< Back to Research

Less unknowns on the pandemic, more unknowns on everything else

Raphael Fiorentino
6th April 2020 - 7 min read

Less unknowns on the pandemic…

Early March, we reported on the staggeringly wide range of possible outcomes from the COVID-19 outbreak due to the uncertainty level surrounding the contagiousness (R0), lethality (CFR) and severity (SCR) of the virus. It turns out that it is:

  • extremely contagious, as most (likely over 80%) infected people are asymptomatic but contagious for an unusually long period (two weeks),
  • but rarely fatal for those without serious chronic conditions (e.g. Italy’s NHI report that 99.2% of covid-19 deaths in Italy affected persons who had one or more serious chronic illnesses like diabetes, hypertension, cardiovascular diseases, cancers)
  • liable to suddenly turn serious (pneumonia-like) for a material (ca. 1 in 8) portion of those who do manifest clear symptoms, and even critical (severe respiratory distress, cardiac arrest, organ failures) for 35% of those who end up admitted in hospital, something that can rapidly overwhelm a country’s healthcare system

It is therefore an entirely manageable outbreak, without the need for total lockdowns, but only if adequate protection (distancing, equipment), diagnostics, tracing and isolation (of at-risk and contagious people), as well as hospital care capacity (especially for critical care) are readily available.

South Korea got most of it right and as a result detected 10 times less infected cases (despite testing more) and incurred daily fatality rates 200 times lower (despite an earlier outbreak) than Europe and the US, all without imposing a country-wide lockdown (see graph below).

Reported covid-19 cases and deaths (from data by CSSE at Johns Hopkins University)

image1

Despite spending more than twice as much as South Korea per capita on public health, Europe and North America got it mostly wrong and the resulting strain on their hospitals multiplies the number of fatalities (see illustration below for France).

France example: 30% of confirmed cases admitted to hospitals, 35% of them requiring intensive care for 10-20 days (from data by Sante Publique France)

image2

...More unknowns on everything else

Unprecedented levels of central banks and governments’ interventions have managed to rescue markets as Big Short Vol trades were spiralling down. The ACWI equity index barely dropped 25% and at -6%, our indicator of global growth expectations (iGDP, derived from the pricing of global financial assets) remains well above the levels reached during 2008 (see graph below).

Market-implied global growth in 2020 somewhat disconnected from the real economy whose condition looks much worse than in 2008

image3

This contrasts with the fact that we have little visibility on the timing of lockdown removals and of peak disruption in global supply chains, on the timeline of supply ramp up, and on the extent of demand contraction (see graph below for illustration of upcoming demand shock).

6.5% of the Active Population in the US have become unemployed in the last two weeks (Federal Reserve Bank of St Louis)

image4

Price discovery remains broken, with infinite central bank puts on the one hand, and on the other hand physical premia on some real assets going vertical (e.g. gold supply shortage vs. plentiful demand) while other prices have no support whatsoever (e.g. plentiful oil supply vs. shortage of holding tanks and soon tankers).

Finally, in addition to the price and supply-demand uncertainties that infinity-QE and total lockdowns create, we can expect several political blowbacks to shape the performance of equity markets going forward, from resentful Italians against useless European institutions (note that there is no Article 50 equivalent for Eurozone members), to destitute Americans against their President, to “populist” uprisings against “global elites”, etc.

 

Stock Selection for Today

With the global macro picture still extremely unsupportive of equities, and market breadth (% of stocks outperforming their regional benchmark) under 30% this past month for all but the most defensive sectors (staples, pharma, telco and utilities), stock selection is both extremely dependant on stock-specifics (e.g. firms that mapped their supply chain networks and have sourcing optionality are at a distinct advantage) and very much tilted toward high recession-resilience (stronger fundamentals and lower GDP sensitivity), as reflected for instance in Butterwire’s latest Stocks’ Heat Maps (see extract below for Large Cap EU Chemicals, where defensive industrial gases and consumer chemicals stocks like Linde and Novozymes stand out on the long side vs. highly pro-cyclical conglomerates like BASF on the short side).

image5

 

Stock Selection for Tomorrow

Butterwire’s Explorer enables to rapidly reflect one’s own interpretation and projection of the above, irrespective of the engine’s current “defensive bias”, by creating a shortlist of “eligible” stocks from which to automatically extract a portfolio. First, we use the top bar of the Explorer page to exclude stocks that we do not feel comfortable investing in at this stage, for instance stocks highly exposed to Italy or from the consumer discretionary sector or with poor share liquidity. Second, we apply filters to remove characteristics that we believe stand little chance to outperform, for instance stocks with high financial and cyclical gearing, or stocks with poor trailing idiosyncratic returns.

Listed below are the criteria and filters that reflect one such interpretation:

  • Applying the following criteria (top bar of the Explorer) extracts 1,273 stocks (from 5,973):

    1. Countries: exclude firms with Italy and Spain listed as country of risk, as well as all Emerging Countries except China
    2. Industries: exclude all Financials and Real Estate (weak fundamentals) as well as Consumer Discretionary, Semiconductors and Tech hardware firms (depressed demand, lean complex supply chains)
    3. Flags and Alerts: exclude red flags as well as Exit? and Check Thesis! alerts
    4. Others: exclude stocks with low (fair or poor) tradability, exclude all Potential Shorts

    image6

  • Further applying the following filters (sliders) reduce the shortlist to 125 stocks:

    1. Fundamentals: exclude high gearing (Net Debt / EV > 60%, market beta > 1.4), high volatility (> 45%), high controversy (> 9.5), low quality (ROC < 10%), low yielding (dividend < 1.5%) firms
    2. Idiosyncratic returns: exclude stocks whose recent share price performance puts them in the bottom 25% (Performance surprise index < -2)
    3. Score outliers: exclude all Fitness and Value scores below 2.5, Momentum scores below 1 (under Base Score), all Pro-Cyclical scores above 8.5 and recession-resilience scores below 1.5 (under Global Macro sensitivities), all Brokers’ View score < 1

image7

Pressing the “Extract” button at the bottom of the page enables to generate portfolios from the resulting shortlist of 125 stocks.

image8

The portfolio selected consists of 27 holdings.

image9

By construction, the holdings are selected to minimise risk factor exposures. As illustrated below, the portfolio has solid style characteristics (e.g. high fitness and recession-resilience scores, good diversity of value/momentum scores), both volatility and market betas are relatively low, and tracking error looks adequate in the current market context.

image10

It is however significantly overweight industrials and metals & mining (44% weight vs. 11% in the benchmark) and underweight US (36% vs. 56% in the benchmark). It turns out that the (unintended) bias introduced in the shortlisting was simply too large to mitigate (64 stocks out of 125 from the industrial and metals & mining sectors, and only 41 from the US). We can either confirm the portfolio if such biases are agreeable (e.g. US most overvalued and to be most hit by Covid-19, Industrial/Mining stocks are most geared to restocking/reflation), or revise it, either manually (Edit function in Portfolios), or by using the Mitigate function (in Portfolios), or by selecting a lower number of holdings, or by adjusting the initial shortlist (in Explorer).

While the above illustrates portfolio construction from cash, the same logic may be applied to portfolio construction from holdings. Our team stands at your disposal (directly via the app's live chat or via zoom video call) to support any of your portfolio-specific request.