NORMALIZED EARNINGS ANALYSIS "Normalized earnings analysis", or "earnings power" analysis is one of the oldest tools in the fundamental analyst toolkit, discussed in depth by Ben Graham in Intelligent investor over 70 year ago. Despite it's longevity, I often find myself
surprised by how much of the analyst community focuses on valuation on near term earnings vs. the much more important "normalized earnings" valuation. I made this mistake myself early in my career. Looking at a stock that seemed "expensive" on near term earnings because the FY1
or FY2 P/E was high without thinking critically about the true underlying drivers of value, I proudly (and dumbly) considered myself a value investor because I didn't want to pay a high headline FY1 P/E.
TSLA is a good example of this dynamic. In 2019 TSLA seemed like an absurdly expensive stock at 3,700x P/E. But in stock selection, the "E" (earnings per share) is more deterministic than the P/E in most cases. TSLA's 1c EPS went to a ~$12.50 estimate by for 2023, and a stock
that looked like it was trading 3,700x (on 2019 EPS) was actually trading at 3x P/E (on forward '23 estimates). This framework is a commonly applied hedge fund framework. Philippe Laffont from Coatue elegantly explained his normalized earnings analysis on Twitter, citing
how Apple traded 30x P/E in '07 but only 3x five years out. The key is building a more accurate 5-7 year view, and the valuation is a simple byproduct of that fundamental view. The P/E formula is simple, but on which "E" we apply the P/E is the trick.
A quick primer on multiples. Investors tend to apply multiples as a "short hand for a DCF" or an intellectual proxy for payback period. Multiples are a blunt, imperfect tool that embed many assumptions, but ultimately are a more helpful applied tool that staring down
a multi-decade DCF and trying to figure out what 15 year out FCF will look like (and in many DCFs a terminal multiple is applied after 5-10 years anyway, which captures the majority of value, so why even go through that dance?). In practice, multiples are applied more often than
DCF's for price target derivation because they are easy to implement and offer easy comparability to other assets and the history of the target asset (though DCFs do have a role). But it is important to understand the pitfalls of a multiple. The largest pitfall, in my opinion,
is that when you apply a multiple, you are making an implicit assumption that the target metric of that multiple sticks around for while. If I apply 15x to a FCF stream that grows 5%, I am saying that FCF stream will be around for 21 years (if it grows 7%, 17 years, if it grows
3%, 29 years). So for simplicity sake, anything you apply a multiple to, assume it is going to 1) grow, 2) be with us for ~20 years. The market should understand this, right? Umm, no. As a 15+ year participant in markets, to me one of the biggest mistakes the market gets wrong
most of the time is the "peak on peak, trough on trough" framework of capitalizing a temporary tailwind or headwind (go study PTON over the last 3 years for proof, or a whole host of COVID winners). So when we think about applying a P/E in an normalized earnings analysis, we
must be very disciplined in applying a multiple on a clean baseline. We cannot multiply anything temporary, or else the temporary nature of those items will distort our valuation. I encourage you to normalized for four buckets. Revenue. Strip out any big revenue headwinds/
tailwinds to create a clean baseline of durable, go-forward revenue. Margin. Think about what the "normal" margin of the business should be. Don't capitalize a peak or trough margin, and don't capitalize an immature margin structure that can normalize higher.
Use an "end state" framework and ask yourself "how profitable can this business be at maturity"? Use mature comps in your sector as a proxy. Once you have identified the durable baseline and have a research-driven view of the future, a simple model can suffice.
Hedge funds like to build highly complex overly detailed models. But often the buy-signal on a name could be sketched out on a cocktail napkin (as Apple likely could have been in '07). My advice - embrace the complexity of the model but don't miss the critical pivot to the
simplicity of what will drive the stock. A simple 5-7 year P&L to identify out year normalized earnings can be the foundation for a fund-making position. I've seen a few ways of thinking about valuation on normalized. Valuation on normalized. What is the P/E level on the
5-7 year out number. If it smacks you in the face as cheap, it's probably cheap (i.e. Apple at 3x, TSLA at 3x). If you have to think about it, it's probably not cheap. What is my end state valuation, applying a mature P/E to mature EPS, then discounting that end state
valuation back by the number of years beyond a normal 12m framework at an internal cost of equity (generally I've used 15%). This can be helpful to benchmark upsides across a portfolio of stocks. What is my IRR/CAGR hurdle, and does waiting this period of time exceed that
hurdle. The Tiger complex is wired to hunt for the "three year double", with the thinking that any stock that can compound at mid-20%+ over 3+ years is a a fabulous place for capital. Does your normalized analysis drive you towards that CAGR?
Is normalized analysis a magic bullet? Absolutely not. This framework does provide a better EPS baseline to capitalize, avoiding the common market pitfall of "peak on peak, trough on trough" and as an analyst learning this process and framework will help you speak the language of
your PM. "Normalized earnings bets" can be volatile. Who knows what EPS will look like in 5 years? A lot can happen. The stocks of these bets can be more volatile than average as if duration is high (i.e. value baked into the future), perception shifts on the baseline can bring
big moves in the stock. As always, you get paid to be right, and punished if you are wrong. This framework has been most helpful to me in four scenarios. Early stage. Of a business, or of a product. TSLA & AAPL. The "goodness" isn't baked into near term earnings and a
longer lens is required. Change. A business model change, margin enhancement story, etc. Not reflected in near term earnings but requires a "look through". M&A / synergy realization story. Near term is messy, but looking out 2-3 years the pro-forma number is compelling
and balance sheet is normalized. Are we compensated to ride through the integration? Temporary impacts. A bolus of COVID-demand. An inflation driven margin pinch. Normalize that. What's the valuation? What's the R/R? Hope that is helpful!