in-principle payments.
Second, in most parts of the world, that contractual payment that you make on debt
is tax-deductible. And third, if you fail on your contractual commitment, bad things happen to you.
You lose control of the firm. Somebody else gets voting rights. So those are the three
characteristics, fixed payment, tax-deductible, loss of control. Those are the three characteristics
I would use to separate debt from non-debt. So let's try a test. Let's go through a typical
balance sheet, and let's think of whether items pass this test. Bank loans, obviously debt.
Corporate bonds, obviously debt. Short-term bank loans, absolutely debt. Here's where you get
tricky. If you look at accounts payable as supplier credit, well, they don't quite pass
muster as debt, and here's why. Neither is explicit interest expenses. Notice the words
I used, explicit. When you take on, when you supply a credit, you often lose discounts, right,
because that supplier will offer you a discount to pay early. So there is an implicit interest
expense, and if you're willing to make that implicit expense explicit, in other words,
if you're willing to go into your cost of goods sold and tell me how much of those costs of goods
sold is lost discounts, then I'm willing to consider accounts payable and supplier credit
as debt. But if you cannot do that, you don't have the information, you're unwilling to go
and take that step, then don't treat accounts payable, supplier credit, or any other non-interest
bearing liability as debt on your balance sheet. So any interest bearing liability I'll count as
debt, short-term as well as long-term. But here's one item that's off the balance sheet that I think
we should consistently treat as debt. If you're a retailer or a restaurant and you open a new
store or a new restaurant, you often lease the space, right? You enter into a contractual agreement
where you make lease payments every period or agree to make those payments every period for
the next 10 or 12 years. That's a contractual commitment, and it's usually tax-deductible,
and if you fail, initially you lose that leasehold, but if you fail on a bunch of leases at the same
time, you go bankrupt. So I would treat lease commitments as debt as well. Operating leases
or capital leases might be a distinction that accountants make, but to me all lease commitments
are debt. We'll talk about how to convert them into debt in a few minutes, but that's the first
part of the process is defining debt. Now let's talk about the cost of debt. Here's what you'd
like to measure. The cost of debt is the rate at which you can borrow money long-term today.
Two key words there, long-term. First, I don't care how short-term your debt is. The fact that
you use commercial paper is interesting, but when I talk about the cost of debt, I'm going to attach
a long-term cost to all your debt even if it's short-term debt. That might strike you as unfair
if you're a company that uses a lot of short-term debt and that short-term debt looks cheaper,
but here's why I'm going to attach a long-term cost to it. I need to make sure that your investments
make more than the rolled over cost of the debt, not just what the rate might be for the next six
months, but that rolled over cost and the long-term cost is a good proxy for that rolled over cost.
Why today? Because I don't care what you borrowed money at two years ago or five years ago.
The cost of debt is the rate at which you can borrow money long-term today. Now defined thus,
there are only two components in your cost of debt. The first is a risk-free rate. Nobody
should be able to borrow at less than the risk-free rate. The second is a default spread.
That's a spread I'm going to add on to the risk-free rate because I see credit risk in your
company. The essence of estimating cost of debt is coming up with a good measure of that default
spread. It's easy in some cases and it's easy because somebody else does the dirty work for you.
Well, I'm talking about the fact that some companies have ratings, bond ratings, estimated
by S&P or Moody's or some other ratings agency. What you might do is use those ratings as a proxy
for default risk and then use that proxy to come up with the default spread and move on.
If you cannot find a rating for your company though, you have to estimate the default spread
and you have to come up with a measure of how risky your company is. It's not that difficult,
but in doing all of this, remember, stick with the same currency that you estimate your cost
of equity in. You cannot estimate your cost of debt in a different currency and expect your
cost of capital to make sense. So let's take the easy route first. Let's look at the three
companies in my sample for which I have a rating. Tata Motors doesn't have a rating.
BookScape does not have a rating, but Valet, Disney and Deutsche Bank do. Disney and Deutsche
Bank have single A ratings and these are S&P ratings and Valet is an A- rating. How does
this help me? At least in November of 2013, the typical default spread for an A-rated bond was
about 1% because I'm going to estimate a cost of debt for Disney and Deutsche Bank. I'm going to
add 1% onto the risk-free rate. In the case of Disney, that risk-free rate is a U.S. dollar
risk-free rate because I'm estimating everything in U.S. dollars. That gives me a pre-tax cost of
debt for Disney of 3.75%. For Deutsche, I'm going to add that same 1% default spread to the euro
risk-free rate, which is 1.75%, which yields a cost of debt for Deutsche of 2.75%. Valet's A-
rating has a default spread of 1.3% associated to it. Since I'm doing everything in U.S. dollars
for Valet, I'm going to add the 1.3% to the risk-free rate of 2.75%, the U.S. dollar risk-free
rate, to come up with a U.S. dollar cost of debt of 4.05%. Now, let's say you were interested still
in estimating a nominal REI cost of debt rather than a U.S. dollar cost of debt. It's very simple
to compute a U.S. dollar cost of debt into a nominal REI cost of debt. Here's how I did it.
It mirrors exactly what we did with the cost of equity. I'm going to take 1 plus the U.S. dollar
cost of debt, which is 4.05%, and I'm going to scale it up to reflect the higher inflation in
Brazil, the 9% inflation as opposed to the 2%. I'm going to use a compounding effect. Some people
might just add the differential. It's not a bad approximation, but in this case, doing it right,
multiplying by 1.09 divided by 1.02, gives me a nominal REI cost of debt for Vale of 11.19%.
Again, it's simple to go across currencies if you remember this differential inflation trick.
So these are the three companies for which I was able to get a rating and use that rating to get
a cost of debt. But let's talk about the more generic problem. What if you cannot find a rating?
To estimate a cost of debt for companies when you cannot find a rating, I'm going to use a
single ratio. That might seem harsh to use one ratio to come up with a rating. Remember,
my objective is not to get a perfect cost of debt or the absolutely right rating. All I need
is an approximate cost of debt and an approximate rating. So I'm willing to accept a much more
simple process to get this, a much simpler process to get this rating and cost of debt.
So I'm going to use an interest coverage ratio to estimate the ratings for each of my companies.
What's an interest coverage ratio? It's the operating income divided by interest expense.
What does it measure? It measures how much safety margin you've built in as a company
to cover your interest expenses. As a lender, you want this number to be a high number.
So I took each of my five companies outside of Deutsche, which I did not compute a synthetic
rating for, and I divided the operating income or earnings before interest and taxes by the
expense for each of the companies. For Disney and Baidu, I get numbers in excess of 20. Basically,
I have more than $20 in income for every dollar in interest expense. For Vale, I get a number above
11. That's a pretty good number still. For Tata Motors and BookScape, I get numbers in the mid
four, five, somewhere around there. Basically, they reflect the fact that these companies
have very different safety margins built in. I still need to go from these interest coverage
ratios to ratings. And to do this, I'm going to use a lookup table. I'll give you a little bit
of history on this lookup table. I constructed it initially by looking at all rated companies,
looking at the ratings, and seeing what kind of interest coverage ratio I would need to get that
rating. So in a sense, I'm using actual rated companies to develop a lookup table that I can
use for non-rated companies. So let's take Disney. With its interest coverage ratio of 22.57,
and the fact that it's a large market cap company, notice that I have two sets of rules.
For large market cap companies, I will allow you to get a higher rating for a much lower interest
coverage ratio. Might not be fair, but that's reality. The rating I would give Disney is AAA.
For Vale, with an interest coverage ratio in excess of 11, even if I treat it as a risky
company because it's an emerging market company, the rating I would give them is AA. So I'm using
the small company coverage ratios now. For Tata Motors, with their interest coverage ratio of less
than five, the rating I would give them is A-. For Baidu, the rating I would give is AAA. I know
part of you is saying, there's no way a company like Baidu will have a AAA rating. You might be
right, but you're going to see in a few minutes why it's not going to matter that much. And finally,
for BookScape, treating them like a public company and giving them a rating with their
interest coverage ratio just in excess of five, the rating I'd give them is A-. So even for
companies without ratings, I can use the interest coverage ratio in this lookup table to come up
with a synthetic rating, synthetic because I'm making it up for each company. Now those ratings
can be used to come up with a cost of debt. Now you might wonder how close these synthetic ratings
are to the actual ratings, because at least for a couple of my companies, I do have an actual
rating. I can compare the synthetic rating to Disney, for instance. My synthetic rating is AAA,
but the actual rating is single A. How do I explain the difference? Well, one reason my rating is so
high is I'm using 2012 and 2013 income, which is much higher than their normal income, the average
income they've earned over time. Maybe the ratings agencies are being more conservative.
Maybe they're reflecting the higher variance of earnings in this business.
But whatever the reason, even if I used to AAA rating in my eventual cost of capital instead
of the actual rating, the effect on my cost of capital will be very small. What I'm trying to
argue for here is don't get too caught up in getting the rating precisely estimated. It doesn't
matter that much. Being within shouting distance of the true rating is all you need to do. Now for
Vale, the synthetic rating I get of AA is much higher than their actual rating of A-, and there's
a reason for that. Vale is a Brazilian company. You might say, so what? Well, the ratings agencies
factor in where you're incorporated in assigning your rating. In fact, they used to have a ceiling
called the sovereign rating ceiling, which prevented a company from ever having a rating
higher than the country in which it was. They no longer have that ceiling explicitly built in,
but they still factor it in. I am not factoring in the Brazil country risk when I give you the
synthetic rating, but the actual rating of A- already reflects it. Now to complete this process,
remember I did not compute a synthetic rating for Deutsche. There's a reason for that. Computing
synthetic ratings for financial service companies is very difficult to do. And with Deutsche, I'm
going to use the actual rating to come up with my cost of debt. So we've essentially got either
actual ratings or synthetic ratings for all our companies. Let's bring this process to fruition
by talking about the after-tax cost of debt for each company. Let me start with BookScape. With
BookScape, I have no choice but to use the synthetic rating. There is no actual rating.
The synthetic rating I came up with for the company, based on its interest coverage ratio is
A-, and the default spread is 1.3%. Adding that to my risk-free rate, I come up with a pre-tax
cost of debt of 4.05%. Now here's my final stop. I need an after-tax cost of debt. What am I
talking about? In most parts of the world, interest is tax deductible, and that's true in the U.S. as
well. So basically, you get an interest tax deduction which allows you to reduce your
effective interest cost. To compute that tax savings, I have to use a tax rate, and the tax
rate I'm going to use is the marginal tax rate. What's a marginal tax rate? It's a tax rate on
your last dollar of income, because interest saves you taxes on your marginal dollar, on the
very last dollars of income. So you're not saving taxes on the middle dollar or the first dollar,
and the marginal tax rate for BookScape is 40%. 4.05%, with the tax benefit netted out,
gives me a 2.43% after-tax cost of debt for BookScape. For the remaining companies,
I'm in position now to compute the after-tax cost of debt. Let me start with Disney. The actual
rating was single A. The pre-tax cost of debt is 3.75%, based on that rating. The marginal tax
rate for Disney is 36.1%. Why so precise? Because Disney very helpfully broke out the marginal tax
rate in their most recent financial statements. You net out the tax benefit, the after-tax cost
of debt for Disney is 2.4%. For DogeChip, the marginal tax rate is based on the German tax
rate, which is 29.5%. You apply that marginal tax rate tax benefit to the pre-tax cost of debt.
The after-tax Euro cost of debt for DogeChip is 1.94%. For Vale, I take their actual rating of
A-, and remember, that already incorporates country risk. I get a cost of debt based upon
that of 4.05%. I apply the 34% marginal tax rate in Brazil to come up with an after-tax cost of
debt for Vale of 2.67%. I have the cost of debt for four of my five companies. The only company
for which I don't have a cost of debt yet is Tata Motors. I do have a rating for Tata Motors from an
Indian ratings agency, and it's double A-, and the default spread for a double A- rated company is
0.70%. But to get a cost of debt for Tata Motors though, I have to take an extra step. And here's
how I get the cost of debt. I start with the risk-free rate in Indian rupees, which is 6.57%.
I do add the 0.7% default spread for Tata Motors, but I also add the default spread for India.
Emerging market companies carry two burdens of risk on their shoulder. One is their own risk,
the company default spread. The other is the country default risk. You add those components
together, you have a rupee cost of debt of 9.62% for Tata Motors. You net out the marginal tax
rate in India, which is 32.5%. You end up with an after-tax cost of debt in rupee terms of
6.5%. Now you might wonder why I did not do this for Vale. The advantage I had with Vale is I was
using an S&P rating, which already should reflect country risk. A Chrysler rating does not, so I do
have to bring in country risk separately. Now those default spreads are central to estimating a
cost of debt, and they do change over time. Here's my January 2014 update. But remember, these numbers
will get dated even as you see them. So it's your job when you compute the cost of debt to get an
updated default spread. And I'll suggest a link that you might find useful. It's called
bondsonline.com. Visit the site. You can get a default spread updated by ratings class for
different maturities, updated to the minute. So use the most updated default spreads you can,
because you need a cost of debt for your company. So when you do get a chance, try this out for a
company. Start off with either an actual rating, or better still, with an interest coverage ratio
to get a synthetic rating. Look up the default spread based on the rating. Check to make sure
your company is not in a country with an additional default spread. Come up with the pre-tax cost of
debt. Then see if you can find a marginal tax rate for the country in which your company is
incorporated. Use it to come up with your after-tax cost of debt. That's the second piece.
And your cost of financing, we can use that to come up with the cost of capital.
Thank you very much for listening.
