|Shares Out. (in M):||2,500||P/E||0||0|
|Market Cap (in $M):||120,000||P/FCF||0||0|
|Net Debt (in $M):||3,100||EBIT||0||0|
|Borrow Cost:||General Collateral|
We are short Novo Nordisk on the basis that approximately 70% of its current revenue base faces a competitive onslaught from biosimilars, generics and competing product approvals over the next couple of years and yet the shares trade on 19x P/E, which is a premium to the peer group and suggests the market fails to discount this huge risk to future earnings.
The key tenet of our short thesis is that insulins/GLP-1s are amongst the most vulnerable drug classes for ‘biologic genericization’ following the expiry of their patents. Insulins/GLP-1s are some of the smallest biologic molecules, they are proteins in comparison to most biologics which are far larger and more complex monoclonal antibodies. As a result, the probability of achieving bio-similarity is higher and production costs are lower as industrialised microbial fermentation methodologies are used, compared to unscalable batch production for monoclonal antibodies which still require live hosts (usually mice). The large addressable market (diabetes is the #1 disease in terms of cost to the US healthcare system), attractive margin profile (we estimate insulin makers earn 97-99% gross margins in the US and around 70% internationally) and relatively low barrier to entry makes this a core target for generic/biosimilar manufacturers.
Biosimilar Insulin Launches
The first biosimilar (Lilly’s glargine ‘Basaglar’), launched earlier this year, is a biosimilar of Sanofi’s Lantus and competes directly in the long acting/basal insulin segment with Novo’s Levemir. Contrary to market expectations it gained significant formulary access in the 2017 negotiations and is widely considered a success. This has resulted in Novo offering significant price cuts for Levemir & Tresiba to maintain formulary access. Sanofi’s Lantus has lost out in this battle and is fighting hard to regain its market position (e.g. by offering some patients the product directly for merely paying the equivalent of their co-pay) which is pressuring prices further. We expect pricing and volume pressure to intensify as the branded companies fight to maintain formulary access ahead of 2019 when Mylan and Samsung are expected to launch biosimilar glargine products (these are expected to feature in the 2018 negotiation). By 2019 we expect Lantus and Levemir to be ‘genericised’ and overall industry revenues to be far below current levels despite scripts steadily increasing in line with long term demographic trends. Levemir and Tresiba account for approx. 20% of Novo’s 2017 revenue.
The second biosimilar to launch is Sanofi’s Admelog which is a biosimilar for Lilly’s Humalog and competes directly in the short-acting/bolus insulin segment with Novo’s Novolog/Novorapid. Sanofi announced FDA approval of Admelog last Friday. As it is still ‘in season’ we expect 2018 formularies to be adjusted. This will cause significant disruption to the Humalog/Novolog duopoly resulting in price erosion and eventually volume erosion as well. In 2019, Mylan is planning to launch a biosimilar Humalog and Samsung is also said to have plans for a 2019 launch. As with the basal segment we expect this to result in significant price pressure and an effective genericization of this segment by 2020. Novolog/Novorapid account for approx. 18% of Novo’s 2017 revenue.
Victoza going generic
Next we have GLP-1s, Novo’s supposed growth engine. The first GLP-1 was Byetta (sold by Astrazeneca). In 2010, Novo launched Victoza which required just one injection a day vs 2 for byetta and was until recently the only GLP-1 approved for use in combination with a long acting insulin. This has allowed Novo to gain significant market share and even after the approval of several new GLP-1s (most notably Trulicity by Lilly) which require injection only once a week, Victoza has managed to maintain approx. 70% share of the category as it is generally considered more efficacious. Here there are 2 upcoming threats. Firstly, Teva is expected to launch a biosimilar Byetta imminently. This is probably not a major deal given Byetta’s significantly inferior profile. The major issue is that Teva has already filed an ANDA for a generic Victoza which we understand they are planning to launch 2021 (Novo argues the relevant IP holds a year longer to 2022). To emphasize, this is a generic, not a biosimilar and hence, if approved, we expect immediate erosion of >90% of Victoza’s sales. To stifle this, Novo aims to transition Victoza users to its new once weekly version (Semaglutide) ahead of the expiry. Although we expect this to succeed somewhat, as Victoza is currently chosen over Trulicity which is a once weekly product as well, we would expect payers to push the generic Victoza to the front once on the market and hence at least a large part of existing Victoza sales will disappear. Victoza accounts for approx.20% of Novo’s 2017 revenue.
In addition to its diabetes franchise, Novo also has a strong haemophilia franchise. The leading product in that segment is NovoSeven. It is a low-volume/very high priced infusion and currently the only medicine on the market for patients requiring factor VII. This will change at the end of this year when Roche introduces AC910 which is expected to capture very significant market share in this segment. ACE910 has achieved significantly better efficacy. In addition, it will require less frequent dosing (once weekly vs 2-4 infusions per week in hospital for Novo) and will be administered as shot rather than an intravenous infusion into the bloodstream making it easier for patients and much cheaper for the healthcare provider. This is particularly important when you consider that new patients are paediatric. We expect the majority of NovoSeven revenue to move over the ACE910 over the next 2 years. It accounts for approx. 8% of Novo’s 2017 revenue.
Norditropin is a growth hormone. Up to last year it was another nicely growing product for Novo (double digit sales growth in 2015 and 2016). However, competition is intensifying significantly with a bunch of new competing launches in the segment (including a biosimilar by Sandoz) resulting in an expected sales drop of 20% for this year. Going forward this pressure will continue as the competitive environment continues and we expect additional new launches by new entrants Versartis, Pfizer and also Genexine and Ascendis. Norditropin still accounts for approx. 6% of Novo’s 2017 revenue.
Total 71% of current sales at risk
Summing up all of the above, at least 71% of Novo’s 2017 sales base will face very material pressure over the next 2-3 years. After giving them the benefit of some of the offsetting factors mentioned and some of the other products in the portfolio (e.g. Saxenda for obesity), we expect sales in constant currency to decline mid-single digits in the near term and this decline to accelerate post 2019 when the Mylan and Samsung biosimilar insulins hit the market and we approach/cross the Victoza IP expiration. Given the high incremental margins we expect profits to decline at 2x the rate of sales. Therefore, in a base case scenario we see EPS declining at a 10% CAGR over the next 3 years. We believe that when the market finally adjusts to all this this, the P/E multiple will contract to low teens (if not lower). If you apply 12.5x P/E to 2019 EPS of 14DKK, you get 40% downside from the current share price over the next 12-18 months…and this is the base case as in reality things could get significantly worse. If insulins and Victoza become fully genericized (we understand there are a whole host of insulin biosimilars in various stages of development in addition to the ones mentioned above) Novo’s entire earnings base will likely erode.
Addressing the risk factors: