Paul and Brett's Alpha

Having committed to writing discursive and expansive factsheets, we do worry about iterance and inanity. Nobody wants to read the same old diatribe, no matter how many novel re-phrasings it may contain, and yet we are also mindful of the persistent and pervasive macro-led dynamics that have influenced the performance of the portfolio and our actions as investment managers.

There is no getting away from the challenges that we have faced over the last 15 months and we have highlighted the arbitrary nature of market moves and significant size factor effects (i.e. SMID healthcare underperforming its Large-Cap and Mega-Cap brethren) and generally irrational price movements.

Consequentially, we have been much more introspective in our actions. This has manifested itself in a lower level of portfolio variance during the calendar year; indeed, the 29 companies in the portfolio are unchanged since early May 2022 (weightings have moved of course, but no stocks have been added or exited). Rather than describe this in qualitative terms once more, we will instead describe the portfolio’s evolution across the year in quantitative terms.

Regular readers will be aware that our investment focus is on multi-year return potential. As a consequence, we have reacted to what we consider to be a temporal aberration in value determination to re-orient the portfolio toward those companies where we see the greatest opportunities based on tried and tested valuation assumptions that have worked over a long period of time.

We are all painfully aware how much the share prices of many of our holdings ‘evolved’ over 2022. However, the stock market is an imprecise mechanism for price discovery and the expression of value, especially in the shorter-term (if it wasn’t, there would be no need for portfolio managers or wealth advisers). The key question then is how much the fundamentals have evolved over this period of time and thus whether or not 2022 represents a fantastic opportunity or the emergence of a new normal. Let us consider this question in more detail...

The best way to determine if fundamentals are changing is via a like-for-like comparison of expectations for investee companies over time. 23 of the 29 stocks in the portfolio at the end of December 2022 were also owned at the end of December 2021. These represented 87.9% and 82.1% of the portfolio’s gross exposure at these respective time points. As regular readers will be aware, we carry detailed internal models for all investee companies and these are continuously updated to reflect news flow, corporate actions, results reporting, foreign exchange rates etc.

Previous versions of these models are archived and we have gone back and compared various outputs from our internal models as at the 2021 and 2022 calendar year end, focusing on our FY2023 and FY2024 revenue forecasts, our fair values and upside to year-end share prices and also the implied reverse discount rate (i.e. what discount rate does one need to apply to our base case forecasts to generate the year end share price as the fair value for the stocks). In each case, we have aggregated the results by the relative portfolio weightings at each time point to give a single point estimate for the portfolio.

Healthcare is a tightly regulated sector driven by demographic trends and characterised by long development timelines. Whilst it contains many complexities, it actually tends to evolve rather slowly and transparently. There may be the odd surprise here and there (a drug has unexpected side effects and gets pulled or a clinical trial fails despite all previous data being positive), but these tend to impact forecasts for one or two companies rather than the whole sector. It is unsurprising then that between the end of 2021 and 2022, our FY2023 and FY2024 revenue estimates barely changed (-2.7% and +0.2% respectively).

The aggregate upside to fair value versus the year end share price stood at 33.2% at the end of 2021 and 57.8% at the end of 2022. These two values align conceptually; if the output of the models did not change much (per the revenue forecasts), then neither will the overall fair value and thus the upside to fair value would grow as share prices fall (recall the Trust’s calendar year end NAV represented a negative total return over 2022 of -11.1% and a number of our holdings saw their share prices punished very severely over the course of the year).

What about the implied discount rate (reverse DCF) analysis? At the end of 2021, this stood at 11.4% and had fallen to 10.3% by the end of 2022. This decrease in the aggregated reverse discount rate is primarily due to the evolution of the portfolio toward stocks with higher terminal contributions (i.e. the portfolio has increased its skew toward companies that are not currently cashflow positive and reflects our active management approach of recycling profits from winners (typically Diversified Therapeutics, i.e. “big pharma” and Managed Care during 2022) into those companies with the most attractive risk/reward profiles (i.e. SMID healthcare which fell out of favour over these past 15 months).

It is worth noting that these fair values include probability adjustments to cashflows arising from any drugs or medical devices that have yet to receive a regulatory approval in the United States or Europe and thus often represent a much higher discount rate to the underlying forecasts than the valuation output from our base case scenario implies.

In summary, the portfolio has evolved, but there is no evidence that the fundamentals for the companies that we continue to invest in have deteriorated to any meaningful extent and certainly not in a manner that is commensurate with the share price movements that many of them have experienced.

Using broadly constant assumptions (again, discussed in more detail below), the upside to base case fair value from current share prices has increased by almost 75% (from 33.2% to 57.8%), so why wouldn’t we continue to own the same securities?

One might try to counter that continuing with the same valuation assumptions is erroneous in the current inflationary environment and we will consider that in due course. Given that we are also using the aforementioned probability discounts, the important point to note is that an implied discount rate of 10.3% is still materially above any commonly conceived equity security discount rate (as described in the following section), presuming of course that many of the widely accepted methods for ‘calculating’ these discount rates are not flawed in and of themselves (also considered in the following section).

The question then is: how should one be thinking about inculcating the risk of persistently higher inflation and contemporaneously higher interest rates into portfolio construction decisions?

If one were an MBA student, now would be a good time to waffle on about the efficient market hypothesis and how share prices are, in effect, the sum of all fears. What assumptions are discounted in the current scenario? MBA types also love a bottom-up theory to justify an approach, so will often apply an equity market risk premium over a risk-free rate.

That risk-free rate will typically be the sovereign bond yield of the country of domicile. Such rates are also presumed to reflect the geo-political risks of investing in these markets because the liquidity in the bond market is such that it is considered to be truly efficient. The evidence does support this contention – we can see in real-time that the bond market effectively discounts geopolitical risks.

Recent examples include the yield spike on UK Gilts after the Truss/Kwarteng delusional “mini-budget” when the 10-year gilt yield increased 44% in seven days and brought down that administration, or the yield on Ukrainian 10-year paper, which was stable around 7% through most of 2021 and now stands at ~33% because Russia is pummelling the country back to the stone age and the economy is tragically in ruins.

However, the choice of risk-free rate is not as easy as it once was. Most of the companies we invest in these days are truly multi-national. They borrow in whatever country and currency optimises their net financing costs and hedge currency risk, using debt interest to depress profits in the most egregiously taxed jurisdictions and increasing profits and inter-company dividend flows where tax is least problematic. In general, revenues (and thus geopolitical risk as regards impact on the company’s future prospects) are now global.

For us, it still makes sense to focus on US Government bonds for the risk-free rate since the US is still the single largest market for healthcare (and thus the largest source of operating cashflows for the portfolio companies), the centre for healthcare innovation and the predominant domicile and reporting currency for the portfolio.

The risk-free rate is the easy part, one should use the Government bond yield over a duration that reflects the expected holding period of the equity securities. In our case as expected long-term holders, we might choose the US 10-Year Treasury Bill (or even the 20-Year). In the current environment, the choice matters little, as their current yields are very close: 3.5% and 3.8% respectively.

The market risk premium is the amount by which the expected return on a market portfolio exceeds the risk-free rate. It represents the financial compensation (i.e. higher profit) investors expect for taking on the additional risk of investing in a market portfolio as opposed to a risk-free investment. Historically, this was calculated through backsolving, by looking at historical equity market dividend returns and comparing them to bond yields.

This is arguably no longer a valid approach; dividends are less popular than they used to be (especially in the US) and most of the positive return from the S&P 500 over the past two decades has come from companies that do not pay dividends (tech, biotech, etc.). Furthermore, the globalisation of the economy has broken down the association of the dividend income stream to the domicile of the company.

How then do we get any validation of a market risk premium assumption? We utilise the tried and tested consensus approach; the market must be discounting whatever figure is most accepted. This can be determined by surveys of academics and finance professionals and there are some robust and publicly available datasets one can scrutinise. Our preferred series is that from Pablo Fernandez at IESE Business School, which has been running since 2007.

The US Pepperdine Graziadio Business School survey series is also interesting (to us, anyway). This is much smaller in terms of respondents but more granular in terms of respondent type and also contains other interesting data points. It has also been running since 2004.

We include the most recent series in Figure 5 below (it is backward-looking so will always lag current risk-free rates if they are moving). The important thing to note from the Fernandez series is that perception of the market risk premium has not really changed very much over the past decade or so (10-year range is 5.3-5.7%).

Some readers may think this is incorrect (the world feels a much riskier place to do business these days, especially if you are exposed to autocratic regimes like China) but that does not really matter overmuch; the market will reflect the consensus view of the participants and this survey confirms that they will continue to behave as if the risks have not changed and thus will continue to believe this remains a perfectly reasonable expectation of future investment returns.

Source: P Fernandez et. al; IESE Business school, SSRN journal May 2022

Some Wall-Street analysts will erroneously backsolve equity discount rates using a WACC (weighted average cost of capital) derived from the Capital Asset Pricing Model (‘CAPM’), which is a project finance tool and not appropriate for valuing public shares as it will tend to increase equity valuation as the proportion of a company’s enterprise value accounted for by debt rises (we accept that it may have value for infrastructure and utilities companies if these are a pure play with regulatory protection on cashflows).

Any leverage buyout whizz will of course tell you that in their models, the risk premium rises with the debt level, as does the cost of debt (think about mortgage rates and loan to value parameters as a proxy). Rarely though do you see sell-side models take account of this reality. Leaving aside the comment about leverage impacts, one could argue that a discount rate of around 9% for the broad market feels about right given where inflation expectations are.

Should healthcare be higher or lower than the market on average? Given that total shareholder returns for the S&P 500 Healthcare Index meaningfully exceed those of the broad S&P500 Index on a 5, 10, 15 and 20 year view we would suggest the answer must be lower (that is what a backsolve would tell you). In addition, the growth of healthcare expenditure in the US bears no relation to growth of the economy; it is classically defensive and thus subject to less economic risk. The Beta of the S&P 500 Healthcare Index is also less than one when compared to the parent index.

Since inception, our approach to all of this has been to construct a proprietary matrix of discount rates that vary by the size, developmental stage and sub-sector (we characterise all companies into 16 separate sub-subsectors based on end customer type, rather than the widely used GICS system) and include additional risk premia for over exposure to certain markets where the regulatory or geo-political situation is more complex (MENA, China, India, etc.). These assumptions were based on extensive back-testing over various time periods. We also apply a range of normalised and terminal growth rates depending on similar variables.

As things stand today, the blended impact of this approach is that our fair values are generated using a weighted average discount rate of 8.0% (range 7.0-9.0%). The terminal value component (2030+) accounts for 47% of the fair value and the weighted average terminal growth rate is 1.8% (range 0.0-3.0%). As noted previously, these base case scenarios include probability weightings as well and generate aggregate upside to fair value of 57% compared to the share prices at the end of December 2022.

Some readers may argue that 8% is too low in the current inflationary environment. Perhaps we should use a higher value? This would be justified if one were to argue that interest rates were likely to stay at current levels in perpetuity. To quote the prolific Prince (the talented and now departed singer-songwriter, not the whingy one all over the papers currently) “forever is a mighty long time”. There are strong arguments in favour of an ageing global population depressing interest rates (cf. Japan) and thus we believe, as many economists do, that global growth will slow in the near-term and remain sluggish for the longer-term.

In the interests of academic fairness though, we have run a simplistic version of the exercise. A 100bp increase in the overall discount rate (i.e. from an average of 8.0% to 9.0% without changing any other assumptions, which surely would have to be the case if inflation were baked in), our fair value upside would decline by 14.4% (i.e. it would be 43% above the year end figure).

Whilst we do not have the time to do a detailed analysis of each company line by line to consider a structurally higher inflation scenario just for this factsheet, most of the companies that we invest in have market leading positions in their respective fields and consequential pricing power. Generally speaking, additional costs from labour, raw materials or running costs have been passed on in the form of higher prices. Thus, were we to undertake such a detailed analysis, it would probably result in higher revenue forecasts, broadly stable margins and thus higher net operating cashflows.

We also have much sympathy with the argument that, at the very least, a higher discount rate should be accompanied by a commensurate increase in the terminal growth rate (this trend was picked up in the most recent Pepperdine Graziadio Business School survey, where terminal growth rates had risen to an average of 3% due to elevated inflation assumptions).

One should probably adjust the normalised growth rate too, but again this gets too complex for this particular exercise. If we make the same 100bp adjustment (so the net terminal growth also moves from +1.8% to +2.8%), then the fair value upside falls from +57% to +52%.

As a final reminder, these figures represent upside to where we think fair value sits today, not at some indeterminate point in the future. Moreover, we expect the portfolio to accrete returns on an annualised double-digit rate so if nothing else happens and the share prices do not improve, that +57% would become at least +63% next year and so on and so on. This might be a growth portfolio, but it now feels very much like a value portfolio to us.

The founding investment proposition of the Bellevue Healthcare Trust can be summarised thus: the healthcare systems of the developed world are fundamentally broken. They are no longer fit for purpose and must be re-imagined and re-engineered to meet the needs of an ageing population burden with chronic (i.e. incurable) diseases.

The companies offering the tools, products and services that will enable this transition are not going to be provided by the large healthcare conglomerates and mega-cap winners of yesteryear, but rather by a new generation of focused, operationally geared companies that are probably off the radar for many generalist investors.

The drivers of adoption are the same demographic demand drivers as for all healthcare services, allied to a compelling additional driver of economic need. We have to bend the healthcare cost curve and improve productivity for frontline staff if we are to have a hope of keeping up with demand.

Let us recapitulate this narrative relative to current evidence. Readers must try not to view these solely through a UK/NHS prism. Whilst the current situation here is awful and the polemics rage about who is to blame and why this is happening, the same signs of stress and excess death are also evident in Germany, Spain, France and the United States, for instance, where healthcare funding (as both a percentage of GDP and on a per capita basis) has been higher over the past decade:

• Has the demand outlook for healthcare diminished in any apparent way in recent years? No.
• Are healthcare systems keeping up with demand and continuing to offer acceptable levels of service? No.
• Are staffing levels managing to scale up to meet this demand? No, quite the opposite in fact.
• Is healthcare expenditure rising faster than GDP in developed countries and still failing to bridge these gaps? Yes, and this is unaffordable, especially if we go into a recession.
• Has the regulatory environment changed in any meaningful way that represents a sustained increase in business risk for healthcare companies? No. At the margin, some funding mechanisms and/or regulatory approaches need work (e.g. home healthcare, social care, human factor testing for US medical products), but overall the situation is largely unchanged.
• Can we continue in this vein for any reasonable length of time? No, the systems will collapse or the electorate will push back on the funding costs.
• Thus, is there any reason to think that the core investment thesis of the Trust is broken in any way? The logical conclusion surely must be: No.

Even if you are more bearish than we are on long-term inflation, the Trust objectively provides operationally geared exposure to a portfolio of undervalued companies that will help to deliver an

urgent and necessary transformation of healthcare. At some point in time, the market will once again appreciate these virtues and re-rate these companies. We are not happy with the recent investment returns but we are happy to stay the course and, in the meantime, continue to buy shares in the Trust for our personal accounts, as we have been doing throughout 2022 and which has continued into 2023.

At the start of the factsheet, we referenced our comments from three years ago and repeated the observation that the world rarely changes quickly and the stock market can be faddish. No doubt some very lucky/clever people have made a lot of money out of Tesla and Beyond Meat by timing their trades well or by shorting them. That said, they went up very far for a long time before reality began to bite.

We are in an unusual position with innovative healthcare at the moment. The market chooses not to see the value at this time, but we know the end user demand is there and that the products work. We can also see and, unlike Tesla boosters, choose to see the competitive dynamics for what they are and factor them in. Three years from now, we fully expect to look back on this moment for what it is – a huge opportunity.

We always appreciate the opportunity to interact with our investors directly and you can submit questions regarding the Trust at any time via:

shareholder_questions@bellevuehealthcaretrust.com

As ever, we will endeavour to respond in a timely fashion and we thank you for your continued support during these volatile months.

Paul Major and Brett Darke