Sunday 27 May 2012

Can Wall Street strategists be trusted?

"What we learn only through the ears makes less impression upon our minds than what is presented to the trustworthy eye."
Horace
Courtesy of our friends at Rcube Global Macro Research, here is a guest post dealing with "Strategists' Capitulation" from the 17th of May.

Can Wall Street strategists be trusted? History clearly suggests that they bring negative value in terms of asset allocation forecasting.

Even a casual look at the graphs shows that strategists are massively wrong
at extremes:

They were the most underweight right before the start of large bull market moves (1997 and 2009), and most overweight in 2001.

This is confirmed by the following quintile analysis (although there haven’t been enough cycles in strategists’ stock allocation cycles to draw strong statistical conclusions):

Currently, strategists are recommending the smallest percentage to equities in the last 15 years, matching 1997 and more ironically March 2009.
What this probably means is that European sovereign issues are already well discounted. For markets to move lower from here, additional catalysts need to arise.

We turned bearish a bit early in February selling European equities, and booked profits a bit early as well. The catalyst that pushed us to first go short and then reverse the positioning is mostly related to the global credit channel, which has now substantially improved.

We feel that investors are now underestimating the fact that, as a consequence of massive liquidity injections throughout the world since Q3 last year, bank funding conditions have improved enough for commercial banks to normalize their lending behavior/conditions. It is true in most EM countries, but also in Europe. In North America, banks are now aggressively easing their lending standards.

In some ways, this is what the recent strength on European and global earnings revision ratios implies.

In the US, yields already seem to be pricing an economic meltdown.




US 10yr yields are trading 3 standard deviations from the 30+ year downtrend.

Relative deviation from 100 days moving averages between bonds and stocks has just hit the 20% historical threshold level.

All this to say that while we are not contrarian just for the sake of it, and believe there are still many large long-term threats out there, a combination of improved fundamentals and large underweights in risky assets sows the seeds for a substantial rebound over the next few months.

However, we'll add to our mildly long positions only after we see a catalyst that would reduce the tail risk in Europe.
"Trust dies but mistrust blossoms." - Sophocles

Stay tuned!

2 comments:

  1. “US 10yr yields are trading 3 standard deviations from the 30+ year downtrend.”

    Asset-price changes are not normally distributed, therefore the standard deviation is not a valid measure of anything. They actually follow a power-law distribution.

    If you’re using variance, standard deviation, mean, or any of the other techniques designed *only* for normally distributed data then you’re going to get incorrect results.

    Please, please stop promulgating the myth that changes in asset prices are normally distributed. We've known for a long time that they are not, thanks to Benoit Mandelbrot.

    "Mandelbrot's key point was that the normal and log normal distributions were best suited for series that exhibited mild and well-behaved randomness, whereas power-law distributions were more suited for series that exhibited large movements and what he termed wild randomness."

    http://books.google.ca/books?id=TJ0dnfed0_wC&pg=PA79&lpg=PA79&dq=benoit+mandelbrot+asset+prices+power+law+distribution&source=bl&ots=9XYYC7E57K&sig=HJ5x38wS3RMAGZIDo0EYBVoiv9o&hl=en&sa=X&ei=BaHET67yGIfD0QXesciqCg&ved=0CEAQ6AEwAA#v=onepage&q=benoit%20mandelbrot%20asset%20prices%20power%20law%20distribution&f=false

    ReplyDelete
  2. Hi Ross,

    Thanks for your comment.

    Point number one:
    -If you read the post carefully, you will see that it is a "guest post" from my friends at Rcube.

    Point number two:
    -I am not promulgating anything and I don't think I need to hear about Mandelbrot and his assertions.
    Because yes, I have also read Mandelbrot, so I am aware of power law-distribution and the many flaws relating to "normal distribution".

    After all, if I wasn't aware of the flaws why would have I used titles such as "Interval of Distrust", or "The Italian Peregrine Soliton"?
    Why would I make many references to rogue waves and non-linearity and the perfection of Schrodinger equations by Howard Peregrine?

    Wild randomness has been a regular feature in my credit conversations.

    Best,

    Martin

    ReplyDelete

 
View My Stats