I was wondering how P123 gets the value of preferred equity when calculating enterprise value.
Just as an example, the actual value of GE’s preferred stock is 5.28 billion dollars, according to their 2016 annual report. The number on their balance sheet, though, is 6 million, one thousandth of its actual value, since there its value is only a dollar per share. Which number is used to calculate enterprise value?
I did some digging into this, and I see two problems with the calculation of EV. Now I’m not a financial expert, so I may be wrong about some of this. But here’s what concerns me.
Before I start, the actual calculation that P123 uses is MktCap + DbtTotQ - CashEquivQ + NonControlIntQ + PfdEquityQ. In every case I’ve seen, that corresponds exactly. If any of these are NA, that defaults to zero.
The first problem is using MktCap rather than Price*SharesFDQ. From what I’ve read, EV should always be calculated using diluted shares. MktCap uses shares outstanding.
The second problem is using PfdEquityQ. Here is where my understanding fails. How can GE preferred stock have a face value of $5.28 billion and a par value of $5.94 million? And which is their liquidation value? PG’s PfdEquityQ is NEGATIVE $238 million. How can that be?
If anyone can give me any insight into how to properly take into account preferred stocks when calculating EV, I’d appreciate it. Maybe P123 is doing the right thing, but it doesn’t make much sense to me.
As far as I know, there is no way to use the market value for the preferred stock. It would be nice if we had a market price and a number of shares but I just use the par value which has no relationship to market. It doesn’t seem to be a major problem for me but I never invest in big companies - no edge when so many people own and follow the stock.
Never had any issue with matching preferred book values versus what’s on P123. According to Capital IQ:
[i]This item represents the net number of preferred shares at year-end multiplied by the par or stated value
per share as presented in the company’s Balance Sheet.
This item includes
Where I do have problems: matching share classes. P123 data seems to really be designed for common equity… all point-in-time data (i.e., share counts) are for common shares. The only exception is “MktCap” which makes adjustments for market prices, share classes, currency adjustments, economic interests, and probably some other things.
It makes sense to account for other share classes – it is easy when economic entitlements are identical, or nearly so, as in the case of RDS.A and RDS.B. But this is not always the case.
Rant over and out.
Yuval, I know this is a very old thread. I’m curious if you ever got to the bottom of this EV question.
There is an even older and lengthy discussion on EV here that raises some questions.
EV uses SharesQ to calculate MktCap while I prefer to use SharesFDQ if it’s higher than SharesQ. EV is NA for some companies with lots of cash and I prefer to use a minimum of $5 million so I don’t exclude companies with negative NAs. Otherwise the formula I use and P123’s formula are the same.
Thanks Yuval,
The official P123 EV formula appears to be this:
((SharesQ * Price) + DbtTotQ + PfdEquityQ + NonControlIntQ - CashEquivQ)
Is this correct?
Which would mean you are using something like this:
isna(((max(SharesQ,SharesFDQ) * Price) + DbtTotQ + PfdEquityQ + NonControlIntQ - CashEquivQ), 5)
That’s correct, but add a few IsNAs in there for companies with, say, NonControllIntQ = NA.
What about those cases where both SharesFDQ = NA and SharesQ = NA?
Is the general best practice to set any NA value to zero?
The only variable in that formula that has no NAs in my universe is DbtTotQ .
Tony
Portfolio123 uses fallback mechanisms to make sure that MktCap is never NA. So if SharesQ or SharesFDQ is NA, go with MktCap instead of Max(SharesQ,SharesFDQ) * Price.
Note that NonControlIntQ is listed at par value, which is an accounting fiction. Does anyone have a way to estimate the actual market value of non-controlling interest?
Yes, if any of those components of EV are dependent on Par value, they may be thrown off- Par value is often a drastic low- a decent explanation here:
It’s surprising to me that in 2023 we still don’t have market value data for these publicly traded companies.
Does anyone have a way to estimate the market value of non-controlling (i.e. minority) interest?
Not sure if this is the most elegant way to do a logical OR operation but I think it covers all the bases except the market value of non-controlling interest dilemma. It results in about 1/3 of a 10K stock universe having a different EV from the P123 EV. And in all but 3 cases, the custom formula has a higher EV value.
isna(eval(SharesQ=NA or SharesFDQ=NA, MktCap, max(SharesQ,SharesFDQ) * Price) + DbtTotQ + isna(PfdEquityQ,0) + isna(NonControlIntQ,0) - isna(CashEquivQ,0), 5)
That is not quite the same since MAX() will return the one that is not NA. Its a positive logical OR instead of a negative logical OR. It only works the same if both are NA.
@abwillingham You’re right, I missed that! How about Eval(SharesQ+SharesFDQ = NA, MktCap, Max(SharesQ,SharesFDQ)*Price)?
BTW, I have found that the real problem with the built-in MktCap calculation is not SharesFD but rather when there are recent changes in share counts after the most recent financial statement date, such as buybacks or shares being sold.
Yep that Eval() works. I didn’t know an NA + nonNA = NA.