Fool.com: A Margins Screen Theorem (Workshop) September 8, 1998

A Margins Screen Theorem

Atlanta, GA. (Sept. 8, 1998) -- In the next few columns, I'm going to investigate two questions. First, can the rising margins screen detect the great investment opportunities? Second, can a rising margins screen filter out mediocre and poor performers among stocks pulled up by the workshop's mechanical models? Because I'm not backtesting years of data for thousands of stocks, my answers will be provisional. But at least we'll be looking at specific examples from this year's mechanical model portfolios.

Today, I propose a theorem: Consumer-oriented stocks should be sold (or avoided) when their margins fall, and they should be bought (or held) when their margins rise.

I'm limiting the theorem (for now) to businesses that sell to average consumers because I think they're the easiest for individual investors to understand. Also, our Fool screen is set up to use earnings per share as a proxy for margins (that is, a stock will make the rising margins screen when its earnings per share grow faster than revenues). That's important because it means that stock buybacks can occasionally help a stock pass muster even when a company's net margins actually fall slightly.

Our examples will be the three top-performing stocks from the model portfolios as of August 26: Dell <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: DELL)") else Response.Write("(Nasdaq: DELL)") end if %>, Best Buy <% if gsSubBrand = "aolsnapshot" then Response.Write("(Nasdaq: BBY)") else Response.Write("(Nasdaq: BBY)") end if %>, and the Gap <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: GPS)") else Response.Write("(NYSE: GPS)") end if %>. Will the theorem produce good results? I dunno. I'm off to run the numbers....

... Ok, I'm back. We'll start with Dell, one of the market's darlings of late and the indisputable Big Daddy of the mechanical models over the past eight months. Going back 10 years, we find Dell trading at $5/16! That's right, just 31 cents per share. (Just to confuse you, my numbers adjust for all splits except for the 2-for-1 split that took effect today.) We see a nice run to $3 1/8 by January 1993 (the end of FY93) then a plunge back to $7/8 by July 1993 (mid FY94). You could track this quarter by quarter, but the annual net margin numbers tell the general story pretty well.

 
 FY90           1.3% 
 FY91           5.0% 
 FY92           5.7% 
 FY93           5.0% 
 FY94          (1.3%) 
 FY95           4.3% 
 FY96           5.1% 
 FY97           6.8% 
 FY98           7.7% 
 1/2FY99:       7.9%

Profit margins began falling at the end of FY93. Of course, it also didn't help that the SEC launched an investigation into Dell's controversial currency hedging practices in November 1992. The declining margins led to the train wreck of FY94, as inventory write-downs and a botched notebook computer effort produced genuine losses. Following the theorem, you would have bailed out (or avoided the stock) at the first sign of eroding margins, missing the high of $3 1/8 but cashing out for around $2 a share. By doing so, you would have escaped the plunge to a low of $7/8 by July 1993.

Year-over-year margin growth returned in the first quarter of FY95. At that point (May '94), Dell shares traded for about $1 1/2. The theorem suggests you should have bought back in at that point. (Using the stronger FY93 results rather than the lame FY94 as your baseline for comparison, margins were again growing by the fourth quarter of FY95, when you still could have acquired Dell at $2 1/2.)

By November '95, Dell had soared to $6 3/16 as margins continued to climb. The stock soon sank yet again, though, hitting a January '96 low of $2 7/8 in anticipation of a 4Q96 dip in margins. There was considerable turmoil in the PC market at the time, but Dell's relative weakness seems to have been caused by a parts shortage and stronger-than-expected demand from lower-margin consumer PCs. In any case, if you waited for the 4Q earnings to be reported in February '96, you would have cashed out of Dell at about $4 1/2, happy with your 200% profit on the shares you had purchased in May 1994.

The next quarter, Dell's EPS soared 58%, outdistancing the 44% sales gain despite just a 33% increase in net income. So while net margins actually fell, Dell made our rising margins screen thanks to aggressive stock buybacks. So one quarter after you sold, you now had a signal to buy. Of course, you would have had to pay up to get back into Dell, which was trading at around $6 by mid-May. Yet it would have been worth it.

Since that 4Q96 shortfall, Dell has appeared on our rising margins screen for ten consecutive quarters. The fourth quarter of last year saw the first substantial decline in average selling prices in the PC biz. And it was clearly a relatively tough quarter for Dell, as net margins fell slightly despite 55% sales growth. Again, though, the stock buybacks saved the day, leaving EPS up 59% and allowing us to hold onto Dell.

Now, a buy-and-hold Dell investor over this entire period would have done better than the investor keying off our margins theorem. Indeed, the buy-and-hold investor in Dell would be so rich that he wouldn't be reading this column. But not every company turns into Dell. The theorem hopefully allows you to avoid some disasters (we'll see if that's true), and to utilize an objective criterion for action. Also, the theorem would have led you to hold your Dell from your $6 repurchase in May 1996 through its recent high of $129, despite the naysayers.

Even assuming you bought Dell as late as May 1996, this would have been a fantastic, world-beating investment. Yet from what I can tell (and please correct me if I'm wrong), the mechanical models would not have picked up Dell quite this early.

What's the point? Since Dell has accounted for most of the excess returns of those model workshop portfolios that have delivered excess returns of late, it's worth thinking about how you could have spotted Dell and spotted it early. Assuming you were working from a list of relatively large consumer-oriented companies and investing according to the proposed theorem, a rising margins screen would have given you Dell early on in the company's growth. Second, while the momentum-oriented mechanical models gave you Dell, they also gave you a whole lot else that (barring some exceptions) didn't do all that well.

Next time, I'll look at Best Buy and the Gap to test the theorem some more. But to allow myself wiggle room, I must add now that I don't think the margins theorem will prove universally successful by itself. In part, that's because Dell's margins story offers just one little window on a complicated tale of how the computer industry's fundamentals have been transformed in just the last few years. Then again, if I can out-black-box heavyweight Robert Sheard, I'll happily change my name to O'Corrigan.

Check out the latest file updates for the Workshop:
New Rankings | 1998 Returns | New Database

What Happened to Robert Sheard?