Depreciation gets little respect nowadays. We all know how much investors have come to dread accounting accruals, so nowadays, many who want to examine capital charges lean toward the real thing; the actual capital expenditure number which is deemed to be superior to the so-called fictional depreciation figure. This is unfortunate. Preferring depreciation over capex will definitely cost you intellectual brownie points, but it might contribute to your dollars-and-cents portfolio performance.
Financial measurement of management performance
Seeking Alpha's thought-provoking December 15 article entitled "Evaluating Good and Bad Management Strategies" by ValueExpectations.com spotlights an interesting metric known as economic margin, which the authors define as:
(Cash Flow - Capital Charge) / Productive Capital
The full definition of this sophisticated ratio can be found in the methodology section of the ValueExpectations web site.
Aside from the details, a casual reading of the article could easily lead one to assume "capital charge" refers to the capital expenditure items from the statement of cash flows; after all cash flow minus capex is widely preferred nowadays as a measure of financial performance that is supposedly better than net income.
I do not share the widespread preference for use of capex instead of depreciation. Capex can chart very bumpy year-to-year or quarter-to-quarter trends as projects ramp up and wind down. Companies that occasionally look bad in terms of cash flow minus cap expenditures might actually be doing a better job for their shareholders over the long term by not scrimping on necessary investment in the asset base. Conversely, the free-cash-flow kings may be doing shareholders a disservice by allowing the asset base to deteriorate.
Depreciation reflects an effort to cope with the reality that often-large single-period expenditures can help the company generate revenue over may additional future periods. It is true that these allocations include a good deal in the way of management assumption (or guesswork if you prefer). But I've long believed that good faith assumptions or educated guesses can present a better picture of management performance than a more solid number that doesn't even try to match revenues and expenditures. Hence my inclination to test the above formula against a version that uses depreciation in lieu of capex.
Before going further, note that the tests here, although inspired by the ValueExpectations article, will not cover the actual metric used by that firm, which, among other things, includes depreciation in cash flow and compares that to a capital charge that is based in part on "Replacement of the firm's Depreciating Assets" (the latter is not specifically defined on the web site). Instead, this test will focus generally on differences between net income and cash flow and more particularly, differences between depreciation and capital spending.
On Portfolio123.com, I created a single-factor ranking system based on:
(Trailing 12 month Cash Flow - Trailing 12 month Capital Spending) / Latest Quarter Total Assets
A screen was set up to exclude financials, OTC stocks, and stocks with market capitalizations below $250 million and prices below $5. Among stocks making this basic screen, I backtested a hypothetical portfolio consisting of the 20 stocks that scored highest in terms of the aforementioned factor. The holdings were rebalanced every four weeks, 0.5 percent slippage was assumed for all transactions, and the test ran from 3/31/01 through 12/15/08 using Portfolio123's point-in-time database (i.e. no survivorship bias).
Figure 1 shows the results of this test:
Oh well. Many who have done this sort of analysis have seen results like that; good performance for most of the 2000s with you-know-what breaking loose in the past year or so. True the relative down-market performance was positive, but much of that stems from the past; the second-half of 2008 contained a disproportionate number of dreadful months.
But measuring this simple metric against the S&P 500 isn't really the point here. What we're really trying to do is consider different ways of calculating the ratio, which, by the way, does seem to closely resemble return on assets. So let's look next at the latter. The denominator of the fraction (total assets) remains unchanged, but in the numerator, we use garden variety net income which reflects a charge for depreciation rather than cash capex. Figure 2 shows the result of this approach:
Ouch. Again, I'm not so much bothered by the inconsistent and too-often lackluster results relative to the S&P 500 (as I would be if this were offered as a full-fledged stock-selection protocol) as I am by the fact that it compares poorly with Figure 1.
Is it possible that investors are completely indifferent to the realities of plant modernization and are locking in on the most short-term of views ("show me the cash now, the heck with the future")?
Before accepting such a depressing conclusion, let's consider one more test.
Look again at the numerator in the first metric I tested: Cash From Operations minus Capital Spending. Computing the second metric, Net Income, is not simply a matter of swapping depreciation for capital spending. There are other accounting accruals involved in net income. Let's get these other accruals off the table by changing the numerator to Cash From Operations minus Depreciation.
Usually, accrual-hating investors eliminate all of these things, while the net-income crowd includes them all. I'm going to steer a middle course by eliminating all accruals except for depreciation. In other words, I'm saying that I, too, generally hate accruals, but I'm making a special exception for depreciation because this particular accounting fiction strikes me as being pretty sensible.
Figure 3 shows the result of a test of the latter approach:
That's interesting. Not only did this middle approach perform best, it was especially noteworthy for its strength in down months (including many of the ones we've been experiencing recently).
I'm not sure at the moment why use of depreciation proved so helpful in bad times. Companies are probably reigning in capital spending now in response to the recession and financial crisis. Perhaps cash flow is being perceived as being inflated by unusually slim capex leaving depreciation as a more reliable measure of ongoing financial performance. I don't know. I'm just thinking out loud. But such reasoning would seem consistent with my initial position to the effect that a reasonable even if imperfect attempt to match revenue with associated plant spending trumps a purer number that doesn't even try to make this crucial match.
Score one for the depreciation accrual
It looks like I'm concluding here that accruals are generally bad except for depreciation, which stands as the one good accrual.
I do think we ought to give depreciation a lot more respect than it deserves. Even value investors who do detailed cash-in and cash-out models have food for thought.
But I'm not at all going to say the other accruals are necessarily worth bashing. This isn't the first time I addressed this general topic. In September, I presented some tests which suggested that problems associated with accruals may really have more to do with items that obstruct an investor's effort to see ongoing trends, free from occasional, even if sometimes large, oddities.
I don't think we're anywhere close to having the last word, one way or the other, on the topic of accruals. But pending a lot more research, it seems that basic accounting concepts may warrant a bit more in terms of a presumption-of-innocence than they often get from the investment community.