Holmberg2020
Holmberg2020
1. Brief Introduction
This paper shows that random numbers generated by the Global
Consciousness Project (GCP) significantly correlate with stock market
returns. This topic is investigated since market prices are the result of
Correspondence:
Ulf Holmberg, PhD, Independent Researcher, Kammakargatan 50, 111 60,
Stockholm, Sweden. Email: [email protected]
Within the field of economics, the market price is a term that refers to
the price at which an asset or service can be bought or sold. Economic
Copyright (c) Imprint Academic
theory further suggests that the market price converges to the ‘clearing
market price’, which is the point where the forces of supply and
demand meet. Since the market price depends on the demand and
supply of a good or service, it is also likely to change regularly in
order to adapt to events that may affect the underlying demand and
supply functions.
Sudden and unexpected macroeconomic and political occurrences
could be classified as such events and arguably such events could
cause the market price to be re-evaluated and changed. This since the
events could affect consumers (demand side shifts), producers (supply
side shifts), or even change the ‘rules of the game’ (i.e. regulatory
changes). However, in the absence of such events, the market price
could be viewed as stable and as an emergent result of the ‘sponta-
neous coordination of the plenty’.1 But for the coordination to occur,
economic agents need to ‘know what they want’, have preferences,
and be able to order and rank the choices presented to them (see e.g.
Mas-Colell, Whinston and Green, 1995). All these factors thus
implicitly assume that some form of conscious behaviour as con-
sumers, for instance, maximizes some underlying utility function
while firms simply seek to maximize profits (and have the cognitive
ability to do so).2 The maximization procedures referred to above are
1 For a discussion on the emergent properties of market prices see Wang et al. (2018).
2 Even though such views on economic agents could be said to grossly oversimplify the
complexity of human behaviour, it should be noted that the underlying economic theory
is flexible enough to accommodate other important aspects of human will. Within e.g.
STOCK RETURNS & THE MIND 33
also likely to occur with ‘one’s own self-interest’ first in mind and, as
famously stated by Adam Smith (1776) in his work The Wealth of
Nations, ‘[markets are] led by an invisible hand to promote an end
which was no part of his intention’. Thus, the individual’s conscious
behaviour is likely to play an important part in the determination of
market prices.
It should, however, be noted that there exists no agreement within
the field of economics as to how the underlying functions that
economic agents seek to maximize should be formalized. Most econo-
mists would, however, agree that a ‘consumer preference set’ of some
sort is needed and that parts of such a preference set could be shared
with other economic agents.3 From this it could be understood that
For personal use only -- not for reproduction
5 Perhaps the problem with explaining what consciousness is originates from the prob-
lems faced in the definition of the concept. Consciousness could, for instance, be
defined as the state of being aware of and responsive to one’s surroundings, but since
such a definition (or similar versions of it) is imprecise, the term has also been defined
in terms of sentience alone, e.g. awareness, qualia, and subjectivity.
36 U. HOLMBERG
6 Physicists have found that even passive observation of quantum phenomena can change
the measured result (see e.g. Buks et al., 1998).
7 It is noted that this interpretation of the observer effect is controversial within the field
of physics.
8 Electromagnetic theories of consciousness branch off into a ‘Cemi’ version and a
‘quantum mind’ version. The former proposes that digital information from neurons is
integrated to form a conscious electromagnetic information field in the brain (see e.g.
McFadden, 2002), and the latter that electric dipoles of water molecules constitute a
quantum field, referred to as the cortical field, with corticons as the quanta of the field.
Thus, the two alternative theories are linked in a sense.
9 One viewpoint here is that the brain may be viewed as a ‘quantum computer’ and that
all psychological phenomena, including consciousness, can be explained using the pro-
cesses of quantum computing (e.g. Penrose, 1994).
STOCK RETURNS & THE MIND 37
by its critics, this has been one of the project’s weaknesses since the
events chosen could be selected in order to fit the data. This criticism
has been addressed by the GCP as they have constructed a predefined
protocol on how to choose the events to be studied. For this study,
however, such criticism is irrelevant as I seek a time series correlation
that does not rely on the importance of single large emotional events.
I chose to study daily stock market returns which thus also requires
the need for a daily aggregate variable from the GCP data. The aggre-
gate GCP data variable derived relies on the the large bulk of publicly
available GCP data (collected every second) and converts it into daily
observations that aims to capture significant deviations from what
should be expected from random numbers. To this end, I use the
For personal use only -- not for reproduction
10 Most GCP studies use XOR filtered data in their studies. Such a filtering procedure will
exclude events that could cause spurious effects (e.g. temperature changes) but, as the
procedure removes some possibly important data, I here use its unfiltered version.
11 Note that the 15-minute intervals underlying the maximum Z-score calculations begin at
00:00:00 UTC. Please visit the GCP website for further details http://noosphere.
princeton.edu/.
12 In particular, all dates with reported bad data are removed. Also, three dates with
unusually high maximum, minimum, and average values are removed.
STOCK RETURNS & THE MIND 39
Figure 1. The GCP data and stock market returns could have a testable
intersection.
Copyright (c) Imprint Academic
6. Results
In this section, the relationship between Max[Z] and stock market
returns (r) is analysed, i.e. the intersection in Figure 1.14 The returns
are obtained from the Dow Jones Global Equity Index as this index is
13 A Taylor series is a series expansion of a function about a point that allows for an
approximation of functional dependence.
14 Returns are defined as rt = (Pt – Pt – 1)/Pt – 1 where Pt is the price/index value at time t.
40 U. HOLMBERG
ideal for capturing global stock market movements and daily changes
in global sentiment. In other words, this index is a good fit for testing
the existence of an intersection as in Figure 1.15
Requiring the data to be balanced, missing data are excluded such
that only dates on which there are both Max[Z] values and some
activity on the Dow Jones Global Equity Index are used. In Table 1
below, descriptive data for the variables used are displayed and, as can
be seen, stock market returns exhibit the ‘usual’ stock market return
properties while Max[Z] is largely skewed and exhibits excess
kurtosis.
15 The index measures the performance of stocks that trade globally, targeting 95%
coverage of markets open to foreign investment. It is float market cap weighted and
quoted in USD.
STOCK RETURNS & THE MIND 41
16 Note that if the usual standard errors are used the qualitative nature of the results is
unaffected.
17 Data on the Dow Jones Global Equity Index began during March 1999.
42 U. HOLMBERG
1.09E-07 – 1.23E-07*
βMax[Z]t-1 – 0.0003** 0.0003*
βMax[Z]t-12 – –4.31E-06** –3.83E-06*
R2 1.79% 1.78% 1.91%
Significance levels: * 10%, ** 5%, and *** 1%.
Note: The sample size (N) is 5237.
18 The 15-minute intervals, from which the daily maximum Z value is calculated (the
Max[Z] variable), begins at 00:00:00 UTC and ends 24 hours thereafter.
STOCK RETURNS & THE MIND 43
ence but, as most variance still remains unexplained, the results open
up an obvious avenue for future research. Note also that the significant
dependence on Max[Z] is unaffected by the inclusion of the lagged
returns and that this variable is only included for the residuals to be
well behaved (and in order to reduce missing variable bias).
The results presented in Table 2 also carry with them some clues
regarding the shape between Max[Z] and stock market returns. This
since it is found that the polynomial in equation (1) has a positive first
term and a negative second term, a result that holds true for both the
Max[Z]t dependence as well as for Max[Z]t – 1. Also, if the current
date’s Max[Z] dependence is studied, both on a stand-alone basis and
in conjunction with Max[Z]t – 1, a third and positive polynomial term is
For personal use only -- not for reproduction
found. This can be seen from the first and last column in Table 2. The
results also suggest that small to moderate values of Max[Z] coincide
with positive returns, which is a result that corresponds nicely to the
Copyright (c) Imprint Academic
analysis.20 The statistical results excluding such large values are pre-
sented in Table 3 and, as can be seen, significant correlations remain.
The ‘open’ return series maintains its dependence on Max[Z]t while its
dependence on Max[Z]t – 1 shifts and becomes linear. Also, a combined
specification suggests such dependence on the ‘open’ series, while for
the ‘close’ series the predating dependence vanishes. These results
thus suggest that large Max[Z] values could, in part, be the cause of
the non-linearity and also that the results support the qualitative con-
clusion that a statistical dependence between these variables exists.
20 A Max[Z] value of 10 is highly unlikely and thus a good proxy for extreme value and
outlier detection.
STOCK RETURNS & THE MIND 45
21 The tables lowest P-value related to Max[Z] is for ‘open’ returns on Max[Z]t –1. Here, the
P-value is about 0.0003, i.e. a one in 3,333.333 result if due to chance.
46 U. HOLMBERG
Asian indexes correlate with Max[Z]t – 1.22 The Japanese Nikkei 225 is
an exception but, taken together, these dependences hint towards the
importance of time.
such a dependence was found to exist for several different world stock
market indexes.
The relationship was found using a newly defined variable, derived
from the GCP data. This new variable was constructed in a way that
captures large changes in the GCP data over the past 24 hours. This
variable was labelled Max[Z] and was found to be correlated with
global stock market returns, as well as various local well-known stock
market indexes. It was also shown that the relationship between
Max[Z] and stock market returns can be approximated with a second-
and third-degree polynomial and that the polynomial dependence is
dependent on whether today’s Max[Z] values are used or if the
Max[Z]t – 1 is studied. The results thus suggest that yesterday’s GCP
For personal use only -- not for reproduction
interesting to follow.
The results also validate some of the claims made by the GCP, and
since the GCP data rest on the assumption that consciousness stretches
out beyond our heads and can affect hardware-generated random
numbers at a distance, the current paradigm with regards to conscious-
ness needs to be discussed. The results actually invite a discussion on
alternative theories of consciousness as the results presented herein
cannot be understood using the prevailing scientific understanding of
consciousness alone. Perhaps the results are better understood through
the lens of some alternative theories as some of these theories allow
for the possibility of consciousness stretching out beyond our heads.
Furthermore, since the random numbers collected by the GCP use
quantum tunnelling techniques to obtain the random numbers,
quantum brain dynamics theories or electromagnetic theories of con-
sciousness are obvious candidates in a pursuit of a deeper under-
standing of the results. Here, this is left as an interesting avenue for
future research.
On a less grandiose level, a better formalization of the functional
form linking large random numbers and stock market returns could be
studied and better daily GCP data variables could be constructed and
tailored to fit individual markets. Also, a more thorough study of the
relationship between different indexes could be done and such a study
should also adjust for time differences in reporting and possible
exchange rate effects. Furthermore, pure market sentiment and market
volatility could be related to the Max[Z] variable and could be studied,
and so could any relationships found between different markets. If
48 U. HOLMBERG
References
Atmanspacher, H. (2015) Quantum theory and consciousness: An overview with
For personal use only -- not for reproduction
Lansing, K.J. & Tubbs, M. (2018) Using sentiment and momentum to predict
stock returns, FRBSF Economic Letter, Federal Reserve Bank of San Francisco.
Llinás, R. (2002) I of the Vortex: From Neurons to Self, Cambridge, MA: MIT
Press.
Mas-Colell, A., Whinston, M. & Green, J. (1995) Microeconomic Theory, Oxford:
Oxford University Press.
McFadden, J. (2002) The conscious electromagnetic information (cemi) field
theory: The hard problem made easy?, Journal of Consciousness Studies, 9 (8),
pp. 45–60.
Nelson, R., Jahn, R. & Dunne, B. (1986) Operator-related anomalies in physical
systems and information processes, Journal of the Society for Psychical
Research, 53 (803), pp. 261–285.
Nelson, R. & Bancel, P. (2011) Effects of mass consciousness in random data
during global events, Explore, 7, pp. 373–383.
Newey, W.K. & West, K.D. (1987) A simple, positive semi-definite, hetero-
For personal use only -- not for reproduction
Small and the Human Mind, pp. 93–143, Cambridge: Cambridge University
Press.
Pockett, S. (2012) The electromagnetic field theory of consciousness: A testable
hypothesis about the characteristics of conscious as opposed to non-conscious
fields, Journal of Consciousness Studies, 19 (11–12), pp. 191–223.
Radin, D., Nelson, R., Dobins, Y. & Houtkooper, J. (2006) Assessing the evidence
for mind–matter interaction effects, Journal of Scientific Exploration, 20 (3), pp.
361–374.
Sharpe, W.F. (1978) Investments, Upper Saddle, River, NJ: Prentice-Hall.
Shiller, R.J. (2017) Narrative economics, American Economic Review, 107 (4), pp.
967–1004.
Smith, A. (1776) An Inquiry into the Nature and Causes of the Wealth of Nations,
Book IV, Chapter II, paragraph IX, London: W. Strahan & T. Cadell.
Taylor, B. (1715) Methodus Incrementorum Directa et Inversa [Direct and
Reverse Methods of Incrementation], in Latin, pp. 21–23, Prop. VII, Thm. 3,
Cor. 2, translated into English by Bruce, I. (trans.), London: William Inny.
Van den Noort, M., Lim, S. & Bosch, P. (2016) On the need to unify neuroscience
and physics, Neuroimmunology and Neuroinflammation, 3, pp. 271–273.
Wang, L., Ahn, K., Kim, C. & Ha, C. (2018) Agent-based models in financial
market studies, Journal of Physics: Conference Series, [Online], doi:
10.1088/1742-6596/1039/1/012022.