Frank Voisin is the author of the popular value focused website Frankly Speaking, found at http://www.FrankVoisin.com
Last month I posted Hugh Hendry’s latest investor letter. Hendry is known for his macro insight, and deservedly so, but there was something in there that I thought was useful to micro investors:
[C]orporate Japan, along with much of the rest of Asia, is guilty of paying too many redundant quasi-state employees anything at all. It is a ruinous policy, symptomatic of a failure ot face tough decisions, and it is slowly but surely eliminating all of the residual equity value in far too many businesses. …
Consider Hitachi. It has a new CEO, one who has committed to improving operational efficiency. That has made the market happy but we do not think the market is concentrating. … Instead let us look at the fact that the firm has 376,000 employees and that what truly moves the needle on margins in these slow moving mega-conglomerates is the ability to reduce headcount. …
Using Sony’s recent restructuring cost as a reference, it looks like it would cost Hitachi between $75k and $100k per redundancy. That is an implied liability of $8bn. No one calculating book value for Hitachi takes this into account. So book value is hugely over stated. …
Another example is Renesas, the slightly less commoditised semiconductor sister of now defunct Elpida. It has some merits (largely that it has outsourced rather than heavily invested in capex) but it still has 44,000 employees. With a market cap of just $2.9bn, this is surely way too many workers. By comparison, note that Qualcomm has a market cap of $114bn, but employs just 21,000 people. Does Renesas have any equity value? We do not think so.
What we see here is Hendry explicitly considering the employee count in a (very rough) valuation. Normally, I would take account of a bloated workforce indirectly via higher expenses leading to lower earnings and free cash flow and thus a lower valuation.
In all honesty, I like Hendry’s approach better. By being explicit about the eventual cost of downsizing, we can be clearer about what assumptions we are making and see how much different variables affect the company’s valuation. With the approach I’ve been using, the expense ratios are often opaque with a lot of variable and a lot of fixed costs thrown together. The potential savings are far from concrete and we usually aren’t explicit about the assumptions being made.
Where I think Hendry misses the mark is his choice of comparing the number of employees to a company’s market cap. The company has no control over its market cap and we see examples all the time of wild fluctuations in a stable company’s market cap over short periods. I think the better ratio is Revenues/Employee because revenues are more directly comparable over time and across companies and revenues tend to be far more stable.
So here’s what I’ve taken to doing: when I analyze a company, my model now has space for the number of employees for each year historically and I calculate the Revenues/Employees ratio. Here I am looking for big shifts over time to give some context to the current numbers. Then I look at a sample of competitors to see what their Revenues/Employee ratios are, for comparison sake. If there is a major difference that cannot be explained (say, by different business lines or sales channels), I’ll create a liability to account for the ultimate downsizing to get closer to industry averages.
What do you think of this approach?