Markets never
repeat themselves but they often rhyme. This rally feels like the same sonnet
we experienced in 1987. As in a sonnet, it is following a strict rhyme
scheme and specific structure.
In 1987 the rally
began gaining steam in the spring when it already seemed overbought and extended.
The rally had initially started in October 1986 at DOW 1400, but during the
spring of 1987 it began to accelerate. It not only didn’t correct, but
continued to gain momentum. Despite all the pundits saying it was about to
correct, it just kept going up. By early fall the bears had capitulated and
the public was scrambling to avoid missing further gains. They were quickly
rewarded as the market moved even higher. No bad news, overextended
fundamentals or technical warnings could stop the rise. The DOW was soon over
2700 for an approximate 93% rise.
Then suddenly in
October 1987, out of nowhere, the crash hit. It was stunning. The
market gave back 22.6% in one day. What was later called Black Monday left a
pale over the US
that was palpable. In a matter of days the market surrendered the entire
gains it had achieved over the previous year.
Before you write
me off too quickly as trying to draw too close a comparison, let me tell you
why it really feels the same. It isn’t just the rise or rhyme;
it’s the reason for both.
Since
March 2009 the current rally has moved from a 666 low on the S & P 500 to
a recent high of 1220, for an 83% increase in just over a year. The rises are
similar to 1987 but so are the critical elements of risk and how risk has
shifted to the innocent. Before I discuss how this risk has been shifted
through Dynamic Hedging, Capital Arbitrage and Regulatory Arbitrage, let me
first briefly talk about the realities of risk versus perceptions.
REALITIES OF RISK
USING STATISTICS
TO SHOW US HOW THE INNOCENT GET LULLED INTO GETTING 'MUGGED'!
Let's say there's
a statistically unlikely event that takes place 1% of the time. As an
example, suppose, just for argument's sake, that if you go for a walk in a
particular Chicago
neighborhood, statistically the police tell you, 1% of the time you will get
mugged in this neighborhood.
So, if you go out
for a walk one time, you have a 99% chance of not getting mugged and a 1%
chance of getting hurt. But suppose you go out for one walk every day for 10
days, the chance that you will get hit on one of those occasions rises. The
way it's calculated is by figuring the odds that the LIKELY event will obtain
at every single iteration and then subtracting that from 100%. The equation is:
D = 1-(1-P)^N
Where:
D=cumulative
percentage chance of disaster
P= Percentage Chance of disaster on each opportunity (iteration)
N=number of iterations
So, if you go out
for 10 walks, your chance of getting into trouble is:
1-(1-.01)^10 =
9.6%.
And if you go out
for a walk every day for, say, 90 days, your chance of getting hurt is
1-(1-.01)^90 = 59.5%.
The graph above
shows what the series looks like. (1)
And in this
scenario, if you go out for a walk every trading day of the year (about 252
times) the odds are about 92% that you will meet your demise.
But the funny
thing about our human nature is that if, say, you went out for 252 walks in
our Chicago neighborhood and came back 252 times, without having had any
violent encounters with our Chicago’s city folk, you would assume that
experience was teaching you that there was very little danger. Indeed it
might be, if you didn't already know the likelihood of getting mugged.
As human beings
we are especially primed to generalize from experience (that's science), but
most especially to generalize from our most recent experiences (which is less
reliable science--or anecdotal evidence). So, the more walks we go on without
getting mugged the less likely we FEEL it to be that we will ever get mugged,
irrespective of what statistics might tell us. As our risk increases,
statistically speaking, we feel safer and safer.
We are all
familiar with the expression "tempting fate." One has to wonder if
we might be somewhere fairly far along on the curve charted above, feeling
safer and safer carrying all these economic loads, but with an ever greater
and greater chance of developing the incidence of one or another severe,
"dislocating," and "unlikely" event. (1)
LAW OF
COMPOUNDING NUMBERS
20% Gains for 3
years then a 20% loss results in = 8.4% CAGR
20% Gains for 3
years then a 35% loss results in = 2.9% CAGR
See the illusion?
Our government is
doing us no favors with an artificial extend and pretend
strategy that makes us feel safer and which consequentially starts the public
spending and investing again. Based on risk, it is both premature and
dangerous to your financial health.
DYNAMIC HEDGING
PORTFOLIO
INSURANCE
When the
investigations were made by the government into the causes of the 1987 crash,
it was discovered that it was primarily because of the wide implementation of
what was then called Portfolio Insurance. It was the rage in the late
80’s as a way of removing risk from portfolios. At its core, Portfolio
Insurance involved trend following methodologies. Consequentially, the more
stocks moved up, the more your portfolio called for more buying. It was self
re-enforcing. It also worked in reverse and consequentially the sudden crash.
The investigations prompted the introduction of circuit breakers into
exchanges to limit downside moves in any given period of time.
SON-OF-PORTFOLIO
INSURANCE
Though Portfolio
lost its appeal after the 1987 crash, it was replaced by what many at the
time referred to as the son-of-portfolio insurance. It was called Dynamic
Hedging.
Dynamic
hedging is a technique that is widely used by derivatives dealers to hedge gamma or vega
exposures. Because it involves adjusting a hedge as the underlier moves—often several times a
day—it is "dynamic. Dynamic hedging is delta
hedging of a non-linear position with linear instruments
like spot positions, futures or forwards. The deltas of the non-linear position and linear
hedge position offset, yielding a zero delta overall. However, as the
underlier's value moves up or down, the delta of the non-linear position
changes while that of the linear hedge does not. The deltas no longer offset,
so the linear hedge has to be adjusted (increased or decreased) to restore
the delta hedge. This continual adjusting of the linear position to maintain
a delta hedge is called dynamic hedging. (2)
Dynamic Hedging
was a major contributor to the tech bubble run-up in the late 1990’s,
the 2002-2007 run-ups and the present rally. It is one of the reasons this
rally feels so similar and is being driven for similar reasons. But there is
more.
The risks are
even greater today because Dynamic Hedging has allowed other advancements to
be layered on top of it.
With the post
tech bubble crash in 2000 and the subsequent advent of the housing bubble
explosion from 2002 to 2007 we saw the emergence of Capital Arbitrage.
CAPITAL ARBITRAGE
I am defining
Capital Arbitrage here (as opposed to the slightly different Regulatory
Arbitrage) as the price difference in the cost of capital through the
reduction of risk via various methods including removing debt obligations
(risk) from the balance sheet. The price difference in capital costs is
reflected in a lower interest coupon or basis point spread.
The advancements
in securitization and financial engineering have allowed this to happen in a
dramatic fashion over the last decade. Consequentially, the ability to extend
credit prior to the financial crisis was almost
exponential in its growth - all of which was hedged through Dynamic Hedging
and through newer techniques such as Credit Default Swaps (CDS). Capital
arbitrage fostered yet another bubble until the reversal once again happened
and we had the expected explosive momentum to the downside.
The table below
is a simplified summary of a lot of the work outlined in recent Extend & Pretend series
articles and the Sultans of Swap series. It
illustrates that almost all forms of standard accounting practices &
procedures have been circumvented through modern Capital Arbitrage
techniques. Whether Corporate accounting with its cost /accrual structure,
Financial and bank accounting with its reserve and capital ratios or Government
accounting with its cash account accounting, it doesn’t matter, they
have all been systematically exploited.
There is only one
goal, obscure debt or financial obligations, commitments, guarantees or
contingent liabilities. This is to allow improvement or maintenance in the
cost of capital and thereby allow further increases and gearing (leverage).
REGULATORY
ARBITRAGE
Today we have
layered yet another layer of risk onto the already existing structure. It is
called Regulatory Arbitrage.
Regulatory
arbitrage is any transaction that has little or no
economic impact on a financial institution while either increasing its
capital or decreasing its required capital. Just as trading arbitrage
identifies and exploits inconsistencies in market prices, regulatory
arbitrage identifies and exploits inconsistencies in capital regulations.
Regulatory arbitrage undermines the effectiveness of capital regulations. It
is one of the primary motivators for regulators to continually improve
capital requirements. (3)
This new strategy
is again intended to remove risks but in this case you are transferring it to
a sovereign government in a number of fashions. Whether debt or contingent
liability obligations, the
strategy involves
the sovereign government assuming responsibility and being forced to create
the credit to further the arbitrage.
It is highly
sophisticated with many elements but like the previous stages, it will end
badly and likely violently. This time the probability is that it ends when
sovereign governments fail or are unable to attract investors at satisfactory
rates (i.e. Greece
now having to pay over 9% on 10 Year Treasuries versus an expected 3.5%).
Credit rating downgrades and forced increases in collateral calls will be the
catalyst. We are now seeing just the tip of this iceberg throughout the
southern European countries (PIIGS).
‘TOO BIG TO
FAIL or are they TOO BIG TO SAVE?’
We have
‘saved’ the following by the public assuming the liabilities
after all the profits were earned and distributed.
1.
Fannie
Mae / Freddie Mac Agencies 1.5 – 2+T
2.
AIG
180B
3.
GM
/ GMAC
45B
4.
TARP
– Banks
700B
5.
FDIC
– Regional Banks
??
===
~ $3 Trillion
CONCLUSION
When markets stop
functioning any algorithm breaks down. Trading algorithms are based on
certain fundamental assumptions that have proven invalid over long periods of
time. The false assumptions include:
1.
Continuous
market liquidity
2.
Continuity
of markets
3.
Counterparty
Risk
The exposure
these ‘discontinuities’ create is well respected but to my
knowledge it is still not able to be modeled effectively. The trick therefore
is to make as much money as possible before ‘time’ delivers the proverbial
‘fat tail’. In layman’s language, as former Citigroup CEO
Charles Prince so famously quipped– it is a game of musical chairs and
“you must get up and dance while the music is playing”. If you
don’t ‘dance’ your competitor will have the competitive
advantage to be able to use an improved stock price to take you out. It
forces fiduciary risk taking. As in a child’s game, it takes enforced
rules or the cheating begets cheating.
If the
Legislators and Regulators won’t address excessive fiduciary risk
taking – then the market will in a violent and unexpected fashion -
with the innocent as the casualties.
For the complete
research report go to: Extend & Pretend
Sign Up for the
next release in the Extend & Pretend series: Commentary
SOURCES:
(1) John
Hussman’s Analysis – (Unable to find link – noted it
down years ago – sorry John)
(2) Dynamic
Hedging – The Risk Glossary.com
(3) Regulatory Arbitrage – The Risk
Glossary.com
The last Extend
& Pretend article: EXTEND & PRETEND -
Uncle Sam, You Sly Devil!
Gordon T. Long
Tipping
Points
Mr. Long is a former senior group
executive with IBM & Motorola, a principle in a high tech public start-up
and founder of a private venture capital fund. He is presently involved in
private equity placements internationally along with proprietary trading
involving the development & application of Chaos Theory and Mandelbrot
Generator algorithms.
Gordon T Long is not a
registered advisor and does not give investment advice. His comments are an
expression of opinion only and should not be construed in any manner
whatsoever as recommendations to buy or sell a stock, option, future, bond,
commodity or any other financial instrument at any time. While he believes
his statements to be true, they always depend on the reliability of his own
credible sources. Of course, he recommends that you consult with a qualified
investment advisor, one licensed by appropriate regulatory agencies in your
legal jurisdiction, before making any investment decisions, and barring that,
you are encouraged to confirm the facts on your own before making important
investment commitments.
© Copyright 2010 Gordon T Long. The information herein was
obtained from sources which Mr. Long believes reliable, but he does not
guarantee its accuracy. None of the information, advertisements, website
links, or any opinions expressed constitutes a solicitation of the purchase
or sale of any securities or commodities. Please note that Mr. Long may
already have invested or may from time to time invest in securities that are
recommended or otherwise covered on this website. Mr. Long does not intend to
disclose the extent of any current holdings or future transactions with
respect to any particular security. You should consider this possibility
before investing in any security based upon statements and information
contained in any report, post, comment or recommendation you receive from him.
|