The Medallion fund has been restricted mainly to RenTech employees since 2005 as the firm took steps to keep its size around $10 billion. The fund has historically averaged annualized returns approaching 80 percent before fees, but such gains can slump when it gets too big. Even employees face annual investment limits, and Medallion also typically distributes its profits every six months instead of reinvesting the gains.
How is that even possible? even warren buffett only averaged around 15-20 a year
Return on capital isn't a super meaningful metric for capacity constrained trades. If a fund earns 80% returns, but has no means to compound the resulting profits through the same mechanism, whoever receives them naturally puts them into something with worse returns, so their wealth still grows slowly over time.
I think Medallion is somewhere between HFT and stat arb, probably a mix of multiple strategies along those time frames. The faster you trade independent opportunities, the more you recycle capital, and prime brokers extend tons of leverage. When I worked in HFT, our profits were bound by other factors way before cash, and my desk's ROC was far higher than this when doing well (even when doing poorly, ROC was quite high. It was the expenses of finding those returns that killed us).
If all the money is employees', Medallion is basically just a prop firm. The employees paid out of the fund are essentially partners/owners, and the rest earn discretionary payouts of management/performance fees.
Even within the fund structure, most profit is return on labor, not capital. I'm sure if you compare margins paid to partners in professional services firms like law or consulting vs. typical publicly traded companies, they're also far higher, but Cravath, McKinsey, etc. won't let you buy in as a passive investor; you have to work for it.
ETA: If you're wondering how it's possible to earn 80% returns, or even more: there are myriad tiny inefficiencies you can trade on given the right research and infrastructure. I'm sure 80% is simply the point at which Medallion makes the optimal $ per year relative to risk. They could probably throttle back, make less $ on smaller capital, but far higher percentage return, if they wanted.
You might notice that the prices of trading A for B, then B for C, then C for A results in more A than you began with if done instantaneously.
But you can’t do that for a trillion dollars, because the “price” of an asset is just the best offer to buy or sell, and those are capped at the quantity of the best offer. Once you exhaust those, the arbitrage might not exist anymore because the remaining best offers will be worse. Say the market has capacity to actively exploit arbitrage for $10,000. You will earn the same amount of dollars if you have $1,000,000 or $10,000 if you exploit it to the max, but your return will be 100x better if your base is $10,000.
i've been thinking about your comment for about 12 hours now (over night after reading it last night). i don't quite understand the force of it.
>If a fund earns 80% returns, but has no means to compound the resulting profits through the same mechanism, whoever receives them naturally puts them into something with worse returns, so their wealth still grows slowly over time.
i don't see why this matters at that AUM. treat it as a fixed annuity for the fund members and it's still fantastically successful (if you know of a savings account that i can safely withdraw 80% of 10b from every year for 20 years please let me know).
>Even within the fund structure, most profit is return on labor, not capital.
i think this is just a matter of "levels of abstraction". if i invest in a mutual fund that has portfolio managers and analysts are my returns thereby ROC or returns on labor? obviously according to GAAP they're returns on capital but inside the mutual fund there are people laboring away. if i in-house that mutual fund along with all of that research infrastructure why does it suddenly become return on labor? it's the same economic activity.
Most of what we know about Rentech's strategies comes from Senate hearings regarding the sketchy things they do. The common theme is that they find strategies with fairly low returns (~3%), and very low risk. They then use massive (and illegal) amounts of leverage, as high as 16x, to turn those low risk, low return strategies into outstanding returns. James Simons and Robert Mercer are among the largest donors ever to both political parties, so it isn't much of a surprise that these hearings haven't really gone anywhere.
Of course it isn't just that simple. Leveraging a strategy should increase it's downside risk along with it's returns. But somehow Rentech does it while never actually experiencing that downside risk. That's their mathematical genius. And the fact that they employ so many brilliant mathematical minds leads me to believe that part may be real.
Not me, The IRA thinks it's illegal. There are limits on how much leverage you are legally allowed to use, and Renaissance has made a name for itself by trying to bypass those laws in every tricky way possible.
> There are limits on how much leverage you are legally allowed to use
No, there aren’t. Reg T limits initial leverage for retail investors in equity securities to 2:1 and FINRA rules maintainence leverage to 4:1 [1]. There are no similar broad-market rules for institutions. (Retail investors, likewise, can leverage much more for FX and—more commonly—real estate.)
They weren't in violation of Reg T. They were in violation of Reg U and Reg X. You are right that there is no law saying you can't use more than x leverage, but there are absolutely regulations governing loans in margin accounts and using securities as collateral for leverage. And The government believes Rentech was in violation of them.
> there are absolutely regulations governing loans in margin accounts and using securities as collateral for leverage
No, there are not for institutional investors. Rentech fell into an interesting dispute (tigger with Deutsche Bank) with the IRS regarding long-term capital gains. If I want to lever my institution 10,000:1, and can find a lender who will lend against it, there is no law prohibiting me from doing so.
Disclaimer: I am not a lawyer. This is not legal advice.
But there are laws preventing a broker from lending you that money. The hearing cites the SEC net capital rules which Rentech and DB worked together to bypass. There are pages dedicated to how RenTech DB and Barclays knowingly and purposefully circumvented rules with some changing of terminology. Starts on page 79 [0].
I definitely should have been more clear with my original statements. Maybe something more along the lines of: at rentech's scale, using a margin account at a prime broker, you cannot leverage your firm 16x.
I understand where you’re coming from. Let me, too, be more clear.
The Senate report is crap. Yes, Reg T and FINRA rules limit the loans B-Ds can provide clients. But leverage, for Reg T’s purposes, is constrained to lending. I can buy a 3x leveraged ETF [1] as a retail trader without violating Reg T. (With options, I could easily increase that leverage without borrowing.) All of this is not only permitted, but common.
Reg T does not exist to protect investors. It exists to keep broker-dealers from going bust from dud margin loans. (And thereby prompting a systemic crisis.)
The leverage RenTech took is in the non-lending and non-systemic (legal) category. The exposure that most closely puts RenTech in the lending bucket is the exposure Deutsche Bank carried on its balance sheet for tax purposes. (This tax avoidance was the core of the scandal.)
Long story short, unless you’re a politician, it doesn’t make sense to talk about RenTech’s illegal leverage. Lots of other market participants, by regular practice, are levered far more.
If a senate report causes issues for you, I don't think it's exactly crap. But I understand your point :) It also keeps mentioning Reg T which I agree is not relevant here. I was wrong to say that RenTech used illegal amounts of leverage, thanks for explaining that. I'm still not convinced it's wrong to say that RenTech worked with BDs to bypass rules that would have stopped them from using the leverage they did in the manner they did.
Was the leverage in the non-lending category though? Isn't that like saying their gains were long-term? That's the whole point of this, no? They used some different terminology to reclassify loans and taxes, in order to use more leverage and pay less tax than they normally would have.
I guess another way to put it: would there have been a way for RenTech to hold the same portfolio, using the same leverage, with the same payout characteristics, that no government agency would have issues with? Maybe the answer is yes, but I doubt it.
> would there have been a way for RenTech to hold the same portfolio, using the same leverage, with the same payout characteristics, that no government agency would have issues with?
Yes, quite easily. In fact, highly-leveraged portfolios like the one RenTech held are an essential feature of market making, which was historically done using banks’ balance sheets. RenTech’s shenanigans were around tax. Everything else is commentary.
(On the Senate report, the whole thing isn’t crap. But that section is crap as in it’s written for political purposes and has limited bearing with respect to the law.)
the subject of this Senate hearing was widely written about wrt rentech back in 2014.
This is more about what rentech does after they've made all their money trading... To avoid capital gains taxes. Sure, it's just another way to make more money, but it's not their core strategy.
It was never about returns exceeding borrowing costs. Rentech is probably the greatest money manager of all time. They can make returns that exceed the borrow costs.
It was always about the amount of borrow available, and the tax paid on the gains. The (illegal) strategies they used allowed for much larger borrow amounts, as well as only paying long term capital gains, when in reality there were millions of trades.
The simple setup: Rentech pays the bank $1B for a call option. The call option is on a $10B pot of money. The bank then hires Rentech to invest the pot of money for them. After a year or two, Rentech then exercises it's call option, which gives it everything in the pot except the bank's $9B plus a fee. The banks loved this fee. And since this was a year long call option, all gains are taxed as long term. They avoided $6B in taxes doing this.
For a modern hedge fund, leverage doesn’t work that way... e.g. if I buy $1000 of AAPL (Apple) and (short) sell $1000 of MSFT (Microsoft), what is my “capital” and what is my “leverage”? My net capital use is $0 - I got $1000 selling MSFT and spent it again buying AAPL... in the end, it comes to a combination of margin required by the broker(s) - i.e. what they estimate is the black swan scenario where AAPL goes down, MSFT goes up, you start losing money on both trades and they can’t close the position fast enough - ultimately it depends on volatility but can be netted across many trades, and the overall volatility / risk of your strategy (i.e. how much you’re potentially willing & able to lose in a year).
When you short you need to post capital as part of the repo trade. You also need to borrow the shares which typically costs more than the interest you're earning on the posted capital.
So no, your net capital use is not $0 in your example
That depends on your borrowing costs. I doubt they'll be paying margin rates lesser mortals pay. ultimately they are screwing their lender because the lender is taking all the risk here without (seemingly) adequate compensation.
Their lenders likely understand the risks pretty well. Their lenders are going to be large banks that have entire departments dedicated to modeling risks like these, and making sure that the risks don't exceed certain limits on these accounts. Nobody's getting screwed, except arguably the government out of some taxes here.
>Their lenders likely understand the risks pretty well. Their lenders are going to be large banks that have entire departments dedicated to modeling risks like these, and making sure that the risks don't exceed certain limits on these accounts.
Three words: Mortgage Backed Securities. Large banks with entire risk modelling departments have made multi-billion-dollar errors in recent memory.
I have a complete lack of faith that any institution whose day-to-day experience completely differs from their catastrophic risk scenario is ever truly prepared.
The personal and organizational strain of maintaining constant vigilance against an invisible enemy is simply too high.
People get lazy, complacent, and think they're more prepared than they are. That's just human nature.
This is absolutely true. Due mainly to wishful thinking as well as some bizarrely bad statistical assumptions, the risk of something like 2008 was simply not forseen. There's absolutely no reason to think that other unforeseen crises couldn't arise in the future.
You're right. But there's very good reasons to think that they won't come from levered quant strategies like Renaissance's. They're extremely diversified and tend to be market neutral.
Yes, you can name one time that that happened. But do you understand the types of strategies RenTech is using? They're diversified across thousands of equities on the long and short side. They're market neutral. Taking on amounts of leverage that seem crazy to you is actually very reasonable in this setting, and it's very easy to show that this is reasonably safe.
Nobody outside of RenTech knows what trading strategies RenTech are using, because they're extremely secretive. There is a strong suspicion that they are generating very high returns by passing on concealed risks to other, less sophisticated players - in the case of RenTech, literally anyone else in the market counts as less sophisticated.
That's really not true. Everyone knows what they're doing. They're doing statistical arbitrage. There's many firms that do it, Two Sigma, Citadel, etc.. The only thing secretive about it is the exact details of each strategy. Renaissance, for instance, pioneered techniques like using satellite imagery to track sales. One of their early strategies noticed that the market tended to go up on days when the weather was nice in NYC. Things like this are what they're doing, but on a grand scale and with a lot of very very smart people working on them.
While it is possible that they're making their profits by shifting risk onto unsophisticated players, it's certainly not necessary for them to do that to make money. They have the smartest people in the world working for them.
It is definitely possible, though it becomes much more difficult at scale. Asymmetries in financial markets are pervasive, and it is technically quite difficult to methodically explore that space analytically. Going deep mathematically into modeling these things is beyond the ken of most funds. For the most part, they are biased to search for patterns in parts of the phase space where it is relatively easy analytically. Motivated individuals that want to search the technically more difficult parts of the phase space still find plenty of gold nuggets because so few people look there or the scale of the opportunity is too small for other funds. RenTech is just really good at systematically mining parts of the phase space that are mathematically out of reach of other funds.
While 80% is extreme, 15-20% is not and I know many people that can do that consistently. Anecdotally, one strategy I developed many years ago that was designed to be easily applied to retirement accounts (so that close friends and family could, and do, use it) has returned an average of 20% annually through boom and bust, outperforming the S&P500 every single year. It didn't require being super-clever, just observant and analytical.
I think we tend to overstate the difficulty of getting a good return to scare people away from pursuing naive and/or stupid strategies in a futile effort to find easy returns.
Particularly when these analytical advantages may compound over time rather than dissipate. For instance, suppose a firm was exceptionally good at pairing world-class mathematicians with a team of mathematically-exceptional programmers who could reconceptualize the math wizard's insights as financial time series models and steer them towards areas of applicability. Now, when the next world-class mathematician comes to the office on a recruiting trip, he recognizes a higher expected value & better colleagues, joins, and the engine gets more repetitions with which to improve itself.
Even smaller funds only average at best 10-15%. 80% is just nuts and to do so with no down years, minimal volatility. I think it's more than just market timing
Again I've seen the same just running a single HFT desk within a larger firm. The only time we ever lost money was from rare technology errors. Trading equities, even if one position spikes 5-10% bad on news, you will still make money, because it's just one little position out of the thousands of tickers you trade. Even guys making far fewer bets in asset classes like FX only ever lost on extreme dislocations like the Euro/Swiss unpeg.
If you make a large number of bets, even with just a tiny statistical edge, you will be consistently profitable. RenTech probably isn't profitable every day, but I bet over a year they make at least as many bets as someone like Virtu makes in a day, so it's not surprising that they never have a down year, provided they have the edge.
1: Now does this mean Virtu the business made a profit above cost every day? Probably not. But it does show that consistent trading profits are achievable.
I'm not sure RenTech and Virtu are comparable though (although I'll admit I don't work in the industry, so correct me if I'm wrong), but based on the article from Matt Levine on his Bloomberg column "Why Do High Frequency Traders Never Lose Money?" [1]:
> Imagine how suspicious it would be if, for instance, your local supermarket made money on every carton of milk that it sold. That just seems too good to be true, doesn't it? How can they know the price of milk before you do? Shouldn't they be losing money on half of their milk, and making it on the other half, so that things balance out? Doesn't the fact that they always make money suggest that they're ripping you off? [...] That is, Virtu (like Goldman) is selling a product, and that product is liquidity, and it charges for that product. High-frequency trading firms are in the business of acting as middlemen, providing a valuable service by letting buyers and sellers trade as soon as they want to, rather than waiting for fundamental sellers/buyers to come in on the other side of the market.
As I understand it, RenTech is taking an investment position which is why the returns are remarkable whereas virtu is (as Levine puts it), selling liquidity.
RenTech doesn't publish much so I'm merely conjecturing based on their volumes and other information that's in legal filings. There are a lot of ways to provide liquidity and fair pricing over diverse time horizons. It doesn't really matter how they do it. If your bets are independent and you have a statistically significant edge, you are basically guaranteed to make money with proper bankroll management.
From what Virtu's published, they make two-way markets, and once filled they scalp a tick, arbitrage in another product, or cross if the market becomes weak. Maybe RenTech does something like buy underpriced oil producers whose prices haven't moved up after bellwether stocks in the industry like XOM and CVX have, then sell once the spread between them converges. I'm sure both firms have loads of different tactics. The key thing is that they're mildly better than chance.
Virtu makes money because they take advantage of a structural setup in the market: they trade retail flow from the likes of Schwab and Robinhood.
This is vastly different than making money in the markets just by connecting to an execution venue and receiving market data.
In the former you have a clear edge over the market because you receive dumb uninformed flow. In the latter example you are left to fend for yourself and challenge the EMH purely through your math skills and insights.
I mean, it's kind of ridiculous to just make blanket statements like that. This is a hedge fund that has a long history of good performance. The whole point of a ponzi scheme is that you have to keep getting more money from outside investors to pay the original investors. Rentec does the exact opposite. They refuse outside money and keep their employees from investing too much in it.
With that out of the way, obviously no one except those who work for them knows what they actually do. What we do know for sure if that they are highly levered, use an entirely systematic/algorithmic investment process, and do a shit load of trading. So even if they don't earn a lot on each trade, they do so many that are just barely profitable and highly levered that it ends up being a lot of money.
It’s all relative. Renaissance is famous because they do it on (low) double digit billions. On the other hand, 80% at a proprietary trading firm - where you might be working with double digit millions - isn’t spectacular. I can tell you that several small outfits in Chicago and the Bay Area return well over 200% consistently. But they are severely capacity constrained and basically disburse profits to the partners and traders each year. Their capital is only in the eight figure range.
I don't think its inconceivable, but comparing it to a hedge fund makes it look pretty crazy. First note that they look more like a prop fund, in which typically returns are labor-constrained rather than cash-constrained. Annual returns there aren't really comparable to investing, as they're more performing services (like arbitrage) for the market than putting in money. Second, the fact that being able to invest is a perk of employment means that part of their pay structure is via Medallion returns. Income that other firms would have to subtract from returns because it is spent on payroll still counts for Rentech even though said returns are going to employees.
Why? You don't need to do a ton of constant trades. While my portfolio is fairly small, I make probably half a dozen year and sometimes less or more but in total it averages to 500-1000 percent a year
A good starting point..if wall Street is consistently shitting on a stock, it's probably something to check out and do some research on.
their gross return is probably just a couple percent but then they leverage the crap out of it, which they can easily do since the strategy is designed to have very low volatility.
I disagree. But more importantly, it doesnt matter.
We cant invest and dont have enough information to assess whether it is a ponzi scheme or not.
Its possible that this one fund has generated excessive outperformance with low volatility, but we should start by assuming it didnt, then prove otherwise
They reportedly don't reinvest returns, which means that the 80%/year is paid out. A ponzi scheme can't return more money than is put in; Medallion does so every 15 months.
How is that even possible? even warren buffett only averaged around 15-20 a year