EVA extrapolation

I'll comment specifically on how I calculate ROIC because I did some work on this and found that there is really no consensus best way to do it. There's obviously somewhat of a standard method in textbooks but it's far from the best.

The way I think about ROIC is that I want to both capture the return that operations generate on the operating assets, but I also want to capture any incremental returns from management’s capital allocation decisions. Last year I built a multifactor model that we now rely on pretty heavily as a screening tool and a reference point in our meetings that scores every stock in a given universe (we have two, SP 500 specific and a broader Russell 3K plus select ADRs). So with that in mind, I needed to calculate it in a way that requires less company specific adjustments and is as one-size fits all as possible.

So our model based calculation is: EBIT + Interest Income / Assets - any tax assets - current liabilities + ST debt + Current portion of LT debt + Any capitalized leases.

A couple things to note that are different from the more textbook way to do it:

I don't back out taxes The common text book method is to take EBIT(1-Tax rate). There are a couple reasons why we don’t back out a tax rate. This is partially a function of we're using this in a model based screening context. You would have to go in and make adjustments for every single company to come up with an accurate tax rate at the EBIT level for each, and that just isn't possible for a screen. The other thing to consider is whether or not tax rate is a result of operations or management capital allocation decisions? If the tax rate is low because they’re domiciled in a country with a lower corporate rate, you could argue that is a structural advantage that they should get credit for. While this was certainly the result of a management decision at one point, it is not one that is likely repeatable, and most likely not one that even came from the current management team. I don’t think that the current management team should be rewarded or punished for this today, nor do I think it is in-line with what we want ROIC to tell us. I think that it makes companies more comparable to not back out taxes or estimate a tax rate.

I include the full cash balance in the IC calculation The argument for backing out cash is that it is not really invested capital, and interest income isn't included in EBIT. We think management should be held accountable for cash and the return it generates, hence we add interest income to EBIT and include the full cash balance in IC. To me, if management is sitting on cash, that is an active decision they are making to not do something with it, and in this environment some companies have massive cash balances (if you back out Apple's true cash balance, they would have negative IC in most years). Second, a lot of companies are going to maintain a minimum cash balance at all times to satisfy debt covenants or liquidity requirements, companies that are not distressed never truly run their cash balance down to 0. So the most accurate thing to do would be to make a determination of how much of that cash is excess, and how much is a required floor. Then you would need to adjust the interest income as well and do this for every company. For us in a screen, this is not possible, so we include the full balance and the return on that balance.

I include Goodwill To me, if you're making truly accretive acquisitions, your ROIC will remain stable or grow as your goodwill balance grows. I understand that strategic acquisitions can be accretive to future periods, but I don't think management should be given a free pass for ballooning goodwill. A counter to this would be that if they take a goodwill write-down in the future, it will artificially inflate their ROIC (same is true of the cash balance, if they pay a large one-time dividend or start buying back shares). This is true, and to counter this we use a 5-year average of ROIC in the model and include the 5-year standard deviation to offset these kinds of movements. Plus in the buyback or dividend scenario for the cash balance, I would view this as more positive than the company sitting on it and think that they can get some credit for it.

There are plenty of other things that could be discussed within the above (include pension liability, use asset based IC vs Liabs + Equity…) but those are some of the bigger ones. Again this is just the way I do it and it’s mostly driven by both the fact that I’m using this in a broad model/screen, and what I want ROIC to tell us, and that is really just:

• What kind of return has this company generated in the past from operations and management’s capital allocation decisions? • What is the trend (ROIC expanding/contracting) (stable and consistent/volatile and uncertain)? • Have these past returns been in excess of their cost of capital? • How does their ROIC compare with peers?

This specifically relates to ROIC in the past, given that my preference is for companies that have high and stable returns, I don’t forecast ROIC forward as much. Happy to debate or discuss any of the above or anything that I didn’t elaborate on in the above.