Jul 30 2013, 14:31 | 11 comments | about: GILD On Friday, the Street.com's healthcare journalist tweeted that two big banks (ISI and Deutsche Bank) sellside analysts in their reports to select institutional investors see peak sales of '7977, Gilead's (GILD) nucleoside analog drug or "nuc", as high as $6-12B! Granted, this could be their upside scenarios and not their base assumptions but it raises the question of how to contextualize this level of sales. In other words, how realistic are these sales numbers really? A quick google search reminded me this is actually an unprecedented level for a drug that is not likely to be used chronically by patients because it offers a potential cure to the disease- in this case to hepatitis C [HCV]. Most of us in the industry are familiar with bigger peak sales figures (as high as $13B for Lipitor at peak) for blockbuster drugs used chronically to treat conditions because they do not solve the underlying problem. This list includes drugs used to treat cholesterol, GI conditions, high blood pressure, and psychiatric conditions. In comparison, as my colleagues noted, the biggest non-chronic drug to date is Avastin- at $6B peak sales.
This is by way of introducing my real discussion of how to value GILD shares at this point. In my mind, the three important conceptual pieces for this discussion are: 1. the outlook for the HCV franchise, 2. the outlook for the HIV franchise, and 3. the value of the late-stage pipeline in cancer. In my own work, I have conducted sensitivity analyses of discounted cash flow (DCF) and P/E multiple valuations (I hope to get around to a sum of parts eventually) of GILD cash flows and earnings, respectively. Assumptions in my analyses are described below. Let me start to frame the discussion by saying that it is very hard to justify the current stock valuation based off of a pure DCF analysis of core HIV and very late-stage pipeline cash flows such as from 7977, unless one assumes a outsized peak sales number for 7977. By saying "hard", I mean that to get to $62 in my DCF, you need to assume nearly $7B in 7977 sales (recall we just said $6B was the record to date) in 2015 AND a favorable HIV scenario of a slow decline in branded GILD sales vs. generic competition. To help contextualize $7B in HCV sales further in terms of eligible patient numbers, GILD has highlighted that the number of HCV patients in the US that were able to be treated with conventional interferon-based treatments was around 150,000. By the way, this is not the number we really want to know, which is the number of HCV patients in the US and EU currently being actively treated by a medical professional, but it's a start. (On its 2Q earnings call, GILD said it is working on building a database to answer this question and may be able to disclose this number later this year.) Nevertheless, this $7B sales level would be greater than what Avastin achieved but not unreasonable if GILD's regimens can capture the lionshare of 150,000 eligible patients in the US - assuming a treatment cost of around $80,000/patient (or per cure)- with more HCV-infected individuals outside the US. However, the critical point I'm making is that a $7B sales potential is just the minimum to justify the current stock valuation. To see attractive stock upside from here in a DCF, one actually needs to see >$8B in 7977 peak sales, which is starting to reach to stratospheric heights for sales of a non-chronic drug (though not quite as fanciful as $12B).
At the moment, it is more common for professional investors to value GILD on a P/E multiple basis vs a DCF basis. The simple reason is probably that the implied '7977 sales look more achievable that way. If one looks at the GILD valuation on a P/E multiple basis and assume GILD earns $4/share in 2015, then the stock is trading at just under 16x, which is in line with the large cap biotech P/E multiple, and implies $6B in 7977 sales, which looks somewhat more achievable than the $7B implied by the DCF. The multiple valuation method should be more favorable than the DCF method, when a lot of sales peak early and erode in later years. In the case of 7977, most professional investors assume sales will peak quickly from pent-up patient demand and then erode dramatically because it should cure more patients than are newly diagnosed in a given year.
DCF / PE
PT Implied 7977 sales ($B) PT / Multiple 2015EPS Implied 7977 sales ($B)
$47 5 $48 / 16x $3.00 4
$62 7 $72/ 16x $4.50 6
$80 9 $109/ 16x $6.80 9
$113 12 $146/ 16x $9.10 12
The bottomline trade:
My call is still bullish on GILD into the 7977 launch (from here in $60s) and I believe GILD still deserves to be a sizable position in a long growth strategy. Sentiment and momentum remain strong and sellside earnings estimates are likely to be revised higher (higher pricing and patient numbers than expected), not lower, into the regulatory approvals. At the margin, as I said in my recent SA article on Vertex (VRTX), news on Thursday of the VX-135 setback has increased the scarcity value of 7977 (was GILD just lucky or smarter than everyone else to find the only safe uridine nuc?). However, I would likely reevaluate any long position once 7977 is launched. At that point, an important factor is that the cancer pipeline would also have generated further data readouts and regulatory milestones in the next 12-18 months to tell us how valuable idelalisib (aka GS-1101) could be long-term.
Details of my analysis:
For the HCV franchise, I assume four levels of sales ranging from $4B to $12B ($4B, $6B, $9B, and $12B). I chose $6B because it's the Avastin benchmark. I looked at $12B level because it's the revenue opportunity for treating 150,000 HCV patients at a cost of $80,000 per patient. I chose $9B simply as a number between those two reference points and then chose one sales level below $6B. In my DCF analysis, 7977 sales/cash flows are probability adjusted (I assume 80%) from the unadjusted levels (i.e., my $6B peak assumption becomes $4.8B with probability adjustment), as 7977 is not actually approved yet and pivotal trials of the 7977+5885/NS5a combination have not yet read out data.
For the HIV outlook, I assume several scenarios with differing rates of sales erosion to GILD's franchise from generic competition from 2018-2021 and beyond. The difference between the best- and worst-case valuation scenarios is substantial-~$15/share in my models. As a reminder, it is known that GILD's $11B HIV franchise will likely face one of the worst patent cliffs ever seen from 2018 to 2021. This would lead Atripla/Truvada/Viread sales to more or less disappear eventually and is the reason GILD has been working furiously on several life cycle strategies, already launching Complera and Stribild and still developing TAF in Phase 3 as the Viread replacement for new fixed dose combination regimens. Stribild and Complera are off to a good start but TAF is still key for retaining much of the market share for Viread containing regimens. I won't argue here whether or not the premise of better renal and bone safety will be compelling enough to favor TAF over generic Viread (available in 2018), the latter of which could be used with superior new agents such as Glaxo's dolutegravir (an integrase inhibitor). Regardless, there is room to debate how fast the sales erosion in the HIV franchise will occur depending on how relatively successful GILD's strategies will be.
For the late-stage cancer pipeline, currently, I assume idelalisib could be a $3B+ drug in 2020 but this is still highly risk adjusted (60%). I do not assume sales yet of 387 (aka CYT387 from the YMI acquisition) in myelofibrosis.
Otherwise, my DCF assumptions are fairly standard.
Link
This is by way of introducing my real discussion of how to value GILD shares at this point. In my mind, the three important conceptual pieces for this discussion are: 1. the outlook for the HCV franchise, 2. the outlook for the HIV franchise, and 3. the value of the late-stage pipeline in cancer. In my own work, I have conducted sensitivity analyses of discounted cash flow (DCF) and P/E multiple valuations (I hope to get around to a sum of parts eventually) of GILD cash flows and earnings, respectively. Assumptions in my analyses are described below. Let me start to frame the discussion by saying that it is very hard to justify the current stock valuation based off of a pure DCF analysis of core HIV and very late-stage pipeline cash flows such as from 7977, unless one assumes a outsized peak sales number for 7977. By saying "hard", I mean that to get to $62 in my DCF, you need to assume nearly $7B in 7977 sales (recall we just said $6B was the record to date) in 2015 AND a favorable HIV scenario of a slow decline in branded GILD sales vs. generic competition. To help contextualize $7B in HCV sales further in terms of eligible patient numbers, GILD has highlighted that the number of HCV patients in the US that were able to be treated with conventional interferon-based treatments was around 150,000. By the way, this is not the number we really want to know, which is the number of HCV patients in the US and EU currently being actively treated by a medical professional, but it's a start. (On its 2Q earnings call, GILD said it is working on building a database to answer this question and may be able to disclose this number later this year.) Nevertheless, this $7B sales level would be greater than what Avastin achieved but not unreasonable if GILD's regimens can capture the lionshare of 150,000 eligible patients in the US - assuming a treatment cost of around $80,000/patient (or per cure)- with more HCV-infected individuals outside the US. However, the critical point I'm making is that a $7B sales potential is just the minimum to justify the current stock valuation. To see attractive stock upside from here in a DCF, one actually needs to see >$8B in 7977 peak sales, which is starting to reach to stratospheric heights for sales of a non-chronic drug (though not quite as fanciful as $12B).
At the moment, it is more common for professional investors to value GILD on a P/E multiple basis vs a DCF basis. The simple reason is probably that the implied '7977 sales look more achievable that way. If one looks at the GILD valuation on a P/E multiple basis and assume GILD earns $4/share in 2015, then the stock is trading at just under 16x, which is in line with the large cap biotech P/E multiple, and implies $6B in 7977 sales, which looks somewhat more achievable than the $7B implied by the DCF. The multiple valuation method should be more favorable than the DCF method, when a lot of sales peak early and erode in later years. In the case of 7977, most professional investors assume sales will peak quickly from pent-up patient demand and then erode dramatically because it should cure more patients than are newly diagnosed in a given year.
DCF / PE
PT Implied 7977 sales ($B) PT / Multiple 2015EPS Implied 7977 sales ($B)
$47 5 $48 / 16x $3.00 4
$62 7 $72/ 16x $4.50 6
$80 9 $109/ 16x $6.80 9
$113 12 $146/ 16x $9.10 12
The bottomline trade:
My call is still bullish on GILD into the 7977 launch (from here in $60s) and I believe GILD still deserves to be a sizable position in a long growth strategy. Sentiment and momentum remain strong and sellside earnings estimates are likely to be revised higher (higher pricing and patient numbers than expected), not lower, into the regulatory approvals. At the margin, as I said in my recent SA article on Vertex (VRTX), news on Thursday of the VX-135 setback has increased the scarcity value of 7977 (was GILD just lucky or smarter than everyone else to find the only safe uridine nuc?). However, I would likely reevaluate any long position once 7977 is launched. At that point, an important factor is that the cancer pipeline would also have generated further data readouts and regulatory milestones in the next 12-18 months to tell us how valuable idelalisib (aka GS-1101) could be long-term.
Details of my analysis:
For the HCV franchise, I assume four levels of sales ranging from $4B to $12B ($4B, $6B, $9B, and $12B). I chose $6B because it's the Avastin benchmark. I looked at $12B level because it's the revenue opportunity for treating 150,000 HCV patients at a cost of $80,000 per patient. I chose $9B simply as a number between those two reference points and then chose one sales level below $6B. In my DCF analysis, 7977 sales/cash flows are probability adjusted (I assume 80%) from the unadjusted levels (i.e., my $6B peak assumption becomes $4.8B with probability adjustment), as 7977 is not actually approved yet and pivotal trials of the 7977+5885/NS5a combination have not yet read out data.
For the HIV outlook, I assume several scenarios with differing rates of sales erosion to GILD's franchise from generic competition from 2018-2021 and beyond. The difference between the best- and worst-case valuation scenarios is substantial-~$15/share in my models. As a reminder, it is known that GILD's $11B HIV franchise will likely face one of the worst patent cliffs ever seen from 2018 to 2021. This would lead Atripla/Truvada/Viread sales to more or less disappear eventually and is the reason GILD has been working furiously on several life cycle strategies, already launching Complera and Stribild and still developing TAF in Phase 3 as the Viread replacement for new fixed dose combination regimens. Stribild and Complera are off to a good start but TAF is still key for retaining much of the market share for Viread containing regimens. I won't argue here whether or not the premise of better renal and bone safety will be compelling enough to favor TAF over generic Viread (available in 2018), the latter of which could be used with superior new agents such as Glaxo's dolutegravir (an integrase inhibitor). Regardless, there is room to debate how fast the sales erosion in the HIV franchise will occur depending on how relatively successful GILD's strategies will be.
For the late-stage cancer pipeline, currently, I assume idelalisib could be a $3B+ drug in 2020 but this is still highly risk adjusted (60%). I do not assume sales yet of 387 (aka CYT387 from the YMI acquisition) in myelofibrosis.
Otherwise, my DCF assumptions are fairly standard.
Link