In Focus

In our ET Professional updates over the last few weeks, we have repeated one particular refrain.

The dominant COVID-19 narrative today is a “short and deep” economic impact.

There’s a corollary narrative here that we think is particularly important for investors, and it’s a narrative that we think is closely tied to real-world data that investors can track on their own.

The corollary COVID-19 narrative today is “pent-up demand” that drives the up-slope of a V-shaped recovery.

The “Short and Deep” Narrative

Here’s on overview of what we’re seeing in the Narrative Machine.

The narrative maps below represents all major outlet US equity markets coverage since March 1, with the bold-faced connections and nodes representing articles with language characterizing the economic effects as “temporary” or “short-term” or “V-shaped” in the first map, or language relating to “lasting” or “long-term” or “L-shaped” economic effects in the second map.

The way to read narrative maps is to look at the number and frequency of bold nodes, yes, but much more importantly you should look at the centrality and the connectivity of the bold nodes.

In the first map, focused on the language of short-term COVID-19 impact, you can see how this language is deeply connected to each of the most central clusters (general market commentary, fiscal and monetary policy, corporate earnings, etc.), and extends from those central clusters into nearly every other peripheral cluster on nearly every market sub-topic. The one exception would be any cluster that focuses on the energy sector. No one is talking about the oil & gas industry with short-term or V-shaped recovery language!

In the second map, however, focused on the language of long-term COVID-19 impact, the only topic with concentrated use of and connections driven by this language is unemployment (dark green cluster on right of graph). There is widespread pessimism about just how “short and deep” the labor market and employment effects will be, but that pessimism is isolated to only those topics.

Language about the short-term economic effects of COVID-19 is pervasive in the way we talk about markets today.

Language about the long-term economic effects of COVID-19 is not.

US Equity Market Coverage Referencing “Temporary / Short-Term” Effects

US Equity Market Coverage Referencing “Lasting / Long-Term” Effects

“Short and Deep” implies “Pent-up Demand”

When you dig into what’s actually being said in these articles and transcripts about the short-term economic impact of COVID-19, you find a corollary narrative about pent-up demand.

A strong narrative structure, like we have with “Short and Deep”, means that NEW information or missionary statements tend to be accepted if they conform to the existing narrative, and tend to be rejected or marginalized if they oppose the existing narrative. This is what people often refer to as “confirmation bias” if you want to take a cognitive or behavioral economics approach, or you can express it as a rational outcome of Bayesian information theory if that’s your cup of tea.

We’re in the middle of earnings season right now, which means that most of the missionary statements we hear come from CEOs and CFOs on their quarterly earnings call, followed by buy-side and sell-side missionary statements about those calls, followed by media amplification of the missionary statements that fit the “Short and Deep” narrative and cartoonification or downplaying of the statements that don’t.

Since these earnings calls are focused on what just happened in Q1 and what is being experienced in Q2 (almost every company has pulled their long-range guidance) it is nearly impossible for these calls to change our minds about “Short and Deep”.

What happens instead is that the market reaction to many Q1 earnings calls is a positive response to news that was “not as bad as expected.” That’s not a prediction about those companies or their specific earnings situations (we really have no idea), but an observation that the narrative structure provided the language to absorb reports that could be aligned with the “depth” of the expected outcome. In other words, there’s a structural narrative asymmetry to positive earnings “surprises” today, not because these companies are doing surprisingly well in real-world, but because they compare well to what’s dominating the way we talk about these companies in narrative-world. Narrative expectations become a strong market tail-wind.

It’s not that market missionaries are lying about what they are experiencing in real-world. On the contrary, virtually every missionary statement we’ve looked at has been consistent in expressing clearly and earnestly the depth of the problem in Q1 and continuing into Q2. But that directness has in turn granted those parties credibility to express optimism about the briefness of the revenue/earnings problem, so that the narrative of “Yay, Pent-up Demand!” and rapid economic repair in Q3 and Q4 becomes common knowledge – what everyone knows that everyone knows.

And that’s the catalyst we think could reverse the current market-positive narrative structure.

If real-world demand for goods and services remains at these abysmal levels … if language about the long-term economic consequences of COVID-19 begins to spread beyond the unemployment cluster … then the common knowledge of “Pent-up Demand” will diminish and the narrative structure of this market can become profoundly skewed to the downside.

What investors should watch for

What would we be looking for to judge whether this narrative structure is reversing? Here’s our list:

• 2H 2020 Guidance: The rubber will begin to meet the road in updated guidance on the second half of 2020 that we should begin to see following the relaxation of stay-at-home orders throughout late May and early June. Because of the “Pent-up demand” narrative, we expect a lot of the initial reports (correctly or incorrectly) to be passed through that existing framing. But the guidance from management that will follow that has a different objective – they want to beat, and they want to get grants with optionality priced to permit those beats to benefit them. We believe there will be more information (in terms of influencing investors to change their minds) in the guidance updates in June/July than in any corporate missionary behavior about COVID-19 to date.

• Real-World Retail Knock-On Narratives: The delay in retail bankruptcies in expectation of liquidation sales that will accompany a relaxation in lockdowns is not a novel observation. There are, however, credit and real estate funds with meaningful exposure here that almost certainly sought and will seek to dampen impacts on Q1 – and yes, Q2 – marks by not accounting for the full coming destruction in value to which they are exposed. The same goes for oil & gas, where hope springs eternal on 1.1-1.3x marks which relax the typical commodity price-driven model approach on the basis of ignoring “short-term volatility.” Real world BKs cause real-world portfolio illiquidity. Be aware.

• Second Half 2020 Lockdowns: As we discussed in a recent ET Live, we think that the fundamental corollary to housing prices in 2007/2008 is the actual, real-world ability of COVID-19 to cause new hot spots that lead to recurrence of stay-at-home orders in some US regions later in 2020. If these re-lockdowns occur, they will place enormous pressure on the “length of time” narrative about COVID-19’s economic effects.

• Democratic Political Narratives: Investors should expect the COVID-19 pandemic to be ruthlessly politicized as part of the 2020 election in ways that it has not been politicized to-date. They should furthermore expect the promotion of new narratives – especially by the Democrats – about the economic reality and the length over which the effects of that reality will be felt. There will be “Trump sank us into a depression” narratives. They do not exist today – at all. They will be new, and they should be monitored.

• GOP Political Policy: Election year politics don’t just show up in campaign slogans and rhetoric about “Trump sinking us into a depression” or “Trump doing everything to rescue Americans.” They show up in policy. The embedded assumption in markets that fiscal policy will keep supporting small businesses and households is part of the narrative structure now, and the White House cannot allow that narrative to falter. This IS their reelection strategy. Yes, the DNC will try to hang a depression narrative around Donald Trump’s neck. The President and the Fed and the GOP-controlled Senate, however, will continue aggressive policy action to ensure a strong S&P 500 and other cartoons of better-than-expected economic outcomes.

The nice thing about this list, I think, is that fundamental investors and allocators are well positioned to evaluate these items. You don’t need AI or a massive team of analysts or even the Narrative Machine to keep up here. Meanwhile, though, we will keep focusing our research on the short-term versus long-term linguistic structure of how we talk about markets. Together, we will get through this!

These observations are a summary of the conclusions we draw from our research into certain of the narrative structures that we believe influence US equity markets. These observations are provided for informational purposes only and do not represent a recommendation or investment advice. These reflect the general views extracted from our research and not our opinion on what you, the reader, should do. Individuals and professionals alike should consider a range of issues before making any investment decision, including any related to the topics described below. There is no guarantee that any decision made using this information will work.

These are sector views and don’t reflect individual companies, but it’s worth repeating because we are so focused on eliminating any potential conflicts: Second Foundation doesn’t provide investment banking or other services to any of these issuers. We don’t permit trading of these instruments by employees, even though they are broad-market ETFs.

Comments

• We think a month of China Trade War-linked volatility broke up any emerging narrative around equity yield. The pieces pushing them disappeared as quickly as they emerged.
• The model appears to be responding positively to an erosion in any consistent narrative for Real Estate and Technology stocks, in particular. Our approach to narrative investing tends to favor more fractured narrative structures with more diverse stories being told about them.
• In contrast, the high attention and high cohesion around Energy continues to be a red flag in our positioning.

ETNA US Sector Model Indications

Energy: Strong Underweight

Utilities: Overweight

Information Technology: Overweight

Materials: Underweight

Industrials: Underweight

Consumer Discretionary: Underweight

Consumer Staples: Underweight

Health Care: Underweight

Financial Services: Underweight

Real Estate: Overweight

Communication Services: Overweight

As part of our narrative monitoring process, we occasionally receive requests for analysis of specific narratives. We also make our own anecdotal observations about what feels to us like an emerging narrative. Both sources have led us to explore the structure of anti-trust and monopoly/oligopoly narratives in the US technology sector.

Definitions first:

• This is an exploration of the existence, affect/sentiment, cohesion and attention being paid to the topic in financial media.

• Cohesion measures how internally similar the language in articles discussing anti-trust and monopoly risks and claims about the tech sector has been over some period.

• Attention measures how similar the language in those articles is to the broader universe of articles discussing the tech sector and tech stocks more broadly.

• Sentiment is a basic measure of the affect value of the language used. In short, are the articles negative or positive?

Volume

While not a part of how we think about narrative structure, it’s still useful to understand how much is being written about a topic. If nothing else, frequency is usually the thing that sparks our awareness that a topic may be part of the Zeitgeist.

If you’ve noticed an uptick in discussion of ‘tech monopolies’, you are not imagining things. The volume of coverage this year has increased. Our dataset includes 708 unique such articles from January 2019. By June 2019 that number had risen to 2,700 before settling slightly at 2,200 in July. The increase has been steady, but most took place in connection with coverage of the 2020 Elections and Democratic debates.

Sentiment

The sentiment of articles published become consistently more negative over the course of 2019, in no small part (we think) to the increasingly aggressive tenor of criticisms from both the political left and right as part of the 2020 campaign.

The scale below runs between -1 and 1, where -1 would indicate that 100% of articles used language which, on balance, carried negative affect. Technology industry coverage tends to be positive, but even if that weren’t the case, a sentiment shift of this magnitude would still be significant.

Fiat News

Fiat News is our measure of the use of explanatory / opinion / causality language in articles about a topic. It is usually very stable, and is most informative – we think – at inflection points. The percentage of articles about “big tech monopolies’ which have included Fiat News language has been creeping higher for most of 2019.

Cohesion

As with negative sentiment and fiat news content associated with Anti-Trust narratives, the cohesion of Big Tech anti-trust content has risen meaningfully in 2019. When people write about the topic, they are increasingly writing the same things, using the same arguments and same language. This is typically our first stop when seeking to identify an emerging narrative.

Attention

The most interesting observation from the narrative structure, however, is that this otherwise negative, fiat news-laden, cohesive story about Big Tech and claims being made about its monopolistic / oligopolistic / anti-competitive behavior, has actually faded from the structure of narrative about these companies in financial markets-focused media.

In other words, when we examine how closely related anti-trust narratives are to the stories being told about Big Tech stocks, the answer we get is: not very, and less than at the beginning of the year.

The graph below shows the narrative structure around tech stocks within broader stock market-focused financial media. The dark/bold nodes are anti-trust / monopoly nodes. In short, this remains a peripheral narrative.

Conclusions / Commentary

• We have some views on the extent to which various technology companies are, in fact, demonstrating monopolistic/oligopolistic behavior. It is not difficult to argue that this is taking place in advertising markets, for example. None of the above reflects these views.

• We likewise have practically zero view (at this stage, anyway) and zero edge on the odds of any action that might be taken against these companies.

• We DO think the lack of attention makes this theme as a catalyst to a portfolio position less attractive than usual.

• On the other hand, for an asymmetry-driven thesis, we would argue that the risk of a increasingly negative, cohesive narrative coalescing around some of the large technology stocks is being substantially underdiscounted. We would expect emergence of this narrative into market common knowledge to have a significant impact, although as noted above, nothing here gives us any edge/insight into predicting the odds of that taking place.

These observations are a summary of the conclusions we draw from our research into certain of the narrative structures that we believe influence US equity markets. These observations are provided for informational purposes only and do not represent a recommendation or investment advice. These reflect the general views extracted from our research and not our opinion on what you, the reader, should do. Individuals and professionals alike should consider a range of issues before making any investment decision, including any related to the topics described below. There is no guarantee that any decision made using this information will work.

These are sector views and don’t reflect individual companies, but it’s worth repeating because we are so focused on eliminating any potential conflicts: Second Foundation doesn’t provide investment banking or other services to any of these issuers. We don’t permit trading of these instruments by employees, even though they are broad-market ETFs.

Comments

• Our July ETNA Sector runs indicate a narrative of increasing coherence and rapidly accelerating attention around “rotation trades for a rate cut cycle!”
• We are seeing the return of dividend and yield-pushing in financial media, sell-side and buy-side sources alike, which manifests in our data through the sectors which tend to be most closely associated with those profiles.
• The result is a pretty stark view fading these now-common-knowledge narratives through their proxies in Staples, Health Care and Utilities in particular.
• To a lesser extent, we think this may represent a fading of some “value rotation” narratives, but we’re less confident in their coherence.

ETNA US Sector Model Indications

Energy: Underweight

Utilities: Underweight

Information Technology: Overweight

Materials: Overweight

Industrials: Neutral

Consumer Discretionary: Neutral

Consumer Staples: Underweight

Health Care: Underweight

Financial Services: Neutral

Real Estate: Underweight

Communication Services: Neutral

To download a PDF of this In Focus note, click here.

As Ben mentioned in his most recent email to Epsilon Theory Professional subscribers, you are a pretty diverse group. But in general, you fall into one of three categories: (1) fund managers and traders, (2) institutional asset allocators and (3) RIAs and adviser teams working with wealthy families. Not everything we publish will be equally useful to these groups. And not everything will be purely about narrative. We think about other things, too.

So it is that this note is really meant for the fund managers and traders, although the rest of you are welcome to read along – I’ll open up the comments for any who want to add them as well. They’ll only be visible to other ET Pro subscribers, and if you would like to ensure your profile is anonymous before doing so, we can help with that to. Instead of writing about narrative, however, this time I want to write about the thing that most of you consider the bane of your existence. Not risk management. Not central banks. Not periods of low volatility or dispersion. Not The Algos. Fundraising.

It is on my mind after reading a piece published a couple days ago in Institutional Investor by Amy Whyte. It’s a pretty good piece, and it is very thorough. It is also written for broad consumption, which means that most of the on-the-record quotes from people in the industry are circumspect and not especially helpful. We don’t have that problem here. I have been on each of the asset owner, fund manager and gatekeeper sides of the table, so let me share my not-nearly-as-circumspect views on how you will and won’t raise institutional money.

How You Won’t Raise Money

Relying on Rolodexes. If there is nothing else you get from this note, please hear me on this point. When you hire your institutional salesperson, they will sell you on their sales skills and personality. Great. They will sell you on their track record. Super. They will sell you on their temperament and drive. Outstanding. They will sell you on how well they understand and can describe your strategies. Perfect. But most of all they will tout their rolodex.

Ignore it.

Here’s the thing: Your sales guy or gal may have made a sale to an institution in the past, or they may have a college buddy, or they may have had enough drinks or dinners over the years that their calls get answered, which will feel like something, at least. Progress. Salespeople have it drilled into their heads that sales success is about getting at-bats, which is not entirely false. But in this case, a meeting taken as a favor is not an at-bat. It’s a fan paying for a Be a Player For a Day experience in which he gets to walk around the bases before the game for $500 or so.

I have never bought from the same salesperson working at two different shops, and I don’t know anyone who has. It’s not because I’m averse to it, but because this isn’t how decisions of this size and magnitude are made. The usual salesperson response will be, “Maybe not at your institution, but…”, but this is a red herring. Let’s suppose that we allowed them to finish that predictable sentence. Even then, “Tell me about your rolodex” is the wrong question. What is the right question?

Pick ten institutions you think should be clients. Maybe you heard their CIO speak. Maybe you know someone else in their portfolio. Maybe you know that you’re in their sweet spot. Now ask your salesperson or candidate to describe how each of those institutions makes decisions. Does their CIO call the shots? Is the consultant an ornamental box-check, or are they a primary idea source? Is the staff glorified ODD or are they the ones populating lists? Who needs to be the internal lead, who needs to buy in, and who needs to provide an ornamental risk-transfer sign-off?

A salesperson who can answer those questions for you will be infinitely more valuable to your fundraising process than someone who can impose on someone to answer their email.

Going Over Staff’s Head. This is a derivative point to the first, but unless you are 100% sure that the CIO and only the CIO calls the shots, there is no more doomed-to-fail strategy than the Call the CIO when the staff don’t return your “I’m coming to town on…” emails strategy. This will get you put on an irrevocable black list with a lot of people, and it’s the one thing you can do that really does make its way into peer discussions. When asset owners gossip over drinks or at conferences, they talk a little bit about managers. They talk a little about returns. And they talk a lot about the numbers like this that salespeople and PMs pull. Names are not omitted.  

By Going the Emerging Manager Route. At a certain point in the fund manager’s life cycle, submitting your materials to an emerging manager search looks promising. There’s a reason why: because these programs are designed by their very nature to take a lot of initial meetings. Because talking about casting a wide net is how the gatekeepers in this space sell themselves to the staffs of asset owners who are forced to implement an emerging manager program that is usually being forced down their throats by a board or other governing body. The entire funnel for these platforms is designed to take a huge number of inputs into the first step – which may involve a significant amount of RFP work, meetings and data preparation on your part – but to graduate very few to anything further.

Once managers are advanced to the next step, the approval process is often a political one, and baby-splitting is extremely common – if there are disagreements, allocations will more often be split between more managers, making ultimate allocation sizes smaller than initially advertised.  From there you will be sold the carrot of graduation to a ‘full manager’ over time, but in practice this is rare (some programs do it better than others). There are obviously success stories, but in general, I expect you will find that the experience of emerging manager program RFPs is a continual exercise in frustration.

How You Will Raise Money

Be Big. Some of the ways you will raise money, like already having AUM, are frustratingly obvious. Once upon a time it used to be $100 million you needed, then $500 million, but these days it’s really $1 billion for most any long or alternative strategy. In many strategies it’s a lot more than even that to be in realistic conversations with consultants for any search. If you don’t have size, you’re going to have to rely on one of the other traits below, or simply good fortune and 3-yards-and-a-cloud-of-dust labor and luck with multi-family or the hollowed out husk of the fund-of-funds industry.

Be Famous Already. Yes, like size, you may not be able to do anything about this, but even after the high profile Goldman prop desk flameouts and the ex-SAC disappointments, name recognition is one of the few things that can get you to the front of the line if you lack the other typical required characteristics. The reason is that it’s one of the few things that allows those of us who are allocators and gatekeepers to pretend that we’re willing to take career risk on new funds and managers without actually taking that risk.

Have Returns that Can’t Be Ignored. Sure, this one is obvious, too, but let me put some figures to it. It’s going to vary by institutions, but one year of insane returns (i.e. think 10% ahead of peers or a benchmark) or three years of really good returns (i.e. >5% annualized excess returns or peer group outperformance) will usually get you inbounds or get your outbounds answered, even if you’re small and just one of the unwashed masses like me and Ben. It is extremely rare that a fund manager will get traction outside of past investors without one of the three above traits. Short of good fortune, the chicken-and-egg problem will be your constant companion, because you can’t really control whether you have any of the above three traits to sell.

There are two traits, however, over which you do have some control.

Be in a Category Where Everyone is Small and Capacity Matters. Your strategy may be your strategy, and it may not be portable to other style boxes, asset classes and categories. But as an asset allocator, there are some strategies where my threshold for a meeting has always been far lower, because anyone who is any good is usually closed:

• Small cap and micro cap, both long only and long/short. Foreign, domestic and global;
• Stat arb;
• Activist equity; or
• Higher vol and higher tracking error strategy variants across a huge number of strategies.

Maybe you can’t do any of these things, although higher vol / higher tracking error / higher concentration strategies are usually a feasible option for most fund managers. But if you don’t have the benefit of any of the first three traits, you can do far worse than considering whether your principles, philosophy and process are compatible with one of these areas.

Go Over the Consultant’s Head. I know I mentioned the meta-game weakness of going over the staff’s head to the CIO, but this is different. As a fund manager who isn’t huge, isn’t famous and doesn’t have knock-you-over returns, identifying the asset allocators who vest almost all of their decision-making authority in an internal professional staff is the single most important thing you can do to improve your odds of fund-raising success.

Consultants are not paid to take risk, but to transfer it away from staffs and boards. They must do just enough to go through the motions of appearing to look for new and interesting strategies, but “just enough” probably doesn’t include looking seriously at your fund. Sorry. Allocators vesting both decision-making power and all functional authority in CIOs can be tempting – the only person you have to convince is me – but the problem is similar. A CIO in which power is vested is a CIO who in most cases has to be judicious about the types of risk to which they subject themselves. Most such individuals are much more likely to spend that currency on illiquidity and asset allocation views, not on the distraction of manager blowups with the potential for the blowback of maverick risk.

Professional staff-driven institutions, however, are often the ones capable of absorbing this kind of risk into their process. They also benefit from the ambitions of staff, who in many cases need to demonstrate decisions and philosophies which distinguish them from peers. This is not ubiquitous. You will find this mindset within many endowments and foundations, some public pensions, few corporate pensions and practically no Taft-Hartley plans.

How do you puzzle this out? For public institutions, including almost all public pensions and a lot of university endowments without private management companies, you should be able to watch the board meetings. Do it. Watch enough and you should develop insights into their processes that will be more valuable than working up a pointless RFP that you won’t win. For other institutions, it is one of the few reasons to attend a conference – not to sell (please don’t) to whomever is there, but to engage the staff, to talk to them about how they make decisions. How they invest. It’s flattering and the kind of thing that the ones who’d be willing to take a risk on you would typically enjoy.

It’s a soapbox, but one I feel comfortable preaching from. If you are submitting RFPs and putting in calls to consultants without having done the same kind of research on a prospective client you would perform on a portfolio company, you’re doing it wrong. And an unhealthy obsession with consultants and gatekeepers – unless in the case of a medium fund manager trying to become a big fund manager – usually tells me that someone is doing it wrong.

“Happy families are all alike; every unhappy family is unhappy in its own way.”

Anna Karenina, by Leo Tolstoy

As we continue to expand the Epsilon Theory Professional research platform, you will soon hear us discuss an additional standardized metric which will supplement our existing monitors. While we will fully introduce the metric in that more detailed update to the FAQ, the start to the year has provided us with an interesting test case. At the least, an introduction is in order.

In short, the measure which we have called “Attention” to-date will shortly be changed to “Cohesion.” The new measure will take over the “Attention” moniker. That is a bit confusing on the surface, but the reason for it should become apparent. The Cohesion measure (previously Attention) is a standardized measure of the internal cohesiveness of a Narrative within a topic; for example, it would measure the average connectedness and similarity of language within an economic sector, or a macroeconomic theme, or one of our ET Pro monitor topics. The idea is to capture the extent to which the people who are writing or talking about a topic are saying the same thing.

The new measure – the new Attention – is slightly different. It is a measure of the influence of stories about a sub-topic on the overall narrative structure within a broader universe of stories concerning financial markets (such as equity markets, fixed income, etc.) or political markets. For example, Cohesion might measure the similarity of all content relating to, say, the Financials Sector in the US. Attention, on the other hand, might measure the influence of the Financials Sector articles on all articles being written about stocks and equity markets.

Our narrative research focus tends to be on sectors, asset classes, strategies and themes, because we think that those dimensions are the ones around which most investment decisions – and narratives – tend to coalesce. For that reason, we have generally published very little about countries, and our analysis has tended to be fairly US-centric. Our new Attention metric, however, gives us an interesting test bed to explore those undercovered areas.

To wit, given the relative drama in markets in December and January, we wanted to explore how the start to 2019 compares to the narratives influencing markets at the beginning of 2018 on the country dimension. Given our typical focus on the US, we instead selected Germany, Japan, the UK, Australia and Italy for a robust sample of developed foreign markets with good English-language newsflow. Below are the changes in Narrative Attention for those markets between Q1 2018 and YTD Q1 2019. Note that the measures cannot always be compared across sub-topics (e.g. Canada will almost always have a higher value because of its geographic and narrative proximity to the U.S.). It is usually most useful to compare a topic to itself over time, or to consider all of the sub-topics together.

So what do we think is going on in major country narratives?

First, we think that discussions of regional equity markets have become increasingly disconnected from one another as each region focuses on and works through pricing the risk of a pronounced idiosyncratic ‘problem.’

In other words, each is an unhappy family, unhappy in its own way.

This is readily apparent when examining the narrative map for each thus far in 2019. Italian equity market coverage is focused on slowing euro-zone growth, economic data supporting that slowdown, and lingering Italian debt crisis questions.

Italy Narrative Map – Q1 2019

By contrast, the UK narrative map is so dominated by discussion of Brexit that the Brexit sub-topic doesn’t even have a cluster – because it exists in and links nearly every other topic!

UK Narrative Map – Q1 2019

Our second view is that Germany is unique among non-US developed markets in its attachment to the main governing narrative of global equity markets right now: trade and tariffs.

This is why Germany’s narrative attention is rising while attention to other foreign equity markets has been declining as a result of rising country-specific issues.

Germany Narrative Map – Q1 2019

Our conclusions from this narrative data are as follows:

• We think that this increased isolation of country-specific narratives is a part of the Zeitgeist. We think it is something that will ebb and flow, but which will remain with as long as populist movements, pushback on globalization and trade disputes are all creators of idiosyncratic narrative structures for individual countries.
• We think this means that country overweights and underweights may produce greater tracking error than backward looking empirical measures would otherwise estimate. Overall market volatility will influence this heavily, of course. We would always be hesitant to take on significant country risk in equity strategies simply because we think that most stock selection strategies have no positive expectation on country bets. We are even more of that opinion today.
• Our view on trade dispute resolution continues to be that it is a game of chicken. In the same way that we believe Apple is being treated as a partial proxy for trade war speculation, we think that Germany and German stocks are probably being targeted for this use by fund managers with a European mandate. We would be hesitant taking meaningful active (or overweight / underweight) positions in German stocks in either direction during this prolonged process, even though most trade newsflow in the US is nominally US/China-focused.

PDF Download: Every Unhappy Family

It is a big week for Aerospace and Defense. After last week, in which United Technologies registered pretty substantial earnings and top-line beats, the week of January 28^th will see reports from Boeing (BA), Lockheed Martin (LMT), Northrop Grumman (NOC), General Dynamics (GD) and Raytheon (RTN). Here we explore the current narratives for the industry and its key members.

Our first stop will be the narrative map over the last quarter for the industry more broadly. A few observations:

• There are some industries in which people writing about the businesses and people writing about the stocks tend to use very different language. This is one. When certain business topics are linked to the stocks / financial language, we think that is worth noticing.
• The temporarily suspended US government shutdown is not well-integrated into the stock price narratives of the industry more broadly.
• The biggest surprise we noted was the apparent lack of attention to trade and tariff issues faced by these companies in comparison to other sectors. What you will see from the stock-specific network maps, however, is that these topics do not coalesce into their own cluster because they are so ubiquitous in almost every topic at present.
• Instead, attention at the market level (bottom left) appears to be much more focused on and connected to discussions of commercial air travel, orders, competition and – somewhat surprisingly – the impact of Brexit.

While the broad industry interests us, we also wanted to explore the two largest aerospace and defense players (other than the multi-industry United Technologies, which has already reported): Boeing and Lockheed Martin.

Boeing

Boeing has a low-attention narrative structure – even more so than the broader industry – by which we mean that people who are writing articles or research about the stock are using language that bears very little similarity to comparable discussions of its business, products, technologies and people. It is our belief and has been our observation that these narrative structures are more susceptible to new event information (i.e. we believe that they reflect asymmetric discounting of fundamental information). Low-attention in our taxonomy does not convey information about sentiment. In this case, sentiment for Boeing is mixed, which reinforces our belief the company is more susceptible than usual to event-related surprise and volatility (and that we would be less inclined toward most trend-following strategies relating to the stock).

As you will observe in the chart below, the data indicate three largely separate types of language being used in discussions about Boeing in the last quarter: (1) Market-focused article language, (2) Commercial Aircraft-focused article language and (3) Defense-focused article language. To be clear, this does not necessarily mean that this is what the articles are about, but that they share language with other stories than tend to be about those topics more broadly. We have highlighted these “meta-clusters” in the oval-shaped areas. The separation of these topics from one another is much stronger than we typically observe. General observations we would make about Boeing’s narrative structure:

• We see very little evidence of a strong central narrative.
• While the meta-clusters are fairly loosely connected to market narratives about the stock, unlike the broader industry, this map shows how closely the China Trade War narrative is to the core of stories being written about a company like Boeing.
• We were surprised to see the (relatively high) connectedness of air travel safety topics as well. To the extent investors believe that such concerns tend to be overstated, we think there would be some reason to believe that those concerns may be overdiscounted.

Lockheed Martin

The Lockheed Martin narrative map is still fairly disconnected, but less so than Boeing’s. We would consider it a map of moderate attention. Discussions of some – although not all – of Lockheed’s various projects and competition use very similar language to what is used in analysis of its stock and financial prospects, and demonstrate much greater interconnectivity. In general, relative to Boeing we think this makes the stock more resilient against event-related surprises (both positive and negative) from topics close to the center of the narrative and its explicitly market-related topics.

Our key observations about the narrative structure for Lockheed Martin:

• Trump’s foreign policy is deeply embedded into the narrative about Lockheed Martin. This is not a novel observation about Lockheed Martin or any company with such a heavy defense focus, but the narrative confirms that intuition: when people talk about Trump’s influence on geopolitical risk, they use a lot of the same language they use to talk about Lockheed’s opportunity set.
• In our judgment, the projects most closely connected to the stock narrative are, in this order:
  • F-35      
  • Aerial refueling (MRTT)
  • Long-range anti-ship missile (LRASM)
  • Hypersonic missile and drone contracts (most articles general, less project-focused)
  • Littoral combat ship (LCS)
• As per usual, F-35 sentiment is almost uniformly negative, but we do note a significant number of articles with Fiat News language that focus on new technology packages, impressive air-to-ground strike tests, etc. Not explicit propaganda, but it is clear to us that there is missionary work about the F-35 taking place to shift the sentiment in a positive direction. Over the past three years, there were 246 stories in our universe of high-circulation sources calling the F-35 program a failure, 144 calling it a disaster and 143 calling it a mistake. Over the last quarter, the frequency of those words has been cut in half, except for ‘disaster’, which made zero appearances. Whether we agreed with those observations or not, we would be very mindful of how they are influencing real narratives, as this has been a significant negative part of the LMT narrative for a very long time.
• Saudi weapon sales and the Kashoggi event are STILL part of the entangled “Trump Foreign Policy” narrative. This hasn’t gone away as an issue for Lockheed Martin.
• Space, rocketry and satellites are very disconnected from the core Lockheed Martin narrative.
• As we noted in our banking note from two weeks ago, gender remains a major topic across industries. With a woman as CEO, this is a very intensely connected part of the Lockheed Martin narrative and makes its way into topics of surprising breadth.

PDF Download: Getting Defensive

With earnings season for banks kicking off with a vengeance this week, we thought it appropriate to dive into the recent structure of narratives around America’s largest banks. Using a broad range of news, sell-side research and other sources between November 1, 2018 and January 13, 2019, we explored the topics of note surrounding Goldman Sachs, Morgan Stanley, JP Morgan and Citigroup as an aggregated entity. As you might imagine, these ngrams (i.e. keywords) also tend to bring back a range of reports being published by those institutions. We have excluded those to attempt to focus on articles that are about them.

The resulting Narrative Map is provided below.

It is a visually complicated map, indicative of little more than the significant number of articles over this period covering these institutions. But even in the visualization, your eye will be drawn to what is apparent in the data. There are really two “central” clusters of topics, and then some loosely attached clusters which don’t appear to be closely related by language and terminology to either.

In the visualization, you can spot these by looking for white space. If you are having trouble, you may think of Metacluster 1 as being the topics in the top-center of the map, beginning with “Mnuchin Calls Bank CEOs” down and to the right through “Poor Market Performance…” You may think of Metacluster 2 as being the topics in the center-bottom of the chart, beginning with “Gender Equity on Wall Street” through “M&A Outlooks and Deal Announcements.” The peripheral topics, which have relatively sparse connectivity to the two metaclusters, are arrayed on the left edge of the map.

The intuitive observation you make from reviewing these metaclusters is similar to what you would draw from the data. Metacluster 1 is a relatively dour sentiment area of the map, which includes reports of market performance over the last quarter, significant concerns about central bank policy, interest rates and the yield curve, and generally negative outlooks for growth in the global economy going into 2019. The center of gravity to this metacluster, and the third most interconnected topic of the entire map, is found in pieces discussing the positioning of funds and recommendations of funds and strategies to weather this environment.

Metacluster 2, on the other hand, is largely positive. In different ways, it is about the opportunities for banks and other financial services companies. It is about the power of new technologies to transform digital banking. It is about exciting IPOs coming up in 2019. It is about transformational acquisitions, new business initiatives and new industries (yes, cannabis again) in need of banking solutions. It is about disruptors (e.g. Robinhood, MS/Fido’s new exchange).

We think missionaries are promoting two different narratives through these two Metaclusters:

1. “Markets are underestimating the transformations going on right now in digital banking, and they are exciting.”
2. “Right when the rotation to banks was supposed to happen, the global economy does this?”

The difference in sentiment between the two is obvious. Below we show the same map, but with coloration based on sentiment of the articles rather than its cluster relationship.

What key observations would we make, and what conclusions would we draw?

• The market stories – positioning, outlook and recommendations – are all in Metacluster 1. Right now, it’s expectations around the macroeconomy, trade fears and monetary policy that are driving the ship.
• Metacluster 2 is only loosely connected to these topics. We’d generally expect positive operational surprises on these sorts of initiatives to be initially underdiscounted.
• What we’d be looking for: watch for bank execs trying to swivel the dominant narrative back toward idiosyncratic opportunities and initiatives.
• M&A and capital markets activity are unusually distant from the core stories being told about banks. As you consider what is being discounted by other participants, we would take this into account.

Three more things:

• Everyone is writing about the Jho Low affair because it’s interesting and salacious, but as a topic it’s almost completely unconnected and isolated. Based on our measure for cluster interconnectedness, it’s has about 9% of the interconnectedness of the “Growth Slowdown, Rates and Fed” cluster.
• Crypto and blockchain are peripheral topics for banks. Sorry, true believers. Just under 15% of the interconnectedness of the “Growth Slowdown, Rates and Fed” cluster.
• Remember how I said that the fund positioning cluster was the third most interconnected cluster? Number two belongs to “Growth Slowdown, Rates and Fed”. Number one? “Gender Equality on Wall Street.” #MeToo ain’t done with banking yet, y’all.

PDF Download: A Tale of Two Tales (Banking)

Sentiment: Very Negative and Stable

Attention: Very Strong and Rising

Leading into Apple’s earnings announcement last November, the product-level narratives for Apple were both remarkably positive and reflective of a consistent narrative: ‘these new phones are going to be a Big Deal’. We noted at the time that the narrative around this product launch was as positive as any since the iPhone 6, which may be among the best-performing products in Apple’s history. Expectations were high.

Company-level narratives incorporated a more complete picture of the various forces facing the company, some good and some bad. In the quarter leading up to the November earnings release, we think there were two active narratives about Apple stock (shown below generally within the two oval shapes):

1. As noted above, a product- and business-line narrative that was positive and idiosyncratic to Apple, portraying higher-priced devices in light of Buffett’s belief (and double-down) that this was an underexploited opportunity to expand margins, alongside news of the company’s increased success and efficiency in its supplier channels.
2. A market and stock-based narrative (the lower oval) of Apple as part of a growing concern about big technology stocks, the influence of federal reserve policy and interest rates, and Apple’s fate as the company that had, to this point, fared the best.

Narrative Structure of Apple Stock – July 2018 – October 2018

Immediately following Apple’s earnings release – and more importantly, its announcement that it would no longer release unit sales figures for its flagship products, two things happened. First, the stock performed very poorly. In short, the company’s announcement flipped the surprisingly positive lens with which the market was taking in unit price information into a decidedly negative one. In this, Apple executives discovered one of two ways to lose the narrative game – to flip a positive, high attention component of the core narrative to a negative one.

The second effect, of course, is that the narrative structure shifted – quickly. The market narratives attached to concerns about technology stocks became central to all of the stories written about Apple. When authors, sell-side analysts, strategists and buy side investor letter writers wrote about Apple, they were also writing about the rest of the wreckage in technology stocks. More importantly, the only two clusters of Apple-specific topics connected to this central narrative about the stock’s performance were (1) fears about China and the trade deal and (2) concerns about the unit sales reporting policy and what that might imply about Apple’s expectations for sales. Both were decidedly negative topics in their treatment of Apple.

Narrative Structure of Apple Stock – November 2018

With this narrative structure in place, in early January 2019, Apple executives discovered the second way to lose on narrative: confirm a negative peripheral component of a high attention narrative, bringing it into the core.

When Apple guided lower on sales and pointed to China in the process, they confirmed the two only high attention, highly connected components of the narrative to the stories being told about Apple, both of which happened to be negative. What’s more, the narrative became more concentrated and more interconnected as the calendar approached 2019. As shown in the over-time chart of attention and sentiment below, our attention measure here reached near-peak levels in December 2018.

What now?

Our view at this time would depend heavily on our time horizon and strategy, but this narrative structure does lead us to a few conclusions:

• For better or worse, Apple is now a stock whose narrative is tied up in the China Trade War narrative. If they aren’t already, we think that traders and investors who are attempting to trade this idea creatively (or do so reflexively) will now be using Apple as part of their thematic “trade war” baskets. As we have written elsewhere, we believe this is a classic game of chicken whose odds cannot be predicted. For shorter-term investors and traders, that means that we think active positions in AAPL are likely to have more or less random outcomes with significant volatility while this broader narrative dominates market attention. Shrink your book.
• For cross-sectional trend-followers, we think that the narrative environment of a strong core narrative and consistent sentiment is more conducive than usual to trend formation over the medium term. We would still be mindful of the higher volatility and potential randomness of the issue highlighted in the first bullet above.
• For longer-term investors less sensitive to short-term volatility, especially long-biased investors or financial advisers, we view the realization of the two highest attention, negative-sentiment topics to Apple’s narrative at once as a positive for active positioning. While any decision should be subject to your own fundamental analysis of potential emerging issues of concern for the company (on which we have no view), barring any such concerns, the major narrative threats to the perception of the company are (for the time being) very much in the open and being discounted to some extent.

PDF Download: Multiple Ways to Lose (AAPL)

Download In Focus PDF HERE.

Sentiment: Positive and Improving

Attention: Weak and Declining

Current Narrative Map Highlights

Current Sentiment and Attention

Commentary

In our broader piece on trend-following (Figaro), we noted that narratives for Tesla had demonstrated moderate-to-high attention (i.e. internal consistency) and very negative sentiment for much of the second half of 2017 into 2018. The causes for this were many, but the most interconnected negative clusters concerned (1) continued underperformance  on production numbers, (2) Elon Musk’s unusual behavior, especially on social media and (3) a number of worrisome key departures in areas that implied potential fraud. We felt that this was a supportive environment for negative trending.

We also noted that if Tesla were to break these narratives, it would be by reinforcing the positively received areas of the narrative (esp. Gigafactory production and Asia) , and by delivering a professional earnings call that might relax the frantic “management problems” components of the strong narrative.

Did they succeed? Yes, in part, we think they did. The aggregate sentiment around Tesla has improved dramatically, but the cohesiveness of the narrative strikes us as still being very mixed, similar to this time last year. We would be less convinced of betting on recent price behaviors as part of a long-term trend until there is more resolution on the negative components of the narrative.

In the medium term (and in addition to whatever long or short thesis we had), as investors in TSLA we would focus most on:

• How market appears to be attaching Tesla to tariff and China issue. This is becoming central to the main narrative and has deeply negative sentiment.
• Perception of CEO behavior (“single digits weeks away” is still a popular n-gram across articles)

PDF Download: Tesla (12/2018)