Is the noise obscuring the signal?
October 19, 2020
Nate Silver, the political pundit wrote a bestselling book in the wake of the 2008 Global Financial Crisis called the Signal and the Noise, with the subtitle ‘Why so many predictions fail, and some don’t”. The book made two important – but sometimes seemingly contradictory – points; that we either don’t take enough account of new data coming in to adapt our modelling and predictions or alternatively that we make our models so overly complex that their ability to deliver insights is drowned out by the ‘noise’.
Interestingly part of the popularity of the book came from the (ex post) analysis of the GFC and part of it came from his calling almost every district right in the US 2012 Election. The fact that he failed to repeat the feat in 2016 or that his own (rather simplistic) stock market predictions have subsequently fallen into a classic value trap do not, in fact, undermine the basic premise. This essentially called for better use of data in making predictions and in particular the use of what are known as Bayesian techniques to not only update our models with new data points but also to apply probability of events happening rather than vague issues around possibility. This is not to review the book, nor to apply some of its insights on modelling and the predictive ability (or not) of Covid models, although we could, but rather to take his terminology and talk about what signals for investors may be being obscured by the noise in markets at the moment, including the Elections and Covid.
To start with currencies – which we frequently refer to as noise trades, since a lot of data points and accompanying hyperbole tends to surround what are actually relatively minor price movements. Sometimes however, we find that the noise is obscuring a potential longer term signal. We note for example that the rally/short covering in the trade weighted US$ has stalled and that, significantly from a technical viewpoint, the offshore Chinese currency the CNY has continued to strengthen, at the time of writing sitting at 6.6884, a fraction above the 2019 level of 6.6872. As discussed last week, we continue to believe that the Chinese government will allow this level to break, while amidst all the noise of the US Election we would also highlight a plausible theory that any new US administration may revisit the “our currency, your problem” devaluation policy of the US$. This is not a certainty, but it has a higher probability in our view than is currently acknowledged.
By contrast, what a year ago many predicted as a long term signal – a so called hard Brexit and the collapse of sterling – looks to be more like noise. Without Covid, the news on the UK/EU trade agreement – which last week was deemed to be moving to the ‘dreaded’ No Deal situation as the UK announced that it couldn’t see the EU making any attempt to agree – would have totally dominated the headlines and probably the ‘noise’ in the FX markets. It hasn’t and indeed, in contrast to the predictions, sterling has in fact been relatively strong. Partly perhaps because many of the more astute political and economic analysts (in our view) seem to be more confident than not that some form of a deal will actually be realised before year end, but also because the simple “No Deal = sell Sterling” prediction model wasn’t very accurate. In fact, since 2016, Sterling has moved in a relatively stable sideways band centred on $1.30, + or – 10 cents. Moreover, the assumed key rate, the Sterling Euro cross rate – which the ‘pundits’ confidently and repeatedly predict will go to parity on a “no deal disaster” actually appears to be heading in the other direction.
On Covid itself, which has taken over the role from Brexit of dominating short term noise, we suspect that even as many of the models appear not to be updating with new data (particularly on the improvement in treatments) as Nate Silver would suggests they should be, the new data actually being incorporated is opinion polling -ironically perhaps his specific area of expertise. Every time governments impose tighter restrictions, the polling appears to ask for more. The public seems to be more frightened of Covid, the less deadly it gets. Markets are pushing predictions of V shaped recessions to W shaped, U shaped or even L shaped. Periodically they get excited about the prospect of a vaccine, but generally come back more pessimistic than the last time.
However, to continue our framework, we believe that this could be similar to the Brexit environment, where predictions were to some extent engineered to produce responses (Project Fear) and also the fact that noise can create feedback loops. We used to see this all the time with Consumer Confidence Surveys; after months of telling people how ‘bad’ everything was going to be, (not coincidentally to support opposition candidates in Elections) the broadcasters then found themselves ‘shocked’ that the surveys appeared to confirm everyone agreed with them. Pundits then incorporated this as ‘new information’.
We continue to believe that US Politics and Covid, the two dominant narratives shaping short term uncertainties, are closely related and this also seems to be confirmed by a fascinating survey from Gallup highlighted over at Lockdown sceptics. To pick just one example, Democrat voting women are almost three times as worried about getting coronavirus as Republican voting women and four times as worried as Republican voting men.
Views on the Virus are highly Partisan
To some extent this could be explained by an urban bias – most big cities are Democrat – but more likely (it seems to us at least) is that this is a “Never Trump’ issue. Telling pollsters you are worried about Covid and going back to work as well as wearing a mask are anti -Trump declarations and to a lesser extent vice versa.
In many ways this is similar to the “I voted Remain because I am not a racist” meme that circulated in 2016, Remain voters denied any or all of their other motives by wrapping themselves in a cloak of virtue. In this Election, Covid has been the main line of attack by the Biden-Harris campaign and ‘fear of Covid’, mask wearing and social distancing have practically become an article of faith for Democrat voters.
The hopeful good news is that this suggests that post November the momentum to this narrative will dissipate – regardless of who wins. Indeed, it will be in the interests of any new administration to deliver a solution rather than simply highlight how bad the problem is. In markets, we become used to how the consensus ‘no brainer’ position and associated narrative can suddenly reverse 180 degrees. Here’s hoping that Covid lockdown policy can do something similar.
Meanwhile in Europe, the seemingly co-ordinated lockdowns continue to be driven by the optics of positive tests and associated ‘worst case forecasts’ rather than actual hospitalisations and deaths. No Bayesian adjustments here. Logic seems largely missing (although logical fallacies abound in arguments in favour of ever tighter restrictions) and the private sector continues to struggle against an oppressive public sector driven policy of (seemingly) zero Covid. In terms of medium term risks therefore this is driving the rich and well funded to further consolidate their economic power and position. However, here again we can perhaps be hopeful that, should the US become more ‘constructive’ on matters, that Europe would follow.
Which brings us to our other potential key signal. Post Covid, we may find when we look, that the structure of capital markets has structurally changed.
As noted a few weeks ago, behind all the political noise, the Hedge Fund Tycoons and Private Equity Titans are busy resetting markets to suit their own agendas. Seemingly boring companies with good cash flows are being taken private, while narrative growth stocks with poor cash flows are being pushed into public markets where a powerful financial PR machine lures in Retail investors. As of today, recent ‘overpriced’ growth IPO stories like THG and Snowflake have not crashed, but neither have they gone anywhere. They may yet end up like Ocado (up 7x in the last 5 years) or alternatively perhaps like Aston Martin (down 90%) but the real point – and the important thing for their ‘promoters’ – is that the insiders have got their money out – and if their ‘private equity carry’ is structured properly it will be tax free to boot.
Meanwhile, the boring cashflows embedded in value stocks like Telcos or resource stocks can be picked up at a very comfortable spread over the incredibly low cost of financing available to Hedge Funds and by using deeply discounted rights issues they can also restructure debt and resume dividend payouts and thus secure themselves powerful income streams into the future in a world where stories are plentiful but cash is scarce. You take the Momentum, I’ll take the Value. We have seen this before and the result is the emergence of powerful conglomerate businesses, very much helped by ‘crisis response measures’ that have essentially allowed powerful companies to borrow money for less than nothing. Make no mistake, a new Oligarchy is being built.
Momentum and ‘Growth’ stories are being pushed into Public Markets, while Value is being taken private.
The problem with this picture for ‘outsiders’ of course is that the answer isn’t really to buy ‘value’ in the public market since it will only get realised when it is private. Existing institutional investors will be reluctant to participate in deep discount rights issues and will likely be focussed more on their short term relative performance than long term returns and will thus sell at only a small premium.
If we combine this with the likelihood, as discussed last week, of an ESG takeover of the majority of the private capital invested in public markets then we risk finding Western Capital markets in a very different place by the time the noise of another US Election comes around.