The Pro-Cyclical US Healthcare Thesis – Impact of ROW Economic Risks

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

Richard Evans / Scott Hinds / Ryan Baum

203.901.1631 /.1632 / .1627

richard@ / hinds@ / baum@sector-sovereign.com

twitter.jpg @SecSovHealth

February 6, 2012

The Pro-Cyclical US Healthcare Thesis – Impact of ROW Economic Risks

  • US healthcare demand growth is slow for cyclical reasons; this implies an acceleration of US healthcare demand as the broader economy improves; we recommend hospitals (e.g. HCA, UHS, THC), HMOs (e.g. UNH, WLP), and non-Rx consumables (e.g. BAX, BDX, COV) as pro-cyclical bets
  • Hospitals and HMOs obviously are US-focused; however non-Rx consumables generally are multinationals with exposure to export market risks, forcing the question of whether ex-US (especially EU) economic risks outweigh pro-cyclical upside in the US
  • We show that: 1) ROW healthcare demand is less elastic than in the US; 2) that the components of ROW demand (volume, price, mix) appear to maintain their relative proportions in upturns and downturns (in contrast, US healthcare demand elasticity is almost entirely a matter of falling per-capita ‘unit’ demand); and, 3) that changes in ROW prices (during downturns) fall disproportionately on high-tech medical imports
  • Thus non-Rx consumables companies face less demand cyclicality from ROW than US sales (ROW is generally less than 50% of sales; and, healthcare demand elasticity is considerably lower ROW than in the US); and, non-Rx consumables face less risk from an ROW economic downturn than their higher-tech multinational healthcare peers (who are disproportionately subject to falling export prices)
  • Despite recent outperformance we see further upside to hospitals, who remain well below their pre-2008 PE-forwards (relative to cap wtd healthcare); this implies hospitals ultimately could see further outperformance even if ACA were overturned
  • We suspect HMOs will underperform as the market prices in increased utilization, though we’re convinced HMOs in fact gain earnings power as the economy improves, and so view weakness as an opportunity to buy. We recommend holding both hospitals and HMOs across the Supreme Court decision on ACA, as the market seems to believe that what (in the context of ACA decisions) is good for HMOs is bad for hospitals, and vice versa

Real growth in US healthcare demand is slow for cyclical reasons[1]; this implies improving healthcare demand growth as the broader US economy improves. Calling the broader economic cycle is hardly our specialty; however because valuations for the more volume-sensitive healthcare names imply an unrealistic continuation of 2010’s 50+ year low in per-capita US healthcare demand through at least 2013, we believe tilting healthcare portfolios toward more volume sensitive names is appropriate

US hospitals are obvious long candidates on this thesis; we believe these stocks have been made all the more attractive by virtue of unrealistic concerns regarding hospitals’ aggregate pricing power[2]. The obvious risk to a long position in hospitals is their volatility around the Supreme Court’s decisions affecting the Affordable Care Act (ACA); these decisions are expected in July of this year, and could drive hospitals’ relative performance sharply in either direction. Being unable to meaningfully handicap whether key provisions of ACA will or will not be upheld[3], we have obvious motive to diversify the pro-cyclical hospitals position

Both HMOs and manufacturers / distributors of non-Rx consumables also are attractive pro-cyclical options. Like hospitals, HMOs offer the advantage of being largely or even exclusively US-focused, and are therefore in geographic terms a pure-play on our US demand thesis. This said, we sense a disconnect between our view of HMOs in an economic up-cycle and the prevailing consensus. To our minds HMOs gain earnings power in an up-cycle: rising per-capita utilization levels are offset by the tendency of marginal enrollees (who grow by virtue of rising employment) to be healthier (i.e. lower claims but identical premiums) than those presently employed and insured. HMOs also benefit from simple operating leverage as membership grows; this leverage is compounded by the fact that HMOs earn a relatively stable gross margin percentage (made more stable by ACA’s medical loss ratio (MLR) ‘floors’) on medical costs, which inflate considerably faster than HMOs’ underlying operating costs. Recognizing the apparent consensus regarding HMOs’ leverage takes the opposite view, we see the potential for HMOs to sell off in an up-cycle, though we obviously view such an event as an improved entry point. Notably, HMOs also are affected by the outcome of the Supreme Court decisions on ACA; and, it’s reasonably clear the market believes that what is good for HMOs (in terms of ACA outcomes) is bad for hospitals, and vice versa. This argues for holding both hospitals and HMOs long across the ACA decision (expected by July 2012), and against holding either sub-sector long without the other

Manufacturers and distributors of non-Rx consumables are the third sub-group we recommend on the basis of our US demand thesis. With the exception of the more US-focused distributors (e.g. OMI) these names tend to operate globally, diluting leverage to our US demand thesis and raising the question of these names’ exposures to ex-US markets. 2012 EU consensus GDP estimates call for significant deceleration of real growth v. 2011; this level of deceleration is far sharper than is forecast for the US, and estimates imply a greater risk of downside in the EU than in the US. Plainly the multinational non-Rx consumables manufacturers are more exposed to these risks than US-focused hospitals or HMOs, but the question remains whether these names are more or less at risk than other US-listed multinational healthcare names. The Appendix lists relevant sub-sectors and names according to US revenues as a percent of total

As compared to the US, ‘ROW’ healthcare demand cyclicality is distinct in three critical ways: 1) ROW healthcare demand is less cyclical; 2) the components of ROW healthcare demand (volume, price, mix) appear to maintain their relative proportions across economic cycles, whereas US healthcare demand cyclicality appears to be largely driven by a change in per-capita volume; and, 3) in down-cycles, changes in the price per episode of ROW healthcare appear to be disproportionately driven by price cuts on high-technology medical imports

Beginning with the relationship between changes in real per-capita GDP and changes in real per-capita health consumption, ‘ROW’ healthcare demand is substantially less reactive to the economic cycle. For example, Exhibit 1 shows the coefficients on change for per-capita healthcare demand and per-capita GDP in the US and ROW (and the largest EU countries). The ROW coefficient of 0.98 essentially argues that ROW healthcare demand and ROW GDP generally move together; contrast this with the US coefficient of 1.1 which shows that changes in US healthcare demand are exaggerated relative to changes in US GDP. The effects of a US GDP down-cycle are readily apparent: in 2010 the cyclical component of US healthcare demand[4] (per-capita ‘elasticity\mix’) hit a 50+ year low. Across 2009 – 2011e this measure fell by (on average) 40bp annually, as compared to average annual growth of just more than 2 percent since 1960 (Exhibit 2). US per-capita elasticity / mix is highly cyclical (Exhibit 3), and wholly accounts for the recent trough in real US healthcare demand

Not only is ROW healthcare demand less cyclical, the components of ROW demand are more stable across the ups and downs of economic cycles. In the US per-capita ‘elasticity / mix’ is the primary driver of cyclical changes in per-capita real healthcare demand; and while this measure includes both units (i.e. ‘elasticity’) as well as mix, the evidence strongly suggests that a fall in US per-capita unit demand is far more important than any change in mix. Crudely, it seems that the marginal US patient simply goes on strike, shrinking the per-capita volume component of US healthcare demand, while mix and real medical pricing remain relatively strong. In contrast, outside of the US per-capita volume (i.e. per-capita ‘units’ of care) and price/mix generally move in unison across the economic cycle

We use per-capita doctor consults as a proxy[5] for ROW per-capita unit demand; in combination with robust data on total real per-capita spending, this allows us to distinguish volume and price/mix changes in real demand. Across multiple economic cycles, per-capita volume accounts for on average about 40% of real ROW per-capita healthcare demand growth. If we separate time periods into those characterized by either accelerating or decelerating per-capita GDP, we see that real demand changes with the economic cycle, but that the components of demand generally do not – whether ROW GDP is decelerating or accelerating, volume and price/mix tend to maintain relatively constant shares of per-capita real healthcare demand growth (Exhibit 4). Falling real per-capita GDP may be an exception to this rule. In periods characterized by an outright drop in per-capita GDP, per-capita volume appears to fall as a proportion of total demand growth (Exhibit 4, again); this may reflect modest queuing

Finally, changes in ROW price appear to fall disproportionately on innovators. We know that imports of consumable medical products by highly developed economies are reactive to the economic cycle, and that the level of reactivity is proportional to the technical sophistication of the specific imports. At one extreme of the technical sophistication scale, commodities (e.g. simple bandages) are priced by competition; at the other extreme high technology medical goods tend to be priced far more subjectively, with the outcome having much to do with the balance of power between seller and buyer. Outside of the US, medical import buyers are governments, who as single payors for their respective health systems have considerable buyer (and thus pricing) power. It follows that an ROW government under economic pressure is a more motivated negotiator than such a government in a stable or improving economy; correspondingly in a weakening economy higher-tech medical imports (whose prices are negotiated) should weaken more than commodity medical imports (whose prices are set by competition)

The evidence fits the thesis; since 1989 higher technology medical imports have weakened more in economic downturns than have less sophisticated medical imports. Exhibit 5 gives the coefficient on change for high-, mid-, and lo-tech medical imports (v. GDP) heading into highly developed economies. The green columns are the coefficients on change for periods of GDP contraction; higher values signify that the category of import changes more, and vice versa. High tech medical imports change the most (and in the same direction, i.e. down) during economic contraction; mid-tech changes ‘on par’ with GDP, and lo-tech changes much less than GDP. It follows that pharmaceuticals, biotech, and innovative (e.g. cardio / ortho) device manufacturers are more at risk of weakening export demand than manufacturers of mid- and lo-tech supplies – such as the non-Rx consumables names (e.g. BAX, BDX, COV) that we favor as a US pro-cyclical play. Notably, coefficients on change during economic expansion (Exhibit 5, light columns) are similar across the three categories. This strongly argues for price and against mix changes as an explanation for the reactivity of higher-tech imports to economic down-cycles[6]

Each of our preferred volume-sensitive sub-sectors have outperformed healthcare over the last several weeks, raising the question of whether our pro-cyclical thesis is priced in. Using long-term relative (to cap wtd healthcare) PE-forwards (next twelve months) as a benchmark, each of the sub-sectors is at least marginally below its 10+ year average (Exhibit 6); at the very least this implies valuations are still unremarkable despite recent performance. We believe it’s more important to read these PE-forwards in the context of how expectations and valuations have changed in the wake of the economic downturn (3-4Q/08) and passage of the ACA (1Q/10), than in a more traditional sense of whether valuations are at historically remarkable levels and thus subject to (at least expectations of) mean reversion. Hospital PE-forwards peaked on passage of ACA, based on perceived improvements in post-2014 patient volumes and uncompensated care costs; expectations then fell sharply as volumes fell to lower than expected levels in the midst of the economic downturn, and fell again as 2013 pricing expectations were lowered due to concerns over Medicare pricing provisions in the Budget Control Act. Hospitals’ immediate volume problems correct as the economy improves, and we have little concern over the sub-sector’s net pricing power; accordingly we would expect hospitals to regain their ‘pre-ACA’ relative (to cap wtd healthcare) PE-forward of roughly 0.95 (current PE-forward is roughly 0.75), even if the Supreme Court over-turned ACA

HMO PE-forwards fell sharply at the beginning of the economic down-cycle; valuations have recovered as the group has priced ahead of trend, and as regulatory concerns stemming from ACA provisions have moderated. As we argued earlier in this note, despite our belief that economic improvement raises HMOs’ earnings power more than current share prices imply, we accept that the sub-sector is likely to sell off as expectations for per-capita (healthcare) unit demand rise. At the risk of cutting the recommendation too finely, this argues in favor of buying hospitals now and HMOs later, but holding both as a 2012 pro-cyclical bet generally, and as a hedge to ACA-related risks (in July) specifically

Non-Rx consumable PE-forwards, since the beginning of the downturn, have essentially traced the arc of falling per-capita utilization. Believing per-capita utilization has every prospect of returning to pre-’08 levels, we correspondingly see the opportunity for the sub-sector to trade at a premium to its healthcare peers as volumes improve. The non-Rx consumables’ relative (to healthcare) earnings power during an upturn is aided by inherently greater operating leverage, less exposure to pricing risks in the event of further EU weakening, and by stable US pricing as we ‘wait’ for per-capita volumes to improve (as contrasted to Device Innovators, e.g. cardio and ortho, where pro-cyclical long positions are ‘taxed’ by falling real US prices)

  1. US Healthcare Demand Slow for Cyclical (i.e. Temporary) Reasons …”, January 12, 2012, Sector & Sovereign Research LLC
  2. Hospitals’ forward earnings expectations fell as news of the Budget Control Act became public; the Act calls for as much as a 2 percent Medicare rate reduction, which plainly should affect hospitals. However hospitals have a long history of using commercial price increases as an offset to Medicare rate cuts. And, because commercial plans need ‘brand name’ hospitals in their networks in order to participate in enrollment gains brought on by the roll-out of health insurance exchanges in 2014, hospitals appear to have more commercial pricing power than ever as 2014 approaches – and to the point, more than enough to offset the effect of Medicare rate cuts. For more please see: “Hospitals’ Stable to Improving Net Pricing Power”, January 24, 2012, Sector & Sovereign Research LLC
  3. ACA at the Supreme Court – What You Should Know”, October 24, 2011, Sector & Sovereign Research LLC
  4. Per-capita ‘elasticity / mix’ is our cyclical measure of US healthcare demand. We find this value by simply backing real medical pricing (which is volatile but not cyclical) and age effects out of per-capita demand
  5. This variable is far less affected by changes in both technology (per-capita prescriptions and implant rates are highly affected by new product flow, which is largely acyclical) and policy (as in the US, but to a lesser degree, ex-US policy has driven shorter lengths of hospital stays). Because these technology and policy changes begin at different time points and unfold at different rates in different economies, they meaningfully confuse the use of prescriptions, implant rates, and days of hospitalization to track ‘simple’ per-capita demand. Tracking hospital discharges instead of days of care controls for length of stay, but discharge rates are available for too few economies across too few economic cycles to be statistically useful
  6. Price cuts during a downturn almost certainly aren’t given back when GDP improves, and this is consistent with higher-tech imports’ strong coefficient on change during economic downturns, and weak coefficient during expansion. Restrictive policies resulting in falling per-capita mix would also cause strong coefficients during down-cycles; however any such restrictions presumably would lift as GDP improved, which would show as strong coefficients on the economic up-cycle; clearly this is not the pattern we see
Print Friendly