Quantifying Illumina's Oncology Upside

Illumina ("ILMN") is the dominant provider of next generation sequencing (NGS) platform. The company provides an entire ecosystem of machines, consumables, software, and services. It's a bit of a razor and razorblade model, with sequencing consumables making up about 60% of revenue.

Analysts tend to worry about how many machines the company will sell this year and next year. I prefer to think in terms of end markets - after all, that is what drives instrument and consumable sales.

I will focus on oncology because I believe that will be the main contributor to the company doubling its revenue.

Oncology is not one, but three separate growth vectors. The three are 1) therapy selection, 2) monitoring, and 3) screening/testing. A couple slides from IR here give you a sense of relative market sizes.  The first is from Illumina, the second from Guardant Health.

The units are different, but we can tell that 1) all three are big markets, and 2) for now, therapy selection is the smallest but most penetrated market (and even then still very early innings); early screening is the largest but least mature market.

Let's think about how ILMN will participate in each of these markets.

1) Therapy selection. Precision medicine is revolutionizing medicine. Old medicine is a lot of trial and error - we don't really know how this drug treats this disease, but we know empirically (through clinical trials) that drug A is correlated with improvement in symptom B, so doctors prescribe it.

New medicine is different. We want to know what exactly is causing cancer. How does it impacts the specific patients in question? How will this specific patient react to this drug?

This is where DNA/RNA sequencing is required. Sequencing of patients' tumor samples allows us to move beyond some homogeneous view of cancer and into specifics of how the tumor works and how best to treat it.

That is the essence of companion diagnostics ("CDx"). CDx is in growth tornado mode and most, if not all, developers of these CDx does sequencing with Illumina machines. An example is Foundation Medicine which, through its parent Roche, has a partnership deal with Illumina to create more companion diagnostics.

Note that even companies who are skeptical of use of sequencing in early screening markets admit that sequencing makes sense for therapy selection. See below note from Exact Sciences (which uses PCR technology for screening, but think sequencing makes sense for therapy selection).

2) Monitoring. I believe the paradigm that prevails will work like Natera's Signatera (see diagram below).

Note that the process starts with sequencing of Tumor tissues. This is where Illumina's platform comes in. Natera would then use the data obtained from that step to create a personalized PCR assay, which is then used for on-going blood test monitoring (less invasive than requiring a solid tumor sample).


3)  Early Screening. I'm not convinced that sequencing will be required for this market and will (at least for now) not give credit to Illumina for this.

Screening is about testing for known diseases, for this PCR would do a cheaper and faster job. Cost of sequencing will likely come down a lot, so I think ultimately it comes down to the speed advantage of PCR.

I also get the sense that PCRs can be more distributed location wise. So instead of sending samples to some central lab for sequencing, it's faster to just have some sort of PCR at a location closer to patients.

To summarize this section, I believe Illumina's sequencing will 1) own the therapy selection market,  2) get parts of the economics in the monitoring market, and 3) get none of screening market (that's my assumption for now anyways).


Quantifying the Upside

So how do we quantify all of this for Illumina? My approach is figure out how much each of the 3 oncology vectors contribute to the company's consumables revenue right now, then scale those numbers up to some estimate of eventual market penetration.

From ILMN's 4Q19 transcript, we know that:

1) Oncology is about 20% of sequencing consumables.
2) Clinical versus research split is about 40%/60%.
3) Within oncology, therapy selection is by far the largest driver. Monitoring is nascent and screening is even earlier.
4) Oncology therapy selection is about 8% penetrated.

So if I take the $2.1bn of of sequencing consumable revenue in 2019, attribute 20% to oncology and 40% to research, that get us to $166mm of oncology clinical consumables revenue. I'll just assume all of that is therapy selection and $0 from monitoring and screening.

This $166mm is the basis of our analysis. I'm going to throw out some numbers just to demonstrating the thinking process and ball park the upside. The output is below.

Allow me to explain.

Ok, so $166mm of oncology treatment selection revenue for 2019. That market is about 8% penetrated now, where do we think it'll be in a few years? I'm assuming 60% here, so that scales up to $1,245mm of treatment/therapy selection revenue 5-10 years out (as shown in table above)

How about monitoring revenue? From the TAM presentation slides above, both Illumina and Guardant Health pegs the monitoring market at about 2.5x that of therapy selection. But remember, I think the Signatera paradigm (as explained above) will become standard, and Illumina only participates in the upfront sequencing and not the on-going monitoring.

Let's say 1/4 of the value from each monitoring treatment accrues to the sequencing platform provider - Illumina. (Plug in your own assumptions).

So my estimated monitoring revenue is the $1,245mm treatment selection revenue * 2.5 * 1/4 = $778mm.

I gave no credit to oncology screening opportunities, but that could change. Finally, for oncology research revenue I just take the present estimate of $249mm and give it a 3x to reflect the early inning nature of overall oncology market.

Layering in other assumptions for NIPT market and other sequencing consumbles, I can see Illumina's sequencing revenue go up to $7bn in a few years. The bulk of these gains come from oncology market shifting toward precision medicine and exploding upward.

Conclusions - about Valuation and Risks

The stock is around $270 at the time of this analysis. I get to about $2.5bn EBIT 5-7 years out and ILMN is thus trading at <15x EV/EBIT in years 5-7 (with interim cash flows lowering that EV). This is a decent price for what is essentially a monopoly that participate in growth markets.

The analysis here implies that Illumina's revenue will re-accelerate at some point - because the end markets will explode upwards.

Those revenues are Illumina's to lose, provided that the company retains its dominant competitive position. For now, Illumina is further entrenching its ecosystem by partnering with Roche and Qiagen to create 3rd party tests.

Risks

The only challenger on the horizon is Thermo Fisher with its Ion Torrent systems. There is also some small chance that the short-read nature of ILMN's machine could become a problem later. (Their failed deal with Pacific Biosciences would have given them strength in long-read market and remove this risk, but the deal failed).

Why Chipotle Stock (CMG) is So Expensive

I read Chipotle ("CMG")'s 10K for the first time the other day. Where have I been!

For years I have ignored the stock because of its very high multiples. I regret that very much. But better late than never.

Here I will share my notes. The 10 second summary is that Chipotle has decades of growth runway ahead of it, and it has not even started pulling some of the upside levers yet.

This is clearly a "buy every dip" stock.

Notes

• Per store revenue and margin normalization - this has been going on the past few years.
• "We are confident we can get back to volumes in the $2.5 million range, which is - those are the volumes we had just a few years ago."
• "We think we can go beyond that. The idea that digital was only 5% or 6% of sales back in 2015 when we hit these peak volumes, digital is now at 18%. It's the fastest-growing part of our business. So we have assets today that we didn't have back when we were doing $2.5 million."

• Margin expansion - increased digital sales drives operating leverage.
• "Ultimately, long-term guidance, of course for AUVs and margins to rise in concert. 
• Management has noted each incremental $100,000 in sales volumes translates into 100 basis points of restaurant margin. 
• Chipotlanes. This is CMG's mobile pick up lanes. This requires a shift in real estate strategy as most of its current real estate is not end-caps but in-line sites. The newer restaurants make good candidates for Chipotlanes though.

• US store expansion. They have about 2,600 stores, mostly in US. Management is confident that it can do 5,000 stores in U.S. I believe them.

• International presence. Almost none right now!
• They only have 39 international stores in Canada, Europe. 
• CMG has no presence in Asia! From my own anecdotal observation, Chipotle will likely be hit among East Asian countries. I have personally observed people who pretty much only eat Asian food becoming big fans of Chipotle's burrito bowl. This is because those cultures are accustomed to rice based diets, and Chipotle's burrito bowl is a familiar format. 

• Franchising. They are doing none of this and they don't need to. For now quality control is key so CMG should own the stores. Nevertheless franchising is something that can really juice ROE down the line.

• No financial leverage. This is another lever that Chipotle can pull when it matures - perhaps 30 years later!

Chipotle is basically sitting on a gold mine, it just have to not mess up! As long as the company keeps up its quality and reputation (which management is carefully doing), the market is theirs to lose. Unlike the burger/pizza/fried chicken fast food market, Chipotle's market is vast with decades of runway ahead, and CMG dominates.

FNF Diworsification (Acquisition of FGL)

On 2/7/2020, Fidelity National Financial ("FNF") announced that it is acquiring FGL Holdings (ticker "FG"), an index annuities provider.

I really do not like this deal.

The Stink of Life Insurance

In fact, I don't like anything that has a whiff of life insurance to it.

Life insurance and annuities are super long duration contracts and your P&L involves projecting out 20 years plus. That high level of of uncertainty means your financial statements are basically made up of layers and layers of assumptions (mortality/longevity, interest rates, equity index levels...etc).

This is why life insurance peers like Prudential (PRU), Metlife (MET), Lincoln (LNC) and so on all trade around 10x P/E. Don't let anyone tell you it's all about the low rates depressing investment income!

No, the very business model of life insurance and annuity is shit, period.

I would actually frown upon growth in this business, as growth would indicate the company is taking on more risk to bring in more business.

Now, I'll admit FGL's index annuities are less risky than the notorious variable annuities with guarantees. In those old GMDB/GMIB/GMDB products, policy holders invest in stock funds and the companies guarantee some minimum amount of return. These companies essentially sell a giant put option, exposing themselves to egregious losses during down markets. Index annuities, on the other hand, are newer derivative products (yes that's what it is). They are less risky because issuers essentially buy a bunch of call options on equity indices, and pass through the benefit to policy holders. Buying calls is less risky than selling puts!

A couple diagrams below show my understanding of how these products work. Notice that both the older GMxB and the new Index Annuities (FGL's products) give customers limited downside, but the latter incurs vastly better risk from insurer perspective.
 

So yes, FGL's annuities are much less risky than those notorious products of old. Still, over the life of a policy a lot of stuff can go wrong. FGL sells a derivative product with all sorts of market risks. I do not trust the financials.

Frankly, I doubt that FGL will ever shake off the stink of life insurance and the black box/ high risk stigma associated with it - even if rates go up. In other words, FGL will likely be a low multiple business forever. Even if it has high growth.


Is FNF Serious?

The first question is: is this actually a strategic acquisition, or it's just Bill Foley doing what he does - bring in some company only to spin it out later?

I told you above I hate the annuity business. So naturally, I hope it's the latter.

"Strategic acquisition" would be a big problem. As a shareholder, I don't want to see FNF deploy its abundant free cash flow toward growing a unrelated and shitty business that will never fetch a high multiple!

Unfortunately, the the 2/7/2020 conference call to discuss the acquisition seem to indicate otherwise. Management spoke of FGL as a strategic diversification and brought up examples of acquisition benefits, all of which are questionable.

As an example, they talk about FGL business smoothing out the combined company's exposure to interest rate changes. This is hogwash and they know it. FNF's refi business are already burned out from years of low rates and can't get hurt much more from higher rates. Also, it's not like FNF doesn't have its own investment operation that will benefit when rates go up!

Management also argued that FGL can benefit from FNF's bank relationship. In that very same call, FNF management actually backtracked from that assertion when challenged by analysts. The benefits will be limited to small time distributions.

I came away from the 2/7/2020 call feeling unsettled, but still hoping this is just Bill Foley playing the spin-off game.

Then came the 4Q19 earning call on 2/14/2020.

It Gets Worse!

In the 4Q19 call, FGL's CEO Chris Blunt tried to address some of my concerns above and tried to argue that FGL is not quite a life insurance business. I think he failed.

The first point:

"We are much more of a spread lender where we can reprice our liabilities on a regular basis".

This is somewhat valid and very important. It corroborates my earlier point that index annuities are less risky than the older variable annuity products.

Repricing liabilities is important because that shortens the duration, making them less sensitive to key macroeconomic factors.

That lessens the pain but doesn't make it go away. I would rather the company NOT take on these liabilities at all! (per FGL's 10K for 2018, liability duration is ~6.2 years).

The second point about improving credit quality brings back nightmares. AAA rated CDO-squared anyone? After the 2008 debacle, how anyone can still equate credit rating with actual risk is beyond me.  

Some of my other notes from the 4Q19 call:

• Talks about FGL doubling its AUM in 5 years with resource of FNF, could be 50% of FNF's overall earnings. (sounds awful!)
• Apparently FGL has ambition in pension risk transfer. (PRT = WTF!!!)
• Talks about giving Blackstone more money to manage. 
• Talks about FGL increasing investment yield without compromising on risk - by switching from BBB corporates to higher rated ABS and CMBS.

Good Fucking God!

I even get the impression Bill Foley is looking at FGL's Chris Blunt as some sort of successor. The latter is not exactly young, But it's hard not to get that impression when Foley talk about FGL growing to 50% of overall company earnings, giving Chris Blunt more money to manage, and basically let Blunt talk nonsense like chasing yield with structured products and pension risk transfer.

FNF is like "yeah it's fine, it's just a spread business". First of all, that's suspect. Unlike your traditional bank lenders with exposure to rates and credit, FGL's business have exposure to rates, credit, equity, lapse rates, and longevity/mortality. It's a spread business in the generic sense that any business is a spread business because it has revenue and cost of goods sold.

Second, even if that's true, why would I trade a steady service business (which is what title insurance actually is) leading an oligopoly, with a "spread business" that has little entry barrier?

Conclusion

I suppose one can argue "of course Foley has to talk like it's a strategic acquisition, of course that's what he says now. Just wait a couple years and he'll spin it out, just watch".

Even if that's the case, FGL is not like FNF's past acquisitions. Black Knight, Ceridian...etc, these are growth companies that has a ready market when the time comes for exit. I don't see that for FGL - it's just not a high multiple business.

After several years of holding FNF (and as my largest position the past 2 years). I will have to exit or at least cut down drastically.

Thankfully it's a long weekend now. I will have time to sleep on it.