First things first: A Sum-of-the-Parts or SOTP valuation is based on t valuation of single parts (assets) of the respective business to finally derive the fair value for the business as a whole, summing it all up and adjusting for debt (and some other stuff).
How does that help You?
It is reasonable to ask the above question, since it sounds that performing a SOTP you multiply the valuation problem by the number of the single parts. Now you have to value each single part of the company seperately, ideally via a DCF (my favourite valuation concept). You could stop reading here and say ‘No thank you!’ but, bear with me:
In general there are many more things that can go wrong than can go right when valuing a company. This of course is also the case when doing a proper DCF, no matter how hard you try and how much research you do. There comes a point where you have to make assumptions (you know: garbage in, garbage out…).
You can set up any model by using unrealistic assumptions for your key inputs like growth, profitability, discount rate (i.e. wacc). This way, you can come up with any (desired) fair value for any business, on purpose or by accident and I can’t tell you which is more fatal…
Simple is better than complex. That usually applies for businesses and for valuation approaches from the point of view of an value investor (at least this applies for me). A simple business with a focused operation is much easier to understand (for in- and outsiders I guess). In a big business (say IBM) it might happen that you study their reports and never get an edge, since the company (most likely executives are at fault here) is changing its reporting structure every year. The definition of segments is often blurred.
Using simple valuation approaches you minimize the errors to be made. The investment proposition is much better to comprehend and the key value drivers can and should be lined out more clearly.
In a perfect world
Assume you want to value a conglomerate. In a perfect world, you (I) would take your time (which is sadly not unlimited in the real world) to perform a very solid DCF on each and every Part of the company. Sum it all up, adjust for debt and voila. Each DCF would be based on a lot of research, both on the respective company and on the industry. But we know that everyone of these valuations is imprecise and uncertain, because there are too many moving parts. Every single DCF will rest on a few key assumptions and external factors. One additional (and very critical implicit assumption) would be on the capital allocation (skills of the managers) within the Holding Company (HoldCo). That is, hopefully the manager(s) allocate capital efficiently (or similar to our underlying assumptions used in our models).
In a less ideal world
Dealing with our time limits in a less ideal world and to mitigate the large number of potential errors in our valuation approach, we could go with valuation short-cuts. We can use these for some or even all of the single parts involved — though we know they are not perfect, but they might be simpler. Thinkable valuation apporaches might be based on accounting values or market values. These again, can be derived differently…
We might use multiples. Performing valuations of assets based on multiples is one possible solution. It is quick and simple: Decide on a common metric (revenue or sales, Ebitda, Ebit, net profit; or nowadays monthly active users, gross merchandise value, …) with comparable companies (peers) and calculate the (trimmed) average/mean multiple of this peer-group with: Multiple = Market-Value / Metric.
Problems with that: Defining the right peer group is somewhat tricky and arbitrary (not discussed here). Value investors believe markets are not (always) efficient, why trusting the market? (read below)
In the real world
We all live in the real world and should adhere to that fact. If we are a bit lucky in this world, some or even most of the parts could be quite simple to value (if we choose it to be). Some might be seperately listed on the stock market, giving us the opportunity to use their current stock market value.
Why trust the market? you might ask. It is indeed true that we value investor strongly believe that stocks are not always priced correctly by ‘Mr Market’. That is not to say, that ‘the market‘ on average is not good in pricing stocks. And there might be some companies that I do not want to value (high effort) or where I would not have a high confidence in my valuation (talent, insights, edge). So sometime it might be a good idea (without much effort) to use current stock market valuations. A common sense reality check is obligatory.
Another strong argument in support of using current market prices or stock market valuations is the following logic: Value investing is build on the believe, that a valuation gap — that is the difference between (conservatively calculated) fair underlying value and current market price — will be closed in the future. This will probably happen over time because of any one or several catalysts taking place (read my page About Catalysts). Assuming the holding company sells off its stakes in publicly traded companies at current market prices valuing these stakes at market prices makes actually sense.
A very obvious catalyst or maybe the most obvious catalyst to close the gap is to sell shares for their current price for cash (which should close the valuation gap). In reality there are some complications involved, like selling a big stake will depress the current market price, perhaps there is capital gains tax to be payed, or any other taxes involved, etc. … . But I hope to give a good example below.
Imagine a company that has gross debt of 100; cash of 10, and the following number (#) of shares of companies A, B, C, all listed on mayor exchanges.
A: #12 x $9 = $108
B: #20 x $5 = $100
C: #15 x $7 = $105
To keep it simple: no costs, no taxes due, no discounts, market prices are not moving. After selling all shares the company has total cash of $10 + $313 = $323 to repay all debt (-$100) and has $223 left. Liquidation of the company will thus result in this cash being distributed to shareholders.
Here we only have one shareholder owning one share (effectively having control over the company). If the share trades currently at 120, this would be a very attractive investment proposition with an underlying simple valuation model and a simple path to close the valuation gap quickly! (order your employees to sell all shares immediately, and fire them).
I would buy instantly. Even if you apply some more realistic assumptions (discounts / taxes / costs / frictions), it might still be an attractive investment opportunity.
The big picture again
In general, a SOTP valuation is nothing else than exactly that, a Sum of the Value of the single Parts. When you decide to go with the SOTP approach, the next question obviously becomes
How to value all the different paths? And it might make sense to take a different approach for each asset. Sometimes it might be best to value a specific asset at zero (imagine a lottery ticket and
you I had no clue about statistics). We all know, that it’s worth more than zero, but the expected value is based on highly skewed outcomes, with the expected value EV being calculated like EV = 99.99%x0 + .0001%xW; with W being the big win.
Valueing it at zero has two advantageous. (1) It keeps things simple and its result does not hinge on extremely positive but highly unlikely outcomes and (2) the SOTP errs on the conservative side. In this case, the specific asset (lottery ticket) would mean pure upside to your derived SOTP result or NAV (and calculated buy price).
The concept in analytical terms is quite simple. You derive the net asset value or NAV by subtracting debt from assets. NAV represents the (theoretical) liquidation value, which might (not) be tangible. You can compare NAV to current market value and calculate if the company trades at a discount to calculated NAV. The you can ask Is the discount too high?
There are many ways to perform a SOTP. But there are better and worse ways in most cases. I prefer valuations to be simple and to err on the conservative side.
You might be interested to take a look at some of my SOTP valuations here.
I hope I was able to explain the basics of SOTP valuations. If you have further questions, please shoot me a message or start a discussion below.