Seeing some interesting posts recently about pods blowing up, stress with drawdowns, etc.
For the pod world, why is this de-grossing / momentum unwind so much worse than other periods. For the uninformed... aren't portfolios supposed to factor neutral + beta neutral?
I recognize that this can only work to a point, and if you get a massive rotation, it doesn't always matter.
But does this also show that the majority aren't actually factor neutral? Hear people poke at this but I would be interested in getting a more detailed explanation here on 1) what a lot of pods do / were doing on factor exposure, and 2) why the supposed hedging doesn't work well in this environment. Is it essentially just crowding?
Just feels surprising that the risk models can allow pods to be max momentum factor. Or is it that you have at least 25% long momentum and hedge as best you can, and then with the rotation and unwind you still get hammered since there was so much crowding and your longs are down way more than any shorts.
Trending Content - Hedge Fund Forum
Career Resources
It’s the latter scenario you mentioned. Let’s running $1B GMV, max 5% of risk to any one factor (maybe that’s a lot, maybe that’s a little). A 1 std dev move might result in a $2M swing on that one factor. If enough factors’ correlations increase then you could be losing $2M multiple times over for what otherwise is expected to be a small move
thanks - but certainly the risk geeks think of this or have been concerned about this right?
Everyone has been talking about crowding and the persistence of the momentum factor for a while. Feel like this should've been on their radar as a highly likely outcome to some type of shock / correction.
But maybe the simple answer is less exciting, and it wasn't just PMs being closet long momentum above all else and doing their best to dupe the risk system
risk doesnt care when youre making money
they have been, but daily factor slippage throws your hedges out of whack in high vol regimes. How often should you readjust your hedges if your expected variance hasn’t changed dramatically?
To be clear - you think portfolio variance is $100, and you lose $400 on factors unrelated to momentum because they all move simultaneously. What should you do in this scenario?
Yeah risk has been flagging this for months, but pods think they're gods when they're riding the wave. FOMO is real.
cause people are bad and disguise beta and momo as alpha
had a great start to the year and so do most of the people i respect/who actually hedge and manage positions - or at least weren't getting killed for being max long RDDT and coinbade and calling it alpha
The issue is that risk factor exposure varies nonlinearly with stock returns. During normal times all the risk models show low factor exposure, but when there is a big drop all factors becomes correlated and the factor exposure spikes up exactly when you don't want it to. The portfolio is usually built to have a max level of linear exposure to risk factors but there is no constraint on higher order exposure, which is impossible to measure or zero out.
Thanks this makes sense
yep same notes as the 07 quant quake imo if anyone wants further granularity on general dynamics that rhymes with the current song. https://web.mit.edu/Alo/www/Papers/august07.pdf
This is the correct answer. Risk systems miss cross-gamma effects that manifest in stress scenarios (>3/4z). They often have cascading effects as cross-gamma effects create feedback loops. A momentum unwind creates volatility, which triggers volatility-based position limits, forcing further selling of momentum names, creating more volatility which the model doesn't capture linearly - especially with the level factor rebalancing that seems to be the norm in most models these days.
It’s a combination of momentum, crowding, and growth factors all collapsing at once. And depending on how a given book was constructed, the magnitude of damage varied.
Let’s say 30% of your book is crowded (net) and that factor is down 5%, that alone is a 150bps draw. And if the said book was also long momentum (likely for an average book) or growth (not necessarily), you easily tag on another 100-200bps draw on top. And would mathematically explain why some teams are down 200-400bps in the last month without any idio risk going against them. And 200-400bps is a lot in MM world.
On top of the above leans, if you also had a few idio calls go against you during earnings season (or a thematic call, maybe you were long cyclicals), that’s a good picture of a pod blowing up right now.
Also every year for the past decade the MM platforms have grown considerably so you get a bunch more PMs all doing a similar thing then you’ll get more issues on the downside by a # factor (more people will just equal more issues surfacing) and a magnitude factor (you get a bunch more dudes running the same surface level strategy whipsaws the downside when they all de-gross when Trump does what he says he’s gonna do).
Last week or two has definitely been very volatile so it’ll shake out any PMs on the edge or rookies in over their head
It really just comes down to crowding. Think about it. Large MMs account for a huge amount of volume, degrossing creates a snowball effect in popular longs/shorts by sector. If every HC pod is long LLY short MRNA and a couple blow up, the moves get worse and worse the longer that degrossing lasts. It’s realistically been ~2 months of this, which is catastrophic. It’s difficult to measure the crowding factor and hence MMs with the same risk model get stuck holding the same bags, have to degross, and that only gets worse with time
Taking risk is actually the beginning of success, you have to take a step in order for you to reach your destination
Agree crowding.
I think it’s easy to miss that this is like being stuck in traffic. You think “god, these cars in front of me are causing such a traffic jam”, while the cars behind you see that YOU are the traffic jam.
If the reason the market went up was because there are positive net marginal dollars, and these are levered (hedge fund dollars etc), then when the market goes down, this marginal, levered dollar needs to be repaid.
Doesn’t matter what type of risk analytics you want to whack on it, you can’t really hide because the returns you were generating in the first place were because of the net levered inflows. The only thing that makes a difference is ability to withstand margin calls (longer term funding), and wait out the gyration.
or put another way, the market is down BECAUSE the pods are blowing up (they are the traffic)
Perspiciatis est nostrum laborum quo nihil. Sed nemo illo earum. Vel dolorum dolorem culpa laudantium enim nam. Esse iusto et ut alias.
Laudantium qui rem non doloremque ad consequatur culpa. Explicabo maiores earum aspernatur sit ullam soluta et. Ut cupiditate eum iste quam vel.
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...