与 Verdad Capital 一起投资生物科技

You're about to listen to the yet another value podcast with your hosts me Andrew Walker. Today I have the team from Verdad Capital on uh Dan and Greg. We are going to talk about their paper they have on kind of value and fundamental quant investing in biotech. Look, I read the paper and I instantly reached out to them. I told I found it fascinating. You're going to hear how excited I am. I'm like a kid in a candy shop asking all these questions. Obviously, I do like more microlevel securities and they're talking macro quantitative like putting everything in, but I I found the paper fascinating. I think you're going to find the conversation fascinating. I learned a lot. Uh, I'm going to include a link to the paper in the show notes, of course, so you should go read the paper, but you're going to really enjoy this conversation. We're going to get there in one second, but first, a word from our sponsors. This podcast is sponsored by TRDA. Look, I've been me, you've heard me talking about TRDA for months on this podcast. There's a reason. It is a really, really good fit for you. If you like this podcast, TRDA is a interviews between two bysiders who are talking about stocks they like. Sometimes you get a bear and a bear. Sometimes you get a bull and a bull. Sometimes you get a bull and bear. Whatever it is, it is two byiders who are interested enough in a stock that they've done research and they want to go on and talk to someone else about the stock and you kind of be get to be a fly on the wall and listen and learn. I'll tell you what, I am recording this in the middle of February. It has been the SAS apocalypse and TRDA has been so so good. So many different companies are covered and you know in real time you're seeing people talk about hey is the AI risk real here? Hey I talked to a CIO of a company who you know they were looking into this and they don't need this anymore. Hey I talked to a CIO who said there is no chance in hell that we will get off of this product. So I I just think TRDA if you have not tried it you should try it. The most frequent feedback I get from people who try it through this podcast they come to me and say hey I really like it. I wish there was more of it. I wish there were more coverage. I love it. So look, if you haven't tried it, you should go to trtrada.com. That's tir trda t r a ta.com and go check it out. All right. Hello and welcome to the yet another value podcast.

I’m your host, Andrew Walker. With me today, I’m excited to have Dan Rasmussen and Greg Obenshain from Verdad Capital. Guys, how’s it going?

Good. Thanks for having us on, Andrew.

Thank you guys so much for coming on. As I told you before, we’re here to talk about the biotech paper you released, “Investing in Biotech.” I’ll include a link to the paper in the show notes. As soon as I read it, I was like, “I’ve got to have these guys on.” I was just so titillated by it, but we’ll start talking there in 1 second.

Just before we get started, the same disclaimer I start every podcast with: nothing on this podcast is investing advice. That’s always true, and maybe particularly true today because we’re just talking about the paper—no stocks in particular. There’s a full disclaimer at the end of the podcast and in the show notes if you want to get there.

You published the biotech paper, “Investing in Biotech.” There’s lots of stuff I want to dive into, but I’ll toss it over to either of you. What is the biotech paper? What does it say, what are the conclusions, and why did you start researching this?

Yeah, I think biotech is fascinating, especially as a value investor, because it’s a huge percentage of the small-cap universe but totally ineligible to us because, in some sense, they’re all money-losing. So you’re like, “Should I just write this entire sector off? But it’s 25% of the Russell 2000, so what the hell do I do?” I think that was really our starting point.

I’m sure you guys know Joel Greenblatt’s The Little Book That Beats the Market. He has you rank things on EV/EBIT on one side and ROC on the other. By the way, you just exclude the entire biotech sector from the entire screen. And he’s not the only one—tons of quants do it. So, yes, I just love that insight.

Yeah. I think, as an investor, you’re always looking for interesting sources of return that are uncorrelated to things that you already own, and biotech is the least correlated sector. It’s really weird.

For us, I sort of said, “Look, this is a huge percentage of the small-cap universe. It’s the least correlated sector, and no one else is really doing it in a systematic way, so we need to go figure it out.” And so I said, “Greg, bonds are boring. Do you want to stop doing bonds for a 1-year project and figure out biotech?” Not stop doing bonds because Greg can’t stop doing his day job, but biotech stocks are just like bonds, I guess.

Very. There’s a lot of cash in both of them. Greg, did you want to add anything to that?

No. Biotech is strange because you can’t really use the financial statements. I will tell you that Dan asked me to do this, and at first I kind of ignored it because I thought there was no way you could build a quantitative strategy in biotech.

But as we started to look at it and think about how biotech works, we decided to meet biotech on its own terms and only use factors that actually made sense in the realm of biotech. We started to see results, and we started to see that it actually worked.

By the end, we really got some terrific results, and I think we’re firmly in the camp that a quantitative approach works in biotech. We’d be happy to dive into more detail on how we made it work.

Let’s do that. I want to start with the first—I think there are lots of things in here. I was really interested in a lot of it, but the one that jumped out to me, and I’m sure it’s going to jump out to people when they read it—and you basically lead with it—is that when you’re investing in biotech, you want to invest in things that sector specialists are heavily invested in, right?

If sector specialists own none of it, I think you say the returns from the things that sector specialists don’t own are basically zero over time. And if multiple sector specialists own it, then the returns are awesome, right?

I’d love to talk about how you guys started thinking about that, because I will just say, having talked to several sector specialists and looked at that, you guys found a quantitative insight into something that I think sector specialists kind of knew in their gut. They’d look at something, and when I talked to them, they’d be like, “Oh, no one owns that stock.” And they would kind of know it was because the science was freaky.

But how did you guys come up with that? And are there any other sectors that you can think of where the ownership itself serves as such a giant signal?

Yeah, I mean, what we learned about that was from talking to sector specialists. One of the nice things about quants is that, at its core, it’s a fundamental exercise. You go out and talk to people about how the industry works and how to think about it, and we talked to people who said, “Hey, we follow what other people own.”

Then we said, “Well, we can do that. We can go out and get that data, download that data, and test it.” Like everything, it was just a thesis we went out and tested. We were surprised at how well it worked.

One of the interesting things that comes out of it is that I think a lot of people say, “Well, I only follow the best specialists.” That helps a lot. There are very successful specialists out there, but what we found is that if you treat specialists as a voting machine and just say, “Consensus matters, right? We want to see more than 1 specialist in there. We want to see several specialists if we can,” that really helps.

Then you need to think about things like, well, if the company is small and it’s unlikely to be owned by a lot of large funds anyway, what do you do? Well, then you say, “Let’s not just look at the number of specialists that own it. Let’s look at the number of specialists relative to all funds that own it,” right?

What you actually care about is the vote of confidence from specialists relative to how many other people are invested in it. And what you find, which is really fun, is that if you look at companies that are owned by a lot of funds but zero specialists, they do terribly.

The specialists are really just a great guide to the industry, and it really ends up being a quality metric. It ends up being an initial screen saying, “Hey, look here. This is going to be a good place to start.”

Are there any other sectors? I want to dive into the specialists in general, but are there any other sectors—and I know you guys do, and I’ve liked your stuff on Japan, and I know you guys do stuff worldwide—sectors, industries, or markets where specialists or insiders have such a heavy weighting to alpha?

We’re looking into that.

Yeah, I think it seems promising, actually. But biotech would be the biggest outlier.

Makes sense. Let’s stick on the specialist conversation for a second. How do you guys define a specialist?

You can define them in a lot of different ways. And we might change how we define specialists because we always go and improve our process, but biotech is always evolving.

But for right now, it’s pretty simple. We just say: Is more than 50% of your portfolio in biotech? What that allows us to do is build a robust set of comps to look at going back in time, and they’ll automatically drop in and out of the universe.

Okay.

Yeah.

Oh, please go ahead.

Yeah. What we’ve seen is that the number of specialist funds has actually increased pretty steadily over time. Right now, in our data set, there are about 70 specialist funds that we’re looking at.

You guys don’t weight them by saying, “This specialist fund has done 1,000% over the past 10 years, while these guys have done negative 10% alpha over the past 10 years.” If they’re a specialist fund, they’re in there, and they show up in the signal.

Yep, correct.

One last question. There’s some level of quant, right? You have to make some sort of simplifying assumptions. Our simplifying assumption is that if they’ve been able to raise a certain amount of money to deploy only in biotech, they must have some biotech expertise—the people working there have to know something about biotech. I’ll come back to some other simplifying assumptions, but you’re saying that, for quant, that’s good enough, right?

We’re going to take 70 of them. If 5 of them are terrible, who cares? In aggregate, they’re specialists. Someone gave them money, and they must know something about the sector in order to be trading it.

It’s funny. Every now and then, I’ll have a guest whom someone didn’t like on the podcast, and they’ll say, “How could you let them on the podcast?” I’ll say, “Dude, it might not be your taste. It might not even be my taste, but these are professionals. People are paying them to invest. Who are you and I to say?” It sounds like you’re taking the same approach.

One more. I realize quant is, as you said, 70 big data points, but what about specialized active ETFs? I think the one that people might think of is the ARK Genomic Revolution ETF, which owns 10% of several biotechs. How would a specialized ETF in general—and I’m really thinking about ARKG in particular—be treated in your data set?

They’d be 1 data point, 1 vote.

That’s completely okay. I wasn’t trying to hate; I was just curious. All right, so you guys discovered that funds and companies in biotech that are highly owned by specialists, particularly when the specialists are concentrated, tend to outperform over time.

I’d love to ask whether there’s any insight into what’s driving that performance. Is it that these firms are better at passing Phase 1, Phase 2, and Phase 3 trials? Is it that, when they’re successful, the drugs have better results, so the stocks pop higher? I can’t say “efficacious,” so I’m not going to say it. Is it that they’re better at getting sold? Is there better cost discipline? I know we’ll talk about cost versus value in a second, but is there anything behind it, or could it be all of the above?

It turns out that I don’t have a direct answer for you, but I do know that those are the companies that tend to get acquired.

Yeah.

That is largely the measure of success in this industry, right? Do you get acquired?

What’s interesting for us is that, as quants, we’re not going to make all our returns from events. We’re not going to model events. If you actually look at the number of events that happen, and think about just having a big sample size, there aren’t enough. We’re not doing event studies.

The way we’re making returns from companies that are highly owned by specialists is that they behave better. Their returns are higher relative to their volatility, and they tend to have higher returns on average. They behave really well in a model when you’re diversifying risk and trying to target companies that are going to increase your return and reduce your volatility.

That’s why they’re good for us. We think about it on a shorter time horizon than 4 years, right? We’re constantly rebalancing our portfolio to take advantage of the relative valuations and risks among companies. The specialist metric is just 1 metric that helps us do that.

We don’t necessarily know exactly why the specialists do better on a week-to-week basis, but they do.

One more on this. I definitely understand, and we’re going to talk about all the other metrics. I just thought the specialist metric was so unique from what I’ve seen.

It occurs to me that you’re getting most of your data—not all, because I’m sure you guys are updating for 13Fs and 13Ds, 13Gs, and stuff between quarters—but most of your data is probably coming from the 13F, right? Is this strategy a lot different from everything else you’ve run, or how do you think about it?

It seems to me like we’re talking in February—is it February 19?—and the 13Fs just got released. Does this result in a lot of turnover? The 13F gets filed, and you’re like, “Oh, 4 specialists dropped out of this thing,” or, “3 specialists came into this one.” Is it a little bit different, or are there ways to smooth that over?

The thing about trading small-cap stocks is that, if you’re a $1 billion hedge fund and you’re putting $50 million into a biotech, you’re completely illiquid. You’re not getting in or out of that thing. You’re stuck.

Our view is that the 13Fs don’t change all that much. It’s not actually that fast-moving a signal. You’ll get some new buys or whatever, but they can’t move in or out very quickly. Because we’re looking at consensus—we’re looking at the things that a lot of people own—the chances that 10 different biotech funds all sell simultaneously in a small-cap stock are low. If that is happening, it’s going to tank the price.

You’re definitely right. RA Capital, RTW, and all these guys own 10% of these small caps. It’s got to be a success or failure. If they’re getting out, it’s probably because the pivotal trial failed, and they’re just like, “Wash my hands of it and be done with it.” There’s just not a lot of salvage value.

Anyway, there are some unique things about specialist funds, and I’m particularly thinking about PIPEs. Specialist funds can raise PIPEs, and they can do penny warrants. A lot of times, I know from trolling through these beneficial ownership filings, I’ll think, “Oh, there’s no specialist fund in here.” Then I’ll look and realize that 2 of the specialist funds own 10% or more of the company through penny warrants.

The other thing is, speaking of warrants, you’ll see a lot of times where a company will do a big raise and say, “Hey, we’re raising $100 million at $10 per share, and all the funds that are participating are getting warrants to buy stock at $15.” I could go on the market and buy the stock at $10 per share, but the specialist funds have definitely got a better deal than me because they’ve got the $15 warrants if things work out.

I’d love to ask how you’re adjusting for warrant ownership. I don’t think the second one matters for the way you guys are structured, but what about the penny-warrant ownership and those complications when it comes to ownership?

We’re working on it. One of the things we do is publish the paper, but we also have a research pipeline that we maintain. This is interesting about how we work in general: That was a question that came up the minute we published the paper. People will come back to us and say, “Hey, what are you doing about this issue?” We’ll put it in our pipeline and research it.

We’re well aware of it, and it’s something that we’re working on. The answer is that I don’t have an answer yet, but we will. Right now, we’re working on it.

Judging from the—people know that I’m really interested in busted biotech—but I tweeted out that I was having you guys on 2 hours ago, and judging from the amount of inbound I got, people are very interested in this paper. Maybe it’s just my circle, but I have 1 more question on sector specialists.

A lot of the big pharmaceutical companies will invest in smaller companies. Pfizer is 1 that I think about. Pfizer actually has to file a 13F. There are about 15 different companies where they partner on drugs and buy 5% to 10% of the equity. Do you guys consider them sector specialists, or is that a question for another day?

No, actually, they do show up on the sector specialist list. One of the things we do is screen to make sure they get screened out if they only own a few biotechs. One of the things we do is screen out companies that are very small—sector specialists that are small—because sometimes those companies will show up as small even though they’re not.

They do show up, and we capture that. We’re also looking into whether having strategic ownership is helpful, but you can imagine that the data is a little messier, harder to obtain, and harder to validate. As a first pass, it’s much easier to get the specialist data, and as we go on, that’s something we’re digging into.

Makes total sense. Pfizer likes to take some equity in them, whereas a lot of other companies might just want to do a straight partnership on the drugs. You guys lead with the famous Pharmasset example, which, if I remember correctly, J&J had JV’d on the key asset there.

They did not own equity, and then Gilead buys them for a fortune. I bet J&J wished they had owned equity when Gilead came over the top and got them, but you have to think about all that. I completely hear you. I want to turn to some other questions on really interesting things in the paper, but I don't want to leave the sector specialist in case there are any burning insights or questions I should ask, or anything on the sector specialist we should discuss.

Yeah, I think the only thing that's sort of fun to note is that when we think about launching this strategy or managing it, one of the things people say is, “Well, why wouldn't I go with the specialists? If the specialists are doing all the work, why would I go with this strategy rather than the specialists?”

I thought one of our thoughts in response to that is to say, “Well, actually, when the specialists have unique ideas, they don't do as well,” right? You'd rather own the thing that everyone agrees on. The consensus is the signal. You don't want us to have independent opinions. You don't want your biotech manager to have independent opinions. Their alpha opinion—the one where they're the only guy who thinks it is—is probably wrong.

It's funny—you know, famously, every fund-of-funds or best-ideas fund that launches fails, and you guys might have found the one sector and the one strategy where it really starts to work. I want to go to shorting. You guys start talking about this especially at the end. It's a long-short strategy, and obviously, for quants, long-short strategies are kind of the nirvana: the shorts underperform the longs. That's how you really capture the alpha and everything.

But biotech is an interesting one to think about on the shorting side because you have these huge catalysts that can make it difficult to rebalance. I want to ask a lot of questions on the shorting side, but how do you guys think about the short side when you're facing the upside risk of, “Hey, if we're wrong?” It's not unheard of for something with zero specialists to get acquired for a huge premium or announce great results out of nowhere and the stock go up 10–20x. How do you guys think about just those dynamics on the short side?

Yeah, so I think the first thing to note is that biotech is really a fertile place for shorting because it's the sector where the largest percentage of individual stocks end up losing money over time. Biotech beta is bad, right? That's the value-investor instinct: “Wait, these companies don't make a profit. They don't even make revenue in many cases.” And you're like, “Yeah, a lot of them should fail, right? Because they're science projects.”

Very few of them are going to work and be lottery-ticket positives, but the majority of them, maybe 60–70% of them, are going to be money losers for you. So I think shorting has to be an important component of a biotech strategy.

Then there's the question of this being a sector that can be highly promotional. It can trade on news, right? We had a positive clinical-trial outcome, and the stock is up 40–50%. Even if the stock eventually fails, that's a pretty painful day for you. So I think our next observation is just around risk management on the short side.

I think what you've seen is that a lot of biotech specialist funds have abandoned shorting. They run 105% long and 5% short, and the 5% short is in XBI. Why are they doing that? Well, they've gotten burned on the shorts, is the truth. They apply the same approach to shorting as they apply to the long side, where they do deep research and say, “Well, I have high conviction that X is a fraud or something, or B is never going to work.” They put a 5% concentrated short position on it, and then they just get annihilated.

Maybe they get annihilated for a month; maybe they end up being right 6 months later. But it's just painful, and they have so many scars from it that they just abandon it altogether. Our view is that this is a place where quant is just really, really good, right? You can look at things like how big the market cap is, what the liquidity is, what the current short interest is, and what the borrow cost is.

You can say, “Hey, gee, you know what? I'd rather be short 70% of the biotech stocks individually than be short 7 of the highest-conviction ones,” right? So it's about position sizing and diversification of risk—all the things that quants are really good at and fundamental managers tend to be really bad at because they get caught up in ideas and double down on things when they go against them.

Or you're just really disciplined in rebalancing frequently and being pretty diversified. By and large, if you're short things that specialists don't own, that are pretty expensive on our value metric, and maybe have a little negative momentum, you're going to do fine on the short side.

Greg, did you want to add anything there? I had some follow-ups, but I want to make sure I gave you a chance if you had anything. You're on mute, but I think Greg's saying no, so I'm going to assume that's a no.

So, I jumped straight to shorting, right? Using this on the long side, another signal that you guys do use—you mentioned short interest—is basically one of the signals. I think you guys have found that if biotech stocks are in the most-shorted percentile, the returns on them are hugely negative, and basically every other quartile is kind of positive, if I remember the chart correctly.

I'd love to ask, just because you mentioned it—I've talked to these guys, and a lot of them are really hesitant to put shorts on—if these companies are ticking up in terms of short interest, who is the short interest, and what signal do you think you're picking up on when these shorts are high? Is it, “Hey, there's a lot of fraud risk,” or, “The science is so bad even journalists can figure it out”? Because it seems strange that it could get that high.

Yeah. Well, I think you have to remember that markets are very efficient, and this short-interest signal works across every sector. The stuff with really high borrow cost and really high short interest just does terribly. I think you can think about that as there being some stuff that everyone sort of knows is bad or dumb or fraudulent, right?

It's like, “They're going to cure cancer and they're out of a strip mall in Miami,” and you're like, “I don't know, probably not.”

You joke, but if you trade cancer for Alzheimer's, I have seen that in the past year.

There you go. There are these things that exist and are sort of obvious to everybody, but the manager is very promotional and makes the stock pop every 3 months with some crazy news item that's totally fraudulent.

But by and large, that signal works across all sectors because markets are efficient. It also creates the challenge that the best things to short have the highest borrow cost, and so you're basically neutralizing the borrow cost versus your expected return. Often, they're sort of neutralized.

So you actually need a really good model to layer in, “What's my expected return versus the borrow cost?” Again, this is a great problem for quants. It might say, “Oh, gee, it's better to be short 50 things we're kind of negative about that have low borrow costs rather than 10 things that we're really negative about that have really high borrow costs, which might be subject to a short squeeze if the sector pops the shorts.”

I'm just curious: Is most of the return on the shorts generated from, “Hey, these guys come out and the drug fails and the stock goes down 90%,” and the shorts were kind of right that the science was either very poor or fraudulent? Or is most of the returns from the shorts, again, as we've talked about, biotechs burning cash, and it's just, “Hey, these guys are kind of the path to nowhere, and they're just always burning their cash balance,” and that type of thing? Is it more slow bleed or just fast drops on the short side?

Yeah, it's more the slow bleed. To think about shorting in general and what it does for a portfolio, it really dampens the volatility of the portfolio. It limits your losses when the market goes down so that you can reinvest profitably. In quant speak, it's taking care of the volatility drag, and so it's limiting losses to allow you to invest and get the upside.

We don't really think about it as shorting for profit. You can actually lose a little bit of money on your shorts over time—the entire time—and be far better off for having shorted. You don't have to make money on your shorts for it to be a huge value addition to the portfolio when you're running a really disciplined, diversified quantitative portfolio.

And past month to 6 weeks aside, I mean, if you were running a large short book on tech stocks and you were like, “Hey, I've lost 5% per year on tech stocks for the past 20 years,” you'd be like, “You are the greatest short seller of all time. Put you in all the portfolios. My God, could we lever you up against the QQQ?”

Yeah, just one last thought. I thought it was so interesting because I've just got this one stuck in my head. Again, I know, as somebody who looks at the individual ones, I probably focus too much on the individual ones, but there was this company, SAVA, and I've seen several other ones like that, right?

They came out with a surprise announcement that the drug works, and the stock went from $8 to $180 overnight, right? So that's a 25x-plus move, and I just keep looking at them and thinking, “Man, if you had asked me to guess, I would have said 99.9% this fails.” When you get a 25x on a success...

I'm like, man, that is just a tough place. Even if you've got all the quant signals, a 25x move on a short is—oh my God—even if you started small, that is life-altering. And it's not like it went up 25x over 10 years, so I could cover on the way. It's like, no, you get hit and it's gone. So I don't know anything about that. Should we go to the next factor we're going to talk about?

Highly diversified. [laughter]

Another factor that you guys mentioned is company insiders. I thought this was interesting from a lot of angles. One of them is, again, I look at these individually, and the alignment of incentives in a lot of them looks so poor because the company insiders get lots of stock options and they're incentivized: invest in the R&D, invest in the R&D. It doesn't matter if it's a terrible EV, because if it pays off, you get a fortune, and if it doesn't pay off, you get zeroed anyway.

You guys just mentioned that the returns here are fantastic for company insiders. So I'd love to talk about that finding and what you see there.

Yeah, this was really fun to look into, and it's hard—it's really hard—to look at company insider transactions, get that data clean, and make it work. The very first thing I'd say about it is that, actually, in biotech and across the industry, sales don't tell you very much, right? They don't tell you who's bearish because everybody sells their stock. That's a standard thing: you're issued stock, and you sell it.

What's really interesting is the buys, right? When you do a non-routine buy, you only buy for 1 reason: you buy because you think your stock is going to do well. Every now and then, it turns out that the CEO always buys whether or not the stock does well, so they're not particularly a good signal.

But the management team—the rest of the management team, especially the CFO—I mean, CFOs are pretty bearish people. So when they start buying, it's a pretty decent signal. When combined with the other signals, it's a pretty decent signal, which is the reason you want your bond guy working on biotech, by the way. It's the same logic.

So, you're actually highly discounting CEO buys. It's really focused on the CFO and the rest of the C-suite. And what about boards of directors?

They actually are mildly predictive, not as much as the C-suite. So, yeah, we count executives excluding the CEO.

Okay. The other thing I think is interesting about the insider buys is that these guys—I haven't been on the inside of a biotech, but I would imagine they have much longer blackout periods than your normal company. From what I have seen, a lot of times you see insider buying after clinical data news, and often the stock is up 200% to 300%.

I've seen CFOs buying, and you guys say in the paper that insider signaling has a lot of signal here. I always think it's worth more when it's, like, hey, it's very rare for us to be able to buy. A lot of times, in the story I told, they're buying when the stock is way up, and sometimes after negative data they buy when the stock is down. I just think it's interesting that they've got such limited windows and that it's actually sending a signal. I don't know if you guys had anything else there, but I find it fascinating.

Yeah, we spent a lot of time looking at how rare the signal was relative to history, and there are ways to do that. There are great papers on that. What I'd say about what we found in the insider-buying signal is that, much like the specialist metric, it was actually a relatively long signal. It had power for months after you could observe it.

Interesting.

This isn't a day-trading signal at all. This is a signal about the power of the company. If people were bullish on what they're doing, they might know a long time before anybody else. It might be for reasons that are independent of earnings.

You go and read—and this is actually sort of an aside—you go and read the academic literature, and it's so focused around whether the insider bought right before earnings were good. I don't think that's how people in a company think at all, right? This is their livelihood. This is their money. If they think the company's good, they're buying it because they think the company's good. They're not buying it because they think the next earnings are going to beat Street expectations, about which they have no idea.

I could think of a few management teams who might, but in general I would agree with you.

Interesting, but there's also a search factor, where you've got to try to find those people. But that's just too hard; it's too rare to play in biotech.

You do something interesting when you talk—I mean, you guys run a value shop, right? I'm a value investor. Most of the time, when you're talking about value stocks, you're talking about a company that trades cheaply to profits—gross profits, EBITDA, whatever you want to use.

You switch it here, right? As Dan said earlier, there are no profits here. There are often no revenues here. So how do you find value? You switch it from profits to spend. I want to ask how you guys came up with that and why you think spend is the right measure of value here. And could you define spend, if you don't mind? Yeah, Greg, you want to define it, and then I can talk through some—

Yeah, it's actually really simple. It's the gap between revenue and cash flow from operations. So it is all the cash out the door. Whether it's called R&D, whether it's called SG&A, whether it's called—I don't know—printer costs, it doesn't matter. It's just how much money you're spending.

We don't try to bucket it in this metric, because spending can—there are a lot of valuable ways to spend, and you spend if you think you have something. So I'll let Dan go on from here.

Yeah. No, but I think I was going to say that, again, Andrew, it's sort of the dumb insight, right? It's not saying, hey, we know which specialists are really good and we're going to copy their portfolios. It's not saying, hey, we really know exactly the right type of spending or how to evaluate that.

But we're saying, hey, if you spent $500 million doing some clinical trials, you must have produced something. Maybe it was a complete waste of money, but surely, in the abstract, on average, a company that spent $500 million on clinical trials is probably worth more than a company that spent $10 million on clinical trials. And if they have the exact same market cap, presumably the one that spent more is worth more.

So it's just saying: take the spending as—you know, who knows what the ROI is—but just assume that there's a constant ROI for all biotech spending. You should value the ones that have spent more more, even though on a traditional value metric, obviously, the companies that lost more money are less valuable and worse.

You're sort of flipping that on its head and saying, no, these are science projects, and what they spent on that matters. Also, the fact that somebody gave them the money to spend on that matters, because those people are going to want to get some return on their investment. Maybe even if that trial that they spent all that money on fails, the people behind it are going to try to figure out a way to repurpose the research or something to make themselves whole.

Yeah. And Andrew, sorry. I was going to say, you might be sitting here—and I think your listeners are probably thinking—well, that specialist metric sounds pretty good, but I don't know about this value metric; that sort of sounds really simple.

Well, guess what? When you run the data and figure out what's driving your return, the value metric works better than the specialist metric, because you need a way to measure how value is changing over time so you can react to it, and the value metric does that. So the value metric is actually one of the most powerful return metrics we have, despite the fact that it's an incredibly simple construction.

So if I'm thinking about it correctly, the value metric is just CFO. The more negative the CFO, the better, right? So it's basically assuming the R&D that's getting spent is getting spent on an at-worst-EV-neutral basis and, hopefully, an EV-positive basis. But that's kind of the way to think about it.

That's the denominator, right? That's one side of it.

And then the numerator is the market cap. Obviously, a $1 billion company that spent $2 billion is weighted lower than a $500 million company that spent $2 billion.

Yeah. So you think about that as an anchor of value. You need an anchor of value when you look at a company, and you need an anchor to compare to the market cap.

No, it's just interesting because, less so today than when you guys started writing and doing this—probably early last year, when the pharma sector was just imploding and everything was trading below cash—that's when I got interested. It sounds like we came at it from different angles, but in the same way.

You would hear people say, “Hey, that company's trading for $10, and they've got $20 of cash on their balance sheet.” And you'd hear people push back and say, “Yeah, but all that cash is going to R&D. They're going to burn it all.”

And you'd say, “Yes, but in the absence of more information, I kind of have to assume that it's getting spent, hopefully somewhat rationally. Are they spending so much that they're getting a negative 50% ROI on that investment? Because that's what it's calling for?”

But I'll pause there. I do have questions on it, but I'll pause there if you want to add anything to my story.

No, that's exactly right. I think part of that is—you got interested in the sector around the same time we did. We just love things that have been really bombed out and destroyed, because I think our general meta-view is that the more pessimistic people are about something, probably the more interesting the opportunity, right? You're looking for places where beliefs are correlated.

Rewind: when we first started doing this, as Greg said, one of the first things we did was find every specialist we knew and ask them a million dumb questions about how they invest, trying to discern what some of these signals are.

The amount of pessimism you heard from these folks was, “Why would you guys look at biotech? This is just a terrible sector.” And you're like, “Well, you've devoted your entire career to it.”

Then we'd be like, “Oh, maybe it'll be market-neutral,” and they'd be like, “Oh, thank God, right? Who would want long biotech beta?” At the time, the specialists were like oil in 2015 or 2016: people just got annihilated. That's always something that piques my interest in the sector.

So you guys go with, “The bigger the spend, the better”? I definitely get that you're adjusting for market cap, but that does seem like it would push you more toward an oncology trial or an Alzheimer's trial. Those are going to cost a lot more than a skin trial or an eczema trial or something.

Did you find that the signal works better in certain subsectors? Oncology has a much bigger spend and a much bigger target market versus some other areas. Or am I imagining too much there? Does it all come out in the wash in the quant data?

It kind of comes out in the wash. But when you talk about the research pipeline, that's the kind of thing that we know in other areas. Across sectors, when you're investing across sectors, using sector-level value can work. It can add something, but it's additive; it's not a replacement. So the answer is yes, we'll probably look into that at some point. We're constantly researching this stuff.

I'm interested because, when I first read the paper, my first thought was, “If they're going to spend, oh boy, they're going to be in a lot of Alzheimer's and oncology drugs.” Maybe what they found is that the past 10 years in oncology and Alzheimer's have been great, but going forward it's a very tough area.

And, rather, the model also diversifies. Remember that even if the model said, “It thinks Alzheimer's is the only thing,” one of your steps at the end, when you're building a risk model, is that it won't put 100% of the portfolio in Alzheimer's.

So even if your value metric is wrong—if it's way too high in Alzheimer's—when you actually create a diversified portfolio, you'll even things out. You'll take the best of Alzheimer's, and then you'll take the best of everything else.

Judging by the returns on Alzheimer's, I don't know if there's a best publicly traded Alzheimer's company.

Speaking of sectors, you mentioned it: you guys also have something really interesting on momentum, where you talk about going trial by trial and classifying firms by what the trial is, then looking at momentum cross-sectionally among companies running the same trials. I want to make sure I understand that piece correctly. What was the methodology? What are you doing? Am I thinking about it correctly?

Yeah, you are. This is the classification problem we're talking about: you need to classify these biotechs and group them. This value problem is exactly that, and we actually did test the value problem you're talking about; we just didn't put it in the model.

What you want to do is classify these things and group them together so you can look at how they move together from a risk perspective. Then you can say, “I want to build a metric within a certain class of companies.”

To do that in biotech, you can do it very simply: categorize them, call them something, and say, “This looks like that; this looks like this; this looks like that.”

But you get a time-series problem, because biotech can change over time. They go through phases, and there are different phases at different times. They might even change what they're targeting. Their lead trial might change; the one that's actually furthest along might change.

What you need to do to do that really well is go back and create a time series of descriptors of the company. The best way in biotech to create a descriptor of the company is to aggregate the clinical trials in which they're involved and aggregate them up to the company level.

You can say, “On average, they're doing these things.” Those clinical trials are great because they have so much data, so much detail, and so much classification data on what the companies are doing. The whole purpose is to build time-series classifications of companies so that we can categorize them and run metrics on them.

One quick thing: you mentioned the number. I do know of companies that will have 1 phase 3 drug, late-stage in Alzheimer's, and then 10 preclinical or phase 1 oncology drugs. The way you described it, it seems like that company would come out as an oncology company, but anyone who knew it would say, “Oh, no, that's an Alzheimer's company with a little bit of oncology—a call option sprinkled on top of it.” How do you adjust for that? Do you give bigger weightings to phase 3 trials versus phase 1? Can you adjust for the amount of spend on the trials? How do you think about that?

That's a really hard problem, and just as hard a problem for a fundamental analyst. This is why no one—this is why quant didn't want to attack biotech, right? All of these are really hard problems, but they're really interesting.

We discussed this in the paper, and there's a cool way that you can classify companies: you can say who they're similar to. Let's use this as a great example. You would say that this company is similar to a company with a phase 3 drug—I can't remember what your phase 3 drug was in—and it's similar to a lot of companies with phase 1 drugs.

Then you have to come up with some sort of algorithm and say how similar it is to each one of the companies in the universe. That number doesn't matter. It could be 1 through 376; we don't care. It's a scale.

You say, “I'm sort of 1-similar to that one. I'm 54-similar to that one. I'm 98-similar to the other one,” based on the phase, the indication, and where my headquarters are. I don't care.

Then you can take an average across that whole thing and say, “I'm most similar to that person.” But what you're not saying is, “I'm most similar to that person.” You're saying, “On average, I'm similar to these, and my average—the average company that looks like me—is this.”

Then I can go out and say, “How are those companies acting? How is my average acting? How's my index acting? What's my value relative to that index? What's my momentum relative to that index?” That's the way you can classify a company without really knowing which one is more important.

So we don't have that analyst sort of input, but we have a really sophisticated way to say, “Hey, we can look at who's similar.”

It strikes me that AI has to be both the most terrifying and the best thing for you guys. AI feels like it’s going to replace the humans who are doing quant, but at the same time, what you described is something you’d never do without AI running a lot of things in parallel with each other.

And just as an aside, I taught myself to code from books, and I’m just really upset that everybody knows how to code now.

I’m kidding. I’ve been thinking about that a lot. Let me stick with momentum. We can have the discussion on AI and vibe coding if you want to, but let me stick with momentum.

If I’m remembering the paper right, in looking at my notes, you guys find positive momentum inside categories. If I’m going to say it bluntly, if Alzheimer’s drugs are doing well, all the Alzheimer’s companies start doing well. That’s my understanding; you can tell me if I’m understanding that correctly, but I’d love to know why you think momentum works.

Because when I look at these companies, what I often see is, “I’m going after bladder cancer, and you’re going after bladder cancer. You announce good bladder cancer results, and my stock goes down because, yes, maybe there’s a good signal that our targets work. But if your drug gets approved, I’m not going to have a monopoly. At best, it’s going to be a duopoly. At worst, it has all sorts of negative implications.”

I was surprised that they all worked together. I’d love to ask why, and what you’re seeing in that data.

Yeah, I think this is interesting because it’s actually a sort of market-wide phenomenon where similarity momentum, or pure momentum, works. In that discrete case, if a clinical trial gets approved for a competitor, that’s bad for you. But the vast majority of the time, if your competitors are—if you think of a car company, if Toyota is going up and you’re Ford, you’re probably going up. These things are influenced by big macro trends.

The thing you were trying to get at is being as precise as possible: What is really the exposure that’s driving this? Ideally, you want to capture almost the thematic thinking. People are really excited about obesity drugs right now, or they’re really off obesity drugs, and so I really want to capture that in my investing.

As long as people like obesity drugs, I want to own obesity, and as long as they don’t like it, I don’t want to own it. Basically, you incorporate that sort of thematic judgment into your investing.

I think that’s actually quite true of the way people trade. People get excited about certain topics or themes, and the market trades things that way, whether it’s for regions, sectors, industries, or peer firms. Often there’s some external driver for why that’s true. Maybe mRNA is having a terrible time and not getting anything approved, and so all our mRNA drugs sell off, or whatever it may be.

No, I think you’re definitely right. I was probably too narrow of an example because you mentioned obesity drugs. I know when Pfizer gets into a bidding war for Metsera, every other obesity-drug player goes up because they say, “Whoever loses that bidding war, they’ve already shown they want to buy, so they’ll probably come buy it.”

When I use bladder cancer as an example, bladder cancer drugs work a lot of the time. All the bladder cancer drugs work because they say, “Hey, at least the target has been proven. At least the mechanism of action has been proven.” So, a lot of them work.

Yeah, Andrew. I was going to say one final thought on that. I think when I talk to people, you’ve asked so many questions that we’ve been asked before—and really good questions.

I was hoping for some unique questions, Greg.

No, no, no. But I talk to people in biotech, and they’re so event-focused. Biotech is events. There’s a lot of movement in biotech stocks that is not event-specific. In fact, most of the movements in biotech are not event-specific.

When you actually do the quant research, you find that you’re not trying to target the events or predict the events. You’re trying to understand the core underlying drivers, how they drive stock prices over time, and how they move together.

So, it’s a very nonfundamental way of looking at it. You build the factors fundamentally, but there’s a very non-fundamental way to look at it, and it’s extremely powerful.

Yeah. I think there’s a famous paper—the title is something like “What Drives the Value Factor?” or “Surprising News About the Value Factor,” or something like that.

The finding is that when an event happens, the event’s impact on a stock is variable, dependent on the valuation coming in. If you’re a really expensive stock, news tends to have a negative impact on your share price. When you’re a really cheap stock, news tends to have a positive impact because, essentially, people think that the world is much more predictable than it is.

They price in— they get very optimistic or very pessimistic at the extremes, and then the news cycle just brings things back to the mean, in some sense. It’s just random dispersal.

If you say, “Hey, some of these events are just random. I don’t know, 30% of trials fail or whatever. I mean, that’s not the right number, but I’m not going to be that good at knowing whether it’s 25% or 35%,” all I should know is that if people think it’s priced like there’s a 90% chance, that’s probably not so smart. If it’s a 10% chance, maybe we should be long it.

I think that’s sort of the idea of quant. It pushes you toward this base-rate-driven approach: Let’s assume the market is sort of normal. Let’s assume that spend has a return. Let’s assume that all these smart people who study biotech know something about what they’re doing, and let’s assume that events are going to unfold unpredictably for everybody.

What really matters is whether you’re positioned in a way so that the events end up playing in your favor.

Yeah, but there’s no space that proves that more than biotech. I know companies that have announced that a drug failed and the stock goes up 100%. You ask why, and they say, “Well, people knew this drug was terrible, and they were worried that they were going to get just enough to continue spending and burn all their cash on it.”

Or a company announces that the drug works, and the stock is down 80%. Why? It was priced like it was going to be the best-in-class, literal cure for cancer, and it came out looking like a me-too cancer drug. There’s no place like biotech for that.

I’ll leave you with my favorite line in the paper. I told Greg before that my favorite line was: “Can a quantitative approach in a sector with idiosyncratic successes that are not reflected in the financials until years after the value is known work?” That was my favorite line in the paper.

I think you guys have done a great job explaining the whole paper and why it probably does work on this podcast. Unless you guys want to say anything else or have any last thoughts, I’m happy to wrap it up here.

No, this is great. Well, thank you for having us on, Andrew. We’re really excited about this. It’s very cool. Biotech is such a weird sector. We did so much work creating new metrics and figuring out a new way to trade it, and we’re excited to see how it develops, to learn more about the sector, and to do more of this research.

I was just blown away by the paper. I thought it was so awesome and so creative to find so many unique signals. Again, some of the signals were things that, if you talked to sector specialists, they’d say, “This is what my gut says.” There were so many unique things, and I really enjoyed it. So, thank you guys.

Last question. Dan, I’ll put you on the spot. You mentioned that you love bombed-out sectors. That’s what Verdad’s focus is on. You had Japan. What’s the most bombed-out sector as you and I speak, February 18th?

Andrew, it’s a little bit frustrating for me right now because Howard Marks says you want to be contrarian and right—have a non-consensus view and be right. Right now, I think we’ve been sitting around and saying, “What has Verdad’s core argument been over the last few years?”

One, we’ve said private equity is in a bubble. Stay the fuck out of it, first and foremost.

KKR this morning would have something to say about that.

Yeah. Look at the front page of the newspapers now. People agree with us. I’m going to have to start—me saying that private equity sucks, and nobody yelling at me, “You’re crazy!” anymore because everybody agrees with me now. KKR bought $25 million of stock this morning, though, so you guys might have to throw them into the insider-purchase screen.

Maybe it’s still good for KKR, but I feel like that was one of our non-consensus bets that’s looking pretty smart. We’ve also been saying, “Hey, Japan’s a great place to invest,” and all of a sudden, over the last few years, Japan’s been probably the best-performing market out there. We’re having trouble refreshing our basket of new ideas.

For a value investor, you start to get worried when you have 2 or 3 good years in a row. You say, “Oh, gosh, maybe it’s overvalued now.” So, we’re refreshing. Maybe Greg and I will just start becoming huge bulls on private-credit BDCs or something. Yeah, they are trading for big discounts. So maybe that is the best.

Hey, can I ask one last question? I’m sorry to prolong it. On insider purchases, the KKR one brought it to mind, right? $25 million is a lot of money. I don’t know if it’s that much money to them. Do you guys adjust for the size of insider purchases, and maybe the size of the insider purchase versus compensation, when you do the adjustments?

In general, when we make metrics, we’ll do a count and a volume, right? So we’ll average them, or look at both of those. In the insider data, we mostly just relied on counts because we wanted the intention. I don’t want to credit somebody more because they just happened to have more money. So we just used counts in that analysis.

Again, this might be quant versus story, but with ServiceNow, the CEO says, “Hey, I’m canceling my 10b5-1 and I’m going to buy $3 million of stock over the next year through a 10b5-1.” You’re like, “Hey, that’s a nice signal, but you get paid $40 million per year and you’ve sold $100 million of stock over the past 3 years. Your stock is down 60%. Maybe you could add a zero to that.” That feels very much like, “Hey, guys, I’m here with you.” So, I don’t know.

Well, the question is, is anybody else following—

Right?

All the execs canceled, but yeah.

Yeah. And was that the CEO? Because the CEO was fine.

It was the CEO. All the other execs canceled their 10b5-1s, which I think is also interesting, but that’s just the one that came to mind where I was like, “Maybe we could add a zero to that or something.”

Guys, it was a great paper. There’s going to be a link in the show notes. I really appreciate you coming on, and hopefully we’ll have you guys again for the next interesting one.

Thanks, Andrew. This was a blast.

A quick disclaimer: nothing on this podcast should be considered investment advice. Guests or the hosts may have positions in any of the stocks mentioned during this podcast. Please do your own work and consult a financial adviser. Thanks.