Over the past year, as headlines (if not reality) began to support the narrative of an improving economy, the Fed found itself in a position where they needed to either raise interest rates or change the narrative.  With over $18 trillion in debt, raising rates is not an option.

Crashing oil prices and a soaring dollar was the only solution [see: Those Wacky Central Bankers] that enabled them to keep both interest rates and inflation under control while preserving the yen carry trade.

But, there was a limit.  If oil prices got too low, the damage to banks who are heavily exposed to the industry would be too great.  And, the dollar was also getting to the point where even Yellen admitted it might be problematic.

The chart below shows the relationship between crude light (red) and the US dollar (purple) over the years.  It has historically shown a strong inverse correlation.

Yesterday’s fedspeak and market manipulation were designed to stabilize oil, the dollar and interest rates — all in one.  And, things were going pretty much according to plan — until the algos got a little out of control.  Actually, make that a lot out of control.

In the process of levitating stocks (to illustrate how wise and beneficial the Fed’s actions were) they forgot to put a lid on oil. Having already bottomed out as we expected two days ago [see: Update on Oil Mar 17] CL popped up through the trend line from 1998, threatening to undo all that nifty price control that got it down to an acceptable level in the first place.

This presented a quandary to the guys pulling the levers and pushing the buttons: how to kill off oil’s rally without also killing stocks?  Because oil and the dollar are so strongly inversely correlated, they needed to quickly ramp up the dollar.

But, the dollar and stocks have been strongly positively correlated lately.  Oil’s occasional ramp jobs have been a big part of algo trading ever since CL first stabilized in January.  So, ramping the dollar would add fuel to stocks and, hence, also oil.

The only solution was to use the yen to bring SPX and, thus, CL back under control.  As we wrote at 3:20 in the members section:

Gotta crash USDJPY (at least until after the close when it can all be reset.)  Maybe CL can be hammered back down then, too.

A cratering yen has been critical to the all-important carry trade which has fueled most of stocks’ rise since 2012.  The movement of USDJPY is the single most important contributor to stock price swings on not only a daily basis, but moment-to-moment as well.  We document its very high positive correlation with SPX on these pages almost every day.

Since mid-January, USDJPY has been advancing within a broad, white channel.  Together with CL (in purple below), is has driven SPX to all-time highs.

In mid-February, when stocks were faltering, USDJPY began constructing a new, steeper channel within the white channel (in gray, above.)  In fact, just two days ago (Mar 17), USDJPY broke out of a perfect little falling red channel above in order to arrest stocks’ sharp decline that day.

By crushing USDJPY (stronger yen, weaker dollar) they could bring stocks and oil back down.  It would mean also crashing the dollar which, as we mentioned above, generally benefits oil.  But, if executed correctly, it would have the necessary effect.

That gray channel had to go — but, preferably in such a way that stocks’ rally wouldn’t suffer any irreparable damage.  The key was timing.

USDJPY had dropped through the gray channel bottom the night before, and the white channel midline shortly after the session opened.  It bounced back above it as Yellen started her double-talk; but, when CL spiked through its overhead resistance, USDJPY was decisively crushed.

It had an immediate effect on stocks, which had just backtested the broken purple channel as we had expected.  We posted the chart below in Tuesday morning’s members section, along with the commentary:

Assuming SPX can get past the SMA20, the midline also intersects with the broken purple channel bottom for a backtest Thursday morning at the red .786 at 2102.49 (or next Tuesday the 24th at the .886 at 2110.48.)

The algos ramped SPX up to 2102 (where we recommended taking profits) and then some.  It had reached 2106.85 before the USDJPY crash finally caught its attention.  It backed off to 2096 by the closing bell, and was threatening more — especially, given the carnage going on with USDJPY, which had just dropped below a critical Fib line.

The white arrow on the USDJPY chart below marks the final 5 minutes of the session, when the pair dropped below 120.11 — the critical .618 Fib retracement of the pair’s decline from its 1998 high of 147 to its low of 75 in 2012.

This was a particularly bad omen for stocks, as previous dips below this line have led to large declines in SPX if they occur during market hours (after-hours, it’s easy to prop up ES in the much lighter volume.)

DX traders took note of the reversal and the crash in USDJPY.  If stocks were going to crater, they certainly didn’t intend to be left holding the bag.

The Fed was able to prop it up until the cash market closed at 4:00 (the white arrow below.  By then, it had fallen back to the purple channel top it had broken out of on Mar 10 (when USDJPY ramped higher to arrest SPX’s plunge.)

After the cash market closed, however, all bets were off.

Stunned by USDJPY’s continuing plunge (it didn’t bottom out until 4:04) and the likelihood that stocks would follow suit, the dollar flash crashed.  It plunged through our initial target at 98.82 and even through our primary target at 96.17 [see: Update on the Dollar Mar 13].

Finally, when it had reached the February lows, the Fed pulled out all the stops to support it.  It bottomed at 4:06, two minutes after USDJPY.  That it miraculously recovered overnight and is now pushing above the purple channel top again (in order to arrest this morning’s sell-off) is no coincidence.

That the USDJPY found support at the white channel bottom is not a coincidence. That EURUSD briefly popped above the midline of a 12-year old channel is not a coincidence.  That CL is suddenly back below the long-term trend line from 1998 (yes, hammered back down overnight) is most definitely not a coincidence.

In fact, there is very little in the way of coincidences in the “market” anymore.  Most of the time, the day-to-day moves are carefully scripted by central banks and the algos and HFT traders that capitalize on their continuing largesse.

The folks writing the script may tell us to pay no attention to “those algos behind the curtain.”  But, as events such as Black Monday, the LTCM debacle, the 2010 Flash Crash and the 2011 US credit downgrade taught us, they are a dangerous and destructive force that is always present — just waiting for the right conditions.

So, every once in a while — particularly when narratives need changing or correlations need resetting — things will get out of control.  As human beings increasingly yield the trading floor to machines, we believe the resulting lower volume and lack of rationale oversight will result in more such events.

Good luck to all.