The Financial & Risk business of Thomson Reuters is now Refinitiv

All names and marks owned by Thomson Reuters, including "Thomson", "Reuters" and the Kinesis logo are used under license from Thomson Reuters and its affiliated companies.

February 2, 2017

(Un)common Knowledge

by Amareos.

Imagine, if you will, the following scenario:

Company A’s stock price is the best performer within its country’s stock market year-to-date (admittedly 2017 has only just started, but still).

In fact, Company’s A’s stock price is not just the best performer; it’s the best by a huge margin having risen an eyewatering 930% in three weeks.

(Starting to get interesting now isn’t it! Nothing attracts investors’ attention like price momentum; it is analogous to blood in shark-infested waters.)

These not being the heady days of the dotcom bubble when a company’s share price surged higher on the merest whiff that it was doing something related to the global interwebby thingy (a phrase we heard more than once in the late 1990s – how far we have come in such a relatively short space of time!) and one might begin to suspect that Company A is either:

  • a takeover target
  • has just solved cold nuclear fusion or announced some other outstanding technological break-through
  • is a mining company that has just struck gold (literally or figuratively)

Sensible speculations though they may be, now imagine that none of the above are true. Furthermore, imagine that the management of Company A, mystified as to the recent surge in its share price, issues a public statement explaining that not only has nothing fundamentally changed in relation to its business model, nothing is expected to change. What’s more, for good measure, it states that the equity of the company, which as a result of the stock price jump has a valuation of USD 7mn, is roughly zero.

Ridiculous and as preposterous as the above sounds, this scenario just happened to a penny stock Danish property company called Victoria Properties A/S, a company that we had never heard of until this week – something we suspect is true of nearly every other investor on the planet[1].

In light of our ignorance, and given we have no crowd-sourced sentiment data relating to such a small company, we have nothing to say specifically with regard to Victoria Properties A/S. Our reason for mentioning this event is that it is yet another (albeit it at the bizarre end of the spectrum) example of the wonderful irrationality of financial markets that once more highlights the dangers of an over-reliance on valuation to drive investment decisions.

As the Victoria Properties episode vividly illustrates the numbers flashing on our screens every day may well accurately reflect what investors are prepared to pay for an asset at a given point in time, but there is no reason to suppose that this has any relation to its fundamental worth (or value).

This distinction between price and value is hardly a new observation. Indeed, Oscar Wilde in his 1892 play Lady Windermere’s Fan had one of his characters say in response to the following question:

What is a cynic? A man who knows the price of everything, but the value of nothing”.

The above quote, or some variant of it, is relatively well-known, but what is less well known is the next line in Wilde’s play. To wit,

“And a sentimentalist, my dear Darlington, is a man who sees absurd value in everything and doesn’t know the market price of any single thing”.

Thinking of financial markets as being comprised of continually shifting numbers of cynics, with their tendency to undervalue assets, and sentimentalists, with their tendency to overvalue assets, is, in our view, a superior model to the economists’ efficient market hypothesis which assumes that the market price constitutes the “best guess” as to the discounted present value of future cash flows[2].

The reason why the efficient markets hypothesis was given any weight at all is because it is extremely difficult to estimate accurately an asset’s fundamental value – there are simply too many moving parts and approaches. However, as we noted in an earlier Market Insight, recent evidence from experimental economics[3] indicates that even when the fundamental value of an asset is easily calculable – and hence knowable to market participants ex ante – the price the asset trades at divergences from its fundamental value for significant periods of time.

This result, which is serious blow to the efficient market hypothesis[4], should come as no surprise to anyone involved in building fundamental valuation models because these models must necessarily fail in order to work. This may sound like an oxymoron but it isn’t.

What we mean by this is that market prices have to periodically diverge from their fundamental value – whatever it may be estimated to be and irrespective of the method – as much as they converge in order to generate trading signals[5]. An absence of divergence, by contrast, would imply that the market price equates to the fundamental value at all times, undermining the usefulness of the valuation approach in its entirety.

We have outlined the Amareos approach to monitoring the changing composition between cynics and sentimentalists in previous Market Insights using crowd-sourced sentiment indicators; information that was not available at the time the efficient market hypothesis was first proposed. How these crowd dynamics unfold over time is something we have been pondering at length over the past several months, particularly in light of the move seen following Trumps’ victory in the November Presidential election. As we noted at the time[6],

“Swift is a gross understatement for describing the speed by which markets reversed direction following the surprise Trump victory last Wednesday. When it became clear in the early hours that Clinton, contrary to pollsters’ predictions, was unlikely to become the first female US President, Wall Street tanked, the USD sold off and Treasuries rallied hard.

Yet, barely six hours later, the materialization of the “greatest political risk event of the year” appeared to transform into one of the most effective reflationary episodes.”.

It was the speed of the change that struck us most forcefully. Heading into the presidential election crowd-sourced sentiments towards US growth and inflation were extremely weak, a perspective fully in accordance with the longstanding view of the experts; Trump would be an economic disaster. Everyone knew this. Just as every knew that Brexit would be an economic disaster for the UK.

It was the fact “that every knew this” that lead us to explore the concept of common knowledge, the best definition of which we came across in a 2013 blogpost by Ben Hunt[7] . According to Ben, common knowledge is,

“information that everyone believes is shared by everyone else”.

The crucial element is the belief that everyone else believes the same information. It is when these generally assumed perceptions are jolted that things get interesting. The US Presidential election is a classic case in point. In the following exhibit we plot the S&P500 index over the election period compared with crowd-sourced sentiment.

Exhibit 1. Crowd-sourced US Equity Sentiment vs. S&P500



As just mentioned, heading into the election the crowd was pessimistic, and sentiment towards US equities was weak and deteriorating. It was Wall Street’s unexpected bounce in the day or two after, which challenged the common knowledge that Trump would be a market disaster that prompted a radical rethink about the outlook for US equities.

What was the catalyst that caused the market to bounce in such an abrupt manner? It is probably hard to prove definitely, but we tend to agree with Bridgewater CEO Ray Dalio[8] when he said,

“That [the] shift was due to the changing complexion of market participants—those who drove the markets after his election were largely those who kept their powder dry until they saw the outcome and chose to process (and bet on) the policies themselves.” 

Although Dalio mentions no names, one rather obvious candidate is the US billionaire, Carl Icahn. Not only did he use the immediate slump in US equities to buy the market he publicly went on record stating his support for Trump and his economic policies; a direct challenge to the perceived common knowledge.

Obviously there are several strategies that one could employ to seek to exploit such situations including keeping a close eye on the narratives coming out from, as Ben calls them in his blog “missionaries” (influencers in more modern parlance). However, often times the catalyst may not be comments from a leading financial market influencer, but rather an unexpected piece of economic data. A great example of this was the strong UK retail sales number last August that confirmed the post Brexit economic environment was not as bad as everyone feared. By challenging the common knowledge of impending economic doom it changed the widely shared bearish GBP outlook leading to a tradeable short squeeze[9].

Similarly one could, just like everyone else can, look at the price action and hope that any countertrend move is not just noise, which unfortunately proves to be the case with a P&L alarming frequency[10]. Crowd-sourced sentiment data offers another alternative, yet complimentary, method.

For any idea to be common knowledge at one point in time it must be widely-held. Such a high degree of polarity in opinion shows up as a strong sentiment skew[11]. Combined with a dramatic reversal in sentiment momentum, confirming that the crowd is in the process of changing its opinion, this mix has the all the hallmarks associated with common knowledge being challenged (“crowd fails” in our lexicon).

To illustrate this effect, we examined all the occasions where crowd-sourced S&P500 sentiment extremes witnessed a sharp reversal over the past decade and compared it with the subsequent move in the price index.

We set two criteria for defining these common knowledge challenges or crowd fails. First, crowd sentiment had to be one standard deviation above or below the long-run average. Second, the change in sentiment over the subsequent week had to be a two standard deviation move in the opposite direction[12]. In situations where crowd sentiment was historically elevated but wanes we generate a sell signal conversely where sentiment was historically low but rises, we generate a buy signal. The underlying crowd-sourced US equity sentiment indicator, the signals and the S&P500 price index are all shown in the lower exhibit.

Exhibit 2. US Equity Sentiment, “Crowd Fail” Signals & Price – S&P500



We then calculated the one-month ahead forward equity price return following these signals as a way to monitor performance [13]. For good measure we repeated the same exercise for the Nikkei and the Dax[14]. The only difference being that in recognition of the high degree of correlation between global bourses we also incorporated the US “crowd fail” signal. The full results are shown in the exhibit below:

Exhibit 3. Crowd Fails vs. One-month Forward Equity Returns


The returns have been combined with the corresponding trading signal such that a positive return in the short column corresponds to the market declining. The number of instances of crowd fail over the past 10 years using the aforementioned criteria were 9 for the US and Germany (we found the US signal alone to be superior) and 15 for Japan.


The results shown in the above exhibit confirm that when strong polarity of opinion is being reversed – common knowledge is being challenged – equity markets typically move in the direction of the sentiment reversal over the following month. This makes intuitive sense as one would not expect everyone to change their mind instantaneously and it is this delayed response that is important.

That the short signals generate positive returns for all three equity markets is encouraging as it implies that the returns from this simple sentiment rule are not just picking up the typical upward drift of equity prices as a result of generally positive nominal GDP growth trajectories. Less positively, due to how crowd-sourced sentiments have evolved over recent years, we have not witnessed many episodes of reversing crowd bearishness (Japan being a small exception), such that based on our criteria the bulk of the long signals occurred in one period during the Great Recession.

The results are tentative and need to be extended beyond these three developed market equity indices, but they are encouraging. It strongly suggests that when the crowd, having had a fairly definitive view as to how the future would unfold (common knowledge), starts to change its mind, the effect upon asset markets is sustained. This not only generates exploitable opportunities for investors but puts yet another nail in the coffin of the efficient markets hypothesis.

Sentiment Analytics are based on MarketPsych indices

[1] See:

[2] Note: It is important to add the “best guess” qualifier to the above definition. As Burton Malkiel, author of the influential book A Randon Walk Down Wall Street pointed out last year the notion that the price exactly matches the discounted value means that “the price is always right” is wrong. Rather the theory implies “the [market] price is always wrong”, just that investors do not know in which direction it is wrong i.e. too low or too high. Unfortunately, this clarification doesn’t help investors very much.

[3] The paper also confirms that the emotional state of investors influences the degree of divergence; the stronger the emotion the great the divergence – see: Andrade, Odean and Lin (2015) “Bubbling With Excitement: An Experiment”.

[4] Oscar Wilde 2, Orthodox Economics 0?

[5] More formally, the residuals of valuation models sum to zero.

[6] Something we noted at the time – see:

[7] See: For the record, prior to writing this Market Insight we had not come across Ben or his work on finance and game theory.

[8] See:

[9] We discussed this in an earlier Market Insight – see:

[10] Unfortunately, some of the biggest one-day up moves recorded in stock prices occur during bear markets. It is never easy is it?

[11] As we outlined in a previous Market Insight, the high degree of polarity of opinion associated with “common knowledge” undermines the predictive power of the many over the few, potentially generating exploitable alpha opportunities; something we labelled “crowd fail” – see:

[12] That is standard deviations of weekly sentiment changes not the level of sentiment. Also because the sentiment data is published on a calendar day basis we used a seven day period.

[13] We used 30 day ahead price returns to be precise. We also incorporated a two-day publication lag for the crowd-sourced sentiment data into our calculations.

[14] Charts available upon request.

Article Topics
Article Keywords

Get In Touch


We have updated our Privacy Statement. Before you continue, please read our new Privacy Statement and familiarize yourself with the terms.×