HF interviews, case studies, and investment pitches always seem to highlight how there must be a variant perception underlying the thesis (and the catalyst to get paid). How often are positions actually taken on this basis in the hedge fund world (thinking of the SM HF, and perhaps even the MM pod shops)
Where I currently work, sometimes positions are simply: this business is a great business, we like the underlying sector and secular growth trends, and its not that expensive at the moment (over simplified but you get the gist). Other times it can be "xyz" is a great defensive position for the portfolio, and offers a steady dividend yield that we think should continue to grow, and anchors the rest of the portfolio.
A lot of times we just run a DCF, see the various projected trajectories for the business, and determine that it is a decent price to own the business based on the potential outcomes.
When i see HFs owning FANG stocks and other super large caps, I often wonder how often is there truly a thesis from a variant perception at play?
Another example: owning travel stocks or office REITs in late 2020 as a play on the COVID recovery seems pretty straightforward and "consensusy" (maybe the variant perception is you expected the recovery to happen at a faster pace than what was being priced in, but often times it doesn't feel like you can really build conviction on those expectations, and you are mostly along for the ride in mostly consensus projections of the recovery)
As a follow up question, does anyone have any good resources/books for someone with principally long only experience looking to learn more about how to run a levered and semi beta neutral strategy? Feel like there needs to be some resources out there for those looking to learn more about how HF PMs build their portfolios and look to factor out market exposures or build market neutral portfolios.
This is a super interesting question that crossed my mind some time ago. Following.
Having a unique view is critical or you're just collecting the expected returns of the stock, which is the purview of long-only. Long-only is like being a curator, while (MM) long/short is more like being a detective.
In the industry there is a gradient of teams/pods, with some being better and others being pretty bad. Generally speaking, a variant perspective should drive almost all of your theses and positions in the book. There can be 10-25% left to tactical (timing related) efforts, but generally this would be tied to the lifting of some negative upcoming catalyst.
When talking about a variant perspective, you need to first understand the distribution of views and underlying arguments for them. E.g. "bull-case" is the cluster of views on the right tail, "bear-case" on the left tail, and "base-case" is around the mean (consensus). Many times, "consensus" is skewed from two primary dimensions: distribution of views (mean vs. average) and timeliness, which is why the most recent views are typically overweight in custom estimate revision indicators.
Once you have identified that there are clusters of views, you can observe if there has been a drift from one side to the other over time (4wk-change), and can spend time breaking down the argument of each side to understand and evaluate the critical factors. You do this by reading equity research reports and talking to analysts. Your goal initially is to capture all of their assumptions so you can test them. You'll do this for each company you cover, which makes building a variant perspective easier, as there are more moving parts, read-throughs, and potential sources of ideas.
Two things primarily drive earnings: revenue growth and margins. As an analyst, you'll spend time building a dashboard of "leading indicators" of demand for your coverage group's product. You'll do the same for major components of margin (sg&a, etc.). You should assume that everyone else on the street is doing the same, so this is not sufficient to create a variant view. However, this helps make you -1day informed (informed as of yesterday).
For beats in the quarter, your index of indicators should be fairly correlated to the direction of revision change across analysts (and you!). For raises, you need to tie in some longer-term indicators (related to your peer group) or identify catalysts that could drive revenue/margins higher. The amount of work you need to do in order to analyze a catalyst is quite high and is very research intensive. Work around a catalyst generally involves a combination of expert calls, private surveys (employees, competitors, and end-market), anecdotal evidence, high frequency data (say weekly credit card spend, sat imagery, etc.), and other sources of edge. There's no single data point that'll give you a sufficient amount of confidence in your thesis; you must stitch together evidence to provide a coherent and cohesive narrative that defines your view and explains the difference in estimates vs. the street.
Thanks this was very informative. I use expert call networks in my research, and while super useful and time efficient, it can also be somewhat anecdotal (even when they do surveys - except for maybe the big sample size ones).
It sounds like a lot of this is playing the earnings moves the majority of the time - what percentage of theses are focused on quarterly earnings vs. other catalysts (news release, new product, M&A, economic indicators, etc.) would you say?.
I know the short-termism of MM's has been covered before, but it really sounds like you are dead focused on being more accurate than the sell-side (and the buy side/whisper number) on each earnings report to just try and play the stock's move that quarter.
Do SM approach things the same way?
PS:any recs on resources that review how to build and manage a portfolio this way- maybe also taking into account hedging out exposures? I do a lot of this work already, although a lot less focused on quarterly results. I generally agree that to be paid 2/20 you need to be capitalizing on variant perceptions, just wondering how often this occurs. Sounds like for MM pods its true the majority of the time, but the focus is primarily on "who can model the next quarter the most accurately"
SM claims to be more like long only, but in reality employs a lot of MM techniques. Nothing wrong with it.
S-curve Capital on Twitter about people at Tiger Cub:
I prefer curating my own expert network for exclusive content. However, paid ones are good for getting up to date with product expertise and general SWOT-analysis-y stuff.
Earnings or corporate action announcements are the "payoff" events. Anything can feed into that; the impact of something like a new product would be a material change to earnings expectations.
The quarter is meaningful, but generally a l/s pod will have a view on a stock for some time and a position may last for multiple quarters or years, especially if the pod is sector focused. Earnings results or major investor presentations are ways information about a company is released... pre-release you are trading the blind based upon the information you have.
Yes at the analyst level, no at the PM level. Most SMs are factor or sector focused.
At the analyst level, modeling the next quarter more accurately is definitely the goal. At the PM level, you need a view on the multiple (more macro), technicals (flows and positioning, not lines on a chart), and timing (recent price action, volatility, etc.). There are no great books on this stuff because it's very niche and the workflow you create is the actual IP. Everything else is just tying in financial research (time varying nature of factors, market anomalies tied to economic outcomes that drive flows) and tools (excel, bloomberg, etc.) into useful insights.
You did a good job of explaining how to source a variant view. The critical thing to add is that it's damn hard to find a situation like that, and so in the end most of a HF's book is populated by straightforward "good business, reasonable valuation, no edge" type of positions.
Everybody likes to claim otherwise, but I can tell you I have access to the best resources and data providers, and 90% of my book is typically pretty straightforward stuff with no special edge.
Haha thanks. I'd say that you're right but that generally there's both a macro and factor timing dimension in addition to "this is a good business and the street likes it". E.g. I think the 10-year is going lower, so I screen software companies down the most that are highly rated and are seeing strong estimate revisions-- voila! Beta timing alpha.![]()
I also want to put this out there -- there seems to be this idea that elite hedge funds are right 80%+ and don't experience pnl volatility.
The reality is that single name pnl volatility is quite high and the vast majority (read: assume all) hedge funds do not generate a profitable pnl on day 1. Through a quarter there can be substantial drawdowns in individual names, which is why a PM "builds a book" of positions to capture returns of an underlying, broader, thesis (long a sector they have a view on, short a macro factor, etc.).
For example, if you're a TMT PM then you pretty much want the beta (because you have a long-term positive view on the sector) of TMT without the volatility and risk. You might start initially by being long all companies that have beat earnings and short companies that have missed. On the surface level your pnl might be relatively smooth (also depends on time of day! lots of SMID caps have lesser liquidity in the morning so pnl gradually becomes more efficient over the day), but single names could be up/down a bunch. Overtime you start to filter your screen more effectively to improve your overall portfolio risk/return by adding additional steps, improving estimate forecasts, and trading more aggressively around catalysts. But that still doesn't mean that each of your positions has a positive pnl slope and little volatility!
At the analyst level you will miss earnings and be wrong nearly half the time...the goal is to 1) be good at identifying the really good opportunities and 2) improve your hit rate over time. The good work you do builds the conviction required to hold through potential drawdowns.
I approach with the mindset of finding things to DISPROVE ie finding a crowded bull case or bear case where certain foundational assumptions of the prevailing thesis can be disproven through creative research or creative use of data. Rather than finding good ideas / bad ideas and proving my case. I think this keeps the discipline on being variant.
I try to start with a variant perception and then filter down to actionable / catalysts. I focus on the short side and I think it’s easier to find dislocations short, because management and banks are incentivized to create dislocations to the positive when things aren’t so rosy. I probably too often leave easy-wins on the table cause it feels too obvious - like with truckers this year, ltl rates rolled, valuations were near highs in very-peak earnings - and I shorted small but it felt too obvious - that extra cushion of variance gives you extra room to be wrong and also I think allows me trade ideas better, capturing alpha when the headline beat/miss is misleading. I have a handful of subsectors where I think there are systemic misunderstandings in unit economics or business mechanics where I can use a different process and data to have repeatably high conviction variant views. Beyond that, over time I think you build pattern recognition - types of situations to look for or even screen for where there are likely to be misleading narratives at play. For example I always keep a list of “hype” words-of-the-month like autonomous/AI-enabled/plant-based/green/sustainable or a little while back 3D-printing/automation/social-network and scan for companies with that in the description plus some financial metrics that are likely to suggest manipulation/unrealistic-outlooks/etc. I think variant view is the most important thing by far and definitely my philosophy starts with fishing in the right ponds to find dislocations. Particularly dislocations that are diligencable and material. In best cases they’re also time-able and repeatable. Just some random thoughts. That said sometimes I throw something on small with too little work cause a really smart friend told me, cause I’m dumb like everyone else.
Thanks for the contribution - the first thing you mentioned seems like it could be prone to confirmation bias (if you aren't careful, not saying that you aren't, but it seems ripe for that type of "finding the information to support a pre-conceived thesis" rather than assessing the information and developing the thesis afterwards).
I think in an invest like the best podcast (can't remember which one), they talk about how many firms talk up the "data" they have collected as a differentiated factor and often times it is worthless, so I'm sure those buzzword situations appear often...
Care to elaborate on the systemic misunderstandings section? - I have a theory to how you unconvered this/these, but just interested in your explanation/process.
To me, it feels like the key to finding variant perceptions is just in the reps and in journalistic like due diligence, but when you're managing +$5bn, it feels hard to keep every position in the portfolio based on a high conviction variant perception, just because of the types of businesses you are limited to at times (large cap) and the depth of their coverage, you might be more prone to "I like FB because its cheap if they keep their growth rates at this level for longer, and I think that is possible because I am taking a view on the world these reasons" (or maybe that is just a variant perception anyways). Maybe PYPL is a better example - everyone owned that thinking their margins could be higher by comparing it to Visa and other payment guys
Macro thematics should drive your large cap views. Keep your variant perspectives for SMID caps up/down their supply chain.
E.g. "I like consumer focused tech (ad-supported, such as FB) because improving PMIs should support ad-spend while interest rate moderation should solidify their multiples." Can go deeper with winners / losers in a sub-industry but your view should be thematic where you are essentially beta timing, and that's ok because the market is much more efficient with large caps. Not saying you can't develop a unique perspective, but it is much harder and less rewarding with large caps.
On the confirmation bias pt I think it’s actually counter to that tendency. Everyone still tell you the 3 reasons and right little story why you should buy a stock. It’s really hard to find a bull case that mgmt or peers aren’t eager to tout to you. I think it’s much simpler to go find those stories that often start with “it’s different this time, because x, y and z” if you can validate x, y and/or z is bullshit, that’s a good start. My favorite right now is “extrapolation problems” I.e. a cyclical business that people think has changed undergone a structural change and will shift to secular growth - for example RVs and boats during covid. Can we apply 2020 margins to ever increasing volumes because people tasted the outdoors and are converted forever and therefore they shouldn’t trade like a cyclical? Or will supply return and/or demand fall off and inventories build and promotions return and the industry blows up like it does every cycle?
Much data is commoditized and most guys are using the same credit card and web traffic and prepackaged data but given how fast data is being generated in all industries you can source your own or process and use public data in unique ways and there’s plenty of alpha - can’t go into more detail than that or I’ll kill the few golden gooses! Takes a lot of trial and error to find truly misunderstood business models with sufficient data to model and track better - best when you can find something to with alot of peers and readthroughs.
Reductio ad absurdum
.
Non sapiente aliquam molestias dolor nisi. Aperiam quisquam aut placeat ad hic. Aspernatur et aut itaque provident reprehenderit. Vel possimus provident iure fugiat ut.
Dolorum deserunt et praesentium tenetur consequatur. Eos et itaque quo ut explicabo.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...