INTERNATIONAL JOURNAL OF FINANCE AND ECONOMICS Int. J. Fin. Econ.

6: 8193 (2001)

IMPORTANCE OF TECHNICAL AND FUNDAMENTAL ANALYSIS IN THE EUROPEAN FOREIGN EXCHANGE MARKET
THOMAS OBERLECHNER* Webster Uni6ersity, A -1220 Vienna, Austria

ABSTRACT This article presents findings of a questionnaire and an interview survey on the perceived importance of chartist/technical and fundamental analysis among foreign exchange traders and financial journalists in Frankfurt, London, Vienna, and Zurich. Results confirm that most traders use both forecasting approaches, and that the shorter the forecasting horizon, the more important chartist/technical analysis is. Financial journalists put more emphasis on fundamental analysis than do foreign exchange traders. Results indicate that the importance of chartism may have increased over the last decade. Regarding the use of chartist/technical and fundamental analysis on seven forecasting horizons, four distinct clusters of traders can be identified. Forecasting styles and the overall importance attached to fundamental versus chartist/technical analysis vary across different trading locations. Foreign exchange traders mention a series of psychological motives and consequences of the use of chartism. Copyright 2001 John Wiley & Sons, Ltd.
KEY WORDS: foreign

exchange; fundamental analysis; psychology; technical analysis

As in all financial markets, a central question in the foreign exchange market is how market participants and currency traders forecast future market developments. Foreign exchange market participants are often classified according to two different forecasting approaches, fundamental and chartist/technical (Frankel and Froot, 1986, 1990). According to this distinction, fundamentalists are market participants who predict exchange rates by analysing the underlying economic conditions upon which they assume changes in exchange rates rest. Chartists, in contrast, study only the exchange rate movements themselves and believe that the history of previous data provides indicators for future exchange rates. Chartists employ technical analysis, i.e. the analysis of past exchange rate movements to guide forecasts and trading decisions in the foreign exchange market (Neely, 1997). Chartism includes the visual search for repeated patterns of data across time. Recent attempts to formulate an alternative to the efficient markets approach and to provide a more complete answer as to how expectations in financial markets are formed, differentiate between arbitrageurs, i.e. rational smart-money investors, and noise traders. Noise is defined as processes leading to shadow information, i.e. market mechanisms which blur observations of the market (Black, 1986). Arbitrageurs and noise traders, however, do not transcend the categories of fundamentalists and chartists described above. Fundamental analysis can be seen as the basis for forecasts and decisions by arbitrageurs (Menkhoff, 1998). Technical analysis can be seen as an example for noise trading since its strategies build on noise or popular models and not on information (Shleifer and Summers 1990, p. 24). The basic distinction between chartists and fundamentalists has served as basis for a demonstration of the existence of heterogeneous market expectations in the foreign exchange market (Frankel and Froot, 1990). Particularly in the short term, efforts to determine future exchange rates on the basis of economic fundamentals have had little success (Harvey, 1996). MacDonald and Taylor (1992) observe that models which are based on a theory of economic fundamentals alone do not offer sufficient explanations of short-term exchange rate movements. Also, Frankel and Froot (1990) note that standard macroeconomic
* Correspondence to: Webster University Vienna, Berchtoldg. 1, A-1220 Vienna, Austria. Tel.: + 43 1 2699293; fax: + 43 1 269929313; e-mail: oberlech@webster.ac.at

Copyright 2001 John Wiley & Sons, Ltd.

82

T. OBERLECHNER

analysis, for example, of growth rates or of trade numbers, cannot predict most short-term exchange rate changes. For example, the value of the dollar in 1984/1985 dramatically departed from what fundamental expectations alone would suggest. In response to the evidence that fundamentals alone do not suffice to explain exchange rate behavior, theoretical models that consider the interaction of chartist and fundamentalist expectations have been developed (e.g. Levin, 1997; Vigfusson, 1997). Despite the growing awareness of non-fundamental factors in the expectations and decisions of foreign exchange participants, there is a shortage of empirical data on the use of technical analysis in financial markets. A small number of previous empirical studies on this topic find that the relative importance of fundamental and technical forecasting methods in the foreign exchange market depends on the trading time horizon assumed. Chartism is used mostly for short-term forecasts in the London foreign exchange market (Allen and Taylor, 1990), a finding confirmed for Hong Kong by Lui and Mole (1998). Menkhoff (1997) finds that also in the German foreign exchange market chartism is used extensively. Previous studies have also shown that the relative importance of chartist and fundamental analysis in the market is subject to a process of change. For example, Frankel and Froot (1990) show that in the period from 1978 to 1988 foreign exchange forecasting services surveyed by Euromoney magazine demonstrate a notable shift in the kind of forecasting techniques used. Whereas during 1983 1985 the percentage of forecasting services employing technical analysis reached a maximum, this percentage then decreased slightly in the three subsequent years. This article attempts to extend the results of previous surveys on the importance of chartism and fundamentalism among foreign exchange traders in London (Taylor and Allen, 1992) and in Hong Kong (Lui and Mole, 1998) to a new geographic location. This article is the first to survey how foreign exchange traders form currency forecasts in different European trading locations, i.e. in Frankfurt, London, Vienna, and Zurich. Since the data for this research was gathered simultaneously in different European trading locations, a direct analysis of local differences in forecasting methods used in different geographic locations of the foreign exchange market is possible. To my knowledge, this is also the first study which determines the importance of different kinds of actors in the market, financial journalists, attribute to chartism and fundamentalism for the foreign exchange market. Finally, for the first time a more differentiated typology of foreign exchange traders based on their forecasting styles which takes different forecasting preferences on several forecasting horizons into account is established. A statistical cluster analysis of traders forecasting approaches on various time horizons goes beyond the simple dichotomy of fundamental and chartist traders and allows for a description of four distinct forecasting styles across time, chartist, ascending, fundamental, ascending, constant chartist, and inverse middle.

1. METHOD The results presented in this article are based on a questionnaire and an interview survey conducted in the European foreign exchange trading centers of Frankfurt, London, Vienna and Zurich.1 The data was collected in spring 1996 from members of three different kinds of organizations in the foreign exchange market: commercial and investment banks, central banks, and financial news providers. A total of 600 foreign exchange trader questionnaires were distributed among the foreign exchange traders of the three largest (in terms of their foreign exchange operations) commercial banks in Austria, Germany, Switzerland and UK.2 Since London is the capital of global foreign exchange trading, a number of other commercial and investment banks in the UK were also included. These additional British banks were included in order to generate more balanced numbers of traders in the comparison of foreign exchange trading in the UK versus in continental Europe. Of the 600 trader questionnaires distributed to these banks and to the participating central banks of Austria, Switzerland and UK, 321 were returned, resulting in a return rate of 54%. Of the responding traders, 18% work in Austria, 12% in Germany, 38% in Switzerland and 32% in the UK. The sample of participating foreign exchange traders includes 92% commercial or investment bank traders, 5% central bank traders, and 3% traders who indicated that they work for a different institution. A large part of the sample consists of senior traders (75%) and a smaller
Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

TECHNICAL AND FUNDAMENTAL ANALYSIS IN EUROPE

83

part of junior traders (22%) and foreign exchange trainees (3%). Fifty-one percent of the responding traders identified themselves as spot traders, 13% as forward traders, 8% as money market traders, 8% as derivatives traders, and 19% as some combination of the above. The sample consists of 91% male and 9% female traders. A total of 200 financial journalist questionnaires were distributed to journalists of the largest financial news providers operating in Europe.3 Fifty-nine questionnaires were returned, resulting in a return rate of 30%. The sample of participating financial journalists includes 64% wire journalists, 14% financial television journalists, 19% daily financial news journalists, and 3% journalists working for financial periodicals. The responding financial journalists work as editors or sub-editors (32%), reporters (56%), or other financial news journalists (12%). The sample consists of 74% male and 26% female journalists. In the questionnaires, traders and journalists were asked background information about their professional and personal status, including the kind of institution worked for, location of the institution, professional rank, work area, age, years of work experience, and gender. In a later part of the questionnaire, participants were asked to evaluate the importance they attach to chartist versus technical analysis on a series of forecasting horizons ranging from intraday to longer than 1 year. The survey also included questions about the traders view on the importance of feelings and of rationality in the foreign exchange market, and in their own decisions in the foreign exchange market. In addition to the questionnaires, confidential semi-structured interviews on selected aspects of the foreign exchange market were conducted with 58 experts from both the trading and the news reporting side. Forty-two bankers in foreign exchange (treasurers, heads of trading divisions, senior trading managers and senior traders) and 16 financial journalists (editors, sub-editors, and senior journalists) in Austria, Germany, Switzerland and UK were interviewed. Interviews were recorded, and were transcribed verbatim after the interview. In the present article, only selected data from these interviews will be quoted in the conclusion in order to contextualize the quantitative findings from the questionnaires. Questionnaires and interviews were designed with the help of foreign exchange experts from the trading and financial journalism sides. In order to ensure high return rates of the questionnaire, an internal contact person assisting in the process of questionnaire distribution was assigned to each participating organization. After the distribution of the questionnaires by the contact person, participating traders and journalists completed these questionnaires, and sealed their answers in attached envelopes. All participants were assured that their answers given in the questionnaires and in the interviews were anonymous, and that the quantitative data would be analysed only on group and not on individual levels. Participants had the choice to return their questionnaires through their institutions contact person or by mailing back the questionnaire directly. In the instructions, participants were asked to fill out the questionnaire individually.

2. RESULTS Table 1 lists the average importance ratings of fundamental analysis versus chartist/technical analysis for foreign exchange traders and for financial journalists on a series of seven forecasting horizons ranging from intraday to longer than 1 year. A multivariate two-factorial ANOVA of these ratings shows that both rater type (i.e. foreign exchange trader or financial journalist; F = 22.38, p B 0.001) and forecasting horizon (F = 75.97, p B 0.001) have a statistically highly significant influence on the perceived importance attached to fundamental analysis and chartist/technical analysis. The interaction effect between both factors was not significant (F = 0.36, p = 0.91). On six of the seven investigated individual forecasting horizons (intraday, 1 week, 1, 3, 6 months, and 1 year), foreign exchange traders attribute a significantly larger role to chartism than do financial journalists, as the results of univariate F -tests presented in Table 1 reveal. Only on the longest forecasting horizon investigated, i.e. on forecasts longer than 1 year, merely a statistically not significant tendency in this same direction can be observed (F = 2.60, p = 0.11).
Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

84

T. OBERLECHNER

Table 1. Importance attached to chartist/technical versus fundamental analysis over different forecast horizons by foreign exchange traders and financial journalists Foreign exchange traders Financial journalists (n = 282) (n = 49) M Intraday 1 week 1 month 3 months 6 months 1 year Longer than 1 year 3.47 3.77 4.40 5.06 5.67 6.35 6.79 S.D. 2.36 1.90 1.71 1.77 2.01 2.50 2.92 M 4.53 4.78 5.49 6.29 6.90 7.37 7.51 S.D. 2.71 2.47 2.07 1.85 1.99 2.32 2.75 8.14** 10.54** 15.91*** 19.83*** 15.52*** 7.07** 2.60 Univariate F -tests

Scale: from 0 = pure chartist analysis to 10 = pure fundamental analysis. M, mean, S.D., standard deviation. Level of significance: ** pB0.01, *** pB0.001.

Figure 1 visualizes that the shorter the forecasting horizon, the more important chartism is both for traders and for financial journalists. On increasingly longer forecasting horizons, both groups assign progressively more importance to fundamental analysis. On all forecasting horizons examined, financial journalists place more emphasis on fundamental analysis than do foreign exchange traders. Accordingly, the forecasting horizon where fundamental analysis becomes more important than chartist analysis in predicting exchange rate movements occurs earlier for financial journalists (forecasts of 1 week to 1 month) than for foreign exchange traders (forecasts of 13 months). On rating scales from 1 = only feelings to 7 = only rationality, foreign exchange traders and financial journalists were then asked to indicate the importance of feelings versus rationality in the foreign exchange market, and in their own decisions in the foreign exchange market. Results of a 2 2 ANOVA presented in Table 2 show that foreign exchange traders and financial journalists also differ in their perceptions of the role of feelings versus rationality (F = 47.08, p B 0.001). Univariate F -tests demonstrate that traders and financial journalists both rate feelings to be slightly more influential than rationality in the foreign exchange market and that their views of the foreign exchange market do not

Figure 1. Importance attached to chartist/technical versus fundamental analysis on seven forecast horizons by foreign exchange traders (n = 282) and by financial journalists (n = 49). Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

TECHNICAL AND FUNDAMENTAL ANALYSIS IN EUROPE

85

Table 2. Perceived importance of feelings versus rationality in the foreign exchange market and in own foreign exchange decisions by foreign exchange traders and by financial journalists Importance of feelings versus rationality Foreign exchange traders (n = 314) M In foreign exchange market In own decisions 3.37 3.46 S.D. 1.02 1.19 Financial journalists (n = 57) M 3.47 5.02 S.D. 1.09 1.28 0.53 81.83*** Univariate F -tests

Scale: from 1 = only feelings to 7 = only rationality. M, mean; S.D., standard deviation. Level of significance: ***p B0.001.

differ from each other. However, foreign exchange traders and financial journalists hold highly significantly different views regarding the role of feelings and rationality in their own decisions. Whereas traders consider feelings to be more influential than rationality in their own trading decisions, financial journalists consider rationality more important than feelings in their own reporting decisions. In the foreign exchange traders views, both of the collective foreign exchange market and of their own decisions, feelings are slightly more important and there is no difference between these two views (t = 1.42, p = 0.16). Financial journalists too consider feelings to be slightly more influential than rationality in the foreign exchange market but show a highly significant difference between this more emotional view of the collective market and a more rational view of their individual reporting decisions (t = 7.42, p B 0.001). In order to establish individual forecasting approaches independent from specific forecasting horizons, an overall fundamentalism versus chartism forecasting approach was determined for each trader by computing the mathematical means of the individual ratings given on the various forecasting horizons. Figure 2 shows the overall distribution of these overall scores. The majority of traders lie somewhere in the middle of the fundamentalist-chartist continuum, with only 3.1% of traders using an exclusively fundamental or an exclusively chartist forecasting approach, i.e. with an average forecasting approach 5 1 or ] 9 on the 0 10 rating scale. Across all forecasting horizons examined, traders usually use a mixture of both fundamental and chartist methods in their forecasting approach. A more detailed examination of the single forecasting horizons contributing to the overall result shown in Figure 2 corroborates this picture: even in intraday forecasts, the shortest forecasting horizon examined, less than one tenth (6.8%) of traders indicate that they use a purely chartist forecasting method. In forecasting horizons longer than 1 year, the longest forecasting horizons examined, less than one fourth of traders (22.6%) indicate that they use a purely fundamental forecasting method.

Figure 2. Overall forecasting approaches of foreign exchange traders (n = 289). Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

86

T. OBERLECHNER

Results also provide evidence that among foreign exchange traders the role of chartism has increased in recent years. Figure 3 compares the present findings for foreign exchange traders in London to the results of a previous survey conducted among chief foreign exchange dealers in London (Taylor and Allen, 1992). The comparison indicates that chartism has become more important to the foreign exchange market in the last decade. Using rating scales from 0 = pure chartism analysis to 10 = pure fundamental analysis, the present sample of London traders judges chartism to be more important on all trading horizons than do the traders surveyed by Taylor and Allen (1992). In order to examine the use of different forecasting approaches by different groups of traders, this study then examined whether there would be a connection between traders overall forecasting approach and their personal and professional background variables. Univariate subgroup analyses of foreign exchange traders overall fundamentalism versus chartism forecasting approaches reveal that there are no significant differences of the forecasting approach used in regard to traders age (F = 0.77, p = 0.57), gender (t = 0.00, p = 0.99), trader type (spot, forward, money market, and derivatives traders; F = 1.87, p = 0.12), position (trainee, junior trader, and senior trader; F = 0.68, p = 0.51), interbank versus customer traders (t = 1.37, p = 0.17), and trading limit (up to US$10 million, US$11 20 million, US$21 50 million, and US$50 million or more; F = 1.66, p = 0.18). However, a statistically significant preference can be found for traders with trading limits of up to US$50 million to use a more chartist forecasting approach than do traders with trading limits of more than US$51 million (t = 2.05, p B 0.05). The simultaneous analysis of foreign exchange traders in various European trading locations, i.e. in Austria, Germany, Switzerland and UK, allows for a comparison of the forecasting approaches used by traders from different trading sites. A 4 7 ANOVA design shows that trading location has a highly significant overall influence on the forecasting approach (F = 2.05, p B 0.01). Table 3 presents the results of detailed univariate analyses of variance between traders from the four different trading locations. On the shorter forecasting horizons of intraday, 1 week, 1, and 3 months, trading location has a statistically significant effect on the forecasting approach used. Only on the longer forecasting horizons of 6 months, 1 year, and longer than 1 year, no such significant influence of the trading location on the forecasting approach used can be found. When the subgroup ratings of traders from Frankfurt, London, Vienna, and Zurich listed in Table 3 are compared with each other on separate time horizons from intraday to 1 year, Bonferroni adjusted subgroup comparisons reveal that traders from London are significantly more fundamental than traders from Zurich in intraday, 1 week, and 1 month forecasting horizons. On the 3 months forecasting horizon,

Figure 3. Shift towards increased importance of chartism. Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

TECHNICAL AND FUNDAMENTAL ANALYSIS IN EUROPE

87

Table 3. Trading approaches used by traders in Frankfurt, London, Vienna, and Zurich Frankfurt (n = 30) M Intraday 1 week 1 month 3 months 6 months 1 year Longer than 1 year 3.73 4.17 4.80 5.83 6.23 6.63 6.87 S.D. 1.82 1.78 1.88 2.12 2.33 2.65 2.92 London (n = 96) Vienna (n = 46) M 3.97 4.18 4.72 5.30 5.78 6.20 6.31 S.D. 2.59 2.05 1.83 1.91 2.19 2.74 3.24 M 3.37 3.65 4.24 4.67 5.26 5.93 6.76 S.D. 2.36 1.89 1.42 1.28 1.41 2.16 2.48 Zurich (n = 107) M 2.94 3.33 4.05 4.78 5.60 6.57 7.21 S.D. 2.17 1.69 1.56 1.60 1.94 2.37 2.76 3.46* 4.07** 3.46* 4.35** 1.58 0.95 1.60 Univariate F -tests

Scale: from 0 = pure chartist analysis to 10 = pure fundamental analysis. M, mean; S.D., standard deviation. Level of significance: * pB0.05, ** pB0.01.

traders from Vienna and traders from Zurich use a significantly more chartist approach than do traders from Frankfurt. When traders from larger foreign exchange trading locations (Frankfurt and London) are compared to traders from smaller trading locations (Vienna and Zurich) in a 2 7 ANOVA design, the size of the trading location emerges as highly significant overall influence on forecasting approach (F = 4.73, p B 0.001). Again, univariate F -tests demonstrate that differences between larger and smaller trading locations exist on the comparatively short forecasting horizons of intraday (F = 9.09, p B 0.01), 1 week (F = 11.26, p B 0.01), 1 month (F = 9.97, p B 0.01), and 3 months (F = 10.77, p B 0.01), but not on the comparatively long forecasting horizons of 6 months (F = 2.65, p = 0.10), 1 year (F = 0.07, p = 0.80), and longer than 1 year (F = 3.22, p = 0.07). In the shorter forecasting horizons, traders from the larger trading centers hold a significantly more fundamental approach than do traders from the smaller trading centers. As described earlier, traders overall forecasting approaches were determined by the mean value of their individual ratings given on the seven time horizons. This approach reduces the information given by an individual trader on the seven different time-scales (intraday to more than 1 year) to one hypothetical overall forecasting approach. All information of each of the seven scales enters into the overall forecasting approach with equal weights. In order to arrive at a more differentiated typology of forecasting styles, a cluster analysis was conducted. This statistical procedure determines homogenous groups of traders using similar forecasting profiles across the various time horizons examined. Unlike the computed overall forecasting approach, a cluster analysis is able to differentiate between traders who arrive at same mean value of ratings by the use of very different forecasting profiles across the seven forecasting horizons. For example, a traders middle of the road overall forecasting approach (i.e. a mean value of 5 on the rating scales from 0 = pure chartism analysis to 10 = pure fundamental analysis) might be based on an equal mixture of chartist/technical and fundamental forecasting methods on each of the seven individual forecasting horizons, but also on very chartist forecasts on the shorter forecasting horizons, and on very fundamental forecasts on the longer horizons. A hierarchical cluster analysis using Wards clustering method and squared Euclidean distance measures suggested a solution of four relatively homogenous clusters of forecasting patterns. A non-hierarchical K -means cluster analysis then was applied to partition traders optimally into the four clusters. Figure 4 gives a picture of the four identified forecasting styles. The largest cluster (43%) of traders represents the forecasting profile termed chartist, ascending which starts with a very chartist approach to intraday forecasts (1.9 on the scale from 0 = pure chartist analysis to 10 = pure fundamental analysis). The longer the forecasting horizon, the more fundamental this forecasting approach becomes, and traders in this cluster progress to a strongly fundamental forecasting approach in forecasting horizons longer than 1 year (M = 8.6 on the 010 scale). Traders in the fundamental, ascending cluster (17%) have a forecasting profile which proceeds parallel to the chartist,
Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

88

T. OBERLECHNER

Figure 4. Four forecasting styles.

ascending profile described above. However, traders in this cluster start with a comparatively more fundamental (i.e. an only slightly chartist) forecasting approach in intraday forecasts (M = 4.0 on the 0 10 scale) and, with increasingly longer forecasting horizons, apply a progressively more fundamental forecasting approach, ending with an almost exclusively fundamental forecasting approach in forecasting horizons longer than 1 year (M = 9.3 on the 0 10 scale). Traders in the constant chartist cluster (16%) apply a persistently chartist forecasting approach across all time horizons (M = 2.6 4.0 on the 010 scales). The remaining fourth of responding traders (23%) is termed inverse middle since their ratings remain in the middle of the continuum. This cluster starts with an approach slightly more on the fundamental side in intraday forecasts (M = 6.4 on the 010 scale) and end with a more chartist approach in long-term forecasts (M = 3.6 on the 0 10 scale). Detailed analyses of these four forecasting styles and traders personal and professional background variables show that like the overall chartism versus fundamentalism approaches, these forecasting styles do also not correlate with traders age ( 2 = 20.54, p = 0.15), gender ( 2 = 2.21, p = 0.53), trader type ( 2 = 17.56, p = 0.13), position ( 2 = 3.84, p = 0.70), interbank versus customer traders ( 2 = 1.25, p = 0.74), and trading limit ( 2 = 8.57, p = 0.48). However, traders forecasting style correlates with trading location to a highly significant degree ( 2 = 23.87, p B 0.01). As Table 4 shows, the four identified forecasting styles are not distributed evenly across the examined trading centers. Relative to the overall sample of responding traders, fundamental, ascending traders are

Table 4. Distribution of forecasting styles in four European trading locations Frankfurt (n = 30) n Inverse middle Fundamentalist, ascending Constant chartist Chartist, ascending 8 9 6 7 % 26.7 30.0 20.0 23.3 London (n = 96) n 27 22 15 32 % 28.1 22.9 15.6 33.3 Vienna (n = 46) n 12 1 9 24 % 26.1 2.2 19.6 52.2 Zurich (n = 107) n 18 16 15 58 % 16.8 15.0 14.0 54.2 p = 0.005 2 = 23.87

Copyright 2001 John Wiley & Sons, Ltd.

Int. J. Fin. Econ. 6: 81 93 (2001)

TECHNICAL AND FUNDAMENTAL ANALYSIS IN EUROPE

89

significantly underrepresented in Austria whereas fundamental, ascending traders are over-represented in Germany to statistically significant degree, and in the UK there is a tendency for fundamental, ascending traders to be over-represented. In Germany and in the UK, chartist, ascending traders are significantly underrepresented. In Switzerland, the inverse middle group of traders is significantly underrepresented. When traders from the trading locations of Frankfurt and London are compared to traders from the trading locations of Vienna and Zurich, a highly significant bias in the distribution of trading styles emerges ( 2 = 17.50, p B 0.001). Whereas fundamental, ascending traders are over-represented in the London/Frankfurt group, chartist, ascending traders are over-represented in the Zurich/Vienna group.

3. DISCUSSION AND CONCLUSIONS This study supports previous findings that foreign exchange market participants do not see fundamental analysis and chartism/technical analysis as mutually exclusive (Taylor and Allen, 1992; Lui and Mole, 1998). Among the foreign exchange traders in the European trading centers of Frankfurt, London, Vienna, and Zurich, fundamentalists and chartists do not emerge as separate and distinct groups. Rather, a computation of traders overall fundamentalism versus chartism approaches results in a bell-shaped distribution of individual forecasting approaches, with a majority of foreign exchange traders using a balanced mix of both forecasting techniques. Only a very small minority of foreign exchange traders demonstrate an exclusively fundamental or exclusively chartist overall forecasting approach. Therefore, rather than being a second-rate forecasting tool employed by a subgroup of market participants who either do not possess the relevant fundamental information or do not know how to interpret this information, chartism seems to be an important forecasting tool employed by a majority of foreign exchange market participants in addition to the analysis of economic fundamentals. Generally, chartism is seen as more important on shorter forecasting horizons, whereas on longer forecasting horizons most market participants put more importance on fundamentals. This finding is consistent with previous survey studies by Taylor and Allen (1992) in London, by Menkhoff (1997) in Germany, and by Lui and Mole (1998) in Hong Kong. The forecasting methods of financial reporters were also examined in this survey because of the interdependent relationship between trading decision-makers and financial news providers in the foreign exchange market (Oberlechner and Hocking, 1997). Financial reporters provide trading decision-makers with news and also analyse possible consequences for market participants. Results show that financial reporters place more emphasis on fundamental analysis than do foreign exchange traders on all forecasting horizons investigated. This difference might be explained by the role played in the market by the media which needs to explain to trading participants reasons for past market events and possible future trends. This emphasis might lead to forecasting approaches to the market which are based more on economic fundamentals than are the forecasting approaches of foreign exchange traders. A difference in task might therefore help explain the divergence between a more academic and intellectual discourse on the market among journalists and a more action- and result-oriented discourse among traders. This assumption is supported by the findings on traders and financial journalists different perceptions of the role of feelings and of rationality in their own decision-making and in the foreign exchange market. Foreign exchange traders and financial journalists do not differ in their perception of the market on the aggregate level but show a marked difference in their view of the role of feelings versus rationality in their own decisions. In their trading decisions, traders consider feelings to be more influential than rationality, whereas financial journalists consider rationality more important than feelings in their reporting decisions. This difference in perception of ones own decisions may be accounted for by the fact that financial journalists often need to purposefully find rationality when reporting on and explaining the foreign exchange market. Already Allen and Taylor (1990, p. 58) concluded that their findings indicate that there should be ample warning to researchers in financial markets who do not allow for non-fundamental influences. The present results underscore this conclusion and may even indicate a continued rise in the importance of
Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

90

T. OBERLECHNER

non-fundamental factors in the foreign exchange market. However, the present results indicating an increased role of chartism need to be interpreted with caution. Whereas Taylor and Allen surveyed chief foreign exchange traders in London, this study included traders on all hierarchical levels. However, the major part of the London traders included in the present study actually were senior traders (73%), and only about one fourth of the London traders were not (junior traders 24%, trainees 3%). A more important limitation of a direct comparison of results is that for the questionnaire used in the present research, the original wording and question format was merely inferred from Taylor and Allens (1992) reports article and it did not offer respondents an answer category that no particular view was held over a certain time horizon. The shift towards chartist analysis may therefore not only be based on a real change of forecasting approaches but also on these methodological differences. However, in a recent survey Cheung et al. (1999) find that foreign exchange traders in the UK indicate that technical-based trading has increased as preferred trading method between 1993 and 1998 while the use of fundamentals-based trading has remained almost unchanged. One remarkable result of this research comes from a classification of foreign exchange traders according to their forecasting profiles across the seven forecasting horizons examined. A statistical cluster analysis results in four relatively homogenous groups of forecasting styles. Two of these groups, chartist, ascending and fundamental, ascending both show a typical linear trend from chartism in short forecasting horizons to fundamental analysis in extended forecasting horizons. At all forecasting horizons, fundamental, ascending traders attach more importance to fundamental analysis than do chartist, ascending traders. More than two out of five participating traders (43%) are chartist, ascending traders and less than one of five traders (17%) belongs to the group of fundamental, ascending traders. A third group of traders, termed constant chartist, applies a clearly more chartist than fundamental approach across all time horizons examined. This group is nearly as large as the group of fundamental, ascending traders (16%). The remaining group of inverse middle traders (23%) shows the most surprising forecasting style. Staying in the middle of the chartism fundamentalism continuum on all seven forecasting horizons, these traders move from a forecasting approach more fundamental than chartist in intraday forecasts to an approach more chartist than fundamental in long-term forecasts. Local factors have a significant effect on foreign exchange traders forecasting approach. This finding supports that expectations and forecasts in the global market are not formed in a homogenous way. Market participants from different trading locations use different forecasting approaches to arrive at their trading decisions. This finding becomes evident both in a comparison of traders overall forecasting approaches (i.e. the means of the ratings on the seven individual forecasting horizons) and in the more differentiated comparison of traders forecasting profiles across the various forecasting horizons examined. A comparison of overall forecasting approaches held by traders from Frankfurt, London, Vienna, and Zurich reveals significant local differences on shorter forecasting horizons from intraday forecasts to forecasting horizons of 3 months. For example, traders from London are significantly more fundamental than traders from Zurich in intraday, 1 week, and 1 month forecasts, and on a 3 months forecasting horizon, traders from Vienna and traders from Zurich use a significantly more chartist approach than do traders from Frankfurt. A comparison between the larger and smaller trading locations shows that on the shorter forecasting horizons of up to 3 months, traders from Frankfurt and London hold a significantly more fundamental overall forecasting approach than do traders from Vienna and Zurich. Reasons for these differences in forecasting styles between trading centers might be rooted in a variety of factors such as the centers geographic location, their market size, currencies chiefly traded, or differences in traders training, and require further examination. The fact that on the longer forecasting horizons differences in forecasting approaches among different trading centers could not be established to a statistically significant degree might indicate a real convergence of forecasting approaches on extended time horizons. However, the majority of surveyed foreign exchange traders can be assumed to have professional forecasting horizons which are shorter than some months. It is these professionally relevant shorter forecasting horizons in which significant local differences in forecasting approaches can be found. The convergence of forecasting approaches on longer
Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

TECHNICAL AND FUNDAMENTAL ANALYSIS IN EUROPE

91

horizons might therefore also be explained by traders lower exposure with longer forecasting horizons, a circumstance which might invite the giving of more similar estimates by traders from different trading locations. Trading location has a significant effect on forecasting style. This finding, already established by a comparison of overall forecasting approaches, could further be corroborated by a detailed analysis of trading styles used in different European trading centers. Most notably, traders from the larger trading locations of Frankfurt and London make more use of a fundamental, ascending forecasting style, and traders from the smaller trading locations of Vienna and Zurich use more a chartist, ascending forecasting style. This finding might be based on a different training of traders or different trading philosophies by market participants in various trading locations. The finding might however also correlate with the observation of some interview participants that the large foreign exchange centers (primarily London but also Frankfurt) are closer to the heart of foreign exchange information than might be Vienna or Zurich. However, in a recent study on the heterogeneity of foreign exchange forecasters MacDonald and Marsh (1996) find that there are only little informational asymmetries among different countries. Whatever the explanation for the systematic variations of forecasting styles dominant in different trading locations may be, findings of the present study suggest that information in the global foreign exchange market is not processed homogeneously by different parts of the market. They also suggest the importance of an even more differentiated understanding of different forecasting styles among market participants, and of the factors underlying these different styles of forming expectations about the market. Such better understanding might help clarify why currency forecasters hold heterogeneous expectations and why available information is interpreted differently by different forecasters (Ito, 1990; MacDonald and Marsh, 1996). The quantitative findings of this study call recurring attention to the importance of a consideration of non-fundamental factors affecting forecasts in the foreign exchange market. These findings are supported by qualitative interview data, where foreign exchange participants discuss chartism repeatedly in order to underscore the importance of behavioral and psychological factors in the foreign exchange market. One trading expert defines charts as visual representations of mass psychology. Traders examples of psychological motives for, and effects of, the use of chartism in the foreign exchange market serve as illustrations of this mass psychology. Traders often presented psychological reasons as co-determinants of why they use charts.4 One trading expert related the growing use of charts to the increasing loss of personal contact with other trading agents. In an effort to make up for the loss of personal contact caused by the use of electronic trading devices, traders would turn to charts in order to regain a sense of contact and safety in their trading decisions. Trading participants also reported that they use charts in order to gain more subjective confidence in the correctness of their judgement. The common use of chartism by foreign exchange decision-makers in the market may generate self-reinforcing market patterns. Such reinforcing patterns on a collective level might endow charts with the power to influence the market by way of a self-fulfilling prophecy. The concept of the self-fulfilling prophecy was first developed by Merton (1948) who characterized the self-fulfilling prophecy as an initially incorrect definition of a situation triggering new actions which then verify the initial false conception. In a recent questionnaire survey of foreign exchange professionals in Germany, Menkhoff (1997) argues that chartism may be seen as a form of self-fulfilling prophecy. This dynamic can already be found in Shillers empirical investigation of investors motives during the dramatic price drops of crashes in the stock market (Shiller, 1989). Also, Shleifer and Summers (1990) observe that many trading strategies based on pseudo-signals, noise, and popular models are correlated, leading to aggregate demand shifts. In the interviews conducted for the present study, trading experts themselves noticed the widespread use of charts in the market to effect strongly self-reinforcing patterns. As one trader observed, a foreign exchange trader may use chart analysis as insurance against the large number of other traders who are known to use charts, thereby providing an even stronger reason for other traders to use charts themselves. Without the chart system you cant deal anymore because everybody uses it, one trader stated. As a consequence, as another trading expert said, if a hundred people believe in something, these hundred people will move the market. So they confirm the chart.
Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

92

T. OBERLECHNER

Economists cannot just rely on assumptions and hypotheses about how speculators and other market agents may operate in theory, but should examine how they work in practice, by first-hand study of such markets, Goodhart (1988, p. 438) declares. The findings of this study shed a differentiated light on the importance of chartism and of fundamental analysis in different locations of the foreign exchange market. Not only is chartism widespread in this market and virtually no foreign exchange traders rate themselves as using fundamental methods exclusively in predicting future exchange rates. Results have also demonstrated a group of markedly different forecasting styles representing different forecasting patterns across the various time horizons examined. This emerging typology of forecasting practitioners in financial markets does not support the assumption that chartist/non-fundamental forecasting methods are used only by a group of ill-informed market participants who run constant danger of being eliminated by rational fundamentalists. A further investigation of local and cultural differences affecting forecasts and decision-making in the foreign exchange market may help shed a more differentiated light on collective market processes. Such research might also be able to explore whether different forecasting styles bring along different information gathering strategies, and different styles of processing information when forming expectations on the market.

ACKNOWLEDGEMENTS

I feel grateful to Sam Hocking for his collaboration and support during an early stage of this research. I also wish to thank all participating individuals and institutions for supporting the study with their cooperation. Research funding was provided by the Oesterreichische Nationalbank and Bank Austria. All remaining errors are my own.
NOTES 1. Together, the countries represented by these trading locations contributed to 42% of the global foreign exchange market activity in 1995 and in 1998 (Bank for International Settlements, 1995, 1998): Austria 1% (1998, 1%) Germany 5% (5%), Switzerland 6% (4%) and UK 30% (32%). 2. This number is an estimate. The number of foreign exchange trader questionnaires actually sent to participating banks was slightly lower. However, after the initial mailing of the questionnaires some banks produced additional copies in order to distribute the questionnaire to even more of their traders than originally agreed upon. 3. Again, the number of financial journalist questionnaires actually mailed is adjusted since some financial news agencies distributed additional copies of the questionnaire among their journalists. 4. It is, however, important to point out that chartism itself represents only a quasi-psychology of the foreign exchange market. The psychology represented by chartist analysis is purely behavioristic in nature and remains on the surface of the observable trading behavior outcomes of the market. Chartism itself does not examine the underlying reasons and decision-making dynamics of market participants. It remains on the level of the observable market movement, inferring laws about future trends from typical shapes of exchange rate patterns.

REFERENCES Allen H, Taylor MP. 1990. Charts, noise and fundamentals in the London foreign exchange market. The Economic Journal 100: 49 59. Bank for International Settlements. 1995. BIS Review No. 214. Basle. Bank for International Settlements. 1998. Central Bank Survey of Foreign Exchange and Derivatives Market Activity in April 1998: Preliminary Global Data. Basle. Black F. 1986. Noise. Journal of Finance 41: 529544. Cheung Y-M, Chinn MD, Marsh IW. 1999. How do UK-based Foreign Exchange Dealers Think Their Market Operates? Centre for Economic Policy Research Discussion Paper No. 2230, London. Frankel JA, Froot KA. 1986. Understanding the US dollar in the eighties: the expectations of chartist and fundamentalists. Economic Record 62: 2438. Frankel JA, Froot KA. 1990. Exchange Rate Forecasting Techniques, Survey Data, and Implications for the Foreign Exchange Market. IMF Working Paper. Goodhart C. 1988. The foreign exchange market: a random walk with a dragging anchor. Economica 55: 437 460. Harvey JT. 1996. Long-term exchange rate movements: the role of the fundamentals in neoclassical models of exchange rates. Journal of Economic Issues 30: 509516. Ito T. 1990. Foreign exchange rate expectations: micro survey data. American Economic Re6iew 80: 434 449. Levin JH. 1997. Chartists, fundamentalists and exchange rate dynamics. International Journal of Finance and Economics 2: 281 290. Copyright 2001 John Wiley & Sons, Ltd. Int. J. Fin. Econ. 6: 81 93 (2001)

TECHNICAL AND FUNDAMENTAL ANALYSIS IN EUROPE

93

Lui Y, Mole D. 1998. The use of fundamental and technical analyses by foreign exchange dealers: Hong Kong evidence. Journal of International Money and Finance 17: 535545. MacDonald R, Marsh IW. 1996. Currency forecasters are heterogeneous: confirmation and consequences. Journal of International Money and Finance 15: 665685. MacDonald R, Taylor MP. 1992. Exchange rates economics: a survey. International Monetary Fund Staff Papers 39: 1 57. Menkhoff L. 1997. Examining the use of technical currency analysis. International Journal of Finance and Economics 2: 307 318. Menkhoff L. 1998. The noise trading approachquestionnaire evidence from foreign exchange. Journal of International Money and Finance 17: 547 564. Merton R. 1948. The self-fulfilling prophecy. Antioch Re6iew 8: 193 210. Neely CJ. 1997. Technical analysis in the foreign exchange market: a laymans guide. Federal Reser6e Bank of St. Louis Re6iew September/October: 2338. Oberlechner T, Hocking S. 1997. Market Psychology and the Dynamics of Information: an Interdisciplinary View of the Foreign Exchange Market. Webster University, Vienna. Shleifer A, Summers L. 1990. The noise trader approach to finance. Journal of Economic Perspecti6es 4: 19 33. Shiller RJ. 1989. Market Volatility. MIT Press: Cambridge, MA. Taylor MP, Allen H. 1992. The use of technical analysis in the foreign exchange market. Journal of International Money and Finance 11: 304 314. Vigfusson R. 1997. Switching between chartists and fundamentalists: a Markov regime-switching approach. International Journal of Finance and Economics 2: 291305.

Copyright 2001 John Wiley & Sons, Ltd.

Int. J. Fin. Econ. 6: 81 93 (2001)