Goldman's theory of "Normalized EBITDA Margins" regarding semi-equipment stocks:
An excerpt from GS report on 1-6-2004
We continue to believe that the stocks will ultimately peak for the cycle when AMAT reaches 15% EBITDA margins, our proxy for normalized industry cash flow. We believe that it is therefore too late in the stock cycle to be overly aggressive on the group.
Although we believe that margin expansion and upward estimate revisions during earnings season could drive a final upward move in the semi equipment stocks, we believe that it is too late in the stock cycle for investors to be significantly overweight the group. This view is driven by our belief that Applied Materials is likely to reach 15% EBITDA margins during its January-quarter, which will be reported in February.
The next logical question investors ask is, why are normalized EBITDA margins an appropriate metric to use as a sell signal? Following, we explain why we believe this methodology makes sense: -We believe the market ultimately will and should gravitate toward normalized numbers. At the bottom of a cycle, bears love to use depressed EPS to highlight that stocks are expensive-and at the top of a cycle, bulls do the same with unsustainably high EPS to argue that stocks are cheap. Both arguments ultimately lead to faulty conclusions. The logical answer to this dilemma is to value stocks on normalized (i.e., sustainable) levels of cash flow. That said, don't just take our word for it. We believe that the market agrees with us, as we discuss following. -The market is smarter than one thinks and ultimately "gets it right." Although the marginal investor/analyst often gets carried away at the bottom of a cycle arguing that an industry is going away, and the same marginal investor/analyst often gets overly bullish about long-term secular growth rates at the top of the cycle, the market as a whole 'gets it right' more often than not. Valuations ultimately do tie in with economic reality-even in the semi equipment space, in which investors and analysts (including us in the past) get carried away by isolated metrics such as sequential orders. For example, consider Applied's July 2002 quarter. In retrospect, many investors argue that the stocks were overvalued when Applied hit $29, but a simple EBITDA calculation shows that at that time, Applied reported an EBITDA margin of 16%. Because cyclical stocks tend to peak at normalized EBITDA levels and Applied peaked at 16% EBITDA, we would argue that the market 'got it right' in the summer of 2002. -Finally, higher-than-normalized cash flow is, by definition, not sustainable in a pure cyclical business. Normalized EBITDA (cash flow) represents the midpoint between the peak cash flow and the trough cash flow. As long as a cycle has begun to bottom, EBITDA margins that are below normalized levels imply that business will eventually trend upward toward the normal levels of profitability. During an upturn when a company moves beyond its normalized cash flow, the growth by definition becomes unsustainable and therefore investors should (and do) sell the stock. Initially, stocks stop outperforming, later stocks stop performing(i.e., they don't go up at all), and then stocks start falling.
How then do we define normalized EBITDA margins for semi equipment? We attempt to derive a normalized margin level for the next cycle by using Applied Materials as an industry proxy. Our analysis is as follows: -We calculate historical EBITDA margins on a calendar-year basis for Applied and then estimate EBITDA margins for calendar 2003 and calendar 2004 based upon our industry growth assumptions of 10% in 2003 and 25%-30% in 2004. We normalize the EBITDA margins over three-, four-, five-, and six-year periods. We normalize the data to smooth the volatile cyclical fluctuations and to mitigate any forecast error that we may have made for 2003 and 2004.
We then plot each set of normalized margins and fit a line to determine the correlation between the data. We find that six-year normalized margins are most highly correlated for Applied (the R2 is 0.84). We determine the equation for the line fit to the six-year normalized margin data of y = -0.012x + 24.4. This equation implies that with each year, x, the normalized margin declines by the slope of the line, in this case by -1.2%. We apply this equation to determine normalized EBITDA estimates for 2004 and 2005 of approximately 16% and 15%, respectively. Because we believe the next peak will occur in 2005, we believe that 15% represents the normalized EBITDA margin level for Applied, and therefore, the industry.
Interestingly, this estimate ties to our bottom-up peak margin analysis for Applied. We estimate that Applied can achieve an operating margin of 30% on a quarterly basis at the peak of the next cycle (slightly below the last peak of 31%). As we highlighted earlier, we believe that Applied's EBITDA margins troughed in the fourth quarter of 2002 at a little more than 5%. The average of 30.0% and 5.0%, or 17.5%, is not too far off our estimated normalized margin of approximately 15.0%. So what about valuation? Given that cyclical stocks peak when an industry reaches normalized levels of profitability, does that mean that investors should just ignore valuation? In a perfect world, Applied Materials would reach our estimated intrinsic value of $20 (based on our normalized free cash flow, discounted peak earnings, and discounted cash flow analyses) in the same quarter that the company achieved normalized EBITDA margins. However, things are not always that lean. Let's look at an example of what happened with Applied's stock price during the head-fake rally in 2002. The stock peaked at $29 in May 2002, roughly $10 above our normalized free cash flow based estimated intrinsic value. While many wondered why valuation was so high during last year's run-up, it just so happened that Applied had 16% EBITDA margins in the April quarter, essentially in line with our normalized EBITDA margin estimate of 15%. Clearly, cyclical stocks can undershoot and overshoot intrinsic value. To underscore this point, if stocks never deviated from their intrinsic value, then Applied would not have declined to $10 last October. Cyclical stocks can and do trade above fair value as an industry is trending upward to normalized levels of free cash flow. Once the industry reaches normalized levels of free cash flow, the stream of cash flows are by definition unsustainable and stocks stop rising as investors rightly stop buying stocks of companies that have an unsustainable stream of cyclical cash flows. We therefore place a greater emphasis on normalized EBITDA margins than on intrinsic value.
ANYBODY HAS ANY THOUGHTS ON THIS? |
