Why don't HFTs just start ramping AGNC into earnings now?
Ive never noticed Y! had message boards before as I source my quotes on BBG and only occasionally use Y! for charts. Very cool to see this ticker in particular has so much interest...
Anyway, as a long time AGNC fan, observer and trader, I find this situation puzzling and wanted to see what the consensus is among watchers:
To begin, its clear that this stock has been subject to a very high amount of HFT manipulation in 2012. Aside from its flash-crashes (well documented on Nanex), the stock for a long time has shown all the hallmarks of HFT patterns- momentum ignition, channel stuffing of quotes and a T/A picture that looked perfectly computer generated. OK. I think most of us have open eyes and saw this coming- the really unexplained crash that caused an absurd discount to book. Its no different than what happened to a lot of other "broken" stocks in 2012- that were left to churn in a pool of dead-beta (look what happened to AAPL and CMG). AGNC was ramped way up on retail getting in without understanding the model and the subsequent drop was to be expected as a stampede ensued when retail got wind of the retarded 'the Fed is gonna drive AGNC out of business' meme. So the HFT's were able to shake out retail VERY quickly and had a lot of fun stomping the stock down on waterfall dips whenever it failed to hold a technical level.
Now here is the part I find really strange: I expected AGNC to keep its dividend, but for the week or so after the most recent EX date to mark the low of the stock (which it seems to have done) as some idiots would naturally take this as the 'last train out of the station'. At that point I took a large position in the expectation of a raise in book value to $33.50+ for next earnings. The snap-back to $31 seemed completely logical and foreseeable in my opinion but this current low-volume churn is something of a mystery to me.
Don't the HFT's want to make money? Then why aren't they ramping AGNC now in the 3 weeks before earnings? Nobody makes out when the stock is just hovering here on pitiful volume day after day. Why aren't we seeing any candle-stick momentum ignition moments? At least that way they could create some vol and get momo players back in. Theres no point in further efforts at crashing the stock to scare idiots into thinking the 'smart money' knows something they don't- whoever had held on through this roller coaster down to $28 certainly isnt going to sell now, all the weak hands have clearly been shaken out- the path of least resistance is up...
So why not ramp now (at least to the last BV of $32.49) and then crash it again after earnings- #$%$ is wrong with these HFT's- why just leave the stock to churn on its own just before earnings? What- are they too busy right now defending the $500 level for AAPL...?
Yup, seems like theyre FINALLY waking up, why they let this stock churn for such a long time is still a mystery, but now that APPL's $500 line is safe, maybe they can finally concentrate on the ramping over here! I cant wait to see the pop on earnings (right before they start to take it down again) as Kain wont fail to deliver. Provided management understands the core concept of buyback/float expansion depending on the spread and discount to BV, this is the ultimate "new normal" investment, its really a money printing machine courtesy of the Fed...
I see, in that case I apologize for misinterpreting your algo-related comments the wrong way. I did get the impression that you thought it was just on those flash crash days.
Also, just to reiterate- im not in any sense advocating TA over fundamentals in anyway. You are completely correct of course to note that fundamentals always win out *over time*. Ever happen to read "expected returns" by antti ilmanen- its one of my favs.
However, I reject the contention entirely that the two waterfall plunges in AGNC were driven by fundamentals. Ill certainly give you the point about the perception of the participants etc. But dont tell me the algos werent there slamming the stock as hard and as fast as they could in those days in order to trigger as many stops as possible. Stop hunting is 90% of their bread and butter, and boy did they hit A LOT of stops on the way down to $28!
I honestly think the idea that such a fast collapse was caused by fundamental uncertainties totally fails to cover the scale of that decline. Sure you could argue fundamental uncertainties caused the collapse but I could just as easily argue the brief rally after QE3 showed confidence in management being able to use the situation to its advantage. No need to reiterate the pros/cons of QE3 for AGNC but to defend the stocks plunge along fundamental lines, your argument would basically have to be: the market just changed its mind about the fundamentals all of a sudden (4 weeks after QE3 announcement) and at the exact time the stock *coincidentally* happened to also break the 50-day MA!
Regarding algos, your statement is totally accurate: "the only people they screw are those unlucky enough to time their transactions out of sync with the bot". Precisely! Just like all those AGNC shareholders who held the stock after it exhausted its bull trend. Look, on Oct. 5th the stock was at $34.90. It broke the 50-day the next Monday and then plunged about 10% in less than 10 days.
Dont give me that BS about time horizons. Time horizons are the ultimate excuse for lack of performance every putz on bloomberg comes on and rightfully notes that on a 3-year window stocks will do just fine...
In the case of AGNC, no shareholder should have sat there and lost 10% of their capital (20% by Nov) when the signals are so obvious that a techincal plunge is coming
How do you know a major plunge is coming? Simple, when a bond fund that has been rallying with bonds fairly closely over the past 6 months, keeps rallying after QE3 even after bonds start falling hard, is a HUGE warning signal. AGNC was rallying with bonds but now rallies with high beta equities- the equivalent of having its cake and eating it too.
If youre not selling at this point, the least you should be doing is setting your stops up at the 50-day MA.
Because algos feed on stop hunting. So you know right away that a stock like this, in a WAY overextended Bull trend will get CRUSHED by stop-hunting algos on the way down.
So yes, at the end of the day you are right: one cant predict algos through geometric patterns (thats about as effective as Elliott waves!) but you CAN predict that when a "winning" stock like AGNC becomes overextended and loses momentum in a bull run, it will fall HARD as the algos pick off stops so fast that it induces a viscious cycle of panic selling fueling more stop hunting.
This isnt exclusive to AGNC, in fact it would be much faster listing stocks this hasnt happened to!
All I am saying is that by following these simple and obvious indicators- someone who is long AGNC for fundamental reasons, can get out strategically and get back in 3 months later with 20% more stock.
And how do you get the signal to get back in? Simple, as I mentioned the correlation to dead-beta stocks broke down quite clearly (see the apple chart) and flashed a clear buy signal shortly after the ex-date. I feel sorry for all those people that waited for a sub-$28 price or that thought the lows would come at the usual opex dates.
Anyway, that really is it from me- all the best to you in your future trading...