How to convert sell-side models to your own models faster?

Subscribe

What's the best way to turn models received from sell-side to your own models faster and most efficiently?

Comments (9)

 
 
Most Helpful
Jul 20, 2020 - 12:18am

This is a great question. I’ve struggled with this for years, so let me share a few things.

The quick answer is to find one or two analysts that model in a way that you would build the model. You will want a full working model that isn’t a mess. One you can update in 10-15 mins for a boring quarter. That was always my approach to modeling on sell side. Hopefully a consistent assumption sheet exists that you can just change if not you might want to create one and link to it. My conversion process is simple. Audit model. Link assumptions to my own deck. Audit again then make changes.

I run a sector specific product so this should apply to a sector specific analyst.

Most important thing with this is getting to know all the analysts that you pay and understand who gets the largest commission. Things might need some tweaks with trading. Last thing you want to do is rely on an analysts models that you aren’t paying enough for.

Ultimately you are trying to understand how flexible the model is, how well it is made, and if you like it. I find that it only takes a few models to figure out if I like an analysts style. I look at top pick, bottom pick, random pick. It isn’t about whether you think the analysts assumptions are right but is this a model you can find what you are looking for? Can you easily change the assumptions? For oil and gas or anything with external inputs I want to be able to adjust those from my own deck of assumptions. If you need a walkthrough talk to the associate that built or maintains the models.

You notice the terrible models quickly.

Some analysts provide partial models to the buyside others the full working model. So make sure you know what you are getting, just ask the analyst or associate (unless it’s a firm policy as it relates to commission $s, I’ve only known one boutique Energy shop that does that). If you find small mistakes in the model don’t be a prick about it. Just fix it for yourself and ask an associate about it later. We’ve all been there before.

Let’s say you find two analysts with models you like. The next item is how much work do you need to do? Are you wanting to move their models into a tear sheet setup with model backups? Will you update their models quarterly (ie make it your own going forward)? Does you PM want models updated fast or can things wait for sell side to update? In which case how quickly does the analyst update their models?

Depending on how many seats you have with a firm, Bloomberg (and others) will take models and convert them - e.g. I’ve had a few analysts that have Capital IQ linked into their model. One was a shipping company with lots of formulas. Bloomberg took the model and updated everything to be compatible with Bloomberg API. I wouldn’t do this often but just in case.

Here is my process, and I’m without an associate right now so I do all this on my own.

  1. I have three analysts whose models I really like. With three firms. So these are my top commissions, for models, research, management access, etc. I like the models because they are clean and very easy to follow. A good model, for me, is one that I can open and have a quarterly report open beside it and move through the report and model in unison. In a pinch I could update them at the quarter.
  2. High priority companies. These are my large overweights and under weights. Some quarters I have analysts send me models (I say analysts because of analyst coverage); after earnings, halfway through quarter and before quarter results. I load each model into my own system and update tear sheets and review model. My goal is to internalize and update them (have a few done), but for now I get the new model make sure nothing major changed then I’ll work on some scenario analysis as needed.
  3. Lower priority companies I get to when I can. These are likely equities that I’m equal weight the benchmark for a variety of reasons. Models mainly used to check key assumptions and standardize to other models as needed (e.g. FX and oil prices). I don’t need to convert these low priority to my own.

Try to find a few analysts that model in a way that makes sense and are consistent with their approach across models. A good analyst uses a template for most companies.

Good luck.

 
Jul 20, 2020 - 8:57am

This is the hedge fund forum, so perhaps the operative time horizons here are materially different than my own as more of an LT long-only, but in my opinion you really should build your own models without any outside influences. Once you have your own independent take on the company, then go in and compare with sell side models. Doing it this way allows you to see where you differ from consensus, and isn't that the whole point, to try and formulate a differentiated view if one exists (no value in being different for the sake of being different)? Personally, the only number from sell side that I compare is top line growth, but even that is often incomparable given I purely model organic growth and sell side estimates often include some kind of incremental addition to the top line from M&A.

I could see where if your shop is more short-term how leaning on sell side models could be applicable and have value, but in most cases I would recommend building your model independently before comparing with external models.

 
Jul 20, 2020 - 9:51am

So out of curiosity, would this be your approach if you were given a coverage universe of 30-40 companies and needed to ramp up? How quickly do you think you could do that? I feel like that is the situation I hear a lot of the analysts at MMs fall into which is why they may be looking for this option. But to be clear I get what you're saying, building from scratch will lead to a better understanding of the business and greater chance of creating a differentiated view. Just doesn't seem like there's enough hours in the day to justify that time commitment at some shops.

 
Jul 20, 2020 - 10:58am

Ya I totally think it will be shop, PM, and time horizon dependent, but if you can do it I think it is well worth the effort and time. Additionally, once you get that initial model setup, maintenance over time becomes easier, it's just really that initial first time build out that is monotonous. Admittedly it probably takes me at least a full day just to get the historicals entered and make disclosure-based adjustments to get it the way I want it. In the scenario you laid out of 30-40 companies for a newer analyst with potentially a time constraint, I think the most value would be spent understanding the businesses themselves vs getting into the weeds of valuation, but as the analyst becomes more seasoned I would for sure recommend building out individual models independently.

Building off that point of becoming more seasoned, my personal coverage represents over 70 companies and I could realistically be expected to answer something about any of them to my PM; however, over time I have communicated a conviction list or a well-followed list of about 15 or so companies that I claim to have deeper and more up-to-date knowledge of. Now it will obviously depend on your PM as to whether they would be on board with you doing something like that, but this has worked well for me and resulted in a much better product when it comes to pitches and just on-going updates of my coverage. I would assume that most analysts cover businesses that are likely not worth owning absent a severe dislocation in pricing, and I would recommend at least trying this approach or asking if it is something your PM or team would be receptive to as a way to manage your workload and improve your research product.

Start Discussion

Total Avg Compensation

September 2020 Hedge Fund

  • Vice President (18) $520
  • Director/MD (10) $359
  • Portfolio Manager (7) $297
  • 3rd+ Year Associate (18) $269
  • 2nd Year Associate (25) $242
  • Engineer/Quant (47) $236
  • 1st Year Associate (58) $189
  • Analysts (175) $167
  • Intern/Summer Associate (13) $133
  • Junior Trader (5) $102
  • Intern/Summer Analyst (186) $80

Leaderboard See all

1
LonLonMilk's picture
LonLonMilk
98.3
2
Jamoldo's picture
Jamoldo
98.3
3
Secyh62's picture
Secyh62
98.2
4
CompBanker's picture
CompBanker
97.8
5
redever's picture
redever
97.7
6
Addinator's picture
Addinator
97.6
7
Edifice's picture
Edifice
97.6
8
frgna's picture
frgna
97.5
9
NuckFuts's picture
NuckFuts
97.5
10
bolo up's picture
bolo up
97.4