Background and Introduction

Pharmaceuticals demand growth is the product of population growth, growth in units (e.g. prescriptions) per person, and growth in price per unit. Price per unit can be further broken down into real pricing power and changes to product mix (i.e. share gains by higher priced new products net of patent losses). With the exception of changes to mix, virtually all elements of drug demand growth are systematic, i.e. shared nearly equally by nearly all entrants. Mix changes are highly idiosyncratic (i.e. stock-specific). It follows that optimizing relative performance (v. peers) of a portfolio of pharma / biotech stocks implies basing stock selection largely on expectations of new product flow. As US real pricing power fades^[1], product flow determines a larger percentage of revenue growth, meaning the influence of product flow on stock selection is likely to become even more important with time

Having examined the relative behavior of therapeutics stocks v. peers in the periods before and after anticipated US product approvals, we’ve developed a set of simple trading rules that produce portfolios with as much beneficial exposure to new product flow as is practically achievable. These rules and the resulting portfolios are the focus of this note

Our portfolios and supporting materials (candidate companies for inclusion, current phase III pipelines, current NDA/BLA submissions, current holdings, buy notifications, sell notifications, transaction logs, and performance metrics) are available by subscription on Bloomberg, at our website (www.sector-sovereign.com), or by e-mail

Summary and Conclusions

Pharma / biotech stocks tend to outperform their peers in the periods immediately preceding and immediately following the anticipated approval of major new products. Expected performance is spread somewhat evenly across the pre- and post-approval periods; however a large percentage of total risk is concentrated right around the expected approval date. In an attempt to optimize risk-adjusted returns, we’ve developed a series of simple buy / sell rules intended to capture relative performance gains in the pre- and post-approval periods, while simultaneously reducing exposure to the disproportionate risks in the days immediately before and after an expected approval

We’ve established three overlapping portfolios of companies with pending product approvals; the portfolios vary only by the ‘relevance threshold’ applied to new products filed for approval. There are three ‘relevance thresholds’: 1%, 5%, and 20%. The percentage values refer to the threshold amount of a company’s consensus revenue forecast attributable to the anticipated new product; e.g. the ‘1%’ portfolio includes US-listed therapeutics stocks with new products representing at least 1 percent of the total company’s consensus revenue forecast for years +2 to +4

Exhibit 1 (first page) compares the risk / return characteristics of the three portfolios with the DRG, BTK, SP500, and our own ‘innovator index’^[2]; the period examined is November 2011 to present. Across this limited period, our portfolios produced risk-adjusted returns superior to each of the comparator indices. Exhibit 2 compares the performance of our previous rules^[3] to the same indices^[4]; however in this case the period examined is considerably longer, from 1996 to 2H2011. Under our ‘previous’ rules, companies were included based on a subjective determination of whether a pending new product was ‘thesis-relevant’; under our current rules inclusion is an objective matter of whether the pending product is a sufficiently large percentage of the company’s total sales forecast. Also, under our previous rules we adjusted the risk of our portfolios by varying the relative weights of companies held in the portfolio; the cap-weighted portfolio was the most conservative, and the equal-weighted portfolio most aggressive. All of our current portfolios are equal-weighted

These results suggest that building therapeutics portfolios with concentrated exposure to new product flow can produce superior risk-adjusted returns. We believe this is particularly true if buy/sell rules mitigate the risks concentrated in the days immediately before and after scheduled regulatory actions, as our rules are designed to do

Building Portfolios around Anticipated Regulatory Actions

As US real pricing power fades, product flow becomes responsible for an ever larger percentage of therapeutics companies’ revenue growth. This, plus our generally bearish view of systematic risks for therapeutics companies, argues for building portfolios that maximize beneficial idiosyncratic risks (e.g. new product flow), and minimize systematic risks. This implies a portfolio consisting entirely of stocks with pending or recent approvals of ‘thesis relevant’ new products; for each of our three portfolios we define thesis relevance as a product that represents at least 1 percent, 5 percent, or 20 percent of a company’s forward total (consensus) sales forecast

Eligible companies include all US-listed therapeutics^[5] stocks, with the exception of OTC-listed stocks having capitalizations below $1B (Appendix 1). The portfolios are overlapping, i.e. companies filing for US approval of new products whose consensus sales represent more than 1, 5, or 20 percent of the filing company’s total consensus forecast are included in all relevant portfolios^[6]

The periods preceding and following the scheduled regulatory action or ‘PDUFA’^[7] date are treated separately. Accordingly each of the three portfolios as defined by product threshold percentages (1, 5, 20) consist of both ‘pre-PDUFA’ and ‘post-PDUFA’ holdings. When we refer to the ‘1 percent’ portfolio we are referring to the ‘combined rules’ portfolio, i.e. the combination of pre- and post-PDUFA holdings

Pre-PDUFA

In the pre-PDUFA portfolios, stocks are included on the date a qualifying product is submitted for approval, and holdings are equally weighted. In the pre-PDUFA period extending from submission to 45 days preceding the scheduled regulatory action, the stocks are held long unless a final regulatory action is announced, or the scheduled regulatory action is indefinitely postponed; in either of these events the position is sold. In the period 45 days prior to the scheduled action date and the scheduled action itself, the stock is held long until the earlier of a final regulatory action, an indefinite postponement of the scheduled action date, a relative gain of more than a threshold magnitude^[8], or the scheduled action date, at which point the position is sold

Beyond the general observation that stocks with major pending regulatory actions outperform their peers, the rationale behind our specific pre-PDUFA rule is two-fold: First, relative performance gains are spread somewhat evenly across the period from submission to scheduled approval; however risks are somewhat concentrated around the scheduled action date. Second, in the period immediately preceding the scheduled regulatory action, we believe the stocks’ relative performances are more likely than not to be defined by expectations related to the product’s approval – thus if we see a substantial relative performance gain in the period between 45 days prior to the PDUFA date and the PDUFA date itself, we choose to realize the gain and limit our exposure to the more concentrated risks in the days immediately surrounding the scheduled decision. Back-tested^[9] across 275 regulatory actions over 15 years, our pre-PDUFA rule reduces both pre-PDUFA relative returns (Exhibit 3^[10]) and pre-PDUFA risks (Exhibits 4, 5^[11]). Our subjective impression is that the performance sacrificed by selling ahead of the PDUFA date is justified by the associated reduction in risk

Post-PDUFA

Stocks tend to outperform their peers following final regulatory actions on major products (both approvals and failures); however timing of entry is crucial. Product approvals tend to lead to periods of relative outperformance; presumably the market has tended to under-appreciate the revenue and/or earnings gains driven by newly approved products. Product rejections also tend to lead to relative outperformance; we interpret this as simple mean-reversion of stocks that were over-sold in the wake of products failing to gain regulatory approval. In both cases, purchasing on or shortly after a regulatory action tends to lead to relative underperformance, whereas purchasing ≥ 30 days following the regulatory action tends to lead to relative outperformance

The post-PDUFA portfolio consists of equally weighted stocks having received a final regulatory action on products whose consensus forecasts represent (or in the case of rejections, represented) at least 1, 5, or 20 percent of the consensus forecast for the entire company. Stocks are included on the earlier of a relative share price reaction (versus the therapeutic index) of a threshold magnitude, or a date 30 days after the final regulatory action. The pre-30 day relative share price reaction trigger is agnostic to the ‘direction’ of relative performance; the purpose of the rule is simply to indicate whether the stock appears to have had its ‘full’ reaction – positive or negative – to the regulatory action, as buying ahead of or during the share price reaction to a final regulatory decision appears to be value-destroying. Stocks included in the post-PDUFA portfolios are held until the earlier of a relative gain above a threshold percentage, or a period 365 days following the final regulatory decision

Current Portfolios and Performance

272 US-listed therapeutics stocks are eligible for inclusion in these portfolios (Appendix 1). We are aware of 76 novel products that these companies have filed for approval with FDA (Appendix 2)^[12] . The ‘1 percent’ pre-PDUFA portfolio currently consists of 20 long positions, all of which are in the submission to 45 day pre-PDUFA period (Exhibit 6). The ‘5 percent’ and ’20 percent’ pre-PDUFA portfolios current consist of 15 and 9 positions each (Exhibit 6, again)

The ‘1 percent’ post-PDUFA portfolio currently consists of 9 positions; the ‘5 percent’ and ’20 percent’ post-PDUFA portfolios consist of 8 and 5 positions each (Exhibit 7)

Since November of 2011, the 1, 5, and 20 percent ‘combined rules’^[13] portfolios have outperformed the therapeutic index by 27, 34, and 33 percent, respectively; the majority of this performance is driven by the pre-PDUFA holdings (Exhibit 8)

Risk-return characteristics of all three (1, 5, 20 percent) portfolios have been superior to all of the DRG, BTK, SP500, and an equal-weighted comparator index of all stocks eligible for inclusion (Exhibit 1, again). Risk-adjusted performance of the 20 percent portfolio has been inferior to the 5 percent portfolio; presumably this is attributable to the facts that the average position in the 20 percent portfolio is more volatile^[14], and the number of positions in the 20 percent portfolio is lower than in either the 1 or 5 percent portfolios

Exhibit 9 summarizes average holding periods and average turnover for the 1, 5, and 20 percent combined rules portfolios. A complete transactions log of purchases and sales over the past year can be found in Appendices 3 (purchases) and 4 (sales)