The 4% rule applied

Transcription

The 4% rule applied
Risk management research
Ineichen Research & Management
(“IR&M”) is an independent research
firm focusing on investment themes
related to absolute returns and risk
management.
The 4% rule applied
Executive summary
26 September 2014

The 4% rule was developed by Ned Davis and popularised by
Martin Zweig in Winning on Wall Street in the 1980s. The basic
idea is that if an index moves by 4%, it is worth our attention.
Something meaningful happened and we should pay attention. We
herein apply this rule more broadly and test whether a variant of
this rule is usable and/or applicable to risk management.

Two aspects about investment management in general and risk
management in particular are meaningfulness of a change in
circumstance and an investor’s conviction in a certain risk exposure.
First, a lot of information every investor is bombarded with on a
daily basis is simple not that important, i.e., meaningless. Second,
position sizing and/or hedging are often a function of an investor’s
conviction, i.e., his confidence in his analysis and reasoning. Large
swings in a variable can help with both, determining meaning and
importance of the information as well as helping the investor to
articulate or quantify the conviction in a certain idea or risk
exposure.

We test the 4% rule on the S&P 500 and were surprised to find
that there is after all a bad backtest.

We apply the idea behind the 4% rule to economic variables. We
find that large changes can mark important turning points but there
are false signals too and not all turning points are highlighted by
large moves. Paying attention to high standard deviation changes
makes sense but the Holy Grail of finance it is not. Large changes
should be treated as an indication of meaningfulness rather than a
solid signal one can blindly adhere to.

When assessing economic variables the most important thing is the
trend, i.e., the economic momentum. The momentum itself is—to a
large extent—a fact, not someone’s opinion or forecast. The second
most important thing to look for is the reversal of the trend. Only
then, third, should we look at the extent of the magnitude, i.e.,
standard deviation of a particular move.
Alexander Ineichen CFA, CAIA, FRM
+41 41 511 2497
[email protected]
www.ineichen-rm.com
“I measure what's going on, and I
adapt to it. I try to get my ego out of
the way. The market is smarter than I
am so I bend.”
—Martin Zweig
The 4% rule applied
26 September 2014
Content
On a personal note .............................................................................................. 3
Introduction ..................................................................................................... 3
The 4% rule ...................................................................................................... 6
Applying the “4% rule” to economic variables .............................................. 11
Testing US PMI ........................................................................................... 11
Germany’s ZEW ......................................................................................... 18
Bottom line .................................................................................................... 19
References ......................................................................................................... 20
Publications ....................................................................................................... 21
Ineichen Research and Management
Page 2
The 4% rule applied
26 September 2014
On a personal note
“I can’t overemphasize the
importance of staying with the trend
in the market, being in gear with the
tape, and not fighting the major
movements. Fighting the tape is an
open invitation to disaster.”
—Martin Zweig (1942-2013),
US investor1
Introduction
I—like so many other market observers, investors, analysts, and practitioners—
have been making fun of the Efficient Market Hypothesis (EMH), rational
expectations, frictionless markets and the normal distributions for a long time.
Some daily market movements are so large, that if the normal distribution would
apply to security prices, the age of the universe would need to be far older for the
event to occur only once. This report excludes the nagging.
“Of all the ways of defining man, the
worst is the one which makes him
out to be a rational animal.”
—Anatole France (1844–1924),
French novelist and winner of the 1921
Nobel Prize for Literature
Nassim Taleb, in the Black Swan, makes the wonderful and very insightful case
that a normal distribution applies to people’s height but not their wealth. This
means that the normal distribution is applicable in the natural world but is not
applicable for economic or socio-economic phenomena. With height there are no
outliers. If the average height is, say, 1.7m then there is no one who is 1.7mm or
1.7km tall. The height of all people is normally distributed around a mean. Wealth
isn’t. Manuel Valls, France’s youthful PM, has an official wealth of, after 31 years
in “business”, EUR1589.88 and a small apartment. 3 Some French people living in
South Kensington or Brussels or Switzerland have, by a factor in the thousands,
much more than that with the normal distribution not even close to being
applicable.
“France remains an attractive
country for investment. We have no
intention to make France
inhospitable, but we are in a period
of crisis. It is logical that the
wealthiest should make a
contribution.”
—Pierre Moscovici, then French
Minister of Finance2
This report is not about all that. In this report I will assume that the normal
distribution actually has its uses when trying to understand what’s going on. I
believe two aspects about investment management in general and risk
management in particular are meaningfulness of a change in circumstance and an
investor’s conviction in a certain risk exposure. First, a lot of information every
investor is bombarded with on a daily basis is simple not that important, i.e.,
meaningless. Second, position sizing and/or hedging are often a function of an
investor’s conviction, i.e., his confidence in his analysis and reasoning. The normal
distribution can help with both, determining meaning and importance of
“Confidence is what you have
before you understand the
problem.”
—Woody Allen
Zweig (1986), p 121.
Financial Times, 31 December 2012
3 NZZ am Sonntag, 31 August 2014
1
2
Ineichen Research and Management
Page 3
The 4% rule applied
26 September 2014
information as well as helping the investor to articulate or quantify the conviction
in a certain idea or risk exposure. Figure 1 shows one example of applying the
normal distribution to, in this case, economic variables.
Figure 1: Applying the normal distribution
Source: IR&M risk management research flash update, 1 September 2014

The presentation of the Purchasing Manager Indices (PMI) in the updates has
recently been changed to reveal more information on one page with less text.
This allows the quick reader to get a picture of what is going on within
seconds but also allows the analyst to examine detail. The red and green
arrows mark the direction of the most recent change. The red and green balls
show whether the change is further from the mean by more than 0.5 standard
deviations. (I probably need to increase that a bit to reduce the number of
balls. The balls should help the quick reader to see what’s going on in a
speedy and efficient fashion; guiding the eye to what is meaningful so to
speak.)

When arguing in standard deviation terms we implicitly assume that these
variables are normally distributed. They most likely are not. However, using
standard deviation allows comparing different reports that have different
methodologies and thereby different degrees of “wobbliness”. It allows
getting a sense for what is meaningful and what is not. It’s a practitioners’
approach.
Ineichen Research and Management
“Keep your eyes on the ball.”
—Ball sport wisdom
Page 4
The 4% rule applied

In the update shown above, both Italy and Australia were mentioned. The PMI
indicators are from two different data providers. The former has a standard
deviation (sd) of 1.3 while the later has one of 4.0. This means Australia’s PMI
jumps about and can be essentially all over the place. In two out of three
months, the PMI is plus or minus four points of the previous month;
approximately.

On 1st September, Italy’s PMI fell by 2.1 points to 49.8 while Australia’s PMI
fell by 3.4 points to 47.3. One important aspect, of course, is that both fell
below 50 as these are diffusion indices and below 50 means the economy is
contracting. (Although 49.8 is obviously not meaningfully below 50.) But given
the difference in variability of these two measures, 1.3 and 4.0 standard
deviations, these changes are not as close. Italy fell by 1.6 standard deviation
(2.1 pts / 1.3 sd) while Australia only fell by 0.9.
The bottom line is that Italy’s fall was much more meaningful than the fall of
Australia’s. Furthermore, according to our nowcasting from the latest regular risk
management research update (29 August), economic momentum in Italy was
declining while rising in Australia. If the investor has a positive opinion on both
countries, it is more important to check one’s view on Italy. One’s conviction in
both bullish views must have fallen more in the case of Italy as the fall was more
meaningful, especially in light of all the other information on the country’s
economic and political prowess.
26 September 2014
“Italy is de facto already out of the
euro. The country is on its knees...
The northern European countries
are only holding onto us until their
banks have recouped their
investments in Italian sovereign
bonds. Then they’ll drop us like a
hot potato.”
—Beppe Grillo1
“The crime is not in being wrong, it
is in staying wrong.”
—Martin Zweig
The motivation for this report came when viewing the following chart in May.
Figure 2: Apple
“There's no chance that the iPhone
is going to get any significant
market share. No chance. It's a $500
subsidized item. They may make a
lot of money. But if you actually
take a look at the 1.3 billion phones
that get sold, I'd prefer to have our
software in 60% or 70% or 80% of
them, than I would to have 2% or
3%, which is what Apple might get.”
—Steve Ballmer, April 2007
[Editor’s note: the iPhone’s market
share is currently around 41.4%.]
Source: stockcharts.com

1
Apple’s share price rose 8.2% on 24th April after the company announced it
would split its share price. This translates into a move of 5.3 standard
Handelsblatt, 13 March 2013
Ineichen Research and Management
Page 5
The 4% rule applied
26 September 2014
deviations.1 If the normal distribution were meaningful, such a daily event
should occur once every 6.6 million years.
With the benefit of hindsight the 5.3 standard deviation move was meaningful.
The funny thing is; it was meaningful not just with the benefit of hindsight. The
event was meaningful as it occurred. For better or worse; it prompted this report:
a revisit of 4%-rule.
The 4% rule
The 4% rule was developed by Ned Davis and popularised by Martin Zweig in
Winning on Wall Street in the 1980s. The basic idea is that if an index moves by
4% in a week, it is worth our attention. Something meaningful happened and we
should pay attention. The idea is that a move of more than 4% is a beginning of a
new trend or the continuation of an existing trend. Martin Zweig used the Value
Line Index which contains 1700 stocks and is equally weighted. 3 One standard
deviation of weekly returns was roughly 2% at the end of the 1980s for a volatility
of 14.4%.
“To convert a model into a
quantitative formula is to destroy its
usefulness as an instrument of
thought.”
—John Maynard Keynes (1883-1946),
British economist2
Martin Zweig on the virtue of the 4% rule:
The virtue of this Four Percent Model, or any trend-following model, is that if
the market makes a very large move, you will be on the right side of the bulk
of it. But there is no free lunch in the stock market. Although you are
guaranteed of being on the right side of major moves, you may get
whipsawed over very short-term movements. If the market were to zig and
zag by moves only a little bit greater than 4%, you might be zagging when
you should be zigging and zigging when you should be zagging. That will cost
you some money, but in the long-run results of the Four Percent Model clearly
show that it is worth that cost. 4
This quote also applies to IR&M’s Momentum Monitor, an add-on to the risk
management research effort. No model or approach is the Holy Grail. It ought to
be helpful to the practitioner. That’s good enough.
Martin Zweig’s most famous quote is most likely his pacifist remark in relation to
the Fed pumping liquidity into the market:
Don’t fight the Fed.5
This means that, according to Zweig’s theory, if interest rates were going down,
stocks would go up, and vice versa. He also claimed the way to make money was
to be risk-averse, rather than taking chances on the upside. He said he was a big
poker player while at Wharton, but had stopped playing when he became a
money manager because he hated losing, even at cards. One of his major pieces of
advice was never to hold stocks, even of the best companies, in a bear market,
“The Fed’s margin account is far
larger than is yours or ours.”
—Dennis Gartman
Volatility is roughly 25% which translates into a daily standard deviation of 1.55%. This means a 8.2% move is a 5.3
standard deviation event. (Mean daily return is roughly zero.)
2 Letter to Roy Harrod written in the summer of 1938, The Collected Writings of John Maynard Keynes, Vol. XIV, The
General Theory and After, Part II: Defence and Development, Macmillan Cambridge University Press 1973.
3 Zweig (1986), p. 92.
4 Ibid., p. 93
5 Ibid, p. 42, title to Chapter 4.
1
Ineichen Research and Management
Page 6
The 4% rule applied
26 September 2014
since even they could disappoint. Since the 2008 financial crises, “Don’t fight the
Fed” has been the most important rule for the US stock market. The equity market
only started to fall when monetary programs (QE1, QE2, etc.) expired. According
to this rule, one ought to remain long equities as long as the Fed keeps pumping
liquidity into the market. The Greenspan put turned into the Bernanke put which is
now, presumably, the Yellen put. Those who have been calling for the day of
reckoning have been waiting in vain.
Newton’s first law of motion, discovered by Galileo and formulised and
popularised by Newton, is often referred to in relation to trends: once they are on
their way, they keep on going until something stops them. Galileo (1564-1642),
who died in the year Newton was born, first discovered the importance of
acceleration in dynamics. He opposed to consensus view of moving bodies ceasing
to move if left alone; as Bertrand Russell put it in 1946:
“The more people who believe
something, the more apt it is to be
wrong. The person who's right often
has to stand alone.”
—Soren Kierkegaard (1813-1855),
Danish philosopher
Galileo held, as against this view, that every body, if left alone, will continue to
move in a straight line with uniform velocity; any change, either in the rapidity
or the direction of motion, requires to be explained as due to the action of
some ‘force’. This principle was enunciated by Newton as the ‘first law of
motion’.1
The 4% rule, one could argue, is the ‘force’. The EUR/USD spot, for example, rose
from 1.2 in July 2012 to 1.4 in May 2014; surprisingly to many (including this
author). However, that was the trend, a measurable fact at the time. In May there
were two high standard deviation moves on the downside. That was the ‘force’
that changed the direction of motion. The trend has been downward ever since.
The aforementioned strong downward moves occurred on 8th and 9th May as a
result of Draghi talking the EUR down. The two moves in standard deviation terms
were 1.1 and 1.9 standard deviation moves. In other words, the change in
direction was spread over two days. The numbers do not matter that much. It is
the violence, the ‘force’, that signals the change in direction. The practical
relevance is that one doesn’t need forecasting; nowcasting is sufficient.
“There is no safety in numbers, or
in anything else.”
—James Thurber (1894-1961),
American author
The turn of the EUR against most currencies was flagged in our Momentum
Monitor from 12th May, as the swings were large enough for the trend reversal
being picked up by our measure for medium-term momentum. The reversal of
long-term momentum of the EUR/USD spot to negative was flagged on 7th July,
i.e., much later at around a level of 1.35. At the time of writing these lines, the
afternoon of 4th September, the EURUSD spot rate was in freefall to 1.2940, after
Draghi unleashed some more of his “magic”.
“It is clear that monetary policy,
when seen from a German
viewpoint, is too expansive for
Germany, too loose. If we pursued
our own monetary policy, which we
don’t, it would look different.”
—Jens Weisman2
Figure 3 shows the Value Line Index with changes that are larger than 4% in both
directions shown as vertical bars.
1
2
Russell (2004), p. 489.
Reuters, 12 July 2014
Ineichen Research and Management
Page 7
The 4% rule applied
26 September 2014
Figure 3: Value Line Index with weekly returns larger than 4%
Source: IR&M, Bloomberg

In the 1980s there were fewer signals as the market make-up and volatility has
changed.

The most recent signal from January 2013 was a positive one.
Figure 4 shows the same chart with rolling standard deviation based on a rolling
one-year window.
Figure 4: Value Line Index with standard deviation
Source: IR&M, Bloomberg

The standard deviation in the 1980s was 2.1. In the 1990s it was 1.7 and in
the 2000s it doubled to 3.4. In this decade to August 2014 the standard
deviation is 2.6. The standard deviation over the full period of the index since
1983 is 2.5. This means an absolute target, of say 4% in a week, will give
different signals in different volatility regimes.
Ineichen Research and Management
Page 8
The 4% rule applied
The main purpose of this report is to elaborate on the thinking, the idea, behind
the 4% rule rather than making the case for using 4%. Nevertheless, a short test is
in order. I use the S&P 500 Index for this test as it is the more familiar index. Figure
5 shows the S&P 500 price index with returns larger than 4% represented as bars.
26 September 2014
“Great minds discuss ideas.
Average minds discuss events.
Small minds discuss people.”
—Eleanor Roosevelt
Figure 5: S&P 500 Index with weekly returns larger than 4%
Source: IR&M, Bloomberg

4% returns in the S&P 500 Index are clustered as they are with the Value Line
Index and probably every other index. (Professor Robert Engle got the Nobel
for formulising this observation.)

The standard deviation of weekly returns since 1980 is 2.3 which annualised
translates into a volatility of 16.6%.
Figure 6 is the equivalent of Figure 4 but for the S&P Index.
Figure 6: S&P 500 Index with standard deviation
Source: IR&M, Bloomberg

Note that the trough of volatility is always in the middle of the decade. Clearly,
this must be a coincident. Crash in 1987, Asian crisis in 1997, Subprime mess
starting in 2007; clearly we need not worry about 2017 approaching.
Ineichen Research and Management
Page 9
The 4% rule applied
26 September 2014
Test
This is odd: a backtest that doesn’t look great. What happened to the saying:
“there’s no such thing as a bad backtest.”
Figure 7 shows two portfolios, a long-only strategy in the S&P 500 and a rulebased strategy. The latter incorporates the 4%-rule. The rule was to go long when
the weekly return was larger than 4% and stay long until the next sell signal; and
to go short when the weekly return was worse than -4%.
Figure 7: Two portfolios (1980 – August 2014)
Source: IR&M, Bloomberg
Based on weekly date. A 20bp fee per transaction was assumed for the rule-based strategy.

The rule worked for a while and then stopped.
I was surprised by how bad the results looked optically, prompting to check
whether I’ve made a mistake. However, there is an unanticipated lessen:
Something that seems working for a while can stop. Figure 8 shows the calendar
year returns of the two portfolios.
“If something cannot go on forever,
it will stop.”
—Herbert Stein’s Law, Herbert Stein
(1916-1999)
Figure 8: Calendar year returns and return differences of the two portfolios (1980 – August 2014)
Year
Long-only Rule-based Difference
Year
Long-only Rule-based
1980
28.2
25.4
-2.8
1998
30.9
-2.0
1981
-10.3
15.3
25.5
1999
19.8
11.8
1982
14.8
14.8
0.0
2000
-10.1
-24.5
1983
17.3
17.3
0.0
2001
-12.1
42.6
1984
0.8
0.8
0.0
2002
-24.6
-26.9
1985
26.1
26.1
0.0
2003
25.2
14.5
1986
17.8
-13.6
-31.4
2004
10.6
10.6
1987
2.1
56.1
54.1
2005
3.0
3.0
1988
10.2
0.9
-9.3
2006
13.6
13.6
1989
27.3
16.8
-10.4
2007
4.2
-9.8
1990
-7.0
-8.5
-1.5
2008
-41.0
-39.8
1991
23.6
23.6
0.0
2009
29.1
14.6
1992
8.2
8.2
0.0
2010
11.6
19.1
1993
6.1
6.1
0.0
2011
0.0
-24.4
1994
-1.5
-1.5
0.0
2012
11.5
-5.9
1995
34.1
34.1
0.0
2013
31.3
19.6
1996
22.9
22.9
0.0
2014
8.8
8.8
1997
23.7
23.7
0.0
Source: IR&M, Bloomberg
Based on weekly date. A 20bp fee per transaction was assumed for the rule-based strategy.
Ineichen Research and Management
Difference
-33.0
-8.0
-14.3
54.6
-2.3
-10.7
0.0
0.0
0.0
-14.1
1.2
-14.5
7.5
-24.4
-17.4
-11.7
0.0
Page 10
The 4% rule applied
26 September 2014
The rule does not result in outperformance when executed passively. However, as
mentioned before, I believe it is the thinking behind the idea that is of value. One
ought to pay attention to large changes in direction.
Applying the “4% rule” to economic variables
Testing US PMI
The following analysis is based on the PMI (Purchasing Manager Index) from the
Institute for Supply Management (ISM) in the United States. It’s a diffusion index
which means it oscillates around 50 without a long-term drift. A PMI reading over
50 indicates growth or expansion as compared to the previous month, while a
reading under 50 suggests contraction.
Another key number to watch in relation to the PMI, according to Investopedia, is
42.2, since a PMI index above this level over a period of time indicates an
expansion of the overall economy. The August 2014 PMI reading of 59.0 indicates
that the U.S. economy expanded for the 63rd consecutive month, as the PMI first
surpassed the 42.2 level in that expansion in June 2009, which—coincidentally or
not—also marked the start of the post-credit crisis U.S. recovery as determined by
the National Bureau of Economic Research.
Figure 9 shows the PMI since its inception, January 1948.
Figure 9: ISM PMI (1948 – August 2014) with monthly changes
Source: IR&M, Bloomberg

70% of observations are at or above 50 and 91% are above 42.2.
Ineichen Research and Management
Page 11
The 4% rule applied

26 September 2014
Standard deviation for the whole time series is 2.7 points. However, the twoyear standard deviation ranges between 5.2 (summer 1951) and 1.2 (autumn
2007).
It is large changes that are of interest in this report. Figure 10 shows the PMI with
monthly changes that are higher than two standard deviations based on the
previous three years. (I use a rolling standard deviation because the range of
variability over this long period is so large.)
Figure 10: US PMI (1948 – August 2014) with monthly changes >= 2 standard deviation
Source: IR&M, Bloomberg

Judging by the graph alone suggests that, yes, large changes can mark
important turning points but there are false signals too and not all turning
points are highlighted by large moves.
This observation is probably consistent with most investor’s intuition, i.e., it might
be helpful but a fool-proof system it is not. The following analysis gives a bit more
colour, literally, to the signals by decade. In the following tables the signal larger
than two standard deviations is highlighted, the change (in points, not standard
deviation) is shown, as well as the PMI after one month, two, three, six and 12
months. As an example: the first signal in the first table worked well: a sharp,
high-standard-deviation fall indicated that things were about to get worse. In a
nutshell, this is what I try to detect when not just highlighting change in the
updates but also try to get a feel for the significance. A large change to the worse
is a red flag; a minor change isn’t. (And I often hope it’s as simple as that.)
“It is [however] always important to
remember that the ability to see
things in their correct perspective
may be, and often is, divorced from
the ability to reason correctly and
vice versa. That is why a man may
be a very good theorist and yet talk
absolute nonsense…”
—Joseph A. Schumpeter (1883-1950),
Austrian economist1
Schumpeter, Joseph A. (2003) “Capitalism, Socialism and Democracy,” Taylor & Francis, p. 76, footnote 3 in relation to
Karl Marx being silly. First published in 1943.
1
Ineichen Research and Management
Page 12
The 4% rule applied
26 September 2014
1950s
Figure 11 shows PMI in the 1950s. Official recessions were from July 1953 to May
1954 and August 1957 to April 1958.
Figure 11: PMI in the 1950s
Date
30/04/1951
31/08/1952
30/06/1955
31/08/1956
31/05/1958
31/12/1959
Change
-12.0
12.1
-6.2
7.3
7.5
7.6
Change to
53.5
60.4
63.3
51.5
46.6
58.2
1M
50.7
56.1
66.2
55.5
51.4
61.5
PMI after x months
2M
3M
6M
45.5
42.1
49.6
56.2
56.8
55.4
64.8
62.4
65.6
52.7
55.0
51.0
54.7
57.3
62.7
52.3
47.8
44.4
12M
36.7
43.5
47.7
45.3
68.2
44.3
Source: IR&M, Bloomberg

The strong move in April 1951 worked very well. The PMI was more than ten
points lower three months later.

The negative signal in June 1955 was a bit early.

The positive signal in May 1958 was spot on. It marked the end of the
recession reasonably precise. It was also the move that cut through 42.2 from
below. However, the three other positive signals were a mixed blessing.
1960s
Figure 12 shows PMI in the 1960s. Official recessions were from April 1960 to
February 1961 and from November 1969 to November 1970.
Figure 12: PMI in the 1960s
Date
29/02/1960
30/04/1961
31/05/1966
Change
-9.2
8.5
-6.5
Change to
52.3
57.6
57.7
1M
47.8
58.9
59.0
PMI after x months
2M
3M
6M
45.3
42.6
47.6
58.1
58.2
62.2
60.3
58.5
53.7
12M
43.6
55.1
44.5
Source: IR&M, Bloomberg

The 2.2 standard deviation negative move from February 1960 was spot on,
two months prior to the start of the recession which always is made official
long after it started.

The positive April 1961 signal was spot on too, informing the investor that the
recession is over before it was made official.

The May 1966 signal was weird as the PMI was only lower than May 1966 six
and twelve months after the signal. However, 1966 was a difficult stock
market year. The two worst returns of the S&P 500 were in May and August
where the market was down 5.4% and 7.8% respectively for an annual loss
of 13.1%. (The report was probably released in the first few days of June
1966, i.e., too late to avoid the May fall. This is a good example of correlation
not proving causality. Economic surveys do not only influence the stock
market; market moves influence the survey participants too. It’s reflexive, as
George Soros might say.)
Ineichen Research and Management
“Market prices have a notorious
habit of fluctuating.”
—George Soros
Page 13
The 4% rule applied
26 September 2014
1970s
Figure 13 shows PMI in the 1970s. Official recessions were from November 1969
to November 1970, as mentioned before, and from November 1973 to March
1975.
Figure 13: PMI in the 1970s
Date
Change
31/12/1970
28/02/1971
31/07/1973
30/09/1974
31/12/1974
5.7
6.9
-7.2
-6.7
-7.0
Change to
45.4
54.8
57.8
46.2
30.9
1M
47.9
51.2
62.7
42.7
30.7
PMI after x months
2M
3M
6M
54.8
51.2
53.8
54.5
54.2
53.6
63.5
66.2
62.1
37.9
30.9
31.6
34.4
31.6
45.1
12M
57.6
60.6
54.8
54.4
54.9
Source: IR&M, Bloomberg

The December 1970 signal worked very well as it was very close with the end
of the recession that is made official typically long after it ended, i.e., the PMI
is faster than the gatekeepers to “official recessionness”. It was also a move
which took the aforementioned 42.2 level from below.

The February 1971 signal, one could argue, suggest continuation of the
expansion.

The negative July 1973 signal was too early. The recession started in
November of that year.

The negative September 1974 signal below 50 was a good signal, implying
continuation of a difficult market environment, i.e., recession.

The negative December 1974 signal to 31 marked either continuation or the
trough of the recession.
1980s
Figure 14 shows PMI in the 1980s. Official recessions were from January 1980 to
July 1980 and July 1981 to November 1981.
Figure 14: PMI in the 1980s
Date
Change
29/02/1980
31/03/1980
30/04/1980
31/05/1980
31/08/1980
28/02/1983
31/01/1984
31/05/1986
31/08/1986
31/01/1987
4.0
-6.6
-6.2
-8.0
10.5
8.4
-9.4
3.7
4.6
4.4
Change to
50.2
43.6
37.4
29.4
45.5
54.4
60.5
53.4
52.6
54.9
1M
43.6
37.4
29.4
30.3
50.1
53.9
61.3
50.5
52.4
52.6
PMI after x months
2M
3M
6M
37.4
29.4
45.5
29.4
30.3
50.1
30.3
35.0
55.5
35.0
45.5
58.2
55.5
58.2
48.8
54.2
56.1
63.1
58.9
61.0
56.1
48.0
52.6
51.2
51.2
51.2
52.6
55.0
55.5
57.5
12M
48.8
49.6
51.6
53.5
48.3
61.3
50.3
57.2
59.3
57.5
Source: IR&M, Bloomberg

The February 1980 jump to 50.2 was a misfire, an outlier. The PMI fell below
50 in August 1979 and stayed below 50, ignoring the outlier, until September
Ineichen Research and Management
Page 14
The 4% rule applied
26 September 2014
1980. The fall below 50 in March 1980 was meaningful and followed by
further falls.

The other signals were worth paying attention to but not much more; they fell
under the nice-to-know category rather belong to the essential-investorsurvival kit. More often than not did the trend follow the direction of the
sharp move.
1990s
Figure 15 shows PMI in the 1990s. The only official recession in the good old
1990s was from July 1990 to March 1991. The Tequila crisis was in December
1994 and the Asian crisis with all its ripple effects started to unfold in July 1997.
Figure 15: PMI in the 1990s
Date
Change
31/07/1990
30/06/1991
29/02/1992
31/12/1994
31/05/1995
31/07/1995
30/06/1996
31/01/1999
-2.6
5.8
5.4
-3.1
-4.8
4.8
4.5
3.8
Change to
46.6
50.3
52.7
56.1
46.7
50.7
53.6
50.6
1M
46.1
50.6
54.6
57.4
45.9
47.1
49.7
51.7
PMI after x months
2M
3M
6M
44.5
43.2
39.2
52.9
54.9
46.8
52.6
55.7
53.4
55.1
52.1
45.9
50.7
47.1
45.9
48.1
46.7
45.5
51.6
51.1
55.2
52.4
52.3
53.6
12M
50.6
53.6
55.2
46.2
49.1
49.7
54.9
56.7
Source: IR&M, Bloomberg

The negative July 1990 signal was spot on, implying recession long before it
became official and was helpful in holding up a red flag ahead of the S&P
500’s fall from 400 to 300.

The June 1991 and February 1992 signals informed about trend continuation
and were probably not that meaningful. The end of the recession was in
March 1991.

The December 1994 signal marked the Tequila crises but was best ignored by
equity investors as the stock market went up in a straight line during 1995.
The same is true for the negative signal in May 1995, a misfire.

The last three signals larger than two standard deviations in the 1990s were
positive but not very meaningful. There was no hint ahead of the 10%
correction in 1997 and the 25% correction in the S&P 500 in 1998. The Asian
crisis was not related to the US economy and the LTCM episode was
endogenous risk, i.e., also not (directly) related to the business cycle of the US
economy.
2000s
Figure 16 shows PMI in the 2000s. Official recessions were from March to
November 2001 and December 2007 to June 2009. Early in the decade, the TMT
bubble burst. Enron filing for bankruptcy and 9/11 were in 2001. The post-TMT
low in the S&P 500 was in October 2002.
Ineichen Research and Management
Page 15
The 4% rule applied
26 September 2014
Figure 16: PMI in the 2000s
Date
Change
31/12/2000
31/10/2001
30/09/2005
29/02/2008
30/09/2008
31/10/2008
-4.6
-5.4
4.4
-2.7
-4.4
-5.9
Change to
43.9
40.8
56.8
47.6
44.8
38.9
1M
42.3
44.1
57.2
48.3
38.9
36.5
PMI after x months
2M
3M
6M
42.1
43.1
43.2
45.3
47.5
52.4
56.7
55.1
54.3
48.8
48.8
49.2
36.5
33.1
36.0
33.1
34.9
39.5
12M
45.3
49.0
52.2
35.5
54.4
56.0
Source: IR&M, Bloomberg

The negative December 2000 signal was worth one’s attention but a bit late.
The S&P 500 had peaked above 1500 once in March and then in September
of that year. With the wonderful benefit of hindsight it was more important to
pay attention to the PMI falling below 50 which it did in August 2000. The
PMI stayed below 50 until February 2002. If there is a lesson to be learnt from
this, it is that three things are worth the investor’s attention: the trend of
economic momentum (improving/declining), above or below 50 (positive
growth/ negative growth), as well as extreme moves. The December 2000
signal confirmed the negative sentiment well.

The negative October 2001 signal marked the trough. The economy was
already officially in recession and everyone knew it, so the signal was not very
meaningful.

The positive September 2005 signal was one of continuation and nothing to
write home about.

The negative February 2008 signal, a fall from 50.3 to 47.6, a two standard
deviation move at the time, was valuable. The PMI was hovering between 49.0
and 51.1 since August 2007 and then broke in February 2008. China was
already in free-fall by then. The other two high-standard deviation signals in
2008 marked the continuation of the decay. They too were most likely
examples of a reflexive relationship, i.e., the PMI affecting the stock market as
well as the stock market moves affecting the responses of the survey
participants behind the PMI report.
2010s
Figure 17 shows PMI in this decade, the 2010s. There have not been any official
recessions; yet.
Figure 17: PMI in the 2010s
Date
Change
31/05/2011
31/01/2014
-5.2
-5.2
Change to
53.7
51.3
1M
56.6
53.2
PMI after x months
2M
3M
6M
52.9
53.0
52.1
53.7
54.9
57.1
12M
53.2
Source: IR&M, Bloomberg

The strong May 2011 fall market the green shoots turning into brown shoots
as QE2 was about to end in June. IR&M’s US economic model, for what it’s
worth, which includes various PMIs turned in June. The S&P 500 fell from
around 1350 to around 1100 by August. In hindsight, the signal and changing
economic momentum was worth paying attention to.
Ineichen Research and Management
Page 16
The 4% rule applied

26 September 2014
The January 2014 signal was a misfire, as the PMI bounced back in the
following months. But then, perhaps it wasn’t. Markets are being manipulated
in a sense that the authorities want inflation, resulting in asset price inflation,
i.e., equities as well as bonds doing reasonably well. Potentially, Herbert
Stein’s Law from page 10 applies.
Summary
Figure 18 shows the PMI will all the signals that are larger than two standard
deviations.
Figure 18: PMI with signals (January 1948 – August 2014)
Source: IR&M, Bloomberg

Not all high standard deviation signals are meaningful and/or important, as
outlined in the previous, more detailed section.

There were 21 negative signals (red dots). 20 (95%) of those signals happened
when the PMI was falling. This is not true for only one, the last one, the sharp
fall from January 2014.

There were 19 positive signals (green dots). Only 12 (63%) occurred when the
PMI was in a rising trend.
This means a sharp fall in the PMI is probably more significant than a sharp rise.
However, we need to make a mental note that a high or low standard deviation
rise crossing 42.2 from below is worth paying attention to.
In the following section, a brief analysis of Germany’s ZEW economic indicator is
shown.
Ineichen Research and Management
Page 17
The 4% rule applied
26 September 2014
Germany’s ZEW
In our update from 14 August 2014 we showed the following slide:
Figure 19: Germany’s ZEW
Source: IR&M risk management update from 14 August 2014
The changes in mid-August were not only in the wrong direction, towards the
lower left hand corner, but they were also material, i.e., standard deviation was
high which warranted a red flag. Now Newton’s first law of motion applies: the
trend is negative until there is a ‘force’ that reverses the trend. The practical
relevance is that the wind has turned and stopped blowing from behind. One
ought to trim the sails accordingly.
Figure 20 shows the vertical axis from the previous graph in chart format, i.e., the
ZEW Germany Assessment of Current Situations Index since inception in 1992. The
moves that are larger than two standard deviations (based on three-year rolling
standard deviation) are highlighted; green positive, red negative. The shaded areas
show the periods where long-term momentum (as defined in the momentum
monitor publication) of the DAX was negative.
Ineichen Research and Management
Page 18
The 4% rule applied
26 September 2014
Figure 20: ZEW Germany Assessment of Current Situations with high standard deviation signals
Source: IR&M, Bloomberg

There are 13 red dots, i.e., negative signals. 11 of those signals occurred when
long-term price momentum in the DAX was already negative.

The 2011 signal was a bit late. The DAX went from around 7500 in July to a
low close to 5000 in August. The 2.8 standard deviation fall of the economic
indicator was released on 16th August when the DAX already had fallen to
6000. Long-term momentum flipped to negative on Friday 12th August, i.e.,
earlier than the signal from the ZEW indicator. The 100-day moving average of
the IR&M Germany Model peaked in May of that year, i.e., gave, in this case,
ample warning.
Bottom line
Paying attention to high standard deviation changes makes sense but the Holy
Grail of finance it is not. Large changes should be treated as an indication of
meaningfulness rather than a solid signal one can blindly adhere to.
When assessing economic variables I believe the most important thing is the trend,
i.e., the economic momentum. The momentum itself is—to a large extent—a fact,
not someone’s opinion or forecast. The second most important thing to look for is
the reversal of the trend. Only then, third, should we look at the extent of the
magnitude, i.e., standard deviation of a particular move.
Ineichen Research and Management
Page 19
The 4% rule applied
26 September 2014
References
Russell, Betrand (2004) “History of Western Philosophy,” Oxon: Routledge Classics. First published 1946; London:
George Allen & Unwin Ltd.
Taleb, Nassim Nicholas (2007) “The Black Swan – The Impact of the Highly Improbable,” New York: Random House.
Zweig, Martin (1986) “Winning on Wall Street,” New York: Warner Business Books.
Ineichen Research and Management
Page 20
The 4% rule applied
26 September 2014
Publications
Risk management research (subscription based)
We’re done
No reform
Time for caution
Bad equilibrium
Make no mistake
Rekindled financial romance
Economic World Cup 2014 (report)
Illegibly incomprehensible
Seriously concerned
Inflationary complacent
Awaiting further accommodation
Sleeper pins (report)
Underpricing geopolitical risk
Small cushion
Fragile improvement
Recovery far from complete
Boring middle years
Boldly going where others have gone
Trends remain in place far longer
12 September 2014
29 August 2014
14 August 2014
18 July 2014
4 July 2014
18 June 2014
6 June 2014
4 June 2014
16 May 2014
1 May 2014
15 April 2014
11 April 2014
28 March 2014
14 March 2014
28 February 2014
14 February 2014
31 January 2014
17 January 2014
6 January 2014
Everything mean reverts
Positive momentum mode
Blue skies and red flags
Walking a tightrope (report)
Tapering off the table
Uncertainty matters
No dope no hope
Change spotting (report)
More fragile than it looks
Fragile growth
In remembrance of caution
Growth surprises, not taper timing
Highly accommodative (report)
Tapering-off talk tapering off
Mispriced bubbles
All they know
Pretty shocking
No understanding – no assurance
IR&M momentum monitor (inaugural issue and tutorial)
A new dimension (report)
A bit of friction here and there
As long as it takes
Keep it coming
The end of a system
Great Unrecovery continues
Showing mettle
Currency wars
Repressionomics (report)
Far from over
20 December 2013
4 December 2013
19 November 2013
6 November 2013
21 October 2013
8 October 2013
26 September 2013
19 September 2013
13 September 2013
30 August 2013
15 August 2013
2 August 2013
17 July 2013
1 July 2013
18 June 2013
3 June 2013
17 May 2013
3 May 2013
3 May 2013
19 April 2013
4 April 2013
22 March 2013
12 March 2013
1 March 2013
22 February 2013
15 February 2013
1 February 2013
18 January 2013
3 January 2013
Wriston’s Law of Capital still at work
A very long process
…
No risk (report)
Wriston’s Law of Capital (report)
What makes bears blush (report)
Europe doubling down (inaugural report)
19 December 2012
5 December 2012
Ineichen Research and Management
IR&M’s risk management research
consists of 25-30 risk management
updates, 20-30 flash updates, 45-50
momentum monitors and 4-8
thematic reports per year.
26 October 2012
10 July 2012
11 April 2012
3 October 2011
Page 21
The 4% rule applied
26 September 2014
Copyright © 2014 by Ineichen Research and Management AG, Switzerland
All rights reserved. Reproduction or retransmission in whole or in part is prohibited except by permission. The information set forth in this document has been obtained from publicly available sources,
unless stated otherwise. All information contained in this report is based on information obtained from sources which Ineichen Research and Management (“IR&M”) believes to be reliable. IR&M
provides this report without guarantee of any kind regarding its contents.
This document is for information purposes only and should not be construed as investment advice or an offer to sell (nor the solicitation of an offer to buy) any of the securities it refers to. The
information has not been independently verified by IR&M or any of its affiliates. Neither IR&M nor any of its affiliates makes any representations or warranties regarding, or assumes any responsibility
for the accuracy, reliability, completeness or applicability of, any information, calculations contained herein, or of any assumptions underlying any information, calculations, estimates or projections
contained or reflected herein. Neither this document nor the securities referred to herein have been registered or approved by any regulatory authority of any country or jurisdiction.
This material is confidential and intended solely for the information of the person to whom it has been delivered and may not be distributed in any jurisdiction where such distribution would
constitute a violation of applicable law or regulation.
While this document represents the author’s understanding at the time it was prepared, no representation or warranty, either expressed or implied, is provided in relation to the accuracy,
completeness or reliability of the information contained herein, nor it is intended to be a complete statement or summary of the securities markets or developments referred to in the document. It
should not be regarded by recipients as a substitute for the exercise of their own judgment.
Investing in securities and other financial products entails certain risks, including the possible loss of the entire amount invested. Certain investments in particular, including those involving structured
products, futures, options and other derivatives, are complex, may entail substantial risk and are not suitable for all investors. The price and value of, and income produced by, securities and other
financial products may fluctuate and may be adversely impacted by exchange rates, interest rates or other factors. Information available on such securities may be limited. The securities described
herein may not be eligible for sale in all jurisdictions or to certain categories of investors. You should obtain advice from your own tax, financial, legal and accounting advisers to the extent that you
deem necessary and only make investment decisions on the basis of your objectives, experience and resources.
Past performance is not necessarily indicative of future results.
Unless specifically stated otherwise, all price information is indicative only.
No liability whatsoever is accepted for any loss (whether direct, indirect or consequential) that may arise from any use of the information contained in or derived from this document. IR&M does not
provide tax advice and nothing contained herein is intended to be, or should be construed as a, tax advice. Recipients of this report should seek tax advice based on the recipient’s own particular
circumstances from an independent tax adviser.
Ineichen Research and Management
Page 22