Today we are going to cover the 29 most common mistakes in discounted cash flow valuation.
And why is this important?
Well, for starters making these mistakes can lead to some really poor decisions.
Case in point?
Well, had Walter White known about DCF he probably wouldn’t have sold his stake in Grey Matters.
And in the real world, Ronald Wayne probably wouldn’t have sold his 10% percent of Apple for $800.
So yeah, this stuff is important.
1. Not doing discounted cash flow because it is “too academic”
The first item on the list.
And with good reason because we love this one.
Some people will say that the DCF is just a garbage in, garbage out model so why bother with the output?
And you know what? We agree with them.
If you treat the DCF model as a garbage in, garbage out model you might as well not bother.
But did Einstein back off with his relativity theory because it was difficult and involved a lot of hard math?
Thankfully, he did not.
And the same applies to the DCF model, which is hands down, the single best way to value a company or an investment.
And the good news? A good discounted cash flow analysis is so, so much easier to make than having to come up with the theory of relativity.
2. Too short a forecast horizon
In general, we advocate using a forecast horizon of at least five years and for fast growing companies, the generally accepted wisdom is 10 years.
In this valuation of Tesla by professor Aswath Damodaran you can see, he uses a ten-year forecast horizon for valuing Tesla.
In the first five years, he assumes a growth rate of 70%!
And no, that is not sustainable.
In years 6-10 the growth rate gradually declines from 25% to 2.75%.
That is sustainable long-term.
The rule is that the forecast period should be as long as it takes to be in a sustainable steady state. This is also sometimes referred to as a normalized state.
3. Too long a forecast horizon
So why not make the forecast horizon 50 years?
You can do that.
But how do you forecast what is going to happen 30 years from now? Can you forecast whether revenue will increase by 3% or 2% in year 30?
The answer is that you probably can’t.
Instead of forecasting things 30 years from now we estimate “the value of all cash flows beyond our forecast period” with the terminal value.
4. Unsustainable growth rate
One of the most common DCF mistakes is forecasting and unsustainable growth rate.
This often happens when a company has been growing by 5-10% for many years.
People start asking: Can this last?
And it might be hard to see why it shouldn’t.
However, history tells us that all companies eventually become mature and grow by 2% in line with the wider economy.
As an example, consider Apple in 2010.
Apple had been growing tremendously over the last 10 years, the company was extremely hot and customers were sleeping in front of Apple stores to get the newest iPhone.
So Apple will keep growing by at least 5% per year?
With the benefit of hindsight (as of writing this in late 2016), Apple’s revenue has grown by 7%, 28% and -8% in 2014, 2015 and 2016 respectively.
Shows how quickly things can change.
So, make sure to set a conservative long term growth rate when arriving at the terminal value.
The norm is for a company to grow at 2% with the wider economy. You can estimate higher growth rates but the arguments for doing so should be very solid.
5. No connection between terminal period growth and forecast period estimates
Imagine you have a five-year forecast period and you estimate revenue to grow by 7% per year.
You have read above that one of the most common DCF mistakes is a too high terminal value growth.
And you estimate terminal growth at 2%.
So, the company went from a 7% growth rate to a 2% growth rate.
Seems rather abrupt doesn’t?
The arguments for this are probably weak.
A common solution here would be to extend the forecast period to 10 years and gradually let growth decline from 7% to 2% over years 6-10.
6. Depreciation higher than capital expenditures in terminal value
The terminal value is a normalized level of cash flow.
Remember, these cash flows will in principle continue in eternity.
And in eternity, depreciation cannot be higher than capital expenditures as the two concepts are intertwined and, to quote Frank Sinatra “go together like a horse and carriage”.
And so, in the terminal value the two should be equal.
Cost of capital
7. Using cost of equity as discount rate
The key to remember when doing a DCF valuation of a company is to use WACC as your foundation for the discount rate.
8. Inaccurate equity risk premium
Another classic beginner’s mistake.
The equity risk premium is given as the expected market return minus the risk free rate.
Historically, the market has returned 8%.
And currently (2016), the risk free rate is basically 0%.
So the market risk premium must be 8%?
No, because we are comparing apples and oranges.
In the example, we take a long term return and subtract a current return. Wrong.
Instead, simply head to over Aswath Damodaran’s site and see what the current equity risk premium is.
As of writing it is 5.5%.
9. Using incorrect Beta
Consider a company that isn’t traded every day.
And you estimate a daily beta for this company.
In other words, four days a week the stock price doesn’t move and on the fifth day, it moves.
But the wider stock market moves every day and thus, the beta calculation… well it is basically worthless.
The solution is to either estimate beta using weekly data or alternatively, use an industry beta.
10. No size discount
Say you are the owner and CEO of a small company that sells software like Microsoft.
And you use Microsoft’s beta as the foundation for your cost of capital.
Is this correct? Yes, you are exposed to the same industry.
Is the risk free rate the same? Yes.
Is the equity risk premium the same? Yes.
But you still have a hard time believing your discount rate should be the same as Microsoft’s.
For example, you could be run over by a truck and there wouldn’t be anyone to take over from you, so you think the risk should be higher.
And you are correct.
That’s where the size premium comes in handy. You basically add the size premium at “the end” of the CAPM formula.
While there are no real academic guidelines, the size premium usually ranges from 1% to 5% depending on the size of the company.
So if Microsoft’s cost of equity is 10%, your cost of equity would be between 11% and 15%.
11. Not doing any financial analysis
Yes, this is actually quite common.
We are forecasting the future so why dwell on the past?
Well, usually the past is a good indicator for the future.
It is thus highly recommended to conduct financial statement analysis for the last three years before diving into forecasting the future.
12. High growth in prior years = high growth in coming years
As established earlier, Apple has seen tremendous growth only to experience a negative growth rate in the last financial year.
It is hard to sustain a high growth rate, as growth rates are relative.
To see this, let’s assume a company’s revenue grow by 50 per year from 100, to 150 to 200.
The corresponding growth rates would be 50% and 33%.
If the trend continues and next year growth is 250, the growth rate would be 25%.
So even though the company has grown by the same absolute amount as in the first year, the growth rate has fallen from 50% to 25%.
And secondarily if a company keeps growing by 50% it would soon own the whole world.
We have yet to see examples of that…
13. Jump in first forecast year
This mistake is somewhat tied to not doing proper financial statement analysis and as such, is a usual beginner’s mistake.
The mistake here is usually, that a lot of fundamental research has been made and boy, it looks good for this company.
They must make a ton of money next year!
So you forecast significantly higher margins compared to prior years.
The best way to validate this is to check whether there have been any major initiatives. If not, then you are probably too optimistic.
14. Unrealistic margin assumptions
Management often say:
“We have implemented new IT systems so we will be much more productive and our EBITDA margin will increase from 10% to 15%. We have also performed relatively bad because of issue this, issue that and an issue there”.
This might be correct.
But an increase from 10% to 15% is quite a lot.
And competitors are also implementing new systems.
And issue that, issue this and the last issue are maybe gone, but will be replaced by new issues in the future.
So, beware of forecasting large margin increases based on what would be characterized as “normal part of doing business”, because competitors will also do this.
15. Last forecast year is better than any years before it
In a growing business, the last forecast year is probably the year with the highest revenue, EBITDA and so forth.
However, if it is also the year with the highest revenue growth and/or the highest margins there should be very solid reasons for this.
Often, beginner’s think: “We have forecasted the business to do better year-on-year and in the final year, all the rewards will be reaped”.
In other words, revenue and margins will soar.
This however, is the wrong way to think about things.
Instead, during the forecast horizon, revenue growth and EBITDA margin should approach the level forecasted in the terminal value and not peak in the final year.
16. Lack of cash to execute business plan
This is a problem often seen in start-up and young companies.
Forecasted revenue growth is high, margins soaring and everything is looking great.
But where are the money financing all this growth going to come from?
Thus, in these situations with high capital expenditures or a large drag in net working capital, special attention should be given to debt and cash levels.
17. Not accounting for M&A
Financial statement analysis for at least the last three years is a large part of doing a DCF.
So, you open the 2015 books and type in the relevant numbers to your spreadsheet.
You open the 2013 and 2014 books and repeat.
The only problem is that you missed a large M&A transaction in 2014 the significantly increased revenue and you forecast this to continue.
However, there was no real growth in the business – all growth came from the M&A transation.
So, you grossly overestimated future growth.
The quick fix?
Luckily this is easily corrected by only using figures from the latest financial report as historical financials in these documents are required to have “like-for-like” items.
In other words, the effect of the M&A activities will be included in these numbers.
18. Unrealistic capital expenditure assumptions
Often you will hear the saying that “we have just invested in a new factory, so we will not have any capital expenditures in many years to come”.
And often that will be incorrect.
The world develops and even though a factory is new, it still needs maintenance and new machines.
Thus, while capital expenditures will likely be lower going forward (remember, we have just bought a new factory) it would be a step too far to not forecast any capital expenditures for the coming years.
19. Miscalculating net working capital
This one is simple: Many people do not know what net working capital is.
The quick fix: Read our guide to financial statement analysis.
20. Depreciation rate
Often, people will forecast the depreciation rate as the average of depreciation rate for the last three years.
This will often be correct.
However, if there are large capital expenditures planned (or have just been finished), depreciation should reflect this and rise following these large investments.
21. Mixing absolute and percentage terms
Before you laugh in my (virtual) face about this one: It happens often.
And if you have made this mistake you now know you are not alone.
Say you estimate revenue to grow by 5% per year and costs to grow by $1m per year for the next five years.
After five years, costs take up a much smaller amount compared to revenue compared to when the forecast period started.
This might be correct, but it probably happened “by accident”.
The solution is to make all your forecasts in relative terms.
22. Tax rate errors
“I am valuing a US company, so I will use the US corporate tax rate of 40%, right”?
Companies have the right to make all different kind of deductions, which impact the tax rate.
The fix usually is to use the average of the last three years’ tax rate as the basis going forward.
23. Using US discount rates for a foreign company
We usually look at things from a US perspective.
And as such, we use the US market rates to determine our WACC.
However, if you are valuing a Brazilian company you should use the Brazilian risk free rate and equity risk premium.
Which, by the way are much higher as Brazil is a high inflation country.
The rule of thumb is to use the same currency as the stock is denominated in.
24. Mixing real and nominal growth rates
This one is quite simple:
A: If you forecast cash flows without incorporating inflation, these should be discounted at a real discount rate
B: If you forecast an expected inflation rate, the discount rate should incorporate the same discount rate.
The vast majority of forecasts use method A.
25. Lack of sensitivity analysis
This is not actually a required part of doing a DCF.
But it still is…
A discounted cash flow valuation is an estimate of cash flows in the future.
And even though it is the most likely future outcome it is by no means certain.
And a DCF is used as a tool for making decisions.
So, one should make a sensitivity analysis when doing a DCF to see what happens if growth rates increases or falls by 1%, see what happens if margins increase or decrease etc.
These things will say something about the robustness of the decision being taking.
If you have a buy signal and it is still a buy signal even after taking more negative scenarios into consideration you have bought yourself a lot more comfort that you are making the right decision.
Which makes you sleep better at night…
No not that kind of models…
We are talking about simple calculation errors, that lead to valuation mistakes.
26. Double counting values
Inadvertently, we have often seen double counting of values.
This is most prevalent in cases where a company is forecasted to repurchase shares.
This will often be built into the model, which makes sense.
However, in these cases the cash flow will often be counted twice.
The first time in the free cash flow formula, as free cash flow is calculated before stock repurchases.
Then it will be counted a second time when the shares are bought back in withdrawn from the market.
Thus, if projecting share repurchases make sure to not count the cash used to purchase shares in free cash flow.
27. Mixing assumptions and calculations
You have probably noticed that links on web pages have a different color or are underlined so they are easy to distinguish from regular text.
The same applies for a good DCF model.
You should easily be able to distinguish between input cells and cells with formulas.
This is usually done by giving the two different colors. It is usually blue for input cells and regular black for cells with formulas.
If you download our DCF valuation spreadsheet you will see these criteria play out as well.
Hard to read example: All cells are the same color.
Easy to read example: Input cells are blue, computing (i.e. do not touch) cells are black.
28. Including free cash flow before valuation date
When you get a job, you do not get the salary of the guy that had the job before.
Likewise, whether you buy a share in a company or the entire company, only cash flows after the purchasing date will belong to you.
So naturally cash flows before the valuation date should not be included in the valuation of the firm.
Therefore, the value of a stock will decline on the day after which a dividend has been paid.
If you bought the stock the day before the dividend you would also pay to get the dividend. And if you buy it after the fact, naturally you would not pay for it.
29. Not sanity checking results
Having done a full DCF one should sanity check the valuation.
For a listed company this is easy to do, as you can benchmark the equity value against the current market cap.
And believe it or not, the first time we did a DCF as undergrad students our first estimated market cap was more than ten times higher than the current market cap of the company.
We could be correct… but we probably weren’t…
In that model, we had made many of the mistakes presented on this page and we knew something needed to be corrected.
This was easy to see as the company was listed.
But what if the company is not listed?
Stable, mature companies normally trade at a range of 6-10x EBITDA.
Young and fast growing can trade at anywhere between 10-25x EBITDA but values above that are rarely seen.
In our model, which you can download here, we have of course included these multiples so they are readily available.
Normally, the most important multiple is the one for the first forecast year.