Can Shuffling the Deck Create Growth?

eytan
Print Friendly
Share on LinkedIn0Tweet about this on Twitter0Share on Facebook0

Richard Evans / Scott Hinds

203.901.1631 /.1632

richard@ / hinds@sector-sovereign.com

May 24, 2011

Can Shuffling the Deck Create Growth?

  • Dividing PFE as it stands (‘total-PFE’) into two or more entities, for example one having older off-patent products (‘established-PFE’) and one having newer products and R&D (‘core-PFE’), infers that such a division either makes core-PFE a growth company, or opens the door to this outcome over time
  • The sales-weighted average age of products on market for core-PFE is just six months less than total-PFE; this suggests core-PFE’s immediate prospects for sales growth are not much different than total-PFE’s
  • The number of Phase III and filed products per dollar of sales for core-PFE is only slightly better than for total-PFE – total PFE ranks 5th among large-caps for both Phase III / sales and filed / sales; core-PFE ranks 5th on the former and 4th on the latter. This suggests prospects for mid-term growth are only slightly improved
  • Most importantly, core-PFE’s R&D / sales ratio is more than 2 percent below the industry average, and at least 4 percent below the spending level necessary for zero real sales growth. Core-PFE is not a growth company, and cannot become one without both higher absolute levels of R&D spending and higher dollar-for-dollar levels of R&D productivity
  • Reducing the sales base to facilitate growth only makes sense if firms are likely to generate sufficient new product flow to create growth. As modeled, core-PFE loses products faster than it creates them. Ironically, firms that lose more products than they create benefit from being larger, since the larger sales base cushions the volatility associated with large product losses
  • Beyond falling well short of the presumed goal of creating growth, divestiture faces a number of practical hurdles: 1) divestiture would tend to reduce asset-utilization rates in the manufacturing base; we estimate rising utilization as a consequence of mergers has reduced COGS / sales by roughly 4%; 2) established-PFE has nearly half of total-PFE’s emerging market sales; while we’re skeptical about the long-term prospects in the EM’s, they do contribute to near- / mid-term growth; and 3) because established-PFE presumably is in ‘planned-decline’ with no investment in R&D we presume it must pay most of its earnings (>= 90%) as dividend; because core-PFE is not a growth company it almost must offer a competitive payout ratio (+/- 50%); together, the combined dividend necessary to support both siblings is larger than that required for the parent

Whether a pharmaceuticals conglomerate creates value by breaking itself into smaller pieces is a multi-layered question, the simplest layer being whether economic gains from the break-up outweigh associated costs. We use PFE as an example in this call for the obvious reason that it has hinted at such a break-up, though we believe the conclusions can be generalized

Economic gains might come in either of two forms – pricing efficiencies with respect to the resulting equities[1], and/or operating efficiencies in the resulting companies. For the moment we’ll focus on the latter. Few operating efficiency gains come from reducing a company’s scale. An exception in the case of a company split along lines of older v. newer products might be that aggregated buyers are then less able to use the conglomerate’s weaker products as negotiating leverage against its stronger products – i.e. less able to demand the company give price breaks on monopoly / oligopoly products in order to avoid having the buyer reduce its purchases of the company’s less-differentiated / commodity products. Another important potential exception is R&D productivity: empiric evidence suggests an ‘inverted-U’ relationship between pharmaceuticals R&D productivity and firm size, so reducing the size of very large R&D operations might produce efficiency gains. The problem here is that in dividing along the lines of older v. newer products, one would expect all of the R&D spending to go with the newer products. Thus while the new product company is in fact smaller than the parent, the R&D spend at the new product company is roughly the same, meaning scale-related gains in R&D efficiency are unlikely[2]

Beyond transaction costs associated with a break-up, we believe that the most important break-up cost is the loss of certain scale-related efficiencies. Most obvious, smaller companies resulting from a break-up should have less negotiating leverage over suppliers than the conglomerate parent, so presumably input costs will rise more quickly for all of the smaller offspring. Less obvious, but we believe far more important, are non-purchasing scale-related operating efficiencies, particularly in manufacturing. For decades, pharmaceutical merger synergies have come primarily from an expansion in NEWCO gross margin, which itself has come from increased asset utilization in the NEWCO’s manufacturing base. PFE has been a primary beneficiary of this dynamic, which raises the important question of how two PFE offspring can be expected to operate separate manufacturing bases and still maintain present high levels of asset utilization (Exhibit 1)

Thus on first look, from a preliminary / high-level perspective and without the benefit of specific guidance as to how PFE might be divided, it’s challenging to see why break-ups should raise the siblings’ earnings power relative to the parent. In fact, our bias would be that the siblings would tend to lose relative earnings power

This leaves the question of whether the siblings’ equities would, even for the same earnings power, be priced more efficiently (hopefully higher) than the parent’s. We recognize that ‘pure-play’ equities allow investors to more precisely and efficiently compose portfolios; though because investors are at least partially able to hedge ‘non-core’ operating exposures that conglomerates bring to a portfolio, we believe the equity pricing efficiency gains of moving to two or more pure-play siblings are small enough to be ignored. In the specific case of PFE, the subtext of a break-up is that the high-yield / value character of the parent equity can be transformed into two separate equities with characteristics that are somehow more desirable, and thus more efficiently (higher) priced – i.e. one ‘core-PFE’ growth equity, and one ‘established-PFE’ value equity. This raises the simple question of whether there is a growth company hidden beneath PFE’s older established products; according to our analysis there is not

A rough proxy for immediate relative growth potential is the sales-weighted average age of products that are currently on-market. In its present form PFE’s sales-weighted average product age is 10.9 years, versus an industry average of 11.4. If we remove established products, PFE’s sales-weighted average product age falls to 10.4 years, a six-month reduction that moves PFE only one spot in rank (Exhibit 2)

A proxy for longer-term growth potential is the R&D to sales ratio, preferably adjusted for company-specific R&D productivity. PFE recently announced its intention to cut R&D spending; if we take the absolute spend implied by consensus (which incorporates PFE’s lowered spending guidance) and compare this to a ‘core-PFE’ sales base after removing established products, the ‘core-PFE’ NEWCO has an R&D / sales ratio of 15.3% in 2011, falling to 13.2% by 2015. Not only is this ratio well below the industry average (Exhibit 4), it is also below the level of spending necessary to maintain the present value of ‘core-PFE’s’ current sales (Exhibit 5). This is true even if we make the assumption that PFE’s R&D productivity going forward is on par with the broader industry’s and that productivity stops declining, neither of which have been true in the past (Exhibit 6)

We conclude that simply removing the emerging products does not make ‘core-PFE’ a growth company, even temporarily. ‘Core-PFE can only produce growth if R&D productivity improves and R&D spending levels are increased to a level sufficient to generate growth. Breaking the company into ‘core’ and ‘established’ siblings achieves neither of these goals

We created pro-forma models of ‘total-PFE’, ‘core-PFE’, and ‘established-PFE’ in order to assess how the core and established siblings’ equities might compare to current market norms. We use consensus projections for most elements of the ‘total-PFE’ income statement through 2015; after 2015 we project ‘core-PFE’ revenues as a function of past R&D spending, i.e. we simply make long-term revenues a function of historically observed R&D sales yields. For established-PFE, we assume no R&D spending, and so assume that these products are in an ongoing gradual decline – specifically, we make the simple assumption that post-2015 sales erode at a rate equivalent to sales erosion between 2011 and 2015e

An immediate complication of dividing PFE into core and established siblings is the overlap of established products with the emerging markets (Exhibit 7). If the emerging markets portion of established product sales goes into established-PFE – and it’s operationally difficult to imagine how they do not – then core-PFE’s near- to mid-term growth prospects[3] suffer from reduced exposure to the emerging markets

We assume all R&D spending goes to core-PFE, and model other costs using branded pharmaceuticals companies with younger product portfolios as comparables for core-PFE, and a blend of specialty and generic companies for established-PFE. Detailed results are provided as Appendix I

Because established-PFE consists of older products, does no R&D, and so is almost certainly a business in steady decline, we believe its equity can be attractive only if it pays out practically all of its earnings as dividends – so we assume a payout ratio for established-PFE of at least 0.90. For core-PFE, because guidance calls for R&D spending below the sustainability rate, we also expect a pattern of declining sales in the longer-term, though the rate of decline is slower than for established-PFE. The average payout ratio for large-cap pharmaceutical companies other than PFE is 43 percent (Exhibit 8). However these companies’ R&D spends are higher than core-PFE’s (Exhibit 4, again), which suggests their opportunities for longer-term growth are greater, which in turn suggests core-PFE’s (at least longer-term) payout ratio should be above the prevailing large-cap pharmaceuticals average

PFE’s current payout ratio is roughly 36%, or 0.80 on an estimated 2011 EPS base of $2.24. If established-PFE were separated from PFE for all of 2011, we would expect a $0.49 dividend at a 90% payout ratio, i.e. established-PFE would pay much more than half of the current dividend. If we assume a 50% payout ratio for core-PFE, we would expect a $0.85 core-PFE dividend. Obviously this produces a $1.34 estimated post-split 2011 dividend, as opposed to the current $0.80 estimated dividend for ‘total-PFE’. Alternatively, if we fix the established-PFE payout ratio at 90% but vary the core-PFE ratio to preserve the current $0.80 expected 2011 payout, then core-PFE can only pay $0.31, which is a payout ratio of only 18%, well below the large cap industry average, and well below the lowest payout ratio in the industry. We conclude that because the established-PFE payout ratio must be quite high in order to make this equity viable, a competitive dividend on core-PFE can only be paid if the aggregate payout ratio of the two siblings approaches $1.34, or a payout ratio of roughly 60%

The preceding assumes all cash returned to future PFE shareholders is in the form of dividends, i.e. it ignores share purchases. If we consider current share repurchases in current payout ratios, a 60% payout is very much in-line with large-cap industry standards. Our point about the two siblings having to pay a higher aggregate dividend than the parent is not about whether they can, but about why they have to – because neither is a growth equity, thus both are highly dependent on yield

The division of a low-multiple / high yield total-PFE into two low-multiple / high yield siblings begs the question of whether anything has been gained by splitting the parent. We accept that core-PFE is theoretically better able to produce growth than total-PFE, simply because core-PFE has a smaller current-sales denominator to weigh down the growth effects of any new product launch. However it makes little sense to reduce the core-PFE sales base for this reason if R&D spending is held below the level necessary to maintain the present value of current sales, much less create growth. Reducing the core-PFE sales base is a helpful step on the path to growth, but in and of itself wildly insufficient. In fact, we’d go so far as to argue that unless R&D productivity and R&D spend levels are sufficient to create growth, a larger sales base is better than a small one, simply because the larger sales base tends to dampen the revenue and earnings volatility associated with patent losses – which at sub-threshold R&D spending levels are a more prominent feature of revenues than new product launches

 

  1. This assumes the strategy is to break the firm into two or more independent companies. We recognize that other approaches are possible, but nevertheless use the simplifying assumption that PFE is broken into two independent firms
  2. PFE’s recent decision to reduce R&D spending in total opens the door to efficiency gains from reducing the absolute scale of R&D operations. However, even if PFE breaks the company into two and puts all of the remaining R&D spending on a smaller sales base, we show that R&D spending relative to sales is insufficient to maintain the present value of future sales. And, we sense that PFE’s ongoing internal review of R&D is more project- (i.e. which projects to keep, close, or re-configure) than process-focused, and we’re wholly convinced that process inefficiencies are at the heart of R&D productivity problems
  3. Notice we’re careful to exclude long-term growth prospects – we believe emerging market growth is a function of wealthier consumers buying Western brands (both on- and off-patent) largely because these consumers cannot trust the drug distribution systems in their home countries. Once these consumers can trust the local pharmacist, they presumably no longer need to rely on brands as an assurance of quality, so we eventually see emerging market demand for Western brands shifting to demand for local generics. (Please see: “Big Pharma’s Tenuous Grip on the Emerging Markets”, Sector & Sovereign Research, July 8, 2010)
Print Friendly