Description
The usefulness of technical analysis, or charting, has been questioned because it flies in the face of the ‘random walk’
and tests present conflicting results. We examine chartists’ decision-making techniques and derive a taxonomy of charting
strategies based on investors’ market ontologies and calculative strategies. This distinguishes between trend-seekers and
pattern-seekers, and trading as a system or an art. We argue that interpretative activity plays a more important role than
previously thought and suggest that charting’s main appeal for users lies in its power as a heuristic device regardless of its
effectiveness at generating returns.
Identi?cation through technical analysis: A study
of charting and UK non-professional investors
Philip Roscoe
1
, Carole Howorth
*
Institute for Entrepreneurship and Enterprise Development, Lancaster University Management School, Lancaster LA1 4YX, United Kingdom
Abstract
The usefulness of technical analysis, or charting, has been questioned because it ?ies in the face of the ‘random walk’
and tests present con?icting results. We examine chartists’ decision-making techniques and derive a taxonomy of charting
strategies based on investors’ market ontologies and calculative strategies. This distinguishes between trend-seekers and
pattern-seekers, and trading as a system or an art. We argue that interpretative activity plays a more important role than
previously thought and suggest that charting’s main appeal for users lies in its power as a heuristic device regardless of its
e?ectiveness at generating returns.
Ó 2008 Elsevier Ltd. All rights reserved.
Introduction
Only days after landing in my new job I’ve found
myself praising such statements from investors
as: ‘I was looking at the 10-day moving average
last night and it is a perfect reverse duck tail
and pheasant. Let’s bet the ranch.’ At this junc-
ture my role was only to shout encouragement:
‘Yeah! Let’s do it.’ (Lewis, 1989, p. 192)
In his entertaining account of life as a Salomon
Brothers bond salesman, Michael Lewis ?nds him-
self bewildered by the investment practice of techni-
cal analysis, commonly known as ‘charting’. This is
a method of identifying investment opportunities
using graphs. Unlike fundamental analysis, charting
requires no information other than price history; it
is not necessary to know the activity – nor even
the name – of the company whose shares are traded;
nor the precise nature of the ?nancial instrument in
question; nor the uses and likely demand for a given
commodity. Chartists are not necessarily schooled
in the staples of fundamental analysis: economics,
accounting, industry expertise and ?nancial model-
ling. Instead, they use methods of varying complex-
ity to extrapolate past price movements into future
predictions. Many researchers share Lewis’ bewil-
derment when encountering charting; for ?nance
researchers especially it ?ies in the face of the ‘ran-
dom walk’ of stock movements and the theory of
e?cient markets (Malkiel, 2003). Moreover, it
ignores the commonsense understanding that secu-
rity prices should re?ect the value of the underlying
asset (Preda, 2007b).
Despite this there is considerable evidence,
o?ered mainly by ?nance researchers but also by
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.05.003
*
Corresponding author. Tel.: +44 1524 594847.
E-mail addresses: [email protected] (P. Roscoe),
[email protected] (C. Howorth).
1
Tel.: +44 7769 737934.
Available online at www.sciencedirect.com
Accounting, Organizations and Society 34 (2009) 206–221
www.elsevier.com/locate/aos
sociologists, that charting is a popular strategy, even
among professional traders (Menkho?, 1997; Tay-
lor & Allen, 1992; Zaloom, 2003). The employment
of its own language, as highlighted in the opening
quote, separates charting from other types of analy-
sis, providing status and legitimacy and identifying
proponents as experts (Batchelor & Ramyar, 2006;
Lo et al., 2000). However, although there are
numerous textbook summaries of charting methods
(Batchelor & Ramyar, 2006), there is a striking
absence of evidence on the way that charting is used
‘in the wild’, and thus little discussion of the calcu-
lative strategies and behaviour of individual chart-
ists. This study proposes a conceptualisation of
chartists through an inductive examination of how
investors who use charting make investment deci-
sions, the techniques that they use, and the way in
which they make sense of the markets.
We focus on non-professional (non-salaried)
investors for several reasons. They are relatively
under researched. They are free from the hierarchi-
cal controls that a?ect professional traders and from
the networks built up between them, proximate or
otherwise (Buenza & Stark, 2004; Knorr Cetina &
Bruegger, 2000; Knorr Cetina & Bruegger, 2002;
Zaloom, 2006), because they are investing their
own funds and often from their own homes. The
result is that the observed decision-making process
is less encumbered by exogenous factors. Moreover,
recent studies have questioned the behaviour of
non-professional (individual) investors, arguing that
they are less sophisticated in comparison to profes-
sionals and that they process information di?erently
(Allee et al., 2007; Frederickson & Miller, 2004). It
is suggested that, where individual investors are
less-informed, they will be over-con?dent in their
knowledge and trade too aggressively (Bloom?eld
et al., 1999). Following Mayall (2006) we de?ne
non-professional investors as individuals investing
their own money in the stocks of companies through
the ?nancial markets; while some may depend on
this for their livelihood, the force of ‘non-profes-
sional’ is to distinguish between these and those sal-
aried investors employed by ?nancial organisations.
This article contributes to the growing Social
Studies of Finance (SSF) research paradigm. It pre-
sents empirical evidence on the activities of non-
professional investors that provides a valuable
insight into their behaviour. A taxonomy is pre-
sented that identi?es four types of charting strategy
based on the market ontology of individual inves-
tors and their calculative autonomy. The taxonomy
highlights a di?ering understanding of how the mar-
ket is structured (ordered or otherwise) among indi-
viduals and di?erent levels of reliance on the
calculative activities of other market agents. In each
category, the importance of individual interpretative
activity and the variations in employment of meth-
ods is stressed. The implication for researchers here
is twofold: some isolated tests of the e?cacy of spe-
ci?c charting methods may lack validity in the real
world, and also a concentration on the ability of
charting methods to develop excess pro?ts neglects
the importance of these methods as heuristic devices
located within the broader interpretative skills of
their users. We do not claim that our ?ndings, based
on a limited group of interviewees, are representa-
tive of the universe of investors as a whole, but sug-
gest a number of exploratory propositions that can
be tested in future research.
Article structure
In the following section we examine previous
studies of technical analysis. These show that
charting is widespread, but o?er mixed evidence
of its success at providing excess pro?ts. We intro-
duce the SSF research project and explain how
this literature provides an avenue for better under-
standing of the calculative activities of chartists.
This is followed by an explanation of the research
method adopted and an interpretive analysis of
interviews detailing the activities of 12 non-profes-
sional investors in the UK who use charting tools
regularly. The post-analysis conceptualisation
derives a four-way taxonomy of investors based
on their calculative methods and market ontology.
The discussion then examines the most important
aspects of this conceptualisation, and suggests a
number of propositions for future research. The
conclusion highlights that this study has impor-
tant implications for earlier and future research,
as well as for practitioners.
Previous studies
Two streams of research have considered chart-
ing. Researchers in ?nance have concentrated on
showing the popularity of the method and testing
its usefulness as a method of investment selection,
while SSF research has focused on documenting
charting as a socially situated practice. This section
explains how these two streams of research are
drawn upon to guide our study.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 207
Finance research on charting
Charting has attracted an increasing amount of
attention from ?nance researchers, exhaustively
reviewed by Park and Irwin (2004). As well as sur-
veys showing the popularity of charting, there are
theoretical explanations based on rational feedback
models (De Long et al., 1990; Shliefer & Summers,
1990), noisy rational expectation models and others.
The majority of studies, though, focus on the possi-
bility of achieving excess returns using charting
methods. This is an important area of research
because demonstrations of this directly contradict
the E?cient Markets Hypothesis (EMH) assump-
tion that prices contain all known information and
that abnormal returns cannot be made either by
fundamental or technical analysis (Fama, 1970,
1991; Malkiel, 2003). Tests of EMH consistently
present con?icting results (see Zuckerman, 2004;
Preda, 2007a for an explanation based on structural
incoherence). Park and Irwin (2004) show that of 92
studies carried out from 1988 to 2004, 58 report that
charting methods consistently generated economic
pro?ts, but Park and Irwin suggest that these stud-
ies remain methodologically ?awed. Their review
highlights that there is, as yet, insu?cient evidence
to determine the e?cacy of charting methods one
way or the other.
Despite the existence of positive studies, many
?nance researchers remain unconvinced of the prin-
ciples underpinning charting, with critics equating it
with ‘voodoo ?nance’ and ‘alchemy’ (Malkiel,
1996). Jegadeesh (2000) suggests that there is ‘no
plausible explanation as to why these patterns
should indeed be expected to repeat.’ Batchelor
and Ramyar (2006) found no evidence of the exis-
tence of Fibonacci series and Elliott wave patterns
in the Dow Jones Industrial Average between 1915
and 2003. Malkiel (2003, p. 78) states: ‘the record
of professionals does not suggest that su?cient pre-
dictability exists in the stock market . . . to produce
excess returns’, commenting also that with enough
time and patience it is possible to tease any pattern
out of a given data set. Researchers also complain
that testing is di?cult because chartists fail to report
their results in a rigorous manner, and because rules
are so vague and complicated as to make replication
impossible (Batchelor & Ramyar, 2006, p. 2).
Within ?nance, a growing literature also exists on
individual investors, where their suboptimal behav-
iour is compared with the norms of professional
?nance practice. Examples of this include De
Bondt’s (1998) survey of individual investor beliefs,
and evidence that individual investors tend to place
more belief in pro-forma disclosures (Allee et al.,
2007), trade too often (Barber & Odean, 2000,
2001), retain losers and sell winners (Odean,
1998b), and buy stocks that have a higher pro?le
in the news (Barber & Odean, 2004). The focus of
this article is on the techniques used by chartists in
general; while we are looking speci?cally at non-
professional investors there is no data to suggest
that the techniques used by non-professional inves-
tors di?er substantially from their professional
counterparts.
The social studies of ?nance (SSF) programme
Many of the tests of charting in the ?nance liter-
ature are based on assumptions of rational eco-
nomic behaviour. Social studies of ?nance have
sought to show that the decisions of market actors
are in?uenced by social surroundings, hierarchies
and networks. For example, Baker’s (1984) study
of a trading ?oor showed that trading pits with
smaller crowds could achieve more stable prices
than those with larger numbers, although it is possi-
ble that these small crowds operated as cartels.
Abola?a (1998) argues that market norms, such as
transparency, arise out of repeated interactions
between traders, rather than existing as a given.
Knorr Cetina and Bruegger (2000) and Zaloom
(2006) show that traders are constrained by their
position in hierarchical organisations, while Mac-
Kenzie (2003a) shows that the relationships of
hedge fund managers located as a social elite in geo-
graphic proximity to each other e?ectively limited
their capacity to act as market arbitrageurs.
Sociological work has also focused on the mech-
anisms of the investment decision. Knorr Cetina
and Bruegger’s work has centred on the screen as
a productive mechanism of the market (Knorr Ceti-
na, 2005; Knorr Cetina & Bruegger, 2000; Knorr
Cetina & Bruegger, 2002) while Buenza and Stark
(2004) have shown how a trading room functions
as a calculative device. Michel Callon’s theoretical
contributions have begun to examine how invest-
ment theory and technique may have a substantive
role in determining security prices (Callon, 1998,
2007; Callon & Muniesa, 2003); some empirical evi-
dence of this has been produced regarding the
implementation of options pricing theory (MacKen-
zie, 2003b; MacKenzie & Millo, 2003) and the tech-
nique of portfolio insurance which helped
208 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
precipitate the 1987 ?nancial market crash (Mac-
Kenzie, 2004). Callon (2007) suggests that evidence
on charting may further contribute to this topic.
The engagement of social studies of ?nance with
charting has been limited. Preda (2007b) has written
about the origins of the practice, giving an account of
how charting developed beyond a simple technique
to an epistemic community represented by ‘simulta-
neously a theory of ?nancial markets, a theory based
technique for forecasting prices, a set of instruments,
a commodity sold by the members of the group, a
commodity around which data processing ?rms
emerged, a media discourse, and a narrative’.
Zaloom (2003) encountered the technique in Lon-
don’s trading rooms, noting that decisions cluster
around signi?cant numbers, conferring on them an
agential property. Godechot (2001) shows that chart-
ing is popular among professional traders in Paris.
Mayall (2006) makes an important ?rst step in
broadening our understanding of individual chartists
with her empirical study of non-professional traders
in Australia. She identi?ed four ‘ideal types’ of trad-
ing: as a ‘scienti?c’ system; as an art based on judge-
ment and intuition; as a game of skill; and as a voyage
of exploration. The ?rst, ‘trading as a system’, is a
category comprising those who strive to remove all
human judgment from their trading activity by dele-
gating all the work to a computerised system, while
the second, ‘trading as an art’, comprises those who
renounce algorithm-based selection entirely in
favour of cognitive, often visual, judgement. In the
third category, where participants consider charting
a game of skill they concentrate on outwitting imag-
ined others, shown at work in the market through
price movements, and in the fourth, those who con-
sider charting a voyage are constantly working to
improve their methods, in a journey towards ever
better means of predicting the market. These four
categories are linked by the common theme of using
charting as a means of ‘seeing’ the market.
Summary of literature
The literature discussed above presents several
important questions. It is clear that charting is pop-
ular, and models explain why this may be so, but it
is not clear that it is capable of generating above
market returns on investment. It is not clear how
chartists actually make investment decisions in the
wild, nor whether individuals using charting tech-
niques beat the market. Finance researchers con-
struct ever more rigorous versions of textbook
models (e.g. Lo et al., 2000), but the assumptions
that underpin such models do not necessarily repre-
sent the day to day activities of chartists accurately.
This study supplies much-needed evidence on the
behaviour of chartists. It will build on Mayall’s ini-
tial contribution to derive a robust conceptualisa-
tion of charting behaviour based on how investors
make sense of the market (their market ontology),
and whether they are autonomous in their construc-
tion and use of charting tools. A four-way taxon-
omy of di?erent types of charting style, and four
propositions for future research are presented.
Research method
Data were obtained from multiple sources and
were mainly qualitative. Aface-to-face questionnaire
was administered at four UK events aimed at non-
professional investors in 2005 and 2006. This pro-
vided a broad overview of the extent of charting
and a screening mechanism for identi?cation of
potential interviewees. The main sources of data
analysed were interviews with 19 non-professional
investors, of whom, 12 had an interest in charting
(see Table 1). The majority of interviewees were iden-
ti?ed from respondents to the questionnaire. Other
sources were employed to introduce greater potential
di?erence between interviewees and to reduce bias.
Two interviewees resulted from an advertisement
on an investment bulletin board and two intervie-
wees were identi?ed through recommendations from
other interviewees. An advertisement on the website
of smaller company market Ofex yielded no replies
and corporate ?nance issuers were unable to assist
in the research. Two interviews, which took place
at evening seminars, were shorter and were not
recorded. Some of the interviewees were interviewed
twice to investigate emerging topics in further detail.
In terms of the interviews with chartists, 18 were con-
ducted with the 12 investors interested in charting;
interviews lasted between 25 and 50 minutes. Table
1 summarises the source of each interviewee, the
number of interviews and basic characteristics. Due
to the geographical dispersion of interviewees, two
interviews were conducted face-to-face and the
remainder were conducted by telephone.
The investor events provided good opportunities
for observing the behaviour of non-professional
investors and the range of products and services
available to them. Observations from these events
were recorded using ?eld notes, and relevant bro-
chures and demonstration material were collected.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 209
Thus, data included interview transcripts, ?eld
notes, responses to the survey questionnaire and
marketing materials for investor products.
Questionnaire data helped to shape interview
outlines and highlight areas of interest. Question-
naires were important in ?rst drawing attention to
the prevalence of charting, as 59% of respondents
cited charting as one of their key decision-making
techniques.
2
Initial interviews were open to allow
themes to emerge (Glaser & Strauss, 1968; Locke,
2001) and then the second round of interviews fol-
lowed a semi-structured schedule. The process of
collecting interviewees and conducting interviews
continued until saturation was achieved. Analysis
was inductive and followed Miles and Huberman’s
(1994) thematically clustered matrices to identify
common elements within the data; questionnaire
data, observations of the investor events, and in
some instances brochures and literature, o?ered tri-
angulation points (Webb et al., 1965) and additional
detail. An ethnomethodological approach to data
collection and analysis (Gar?nkel & Sachs, 1970)
treats interviewees’ accounts of their practices and
beliefs as meaningful phenomena in their own right
(Lynch, 1993), as is appropriate for an empirical
project in a relatively unresearched area.
We recognise some limitations in the data. The
majority of interviewees were recruited at investor
fairs and this could lead us to overemphasise the
importance of these events for investors. The inter-
viewees are not a representative sample and we do
not claim that the ?ndings are representative of
the universe of investors as a whole, although we
note that the data displays a number of parallels
with that gathered in Australia by Mayall (2006).
In the light of these limitations we do not generalise
the ?ndings but suggest a number of propositions
that can be tested by future research projects.
Initial data analysis
Among the chartists there was a distinction
between those who used charting as the mainstay
Table 1
Characteristics of interviewees
Pseudonym Gender Source of interview # of interviews Age Portfolio size Investing style
Albert M IX 06 1 60+ £100k–£149 C
Anne F IX 05 2 60+ £200k+ F
Chris M IX 06 1 50–59 £150k–£299k C
George M Advertisement 2 40–49 £200k+ F,I
James M IX 05 1 60+ £100k–£149 C
Karl M IX 06 1 30–39 £50k–£99k I
Max M IX 06 1 40–49 Less than £50k C
Mickey M Referral 2 60+
*
C
Mike M IX 05 1 60+ £100k–£149 F,I
Nigel M Referral 2 30–39 £200k+ F,I
Peter M Advertisement 1 30–39 £100k–£149 F
Robert M IX 05 2 60+ £50k–£99k C
Simon M IX 05 2 40–49
*
C
Stewart M IX 06 2 60+ £200k+ F
Sunil M Seminar 1 40–49
*
F
Terry M IX 06 1 30–39 Less than £50k C
Tony M IX 05 1 60+ £150k–£299k C
Trevor M IX 06 1 60+ £200k+ F
William M Seminar 1
*
F
Key: : Short interview, not taped.
*
: Undisclosed.
C: Chartist, makes use of charts as primary decision-making tool.
F: Fundamentalist, makes use of fundamental analysis as primary decision tool.
I: Incidental chartist, makes substantial use of charts as part of mixed decision-making strategy.
2
It may be the case that chartists are more likely to be
encountered at investor fairs, as these are one of the places where
charting methods are sold and promoted. This could be because
investors are converted to charting at fairs, or chartists attend out
of an existing interest. However, several non-charting intervie-
wees were also recruited at investor fairs (see Table 1). There is,
therefore, insu?cient evidence to demonstrate that interviewees
recruited at investor fairs are more likely to be predisposed to
charting.
210 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
of their investment decision-making and those who
used it alongside other methods. The analysis con-
siders all those who used charting to some extent
and made investment decisions based on the results
and interpretation of charting. However, three
interviewees (George, Mike and Nigel) explicitly
rejected the title ‘chartist.’ They made use of the
charts in a less systematic way and can be classi?ed
as ‘incidental chartists.’ They are characterised by
the use of basic charting tools and the language
and concepts of charting, alongside other methods.
For example, George visits companies and meets
the management personally before investing rela-
tively large sums in illiquid small cap companies.
He does, however, make use of charts before buying
a stock, using the programme to check the timing of
his investment in order to avoid trends that he
expects to persist:
If there’s something that looks as if it’s in a fairly
nasty downtrend then I wouldn’t want to buy it
until there is some indications that downtrend
is ?nished. (George)
Table 1 highlights that ‘incidental chartists’
George and Nigel were in the largest portfolio
bracket (over £200k), and that all the other intervie-
wees investing more than £200,000 were non-chart-
ists. While it is tempting to regard this fact as
signi?cant, there is little to suggest that this is the
case. Absolute portfolio size need not bear any rela-
tionship to investment success; in the case of
George, for example, it re?ects career success as
an entrepreneur, while Mike, whose portfolio falls
in the £150k to £199k bracket, professed a poor
track record as an investor. Moreover, those with
larger portfolios are likely to be older investors
who have had time to accumulate substantial assets
(Anne, Stewart and Tony)
3
; interviews showed that
these investors learned more ‘traditional’ investment
methods many years previously. Interviews also
indicated that charting has a close linkage with lev-
eraged products such as spread-betting and CFDs,
which require smaller portfolios, and therefore
attract younger or less capital-rich investors. There
is also the possibility that chartists utilise a smaller
proportion of their funds in an activity perceived
as more risky. With the exception of Max, and
Simon (who uses charting for spread-betting trades
that require less capital for the risk accepted), the
interviewees all had committed the majority of their
liquid (non-property, non-pension) assets to their
share-trading activity.
Developing a taxonomy of charting styles
Analysis revealed some evidence of each of May-
all’s four ideal types of trading among our intervie-
wees: i.e. as a ‘scienti?c’ system; as an art based on
judgement and intuition; as a game of skill; and as a
voyage of exploration. However, the ?rst two types
were of particular importance and we found ele-
ments of the third and fourth styles distributed
among chartists in both the ?rst and second catego-
ries. In a scienti?c system, chartists will strive to
remove all elements of human judgement and emo-
tion, seen as a weakness, from their trading. Those
who consider charting an art, aim to increase their
intuitive, often visual, understanding of the mar-
ket’s movements as a means of divining price move-
ments. We consider that the categories of system
and art are important for understanding the activi-
ties of chartists, but that they require deepening
and strengthening. In our scheme, system encom-
passes those chartists who use a system based on
the interpretative skills of others; these individuals
are characterised by their purchase and implementa-
tion of whole charting methodologies, from the very
simple to extremely complex, where their own activ-
ity is limited to the carrying out of the processes
required by these methodologies. The second cate-
gory, art, designates those for whom their own
interpretative input is of primary importance in
the charting process, developing their own methods
in the process.
Our analysis identi?ed a further distinction
between the charting styles of interviewees, related
to the individuals’ understanding of stock market
ontology, i.e. how the market is. While some inter-
viewees believed that the market is chaotic and dis-
ordered, others held that while super?cially chaotic,
the market is organised along certain predictable
principles, often related to patterns or cycles of
numbers. This distinction in ontology necessarily
in?uences calculative techniques. The ?rst group
believe only in the persistence of trends (share price
momentum), and are termed here ‘trend-seekers’.
These trend-seekers hunt for stocks where the price
is moving steadily in a particular direction, or
3
Initially, there were no observable patterns with regard to the
age of the investors, the size of investment or wealth. There was a
distinct lack of women among the non-professional investors and
we were only able to obtain one woman interviewee, who was not
a chartist.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 211
‘trending’, and buy or sell short accordingly. The
trend, identi?ed in a variety of ways, is followed
until it runs out of momentum. Trend-seekers make
no assumption about the underlying nature of the
market, beyond the observation that trends tend
to have some persistence. As Simon, one trend see-
ker, said: ‘There’s no way you can predict share
prices, so all I try to do is latch on to price
movements’.
The second group are termed ‘Pattern-seekers’.
These pattern-seekers believe that it is possible to
predict share price movements and changes in price
direction. Number series, such as Fibonacci, are
used to describe shapes which can be searched for
in the market either visually or using an automated
system. The pattern will determine not only the
change in price direction but also the amount that
it can be expected to run in that direction. This
method involves the ontological assumption that
there is some order hidden behind the chaotic
appearance of the market, and that these series of
numbers have some kind of determining power over
the actions of market actors. This belief, clearly
articulated by interviewees, links market movement
to numbers found in tides, waves, pine cones, and
the human body. For example
Fibonacci ratios exist everywhere, they exist in
art, they exist in the human body. If you measure
the distance from your shoulder to your ankles,
and then you measure the distance of your arm
you’ll see that that is a Fibonacci ratio, I think
it’s about 1.618, or .618, or your arm is a ratio
of your body. When something looks aestheti-
cally pleasing to us, very often we will ?nd that
Fibonacci ratios are existing in the relationships
between objects in a painting, let’s say. (Terry)
Other interviewees used systems based on Elliott
Waves and the more complex Delta wave model. In
these cases share prices are governed by ‘impulse’
and ‘retracement’ waves in groups of ?ve and two,
often broken down into substructures. Interviewees
also employed ‘pivot levels’, described by Chris as
‘calculators for the day ahead based on the high,
low and close of the day before.’ Our interviewees’
descriptions of these patterns tally closely with those
o?ered by Batchelor and Ramyar (2006), summaris-
ing textbook charting methods. However, the strik-
ing di?erence here is the unpredictability of the
eventual decision-making process, which is driven
by individual chartist’s own interpretative activity
and methodological bricolage.
Further analysis indicated that the distinction
between trend-seeking and pattern-seeking could
be combined with the system/art distinction to give
a four-way taxonomy of charting behaviour, which
is presented in Fig. 1. Vertically, Fig. 1 distinguishes
between charting styles based on art (interpretative)
and systems (non-interpretative). Horizontally,
charting styles are separated into trend-seeking
and pattern-seeking. This provides the basis for a
taxonomy that identi?es four distinct types of chart-
ing style. Each of the interviewed chartists was
empirically mapped onto the resulting taxonomy.
There were no overlapping or missing categories.
Each type of charting style was given a descrip-
tive identifying name. ‘Number-Crunching’ chart-
ists follow trends and use their own methods,
construct their own indicators or even build their
own systems: Simon, Robert and Tony. ‘Black-Box-
ing’ chartists follow trends but prefer someone else
do the interpretative work: Albert, James, and the
‘incidental’ chartists, Mike, Nigel and George.
Chris, Mickey and Terry are the ‘Wave-Sur?ng’
chartists, searching for patterns using automated
systems and making use of purchased methodolo-
gies. Finally, ‘Chart-Seeing’ investors search for
patterns through creative visualisation and search-
ing by eye; while envious stories of these individuals
are common, only one of our interviewees (Max) ?t-
ted this category. The following section considers
each type in more detail.
Black-Boxing
‘Black-Boxing’ makes use of an entirely systema-
tized method that requires little or no interpretative
activity. Chartists in this group do not commit sig-
ni?cant resources to developing their skills as chart-
ists, and instead rely on simple analysis provided by
others. The interpretative system need not be based
on the investor’s computer; two of the interviewees,
Albert and James, subscribe to a newsletter which
provides identi?ed trends. Albert, who said he was
a ‘busy family man’ stated that the simplicity and
convenience of this approach means that it is the
best method he has found in 15 years of investing.
All the same, he does not buy without performing
his own checks, watching the recommended shares
on a simple end of day charting programme. At this
point
‘if the graph is going like the north face of the
Eiger I may well buy, and I’ll hold those until
212 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
they drop. I only buy shares when they’re going
up’ (Albert)
George, Nigel and Mike all use simple systems in
combination with fundamental analysis. George
uses the same programme as Albert. Nigel makes
use of one particular indicator, the RSI, before
making a purchase, although he has only a rough
idea of what this indicator does, while Mike, who
tends to invest in smaller companies on the basis
of share tips, also makes a casual reference to the
charts before placing his order. Mike stated that if
he spots a trend forming – when a share suddenly
starts to ‘tip up’ – he will invest.
Accessibility and ease of use is taken to an
extreme by ‘black box’ programs which automati-
cally process data through a pre-programmed algo-
rithm and generate investment recommendations,
which can be transmitted directly to a broker. How-
ever, although this kind of software was seen on
sale, and interviewees remarked that it existed, none
of the interview or survey respondents used this
kind of software. It may be suggested that even
among the Black-Boxing chartists, those who are
prepared to place their investment decisions entirely
in the hands of an automated system are relatively
rare.
Number-Crunching
Trend-seeking chartists who take decisions into
their own hands by constructing their own systems
and methods provided much richer and more
involved descriptions than their peers in the ?rst cat-
egory . Number-Crunching uses a barrage of devices
for spotting trends; collectively termed ‘indicators’,
these range from the simple moving average, to
the recherche´ Ichimoku (a Japanese technique that
draws a cloud beneath or above a stock chart, and
from which the user can deduce the existence of a
trend). Simon described his initial method as
follows:
The ?rst thing that I actually looked at was a
breakout system. . .you’d buy a stock when it
went through the 20 day moving average and
keep moving the stop[loss] up, and sell it when
it went through the 50 day moving average, and
the idea is to capture the sort of wave move-
ment. . . (Simon)
While this method proved unsatisfactory because
small-scale ?uctuations in stock prices kept trigger-
ing the stop-loss, it served as a basis for Simon to
start developing more complex methods. Building
new and testing techniques, usually through a brico-
lage of existing methods, is a crucial part of this
group’s charting activity.
A particularly important tool for this process is
‘back-testing’. This process involves the chartist
running a retrospective simulation with historical
price data to ascertain whether a particular combi-
nation of indicators would have identi?ed a su?-
cient number of trends. The process is usually
Art
(interpretative)
System (non-
interpretative)
Trend-seekers
(order-less market)
Pattern-seekers
(ordered market)
Self built calculative systems;
commercial systems running
self assembled indicators;
mathematical (curve-fitting)
back-testing
(Simon, Robert, Tony)
Programme trading (black box
systems); commercial systems
running pre-assembled indicators;
purchased analysis; no back-
testing
(Albert, George, James, Mike,
Nigel)
Pattern search by eye; creative
visualisation; pattern-seeking
back-testing
(Max)
Automated pattern search via
commercial packages;
Fibonnaci/Eliott computation
programmes; proprietary
methodologies; no back-testing
(Chris, Mickey,Terry)
‘Number-Crunching’
‘Black-Boxing’
‘Chart-Seeing’
‘Wave-Surfing’
Fig. 1. A taxonomy of charting styles.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 213
computerised and back-testing functionality is an
important part of more complex charting software.
Robert, discussing his most recent experiment, gave
an example of this process
I can look for shares on the weekly chart where
the 14 day RSI is crossing over 50 and the slope
of the 40 week exponential moving average is
greater than a certain value, and the slope of
the 10-week moving average is greater than
another value, or certain other conditions, and
so on. . . I can make it [the software package]
automatically put little buy and sell signals on
the charts, then I can go and look at those charts
and say to myself would those have been good
buy or sell signals, and if they’re not quite right
I can adjust them slightly. . . (Robert)
This process of testing and adjustment is
repeated until a retrospectively successful strategy
is discovered, at which point it will be applied in real
time. The intention of back-testing is to give the user
con?dence that a particular method will lead to the
discovery of successful trends by the demonstration
that it would have done so in the past. The back-
testing process is often operated alongside a dummy
portfolio, where new methods are tested on a real-
time basis, either without risking any money or by
using very small stakes. Tony and Simon both use
this method, with Simon operating a home-made
trading simulator constructed from an Excel spread-
sheet and downloaded end of day data.
These systems need not deliver a successful result
in every instance. Chartists will accept surprisingly
low probabilities of success in their simulations;
Robert, for example, is aiming at a success rate of
just over 50%. This is acceptable because of the inte-
gral role that the stop loss (an order placed at the
brokerage to automatically exit a position should
the share price move by a certain amount against
the trade) plays in the strategies of these chartists.
Chartists view the stop loss as a means of increas-
ing the probability of success across their trades; by
setting the stop loss relatively tightly they improve
their probability of success. Against this, many
chartists argue that stop-losses can be automatically
triggered by minor moves in the market causing
positions to be exited at a loss when, had they been
held longer, they would have resulted in a pro?t.
Simon encountered this in his early methods and
confessed that it had presented a serious problem.
He ascribes the phenomenon to market volatility.
Other interviewees (Tony, Chris, and Mickey)
argued that stop-losses are manipulated by those
professional players with the power to move prices,
who force the market down in order to trigger auto-
matic sales and buy up stock cheaply. References to
an imagined and malevolent other are reminiscent
of the ‘spoofer’ identi?ed by Zaloom (2003), and
also recall Mayall’s category of charting as a com-
bat or game. Anecdotal evidence indicates that the
practice of searching out and triggering stop-losses
does indeed take place.
4
The stop loss is therefore a crucially important
device for chartists, a fact which has not been fully
recognised in the ?nance literature. Park and Irwin
(2004) discuss two early studies based on a stop-loss
?lter role, which showed the method to substantially
outperform a buy-and-hold strategy, but there
appears to be no work combining stop-losses with
more complex charting strategies. Number-Crunch-
ing strategies highlight just how di?cult it might be
to test this. As Batchelor and Ramyar (2006) note,
the individual strategies of chartists are vague, com-
plex, and badly documented, but it is precisely this
variety of indicator bricolage, overlaid with per-
sonal, often unrecognised, interpretative activity
on the part of the user, that appears to give Num-
ber-Crunching its distinctive character.
In summary, the Number-Crunching chartists
make creative use of their resources to try and
develop ever more accurate means of spotting price
trends in the market. In this category, and indeed in
the case of Black-Boxing, the charts are heuristic
devices, providing a way of seeing the market (May-
all, 2006) and of managing market dynamics. The
charts provide a means of searching through an
inaccessibly large universe of stocks, so that a ?nal,
often visual interpretation can be made. As Tony
said:
The charts will tell me what I need to know, and
if it looks interesting, either because it’s going up
or going down, whether it’s worth investigating
further. (Tony)
In Number-Crunching and Black-Boxing calcu-
lative strategies charts are therefore devices for
managing market chaos, seeking out and visualising
investment opportunities, and reinforcing the calcu-
lative skills of their users. For the following two
strategies they are tools for revealing the hidden
4
Donald Mackenzie recalls a professional trader remarking
that on a quiet day he and his colleagues would ‘bounce the price
around’ to take out the stop-losses.
214 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
order in markets. The problem is once again that of
search, and again there are two clearly identi?able
strategies: automation, and a visual, art-based
method.
Wave-Sur?ng
Wave-Sur?ng uses automated methods involving
proprietary methods and systems to search for spe-
ci?c patterns of price movements. In this category
are Chris, Mickey and Terry. Once patterns have
been discovered, the deterministic nature of number
series makes the investment decision relatively easy.
Having identi?ed a share that is demonstrating a
Fibonacci zigzag, the chartist can determine not
only the correct moment to invest, but the distance
that will be travelled in a given direction. They can
then use the stop loss to quantify potential risk and
arrive at a precise risk/reward measure on which to
base the decision to trade.
Wave-Sur?ng methods are more complex than
those of Black-Boxing – the other system-based cat-
egory – and are predicated on the existence of pat-
terns in the market. So ‘system’ here takes on a
wider de?nition to include not just the computer
software but also the theory and associated tech-
nique, which together form an integrated discursive
scheme, purchased in its entirety. Terry’s story
exempli?es the dedication and strong beliefs of the
Wave-Sur?ng chartists. Terry is convinced that
there is a code to be cracked in the markets and
has dedicated considerable expense, both in ?nan-
cial and personal terms, to learn charting skills.
He has spent several thousand pounds on purchas-
ing training CDs and charting software and attend-
ing courses. At the time of interview he had spent all
his spare time for nine months working on it and
intended going full-time after another four months.
Terry’s method was to attend a course or purchase a
training programme, and make use of the method
suggested until he decided that it was not accurate
enough, or that it failed in some other way, and then
repeat the process with a more complex method. He
had worked through Elliott waves, found to be too
inexact for short-term trading; then a method which
integrated Elliott and Fibonacci; then Delta. He
believed Delta to be inaccurate in the short term,
although he believes that the market does obey the
Delta cycles in the longer term, so he moved to a
system of Fibonacci laid over Delta; and ?nally a
method integrating other technical indicators with
Delta and Elliott which he thinks is capable of giv-
ing accurate short-term predictions of market
movements.
The Wave-Sur?ng process of investment search is
simpli?ed and automated by using software pro-
grammes, often sold in conjunction with training
materials or at expert seminars hosted by the soft-
ware manufacturer. These require substantial com-
puting power, usually two PCs and two or more
screens. When interviewed, Terry was using a soft-
ware programme developed by the presenter of the
course that he had most recently attended. The pro-
gramme is able, he said, to search for 9000 combina-
tions of Elliott wave structures. From a list of 300
shares, scanned overnight, it usually produces four
or ?ve potential trades to be assessed for risk and
return. Mickey also uses a charting package to
search for Fibonacci calculations, from which he
establishes points of support and resistance for mar-
ket turns. In practical terms, he says:
it’s dead easy . . . I select Fibonacci, and I click
once on a high point with my mouse and click
once on the low point and the lines are automat-
ically drawn. (Mickey)
Chris uses two PCs, each with two screens run-
ning two separate charting programmes. He scans
the market automatically every minute looking for
stocks that are approaching the bounds of trade cal-
culated by his Pivot formula. On spotting a suitable
candidate he checks for any outstanding bids or
o?ers, and on being satis?ed that the buying or sell-
ing that has caused this movement has run out, he
will take a position.
Chart-Seeing
All three of the ‘Wave-Surfers’ talked about oth-
ers who had managed to elevate these complex pro-
cedures into an intuitive act and who made their
judgements by eye. We term this calculative strategy
Chart-Seeing. Terry and Mickey seemed envious of
this ability, referred to as ‘intuition’ or a ‘native
skill’. They provided examples of investors they
admired who had this ability: in the case of Mickey
an eminent City analyst who scanned his charts by
eye on the train home, or in the case of Terry a
famous trader who over the course of 11 years has
become ‘attuned’ to the market and can ‘see
instantly’ where it is heading. Of the chartists inter-
viewed, only Max has attempted to learn or exercise
this skill. Like his Wave-Sur?ng peers, he began to
learn about charting by attending a number of
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 215
courses, where he was taught Fibonacci-based
(Wave-Sur?ng) methods. He tried these in the mar-
ket, and became increasingly frustrated as his losers
outnumbered his winners. Eventually, he came to
the conclusion that no system could su?ce in its
entirety, realising ‘you’ve actually got to lay on to
that [the most systematic of approaches] a feel for
what the markets are doing’. Being a relative begin-
ner, with less than a year’s trading experience, Max
embarked on a rigorous programme of back-testing
to develop his method of identifying patterns visu-
ally. He has ?nally reached the point where he scans
a universe of 200 stocks, limited by his chosen spe-
cialisation in US equities and his choice of broker,
?rst weekly and then daily to look for patterns
forming.
Max’s ‘feel for what the markets are doing’ is
based on interpretative activity. Interpretation is
crucial for Chart-Seeing. Observation conducted
as part of our research included a seminar given
by the famous trader mentioned by Terry above.
This trader suggested that his method was a
‘straight-forward modi?ed Delta pattern’ but as he
explained the method he employed a considerable
amount of interpretation as he showed his charts,
explaining how he had adjusted the model in
responding to various situations to make a pro?t.
Discussion
Our analysis of charting styles con?rms previ-
ous research that technical analysis encompasses
a variety of methods and techniques. Our classi?-
cation of these techniques provides an improved
understanding of variations in investor decision-
making and raises a number of important points
that are discussed in this section. This section
?nally suggests four propositions for testing by
future research.
The ?rst point to stress is the clear distinction
between the trend-seekers and the pattern-seekers.
This is based on fundamental di?erences in the
investors’ ontology of ?nancial markets. Trend-
seekers see the market as chaotic and unpredictable,
believing only that price trends have some persis-
tence. Pattern-seekers believe that any observed
chaos is only super?cial and beneath it the market
has hidden structures. Our interviews highlighted a
clear distinction; chartists identi?ed strongly with
either pattern-seeking or trend-seeking and ?rmly
believed themselves to be right and the other posi-
tion wrong. This distinction serves to clarify two
assumptions that have appeared in the literature.
The ?rst of these is the elision of charting with
momentum investing, and then of momentum
investing with the search for patterns in the market
(Buenza & Stark, 2004, p. 375). We show that not
all chartists are momentum investors, and that those
who are, the trend-seekers, do not search for pat-
terns. Park and Irwin’s (2004) claim that all charting
focuses on the search for trends is equally mistaken.
The second common misconception is that everyone
necessarily believes in an overarching, transcendent
order in the market (Mayall, 2006, p. 124). This
research clearly shows that this applies only to a sec-
tion of the charting community.
This distinction perhaps helps us in responding
to Menkho?’s (1997) challenge to justify charting
as a rational occupation. Some arguments can
be advanced for the rationality of trend-seeking.
Charts may serve as a rational search method
for a momentum strategy; the usefulness of
momentum investing has been demonstrated in
the US (Jegadeesh & Titman, 2001), and in the
UK (Liu, Strong, & Xu, 1999; Tonks & Hon,
2003). It is also rational to act in a particular
way in the expectation that others will do the
same. Models of herding (Banerjee, 1992; Bikh-
shandani & Sharma, 2001; Hirshleifer & Teoh,
2003) and of asset price bubbles (De Long
et al., 1990) demonstrate this. Jegadeesh (2000)
suggests that charting methods are used in con-
junction with other investment selection methods,
paralleling the activities of the incidental chartists,
although this was clearly not the case with the
other chartists studied.
It is more di?cult to make a case for rationality
among the pattern-seekers. The arguments above
hinge on the assumption that certain charting styles
may have some value as investment methods, and
the same can be said here only in a limited way.
While some numbers do have support and resis-
tance e?ects (Donaldson & Kim, 1993), o?ering
some justi?cation for methods based on important
numbers, more sophisticated pattern systems can
claim no such evidence. Batchelor and Ramyar
(2006) demonstrations of the absence of Fibonacci
support levels is particularly problematic, while
Malkiel (1996) and Jegadeesh (2000) are both insis-
tent that there is no reason for patterns to form in
stock prices.
Psychological biases may also help to account for
the popularity of charting. It can be argued that sys-
tematic overcon?dence (Barber & Odean, 2001;
216 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
Daniel et al., 2001) may explain the con?dence that
individuals show in their own methods and their
willingness to invest on the basis of these, even in
the face of poor investment performance. The charts
themselves, as a psychologically available means of
presenting decisions in a visual format (Kahneman
& Tversky, 1974), may add to the appeal of the
investment technique, the format making decisions
easy and accessible. Finally the pursuit of trends
has a demonstrated psychological appeal (Andreas-
sen & Kraus, 1998) and this too may contribute to
the popularity of charting.
Our analysis suggests that it is mistaken to regard
charting as a means for achieving above market
returns. Finance research provides scant support
for its ability to do so, as do the testimonies of sev-
eral of the chartists. Yet charting is popular among
non-professional investors. Menkho?’s challenge
can, perhaps, be answered in another way. The
research programme of SSF directs attention to
the calculative strategies and technologies of indi-
vidual market agents; Hutchins (1995) shows how
sophisticated processes of calculation can be distrib-
uted among human actors and technological
devices. Buenza and Stark (2004) show this distribu-
tion in action in an investment bank’s trading room.
Charting is amenable to this analysis, as calculation
is distributed to computer programmes and their
designers, to newsletter editors, to fellow chartists,
to imagined others such as the ‘spoofer’, or the
malevolent professional traders making fun of
stop-losses. The charting apparatus in each of the
four categories above serve to produce the market,
to make it visible for individual investors working
alone in their homes, and to allow interaction
between them and the market. In each instance,
the charting apparatus serves to narrow the range
of decisions presented to the individual. Instead of
the unfathomable possibilities of the market, possi-
bilities are constrained, and decisions framed and
disentangled (Callon, 1998). In this analysis, the
importance of charting lies less in its e?cacy as an
investment selection method, and more in its power
to make markets visible and accessible and to sim-
plify and support the huge burden of calculative
activity that falls on individual market agents.
Charting may, in this sense, be considered rational,
as it allows individual non-professional investors to
participate as market actors. This is a theme that
has been strongly articulated across all four catego-
ries: Black-Boxing is quick and e?cient; the com-
plex calculative technologies of Wave-Sur?ng
allow glimpses of the order behind market chaos;
Number-Crunching assembles indicators to pick
out the trends which it seeks; and Chart-Seeing pre-
sents the individual with available raw materials for
visual analysis.
We attempted to gauge the extent to which indi-
viduals did manage to beat the market. However, in
interviews it is di?cult to gain more than subjective
estimates of individual success. The problem is exac-
erbated by the fact that interviewees tended to com-
pare their returns to their starting point, rather than
the market, and that data collection took place dur-
ing a prolonged bull-market. Where possible, inter-
viewees’ returns on investment were ascertained and
classi?ed as high (above market), medium (in line
with market) and low (below market). This is shown
in Table 2.
When considering charting purely as a means of
gaining above market returns, our initial discussion
provided more support for the rationality of trend-
seekers than pattern-seekers and therefore trend-
seekers might be expected to show higher returns
than pattern-seekers. However, we failed to ?nd evi-
dence to support this expectation; some pattern-
seekers were very successful, and some trend-seekers
were not. Additionally, our analysis indicated that
art-based charting styles have a higher status than
system-based ones, and therefore it might be
expected that art-based charting styles would be
associated with higher returns. Again, there was
no evidence to support this. Overall, there was no
evidence to indicate that there was an association
between charting style and returns on investment;
instead our analysis suggested that returns may be
associated with the interpretative skills and activi-
ties of the individual chartist. Therefore, our ?rst
proposition is as follows:
Table 2
Charting style and success level
Pseudonym Charting style Success
Albert Black-Boxing High
George Black-Boxing High
James Black-Boxing High
Tony Number-Crunching High
Chris Wave-Sur?ng High
Nigel Black-Boxing Medium
Max Chart-Seeing Medium
Simon Number-Crunching Medium
Terry Wave-Sur?ng Medium
Mike Black-Boxing Low
Robert Number-Crunching Low
Mickey Wave-Sur?ng Low
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 217
P1: There is no association between charting style
and the ?nancial return on investments.
The lack of association between charting style
and ?nancial return appears to underscore the cen-
tral claim of this article – that, while the charts form
part of the toolkit of heuristic devices that the indi-
vidual investor uses to manage and control the mar-
ket, individual investment performance is largely
dependent on the tacit skills of the chartist. The
clearest example of this was given in a seminar we
attended, where a chart-seeing professional trader
set out a theory of markets based on a four day
cycle, illustrated by charts. The charts frequently
bore no clear relation to either the cycle’s predic-
tions, or the trader’s comments, and instead
appeared to o?er a heuristic support for the trader’s
own tacit interpretative skills, developed over a long
period of time. In this sense we can perhaps answer
Menkho?’s challenge: it is entirely rational for an
individual to use tools that support their activities,
where the tools serve productively to support calcu-
lative activity but not determine the eventual invest-
ment decision. In this instance, we can see that there
may be a distinction between the ‘art’ and ‘science’
categories of chartist, but the argument can equally
be made that calculation is distributed (Hutchins,
1995) to the compilers of newsletters or charting
software, who themselves rely on charts as heuristic
devices.
There are parallels here with discussion of valua-
tion models used by professional securities analysts.
Barker (1999) and Demirakos et al. (2004) show
that analysts consistently prefer simple models such
as price to earnings ratio and dividend yield to
sophisticated models of discounted cash-?ow and
residual income, the latter considered to be too
complicated for real life situations. Barker identi?es
a need for further research in understanding why
simple models are preferred; the implication of
chartists’ experience is that the analysts’ tacit exper-
tise is of primary importance when grappling with
the complexities of real life company valuation,
and the simpler ratios form a more useful, ?exible
and available heuristic device in dealing with this
process.
Our analysis also highlighted the importance of
the stop-loss in the investment strategies of chart-
ists. The stop-loss limits the potential downside of
trades, but may also lead to closing a trade too early
and crystallizing a loss rather than achieving an
eventual pro?t. Further research is required to
establish the e?ect of stop-losses on the pro?tability
of investment strategies. Hence our second
proposition:
P2: There is a positive association between the use
of stop-losses and investment returns.
As with P1, this proposition has another implica-
tion for future research. We suggest that the stop-
loss often performs a markedly di?ering function
for individual chartists than is supposed in ?nance
literature. Rather than being simply another part
of the chartist’s tool kit, it appears that the utility
of the stop-loss is closely related to the signi?cance
of the charts as heuristic devices for investors. The
stop-loss forms an action-based ruling on the valid-
ity of any given trading set up, informing the chart-
ist at an early stage whether he is ‘right’ in his
combination of rules and indicators. Terry alludes
to this with the comment that the stop-loss resem-
bles a ‘slap on the wrist’ that will ‘invalidate your
analysis’. In this sense the power of the stop-loss lies
in its ability as a corrective device for ill-conceived
conceptual arrangements on the part of the individ-
ual, not only improving investment returns but also
continuously enhancing the chartist’s heuristic
apparatus and potentially improving her investment
performance. However, this aspect of the stop-loss
is more di?cult to test than its relationship to
pro?tability.
Our analysis also indicated that there was rela-
tively little movement between charting styles. This
is particularly surprising in the case of the art/sys-
tem divide. As art has a higher status than system,
it might be expected that there would be a progres-
sion from system to art. This was certainly the case
with Max, who moved to a visual style of pattern
spotting after an unsuccessful attempt at a system-
based approach. Apart from this there is little
mobility among our interviewees and it appears that
the choice between system and art-based charting
styles is associated with personal circumstances
and preference. System-based styles are chosen for
reasons of convenience and practicality; its methods
save time and o?er the user an e?cient means of
sifting through large volumes of data. Art-based
styles are popular with those who enjoy experimen-
tation and are prepared to devote time and e?ort to
discovering new methods; Max has, by means of a
personal windfall, been given a period of time where
he could devote extensive e?ort to perfecting his
charting style. There is less conceptual support for
movement between trend-seekers and pattern-seek-
ers as this would require a shift in market ontology,
a change in the individual’s understanding of how
218 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
markets function. However, Chris moved from
‘Number-Crunching’ to ‘Wave-Sur?ng’, making an
ontological leap as he became convinced of the use-
fulness of number series and other predictive indica-
tors. This raises some interesting issues for further
investigation. Our third proposition, then, is
P3: Individuals’ charting styles are relatively static
and are determined by exogenous (non-market)
factors.
There are again parallels with the methods used
by professional securities analysts, where choice of
methodology appears to be in?uenced by the indus-
try sector, type of ?rm, and expectations of the mar-
ket (Barker, 1999; Block, 1999; Demirakos et al.,
2004). It may be the case here too that the evidence
of chartists can illuminate new avenues of research
on professional analysts; perhaps truly exogenous
factors such as personal preferences and skills, as
well as the circumstances and customs of employ-
ment, may be shown to in?uence valuation
methodologies.
Propositions one to three underscore the central
argument of this study – that the use of charting is
not solely associated with generating investment
returns. We have suggested that its power as a heu-
ristic device, enabling individuals to organize, man-
age and understand the complexities of the market
is equally, if not more, important to users. We noted
during interviews that the charts provided a frame
of reference for seeing, recalling and talking about
the market; interviewees would often refer to a chart
as they illustrated a point or substantiated a claim.
We also note that charting o?ers individuals other
non-market returns. The experimentation of the
Number-Crunching investors provided them with
enjoyment, pride, and a sense of mastery over the
market. Across all categories mastery of techniques
provided investors with prestige. In summary,
returns from charting are much broader than a sim-
ple return on investment. Hence, we suggest a ?nal
proposition for future investigation:
P4: Investors use charting for its heuristic power
rather than its ability to generate ?nancial returns.
Conclusion
In this article we have conducted an analysis of
charting styles, and classi?ed them into four groups
depending on whether they were art or system based
and whether they were seeking trends or patterns.
We argue that the visualisation of a ?nancial market
is inexorably linked to, if not driven by, chartists’
views on market ontology and e?ciency. Our taxon-
omy has shown how di?erences in seeing the mar-
ket, driven by fundamental di?erences in chartists’
conception of what the market is, relate to varied
calculative strategies. An investor who believes that
markets are e?cient has little need of this kind of
graphic visualisation, and will simply buy a tracking
fund. The purpose of charting appears to be the
opposite, providing investors with visual illustra-
tions of the structure of markets and showing the
opportunities for pro?t amongst market chaos. A
chartist, almost by de?nition, may not believe in
e?cient markets, whatever the evidence to the con-
trary; chartists who were less successful blamed
themselves for deviation from their methods, and
looked forward to making ‘big money one day’
(Chris). We did not interview, meet, or hear of, any-
one who was an ex-chartist.
Two major themes interlink to underpin the ?nd-
ings in this study. These themes are extremely
important because they a?ect the validity of
assumptions of some earlier research. The ?rst is
that the use of charting and the choice of charting
style do not appear to be associated with ?nancial
returns. The second is the importance of interpreta-
tive and tacit skills in the activities of chartists. We
have shown no clear relationship between charting
style and individual investors’ returns and have sug-
gested that returns may be dependent on individu-
als’ own interpretative skills. This presents
problems for researchers attempting to determine
the usefulness of charting, or its popularity with
investors, on the basis of textbook methods. While
the similarity between our interviewees’ accounts
and textbook presentations of charting method
shows that these are accepted as a basis for individ-
ual charting methods, it is clear that the strategies
pursued by individuals may in fact be very di?erent
from those methods. We therefore suggest that
attention could usefully be given to the way in
which individuals develop their own interpretative
strategies, and that it would be bene?cial for some
future research to focus on individual rather than
generic charting strategies.
There may also be implications for practitioners
such as investors, company managers and market
supervisors. In particular, practitioners should be
aware of the popularity of charting as an investment
method. Rather than dismissing it as irrational or
speculative, it may help practitioners to understand
that it is employed by relatively sophisticated inves-
tors, and appreciate that it may have its roots in
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 219
attempts to manage and make sense of the market
in a rational manner.
Finally, we suggest that previous studies that aim
to demonstrate or refute the e?cacy of charting
methods have only limited scope for developing
our understanding of the popularity of charting
among investors. This study has shown that the
appeal of charting lies in the way that it allows
investors to make sense of, manage, and participate
in the markets. As an integrated discursive scheme,
supplying a way of understanding ?nancial markets
and the tools to turn this understanding into invest-
ment practice, charting can be considered a perfor-
mative technique in the sense of Callon (1998,
2007). This aspect of the appeal of charting is one
which may be investigated most fruitfully by future
research.
Acknowledgements
The authors are very grateful to two anonymous
reviewers and Anthony G. Hopwood for providing
very positive and constructive comments. Opinions
(and the potential errors) expressed in this paper
are, of course, the authors alone.
References
Abola?a, M. (1998). Markets as cultures: An ethnographic
approach. In M. Callon (Ed.), The laws of markets
(pp. 69–85). Oxford: Oxford University Press.
Allee, K. D., Battacharya, N., et al. (2007). Proforma disclosure
and investor sophistication: External validation of experi-
mental evidence using archival data. Accounting Organiza-
tions and Society, 32(3), 201–222.
Andreassen, P., & Kraus, S. (1998). Judgemental prediction by
extrapolation. Mimeo, Harvard.
Baker, W. (1984). The social structure of a national securities
market. American Journal of Sociology, 89(4), 775–811.
Banerjee, A. (1992). A simple model of herd behaviour. Quarterly
Journal of Economics, 107(3), 797–817.
Barber, B., & Odean, T. (2000). Trading is Hazardous to your
Wealth: The common stock performance of individual
investors. Journal of Finance, 55(2), 773–799.
Barber, B., & Odean, T. (2001). Boys will be boys: Gender,
overcon?dence and common stock investment. Quarterly
Journal of Economics(February), 261–292.
Barber, B. M., & Odean, T. (2004). All that glitters: The e?ect of
attention and news on the buying behaviour of individual and
institutional investors. University of California.
Barker, R. G. (1999). Survey and market-based evidence of
industry-dependence in analysts’ preferences between the
dividend yield and price-earnings ratio valuation models.
Journal of Business Finance & Accounting, 26(3–4), 393–418.
Batchelor, R., & Ramyar, R. (2006). Magic numbers in the Dow.
London: Cass Business School.
Bikhshandani, S., & Sharma, S. (2001). Herd behaviour in
?nancial markets. IMF Sta? Papers, 47(3).
Block, S. B. (1999). A study of ?nancial analysts: Practice and
theory. Financial Analysts Journal, 55(4), 86.
Bloom?eld, R., Libby, R., et al. (1999). Con?dence and the
welfare of less-informed investors. Accounting Organizations
and Society, 24(8).
Buenza, D., & Stark, D. (2004). Tools of the trade: The socio-
technology of arbitrage in a Wall Street trading room.
Industrial and Corporate Change, 13(2), 369–400.
Callon, M. (1998). The embeddedness of economic markets in
economics. In M. Callon (Ed.), The laws of markets
(pp. 1–58). Oxford: Oxford University Press.
Callon, M. (2007). What does it mean to say that economics is
performative?. In D. MacKenzie F. Muniesa, & L. Siu (Eds.),
Performing economics. Princeton University Press.
Callon, M., & Muniesa, F. (2003). Les marche´s e´conomiques
commes dispositifs collectifs de calcul. Re´ seaux, 21(122),
189–233.
Daniel, K., Hirshleifer, D., et al. (2001). Investor psychology in
capital markets: Evidence and policy implications. Journal of
Monetary Economics, 49, 139–209.
De Bondt, W. (1998). A portrait of the individual investor.
European Economic Review, 42, 831–844.
De Long, J., Shleifer, A., et al. (1990). Positive feedback
investment strategies and destabilising rational speculation.
Journal of Finance, 45(2), 379–395.
Demirakos, E. G., Strong, N. C., et al. (2004). What valuation
models do analysts use?. Accounting Horizons 18(4), 221.
Donaldson, R., & Kim, H. (1993). Price barriers in the Dow
Jones industrial average. Journal of Financial and Quantitative
Analysis, 28(3), 313–330.
Fama, E. (1970). E?cient capital markets: A review of theory and
empirical work. Journal of Finance, 25, 383–417.
Fama, E. (1991). E?cient capital markets II. Journal of Finance,
46(5), 1575–1617.
Frederickson, J. R., & Miller, J. S. (2004). The e?ects of proforma
earnings disclosures on analysts’ and non-professional inves-
tors’ equity valuation judgements. The Accounting Review,
79(3), 667–686.
Gar?nkel, H., & Sachs, H. (1970). On the formal structure of
practical actions. In J. McKinney & E. Tiryakian (Eds.),
Theoretical sociology (pp. 337–366). New York: Appleton
Century Crofts.
Glaser, B., & Strauss, A. (1968). The discovery of grounded
theory: Strategies for qualitative research. London: Weiden-
feld and Nicholson.
Godechot, O. (2001). Les traders. Paris: la De´couverte.
Hirshleifer, D., & Teoh, S. (2003). Herd behaviour and cascading
in capital markets: A review and synthesis. European Financial
Management, 9(1), 25–66.
Hutchins, E. (1995). Cognition in the wild. Cambridge, MA: The
MIT Press.
Jegadeesh, N. (2000). Foundations of technical analysis: Com-
putational algorithms, statistical inference, and empirical
implementation: Discussion. Journal of Finance, 60(4).
Jegadeesh, N., & Titman, S. (2001). Pro?tability of momentum
strategies. Journal of Finance, 56, 699–720.
Kahneman, A., & Tversky, D. (1974). Judgement under uncer-
tainty. Science, 185, 1124–1131.
Knorr Cetina, K. (2005). From pipes to scopes: The ?ow
architecture of ?nancial markets. In A. Barry & D. Slater
220 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
(Eds.), The technological economy (pp. 122–141). Abingdon:
Routledge.
Knorr Cetina, K., & Bruegger, U. (2000). The Market as an
object of attachment: Exploring postsocial relations in ?nan-
cial markets. Canadian Journal of Sociology, 25(2), 141–168.
Knorr Cetina, K., & Bruegger, U. (2002). Global microstruc-
tures: The virtual societies of ?nancial markets. American
Journal of Sociology, 107(4), 905–950.
Lewis, M. (1989). Liar’s Poker. London: Coronet.
Liu, W., Strong, N., & Xu, X. (1999). The pro?tability of
momentum investing. Journal of Business Finance and
Accounting, 26, 1043–1091.
Locke, K. (2001). Grounded theory in management research.
London: Sage Publications.
Lo, A. W., Mamaysky, H., et al. (2000). Foundations of technical
analysis: Computational algorithms, statistical inference, and
empirical implementation. The Journal of Finance, 55(4),
1705–1765.
Lynch, M. (1993). Scienti?c practice and ordinary action. Cam-
bridge: Cambridge University Press.
MacKenzie, D. (2003a). An equation and its worlds. Social
Studies of Science, 33(6), 831–868.
MacKenzie, D. (2003b). Long term capital management and the
sociology of arbitrage. Economy and Society, 32(3), 349–380.
MacKenzie, D. (2004). The big bad wolf and the rational market:
Portfolio insurance, the 1987 crash and the performativity of
economics. Economy and Society, 33(3), 303–330.
MacKenzie, D., & Millo, Y. (2003). Constructing a market,
performing theory: The historical sociology of a ?nancial
derivatives exchange. American Journal of Sociology, 109(1),
107–145.
Malkiel, B. G. (1996). A random walk down wall street. New
York: W.W. Norton.
Malkiel, B. G. (2003). The e?cient market hypothesis and its
critics. The Journal of Economic Perspectives, 17(1), 59–82.
Mayall, M. (2006). Seeing the market: Technical analysis in
trading styles. Journal for the Theory of Social Behaviour,
36(2), 119–140.
Menkho?, L. (1997). Examining the use of technical currency
analysis. International Journal of Financial Economics, 2,
307–318.
Miles, M., & Huberman, A. (1994). Qualitative data analysis.
London: Sage Publications.
Odean, T. (1998b). Are investors reluctant to realize their losses?.
Journal of Finance 53(5).
Park, C.-H., & Irwin, S. H. (2004). The pro?tability of technical
analysis: A review. AgMAS Project Research Report. Uni-
versity of Illinois at Urbana-Champaign: 04.
Preda, A. (2007a). The sociological approach to ?nancial
markets. Journal of Economic Surveys, 21(3), 506–533.
Preda, A. (2007b). Where do analysts come from? The case of
?nancial chartism. In M. Callon, F. Muniesa, & Y. Millo
(Eds.), Market devices. Oxford: Blackwell.
Shliefer, A., & Summers, C. (1990). The noise trader approach to
?nance. Journal of Economic Perspectives, 4(2), 19–33.
Taylor, M., & Allen, H. (1992). The use of technical analysis in
the foreign exchange market. Journal of International Money
and Finance, 11(3), 304–314.
Tonks, M. T., & Hon, I. (2003). Momentum in the UK stock
market. Journal of Multinational Financial Management,
13(1), 43–70.
Webb, E., Campbell, D., et al. (1965). Unobtrusive measures.
Chicago: Rand McNally.
Zaloom, C. (2003). Ambiguous numbers: Trading technologies
and interpretation in ?nancial markets. American Ethnologist,
30(2), 258–272.
Zaloom, C. (2006). Out of the pits. Chicago: University of
Chicago Press.
Zuckerman, E. (2004). Structural incoherence and stock market
activity. American Sociological Review, 69, 405–432.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 221
doc_970746787.pdf
The usefulness of technical analysis, or charting, has been questioned because it flies in the face of the ‘random walk’
and tests present conflicting results. We examine chartists’ decision-making techniques and derive a taxonomy of charting
strategies based on investors’ market ontologies and calculative strategies. This distinguishes between trend-seekers and
pattern-seekers, and trading as a system or an art. We argue that interpretative activity plays a more important role than
previously thought and suggest that charting’s main appeal for users lies in its power as a heuristic device regardless of its
effectiveness at generating returns.
Identi?cation through technical analysis: A study
of charting and UK non-professional investors
Philip Roscoe
1
, Carole Howorth
*
Institute for Entrepreneurship and Enterprise Development, Lancaster University Management School, Lancaster LA1 4YX, United Kingdom
Abstract
The usefulness of technical analysis, or charting, has been questioned because it ?ies in the face of the ‘random walk’
and tests present con?icting results. We examine chartists’ decision-making techniques and derive a taxonomy of charting
strategies based on investors’ market ontologies and calculative strategies. This distinguishes between trend-seekers and
pattern-seekers, and trading as a system or an art. We argue that interpretative activity plays a more important role than
previously thought and suggest that charting’s main appeal for users lies in its power as a heuristic device regardless of its
e?ectiveness at generating returns.
Ó 2008 Elsevier Ltd. All rights reserved.
Introduction
Only days after landing in my new job I’ve found
myself praising such statements from investors
as: ‘I was looking at the 10-day moving average
last night and it is a perfect reverse duck tail
and pheasant. Let’s bet the ranch.’ At this junc-
ture my role was only to shout encouragement:
‘Yeah! Let’s do it.’ (Lewis, 1989, p. 192)
In his entertaining account of life as a Salomon
Brothers bond salesman, Michael Lewis ?nds him-
self bewildered by the investment practice of techni-
cal analysis, commonly known as ‘charting’. This is
a method of identifying investment opportunities
using graphs. Unlike fundamental analysis, charting
requires no information other than price history; it
is not necessary to know the activity – nor even
the name – of the company whose shares are traded;
nor the precise nature of the ?nancial instrument in
question; nor the uses and likely demand for a given
commodity. Chartists are not necessarily schooled
in the staples of fundamental analysis: economics,
accounting, industry expertise and ?nancial model-
ling. Instead, they use methods of varying complex-
ity to extrapolate past price movements into future
predictions. Many researchers share Lewis’ bewil-
derment when encountering charting; for ?nance
researchers especially it ?ies in the face of the ‘ran-
dom walk’ of stock movements and the theory of
e?cient markets (Malkiel, 2003). Moreover, it
ignores the commonsense understanding that secu-
rity prices should re?ect the value of the underlying
asset (Preda, 2007b).
Despite this there is considerable evidence,
o?ered mainly by ?nance researchers but also by
0361-3682/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aos.2008.05.003
*
Corresponding author. Tel.: +44 1524 594847.
E-mail addresses: [email protected] (P. Roscoe),
[email protected] (C. Howorth).
1
Tel.: +44 7769 737934.
Available online at www.sciencedirect.com
Accounting, Organizations and Society 34 (2009) 206–221
www.elsevier.com/locate/aos
sociologists, that charting is a popular strategy, even
among professional traders (Menkho?, 1997; Tay-
lor & Allen, 1992; Zaloom, 2003). The employment
of its own language, as highlighted in the opening
quote, separates charting from other types of analy-
sis, providing status and legitimacy and identifying
proponents as experts (Batchelor & Ramyar, 2006;
Lo et al., 2000). However, although there are
numerous textbook summaries of charting methods
(Batchelor & Ramyar, 2006), there is a striking
absence of evidence on the way that charting is used
‘in the wild’, and thus little discussion of the calcu-
lative strategies and behaviour of individual chart-
ists. This study proposes a conceptualisation of
chartists through an inductive examination of how
investors who use charting make investment deci-
sions, the techniques that they use, and the way in
which they make sense of the markets.
We focus on non-professional (non-salaried)
investors for several reasons. They are relatively
under researched. They are free from the hierarchi-
cal controls that a?ect professional traders and from
the networks built up between them, proximate or
otherwise (Buenza & Stark, 2004; Knorr Cetina &
Bruegger, 2000; Knorr Cetina & Bruegger, 2002;
Zaloom, 2006), because they are investing their
own funds and often from their own homes. The
result is that the observed decision-making process
is less encumbered by exogenous factors. Moreover,
recent studies have questioned the behaviour of
non-professional (individual) investors, arguing that
they are less sophisticated in comparison to profes-
sionals and that they process information di?erently
(Allee et al., 2007; Frederickson & Miller, 2004). It
is suggested that, where individual investors are
less-informed, they will be over-con?dent in their
knowledge and trade too aggressively (Bloom?eld
et al., 1999). Following Mayall (2006) we de?ne
non-professional investors as individuals investing
their own money in the stocks of companies through
the ?nancial markets; while some may depend on
this for their livelihood, the force of ‘non-profes-
sional’ is to distinguish between these and those sal-
aried investors employed by ?nancial organisations.
This article contributes to the growing Social
Studies of Finance (SSF) research paradigm. It pre-
sents empirical evidence on the activities of non-
professional investors that provides a valuable
insight into their behaviour. A taxonomy is pre-
sented that identi?es four types of charting strategy
based on the market ontology of individual inves-
tors and their calculative autonomy. The taxonomy
highlights a di?ering understanding of how the mar-
ket is structured (ordered or otherwise) among indi-
viduals and di?erent levels of reliance on the
calculative activities of other market agents. In each
category, the importance of individual interpretative
activity and the variations in employment of meth-
ods is stressed. The implication for researchers here
is twofold: some isolated tests of the e?cacy of spe-
ci?c charting methods may lack validity in the real
world, and also a concentration on the ability of
charting methods to develop excess pro?ts neglects
the importance of these methods as heuristic devices
located within the broader interpretative skills of
their users. We do not claim that our ?ndings, based
on a limited group of interviewees, are representa-
tive of the universe of investors as a whole, but sug-
gest a number of exploratory propositions that can
be tested in future research.
Article structure
In the following section we examine previous
studies of technical analysis. These show that
charting is widespread, but o?er mixed evidence
of its success at providing excess pro?ts. We intro-
duce the SSF research project and explain how
this literature provides an avenue for better under-
standing of the calculative activities of chartists.
This is followed by an explanation of the research
method adopted and an interpretive analysis of
interviews detailing the activities of 12 non-profes-
sional investors in the UK who use charting tools
regularly. The post-analysis conceptualisation
derives a four-way taxonomy of investors based
on their calculative methods and market ontology.
The discussion then examines the most important
aspects of this conceptualisation, and suggests a
number of propositions for future research. The
conclusion highlights that this study has impor-
tant implications for earlier and future research,
as well as for practitioners.
Previous studies
Two streams of research have considered chart-
ing. Researchers in ?nance have concentrated on
showing the popularity of the method and testing
its usefulness as a method of investment selection,
while SSF research has focused on documenting
charting as a socially situated practice. This section
explains how these two streams of research are
drawn upon to guide our study.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 207
Finance research on charting
Charting has attracted an increasing amount of
attention from ?nance researchers, exhaustively
reviewed by Park and Irwin (2004). As well as sur-
veys showing the popularity of charting, there are
theoretical explanations based on rational feedback
models (De Long et al., 1990; Shliefer & Summers,
1990), noisy rational expectation models and others.
The majority of studies, though, focus on the possi-
bility of achieving excess returns using charting
methods. This is an important area of research
because demonstrations of this directly contradict
the E?cient Markets Hypothesis (EMH) assump-
tion that prices contain all known information and
that abnormal returns cannot be made either by
fundamental or technical analysis (Fama, 1970,
1991; Malkiel, 2003). Tests of EMH consistently
present con?icting results (see Zuckerman, 2004;
Preda, 2007a for an explanation based on structural
incoherence). Park and Irwin (2004) show that of 92
studies carried out from 1988 to 2004, 58 report that
charting methods consistently generated economic
pro?ts, but Park and Irwin suggest that these stud-
ies remain methodologically ?awed. Their review
highlights that there is, as yet, insu?cient evidence
to determine the e?cacy of charting methods one
way or the other.
Despite the existence of positive studies, many
?nance researchers remain unconvinced of the prin-
ciples underpinning charting, with critics equating it
with ‘voodoo ?nance’ and ‘alchemy’ (Malkiel,
1996). Jegadeesh (2000) suggests that there is ‘no
plausible explanation as to why these patterns
should indeed be expected to repeat.’ Batchelor
and Ramyar (2006) found no evidence of the exis-
tence of Fibonacci series and Elliott wave patterns
in the Dow Jones Industrial Average between 1915
and 2003. Malkiel (2003, p. 78) states: ‘the record
of professionals does not suggest that su?cient pre-
dictability exists in the stock market . . . to produce
excess returns’, commenting also that with enough
time and patience it is possible to tease any pattern
out of a given data set. Researchers also complain
that testing is di?cult because chartists fail to report
their results in a rigorous manner, and because rules
are so vague and complicated as to make replication
impossible (Batchelor & Ramyar, 2006, p. 2).
Within ?nance, a growing literature also exists on
individual investors, where their suboptimal behav-
iour is compared with the norms of professional
?nance practice. Examples of this include De
Bondt’s (1998) survey of individual investor beliefs,
and evidence that individual investors tend to place
more belief in pro-forma disclosures (Allee et al.,
2007), trade too often (Barber & Odean, 2000,
2001), retain losers and sell winners (Odean,
1998b), and buy stocks that have a higher pro?le
in the news (Barber & Odean, 2004). The focus of
this article is on the techniques used by chartists in
general; while we are looking speci?cally at non-
professional investors there is no data to suggest
that the techniques used by non-professional inves-
tors di?er substantially from their professional
counterparts.
The social studies of ?nance (SSF) programme
Many of the tests of charting in the ?nance liter-
ature are based on assumptions of rational eco-
nomic behaviour. Social studies of ?nance have
sought to show that the decisions of market actors
are in?uenced by social surroundings, hierarchies
and networks. For example, Baker’s (1984) study
of a trading ?oor showed that trading pits with
smaller crowds could achieve more stable prices
than those with larger numbers, although it is possi-
ble that these small crowds operated as cartels.
Abola?a (1998) argues that market norms, such as
transparency, arise out of repeated interactions
between traders, rather than existing as a given.
Knorr Cetina and Bruegger (2000) and Zaloom
(2006) show that traders are constrained by their
position in hierarchical organisations, while Mac-
Kenzie (2003a) shows that the relationships of
hedge fund managers located as a social elite in geo-
graphic proximity to each other e?ectively limited
their capacity to act as market arbitrageurs.
Sociological work has also focused on the mech-
anisms of the investment decision. Knorr Cetina
and Bruegger’s work has centred on the screen as
a productive mechanism of the market (Knorr Ceti-
na, 2005; Knorr Cetina & Bruegger, 2000; Knorr
Cetina & Bruegger, 2002) while Buenza and Stark
(2004) have shown how a trading room functions
as a calculative device. Michel Callon’s theoretical
contributions have begun to examine how invest-
ment theory and technique may have a substantive
role in determining security prices (Callon, 1998,
2007; Callon & Muniesa, 2003); some empirical evi-
dence of this has been produced regarding the
implementation of options pricing theory (MacKen-
zie, 2003b; MacKenzie & Millo, 2003) and the tech-
nique of portfolio insurance which helped
208 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
precipitate the 1987 ?nancial market crash (Mac-
Kenzie, 2004). Callon (2007) suggests that evidence
on charting may further contribute to this topic.
The engagement of social studies of ?nance with
charting has been limited. Preda (2007b) has written
about the origins of the practice, giving an account of
how charting developed beyond a simple technique
to an epistemic community represented by ‘simulta-
neously a theory of ?nancial markets, a theory based
technique for forecasting prices, a set of instruments,
a commodity sold by the members of the group, a
commodity around which data processing ?rms
emerged, a media discourse, and a narrative’.
Zaloom (2003) encountered the technique in Lon-
don’s trading rooms, noting that decisions cluster
around signi?cant numbers, conferring on them an
agential property. Godechot (2001) shows that chart-
ing is popular among professional traders in Paris.
Mayall (2006) makes an important ?rst step in
broadening our understanding of individual chartists
with her empirical study of non-professional traders
in Australia. She identi?ed four ‘ideal types’ of trad-
ing: as a ‘scienti?c’ system; as an art based on judge-
ment and intuition; as a game of skill; and as a voyage
of exploration. The ?rst, ‘trading as a system’, is a
category comprising those who strive to remove all
human judgment from their trading activity by dele-
gating all the work to a computerised system, while
the second, ‘trading as an art’, comprises those who
renounce algorithm-based selection entirely in
favour of cognitive, often visual, judgement. In the
third category, where participants consider charting
a game of skill they concentrate on outwitting imag-
ined others, shown at work in the market through
price movements, and in the fourth, those who con-
sider charting a voyage are constantly working to
improve their methods, in a journey towards ever
better means of predicting the market. These four
categories are linked by the common theme of using
charting as a means of ‘seeing’ the market.
Summary of literature
The literature discussed above presents several
important questions. It is clear that charting is pop-
ular, and models explain why this may be so, but it
is not clear that it is capable of generating above
market returns on investment. It is not clear how
chartists actually make investment decisions in the
wild, nor whether individuals using charting tech-
niques beat the market. Finance researchers con-
struct ever more rigorous versions of textbook
models (e.g. Lo et al., 2000), but the assumptions
that underpin such models do not necessarily repre-
sent the day to day activities of chartists accurately.
This study supplies much-needed evidence on the
behaviour of chartists. It will build on Mayall’s ini-
tial contribution to derive a robust conceptualisa-
tion of charting behaviour based on how investors
make sense of the market (their market ontology),
and whether they are autonomous in their construc-
tion and use of charting tools. A four-way taxon-
omy of di?erent types of charting style, and four
propositions for future research are presented.
Research method
Data were obtained from multiple sources and
were mainly qualitative. Aface-to-face questionnaire
was administered at four UK events aimed at non-
professional investors in 2005 and 2006. This pro-
vided a broad overview of the extent of charting
and a screening mechanism for identi?cation of
potential interviewees. The main sources of data
analysed were interviews with 19 non-professional
investors, of whom, 12 had an interest in charting
(see Table 1). The majority of interviewees were iden-
ti?ed from respondents to the questionnaire. Other
sources were employed to introduce greater potential
di?erence between interviewees and to reduce bias.
Two interviewees resulted from an advertisement
on an investment bulletin board and two intervie-
wees were identi?ed through recommendations from
other interviewees. An advertisement on the website
of smaller company market Ofex yielded no replies
and corporate ?nance issuers were unable to assist
in the research. Two interviews, which took place
at evening seminars, were shorter and were not
recorded. Some of the interviewees were interviewed
twice to investigate emerging topics in further detail.
In terms of the interviews with chartists, 18 were con-
ducted with the 12 investors interested in charting;
interviews lasted between 25 and 50 minutes. Table
1 summarises the source of each interviewee, the
number of interviews and basic characteristics. Due
to the geographical dispersion of interviewees, two
interviews were conducted face-to-face and the
remainder were conducted by telephone.
The investor events provided good opportunities
for observing the behaviour of non-professional
investors and the range of products and services
available to them. Observations from these events
were recorded using ?eld notes, and relevant bro-
chures and demonstration material were collected.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 209
Thus, data included interview transcripts, ?eld
notes, responses to the survey questionnaire and
marketing materials for investor products.
Questionnaire data helped to shape interview
outlines and highlight areas of interest. Question-
naires were important in ?rst drawing attention to
the prevalence of charting, as 59% of respondents
cited charting as one of their key decision-making
techniques.
2
Initial interviews were open to allow
themes to emerge (Glaser & Strauss, 1968; Locke,
2001) and then the second round of interviews fol-
lowed a semi-structured schedule. The process of
collecting interviewees and conducting interviews
continued until saturation was achieved. Analysis
was inductive and followed Miles and Huberman’s
(1994) thematically clustered matrices to identify
common elements within the data; questionnaire
data, observations of the investor events, and in
some instances brochures and literature, o?ered tri-
angulation points (Webb et al., 1965) and additional
detail. An ethnomethodological approach to data
collection and analysis (Gar?nkel & Sachs, 1970)
treats interviewees’ accounts of their practices and
beliefs as meaningful phenomena in their own right
(Lynch, 1993), as is appropriate for an empirical
project in a relatively unresearched area.
We recognise some limitations in the data. The
majority of interviewees were recruited at investor
fairs and this could lead us to overemphasise the
importance of these events for investors. The inter-
viewees are not a representative sample and we do
not claim that the ?ndings are representative of
the universe of investors as a whole, although we
note that the data displays a number of parallels
with that gathered in Australia by Mayall (2006).
In the light of these limitations we do not generalise
the ?ndings but suggest a number of propositions
that can be tested by future research projects.
Initial data analysis
Among the chartists there was a distinction
between those who used charting as the mainstay
Table 1
Characteristics of interviewees
Pseudonym Gender Source of interview # of interviews Age Portfolio size Investing style
Albert M IX 06 1 60+ £100k–£149 C
Anne F IX 05 2 60+ £200k+ F
Chris M IX 06 1 50–59 £150k–£299k C
George M Advertisement 2 40–49 £200k+ F,I
James M IX 05 1 60+ £100k–£149 C
Karl M IX 06 1 30–39 £50k–£99k I
Max M IX 06 1 40–49 Less than £50k C
Mickey M Referral 2 60+
*
C
Mike M IX 05 1 60+ £100k–£149 F,I
Nigel M Referral 2 30–39 £200k+ F,I
Peter M Advertisement 1 30–39 £100k–£149 F
Robert M IX 05 2 60+ £50k–£99k C
Simon M IX 05 2 40–49
*
C
Stewart M IX 06 2 60+ £200k+ F
Sunil M Seminar 1 40–49
*
F
Terry M IX 06 1 30–39 Less than £50k C
Tony M IX 05 1 60+ £150k–£299k C
Trevor M IX 06 1 60+ £200k+ F
William M Seminar 1
*
F
Key: : Short interview, not taped.
*
: Undisclosed.
C: Chartist, makes use of charts as primary decision-making tool.
F: Fundamentalist, makes use of fundamental analysis as primary decision tool.
I: Incidental chartist, makes substantial use of charts as part of mixed decision-making strategy.
2
It may be the case that chartists are more likely to be
encountered at investor fairs, as these are one of the places where
charting methods are sold and promoted. This could be because
investors are converted to charting at fairs, or chartists attend out
of an existing interest. However, several non-charting intervie-
wees were also recruited at investor fairs (see Table 1). There is,
therefore, insu?cient evidence to demonstrate that interviewees
recruited at investor fairs are more likely to be predisposed to
charting.
210 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
of their investment decision-making and those who
used it alongside other methods. The analysis con-
siders all those who used charting to some extent
and made investment decisions based on the results
and interpretation of charting. However, three
interviewees (George, Mike and Nigel) explicitly
rejected the title ‘chartist.’ They made use of the
charts in a less systematic way and can be classi?ed
as ‘incidental chartists.’ They are characterised by
the use of basic charting tools and the language
and concepts of charting, alongside other methods.
For example, George visits companies and meets
the management personally before investing rela-
tively large sums in illiquid small cap companies.
He does, however, make use of charts before buying
a stock, using the programme to check the timing of
his investment in order to avoid trends that he
expects to persist:
If there’s something that looks as if it’s in a fairly
nasty downtrend then I wouldn’t want to buy it
until there is some indications that downtrend
is ?nished. (George)
Table 1 highlights that ‘incidental chartists’
George and Nigel were in the largest portfolio
bracket (over £200k), and that all the other intervie-
wees investing more than £200,000 were non-chart-
ists. While it is tempting to regard this fact as
signi?cant, there is little to suggest that this is the
case. Absolute portfolio size need not bear any rela-
tionship to investment success; in the case of
George, for example, it re?ects career success as
an entrepreneur, while Mike, whose portfolio falls
in the £150k to £199k bracket, professed a poor
track record as an investor. Moreover, those with
larger portfolios are likely to be older investors
who have had time to accumulate substantial assets
(Anne, Stewart and Tony)
3
; interviews showed that
these investors learned more ‘traditional’ investment
methods many years previously. Interviews also
indicated that charting has a close linkage with lev-
eraged products such as spread-betting and CFDs,
which require smaller portfolios, and therefore
attract younger or less capital-rich investors. There
is also the possibility that chartists utilise a smaller
proportion of their funds in an activity perceived
as more risky. With the exception of Max, and
Simon (who uses charting for spread-betting trades
that require less capital for the risk accepted), the
interviewees all had committed the majority of their
liquid (non-property, non-pension) assets to their
share-trading activity.
Developing a taxonomy of charting styles
Analysis revealed some evidence of each of May-
all’s four ideal types of trading among our intervie-
wees: i.e. as a ‘scienti?c’ system; as an art based on
judgement and intuition; as a game of skill; and as a
voyage of exploration. However, the ?rst two types
were of particular importance and we found ele-
ments of the third and fourth styles distributed
among chartists in both the ?rst and second catego-
ries. In a scienti?c system, chartists will strive to
remove all elements of human judgement and emo-
tion, seen as a weakness, from their trading. Those
who consider charting an art, aim to increase their
intuitive, often visual, understanding of the mar-
ket’s movements as a means of divining price move-
ments. We consider that the categories of system
and art are important for understanding the activi-
ties of chartists, but that they require deepening
and strengthening. In our scheme, system encom-
passes those chartists who use a system based on
the interpretative skills of others; these individuals
are characterised by their purchase and implementa-
tion of whole charting methodologies, from the very
simple to extremely complex, where their own activ-
ity is limited to the carrying out of the processes
required by these methodologies. The second cate-
gory, art, designates those for whom their own
interpretative input is of primary importance in
the charting process, developing their own methods
in the process.
Our analysis identi?ed a further distinction
between the charting styles of interviewees, related
to the individuals’ understanding of stock market
ontology, i.e. how the market is. While some inter-
viewees believed that the market is chaotic and dis-
ordered, others held that while super?cially chaotic,
the market is organised along certain predictable
principles, often related to patterns or cycles of
numbers. This distinction in ontology necessarily
in?uences calculative techniques. The ?rst group
believe only in the persistence of trends (share price
momentum), and are termed here ‘trend-seekers’.
These trend-seekers hunt for stocks where the price
is moving steadily in a particular direction, or
3
Initially, there were no observable patterns with regard to the
age of the investors, the size of investment or wealth. There was a
distinct lack of women among the non-professional investors and
we were only able to obtain one woman interviewee, who was not
a chartist.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 211
‘trending’, and buy or sell short accordingly. The
trend, identi?ed in a variety of ways, is followed
until it runs out of momentum. Trend-seekers make
no assumption about the underlying nature of the
market, beyond the observation that trends tend
to have some persistence. As Simon, one trend see-
ker, said: ‘There’s no way you can predict share
prices, so all I try to do is latch on to price
movements’.
The second group are termed ‘Pattern-seekers’.
These pattern-seekers believe that it is possible to
predict share price movements and changes in price
direction. Number series, such as Fibonacci, are
used to describe shapes which can be searched for
in the market either visually or using an automated
system. The pattern will determine not only the
change in price direction but also the amount that
it can be expected to run in that direction. This
method involves the ontological assumption that
there is some order hidden behind the chaotic
appearance of the market, and that these series of
numbers have some kind of determining power over
the actions of market actors. This belief, clearly
articulated by interviewees, links market movement
to numbers found in tides, waves, pine cones, and
the human body. For example
Fibonacci ratios exist everywhere, they exist in
art, they exist in the human body. If you measure
the distance from your shoulder to your ankles,
and then you measure the distance of your arm
you’ll see that that is a Fibonacci ratio, I think
it’s about 1.618, or .618, or your arm is a ratio
of your body. When something looks aestheti-
cally pleasing to us, very often we will ?nd that
Fibonacci ratios are existing in the relationships
between objects in a painting, let’s say. (Terry)
Other interviewees used systems based on Elliott
Waves and the more complex Delta wave model. In
these cases share prices are governed by ‘impulse’
and ‘retracement’ waves in groups of ?ve and two,
often broken down into substructures. Interviewees
also employed ‘pivot levels’, described by Chris as
‘calculators for the day ahead based on the high,
low and close of the day before.’ Our interviewees’
descriptions of these patterns tally closely with those
o?ered by Batchelor and Ramyar (2006), summaris-
ing textbook charting methods. However, the strik-
ing di?erence here is the unpredictability of the
eventual decision-making process, which is driven
by individual chartist’s own interpretative activity
and methodological bricolage.
Further analysis indicated that the distinction
between trend-seeking and pattern-seeking could
be combined with the system/art distinction to give
a four-way taxonomy of charting behaviour, which
is presented in Fig. 1. Vertically, Fig. 1 distinguishes
between charting styles based on art (interpretative)
and systems (non-interpretative). Horizontally,
charting styles are separated into trend-seeking
and pattern-seeking. This provides the basis for a
taxonomy that identi?es four distinct types of chart-
ing style. Each of the interviewed chartists was
empirically mapped onto the resulting taxonomy.
There were no overlapping or missing categories.
Each type of charting style was given a descrip-
tive identifying name. ‘Number-Crunching’ chart-
ists follow trends and use their own methods,
construct their own indicators or even build their
own systems: Simon, Robert and Tony. ‘Black-Box-
ing’ chartists follow trends but prefer someone else
do the interpretative work: Albert, James, and the
‘incidental’ chartists, Mike, Nigel and George.
Chris, Mickey and Terry are the ‘Wave-Sur?ng’
chartists, searching for patterns using automated
systems and making use of purchased methodolo-
gies. Finally, ‘Chart-Seeing’ investors search for
patterns through creative visualisation and search-
ing by eye; while envious stories of these individuals
are common, only one of our interviewees (Max) ?t-
ted this category. The following section considers
each type in more detail.
Black-Boxing
‘Black-Boxing’ makes use of an entirely systema-
tized method that requires little or no interpretative
activity. Chartists in this group do not commit sig-
ni?cant resources to developing their skills as chart-
ists, and instead rely on simple analysis provided by
others. The interpretative system need not be based
on the investor’s computer; two of the interviewees,
Albert and James, subscribe to a newsletter which
provides identi?ed trends. Albert, who said he was
a ‘busy family man’ stated that the simplicity and
convenience of this approach means that it is the
best method he has found in 15 years of investing.
All the same, he does not buy without performing
his own checks, watching the recommended shares
on a simple end of day charting programme. At this
point
‘if the graph is going like the north face of the
Eiger I may well buy, and I’ll hold those until
212 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
they drop. I only buy shares when they’re going
up’ (Albert)
George, Nigel and Mike all use simple systems in
combination with fundamental analysis. George
uses the same programme as Albert. Nigel makes
use of one particular indicator, the RSI, before
making a purchase, although he has only a rough
idea of what this indicator does, while Mike, who
tends to invest in smaller companies on the basis
of share tips, also makes a casual reference to the
charts before placing his order. Mike stated that if
he spots a trend forming – when a share suddenly
starts to ‘tip up’ – he will invest.
Accessibility and ease of use is taken to an
extreme by ‘black box’ programs which automati-
cally process data through a pre-programmed algo-
rithm and generate investment recommendations,
which can be transmitted directly to a broker. How-
ever, although this kind of software was seen on
sale, and interviewees remarked that it existed, none
of the interview or survey respondents used this
kind of software. It may be suggested that even
among the Black-Boxing chartists, those who are
prepared to place their investment decisions entirely
in the hands of an automated system are relatively
rare.
Number-Crunching
Trend-seeking chartists who take decisions into
their own hands by constructing their own systems
and methods provided much richer and more
involved descriptions than their peers in the ?rst cat-
egory . Number-Crunching uses a barrage of devices
for spotting trends; collectively termed ‘indicators’,
these range from the simple moving average, to
the recherche´ Ichimoku (a Japanese technique that
draws a cloud beneath or above a stock chart, and
from which the user can deduce the existence of a
trend). Simon described his initial method as
follows:
The ?rst thing that I actually looked at was a
breakout system. . .you’d buy a stock when it
went through the 20 day moving average and
keep moving the stop[loss] up, and sell it when
it went through the 50 day moving average, and
the idea is to capture the sort of wave move-
ment. . . (Simon)
While this method proved unsatisfactory because
small-scale ?uctuations in stock prices kept trigger-
ing the stop-loss, it served as a basis for Simon to
start developing more complex methods. Building
new and testing techniques, usually through a brico-
lage of existing methods, is a crucial part of this
group’s charting activity.
A particularly important tool for this process is
‘back-testing’. This process involves the chartist
running a retrospective simulation with historical
price data to ascertain whether a particular combi-
nation of indicators would have identi?ed a su?-
cient number of trends. The process is usually
Art
(interpretative)
System (non-
interpretative)
Trend-seekers
(order-less market)
Pattern-seekers
(ordered market)
Self built calculative systems;
commercial systems running
self assembled indicators;
mathematical (curve-fitting)
back-testing
(Simon, Robert, Tony)
Programme trading (black box
systems); commercial systems
running pre-assembled indicators;
purchased analysis; no back-
testing
(Albert, George, James, Mike,
Nigel)
Pattern search by eye; creative
visualisation; pattern-seeking
back-testing
(Max)
Automated pattern search via
commercial packages;
Fibonnaci/Eliott computation
programmes; proprietary
methodologies; no back-testing
(Chris, Mickey,Terry)
‘Number-Crunching’
‘Black-Boxing’
‘Chart-Seeing’
‘Wave-Surfing’
Fig. 1. A taxonomy of charting styles.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 213
computerised and back-testing functionality is an
important part of more complex charting software.
Robert, discussing his most recent experiment, gave
an example of this process
I can look for shares on the weekly chart where
the 14 day RSI is crossing over 50 and the slope
of the 40 week exponential moving average is
greater than a certain value, and the slope of
the 10-week moving average is greater than
another value, or certain other conditions, and
so on. . . I can make it [the software package]
automatically put little buy and sell signals on
the charts, then I can go and look at those charts
and say to myself would those have been good
buy or sell signals, and if they’re not quite right
I can adjust them slightly. . . (Robert)
This process of testing and adjustment is
repeated until a retrospectively successful strategy
is discovered, at which point it will be applied in real
time. The intention of back-testing is to give the user
con?dence that a particular method will lead to the
discovery of successful trends by the demonstration
that it would have done so in the past. The back-
testing process is often operated alongside a dummy
portfolio, where new methods are tested on a real-
time basis, either without risking any money or by
using very small stakes. Tony and Simon both use
this method, with Simon operating a home-made
trading simulator constructed from an Excel spread-
sheet and downloaded end of day data.
These systems need not deliver a successful result
in every instance. Chartists will accept surprisingly
low probabilities of success in their simulations;
Robert, for example, is aiming at a success rate of
just over 50%. This is acceptable because of the inte-
gral role that the stop loss (an order placed at the
brokerage to automatically exit a position should
the share price move by a certain amount against
the trade) plays in the strategies of these chartists.
Chartists view the stop loss as a means of increas-
ing the probability of success across their trades; by
setting the stop loss relatively tightly they improve
their probability of success. Against this, many
chartists argue that stop-losses can be automatically
triggered by minor moves in the market causing
positions to be exited at a loss when, had they been
held longer, they would have resulted in a pro?t.
Simon encountered this in his early methods and
confessed that it had presented a serious problem.
He ascribes the phenomenon to market volatility.
Other interviewees (Tony, Chris, and Mickey)
argued that stop-losses are manipulated by those
professional players with the power to move prices,
who force the market down in order to trigger auto-
matic sales and buy up stock cheaply. References to
an imagined and malevolent other are reminiscent
of the ‘spoofer’ identi?ed by Zaloom (2003), and
also recall Mayall’s category of charting as a com-
bat or game. Anecdotal evidence indicates that the
practice of searching out and triggering stop-losses
does indeed take place.
4
The stop loss is therefore a crucially important
device for chartists, a fact which has not been fully
recognised in the ?nance literature. Park and Irwin
(2004) discuss two early studies based on a stop-loss
?lter role, which showed the method to substantially
outperform a buy-and-hold strategy, but there
appears to be no work combining stop-losses with
more complex charting strategies. Number-Crunch-
ing strategies highlight just how di?cult it might be
to test this. As Batchelor and Ramyar (2006) note,
the individual strategies of chartists are vague, com-
plex, and badly documented, but it is precisely this
variety of indicator bricolage, overlaid with per-
sonal, often unrecognised, interpretative activity
on the part of the user, that appears to give Num-
ber-Crunching its distinctive character.
In summary, the Number-Crunching chartists
make creative use of their resources to try and
develop ever more accurate means of spotting price
trends in the market. In this category, and indeed in
the case of Black-Boxing, the charts are heuristic
devices, providing a way of seeing the market (May-
all, 2006) and of managing market dynamics. The
charts provide a means of searching through an
inaccessibly large universe of stocks, so that a ?nal,
often visual interpretation can be made. As Tony
said:
The charts will tell me what I need to know, and
if it looks interesting, either because it’s going up
or going down, whether it’s worth investigating
further. (Tony)
In Number-Crunching and Black-Boxing calcu-
lative strategies charts are therefore devices for
managing market chaos, seeking out and visualising
investment opportunities, and reinforcing the calcu-
lative skills of their users. For the following two
strategies they are tools for revealing the hidden
4
Donald Mackenzie recalls a professional trader remarking
that on a quiet day he and his colleagues would ‘bounce the price
around’ to take out the stop-losses.
214 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
order in markets. The problem is once again that of
search, and again there are two clearly identi?able
strategies: automation, and a visual, art-based
method.
Wave-Sur?ng
Wave-Sur?ng uses automated methods involving
proprietary methods and systems to search for spe-
ci?c patterns of price movements. In this category
are Chris, Mickey and Terry. Once patterns have
been discovered, the deterministic nature of number
series makes the investment decision relatively easy.
Having identi?ed a share that is demonstrating a
Fibonacci zigzag, the chartist can determine not
only the correct moment to invest, but the distance
that will be travelled in a given direction. They can
then use the stop loss to quantify potential risk and
arrive at a precise risk/reward measure on which to
base the decision to trade.
Wave-Sur?ng methods are more complex than
those of Black-Boxing – the other system-based cat-
egory – and are predicated on the existence of pat-
terns in the market. So ‘system’ here takes on a
wider de?nition to include not just the computer
software but also the theory and associated tech-
nique, which together form an integrated discursive
scheme, purchased in its entirety. Terry’s story
exempli?es the dedication and strong beliefs of the
Wave-Sur?ng chartists. Terry is convinced that
there is a code to be cracked in the markets and
has dedicated considerable expense, both in ?nan-
cial and personal terms, to learn charting skills.
He has spent several thousand pounds on purchas-
ing training CDs and charting software and attend-
ing courses. At the time of interview he had spent all
his spare time for nine months working on it and
intended going full-time after another four months.
Terry’s method was to attend a course or purchase a
training programme, and make use of the method
suggested until he decided that it was not accurate
enough, or that it failed in some other way, and then
repeat the process with a more complex method. He
had worked through Elliott waves, found to be too
inexact for short-term trading; then a method which
integrated Elliott and Fibonacci; then Delta. He
believed Delta to be inaccurate in the short term,
although he believes that the market does obey the
Delta cycles in the longer term, so he moved to a
system of Fibonacci laid over Delta; and ?nally a
method integrating other technical indicators with
Delta and Elliott which he thinks is capable of giv-
ing accurate short-term predictions of market
movements.
The Wave-Sur?ng process of investment search is
simpli?ed and automated by using software pro-
grammes, often sold in conjunction with training
materials or at expert seminars hosted by the soft-
ware manufacturer. These require substantial com-
puting power, usually two PCs and two or more
screens. When interviewed, Terry was using a soft-
ware programme developed by the presenter of the
course that he had most recently attended. The pro-
gramme is able, he said, to search for 9000 combina-
tions of Elliott wave structures. From a list of 300
shares, scanned overnight, it usually produces four
or ?ve potential trades to be assessed for risk and
return. Mickey also uses a charting package to
search for Fibonacci calculations, from which he
establishes points of support and resistance for mar-
ket turns. In practical terms, he says:
it’s dead easy . . . I select Fibonacci, and I click
once on a high point with my mouse and click
once on the low point and the lines are automat-
ically drawn. (Mickey)
Chris uses two PCs, each with two screens run-
ning two separate charting programmes. He scans
the market automatically every minute looking for
stocks that are approaching the bounds of trade cal-
culated by his Pivot formula. On spotting a suitable
candidate he checks for any outstanding bids or
o?ers, and on being satis?ed that the buying or sell-
ing that has caused this movement has run out, he
will take a position.
Chart-Seeing
All three of the ‘Wave-Surfers’ talked about oth-
ers who had managed to elevate these complex pro-
cedures into an intuitive act and who made their
judgements by eye. We term this calculative strategy
Chart-Seeing. Terry and Mickey seemed envious of
this ability, referred to as ‘intuition’ or a ‘native
skill’. They provided examples of investors they
admired who had this ability: in the case of Mickey
an eminent City analyst who scanned his charts by
eye on the train home, or in the case of Terry a
famous trader who over the course of 11 years has
become ‘attuned’ to the market and can ‘see
instantly’ where it is heading. Of the chartists inter-
viewed, only Max has attempted to learn or exercise
this skill. Like his Wave-Sur?ng peers, he began to
learn about charting by attending a number of
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 215
courses, where he was taught Fibonacci-based
(Wave-Sur?ng) methods. He tried these in the mar-
ket, and became increasingly frustrated as his losers
outnumbered his winners. Eventually, he came to
the conclusion that no system could su?ce in its
entirety, realising ‘you’ve actually got to lay on to
that [the most systematic of approaches] a feel for
what the markets are doing’. Being a relative begin-
ner, with less than a year’s trading experience, Max
embarked on a rigorous programme of back-testing
to develop his method of identifying patterns visu-
ally. He has ?nally reached the point where he scans
a universe of 200 stocks, limited by his chosen spe-
cialisation in US equities and his choice of broker,
?rst weekly and then daily to look for patterns
forming.
Max’s ‘feel for what the markets are doing’ is
based on interpretative activity. Interpretation is
crucial for Chart-Seeing. Observation conducted
as part of our research included a seminar given
by the famous trader mentioned by Terry above.
This trader suggested that his method was a
‘straight-forward modi?ed Delta pattern’ but as he
explained the method he employed a considerable
amount of interpretation as he showed his charts,
explaining how he had adjusted the model in
responding to various situations to make a pro?t.
Discussion
Our analysis of charting styles con?rms previ-
ous research that technical analysis encompasses
a variety of methods and techniques. Our classi?-
cation of these techniques provides an improved
understanding of variations in investor decision-
making and raises a number of important points
that are discussed in this section. This section
?nally suggests four propositions for testing by
future research.
The ?rst point to stress is the clear distinction
between the trend-seekers and the pattern-seekers.
This is based on fundamental di?erences in the
investors’ ontology of ?nancial markets. Trend-
seekers see the market as chaotic and unpredictable,
believing only that price trends have some persis-
tence. Pattern-seekers believe that any observed
chaos is only super?cial and beneath it the market
has hidden structures. Our interviews highlighted a
clear distinction; chartists identi?ed strongly with
either pattern-seeking or trend-seeking and ?rmly
believed themselves to be right and the other posi-
tion wrong. This distinction serves to clarify two
assumptions that have appeared in the literature.
The ?rst of these is the elision of charting with
momentum investing, and then of momentum
investing with the search for patterns in the market
(Buenza & Stark, 2004, p. 375). We show that not
all chartists are momentum investors, and that those
who are, the trend-seekers, do not search for pat-
terns. Park and Irwin’s (2004) claim that all charting
focuses on the search for trends is equally mistaken.
The second common misconception is that everyone
necessarily believes in an overarching, transcendent
order in the market (Mayall, 2006, p. 124). This
research clearly shows that this applies only to a sec-
tion of the charting community.
This distinction perhaps helps us in responding
to Menkho?’s (1997) challenge to justify charting
as a rational occupation. Some arguments can
be advanced for the rationality of trend-seeking.
Charts may serve as a rational search method
for a momentum strategy; the usefulness of
momentum investing has been demonstrated in
the US (Jegadeesh & Titman, 2001), and in the
UK (Liu, Strong, & Xu, 1999; Tonks & Hon,
2003). It is also rational to act in a particular
way in the expectation that others will do the
same. Models of herding (Banerjee, 1992; Bikh-
shandani & Sharma, 2001; Hirshleifer & Teoh,
2003) and of asset price bubbles (De Long
et al., 1990) demonstrate this. Jegadeesh (2000)
suggests that charting methods are used in con-
junction with other investment selection methods,
paralleling the activities of the incidental chartists,
although this was clearly not the case with the
other chartists studied.
It is more di?cult to make a case for rationality
among the pattern-seekers. The arguments above
hinge on the assumption that certain charting styles
may have some value as investment methods, and
the same can be said here only in a limited way.
While some numbers do have support and resis-
tance e?ects (Donaldson & Kim, 1993), o?ering
some justi?cation for methods based on important
numbers, more sophisticated pattern systems can
claim no such evidence. Batchelor and Ramyar
(2006) demonstrations of the absence of Fibonacci
support levels is particularly problematic, while
Malkiel (1996) and Jegadeesh (2000) are both insis-
tent that there is no reason for patterns to form in
stock prices.
Psychological biases may also help to account for
the popularity of charting. It can be argued that sys-
tematic overcon?dence (Barber & Odean, 2001;
216 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
Daniel et al., 2001) may explain the con?dence that
individuals show in their own methods and their
willingness to invest on the basis of these, even in
the face of poor investment performance. The charts
themselves, as a psychologically available means of
presenting decisions in a visual format (Kahneman
& Tversky, 1974), may add to the appeal of the
investment technique, the format making decisions
easy and accessible. Finally the pursuit of trends
has a demonstrated psychological appeal (Andreas-
sen & Kraus, 1998) and this too may contribute to
the popularity of charting.
Our analysis suggests that it is mistaken to regard
charting as a means for achieving above market
returns. Finance research provides scant support
for its ability to do so, as do the testimonies of sev-
eral of the chartists. Yet charting is popular among
non-professional investors. Menkho?’s challenge
can, perhaps, be answered in another way. The
research programme of SSF directs attention to
the calculative strategies and technologies of indi-
vidual market agents; Hutchins (1995) shows how
sophisticated processes of calculation can be distrib-
uted among human actors and technological
devices. Buenza and Stark (2004) show this distribu-
tion in action in an investment bank’s trading room.
Charting is amenable to this analysis, as calculation
is distributed to computer programmes and their
designers, to newsletter editors, to fellow chartists,
to imagined others such as the ‘spoofer’, or the
malevolent professional traders making fun of
stop-losses. The charting apparatus in each of the
four categories above serve to produce the market,
to make it visible for individual investors working
alone in their homes, and to allow interaction
between them and the market. In each instance,
the charting apparatus serves to narrow the range
of decisions presented to the individual. Instead of
the unfathomable possibilities of the market, possi-
bilities are constrained, and decisions framed and
disentangled (Callon, 1998). In this analysis, the
importance of charting lies less in its e?cacy as an
investment selection method, and more in its power
to make markets visible and accessible and to sim-
plify and support the huge burden of calculative
activity that falls on individual market agents.
Charting may, in this sense, be considered rational,
as it allows individual non-professional investors to
participate as market actors. This is a theme that
has been strongly articulated across all four catego-
ries: Black-Boxing is quick and e?cient; the com-
plex calculative technologies of Wave-Sur?ng
allow glimpses of the order behind market chaos;
Number-Crunching assembles indicators to pick
out the trends which it seeks; and Chart-Seeing pre-
sents the individual with available raw materials for
visual analysis.
We attempted to gauge the extent to which indi-
viduals did manage to beat the market. However, in
interviews it is di?cult to gain more than subjective
estimates of individual success. The problem is exac-
erbated by the fact that interviewees tended to com-
pare their returns to their starting point, rather than
the market, and that data collection took place dur-
ing a prolonged bull-market. Where possible, inter-
viewees’ returns on investment were ascertained and
classi?ed as high (above market), medium (in line
with market) and low (below market). This is shown
in Table 2.
When considering charting purely as a means of
gaining above market returns, our initial discussion
provided more support for the rationality of trend-
seekers than pattern-seekers and therefore trend-
seekers might be expected to show higher returns
than pattern-seekers. However, we failed to ?nd evi-
dence to support this expectation; some pattern-
seekers were very successful, and some trend-seekers
were not. Additionally, our analysis indicated that
art-based charting styles have a higher status than
system-based ones, and therefore it might be
expected that art-based charting styles would be
associated with higher returns. Again, there was
no evidence to support this. Overall, there was no
evidence to indicate that there was an association
between charting style and returns on investment;
instead our analysis suggested that returns may be
associated with the interpretative skills and activi-
ties of the individual chartist. Therefore, our ?rst
proposition is as follows:
Table 2
Charting style and success level
Pseudonym Charting style Success
Albert Black-Boxing High
George Black-Boxing High
James Black-Boxing High
Tony Number-Crunching High
Chris Wave-Sur?ng High
Nigel Black-Boxing Medium
Max Chart-Seeing Medium
Simon Number-Crunching Medium
Terry Wave-Sur?ng Medium
Mike Black-Boxing Low
Robert Number-Crunching Low
Mickey Wave-Sur?ng Low
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 217
P1: There is no association between charting style
and the ?nancial return on investments.
The lack of association between charting style
and ?nancial return appears to underscore the cen-
tral claim of this article – that, while the charts form
part of the toolkit of heuristic devices that the indi-
vidual investor uses to manage and control the mar-
ket, individual investment performance is largely
dependent on the tacit skills of the chartist. The
clearest example of this was given in a seminar we
attended, where a chart-seeing professional trader
set out a theory of markets based on a four day
cycle, illustrated by charts. The charts frequently
bore no clear relation to either the cycle’s predic-
tions, or the trader’s comments, and instead
appeared to o?er a heuristic support for the trader’s
own tacit interpretative skills, developed over a long
period of time. In this sense we can perhaps answer
Menkho?’s challenge: it is entirely rational for an
individual to use tools that support their activities,
where the tools serve productively to support calcu-
lative activity but not determine the eventual invest-
ment decision. In this instance, we can see that there
may be a distinction between the ‘art’ and ‘science’
categories of chartist, but the argument can equally
be made that calculation is distributed (Hutchins,
1995) to the compilers of newsletters or charting
software, who themselves rely on charts as heuristic
devices.
There are parallels here with discussion of valua-
tion models used by professional securities analysts.
Barker (1999) and Demirakos et al. (2004) show
that analysts consistently prefer simple models such
as price to earnings ratio and dividend yield to
sophisticated models of discounted cash-?ow and
residual income, the latter considered to be too
complicated for real life situations. Barker identi?es
a need for further research in understanding why
simple models are preferred; the implication of
chartists’ experience is that the analysts’ tacit exper-
tise is of primary importance when grappling with
the complexities of real life company valuation,
and the simpler ratios form a more useful, ?exible
and available heuristic device in dealing with this
process.
Our analysis also highlighted the importance of
the stop-loss in the investment strategies of chart-
ists. The stop-loss limits the potential downside of
trades, but may also lead to closing a trade too early
and crystallizing a loss rather than achieving an
eventual pro?t. Further research is required to
establish the e?ect of stop-losses on the pro?tability
of investment strategies. Hence our second
proposition:
P2: There is a positive association between the use
of stop-losses and investment returns.
As with P1, this proposition has another implica-
tion for future research. We suggest that the stop-
loss often performs a markedly di?ering function
for individual chartists than is supposed in ?nance
literature. Rather than being simply another part
of the chartist’s tool kit, it appears that the utility
of the stop-loss is closely related to the signi?cance
of the charts as heuristic devices for investors. The
stop-loss forms an action-based ruling on the valid-
ity of any given trading set up, informing the chart-
ist at an early stage whether he is ‘right’ in his
combination of rules and indicators. Terry alludes
to this with the comment that the stop-loss resem-
bles a ‘slap on the wrist’ that will ‘invalidate your
analysis’. In this sense the power of the stop-loss lies
in its ability as a corrective device for ill-conceived
conceptual arrangements on the part of the individ-
ual, not only improving investment returns but also
continuously enhancing the chartist’s heuristic
apparatus and potentially improving her investment
performance. However, this aspect of the stop-loss
is more di?cult to test than its relationship to
pro?tability.
Our analysis also indicated that there was rela-
tively little movement between charting styles. This
is particularly surprising in the case of the art/sys-
tem divide. As art has a higher status than system,
it might be expected that there would be a progres-
sion from system to art. This was certainly the case
with Max, who moved to a visual style of pattern
spotting after an unsuccessful attempt at a system-
based approach. Apart from this there is little
mobility among our interviewees and it appears that
the choice between system and art-based charting
styles is associated with personal circumstances
and preference. System-based styles are chosen for
reasons of convenience and practicality; its methods
save time and o?er the user an e?cient means of
sifting through large volumes of data. Art-based
styles are popular with those who enjoy experimen-
tation and are prepared to devote time and e?ort to
discovering new methods; Max has, by means of a
personal windfall, been given a period of time where
he could devote extensive e?ort to perfecting his
charting style. There is less conceptual support for
movement between trend-seekers and pattern-seek-
ers as this would require a shift in market ontology,
a change in the individual’s understanding of how
218 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
markets function. However, Chris moved from
‘Number-Crunching’ to ‘Wave-Sur?ng’, making an
ontological leap as he became convinced of the use-
fulness of number series and other predictive indica-
tors. This raises some interesting issues for further
investigation. Our third proposition, then, is
P3: Individuals’ charting styles are relatively static
and are determined by exogenous (non-market)
factors.
There are again parallels with the methods used
by professional securities analysts, where choice of
methodology appears to be in?uenced by the indus-
try sector, type of ?rm, and expectations of the mar-
ket (Barker, 1999; Block, 1999; Demirakos et al.,
2004). It may be the case here too that the evidence
of chartists can illuminate new avenues of research
on professional analysts; perhaps truly exogenous
factors such as personal preferences and skills, as
well as the circumstances and customs of employ-
ment, may be shown to in?uence valuation
methodologies.
Propositions one to three underscore the central
argument of this study – that the use of charting is
not solely associated with generating investment
returns. We have suggested that its power as a heu-
ristic device, enabling individuals to organize, man-
age and understand the complexities of the market
is equally, if not more, important to users. We noted
during interviews that the charts provided a frame
of reference for seeing, recalling and talking about
the market; interviewees would often refer to a chart
as they illustrated a point or substantiated a claim.
We also note that charting o?ers individuals other
non-market returns. The experimentation of the
Number-Crunching investors provided them with
enjoyment, pride, and a sense of mastery over the
market. Across all categories mastery of techniques
provided investors with prestige. In summary,
returns from charting are much broader than a sim-
ple return on investment. Hence, we suggest a ?nal
proposition for future investigation:
P4: Investors use charting for its heuristic power
rather than its ability to generate ?nancial returns.
Conclusion
In this article we have conducted an analysis of
charting styles, and classi?ed them into four groups
depending on whether they were art or system based
and whether they were seeking trends or patterns.
We argue that the visualisation of a ?nancial market
is inexorably linked to, if not driven by, chartists’
views on market ontology and e?ciency. Our taxon-
omy has shown how di?erences in seeing the mar-
ket, driven by fundamental di?erences in chartists’
conception of what the market is, relate to varied
calculative strategies. An investor who believes that
markets are e?cient has little need of this kind of
graphic visualisation, and will simply buy a tracking
fund. The purpose of charting appears to be the
opposite, providing investors with visual illustra-
tions of the structure of markets and showing the
opportunities for pro?t amongst market chaos. A
chartist, almost by de?nition, may not believe in
e?cient markets, whatever the evidence to the con-
trary; chartists who were less successful blamed
themselves for deviation from their methods, and
looked forward to making ‘big money one day’
(Chris). We did not interview, meet, or hear of, any-
one who was an ex-chartist.
Two major themes interlink to underpin the ?nd-
ings in this study. These themes are extremely
important because they a?ect the validity of
assumptions of some earlier research. The ?rst is
that the use of charting and the choice of charting
style do not appear to be associated with ?nancial
returns. The second is the importance of interpreta-
tive and tacit skills in the activities of chartists. We
have shown no clear relationship between charting
style and individual investors’ returns and have sug-
gested that returns may be dependent on individu-
als’ own interpretative skills. This presents
problems for researchers attempting to determine
the usefulness of charting, or its popularity with
investors, on the basis of textbook methods. While
the similarity between our interviewees’ accounts
and textbook presentations of charting method
shows that these are accepted as a basis for individ-
ual charting methods, it is clear that the strategies
pursued by individuals may in fact be very di?erent
from those methods. We therefore suggest that
attention could usefully be given to the way in
which individuals develop their own interpretative
strategies, and that it would be bene?cial for some
future research to focus on individual rather than
generic charting strategies.
There may also be implications for practitioners
such as investors, company managers and market
supervisors. In particular, practitioners should be
aware of the popularity of charting as an investment
method. Rather than dismissing it as irrational or
speculative, it may help practitioners to understand
that it is employed by relatively sophisticated inves-
tors, and appreciate that it may have its roots in
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 219
attempts to manage and make sense of the market
in a rational manner.
Finally, we suggest that previous studies that aim
to demonstrate or refute the e?cacy of charting
methods have only limited scope for developing
our understanding of the popularity of charting
among investors. This study has shown that the
appeal of charting lies in the way that it allows
investors to make sense of, manage, and participate
in the markets. As an integrated discursive scheme,
supplying a way of understanding ?nancial markets
and the tools to turn this understanding into invest-
ment practice, charting can be considered a perfor-
mative technique in the sense of Callon (1998,
2007). This aspect of the appeal of charting is one
which may be investigated most fruitfully by future
research.
Acknowledgements
The authors are very grateful to two anonymous
reviewers and Anthony G. Hopwood for providing
very positive and constructive comments. Opinions
(and the potential errors) expressed in this paper
are, of course, the authors alone.
References
Abola?a, M. (1998). Markets as cultures: An ethnographic
approach. In M. Callon (Ed.), The laws of markets
(pp. 69–85). Oxford: Oxford University Press.
Allee, K. D., Battacharya, N., et al. (2007). Proforma disclosure
and investor sophistication: External validation of experi-
mental evidence using archival data. Accounting Organiza-
tions and Society, 32(3), 201–222.
Andreassen, P., & Kraus, S. (1998). Judgemental prediction by
extrapolation. Mimeo, Harvard.
Baker, W. (1984). The social structure of a national securities
market. American Journal of Sociology, 89(4), 775–811.
Banerjee, A. (1992). A simple model of herd behaviour. Quarterly
Journal of Economics, 107(3), 797–817.
Barber, B., & Odean, T. (2000). Trading is Hazardous to your
Wealth: The common stock performance of individual
investors. Journal of Finance, 55(2), 773–799.
Barber, B., & Odean, T. (2001). Boys will be boys: Gender,
overcon?dence and common stock investment. Quarterly
Journal of Economics(February), 261–292.
Barber, B. M., & Odean, T. (2004). All that glitters: The e?ect of
attention and news on the buying behaviour of individual and
institutional investors. University of California.
Barker, R. G. (1999). Survey and market-based evidence of
industry-dependence in analysts’ preferences between the
dividend yield and price-earnings ratio valuation models.
Journal of Business Finance & Accounting, 26(3–4), 393–418.
Batchelor, R., & Ramyar, R. (2006). Magic numbers in the Dow.
London: Cass Business School.
Bikhshandani, S., & Sharma, S. (2001). Herd behaviour in
?nancial markets. IMF Sta? Papers, 47(3).
Block, S. B. (1999). A study of ?nancial analysts: Practice and
theory. Financial Analysts Journal, 55(4), 86.
Bloom?eld, R., Libby, R., et al. (1999). Con?dence and the
welfare of less-informed investors. Accounting Organizations
and Society, 24(8).
Buenza, D., & Stark, D. (2004). Tools of the trade: The socio-
technology of arbitrage in a Wall Street trading room.
Industrial and Corporate Change, 13(2), 369–400.
Callon, M. (1998). The embeddedness of economic markets in
economics. In M. Callon (Ed.), The laws of markets
(pp. 1–58). Oxford: Oxford University Press.
Callon, M. (2007). What does it mean to say that economics is
performative?. In D. MacKenzie F. Muniesa, & L. Siu (Eds.),
Performing economics. Princeton University Press.
Callon, M., & Muniesa, F. (2003). Les marche´s e´conomiques
commes dispositifs collectifs de calcul. Re´ seaux, 21(122),
189–233.
Daniel, K., Hirshleifer, D., et al. (2001). Investor psychology in
capital markets: Evidence and policy implications. Journal of
Monetary Economics, 49, 139–209.
De Bondt, W. (1998). A portrait of the individual investor.
European Economic Review, 42, 831–844.
De Long, J., Shleifer, A., et al. (1990). Positive feedback
investment strategies and destabilising rational speculation.
Journal of Finance, 45(2), 379–395.
Demirakos, E. G., Strong, N. C., et al. (2004). What valuation
models do analysts use?. Accounting Horizons 18(4), 221.
Donaldson, R., & Kim, H. (1993). Price barriers in the Dow
Jones industrial average. Journal of Financial and Quantitative
Analysis, 28(3), 313–330.
Fama, E. (1970). E?cient capital markets: A review of theory and
empirical work. Journal of Finance, 25, 383–417.
Fama, E. (1991). E?cient capital markets II. Journal of Finance,
46(5), 1575–1617.
Frederickson, J. R., & Miller, J. S. (2004). The e?ects of proforma
earnings disclosures on analysts’ and non-professional inves-
tors’ equity valuation judgements. The Accounting Review,
79(3), 667–686.
Gar?nkel, H., & Sachs, H. (1970). On the formal structure of
practical actions. In J. McKinney & E. Tiryakian (Eds.),
Theoretical sociology (pp. 337–366). New York: Appleton
Century Crofts.
Glaser, B., & Strauss, A. (1968). The discovery of grounded
theory: Strategies for qualitative research. London: Weiden-
feld and Nicholson.
Godechot, O. (2001). Les traders. Paris: la De´couverte.
Hirshleifer, D., & Teoh, S. (2003). Herd behaviour and cascading
in capital markets: A review and synthesis. European Financial
Management, 9(1), 25–66.
Hutchins, E. (1995). Cognition in the wild. Cambridge, MA: The
MIT Press.
Jegadeesh, N. (2000). Foundations of technical analysis: Com-
putational algorithms, statistical inference, and empirical
implementation: Discussion. Journal of Finance, 60(4).
Jegadeesh, N., & Titman, S. (2001). Pro?tability of momentum
strategies. Journal of Finance, 56, 699–720.
Kahneman, A., & Tversky, D. (1974). Judgement under uncer-
tainty. Science, 185, 1124–1131.
Knorr Cetina, K. (2005). From pipes to scopes: The ?ow
architecture of ?nancial markets. In A. Barry & D. Slater
220 P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221
(Eds.), The technological economy (pp. 122–141). Abingdon:
Routledge.
Knorr Cetina, K., & Bruegger, U. (2000). The Market as an
object of attachment: Exploring postsocial relations in ?nan-
cial markets. Canadian Journal of Sociology, 25(2), 141–168.
Knorr Cetina, K., & Bruegger, U. (2002). Global microstruc-
tures: The virtual societies of ?nancial markets. American
Journal of Sociology, 107(4), 905–950.
Lewis, M. (1989). Liar’s Poker. London: Coronet.
Liu, W., Strong, N., & Xu, X. (1999). The pro?tability of
momentum investing. Journal of Business Finance and
Accounting, 26, 1043–1091.
Locke, K. (2001). Grounded theory in management research.
London: Sage Publications.
Lo, A. W., Mamaysky, H., et al. (2000). Foundations of technical
analysis: Computational algorithms, statistical inference, and
empirical implementation. The Journal of Finance, 55(4),
1705–1765.
Lynch, M. (1993). Scienti?c practice and ordinary action. Cam-
bridge: Cambridge University Press.
MacKenzie, D. (2003a). An equation and its worlds. Social
Studies of Science, 33(6), 831–868.
MacKenzie, D. (2003b). Long term capital management and the
sociology of arbitrage. Economy and Society, 32(3), 349–380.
MacKenzie, D. (2004). The big bad wolf and the rational market:
Portfolio insurance, the 1987 crash and the performativity of
economics. Economy and Society, 33(3), 303–330.
MacKenzie, D., & Millo, Y. (2003). Constructing a market,
performing theory: The historical sociology of a ?nancial
derivatives exchange. American Journal of Sociology, 109(1),
107–145.
Malkiel, B. G. (1996). A random walk down wall street. New
York: W.W. Norton.
Malkiel, B. G. (2003). The e?cient market hypothesis and its
critics. The Journal of Economic Perspectives, 17(1), 59–82.
Mayall, M. (2006). Seeing the market: Technical analysis in
trading styles. Journal for the Theory of Social Behaviour,
36(2), 119–140.
Menkho?, L. (1997). Examining the use of technical currency
analysis. International Journal of Financial Economics, 2,
307–318.
Miles, M., & Huberman, A. (1994). Qualitative data analysis.
London: Sage Publications.
Odean, T. (1998b). Are investors reluctant to realize their losses?.
Journal of Finance 53(5).
Park, C.-H., & Irwin, S. H. (2004). The pro?tability of technical
analysis: A review. AgMAS Project Research Report. Uni-
versity of Illinois at Urbana-Champaign: 04.
Preda, A. (2007a). The sociological approach to ?nancial
markets. Journal of Economic Surveys, 21(3), 506–533.
Preda, A. (2007b). Where do analysts come from? The case of
?nancial chartism. In M. Callon, F. Muniesa, & Y. Millo
(Eds.), Market devices. Oxford: Blackwell.
Shliefer, A., & Summers, C. (1990). The noise trader approach to
?nance. Journal of Economic Perspectives, 4(2), 19–33.
Taylor, M., & Allen, H. (1992). The use of technical analysis in
the foreign exchange market. Journal of International Money
and Finance, 11(3), 304–314.
Tonks, M. T., & Hon, I. (2003). Momentum in the UK stock
market. Journal of Multinational Financial Management,
13(1), 43–70.
Webb, E., Campbell, D., et al. (1965). Unobtrusive measures.
Chicago: Rand McNally.
Zaloom, C. (2003). Ambiguous numbers: Trading technologies
and interpretation in ?nancial markets. American Ethnologist,
30(2), 258–272.
Zaloom, C. (2006). Out of the pits. Chicago: University of
Chicago Press.
Zuckerman, E. (2004). Structural incoherence and stock market
activity. American Sociological Review, 69, 405–432.
P. Roscoe, C. Howorth / Accounting, Organizations and Society 34 (2009) 206–221 221
doc_970746787.pdf