In the last post, I laid out how CSCO is gracefully growing up. Its highly profitable core “switches and routers” business (the plumbing of the internet) is being challenged by new competitors, and the usual forces that drive technological irrelevance. CSCO – the survivor – is entering new businesses, even while maintaining healthy overall margins.
In this post, I will practice applying some of the lessons I learned from Aswath Damodaran’s book and come up with an estimate of CSCO’s current value.
Over the past 7 years, CSCO’s Sales/Capital Employed ratio (SCE) has varied between 2.5 and 2.0, where capital employed is defined as debt + equity – cash. I will assume SCE = 2.0 for the purposes of this valuation. This ratio tells me that for every $1 that CSCO is able to plough back into the business, its sales go up by $2.
Since 2009, CSCO’s tax-rate has consistently been around 20%. It is certainly possible that this rate will go up in the future, but I will assume its well-paid tax lawyers won’t let it rise.
CSCO’s debt/capital ratio has increased to nearly 30% in recent years.
The TTM revenue in June 2016 is $49.5B.
I project cash flows for the next 10 years under the following conservative assumptions:
- revenue growth tapers from 3.5% to 2.5% in year 10, as it matures and approaches the growth rate of the economy as a whole
- EBIT margin goes from the current 20% to 16% in year 10, as its core business declines, and new business are not as profitable
- tax rate remains 20%
In the spreadsheet below, the dollar numbers are all in millions.
Cost of Capital
CSCO is a mature profitable company. Overall, I assume:
- beta declines steadily from 1.2 to 1.1 over 10 years, more befitting an even more mature company
- risk-free rate is 2%, equity risk premium (ERP) is 5%
- the balance sheet is pristine and its debt is really high quality (spread = 0.75%)
- a marginal tax rate of 35%
- I model a debt ratio increasing modestly from the current 30% to 35%
The cost of equity = risk-free rate + beta * ERP, and the cost of debt = (risk-free rate + spread) * (1 – marginal tax rate). I determine the weighted cost of capital by combining these costs, using the projected debt-ratio.
The cost of capital profile hovers around 6%, reflecting the low interest rates, CSCO fortress balance sheet, and judicious use of debt to lower the cost of capital.
I use the cash flows table above to project out FCFF in year 11 using the same revenue growth rate and operating margins as in year 10. I further assume:
- a stable revenue and earnings growth of 2.5% keeping up with the growth of the economy
- CSCO is somehow able to maintain its stellar return on capital employed (ex-cash) of nearly 30%. This implies a reinvestment rate of 2.5%/30% = 8% (approx). This seems like a reasonable number for our base line assumption of CSCO as a profitable, limited growth, mature company. If any of the strategic forays plays out, and CSCO enters a new moderate growth phase, it will surprise to the upside.
- A stable cost of capital of 5.5%.
The Gordon growth formula suggests:
Terminal Value = NOPAT in year 11 * (1. – stable reinvestment rate)/(stable cost of capital – stable growth) = $262 B.
Discounting Cash Flows
If I assume that the discount rate is the cost of capital, I can discount all the FCFF numbers and the terminal value. This gives me a net present value of cash-flows to the firm as $204 billion.
I add back net cash, assuming a 20% haircut due to repatriation taxes, (0.8*$63.5B) and subtract debt and options ($28.6B + $1.3B) to find the value of the equity. Dividing by the fully diluted number of shares 5.146B, I get an intrinsic value per share of about $43.
The current share price of about $28-$29 seems cheap. At $43, CSCO would trade for 18-19x this years projected EPS, which seems a tad rich.
I am still unconvinced that I should be using the cost of capital as the discount rate in discounted cash flows. For I know this: if I buy CSCO at $43/share, and all my assumptions prove spot on, then my return on investment will be the discount rate (approximate 5.5-6.5%).
I would definitely like a higher hurdle rate; say something like 10% or better.
I redid the discounting with a discount rate of 10%. When I do this, I get a net present value of cash-flows to the firm as $147.5 billion. Adjusting for cash, debt, and options, I get a more “reasonable” intrinsic value near $32/share, around 10% above prevailing prices.
In the valuation, the projections are all smooth functions of time. We expect reality to be bumpy: for example, year to year revenues may decline, but be offset by unexpected increases in year 7.
To characterize the uncertainty in IV, I ran 10,000 MC simulations with the same the same baseline assumptions as above, but threw in some noise. In particular:
- I assumed a revenue growth in any year was a normally distributed function with the same mean but a standard deviation of 1%
- Likewise, I dialed in a 2% standard deviation in the operating margin, a 10% standard deviation in the sales/capital employed ratio
Note that I used independent normal distributions to model “uncertainty”. These distributions are symmetric. Perhaps, I should use smarter distributions in the future. In any case, I just wanted to see how sensitive the IV is to small perturbations in my particular assumptions. I get a distribution that looks like:
So a reasonable range looks like $30-$35. Of course, unanticipated events could derail the entire thesis. Examples of “known unknowns” might be technological obsolescence due to new innovations (-ve), some non-core business takes off (+ve), Congress allows trapped cash to be repatriated at low or zero rates, etc.