Tuesday, November 23, 2010

Stocks Tumble as the Credit Markets Rumble

Stocks sold of hard as debt concerns continue to haunt the market.
Investors initially ran to the safety of bonds until the 5 year Treasury auction was announced today:

"The bond market rally eased in the afternoon session as lower yields tempered demand on a $35 billion sale in five-year notes. Dealers pushed up bond yields from session lows to set up for $29 billion seven-year notes sale due Wednesday afternoon.

The Treasury Department sold the five-year notes at a yield of 1.411%, compared to 1.396% right before the auction. The higher yield signals weaker bidding prices. The auction was 2.65 times oversubscribed, compared with the average of 2.92 for the previous four sales."

My Take:

The bond market didn't like this 5 year auction at all and the 10 year sold off as a result:


The selling in treasuries could have been much worse if stocks hadn't performed so badly today.  At this point, investors really don't know where to run  hide. 

Bonds don't look attractive right now because there are fears that the debt contagion could eventually hit our bond market.  After all, the reality is we are in no better fiscal shape than Greece or Ireland. 

However, risk is relative so the money still flocks to the US when it looks like the crap is about to hit the fan because it appears that we will be the last country that goes down for the count. 

The dollar soared today as the European financial system continues to hang on by a thread:


The US is basically the least worst choice of all investing options at this point.  The European disaster has created a nice positive for the Fed:  It has taken a lot of the QE pressure off as the dollar strengthens.

How scary is it that our dollar is rising sharply only a few weeks after the Fed announced that they are money printing? 

I guess a little money printing doesn't look so bad when you live on a continent with 4 countries that appear to be on the brink of default.

Investors also flocked to gold today as stocks and bonds continue to wobble.

The Bottom Line

I don't know where this all leads to folks.  This all feels surreal to me at times.  It's like a bad dream that you never wake up from.

The Fed minutes came out today predicting that the unemployment rate will fall to 8% with 4% GDP growth by 2012. 

Yeah OK says who?

They must smoke some seriously awesome crack when they sit in these meetings.  I would love to know which Fed committee's members ass they pulled these numbers from.

Aren't these the same clowns that told us "subprime is contained" and used 10% as the maximum possible unemployment rate when they conducted the banking stress tests?

Excuse me for being a cynic but I call bull on these numbers.  The market did too as stocks sold off a bit on the release of the news.

When it comes to investing in this market all I can say is stay diversified.  There are no good options at this point.  Cash could crash as fast as gold could depending on what monetary policy path we take to get out of this mess.

The Fed's zero interest rates policy have put almost every asset class at risk because bubbles have been blown everywhere...Stocks, bonds, gold anyone?

IMO, Inflation still appears to me to be the end result if rates stay this low, but we could see a bout of nasty deflation before we get there.

I continue to sit in cash, gold, and some short hedges.  I continue to drool over the idea of shorting treasuries, but the European mess could create a short term move higher in bonds so I will sit on my hands before making this trade.

The price action this week has been amazing given how close we are to a major holiday.  Futures are down a tad again tonight so tomorrow could be interesting.

9 comments:

Herb said...

How can anyone put any stock in the Fed right now? Have they gotten anything right in the past 10 years?

Jeff said...

Herb

I agree!

I guess if you like bubbles that end in calamity they have been great for you.

flipdippy said...

Jeff - did you move out of bmore?

Glad you came around on the quant aspect of trading. I got an insiders view of how it actually works a few weeks ago, and let me tell you, anyone who is trading here is completely insane. You need at least a masters in math to be able to understand the tools being applied to markets.

Jeff said...

Flip

Yeah I did move for personal reasons.

I'll be back though. Just taking care of some things.

Yeah, the whole trading algo thing is a total sham. I can only imagine what you were able to learn.

These computers are several steps ahead of us and look at the markets in a totally different way.

Fundementals mean nothing now. These algos seem to be predatory and look for stocks that are vulnerable to short term price moves.

I just hope this insider trading investigation puts an end to this stuff because there will be no investors left that will want to invest if they let these games continue.

Looks like it's risk on this morning. DOW up almost 100.

flipdippy said...

Roger that - hope you're back soon for the local commentary.

What was interesting is how the trading aspect is about 100% human free, even at this point.

Engineers write software. They have massive computing resource access and data they then run simulation after simulation after simulation. They do 'QA' by simulating the market at various points in history to see how well the algo would have performed against major indices and benchmarks and well known traders.

If the algo beats the market, it goes into pre-production, basically running in a full scale simulation using current market data. If it continues to work there, it's pushed out onto their trading platforms, and basically just runs.

Even at the retail level, the advent of public clouds like force.com and amazon EC2 has enabled software engineers with some spare time to write their own software and backtest and do QA with the same amount of computing resources available to them.

The end of the day is like discovery. It may be "oh wow, that S&P options algo I wrote traded 371 times today and beat the S&P by 9.3% and I earned $5,319 for myself". Or it could be "hmm, S&P was up 1.32% and my algo only earned me .37%, time to go back and see why it didn't work".

We saw what happens when exception handling isn't robust enough to handle a sudden change - the flash crash was one. But like any software it merely exposed a flaw that was fixed in later releases. There will probably be another flash crash but who knows when and who knows what will cause it.

And you're right - because they can monitor so much information in parallel, there are many dimensions of weakness or strength they watch for triggers to hit and then kickoff other algos, even if they are competing against other algos that same trader or hedge fund or GS has running at the same time. There's your liquidity.

Technicals and fundamentals don't mean anything as we all know by now. The window in which a large majority of these programs trade is like at the quantum level (milliseconds to seconds), and it has completely blown up all the tools we humans have at our disposal.

The ultimate irony is in the end, bears and bulls will be screwed because software engineering is something of a herd mentality. It would not surprise me if we crack the 2007 market highs next year and then due to some flaw in all that code we also see the Denninger/Prechter lows. But as we see, it probably won't play out over years like it did during the depression, we will probably see both in a short time and then the market will revert to the mean.

flipdippy said...

Roger that - hope you're back soon for the local commentary.

What was interesting is how the trading aspect is about 100% human free, even at this point.

Engineers write software. They have massive computing resource access and data they then run simulation after simulation after simulation. They do 'QA' by simulating the market at various points in history to see how well the algo would have performed against major indices and benchmarks and well known traders.

If the algo beats the market, it goes into pre-production, basically running in a full scale simulation using current market data. If it continues to work there, it's pushed out onto their trading platforms, and basically just runs.

Even at the retail level, the advent of public clouds like force.com and amazon EC2 has enabled software engineers with some spare time to write their own software and backtest and do QA with the same amount of computing resources available to them.

The end of the day is like discovery. It may be "oh wow, that S&P options algo I wrote traded 371 times today and beat the S&P by 9.3% and I earned $5,319 for myself". Or it could be "hmm, S&P was up 1.32% and my algo only earned me .37%, time to go back and see why it didn't work".

We saw what happens when exception handling isn't robust enough to handle a sudden change - the flash crash was one. But like any software it merely exposed a flaw that was fixed in later releases. There will probably be another flash crash but who knows when and who knows what will cause it.

And you're right - because they can monitor so much information in parallel, there are many dimensions of weakness or strength they watch for triggers to hit and then kickoff other algos, even if they are competing against other algos that same trader or hedge fund or GS has running at the same time. There's your liquidity.

Technicals and fundamentals don't mean anything as we all know by now. The window in which a large majority of these programs trade is like at the quantum level (milliseconds to seconds), and it has completely blown up all the tools we humans have at our disposal.

The ultimate irony is in the end, bears and bulls will be screwed because software engineering is something of a herd mentality. It would not surprise me if we crack the 2007 market highs next year and then due to some flaw in all that code we also see the Denninger/Prechter lows. But as we see, it probably won't play out over years like it did during the depression, we will probably see both in a short time and then the market will revert to the mean.

flipdippy said...

Roger that - hope you're back soon for the local commentary.

What was interesting is how the trading aspect is 100% human free, even now.

Engineers write software. They have massive computing resource access and data they then endless simulations. They 'QA' by simulating the market at various points in history to see how well the algo performed against major indices and benchmarks.

If the algo beats the market, it goes into pre-production, running a full scale real time simulation. If it continues to work there, it's pushed onto their trading systems, and just cranks.

Even at the retail level, public clouds like force.com and amazon EC2 have enabled software engineers with spare time to write their own algos and do the same backtests/QA with the same amount of computing resources as the big boys.

The end of the trading session is discovery. It may be "oh wow, that S&P put options algo I wrote traded 73 times today and beat the S&P by 9.3% and I earned $5,319 for myself". Or it could be "hmm, S&P was up 1.32% and my algo only earned me .37%, back to the lab with you".

We saw what happens when exception handling isn't robust enough to handle a sudden change (flash crash). But like any code it only exposed a flaw that was fixed in later releases. There will likely be another flash crash but who knows when and who knows what will cause it.

The ultimate irony is in the end, bears and bulls will be screwed because software engineering is something of a herd mentality. It would not surprise me if we crack the 2007 market highs next year and then due to some flaw in all that code we also see the Denninger/Prechter lows. But as we see, it probably won't play out over years like it did during the depression, we will probably see both in a short time and then the market will revert to the mean.

ManHands said...

"We saw what happens when exception handling isn't robust enough to handle a sudden change - the flash crash was one. But like any software it merely exposed a flaw that was fixed in later releases. There will probably be another flash crash but who knows when and who knows what will cause it."

By that same reasoning, we should also look out for a flash "spike". Either way, the whole thing is such a sham. It would have been one thing if the flash crash stayed crashed, but it retraced 70-80% of its pre crash levels in the next 15 minutes. Just think of how much money was made by some (and lost by others) due to some guy having a screwy software program!!!

Jeff said...

Flip

WOW

I ws talking to my credit trader friend yesterday and he said the same thing. The human element is totally out of it at this point.

Thanks for sharing the details.

That is some scary sh*t if you ask me. There are always "black swans" in the markets from time to time and you have no clue how the algos would react to such an event.

A market where the the liquidity is created by little black boxes trading with one another is a market I want no part of.

In terms of scenarios who knows where these robots takes us.

I want no part of it because the 20% potential gain on the upside is not worth risking a potential 70% loss when the whole scam comes crashing down.

I feel bad for the little guy who is trying to invest during these times.

They seem to be paralyzed at this point interms of what they should do.

Crazy cra. Thanks for sharing!!!