Taylor Menkhoff Technical Analysis

dec07_Article2 11/29/07 10:20 AM Page 936
Journal of Economic Literature

Vol. XLV (December 2007), pp. 936–972
The Obstinate Passion of

Foreign Exchange Professionals:
Technical Analysis
LUKAS MENKHOFF AND MARK P. TAYLOR∗
Technical analysis involves the prediction of asset price movements from inductive
analysis of past movements. We establish a number of stylized facts, including that
technical analysis is widespread in the foreign exchange market and that it may be
profitable. We then analyze four arguments that have been put forward to explain
this: that the market may not be fully rational; that technical analysis may exploit the
influence of official interventions; that it may be an efficient form of information pro-
cessing; and that it may inform on nonfundamental influences. While each may have
some validity, the latter is the most plausible.
As for the foreign exchange, it is almost as romantic as young love, and quite as
resistant to formulae.
H. L. Mencken
1. Introduction professional economists, the widespread,

continuing use of these techniques in the
T echnical analysis involves the prediction
of future exchange rate (or other asset-
price) movements from an inductive analysis
foreign exchange market (Mark P. Taylor
and Helen L. Allen 1992; Yin-Wong Cheung
and Menzie D. Chinn 2001) is somewhat
of past movements, using either qualitative
puzzling since technical analysis eschews
methods (e.g., the recognition of certain pat-
scrutiny of economic fundamentals and
terns in the data from visual inspection of a
relies only on information on past exchange
time-series plot) or quantitative techniques
rate movements that, according to the weak-
(e.g., based on an analysis of moving aver-
est notion of market efficiency, should already
ages), or a combination of both. For many
be embedded in the current exchange rate,
making its use either unprofitable or
∗
Menkhoff: University of Hannover. Taylor: University implying that any positive returns that are
of Warwick, Barclays Global Investors, and the Centre for generated are accompanied by an unac-
Economic Policy Research. The authors are grateful to ceptable risk exposure.1 On the other hand,
Carol Osler, Stephan Schulmeister, Roger Gordon (the
editor), and five anonymous referees for helpful com- despite an apparent emerging consensus
ments on an earlier version of this paper. Menkhoff grate-
fully acknowledges financial support from the German
Research Foundation (Deutsche Forschungsgemein- 1 In other words, the ratio of expected return to risk
schaft DFG). (the volatility of returns) is unacceptably low.
936
dec07_Article2 11/28/07 12:59 PM Page 937
Menkhoff and Taylor: The Obstinate Passion of Foreign Exchange Professionals 937
that fundamentals such as relative prices or in the world.3 Second, foreign exchange mar-
relative monetary velocity are capable of kets consist almost entirely of professional
explaining long-term exchange rate move- traders (Michael J. Sager and Taylor 2006), so
ments (Alan M. Taylor and Taylor 2004), that the impact of individual private investors
there is still no fundamentals-based may be neglected without loss of generality
exchange rate model available that is capa- (in contrast to equity markets—see, e.g., John
ble of forecasting exchange rate behavior M. Griffin, Jeffrey H. Harris, and Selim
over the shorter term (e.g., over a horizon Topaloglu 2003). Third, the share of short-
of twelve months or less: Jeffrey A. Frankel term interdealer trading is much higher in the
and Andrew K. Rose 1995; Taylor 1995b). foreign exchange market than it is in stock
Hence, the suggestion of Burton G. markets (Richard K. Lyons 2001). Finally, one
Malkiel (1996, p. 154), that “technical can probably say that there is less confidence
strategies are usually amusing, often com- among traders in models of fair value in the
forting, but of no real value” is perhaps a foreign exchange market compared to equity
little too dismissive, and this has been rec- markets (Frankel and Rose 1995; Taylor
ognized by a number of researchers. 2 1995b; John Y. Campbell, Andrew W. Lo, and
Indeed, over the past twenty years or so, A. Craig MacKinlay 1996). The greater lack of
international financial economists have consensus in models of fair value in the for-
increasingly turned their attention to the eign exchange market and the greater con-
study of technical analysis in an attempt to centration on shorter trading horizons might
understand both the behavior of foreign suggest that the use of technical analysis
exchange rates and the behavior of foreign would be more popular in the foreign
exchange market participants; so much so, exchange market, although the high propor-
in fact, that quite an extensive literature has tion of professional traders and deeper liquid-
developed on this topic. ity of the foreign exchange market might
Although the literature on the application suggest the opposite. However, we do not
of technical analysis to the foreign exchange want to dwell too long on the differences
market is sufficiently developed to warrant a between the foreign exchange market and
survey of its own, however, the foreign equity markets, but rather to emphasize the
exchange market cannot be viewed in total fact that foreign exchange is increasingly seen
isolation from other financial markets and as a separate asset class (Mark Snyder 2005).
so we shall occasionally stray into the litera- In this paper, we provide a selective and
ture on the application of technical analysis critical overview of the literature on technical
to financial markets more generally and to
equity markets in particular. The foreign
exchange market differs from equity mar- 3 Comparing spot market turnover yields a ratio of about
kets in some important aspects, however. three in favor of the foreign exchange market compared to
equities. We calculate this by taking daily spot trading in
First, total turnover in the global foreign equities and currencies in 2004–05 in the seven largest
exchange market is very high, at some 2,000 asset trading centers in the world. For equities, these were
billion U.S. dollars per day (Bank for the New York Stock Exchange, Nasdaq, the London Stock
Exchange, the Tokyo Stock Exchange, Euronext (Paris,
International Settlements 2005), which is Brussels, Amsterdam, Lisbon), the Deutsche Börse
several times greater than the combined (Frankfurt) and the Bolsas y Mercados Españoles (Spain)
daily turnover of the largest stock exchanges (see the website www.world-exchanges.org), while for for-
eign exchange the major centers were London, New York,
Tokyo, Singapore, Frankfurt, Hong Kong, and Sydney
(Bank for International Settlements 2005). In these cen-
2 As we discuss in more detail below, Malkiel’s dismis- ters, daily spot market turnover was about 485 billion U.S.
sive treatment of technical analysis is at odds with the evi- dollars in foreign exchange versus 160 billion U.S. dollars in
dence that technical analysis is widely used by financial equities; both figures represent more than 75 percent of
market traders. the respective world markets.
938 Journal of Economic Literature, Vol. XLV (December 2007)
analysis in the foreign exchange market. At The more qualitative aspect of technical
the forefront of our discussion throughout is analysis involves recognizing patterns in the
the question as to why technical currency data that are thought to herald trend rever-
analysis is so intensively and widely used by sals, such as “flags,” “head and shoulders”
foreign exchange professionals. As an organ- patterns, and so on (see, e.g., Allen and
izing device, we develop a set of stylized facts Taylor 1992).
concerning the importance and profitability The more widely used quantitative forms
of technical currency analysis. We then offer of technical analysis generally involve meth-
four different kinds of explanation for the ods such as moving averages in order to
persistent use of technical analysis and ana- exploit trends in the data. They thus attempt
lyze the supporting evidence in each case. to distinguish trends from noise, i.e., fluctua-
tions around a trend, by smoothing currency
2. The Nature of Technical Analysis returns.
A simple moving average rule would sig-
Technical analysis or, as it is also some-
nal an imminent break in trend, or the emer-
times called, “chartist analysis,” is a set of
gence of a new trend, when the moving
techniques for deriving forecasts of financial
average is crossed by the spot rate or by a
prices exclusively by analyzing the history of
shorter moving average. Thus, an imminent
the particular price series plus perhaps
upward break in trend for the spot rate, st,
transactions volumes.4 This analysis can be
might be signaled by a short moving average
performed in a qualitative form, relying
of length m > 1, MAt(m), intersecting from
mainly on the analysis of charts of past price
below a longer moving average of length
behavior and loose inductive reasoning, or it
n(n > m), MAt(n), i.e.,
can be strictly quantitative, by constructing
trading signals or forecasts through a quanti- MAt−1 (m) < MAt−1 (n) and MAt (m)
tative analysis of time series data. In prac-
> MAt (n), m < n,
tice, technical analysts often employ a
combination of both qualitative and quanti- where
tative techniques. 1
∑ i=0 st− i , j = m,n .
j −1
Clearly, technical analysis assumes that MAt ( j)
price developments display regular, recurring j
patterns, otherwise such a purely inductive Conversely, a downward break in trend would
technique would be useless. A second condi- be signaled by the short moving average
tion for the profitability of technical analysis is crossing the long moving average from
that these patterns must last long enough, above.5 Indicators of this kind will tend to be
first to be recognized, and second to make up profitable in markets exhibiting definite
for transaction costs and false signals. trends and so they are generically known as
“trend-following” or “momentum” indicators.
4 In this paper, we use the terms “technical analysis”
Another widely used device is the “over-
and “chartist analysis” and their derivatives largely as syn- bought/oversold” indicator, or oscillator.
onyms. This usage is not universal, although it is not Oscillators are measures designed to indi-
unusual among practitioners (see, e.g., Callum Henderson cate that price movements in a particular
2002) and has some precedence in the academic literature
(e.g., Charles Goodhart 1988; Frankel and Kenneth A.
Froot 1990b; Taylor and Allen 1992). It should be noted,
however, that some practitioners and some authors differ- 5 A variant would be to use exponential moving aver-
entiate “chartist analysis” as denoting the use of largely ages rather than simple arithmetic moving averages. Also,
visual analysis of charts and therefore see it as a subset of analysts may smooth the data prior to any analysis by
the methods denoted by “technical analysis” (see, e.g., applying very short-run (e.g., one-day) moving averages or
Christopher J. Neely 1997). See Allen and Taylor (1992) exponential moving averages to the data, in order to
for an outline of the origins of technical analysis. reduce the effect of noise on trading signals.
direction have recently been too rapid and since they are designed to anticipate a
that a correction in the opposite direction is reversal in trend.
imminent; they may take a number of pre- A third standard quantitative technique of
cise forms. One popular form is the relative technical analysis, the filter rule, involves
strength indicator (RSI) (J. W. Wilder 1978), buying a currency against another currency
for example, which is defined as: whenever the exchange rate has risen by
Ut more than a given percentage above its most
RSI t = 100 , recent low and selling it when the rate drops
U t + Dt by more than the same percentage below its
where Ut denotes the cumulated “up move- most recent high. An x-percent filter rule
ment” (i.e., the close-to-close increase on a may be expressed thus:8
day when the exchange rate has closed high- st − {st− i | i = min[i > 0 | (st− i − st )
er than the previous day’s closing rate) over a Buy if: 100
{st{−si t|− ii |=i min[
= min[
i >i 0> |0(s|t(−si t− s− )s<)
−i t t
certain period, and Dt denotes the cumulated
absolute “down movement” (the absolute s<t ) 0< &
0&(st−(si t−− i s−t−1s−t−i )1−<i ) 0<]}0 ]}
close-to-close decrease on a day when the > x%
exchange rate has closed lower than the pre- < 0 & (st− i − st−1− i ) < 0 ]}
vious day’s closing rate) over the same period {st{−si t|− ii |=i min[
= min[
i >i 0>|0(s| t(−si t− s−)s>)
−i t t
(often fourteen days):6 Sell if: 100
{st− i | i = min[i > 0 | (st− i − st )
U t = ∑ i=1 τι(st− i − st−1− i > 0)(st− i − st−1− i )
m
> 0 & (st− i − st−1− i ) > 0 ]} − st
> x%.
st>) >0 0&&(st(−sit−−i −st−s1t−−i1)− i>) >0 0]} ]}
Dt = ∑ i=1 τι(st− i − st−1− i < 0) | st− i − st−1− i | ,
m
Obviously, the variety of both qualitative

where τ (•) is an indicator variable that takes and quantitative techniques varies enor-
the value one when the statement in paren- mously—a fact which makes it quite difficult
theses is true, and zero otherwise.7 The RSI to provide a systematic assessment of techni-
thus attempts to measure the strength of cal analysis. Moreover, empirical tests of
“up movements” relative to the strength of specific rules and their associated trading
“down movements”, and is normalized to lie signals are not fully satisfactory as tests of the
between 0 and 100; common values at efficacy of technical analysis more generally,
which a particular currency is deemed to since users typically do not apply a single
have been overbought (signaling an immi- rule but rather a range of technical indica-
nent downward correction) or oversold (sig- tors which they update on a nonregular
naling an imminent upward correction) are
70 and 30, respectively (see, e.g., 8 Note that the “min” conditions in these filter-rule for-
Henderson 2002). Indicators of this kind are mulae minimize with respect to the time period rather
also referred to as “reversal” indicators, than the exchange rate, so that they find the most recent
period when the conditions indicated are met. For exam-
6 Some expositions define U and D in terms of average ple, the formula for the buy signal may be expressed thus:
t t
rather than cumulated up and down movements. This is “Starting at time t, find the most recent period in which
equivalent to our definition, however, since it just involves the exchange rate was less than it is at time t but where it
dividing by the total number of days and this factor can- had been falling compared to the previous period (i.e., the
cels out when the RSI is calculated. exchange rate’s most recent low) and if the exchange rate
7 The RSI is sometimes equivalently defined as has risen by more than x percent since then and time t,
buy.” The nonnegativity condition on the i subscripts is to
⎛ 1 ⎞ ensure that the formulae apply to lags rather than leads.
RSI t ≡ 100 − 100 ⎜ ⎟ ,
⎝ 1 + RSt ⎠ (Naturally, it is understood that “buy” means “buy the cur-
where RS, or “relative strength”, is defined as rency whose price in terms of the second currency is rep-
resented by the exchange rate in question” and “sell” is to
RSt ≡ U t / Dt . be interpreted similarly.)
basis. In addition, many technical analysts and Taylor and Allen (1992), was carried out
will also apply considerable market intuition among chief foreign exchange dealers at
to complement their quantitative conclu- financial institutions located in London in
sions, so there remains always a major ele- 1988; the most recent was conducted in
ment of subjectivity with the application of 2001 by Thomas Gehrig and Lukas
technical analysis. Menkhoff (2004) among foreign exchange
dealers and fund managers located in
3. The Importance of Technical Analysis in Germany. In all, the various surveys have
the Foreign Exchange Market covered foreign exchange professionals
based in London, Frankfurt, Hong Kong,
The widespread use of technical analysis
Singapore, Tokyo, New York and Zurich. In
by foreign exchange professionals was first
1995—the midpoint between the earliest
brought to the attention of academic
and latest studies—the Bank for Inter-
researchers by Stephen H. Goodman (1979),
national Settlements (2005) ranked the
Group of Thirty (1985), Frankel and Froot
seven locations covered by the survey stud-
(1986, 1990a, 1990b), and Goodhart (1988).9
ies as first to seventh in terms of daily
However, the existence of technical analysis
turnover in foreign exchange dealing, mak-
and even its use did not provoke sustained
ing up about 78 percent of the total global
academic interest as long as the available
turnover; the combined market share of
evidence was not of a more systematic
these centers was virtually unchanged until
nature. The skepticism with which academic
2004. Although the response rates of the
economists initially viewed (and to some
studies differ markedly, the results are
extent continue to view) technical analysis
remarkably invariant.
was derived largely from the intellectual
The studies of Allen and Taylor (1990) and
standing of the efficient markets hypothesis
Taylor and Allen (1992) documented for the
(EMH), which, even in its “weak form”
first time systematically that technical analy-
(Eugene F. Fama 1970), maintains that all
sis is, indeed, an important tool in decision
relevant information should already be
making in the foreign exchange market.
embodied in asset prices, making it impossi-
They further established a perceived com-
ble to earn excess returns on forecasts based
plementarity among market practitioners in
on historical price movements, once suitable
the use of technical and fundamental analy-
risk-adjustment is made.10
sis, and showed that reliance on technical
Nevertheless, during the 1990s, beginning
analysis was skewed toward shorter trading
with the work of Allen and Taylor (1990), a
or forecast horizons. These three basic find-
number of academic studies appeared that
ings are also features of the results reported
reported the results of surveys of foreign
in the remaining six survey studies (see table
exchange market participants concerning
2), and therefore form the first three of our
the use of technical analysis. The salient
stylized facts—SF1, SF2, and SF3:
characteristics of these studies—in terms of
Stylized Fact 1 (SF1): Almost all foreign
survey year, target group, response rate,
exchange professionals use technical analysis
location, etc.—are given in table 1. The first
as a tool in decision making at least to some
survey, discussed in Allen and Taylor (1990)
degree.
Stylized Fact 2 (SF2): Most foreign
9 The early study of technical analysis in the foreign exchange professionals use some combina-
exchange market of William Poole (1967a) can be seen as tion of technical analysis and fundamental
very much ahead of its time. analysis.
10 The more extreme form of the EMH assumes that
agents are risk neutral so that significantly nonzero excess Stylized Fact 3 (SF3): The relative weight
returns cannot be earned. given to technical analysis as opposed to
TABLE 1
INFORMATION ON QUESTIONAIRE SURVEY STUDIES
Study time financial target group number of response

of survey center responses rate
Taylor and 1988 London chief 213 60.3%
Allen (1992) FX dealers
Menkhoff 1992 Germany FX dealers; 205 41.3%
(1997) int’l fund
managers
Lui and Mole 1995 Hong Kong FX dealers 153 18.8%
(1998)
Cheung and 1995/96 Hong Kong, FX dealers 392 20.0%
Wong (2000) Singapore,
Tokyo
Oberlechner 1996 Switzerland, FX dealers; 321 53.5%
(2001) United Kingdom, (financial (59) (29.5%)
(Austria, Germany) journalists)
Cheung and 1996/97 United States FX dealers 142 8.1%
Chinn (2001)
Cheung, Chinn 1998 United Kingdom FX dealers 110 5.8%
and Marsh (2004 )
Gehrig und 2001 Germany, (Austria) FX dealers; 203 51.9%
Menkhoff (2004) int’l fund managers
Notes: “FX” stands for “foreign exchange.”
fundamental analysis rises as the trading or a high average score could mask its concen-
forecast horizon declines. tration in small subgroups in the market.
SF1 can be established from the third col- Table 3 reveals, however, that technical
umn in table 2. Those surveys which asked analysis is perceived as important relative to
respondents whether or not they used tech- fundamentals across a range of practitioner
nical analysis at some horizon found that groups such as chief foreign exchange deal-
around 90 percent or more did so. The fact ers, international portfolio managers, and
that practitioners use technical analysis does others, whatever their specific role in foreign
not by itself, however, establish that they exchange trading may be.11
regard it as of major importance—they may Early analytical studies of the foreign
attach some weight to it, but only a low exchange market that allocated a role to
weight. Although not identical in design, technical analysts or chartists tended to view
most of the surveys therefore also asked chartists and fundamentalists as competing
respondents to quantify the weight given to factions, either in their own right as traders
technical analysis relative to fundamental
analysis at various horizons (SF2); the aver- 11 Overall foreign exchange trading operations will be
age relative weight assigned to technical headed by a chief dealer who is, however, due to his man-
analysis is shown in column five of table 2, agement role, not necessarily the most active trader. Then
there will be core traders, such as those responsible for
and ranges from around 30 percent to a little spot trades in a given exchange rate, and finally there are
over 50 percent. other foreign exchange traders who may focus on further
A further aspect of the importance of objectives such as trading forwards and futures, or are
junior and thus have less leeway in their decision making.
technical analysis concerns its use among See Sager and Taylor (2006) for a comprehensive analysis
various groups of market participants, since of the structure of the foreign exchange market.
TABLE 2
THE IMPORTANCE OF TECHNICAL ANALYSIS ACCORDING TO QUESTIONNAIRE SURVEYS
Study form of some use share of share of the relation

analysis for of technical technical plus technical analysis between the
decision making analysis fundamental to technical plus weight of technical
analysis to total fundamental analysis and
forms(2) analysis(2) horizon
Taylor and fundamental 89.4% 100% 32%(4) strictly negative
Allen (1992) analysis;
technical analysis
Menkhoff fundamental; >90% 82% 45% weakly
(1997) technical; hump-shaped
flow analysis
Lui and fundamental; ~100% 100% 51%(5) strictly negative
Mole (1998) technical
Cheung and fundamental; n.a. 62%(3) 40%(3) strongly
Wong (2000)(1) technical; hump-shaped
bandwagon;
overreaction;
speculative forces
Oberlechner fundamental; >98% 100% 49% strictly negative
(2001) technical
Cheung and see Cheung and n.a. 56% 29% strongly
Chinn (2001)(1) Wong (2000) hump-shaped
Cheung, Chinn see Cheung and n.a. 49%(6) 47%(6) strongly
and Marsh Wong (2000) 54%(7) 29%(7) hump-shaped
(2004)
Gehrig and fundamental; >90% 77% 53% weakly
Menkhoff (2004) technical; hump-shaped
flows analysis
(1)
Notes: These studies do not directly ask for analytical tools but for “factors determining exchange rate move-
ments.”
(2)
Unweighed averages of values for different horizons.
(3)
Data based on Hong Kong only (values for Singapore and Tokyo are similar).
(4)
Share is calculated as ratio of scale values 0 to 4 / scale values 0 to 4 plus 6 to 10 (i.e., preference for
technical analysis to total preferences); weighed with share of respondents at respective horizon (see
Taylor and Allen 1992, table 3B, first column).
(5)
Share is calculated as ratio of importance given to technical analysis to total.
(6)
Traders were asked to select the technique which best characterizes their dealing method.
(7)
This value is a more indirect indication and is derived from the same question as mentioned in foot-
note (1).
or as advisers to traders (Goodhart 1988; (1992). In particular, the weight given to

Frankel and Froot 1990a, 1990b). SF2, strong mutual exclusiveness of chartism and
however, (“Most foreign exchange profes- fundamentalism, i.e., a reliance on either
sionals use some combination of technical fundamental or technical analysis (repre-
analysis and fundamental analysis”) chal- sented in figure 1 by values 9 and 10), is at
lenges this adversarial view of chartism and most ten percent of respondents in all stud-
fundamentalism. Figure 1 demonstrates ies.
that other studies have basically reproduced Finally, SF3 states that technical analysis
this finding of perceived complementarity tends to be perceived as less important at
(i.e., a reliance on fundamental and techni- longer horizons in comparison with funda-
cal analysis) established by Taylor and Allen mental analysis (see explicitly Taylor and
TABLE 3
THE IMPORTANCE OF TECHNICAL ANALYSIS IN SEVERAL SUB-GROUPS AND AT TYPICAL FORECASTING HORIZONS
Question: “Please evaluate the importance of the three following information types for your typical decision
making, by distributing a total of 100 points. For information types which you do not use, please give
0 points.”
. . . Fundamentals (economic, political)
. . . Technical Analysis (charts, quantitative methods)
. . . Flows (who is doing what, which customer orders are existing)
Menkhoff (1997, 1998) Gehrig and Menkhoff (2006)

Horizon chief FX core FX other FX int’l fund chief FX other FX int’l fund
dealers dealers dealers managers dealers dealers managers
Intraday 30.5 36.6 23.2 n.a. 45.0 37.3 n.a.
Few days 37.8 38.6 44.0 45.0 45.9 45.1 52.5
Few weeks 34.3 42.5 40.6 35.9 46.9 37.3 32.8
2 to 6 months 42.6 50.0 29.3 36.1 28.3 31.7 31.7
6 to 12 months (20) n.a. (20) 30.0 (0) n.a. (15.0)
>12 months n.a. n.a. (40) n.a. (100) (30) n.a.
Mean 35.4 38.4 39.9 36.1 44.9 40.0 37.0
n 44 66 39 50 42 102 58
Notes: Data are from the studies Menkhoff (1997, 1998) and Gehrig and Menkhoff (2006). The first value of 30.5
says that chief FX dealers who have a typical intraday forecasting horizon give technical analysis a weight of 30.5%
(out of 100% for fundamental, technical and order flow analysis). Shaded cells mark the typical horizon (median
value) for decision making of the respective group (e.g. 49% of core FX dealers mark intraday as their typical
horizon). Numbers in parenthesis refer to groups with 1 to 3 responses. “FX” stands for “foreign exchange.”
Allen 1992, table 3B). A graphical presenta- foreign exchange transactions signed accord-
tion of the research regarding SF3 can be ing to the originator of the trade—see, e.g.,
seen from the work of Taylor and Allen Takatoshi Ito, Lyons, and Michael T. Melvin
(1992), Yu-Hon Lui and David Mole (1998), 1998; Lyons 2001; Lucio Sarno and Taylor
and Thomas Oberlechner (2001) in figure 2A. 2001b; Martin D. D. Evans and Lyons 2002).
These three studies relate the perceived rela- The inclusion of factors other than technical
tive importance of technical and fundamental and fundamental analysis in the menu of
analysis with forecast or trading horizon. If choices offered to survey participants thus
one takes, however, the other studies—fea- dilutes the relative score given to technical
tured in figure 2B—into account, the result analysis in the shorter-term domain (see on
remains unchanged for the medium and these factors columns two and four in table
longer horizons but becomes less clear for the 2).12 Considering only fundamental and
very short horizon. There is, however, an
obvious reason for this apparent difference in 12 It is possible that order flow might better be inter-
perceived relative importance at the very preted as fundamental rather than as technical informa-
short horizon, in terms of their coverage of tion since, although it is clearly not on the list of standard
analytical tools or price-determining factors. macroeconomic fundamentals, it may in some sense
embody the net effect of fundamental influences on the
In particular, studies featured in figure 2B foreign exchange market (Lyons 2001; Evans and Lyons
also take into account other factors such as 2005b). Nevertheless, we rely here on studies in which
the perceived importance attached by market order flow forms a third category and which may to some
extent represent the current perception of order flow by
practitioners to information on order flow foreign exchange professionals (Henderson 2002; Gehrig
(i.e. on information relating to the value of and Menkhoff 2004).
25 percentage
of responses
Taylor and Allen (1992; Table 4)

20 Menkhoff (1997)
Oberlechner (2001, Figure 2)
Lui and Mole (1998, Figure 1)
Gehrig and Menkhoff (2004)
15
10
strongly
complementary
0
0 1 2 3 4 5 6 7 8 9 10
mutually
exclusive
−5
Figure 1. On the Perceived Complementarity of Technical and Fundamental Analysis
Note: The regressions are calculated as best fit of a polynomial of second order. The data from Menkhoff (1997) and
Gehrig and Menkhoff (2004) are transformed in the following way: the individual weight given to fundamental analysis
(f) and to technical analysis (t) is put into one measure x. x = | f − t | : (f + t) •100. The percentage is then put into the scale
0 to 10 according to: x < 10% → 0; 10% ≤ x < 20% → 1; . . .; 90% ≤ x < 100% → 9; 100% → 10. The data from
Oberlechner (2001) are transformed as follows: < 6 (strong complementarity) → 0, < 5 → 2, < 4 and < 7 → 4, < 3 and
< 8 → 6, < 2 and < 9 → 8, < 1 and < 10 → 10 (each value multiplied by 0.6 to account for different scaling).
technical analysis for the purpose of com- dealers in an emerging market (N. R.
parison indeed reproduces the earlier find- Bhanumurthy 2004).
ing in figure 2C (also supported by Gehrig
and Menkhoff 2006 with a different 4. Profitability of Technical Analysis:
methodology). Measures and Results
As a final remark it may be reassuring that
the stylized facts shown for foreign exchange The evidence concerning the profitability
dealers and fund managers from the main of technical currency analysis tends to be
financial centers by and large also hold for inconclusive. From a theoretical point of
financial journalists (Oberlechner 2001) and view, this is perhaps unsurprising, since if
A. Studies considering technical and fundamental analysis only

80%
Lui and Mole
relative 70% Oberlechner
weight of 60% Taylor and Allen
technical
50%
analysis
40%
30%
20%
10%
0%
Intraday 1 week 1 month 3 months 6 months 1 year > 1 year
B. Studies considering also further influences Cheung and Chinn

50% Cheung, Chinn and Marsh
relative 45% Cheung and Wong
Gehrig and Menkhoff
weight of 40% Menkhoff
technical
35%
analysis
30%
25%
20%
15%
10%
5%
0%
Intraday 1 week 1 month 3 months 6 months 1 year
C. Studies in B, here without further influences, i.e., relating only the weight of
technical and fundamental analysis to each other
Cheung and Chinn
100% Cheung, Chinn and Marsh
relative 90% Cheung and Wong
Gehrig and Menkhoff
weight of 80% Menkhoff
technical 70%
analysis 60%
50%
40%
30%
20%
10%
0%
Intraday 1 week 1 month 3 months 6 months 1 year
Figure 2. The Relative Importance of Technical Analysis Depending on the Horizon of Decision Making
Notes: The relative weight of technical analysis is calculated as the weight of technical analysis to the sum of technical
plus fundamental analysis. Horizons are taken from Taylor and Allen (1992) and Lui and Mole (1998); importance (scale
0–10) in Oberlechner (2001); Figure is transformed into percentage points; Menkhoff (1998) and Gehrig and Menkhoff
(2004) are transformed: “few days” into “1 week,” “few weeks” into “1 month,” “2 to 6 months” into “3 months” and “6
to 12 months” into “1 year,” data for “6 months” is interpolated; Cheung and Wong (2000), Cheung and Chinn (2001),
and Cheung, Chinn and Marsh (2004) are transformed: “medium run” (< 6 months) into “1 month” and “long run” (> 6
months) into “1 year,” data for “1 week,” “3 months,” and “6 months” are interpolated.
technical analysis was never profitable, its Third, it is probably the form of risk con-
widespread use (see section 2) would be hard sideration—an essential element of Fama’s
to understand; if, on the other hand, techni- “equilibrium expected profits”—that divides
cal analysis was always profitable, it would advocates and opponents of technical analy-
perhaps imply that the foreign exchange sis in the interpretation of their empirical
market is inefficient to a degree that many work.
economists would not find credible. If one compares the overview of earlier
Indeed, the EMH does not imply in this studies in table 4 with the three require-
respect that returns to applying a technical ments just mentioned, it becomes clear that
trading strategy have to be zero. Rather, effi- these studies are all characterized by short-
cient markets “rule out the possibility of comings to a greater or lesser extent: many
trading systems based only on [current and of them examine only one currency, some do
past] information [having] expected profits not consider all kinds of costs and most are
or returns in excess of equilibrium expected handicapped by a short period of investiga-
profits or returns” (Fama 1970, p. 385). In tion which does not allow for appropriate
this context, equilibrium expected returns out-of-sample calculations. Thus, only the
must be calculated after allowing for a rea- studies of Michael P. Dooley and Jeffrey R.
sonable return to risk and after allowance for Shafer (1983) and Richard James Sweeney
transactions costs.13 (1986) have been consistently cited in the
In assessing the profitability of technical literature (see, e.g., Ramazan Gencay 1999;
analysis, therefore, three methodological Blake LeBaron 1999; Neely 2002; Dennis
aspects have to be addressed carefully. First, Olson 2004).
any examination should define appropriate It is interesting to note that most of the
alternatives, i.e., on the one hand the techni- stylized facts that can be drawn from prof-
cal analysis strategies and on the other hand itability examinations are already found in
a strategy relying on the EMH. Important this early literature (table 4) and that they
features of this comparison must include are supported by later studies (e.g.,
transaction costs and interest rate carry Patchara Surajaras and Sweeney 1992;
costs.14 Menkhoff and Manfred Schlumberger
Second, the issue of statistical significance 1995; Keith Pilbeam 1995; Neely 1997;
has to be tackled. Independently of the dis- LeBaron 1999; Peter Saacke 2002). They
tribution of exchange rate returns, there can be gathered together here as Stylized
must always exist a technical analysis strate- Facts 4, 5 and 6:
gy that is able to exploit characteristics of the Stylized Fact 4 (SF4): The consideration of
time series in any particular sample. Thus, it transaction costs and interest rate costs actu-
is not profitability per se that is interesting ally faced by professionals does not necessar-
but the possible significance of the result ily eliminate the profitability of technical
that challenges the EMH. currency analysis.
13 It may also allow for tax payments, where tax treat-

ment differs across investor groups. how the information would be imparted into market
14 It must also constitute a study of profitability from an prices at all, so that the proposed “no-profit-from-costly-
ex ante rather than an ex post perspective, so that there is information equilibrium” implies a contradiction and so
perceived profit from gathering and utilising information cannot in fact be an equilibrium. This is the so-called
in a superior fashion. This is important because it could be Grossman–Stiglitz paradox, and in the resolution of this
argued that, in equilibrium, there should be no gain from paradox, both Sanford J. Grossman and Joseph E.
utilising superior information since such information Stiglitz (1980) and W. Bradford Cornell and Richard Roll
would already be embodied in market prices. If this were (1981) have shown that a sensible financial market equi-
the case, however, traders would presumably not bother to librium must leave some incentive for costly information
gather costly information, in which case it is difficult to see acquisition and analysis.
TABLE 4
EARLIER STUDIES EXAMINING THE PROFITABILITY OF TECHNICAL ANALYSIS IN FOREIGN EXCHANGE MARKETS
Study period number form of consideration of excess

covered of exchange technical returns of
rates analysis transaction interest risk technical
costs rates analysis
Poole (1967) 1919– 9 10 filters no no no +
24/29
Poole (1967a) 1950–62 1 12 filters no no no +
Dooley and 1973–75 8 Filter +
Shafer (1976)
Logue and 1970–74 1 14 filters yes no no +
Sweeney (1977)
Logue, Sweeney and 1973–76 7 11 filters no no no +
Willett (1978)
Cornell and 1973–75 6 13 filters, yes no yes +
Dietrich (1978) 27 mov.
Averages
Dooley and 1973–81 9 7 filters yes yes no +
Shafer (1983)
Sweeney (1986) 1973–80 10 7 filters yes partially yes +
Schulmeister (1987) 1973–86 1 9 filters, yes partially no +
9 mov. av.,
5 momentum,
1 point & figure
Stylized Fact 5 (SF5): Technical analysis time (SF6) include Dennis E. Logue,
tends to be more profitable with volatile cur- Sweeney, and Thomas W. Willett (1978),
rencies. Dooley and Shafer (1983), Menkhoff and
Stylized Fact 6 (SF6): The performance of Schlumberger (1995), LeBaron (2000),
technical trading rules is highly unstable Michael Dueker and Neely (2007), and
over time. Neely, Weller, and Ulrich (forthcoming).15
Evidence on SF4—the profitability of The rapidly growing empirical literature
technical trading rules after allowance for on the profitability of technical analysis in the
transactions costs and interest rate carry—is foreign exchange market that has appeared
provided by, among others, Cornell and J. since the late 1990s has, if anything, further
Kimball Dietrich (1978), Sweeney (1986), substantiated these older stylized facts (see,
Stephan Schulmeister (1987), LeBaron
(1999), Saacke (2002), and Neely, Paul A. 15 As an anonymous referee has pointed out that the
Weller, and Joshua M. Ulrich (forthcoming). evidence for the instability of technical trading rules
should be interpreted with care, however, since exchange
Studies supporting the hypothesis that tech- rate returns are noisy relative to sample length, tests for
nical trading rules are more profitable for unknown structural breaks have notoriously low power
currencies experiencing relatively higher and test for structural breaks at known breakpoints are
subject to data snooping bias. Moreover, the popularity of
volatility (SF5) include Cornell and Dietrich technical analysis may be sustained not by its consistent
(1978), Dooley and Shafer (1983), Chun I. performance but by its perceived performance across var-
Lee and Ike Mathur (1996), and Neely and ious prominent episodes and instability in performance
over time is also a characteristic of fundamentals-based
Weller (1999). Work suggesting that techni- exchange rate models (Cheung, Chinn, and Antonio
cal trading rule performance is unstable over Garcia Pascual 2005).
e.g., Cheol-Ho Park and Scott H. Irwin allowing for a reduction in transactions costs
2007). In addition, it is possible to discern a over time (Neely, Weller, and Ulrich forth-
number of developments among the more coming). A possible reason for the decline
recent literature. in profitability of technical trading rules in
First, a major methodological innovation recent years is the growth in the hedge fund
has been the introduction of the bootstrap industry to a level estimated to be of the
approach addressing the problem of insignif- order of more than $1 trillion in assets under
icant evidence (Richard M. Levich and Lee management in the early to mid 2000s and
R. Thomas 1993; LeBaron 1999, 2000; Carol still growing (Securities and Exchange
L. Osler 2000, 2003) and, more recently, the Commission 2003).17 Foreign exchange has
introduction of methods for testing for been increasingly seen as a separate asset
potential data-snooping bias (Park and Irwin class by hedge funds (Deutsche Bank 2005)
2005; Min Qi and Yangru Wu forthcoming). and many hedge funds apply quantitative
Second, the range of technical analysis strategies that may employ one or more
tools and trading rules considered has been technical trading rules (Sager and Taylor
increased far beyond filter rules, moving 2006); insofar as technical analysis provides
averages, or point-and-figure indicators, and a short-term informational advantage, a
now includes the possible psychological bar- large amount of trades employing similar
riers of round figures, the closely related technical trading rules will clearly erode
issue of support and resistance levels (Paul that advantage much faster. Nevertheless,
De Grauwe and Danny Decupere 1992; significant evidence of profitability—albeit
Goodhart and Riccardo Curcio 1992; Osler on a reduced level—remains and may even
2000, 2003, 2005), or of momentum-based have been increasing during recent years in
strategies (John Okunev and Derek White euro-dollar trading (Park and Irwin 2005;
2003).16 Schulmeister 2005). Also, there may be
Third, the longer span of data available more complex forms of technical analysis
for the floating rate period since the early that did not become less profitable over
1970s has stimulated the question as to time (e.g., Okunev and White 2003; Neely,
whether profits from technical analysis are Weller, and Ulrich forthcoming).18
declining over time. There is indeed evi- Fourth, there have been attempts to avoid
dence that the foreign exchange market has potential selection bias by letting actors state
become more efficient over time in the their preferred rules prior to any profitabili-
sense that the application of traditional ty analysis (Allen and Taylor 1990; Goodhart
moving average rules, which was shown to
be profitable for the 1970s (e.g., Logue and
Sweeney 1977; Cornell and Dietrich 1978; 17 A hedge fund may be defined as a lightly regulated
Dooley and Shafer 1983; Sweeney 1986), investment fund that engages in active and leveraged
financial portfolio management, usually involving a range
became much less profitable in the 1990s of securities in which they may employ complex trading
(LeBaron 2000; Olson 2004), even after strategies (not just “buy and hold”) in which both long and
short positions are typically held, and whose objective is to
provide a return much greater than the return from a pas-
16 The technical trading rules that have been examined sive investment strategy. They are lightly regulated large-
in this literature include those based on head and shoul- ly because, in most countries, they are only open to
ders patterns (P. H. Kevin Chang and Osler 1999; Bernd institutional investors and wealthy accredited investors,
Lucke 2003), candlestick formations (Norbert M. Fiess rather than the general public.
and Ronald MacDonald 2002), neural networks (Gencay 18 Similar results are reported by Po-Hsuan Hsu and
1999), genetic programming (Neely, Weller, and Robert Chung-Ming Kuan (2005) for stock markets, providing
Dittmar 1997; Neely and Weller 1999, 2001), Markov support to the interpretation of Neely, Weller, and Ulrich
switching models (Ian W. Marsh 2000; Dueker and Neely (forthcoming) that markets may need time to become
forthcoming) and real-time trading models (Gencay et al. aware of and then to arbitrage away profit opportunities
2003). generated by technical trading rules.
and Curcio 1992, Curcio and Goodhart known that positive results are generally
1993; Osler 2000). much easier to report and publish than neg-
Fifth, studies have explored the relation ative results. In particular, these studies may
between nonlinear exchange rate modeling be subject to “data snooping” (Halbert White
and technical analysis (William C. Clyde and 2000). Data snooping occurs when a given
Osler 1997; Fiess and MacDonald 1999; set of data is used more than once for pur-
Lutz Kilian and Taylor 2003; De Grauwe poses of inference or model selection, so that
and Marianna Grimaldi 2006a, 2006b; the possibility arises that any satisfactory
Stefan Reitz and Taylor forthcoming). results obtained may simply be due to
Sixth, stimulated by the apparent success chance rather than to any merit or skill inher-
of longer-term exchange rate modeling via a ent in the method yielding the good results.
Markov switching approach (Charles Engel White (2000) develops a bootstrap simula-
and James D. Hamilton 1990), studies have tion technique—the “reality check”—for
found some links between regime switches examining whether it is inherent skill or pure
and technical trading rules (Hans chance that leads to the best rule being cho-
Dewachter 1997, 2001; Robert Vigfusson sen out of any given universe of rules.
1997). However, profitability does not seem Intuitively, a reality check involves replicat-
to be better than for simple moving average ing, by Monte Carlo methods, many artificial
rules (Dueker and Neely 2007) although an data sets that in some sense match the prop-
advantage may be gained by the fact that erties of the original data sets, and testing the
profits remain more stable over time (Neely, various trading rules for profitability on each
Weller, and Ulrich forthcoming). data set. If there is a tendency for the same
Finally, some studies (Curcio et al. 1997; rule to be selected as the most profitable for
Osler 2000, 2003; Neely and Weller 2003; each data set, then this suggest that it really
Roman Kozhan and Mark Salmon 2006) is a good rule; if there is no tendency to
have examined the profitability of technical select that particular rule for each of the arti-
analysis on a very high-frequency (intraday) ficial data sets, then this indicates that it was
basis, with mixed results. selected as the most profitable rule in the
On balance, however, the literature on the original data set purely by chance.19
profitability of technical trading rules tends The first application of White’s reality
to support the existence of significant profits check to technical trading rules in the for-
to be had by employing these rules in the foreign exchange market is due to Qi and Wu
eign exchange market (see also Park and (forthcoming). These authors examine a
Irwin 2007 for a concurring reading of the large number of technical trading rules and
literature). Of course, this in itself raises a apply them to daily data on seven dollar
sample selection bias issue, since it is well exchange rates over the period 1973–98. The
19 An anonymous referee has pointed out a number of

issues that may be raised with respect to White’s reality a given nominal test size while the performance of the
check. In particular, while the reality check is clearly an benchmark trading rule does not change. An alternative
improvement over earlier approaches that ignored data- approach would be to carry out an ex ante search for prof-
snooping bias, the group of trading rules making up the itable trading rules using artificial intelligence such as a
universe within which the reality check is carried out must genetic algorithm that “learns” trading rules and applies
still be chosen and that brings back the danger of data them, as in the equity-market study of Franklin Allen and
snooping in a different guise. Indeed, there may even be Risto Karjalainen (1999), although this approach would
a systematic bias involved as researchers may, consciously also potentially be subject to sample-selection bias.
or unconsciously, rely on rules that have been implicitly Alternatively, one can perform true out-of-sample tests by
tested on similar data in previous research. Moreover, retesting rules that have found to be profitable in earlier
merely adding a large number of poor rules into the real- studies—as for example in LeBaron (2000) or Neely,
ity check universe will tend to raise the critical values for Weller, and Ulrich (forthcoming).
technical rules are various calibrations of beta of foreign currencies with this index
four classes of rules: filter rules (buy or sell a rather than the beta of currency positions that
currency if it moves more than a certain per- result from technical trading rules (for the lat-
cent from its most recent high or low); mov- ter, see, e.g., Stephen J. Taylor 1992 with the
ing average rules (as discussed above); same result).20 Nevertheless, their very low
support and resistance rules (buy or sell a beta estimates suggest that investing in for-
currency when it breaks above or below the eign currency provides a good hedge for an
maximum or minimum level, the resistance investor whose portfolio is primarily centered
level, over a stipulated recent period); and on U.S. stocks (see also Neely 1997).
channel breakout rules (but or sell a curren- The first study systematically integrating
cy when it breaks out of a channel, defined risk-adjustment into the empirical examina-
as occurring when the high price of a foreign tion of certain rules, however, was due to
currency over the previous n days is within x Sweeney (1986). This study characterizes a
percent of the low over the previous n days). quite different approach to that of Cornell
Using standard tests, Qi and Wu’s results and Dietrich (1978), as it compares trading
indicate significant profitability of moving rules based on technical currency analysis
average and channel breakout rules for rules to buy-and-hold strategies. If, for
seven dollar exchange rates. They then apply example, deviations from uncovered interest
White’s (2000) reality check bootstrap rate parity (the condition that the expected
methodology to evaluate these rules and to excess return, net of interest rate carry, from
characterize the effects of potential data- buying and holding foreign currency should
snooping biases. They find significant prof- be zero) simply represent risk premia, then a
itability at the one percent level for all seven possible implication is that apparently prof-
currencies even, after data-snooping biases itable technical trading rules are simply pick-
(as well as transactions costs) are properly ing up these risk premia. Sweeney indeed
taken into account (Park and Irwin 2005 find calibrated his work under the assumption of
a similar result for euro and yen futures). a constant risk premium, i.e., the average
Moreover, employing the Japanese yen or return on foreign exchange holdings is
the German mark as a vehicle currency adjusted by the foreign–domestic interest
(instead of the U.S. dollar) yields even rate differential. Then the excess return on
stronger results. following the technical analysis rule, i.e.,
Even if the existence of significantly prof- gross return minus return from buy-and-
itable technical trading rules can be estab- hold, is adjusted by the share of days that the
lished, however, there is still the possibility trading rule is invested in foreign currency
that all that is being measured is a risk premi- and has thus to earn a risk premium.
um, so that the risk-adjusted returns from the According to this procedure, Sweeney did
rule would on average be non-positive. Table not find a risk-based explanation for excess
4 revealed already that earlier studies usually returns. Levich and Thomas (1993), apply-
ignored this issue but more recent studies ing a similar methodology, found a similar
have elaborated on it (see table 5). The pio- result.
neering attempt in this respect is Cornell and However, these results may be ques-
Dietrich (1978) who suggest a risk adjustment tioned on at least two grounds. First, it is
according to the international capital asset not clear why one should expect a positive
pricing model (ICAPM). Their empirical
realization is limited, however, by practical
20 From today’s perspective, the choice of a U.S. port-
constraints: first, the world portfolio is prox-
folio may seem less of a shortcoming, taking account of
ied by a U.S. market stock index (the S&P the well-documented preference of investors for home
500) and, second, they generally calculate the assets (Karen K. Lewis 1999).
TABLE 5
SUGGESTED RISK ADJUSTMENTS IN ASSESSING TECHNICAL ANALYSIS’ EXCESS RETURNS
Study period number standard of risk adjustment risk-adjusted

covered of cases(1) comparison excess returns
Cornell and 1973–75 6 S&P 500 beta of currency with +
Dietrich (1978)(2) S&P 500
Sweeney (1986) 1973–80 70 B&H (buy constant risk premium +
and hold) equivalent to uncovered
interest parity-notation
Taylor (1992) 1981–87 16 S&P 500 B&H beta with S&P 500; +
time-varying risk premia +
estimated on AR(1) premia
processes and the UIP
Menkhoff and 1981–91 129 B&H Sharpe ratio; risk–return +
Schlumberger ratio of monthly return −
(1995) differences against B&H
Kho (1996) 1980–91 72 MSCI (in excess covariation of currency −
of one week $ returns with world market
interest rates) portfolio excess returns
Chang and 1973–94 24 S&P 500, Sharpe ratio with S&P 500; +
Osler (1999) Nikkei, DAX beta with national index +
Neely (1997) 1974–97 40 S&P 500 Sharpe ratio; beta +
with S&P 500 +
(1)
Notes: Cases are the product of currencies times rules times models (if applicable).
(2)
Incomplete documentation of results; favorable outcomes refer to ex post selection of best technical
analysis rules.
risk premium for investing in one currency be a reasonable constellation of parameters

in a bilateral exchange rate as this implies for the time-varying risk premium which
that investment in the second currency (or would be needed to explain observed
a short position in the first currency) earns returns as a compensation for risk (see also
a negative risk premium.21 Second, the Okunev and White 2003). On the other
assumption of a constant foreign exchange hand, all that this evidence may be reveal-
risk premium is not a very realistic one ing is that the wrong parameterization of
(Taylor 1995b). The first study to relax this the risk premium was assumed.
assumption in the context of technical trad- A much more extensive approach in deriv-
ing rules was Taylor (1992), who allows for ing time-varying foreign exchange risk pre-
a time-varying risk premium in the form of mia in this context is adopted by Bong-Chan
a first-order autoregressive process. Kho (1996). He relates possible excess
Parameter values of this process, justified returns to a world stock portfolio (the MSCI
by results from other studies, are used to index) in a conditional ICAPM framework.
enter into a pricing model. For several Within his framework, there are basically
combinations of parameter values, hun- three factors which are assumed to deter-
dreds of time series are then simulated on mine world excess returns: interest rate dif-
which technical analysis rules are evaluat- ferentials against the U.S. dollar, the
ed. It is found that there does not seem to conditional variance of world excess returns
and a moving average term. The empirical
21 We thank two of the anonymous referees for encour- work uses econometric models in which the
aging us to make this argument. conditional variance is allowed to affect the
conditional mean (i.e., GARCH-m models) holding a broad portfolio index such as the
in order to calculate expected risks. Kho market index (e.g., Neely 1997; Chang and
finds that much of the technical analysis Osler 1999; LeBaron 2000; Saacke 2002).23
returns can indeed be explained as compen- The results of these studies show higher risk-
sation for the high risks involved. adjusted returns to technical analysis rules
The above approaches to incorporate risk than to the benchmark portfolios.24
into profitability measurement implicitly The popular Sharpe or information ratio
need a benchmark model of asset pricing in (IR) has its own problems, however, when it
equilibrium. Obviously, the ICAPM is most is used as a criterion by which to measure
popular in this respect although empirical the performance of trading rules. Suppose
finance may tentatively prefer multifactor that the mean excess return of the trading
models, such as the Fama and Kenneth R. rule over a period of T years is −, with stan-
French (1996) approach. From a theoretical dard deviation . Then the IR will be
point of view consumption-based asset pric- defined as
− .
ing seems more advantageous to the CAPM
(John H. Cochrane 2005). However, neither IR ≡ ⎯
of these approaches has been applied to the
foreign exchange market.22 Given the failure Now, it can easily be shown that
in identifying meaningful time-varying risk τ = √T × IR is approximately equal to a t-ratio
premia in international finance in general for a test of the hypothesis that the excess
(Taylor 1995b), this shortcoming may be return is zero.25 A common benchmark for a
excusable. This lack of knowledge has, more- “good” trading rule in the finance industry in
over, fueled other ways of addressing the general is an IR of 0.5 (see, e.g., Richard C.
riskiness inherent in the use of technical Grinold and Ronald N. Kahn 2000). But this
analysis. means that an IR of 0.5 must be sustained
Some studies circumvent the problem of over about eleven years before the trading
measuring the world portfolio and deriving rule can be said to have generated excess
risk premia. Instead, they directly compare returns significantly greater than zero at the 5
the return–risk profile of a speculative cur- percent significance level, since this would
rency portfolio to a benchmark portfolio by give a value of τ(= √11 × 0.5 = 1.658,)
using the ratio of annualized excess returns greater than the critical value for a one-
(relative to a benchmark strategy) to the sided test at the 5 percent level (i.e., 1.645).
standard deviation of those returns, i.e., the Suppose that a trader selects a certain rule
Sharpe or information ratio (William F. because it has an IR of 0.5 according to a
Sharpe 1966). Alternative benchmarks in “backtest” with ten or more years of data.
this respect are either a buy-and-hold cur- As we discussed earlier, there is a strong
rency strategy (e.g., Menkhoff and likelihood that the rule will be subject to
Schlumberger 1995) or the return from data-snooping, and a true out-of-sample
22 It is interesting to note in this respect that these alternative to another investment. Accordingly, such
more advanced approaches are also confronted with evi- benchmarks should not be taken literally.
dence that questions their explanatory power (Jonathan 24 In this vein, Dewachter and Marco Lyrio (2005) find
Lewellen, Stefan Nagel, and Jay Shanken 2006). that the application of moving average rules can provide a
23 These alternative benchmarks are second-best solu- significant return to investors.
tions adopted from the equity market literature. Thus, the 25 This result is independent of the distribution of
buy-and-hold benchmark implicitly uses a national, one- excess returns and follows from the Central Limit
sided perspective whereas trading rules in foreign Theorem, which states that whenever a random sample
exchange are typically symmetric with respect to the two of size T is taken from any distribution with mean and
currencies involved. Regarding the index benchmarks, variance ζ2, then the sample mean will be approximately
they implicitly assume that the trading rule would be an normally distributed with mean µ and variance ζ2/T.
test would require that the trader keeps the Maureen O’Hara, and P. S. Srinivas 1998).
rule and monitors its performance over the Some evidence supporting this view is pro-
ensuing ten or eleven years or so—which is a vided by the fact that when the authors
very long time in the financial markets. exclude the fifty largest absolute daily
This picture changes, however, if one returns, all of the trading rules incorporating
measures risk not in the traditional sense of the open interest differential become less
the variability of returns but if one tries profitable (implying that the excess returns of
instead to integrate the professionals’ per- technical trading rules to some extent reflect
ception of risk as relating to relative per- compensation for risk), but many of them
formance in comparison with the market nevertheless remain strongly profitable.
(see, e.g., Goodhart 1988, p. 457). Here, SF6 In summary, looking at the last column of
comes into play, namely that profitability is table 5, where risk-adjusted profitability is
unstable over time. In summary, applying displayed, the majority of studies conclude
technical analysis involves a high probability that the profitability of technical currency
of making “wrong” decisions, i.e., perform- analysis holds in a risk-adjusted sense. Going
ing below the market, at least during some more into detail, there is, first, evidence that
periods (see, e.g., William L. Silber 1994, p. time-varying risk premia might explain some
44; Neely 1997). Thus, Menkhoff and of the excess return of technical analysis but
Schlumberger (1995) suggest addressing the not all or even most of it (Taylor 1992; Kho
risk inherent in using technical analysis by 1996).26 Furthermore, even the correct
focusing on the monthly difference between determination of appropriate risk premia is
the rules’ profitability performance and a questionable with the present state of knowl-
buy-and-hold performance (this may be edge. A second line of argument is that it is
understood as a form of myopic loss aver- perhaps possible to explain some of the
sion, see Shlomo Benartzi and Richard H. excess returns using a measure of risk as per-
Thaler 1995). Due to the high level of insta- ceived by market participants. Indeed, the
bility of technical analysis’ returns, the available evidence indicates that technical
“excess return”—as shown by raw returns or analysis is quite risky in this respect.
by a Sharpe ratio criterion—ceases to be sig-
nificant at the 5 percent level. 5. Explaining the Continued Use of
In a recent paper, Maxime Charlebois and Technical Analysis in the Foreign Exchange
Stephen G. Sapp (2007), using daily data on Market
dollar–mark over the period 1988–98, find
Technical analysis is an important tool in
that moving-average trading rules generate
real-world decision making in foreign
significant excess returns and that the excess
exchange markets (SF1, SF2, and SF3). In
returns increase when information is includ-
addition, it appears that applying certain
ed on the open interest differential on cur-
technical trading rules to volatile foreign
rency options (i.e., the net difference
exchange markets over a sustained period
between the cumulative value in dollar terms
may lead to significant positive excess
of all put options that are still active on a
given day less the cumulative value of all
active call options). They interpret this as 26 One must admit, however, that there is not much
partly reflecting risk premia and partly as guidance as to whether the measures of risk premia used
in the empirical literature are fully convincing from a the-
reflecting extra fundamental information that oretical point of view. For example, Taylor’s (1992) AR(1)
is reflected in options prices, since options risk premium model may simply be too restrictive,
may be the instrument of trading of choice of while—as an anonymous referee has commented—the
Kho (1996) study is based on a limited sample and has not
more informed traders because of the lever- to date been replicated for other currencies or sample
age advantage provided (David Easley, periods.
returns, although it is not clear that the per- directly interested in generating profit but
formance of these rules is stable over time or nevertheless have a significant influence on
that the excess returns earned significantly the market, then these participants may
outweigh the associated risk premia (SF4, generate profit-making opportunities for
SF5, and SF6). technical analysts over sustained periods of
There remains a need for further explana- time, allowing them to survive in the mar-
tion of the continued and passionate obses- ket. One such group that has been proposed
sion of foreign exchange professionals with in this context is comprised of the major
technical analysis, however, as profitability central banks, and the behavior of central
studies have not to date arrived at a clear banks in intervening in the foreign exchange
verdict. The organizing idea here is to allow market has been posited as a second expla-
explicitly for heterogeneous agents and nation for the persistence of technical
asymmetric information in the foreign analysis.
exchange market, which makes market effi- A third position is that if it takes time for
ciency a more complex concept. It may, the effects of economic fundamentals to
however, be reassuring in this context that feed through fully into market exchange
this complexity can indeed by rooted in rates, then technical analysis may serve as a
Fama’s (1970, p. 388) seminal paper on means of detecting these kinds of influ-
financial market efficiency, as he discusses ences earlier than would otherwise be the
“disagreement among investors about the case.
implications of given information” as a Fourthly and finally, it has often been
potential source of “inefficiency.” So, in what argued that financial prices may not only
way may disagreement (i.e., heterogeneity) reflect the information from fundamentals
help in resolving our puzzle? We group the but also influences from other sources, such
various explanations that have been suggest- as the influence of noise traders or the self-
ed into four positions, which we shall briefly fulfilling influences of technical analysis
describe before we relate them to rational itself.
behavior of agents and efficient markets (see Among these four explanations, it is only
figure 3 for an overview). the first that directly refers to irrational
If one follows the traditional understand- behavior of agents: either the users of tech-
ing of the EMH and regards foreign nical analysis are simply irrational and will be
exchange markets as at least weakly efficient driven out of the market (as suggested by
in the sense of Fama (1970)—i.e., in the Milton Friedman 1953) or they systematical-
sense that significant profits cannot be gen- ly underestimate risk (as suggested by J.
erated using forecasts based on past price Bradford De Long et al. 1990). The other
movements alone—then one would assess three explanations do not rely on techni-
the use of technical analysis as evidence of cians’ irrationality but on Fama’s (1970)
irrational behavior. This is the first explana- argument that not all market participants
tion for the continued use of technical analy- need to interpret all information at the same
sis in the foreign exchange market. time in the same way.
However, the assumption that most pro- With regard to the foreign exchange inter-
fessionals in the market behave consistently vention explanation, it would be the central
irrationally does not fit the EMH either: bank that distorts markets and technical
according to the EMH, they should quickly traders profit from this “inefficiency.”
be driven out of the market as they make Regarding the third explanation, technical
losses at the expense of rational traders. analysis is seen as an instrument via which to
But, if there is an important set of foreign learn about the revelation of fundamentals
exchange market participants who are not that cannot be recognized from observing
temporarily suboptimal
behavior
as an indication of not- users underestimate

fully-rational behavior the risk involved
a marketing instrument to
impress noninformed
clients
as a means of profiting
from foreign exchange
intervention
four positions
explaining the use
of technical
currency analysis time-consuming revelation
of fundamentals
as a means of
processing
fundamental using patterns from order
influences on flows
exchange rates
sentiment, psychological
as a means of influences on prices
processing information
on nonfundamental
influences on technical analysis as a self-
exchange rates fulfilling decision process
Figure 3. An Overview of Explanations for the Use of Technical Analysis on Foreign Exchange Markets
fundamentals directly.27 Here, neither tech- there are not-fully-rational traders in the mar-
nical traders nor the market need be ineffi- ket who have price impact and whose behav-
cient except according to a very strict form of ior can be detected and exploited by technical
the EMH requiring that market prices should analysis. Obviously, in this case markets are
reflect new information instantaneously; in not efficient and technicians who are rational
the real world, it takes time to learn and tech- in the sense of exploiting all available infor-
nical analysis may be one method of learning. mation for trading purposes (whether it is
Finally, in the fourth strand of explanations, information about fundamentals or nonfunda-
mental influences) will profit at the expense
of noise traders who are irrational in the sense
27 We agree with an anonymous referee that order flow
of not using all available information.
seems to have a fundamental component and appears to
be related to movements in fundamentals (Evans and It seems noteworthy that, in the three
Lyons 2005b). latter explanations discussed here, the
commonality is that there are price-rele- Regarding the first argument, concerning
vant influences that cannot be addressed temporarily irrational behavior, one would
by conventional fundamental analysis, need information about the behavior of par-
either because the central bank intervenes ticipants in the time-series domain to test
or because fundamentals cannot be directly whether behavior changes over
observed or else because nonfundamentals time—this data is not available so far. An
impact on prices.28 By using these pieces alternative is to test the cross-sectional impli-
of “information” technical analysis does not cations of this approach, for example that
necessarily yield excess returns. traders relying on technical analysis tend to
We review below the available empirical be less experienced and will in some sense
evidence with respect to each of these four learn to use fundamental analysis as their
positions. experience grows over time. Learning means
here the same as learning the “right model,”
5.1 Technical Analysis as Reflecting
i.e., the lesson of avoiding technical analysis
Irrational Behavior
in the future. Another implication seems to
The charge of not-fully-rational behavior be that not-fully-rational behavior will not
on the part of those applying technical analy- lead to market success, so that chartists do
sis is probably the most common position in not reach senior positions as often as others.
explaining its use, since in its reliance on Finally, nonrationality may be a consequence
extrapolation and/or visual pattern recogni- of a lower level of education, since it may be
tion, technical analysis is inconsistent with argued that technical analysis does not
weak efficiency of the foreign exchange mar- require any level of economic understand-
ket. However, as mentioned earlier, this ing but is—quite the contrary—easily
position has the paradoxical implication that understandable on an intuitive basis.
the market is in fact not efficient since tech- These three implications of the assump-
nical analysis is so widely used in the market tion of temporary irrationality on the part of
(SF1). Thus, there must be more subtle rea- traders using technical analysis have been
sons for using technical analysis rather than tested with survey data of Gehrig and
just an outright lack of rationality. In effect, Menkhoff (2006). The details, given in table
there seem to have been three arguments 6, reveal that those market participants who
put forward in the literature. prefer the use of technical analysis are not,
First, that the irrational behavior may be in fact, characterized by symptoms of a pos-
of a largely temporary nature. sibly suboptimal behavior (see also
Second, that it may be the case that users Menkhoff 1997; Cheung and Clement Yuk-
of technical analysis systematically underes- Pang Wong 1999; Cheung, Chinn, and
timate the risks involved in its use. Marsh 2004).
Third, that the application of technical This leads to the next argument, put for-
analysis may in fact be a form of marketing ward by De Long et al. (1990) in the more
or “window dressing” on the part of financial general context of noise traders, that the
institutions in order to impress and attract application of technical analysis may be
less-informed clients. related to an underestimation of the risk
involved by its users. Again, there is no
28 It is thus only the fourth explanation that requires the
direct evidence available which could inform
presence of outright nonfundamental forces in the market. about risk preferences and risk perception of
Intervention itself, referring to the second explanation, chartists. Moreover, the studies examining
may react on nonfundamental prices or create them. The risk-adjusted profitability do not come to a
third explanation may be related to nonfundamental prices
when market participants are affected by a preference for unanimous conclusion (see section 4). The
certain figures. only study that directly compares the
TABLE 6
THE USE OF TECHNICAL ANALYSIS AS A SIGN OF TEMPORARILY SUBOPTIMAL BEHAVIOR
Hypothesis being tested aggregated figures Pearson χ2 probability

for chartists v. others
1. Chartists have the same younger than 35 years: 0.645 (0.419)
age as other market chartists 55.56% v.
participants. others 49.61%
2. Chartists reach senior senior positions reached: 1.395 (0.237)
positions as often as chartists 31.94% v.
other market participants. others 24.22%
3. Chartists have achieved the university level achieved: 1.254 (0.263)
same level of education as chartists 24.64% v.
other market participants. others 32.28%
Notes: The source is Gehrig and Menkhoff (2006). Chartists are defined as respondents who attach a greater
weight to technical analysis than to either fundamental or flow information. The number of chartists according to
this criterion was 72 and the number of other market participants was 129 (exact numbers may differ slightly due
to incomplete replies). The achieved university level compounds graduation from university as well as from uni-
versity of applied sciences. The χ2-test exploits not only the aggregated figures being presented here, but all
available information.
consequences of relying on technical rather There is, finally, a third argument in favor
than fundamental analysis is that of Curcio of not-fully-rational behavior on the part of
and Goodhart (1993). In their experiment, technical traders—the marketing argument.
the profit of technical traders and of funda- The claim here is not that technical analysis
mentalists was similar but the volatility was can provide any useful information in fore-
lower for the users of technical analysis. This casting but that it generates buy and sell sig-
might indicate, therefore, that technical nals which translate into fee and commission
analysis can in fact serve as a risk-reducing income for financial intermediaries (see,
instrument. As this is, however, only a single e.g., Richard Sylla 1992, p. 343). This view
study, the significance of this result should may characterize the motivation of those
not be overstated. selling technical analysis, but it does not
Fortunately, there is another piece of evi- explain why others buy such services. If
dence which can be drawn from the form of technical analysis were particularly popular
technical analysis that is preferred. Both with small investors or other less profession-
studies asking this question come to the al market participants (e.g., “day traders”),
same conclusion that trend-following forms this argument would come close to the first
dominate rate of change indicators (see argument discussed above, i.e., that of sub-
Taylor and Allen 1992, table 1A; Lui and optimal behavior. Unfortunately, there is no
Mole 1998, table 4). If one assumes that evidence that small investors are in fact par-
most people would regard “going with the ticularly heavy users of technical analysis,
wave” as less risky than betting against it, this although it is known that a large number of
preference of available instruments does not professionals adhere to this tool. In addition
indicate risk-loving behavior. to the evidence presented in section 3 it can
Overall, therefore, the evidence presented be said that, according to the survey of
is unavoidably thin. Nevertheless, available Taylor and Allen (1992, table 1), most insti-
information does not support the notion of tutions subscribe to some form of external
chartists being a selection of people who chartist advice. Moreover, about 25 percent
underestimate risk in general. employ an in-house technical analyst in
comparison to 39 percent who employ an in- to address the issue of a possible third factor
house economist (Taylor and Allen 1992, in this analysis, LeBaron undertakes several
table 2). checks to investigate the existence of com-
In summary, the evidence regarding the mon factors that might drive interventions
not-fully-rational behavior position in explain- and profitability of technical analysis at the
ing the use of technical analysis is mostly same time. One finding is that periods of
quite indirect. Nevertheless, it is interesting highest expected volatility (calculated using
to note that available information points GARCH models) are not those of highest
against rather than in favor of this position.29 profitability of the moving average rules.
The thrust of this literature suggests that
5.2 Technical Analysis as Exploiting the
official interventions may distort the rela-
Impact of Central Bank Interventions
tionship between standard fundamentals
and exchange rate movements and thereby
As the central bank is not part of the reg-
disadvantage fundamentals-oriented traders
ular market process and, in particular, for-
while possibly favoring technical traders, for
eign exchange intervention is not generally
example if the intervention creates trends in
motivated by profit considerations, the
the exchange rate or support and resistance
process of central bank intervention in the
levels.
foreign exchange market may provide an
LeBaron’s (1999) analysis has been
explanation as to why financial markets are
extended by Saacke (2002). Saacke not only
actually efficient although excess returns
considers U.S. data but also interventions by
can be earned. This idea was formulated
the Deutsche Bundesbank. Moreover, this
long ago—see, for example, Dooley and
study covers two additional years and con-
Shafer (1983, p. 65), Levich (1985), or
siders a wide range of technical analysis rules
Sweeney (1986)—but it had not been tested
and confirms LeBaron’s findings.
in a rigorous way until quite recently.
Other studies that make similar argu-
The seminal paper in this respect is the
ments concerning the influence of central
study by LeBaron (1999). He applies a sim-
bank intervention on technical analysis prof-
ple moving average rule to a fourteen-year
itability include Silber (1994) and Andrew C.
period (1979 to 1992) of daily as well as
Szakmary and Mathur (1997). Silber (1994)
weekly time series of D-Mark/U.S.-Dollar
generally links markets where technical
and Yen/U.S.-Dollar exchange rates. This
analysis is profitable with the fact that these
rule generates considerable returns—of the
are the markets where central banks inter-
order of more than 5 percent per year
vene. Szakmary and Mathur (1997) examine
(LeBaron 1999, table 4). LeBaron calcu-
five major foreign exchange markets but rely
lates, however, the effect of removing days
on the IMF’s International Financial
when official foreign exchange interventions
Statistics to infer the degree of intervention
took place. The result is that the formerly
from data on foreign exchange reserves.
highly profitable technical analysis rules
These monthly figures cannot reveal higher
diminish in attractiveness (LeBaron 1999,
frequency interventions and are influenced
figure 2). This indicates that intervention
by nuisance components, such as revalua-
“has something to do with the observed pre-
tions or interventions in third currencies.
dictability” (LeBaron 1999, p. 137). In order
Interestingly, however, they nevertheless
reach basically the same conclusion as
LeBaron (1999).
29 Despite this interpretation of the available systemat-
This interpretation—i.e., that central
ic evidence, we do not wish to claim that there is no irra-
tionality in the market (Oberlechner 2004; Oberlechner bank intervention may be the source of the
and Osler 2006). profitability of technical analysis—has been
cautiously questioned by Neely (1998). He exchange intervention data from the

stresses the point that, due to LeBaron’s Federal Reserve and the Bundesbank, Reitz
methodology, most of the profits from tech- and Taylor (forthcoming) analyze the inter-
nical analysis rules occur concurrently with action of chartism, fundamentalism, and
intervention operations (see Neely 1998, p. central bank intervention and provide evi-
7f). If official interventions tend to occur dence that intervention is most likely to
when markets are trending, this would also occur and to be effective after a period of
explain the findings of LeBaron and others. sustained trending away from the equilibri-
In particular, if intervention days are those um level suggested by purchasing power
where markets move heavily (and might parity. They argue that this is evidence of
possibly move even more strongly without the “coordination channel” of intervention
interventions), then interventions and tech- effectiveness, which has been put forward
nical analysis profitability may be positively by Taylor (1995a, 2004) and Sarno and
correlated. The decisive step necessary to test Taylor (2001a). According to the coordina-
this competing interpretation is the use of tion channel, if technical analysts are capable
intradaily as opposed to daily data. of driving the exchange rate away from its
Neely (2002) has performed this task by fundamental equilibrium level over a sus-
combining several sources of daily data avail- tained period, then fundamentalist analysis
able at different times during the trading will not be profitable and fundamentalists
day. He checked the timing of technical will lose credibility in the market, or confi-
analysis profitability and intervention for five dence in the fundamentals. Hence, funda-
exchange rates, mostly over the period from mentalists will reduce their trades based on
1983 to 1998. Neely finds that “intervention fundamental analysis and the exchange rate
reacts to the same strong short-run trends will tend to stick away from (and perhaps still
from which the trading rules have recently trending away from) the fundamental equi-
profited” (Neely 2002, p. 230). The result is librium. (This is an example of the “limits of
confirmed for a high-frequency analysis of arbitrage” effect, as suggested in a more gen-
Bundesbank interventions (Michael Frenkel eral setting by Andrei Shleifer and Robert W.
and Georg Stadtmann 2004). It is also com- Vishny 1997.) When this occurs, the central
patible with Neely and Weller’s (2001) result bank may at some point intervene publicly in
on daily data, namely that their genetic pro- the hope that the intervention will act as a
gramming rules are most profitable on the coordinating signal to fundamentalists to
day before interventions take place.30 enter the market at the same time and so
Moreover, information about central bank return the exchange rate to its fundamental
information does not increase profitability. equilibrium level. To the extent that funda-
In a recent study using daily data on the mentalists rally to the central bank’s clarion
mark–dollar exchange rate and foreign call, the intervention will then be effective.
Using a nonlinear microstructural model of
exchange rate behavior, Reitz and Taylor
30 One reason that technical analysis may be profitable
(forthcoming) find evidence supportive of
before interventions was revealed by Bettina Peiers
(1997), indicating that one bank had superior forecasting the existence of a coordination channel of
performance with respect to later interventions. It seems intervention effectiveness.
plausible that this bank had an information advantage, so The coordination channel therefore pro-
that these profits may be due to private information—this
would not reject market efficiency in its semistrong form, vides a rationale as to why intervention, the
i.e., relying on the use of publicly available information. use (or profitability) of technical analysis,
Probably, this finding should not be generalized to other and trending exchange rates may all coin-
banks as the persistent forecasting power could not be
found for other situations (Sapp 2002; Kathryn M. E. cide. Note, however, that the coordination
Dominguez 2003). channel implies that intervention may be
effective because technical analysis is effec- David P. Brown and Robert H. Jennings
tive in generating a sustained trend away 1989).31
from fundamentals, not vice versa. The decisive point in this connection, how-
A final piece of evidence on the relation ever, is whether or not this reasoning has any
between intervention and technical analysis resemblance with the real-world conditions
profitability is provided by Sapp (2004). He of foreign exchange markets. Some evidence
finds that market uncertainty—measured by that this is indeed the case is provided by
spread and volatility—is high before inter- Sager and Taylor (2004) and Melvin, Sager,
ventions and lower afterwards. This provides and Taylor (2006), who show, using five-
an economic rationale for interventions (see minute data on dollar–sterling and
also Alain Chaboud and LeBaron 2001) and dollar–euro exchange rates, that there is an
indicates that profits earned by technical upward shift in exchange rate volatility fol-
analysis during these periods may be a lowing the interest rate announcements of
compensation for risk. the European Central Bank and the Bank of
England, suggesting a period of learning.
5.3 Technical Analysis as a Method of Earlier work supporting this notion includes
Information Processing Goodhart’s (1988) examination of exchange
rate changes in reaction to major fundamen-
Another explanation for the continued use tal news, which he assesses as initial underre-
of technical analysis is that it is in fact simply action (see Evans and Lyons 2005a). Further,
an instrument for processing and assimilat- a number of authors have recently analyzed
ing market information that is contained in the high-frequency reaction of foreign
exchange rates. The question concerning exchange markets to news announcements
how fundamental information is imparted more generally, and this work reveals that
into financial prices has long been a field of markets do react very quickly: most price
debate. If one leaves the macroeconomic reaction to scheduled news is in the form of
level and goes down to the actions of indian immediate jump (Torben G. Andersen et
vidual market participants at the microstruc- al. 2003). Unfortunately, however, these
tural level, it becomes clear that it will often price changes explain only a marginal frac-
be single entities or limited groups that rec- tion of overall price variability and even after
ognize or correctly interpret fundamental such marked jumps volatility remains persist-
changes earlier than others. These others ent for at least an hour—indicating that
may for some time interpret the actions of much more is going on in the market.32 Over
the better informed group as liquidity or longer horizons, by contrast, it is known that
noise trading, so learning takes time. exchange rates converge toward fundamental
Assuming that the fundamentally correct values (see, e.g., Nelson C. Mark 1995;
view succeeds in the end implies that there James R. Lothian and Taylor 1996).
exists an intermediate period during which
exchange rates move from the “incorrect” to
31 Nevertheless, there is always the possibility—as an
the “correct” level. Martin F. Hellwig (1982)
anonymous referee has pointed out—that information
was among the first to model this process processing can sometimes be linked to the “simple heuris-
and to note that it allows less informed tics” side of the psychology literature (e.g., Gerd
traders to infer information from observing Gigerenzer and Peter M. Todd 1999).
32 An anonymous referee has pointed out that the lim-
past price movements. Thus, on this argu- ited power of such studies to explain exchange rate behav-
ment, inferring future price movements ior even within short intervals may suggest that other
from past price movements, as in technical forces are at work. The role of technical analysis is unclear
in this respect as it could either be used as an instrument
analysis, may not be so irrational after all via which to assimilate information, or itself be a factor
(Jack L. Treynor and Robert Ferguson 1985; impeding the incorporation of fundamentals into prices.
Taken together, there emerges from this in the shorter term (see, e.g., Saacke 2002).
evidence a pattern whereby exchange rates This implies that fundamentals do not neces-
tend to react quickly but nevertheless may sarily feed immediately into exchange rates
underreact on the announcement of funda- and that technicians may try to exploit what
mental news. Despite their reversion to the they can learn as central banks intervene.
fundamental value over longer horizons, Third, it is a stylized fact that fundamental
there is an intermediate period where price exchange rate models fail empirically at short-
changes are imperfectly understood. The role er-term horizons (Sarno and Taylor 2002;
of technical analysis—in the form of trend- Cheung, Chinn, and Garcia Pascual 2005).
following signaling rules (e.g., moving average Nevertheless, recent research demonstrates
rules)—may therefore be to detect emerging the predictability of exchange rates over
shorter-term trends. There is some empirical shorter-term horizons (see, e.g., Richard H.
evidence consistent with this interpretation.33 Clarida and Taylor 1997; Clarida et al. 2003).
First, foreign exchange professionals show Interestingly, exchange rate predictability
a pattern in their expectations formation that appears to depend on two elements: first, the
clearly resembles this stylized pattern of term structure of interest rates (and therefore
reaction to fundamental news. In particular, the term structure of forward exchange rates)
they reveal bandwagon (highly extrapolative) may capture complex expectations and, sec-
expectations over horizons of a week to a few ond, the regime switching process may be
months tendency toward regressive expecta- related to different “environments” of
tions over longer horizons (Froot and Ito exchange rate determination. A fundamental
1989; Frankel and Froot 1990a, 1990b; Ito interpretation of these influences seems much
1990). Thus, evidence that appears hard to more difficult then the “direct” and athe-
reconcile with rational expectations may, oretical approach via technical analysis rules.
indeed, be evidence of learning. If the learn- Fourth, technical analysis rules have most
ing process means for example that informa- often been examined for shorter-term reac-
tion is increasingly imparted into prices, tions—for example, in the case of long-short
then extrapolative expectations and respec- moving average combinations, in the band
tive technical trading rules may have a between five to ten days for the short moving
rational basis. average and up to about 150 days for the long
Second, central banks that do not explicit- moving average. Saacke (2002, p. 464)
ly intend to make profits by intervening may demonstrates that this combination, both in
nevertheless do so (Sweeney 1997; Sarno and application by market practitioners and in aca-
Taylor 2001a). Insofar as central banks inter- demic studies, appears to fit the range within
vene, one may thus interpret their behavior which these rules are most profitable. Their
as tantamount to possessing knowledge application does not make sense at the very
about a true fundamental equilibrium short-term end or over longer horizons. This
exchange rate (or, at least, a range within is consistent with the position that technical
which the fundamental must lie)—thus “buy- analysis may be able to catch a sluggish and
ing low” and “selling high” to correct mis- then overshooting shorter-term adjustment of
alignments of fundamentals—and from exchange rates to fundamentals.
which they can profit in the long run but not A further strand of the literature has
focused on the information contained in
33 There are theoretical papers, such as Nicholas order flow that may help in understanding
Barberis, Shleifer, and Vishny (1998), which are capable exchange rate movements (e.g., Ito, Lyons,
of explaining the coexistence of short-term trends and and Melvin 1998; Lyons 2001; Evans and
longer-term mean reversion, although the “behavioral”
elements of these models have been criticised by, e.g., Lyons 2002). An important study in this con-
Fama (1998) and G. William Schwert (2002). nection by Osler (2003) demonstrates that
customer orders can be usefully linked to on a superficial level, it is order flow which
technical analysis. Her study uses data on generates the basis for the success of techni-
almost ten thousand conditional customer cal analysis but on a deeper level it is really
orders at a large U.S. bank over a period of customers’ preference for round numbers in
more than seven months in 1999–2000.34 In this context.36 If decisionmakers’ behavior is
particular, Osler constructs a limit-order subject to other habits and rules of thumb,
book, defined as the set of currency stop-loss then this may provide a basis for the use of
and take-profit orders existing at any point in other forms of technical analysis as means of
time, and finds that orders are not placed ran- exploiting movements in exchange rates gen-
domly but concentrate near “round” erated by non-fundamental influences (see
exchange rate values at “big figures” or “half the next subsection).
big figures” (such as a rate of 1.6100 or 1.6150 However, patterns in exchange rate move-
dollars per pound, rather than, say 1.6125 or ments may also reflect institutional design.
1.6133). The clustering of orders and the Thus it is known that some technical trading
respective behavior of three exchange rates is rules rely heavily on very specific prices dur-
indeed consistent with the predictions of ing the trading day—in particular the open-
technical analysis that, first, downtrends tend ing, closing, high, and low prices. One can
to be reversed at support levels and vice versa link these prices to order flows in the sense
and, second, that trends gain momentum that opening and closing prices may reveal
when support or resistance levels are crossed. more permanent demand and supply imbal-
A fruitful extension of this work, which has ances due to the need of many dealers to
not so far been examined for the foreign square their positions at the end of their day,
exchange market, refers to the ability of tech- whereas high and low prices may reveal a
nical analysis to locate order-book depth. In mismatch of buy and sell orders. The econo-
particular, in a study of order flow on the metric study of Fiess and MacDonald (2002)
New York Stock Exchange, Kenneth A. shows that analyzing these specific prices
Kavajecz and Elizabeth R. Odders-White can generate useful forecasts of exchange
(2004) show, first, that support and resistance rates (and volatility).37
levels of technical analysis are related to One could thus speculate whether similar
price levels in the order book where liquidity institutionally motivated effects might be
is very high (echoing Osler’s 2003 results) detected in the behavior of international
and, second, that simple moving average fund managers. It has been argued that equi-
indicators inform about the relative depth of ty market fund managers may have incen-
the order book on one or the other side, tives for herding and, moreover, that there is
which they call the “skewness of liquidity.” shorter-term momentum in the returns of
Overall, the conclusion that emerges from stocks which may be caused by herding
the research on limit-order books is that behavior (e.g., Mark Grinblatt, Sheridan
“technical analysis works because orders are Titman, and Russ Wermers 1995). Further,
clustered” (Osler 2003).35 More specifically, there are indications of short-term underre-
action to news and medium-term overreac-
34 Accordingly, this study focuses on higher frequencies tion, so that shorter-term momentum and
than most studies covered in section 4 on profitability
which use daily data, which apply rules changing positions
36 An anonymous referee has commented that a feed-
typically after weeks and which analyze years of trading.
35 Another clustering, one in the time dimension, is back channel may also exist if technical analysts use “big fig-
analyzed by Fabrizio Lillo and J. Doyne Farmer (2004) for ures” or “round numbers” in foreign exchange prices,
the London Stock Exchange. They find that the sign of which may motivate customers to place orders accordingly.
37 Popular technical trading rules relying on these
orders, i.e., either buy or sell, is positively correlated. It
seems intuitively possible to formulate technical trading specific prices are so-called candlestick formations or
rules in order to exploit such properties. stochastics indicators.
longer-term contrarian investment strategies inefficient in the sense that prices do not
appear to promise excess returns (e.g., only reflect fundamental information. A
Narasimhan Jegadeesh and Titman 2001). If prominent model in this vein is DeLong et
this behavior translates from stock prices to al. (1990), where not-fully-rational noise
foreign exchange, it may be responsible for traders create risks for rational investors
generating shorter-term trends, i.e., momen- with limited arbitrage capacity. The role of
tum, which may be detected by technical technical analysis in this environment is to
analysis rules (see Okunev and White 2003). provide an instrument to analyze and possi-
As these considerations are necessarily spec- bly forecast the behavior of noise traders.
ulative, it may be reassuring that the reverse Therefore, technicians are seen as rational
chain of argument has some substantiation: if agents who exploit noise traders without
this kind of behavior were able to generate necessarily bringing exchange rates closer
successful technical trading in the foreign to economic fundamentals.
exchange market, it should do so in the equi- The notion that there may be social psy-
ty market. Some studies lend support to this chological influences on financial markets
view of equity markets (e.g., William Brock, was noted over twenty years ago by Robert J.
Josef Lakonishok, and LeBaron 1992; Shiller (1984) and it has been investigated
Lawrence Blume, Easley, and O’Hara 1994; more systematically in later research.
Lo, Harry Mamaysky, and Jiang Wang 2000; Regarding foreign exchange markets, the
Kavajecz and Odders-White 2004). It is also conceptually similar survey studies of
interesting to note that “round” figures seem Cheung and Wong (2000), Cheung and
to play a prominent role in stock markets (R. Chinn (2001), and Cheung, Chinn, and
Glen Donaldson and Harold Y. Kim 1993). Marsh (2004) directly asked market practi-
This is, however, clearly an avenue for future tioners about their assessment of the price
foreign exchange market research. relevance of several factors that can be
linked to psychological influences on the
5.4 Technical Analysis as Providing
market. The result, presented graphically in
Information about Nonfundamental
figure 4, clearly shows the importance that
Exchange Rate Determinants
foreign exchange dealers attach to psycho-
The last position analyzed here—i.e., that logical forces in the very short run.
technical analysis may provide information Whereas these three survey studies ask for
about nonfundamental influences on the perceived importance of technical fac-
exchange rates—is quite common both in tors in competition with psychological fac-
the literature and among market practition- tors, the study by Gehrig and Menkhoff
ers. For example Taylor and Allen (1992, p. (2006) allows an analysis of the implicit rela-
311) mention two recurrent groups of com- tions between these two factors (the results
ments made by respondents to their survey hold for data from Menkhoff 1997). In par-
of London foreign exchange dealers, name- ticular, if one relates the weight given by
ly: “a belief that charts essentially measure traders to the use of technical analysis and
swings in market psychology” and “that chart the perceived importance of psychological
analysis may be largely self-fulfilling.” Both price influences, a positive correlation
views imply that exchange rates may not be becomes obvious (see the lightly shaded bars
exclusively dependent on the course of fun- in figure 5).
damentals but may also react to additional, A similar approach can be applied to learn
nonfundamental factors that are sometimes about the possibility of the self-fulfilling
loosely termed “market sentiment.” nature of technical analysis, which, to date,
This position clearly views the foreign has not been examined systematically. One
exchange market to a certain degree as reason why this issue has not been examined
90
weight of 80 unweighted average taken from Hong
psychological , Singapore,
Kong, , , Tokyo, United
influence 70 ,
Kingdom, and United States
in % extreme values
60
50
40
30
20
10
0
intraday medium long
Figure 4. The Perceived Importance of Psychological Influences on Exchange Rate Determination
Note: Data are taken from Cheung and Wong (2000, table 4C), Cheung and Chinn (2001, figure 8C) and Cheung,
Chinn and Marsh (2004, p. 305). Psychological influences is defined as the sum of “bandwagon effects,” “overreaction to
news,” and “speculative forces.”
may be an ex ante skepticism against this that the use of technical analysis has also
possibility, as it is well-known that chartists become more widespread over time, this
differ markedly in the instruments they use suggests that traders may indeed alter the
and even more so in their respective calibra- weight attached to technical analysis in
tion (e.g., the precise number of days used in accordance to its perceived forecasting
moving average rules), and may also display power (as originally suggested by Frankel
significant heterogeneity in their forecasts and Froot 1986, 1990a).
(Allen and Taylor 1990). Nevertheless, asking Additional indirect support for nonfunda-
those who use technical analysis their opin- mental price influences in foreign exchange
ion as to the self-fulfilling hypothesis leads to markets is provided by recent models with
a bimodal relation—i.e., there emerge two heterogeneous agents (Cars Hommes 2005).
views: either an “opportunists’ view” where- These models have in common that complex
by chartism is used because it is perceived as interaction between heterogeneous groups
self-fulfilling, or a “believers’ view” whereby of actors can successfully generate stylized
chartism is seen as an intrinsically valuable facts in financial markets, such as fat tails in
methodology rather than merely self- the distribution of returns and volatility clus-
fulfilling in its predictions (see the dark tering. What is remarkable here is the fact
shaded bars in figure 5). Interestingly, the that the most prominent way of describing
opportunists’ view appears to have gained the behavior of nonfundamentalists is to
ground over time (Menkhoff 1997).38 Given assume trend-following behavior, which is
often motivated by applying technical analy-
38 Of course, the self-fulfilling nature of technical
sis (Thomas Lux 1998; Frank H. Westerhoff
analysis cannot be an example of perpetual motion—
returns must ultimately be generated by trading with oth- 2003; De Grauwe and Grimaldi 2006b; Reitz
ers. Thus, this view is basically informative about the and Taylor forthcoming).
motivation of users. One may, however, speculate whether In a related study modeling the interaction
users of technical analysis who are motivated solely by the
fact that other market participants are using it may in of agents with different horizons, LeBaron
effect end up as market followers often do, i.e., as losers. (2001) finds that such a computational stock
Question: “How much importance do fundamentals and psychology have for exchange rate movements?”
• People are not machines; thus psychology is clearly more important than fundamentals.
Question: “What is in your opinion the value of technical analysis?”
• I regard technical analysis only because others regard it.
weight 50
given to psychology is technical analysis is
technical important self-fulfilling
analysis
45
(in %)
40
35
30
25
20
1 2 3 4 5 6
agree disagree
completely completely
Coefficient of rank correlation of higher weight given to technical analysis with
• “Psychology is important”: −0.220∗∗ (P = 0.002), n = 200
• “Technical analysis is self-fulfilling”: −0.069 (P = 0.334), n = 197
Figure 5. The Weight Given to Technical Analysis in Relation to a Psychological Rationale for Its Use and
the ‘Self-fulfilling’ Rationale for Its Use
Notes: Data are taken from Gehrig and Menkhoff (2006). Note that the y axis starts at 20 percent.
market produces some well-known charac- news. The most prominent relation of non-
teristics of foreign exchange as well. fundamentals news—derived from newswire
Different time scales of heterogeneous reports—is to technical analysis. This evi-
agents are an important subject in foreign dence indicates both that nonfundamental
exchange (Michael M. Dacorogna et al. influences may have significant short-run
2001) and seem to be intuitively directly effects on exchange rates and that they may
related to the relatively short-term oriented be related to technical analysis.
chartists as discussed above (stylized fact 3). In another recent innovative study,
In a recent study using intraday exchange Schulmeister (2006) analyzes marketwide
rate data, Dominguez and Freyan Panthaki relations between signals from technical
(2006) distinguish three categories of factors analysis (using 1024 different trading rules),
generating exchange rate movements: fun- exchange rates, and order flow. He finds that
damental news, order flow (reflecting pri- during most periods, the trading rules under
vate information), and nonfundamental consideration tend to be on the same side of
the market, and so may possibly be pushing fits well with the puzzle mentioned in the
exchanges rates by generating similar trad- introduction that shorter-term exchange rate
ing signals. The analysis is then extended to movements cannot be explained with exist-
include order flow for a three-month period ing fundamental models (the so-called “dis-
(the sample period being limited by data connect puzzle”). If there are other forces at
availability). During this three-month peri- work (see, e.g., Dominguez and Panthaki
od, the vast majority of the technical analysis 2006), this would help us solve the puzzle
considered tended to generate trading posi- and at the same time provide a rationale for
tions similar to those generated using signals the use of technical analysis.
based on order flow. Consequently, this
(admittedly short-sample) study suggests 6. Conclusion
that order flow could be the result of techni-
cal trading, in addition to the earlier exam- A reading of the literature on the nature
ined sources of information revelation and and use of technical analysis in the foreign
liquidity trading (Evans and Lyons 2005a). exchange market allows us to draw up a set
Thus, Schulmeister conjectures that techni- of stylized facts concerning its nature and
cal analysis may magnify otherwise small use, and also to distinguish a number of
exchange rate changes, contributing to the arguments that have typically been adduced
kind of short-term overreactions mentioned to explain its continued use.
above in a self-fulfilling fashion (see also Indeed, first and foremost among these
Farmer and Shareen Joshi 2002). stylized facts lies the continued and wide-
Although there seems to be some support spread use of technical analysis in the for-
for the position that technical analysis is an eign exchange market. Research conducted
instrument for processing and assimilating in most of the major foreign exchange mar-
information about nonfundamental price kets during the last decade or so reveals
determinants, it should be noted that the clearly that the use of technical analysis is an
evidence so far only reveals three facts. important and persistent phenomenon
First, that a significant proportion of foreign which is highly influential in the decision
exchange market participants believe in making of foreign exchange professionals. A
nonfundamental influences on exchange similar situation emerges with respect to the
rates. Second, that those who believe more profitability of technical analysis. It is
strongly rely more heavily on technical beyond question that, for major flexible
analysis. Third, that participants believing in exchange rates and over longer time periods,
technical analysis often ascribe its impor- the use of technical analysis may be used to
tance to its self-fulfilling nature. These facts provide very high returns. What is disputed,
would also be consistent with the view that however, is whether the realization of these
chartists do not understand the fundamen- profits has to be bought at the cost of taking
tal nature of exchange rates. However, there large risks and whether the profits can fully
are two further pieces of evidence that give compensate for this additional risk.
weight to the proposition under review in A contribution that we have sought to
this subsection. make in this paper is in relating the available
In particular, as discussed above in sub- empirical evidence to several positions that
section 5.1, there is no evidence that the use have been developed in order to explain the
of nonfundamental information—among continued use of technical analysis.
which technical analysis is perhaps the most The first of these—interpreting the use of
important—could be related to indicators of technical analysis as an indication of not-fully-
reduced rationality (see Menkhoff 1998). rational behavior—is difficult to reconcile
Moreover, the position under review here with the fact that virtually all professionals in
the market rely on this tool at least to a small fits well with the stylized fact on the higher
degree. Moreover, there is no hard evidence profitability of technical analysis in flexible
showing that chartists are characterized by exchange rate markets, as there is some indi-
temporarily suboptimal behavior, or underes- cation that these markets may be character-
timate the risk involved or accept technical ized by a degree of volatility that is hard to
analysis as a marketing instrument. explain by fundamentals alone (Robert P.
The second position, relating profitability Flood and Rose 1995).
to foreign exchange interventions by the This still leaves open, however, the ques-
monetary authorities, is a little more satisfy- tion of risk-adjusted profitability. If technical
ing in the sense that it suggests a more solid analysis has some rationale in the sense of
rationale for the use of technical analysis by being able to generate profitable trading
rational agents. Also, some stylized facts con- rules, why does the market process not
cerning the profitability of technical analy- assimilate or arbitrage these profit opportu-
sis—namely that it tends to be more nities away? The answer may be the same as
profitable during periods of official interven- with fundamental analysis: in well function-
tion—fit well with this position. There is, ing markets one would expect that profit
however, more recent evidence that suggests opportunities will be exploited up to an
that it may be large exchange rate move- extent where agents feel appropriately com-
ments themselves that may be leading both pensated for their risk. To take open posi-
intervention and technical analysis prof- tions is inherently risky, whether the
itability or, equivalently, that the influence of decision is based on fundamental or techni-
technical analysis, by driving the exchange cal considerations.
rate away from the level consistent with the What is perhaps most striking from our
fundamentals, may generate a rationale for reading of the literature, however, is that
official intervention, rather than vice versa, technical analysis remains a passionate
through the coordination channel of inter- obsession of many foreign exchange market
vention effectiveness. professionals; it is clearly an intrinsic part of
The third position, namely that technical this market. For academic researchers, this
analysis is simply an instrument in the pro- means that technical analysis must be under-
cessing and assimilation of market informa- stood and integrated into economic reason-
tion, can also reconcile the importance of ing at both the macroeconomic and the
order flows and technical analysis to some microstructural levels. For market practi-
degree. The main problem with this posi- tioners, it means that technical trading
tion, however, is that it does not explain the strategies should be constantly evaluated as
reason behind sluggish adjustment to news, potentially important tools in the search for
preferences for round figures in order excess returns.
placement, etc. REFERENCES
Overall, therefore, perhaps the most satis- Allen, Franklin, and Risto Karjalainen. 1999. “Using
fying explanation concerning the continued Genetic Algorithms to Find Technical Trading
use of technical analysis seems to be position Rules.” Journal of Financial Economics, 51(2):
245–71.
four, whereby technical analysis is seen as an Allen, Helen L., and Mark P. Taylor. 1990. “Charts,
instrument informing traders about nonfun- Noise and Fundamentals in the London Foreign
damental price determinants. These forces Exchange Market.” Economic Journal, 100(400):
49–59.
are more important in the shorter-run, so for Allen, Helen L., and Mark P. Taylor. 1992. “Chartist
a full understanding of exchange rate Analysis.” In The New Palgrave Dictionary of Money
dynamics, professionals need a combination and Finance, ed. Peter Newman, Murray Milgate,
and John Eatwell. London: Macmillan, 339–42.
of several tools, in particular both technical Andersen, Torben G., Tim Bollerslev, Francis X.
and fundamental analysis. This position also Diebold, and Clara Vega. 2003. “Micro Effects of
Macro Announcements: Real-Time Price Discovery Exchange Rate Dynamics.” Journal of International
in Foreign Exchange.” American Economic Review, Economics, 51(2): 401–19.
93(1): 38–62. Clarida, Richard H., Lucio Sarno, Mark P. Taylor, and
Bank for International Settlements. 2005. Triennial Giorgio Valente. 2003. “The Out-of-Sample Success
Central Bank Survey: Foreign Exchange and of Term Structure Models as Exchange Rate
Derivatives Market Activity in April 2004. Basel: Predictors: A Step Beyond.” Journal of International
Bank for International Settlements. Economics, 60(1): 61–83.
Barberis, Nicholas, Andrei Shleifer, and Robert W. Clarida, Richard H., and Mark P. Taylor. 1997. “The
Vishny. 1998. “A Model of Investor Sentiment.” Term Structure of Forward Exchange Premiums and
Journal of Financial Economics, 49(3): 307–43. the Forecastability of Spot Exchange Rates:
Benartzi, Shlomo, and Richard H. Thaler. 1995. Correcting the Errors.” Review of Economics and
“Myopic Loss Aversion and the Equity Premium Statistics, 79(3): 353–61.
Puzzle.” Quarterly Journal of Economics, 110(1): Clyde, William C., and Carol L. Osler. 1997. “Charting:
73–92. Chaos Theory in Disguise?” Journal of Futures
Bhanumurthy, N. R. 2004. “Microstructures in the Markets, 17(5): 489–514.
Indian Foreign Exchange Market.” Unpublished. Cochrane, John H. 2005. Asset Pricing: Revised
Blume, Lawrence, David Easley, and Maureen Edition. Princeton: Princeton University Press.
O’Hara. 1994. “Market Statistics and Technical Cornell, W. Bradford, and J. Kimball Dietrich. 1978.
Analysis: The Role of Volume.” Journal of Finance, “The Efficiency of the Market for Foreign Exchange
49(1): 153–81. under Floating Exchange Rates.” Review of
Brock, William, Josef Lakonishok, and Blake LeBaron. Economics and Statistics, 6(1): 111–20.
1992. “Simple Technical Trading Rules and the Cornell, W. Bradford, and Richard Roll. 1981.
Stochastic Properties of Stock Returns.” Journal of “Strategies for Pairwise Competitions in Markets
Finance, 47(5): 1731–64. and Organizations.” Bell Journal of Economics,
Brown, David P., and Robert H. Jennings. 1989. “On 12(1): 201–13.
Technical Analysis.” Review of Financial Studies, Curcio, Riccardo, and Charles Goodhart. 1993.
2(4): 527–51. “Chartism: A Controlled Experiment.” Journal of
Campbell, John Y., Andrew W. Lo, and A. Craig International Securities Markets, 7: 173–86.
MacKinlay. 1996. The Econometrics of Financial Curcio, Riccardo, Charles Goodhart, Dominique
Markets. Princeton: Princeton University Press. Guillaume, and Richard Payne. 1997. “Do Technical
Chaboud, Alain, and Blake LeBaron. 2001. “Foreign- Trading Rules Generate Profits? Conclusions from
Exchange Trading Volume and Federal Reserve the Intra-day Foreign Exchange Market.”
Intervention.” Journal of Futures Markets, 21(9): International Journal of Finance and Economics,
851–60. 2(4): 267–80.
Chang, P. H. Kevin, and Carol L. Osler. 1999. Dacorogna, Michael M., Ramazan Gencay, Ulrich A.
“Methodical Madness: Technical Analysis and the Mueller, Richard B. Olsen, and Olivier V. Pictet.
Irrationality of Exchange-Rate Forecasts.” Economic 2001. An Introduction to High-Frequency Finance.
Journal, 109(458): 636–61. San Diego: Academic Press.
Charlebois, Maxime, and Stephen G. Sapp. 2007. De Grauwe, Paul, and Danny Decupere. 1992.
“Temporal Patterns in Foreign Exchange Returns “Psychological Barriers in the Foreign Exchange
and Options.” Journal of Money, Credit and Banking, Market.” Journal of International and Comparative
39(2–3): 443–47. Economics, 1(2): 87–101.
Cheung, Yin-Wong, and Menzie D. Chinn. 2001. De Grauwe, Paul, and Marianna Grimaldi. 2006a.
“Currency Traders and Exchange Rate Dynamics: A “Exchange Rate Puzzles: A Tale of Switching
Survey of the US Market.” Journal of International Attractors.” European Economic Review, 50(1):
Money and Finance, 20(4): 439–71. 1–33.
Cheung, Yin-Wong, Menzie D. Chinn, and Antonio De Grauwe, Paul, and Marianna Grimaldi. 2006b. The
Garcia Pascual. 2005. “Empirical Exchange Rate Exchange Rate in a Behavioral Finance Framework.
Models of the Nineties: Are Any Fit to Survive?” Princeton and Oxford: Princeton University Press.
Journal of International Money and Finance, 24(7): De Long, J. Bradford, Andrei Shleifer, Lawrence H.
1150–75. Summers, and Robert J. Waldmann. 1990. “Noise
Cheung, Yin-Wong, Menzie D. Chinn, and Ian W. Trader Risk in Financial Markets.” Journal of
Marsh. 2004. “How Do UK-Based Foreign Political Economy, 98(4): 703–38.
Exchange Dealers Think Their Market Operates?” Deutsche Bank. 2005. Currency Alpha: An Investor’s
International Journal of Finance and Economics, Guide. London: Deutsche Bank, AG.
9(4): 289–306. Dewachter, Hans. 1997. “Sign Predictions of Exchange
Cheung, Yin-Wong, and Clement Yuk-Pang Wong. Rate Changes: Charts as Proxies for Bayesian
1999. “Foreign Exchange Traders in Hong Kong, Inferences.” Weltwirtschaftliches Archiv, 133(1):
Tokyo, and Singapore: A Survey Study.” In Advances 39–55.
in Pacific Basin Financial Markets, Volume 5, ed. T. Dewachter, Hans. 2001. “Can Markov Switching
Bos and T. A. Fetherston. Stamford, Conn.: JAI Models Replicate Chartist Profits in the Foreign
Press, 111–34. Exchange Market?” Journal of International Money
Cheung, Yin-Wong, and Clement Yuk-Pang Wong. and Finance, 20(1): 25–41.
2000. “A Survey of Market Practitioners’ Views on Dewachter, Hans, and Marco Lyrio. 2005. “The
Economic Value of Technical Trading Rules: A Multinational Finance Journal, 3(3): 147–72.
Nonparametric Utility-Based Approach.” Fiess, Norbert M., and Ronald MacDonald. 2002.
International Journal of Finance and Economics, “Towards the Fundamentals of Technical Analysis:
10(1): 41–62. Analysing the Information Content of High, Low
Dominguez, Kathryn M. E. 2003. “The Market and Close Prices.” Economic Modelling, 19(3):
Microstructure of Central Bank Intervention.” 353–74.
Journal of International Economics, 59(1): 25–45. Flood, Robert P., and Andrew K. Rose. 1995. “Fixing
Dominguez, Kathryn M. E., and Freyan Panthaki. Exchange Rates: A Virtual Quest for Fundamentals.”
2006. “What Defines ‘News’ in Foreign Exchange Journal of Monetary Economics, 36(1): 3–37.
Markets?” Journal of International Money and Frankel, Jeffrey A., and Kenneth A. Froot. 1986.
Finance, 25(1): 168–98. “Understanding the U.S. Dollar in the Eighties: The
Donaldson, R. Glen, and Harold Y. Kim. 1993. “Price Expectations of Chartists and Fundamentalists.”
Barriers in the Dow Jones Industrial Average.” Economic Record, 0(0): 24–38.
Journal of Financial and Quantitative Analysis, Frankel, Jeffrey A., and Kenneth A. Froot. 1990a.
28(3): 313–30. “Chartists, Fundamentalists and the Demand for
Dooley, Michael P., and Jeffrey R. Shafer. 1976. Dollars.” In Private Behaviour and Government
“Analysis of Short-Run Exchange Behavior: March, Policy in Interdependent Economies, ed. A. S.
1993 to September, 1975.” Board of Governors of Courakis and M. P. Taylor. Oxford; New York;
the Federal Reserve System, International Finance Toronto and Melbourne: Oxford University Press,
Discussion Papers, no. 123. Clarendon Press, 73–126.
Dooley, Michael P., and Jeffrey R. Shafer. 1983. Frankel, Jeffrey A., and Kenneth A. Froot. 1990b.
“Analysis of Short-Run Exchange Rate Behaviour: “Chartists, Fundamentalists, and Trading in the
March 1973 to November 1981.” In Exchange Rate Foreign Exchange Market.” American Economic
and Trade Instability, ed. David Bigman and Teito Review, 80(2): 181–85.
Taya. Cambridge, Mass: Ballinger, 43–69. Frankel, Jeffrey A., and Andrew K. Rose. 1995.
Dueker, Michael, and Christopher J. Neely. 2007. “Can “Empirical Research on Nominal Exchange Rates.”
Markov Switching Models Predict Excess Foreign In Handbook of International Economics, Volume 3,
Exchange Returns?” Journal of Banking and ed. G. M. Grossman and K. Rogoff. Handbooks in
Finance, 31(2): 279–96. Economics, vol. 3. Amsterdam; New York and
Easley, David, Maureen O’Hara, and P. S. Srinivas. Oxford: Elsevier, North-Holland, 1689–1729.
1998. “Option Volume and Stock Prices: Evidence Frenkel, Michael, and Georg Stadtmann. 2004.
on Where Informed Traders Trade.” Journal of “Trading Rule Profitability and Central Bank
Finance, 53(2): 431–65. Interventions in the Dollar–Deutschmark Market.”
Engel, Charles, and James D. Hamilton. 1990. “Long Jahrbücher für Nationalökonomie und Statistik,
Swings in the Dollar: Are They in the Data and Do 224(6): 653–72.
Markets Know It?” American Economic Review, Friedman, Milton. 1953. “The Case for Flexible
80(4): 689–713. Exchange Rates.” In Essays in Positive Economics.
Evans, Martin D. D., and Richard K. Lyons. 2002. Chicago: University of Chicago Press, 157–203.
“Order Flow and Exchange Rate Dynamics.” Journal Froot, Kenneth A., and Takatoshi Ito. 1989. “On the
of Political Economy, 110(1): 170–80. Consistency of Short-Run and Long-Run Exchange
Evans, Martin D. D., and Richard K. Lyons. 2005a. Rate Expectations.” Journal of International Money
“Do Currency Markets Absorb News Quickly?” and Finance, 8(4): 487–510.
Journal of International Money and Finance, 24(2): Gehrig, Thomas, and Lukas Menkhoff. 2004. “The Use
197–217. of Flow Analysis in Foreign Exchange: Exploratory
Evans, Martin D. D., and Richard K. Lyons. 2005b. Evidence.” Journal of International Money and
“Meese–Rogoff Redux: Micro-based Exchange-Rate Finance, 23(4): 573–94.
Forecasting.” American Economic Review, 95(2): Gehrig, Thomas, and Lukas Menkhoff. 2006.
405–14. “Extended Evidence on the Use of Technical
Fama, Eugene F. 1970. “Efficient Capital Markets: A Analysis in Foreign Exchange.” International
Review of Theory and Empirical Work.” Journal of Journal of Finance and Economics, 11(4): 327–38.
Finance, 25(2): 383–417. Gencay, Ramazan. 1999. “Linear, Non-linear and
Fama, Eugene F. 1998. “Market Efficiency, Long- Essential Foreign Exchange Rate Prediction with
Term Returns, and Behavioral Finance.” Journal of Simple Technical Trading Rules.” Journal of
Financial Economics, 49(3): 283–306. International Economics, 47(1): 91–107.
Fama, Eugene F., and Kenneth R. French. 1996. Gencay, Ramazan, Michael M. Dacorogna, Richard B.
“Multifactor Explanations of Asset Pricing Olsen, and Olivier V. Pictet. 2003. “Foreign
Anomalies.” Journal of Finance, 51(1): 55–84. Exchange Trading Models and Market Behavior.”
Farmer, J. Doyne, and Shareen Joshi. 2002. “The Price Journal of Economic Dynamics and Control, 27(6):
Dynamics of Common Trading Strategies.” Journal 909–35.
of Economic Behavior and Organization, 49(2): Gigerenzer, Gerd, and Peter M. Todd. 1999. “Fast and
149–71. Frugal Heuristics: The Adaptive Toolbox.” In Simple
Fiess, Norbert M., and Ronald MacDonald. 1999. Heuristics That Make Us Smart, ed. Gerd
“Technical Analysis in the Foreign Exchange Gigerenzer, Peter M. Todd, and the ABC Research
Market: A Cointegration-Based Approach.” Group. Oxford: Oxford University Press, 3–37.
Goodhart, Charles. 1988. “The Foreign Exchange Difficult to Beat the Random Walk Forecast of
Market: A Random Walk with a Dragging Anchor.” Exchange Rates?” Journal of International
Economica, 55(220): 437–60. Economics, 60(1): 85–107.
Goodhart, Charles, and Riccardo Curcio. 1992. “When Kozhan, Roman, and Mark Salmon. 2006. “On the
Support/Resistance Levels Are Broken, Can Profits Predictability of Tick by Tick Exchange Rates Using
Be Made? Evidence from the Foreign Exchange the Structure of the Order Book.” Unpublished.
Market.” Financial Markets Group Discussion LeBaron, Blake. 1999. “Technical Trading Rule
Paper, no. 142. Profitability and Foreign Exchange Intervention.”
Goodman, Stephen H. 1979. “Foreign Exchange Rate Journal of International Economics, 49(1): 125–43.
Forecasting Techniques: Implications for Business LeBaron, Blake. 2000. “Technical Trading Profitability
and Policy.” Journal of Finance, 34(2): 415–27. in Foreign Exchange Markets in the 1990’s.
Griffin, John M., Jeffrey H. Harris, and Selim Unpublished.
Topaloglu. 2003. “The Dynamics of Institutional and LeBaron, Blake, 2001. “Empirical Regularities from
Individual Trading.” Journal of Finance, 58(6): Interacting Long- and Short-Memory Investors in an
2285–2320. Agent-Based Stock Market.” IEEE Transactions on
Grinblatt, Mark, Sheridan Titman, and Russ Wermers. Evolutionary Computation, 5(5): 442–55.
1995. “Momentum Investment Strategies, Portfolio Lee, Chun I., and Ike Mathur. 1996. “Trading Rule
Performance, and Herding: A Study of Mutual Fund Profits in European Currency Spot Cross-Rates.”
Behavior.” American Economic Review, 85(5): Journal of Banking and Finance, 20(5): 949–62.
1088–1105. Levich, Richard M. 1985. “Empirical Studies of
Grinold, Richard C., and Ronald N. Kahn. 2000. Active Exchange Rates: Price Behaviour, Rate Determination
Portfolio Management: A Quantitative Approach for and Market Efficiency.” In Handbook of International
Producing Superior Returns and Controlling Risk. Economics, Volume 2, ed. Ronald W. Jones and Peter
New York: McGraw-Hill. B. Keenen. Amsterdam: North-Holland, 979–1040.
Grossman, Sanford J., and Joseph E. Stiglitz. 1980. “On Levich, Richard M., and Lee R. Thomas III. 1993.
the Impossibility of Informationally Efficient “The Significance of Technical Trading-Rule Profits
Markets.” American Economic Review, 70(3): in the Foreign Exchange Market: A Bootstrap
393–408. Approach.” Journal of International Money and
Group of Thirty. 1985. The Foreign Exchange Market Finance, 12(5): 451–74.
in the 1980s. New York: Group of Thrity. Lewellen, Jonathan, Stefan Nagel, and Jay Shanken.
Hellwig, Martin F. 1982. “Rational Expectations 2006. “A Skeptical Appraisal of Asset-Pricing Tests.”
Equilibrium with Conditioning on Past Prices: A NBER Working Papers, no. 12360.
Mean–Variance Example.” Journal of Economic Lewis, Karen K. 1999. “Trying to Explain Home Bias in
Theory, 26(2): 279–312. Equities and Consumption.” Journal of Economic
Henderson, Callum. 2002. Currency Strategy: The Literature, 37(2): 571–608.
Practitioner’s Guide to Currency Investing, Hedging Lillo, Fabrizio, and J. Doyne Farmer. 2004. “The Long
and Forecasting. Hoboken, N.J.: John Wiley and Memory of the Efficient Market.” Studies in
Sons. Nonlinear Dynamics and Econometrics, 8(3): 1–33.
Hommes, Cars. 2005. “Heterogeneous Agent Models Lo, Andrew W., Harry Mamaysky, and Jiang Wang.
in Economics and Finance.” Tinbergen Institute 2000. “Foundations of Technical Analysis:
Discussion Paper, no. 056. Computational Algorithms, Statistical Inference, and
Hsu, Po-Hsuan, and Chung-Ming Kuan. 2005. Empirical Implementation.” Journal of Finance,
“Reexamining the Profitability of Technical Analysis 55(4): 1705–65.
with Data Snooping Checks.” Journal of Financial Logue, Dennis E., and Richard James Sweeney. 1977.
Econometrics, 3(4): 606–28. “‘White-Noise’ in Imperfect Markets: The Case of
Ito, Takatoshi. 1990. “Foreign Exchange Rate the Franc/Dollar Exchange Rate.” Journal of
Expectations: Micro Survey Data.” American Finance, 32(3): 761–68.
Economic Review, 80(3): 434–49. Logue, Dennis E., Richard James Sweeney, and
Ito, Takatoshi, Richard K. Lyons, and Michael T. Thomas W. Willett. 1978. “Speculative Behaviour of
Melvin. 1998. “Is There Private Information in the Foreign Exchange Rates during the Current Float.”
Fx Market? The Tokyo Experiment.” Journal of Journal of Business Research, 6(2): 159–74.
Finance, 53(3): 1111–30. Lothian, James R., and Mark P. Taylor. 1996. “Real
Jegadeesh, Narasimhan, and Sheridan Titman. 2001. Exchange Rate Behavior: The Recent Float from the
“Profitability of Momentum Strategies: An Perspective of the Past Two Centuries.” Journal of
Evaluation of Alternative Explanations.” Journal of Political Economy, 104(3): 488–509.
Finance, 56(2): 699–720. Lucke, Bernd. 2003. “Are Technical Trading Rules
Kavajecz, Kenneth A., and Elizabeth R. Odders-White. Profitable? Evidence for Head-and-Shoulder Rules.”
2004. “Technical Analysis and Liquidity Provision.” Applied Economics, 35(1): 33–40.
Review of Financial Studies, 17(4): 1043–71. Lui, Yu-Hon, and David Mole. 1998. “The Use of
Kho, Bong-Chan. 1996. “Time-Varying Risk Premia, Fundamental and Technical Analyses by Foreign
Volatility, and Technical Trading Rule Profits: Exchange Dealers: Honk Kong Evidence.” Journal
Evidence from Foreign Currency Futures Markets.” of International Money and Finance, 17(3): 535–45.
Journal of Financial Economics, 41(2): 249–90. Lux, Thomas. 1998. “The Socio-Economic Dynamics
Kilian, Lutz, and Mark P. Taylor. 2003. “Why Is It So of Speculative Markets: Interacting Agents, Chaos,
and the Fat Tails of Return Distributions.” Journal of Hypothesis: Evidence from the Foreign Exchange
Economic Behavior and Organization, 33(2): Market.” Journal of Financial and Quantative
143–65. Analysis.
Lyons, Richard K. 2001. The Microstructure Approach Oberlechner, Thomas. 2001. “Importance of Technical
to Exchange Rates. Cambridge and London: MIT and Fundamental Analysis in the European Foreign
Press. Exchange Market.” International Journal of Finance
Malkiel, Burton G. 1996. A Random Walk Down Wall and Economics, 6(1): 81–93.
Street: Including a Life-Cycle Guide to Personal Oberlechner, Thomas. 2004. The Psychology of the
Investing. Sixth edition. New York and London: Foreign Exchange Market. Hoboken, N.J.: John
Norton. Wiley and Sons.
Mark, Nelson C. 1995. “Exchange Rates and Oberlechner, Thomas, and Carol L. Osler. 2006.
Fundamentals: Evidence on Long-Horizon “Overconfidence in Currency Markets.” Unpublished.
Predictability.” American Economic Review, 85(1): Okunev, John, and Derek White. 2003. “Do
201–18. Momentum-Based Strategies Still Work in Foreign
Marsh, Ian W. 2000. “High-Frequency Markov Currency Markets?” Journal of Financial and
Switching Models in the Foreign Exchange Market.” Quantitative Analysis, 38(2): 425–47.
Journal of Forecasting, 19(2): 123–34. Olson, Dennis. 2004. “Have Trading Rule Profits in the
Melvin, Michael T., Michael J. Sager, and Mark P. Currency Markets Declined over Time?” Journal of
Taylor. 2006. “Bank of England Interest Rate Banking and Finance, 28(1): 85–105.
Announcements and the Foreign Exchange Market.” Osler, Carol L. 2000. “Support for Resistance:
CEPR Discussion Paper. Technical Analysis and Intraday Exchange Rates.”
Menkhoff, Lukas. 1997. “Examining the Use of Federal Reserve Bank of New York Economic Policy
Technical Currency Analysis.” International Journal Review, 6(2): 53–68.
of Finance and Economics, 2(4): 307–18. Osler, Carol L. 2003. “Currency Orders and Exchange
Menkhoff, Lukas. 1998. “The Noise Trading Rate Dynamics: An Explanation for the Predictive
Approach—Questionnaire Evidence from Foreign Success of Technical Analysis.” Journal of Finance,
Exchange.” Journal of International Money and 58(5): 1791–1819.
Finance, 17(3): 547–64. Osler, Carol L. 2005. “Stop-Loss Orders and Price
Menkhoff, Lukas, and Manfred Schlumberger. 1995. Cascades in Currency Markets.” Journal of
“Persistent Profitability of Technical Analysis on International Money and Finance, 24(2): 219–41.
Foreign Exchange Markets?” Banca Nazionale del Park, Cheol-Ho, and Scott H. Irwin. 2005. “A Reality
Lavoro Quarterly Review, 48(193): 189–216. Check on Technical Trading Rule Profits in US
Neely, Christopher J. 1997. “Technical Analysis in the Futures Markets.” Unpublished.
Foreign Exchange Market: A Layman’s Guide.” Park, Cheol-Ho, and Scott H. Irwin. 2007. “What Do We
Federal Reserve Bank of St Louis Review, 79(5): Know about the Profitability of Technical Analysis?”
23–38. Journal of Economic Surveys, 21(4): 786–826.
Neely, Christopher J. 1998. “Technical Analysis and the Peiers, Bettina. 1997. “Informed Traders, Intervention,
Profitability of U.S. Foreign Exchange and Price Leadership: A Deeper View of the
Intervention.” Federal Reserve Bank of St Louis Microstructure of the Foreign Exchange Market.”
Review, 80(4): 3–17. Journal of Finance, 52(4): 1589–1614.
Neely, Christopher J. 2002. “The Temporal Pattern of Pilbeam, Keith. 1995. “The Profitability of Trading in
Trading Rule Returns and Exchange Rate the Foreign Exchange Market: Chartists, Funda-
Intervention: Intervention Does Not Generate mentalists, and Simpletons.” Oxford Economic
Technical Trading Profits.” Journal of International Papers, 47(3): 437–52.
Economics, 58(1): 211–32. Poole, William. 1967a. “Speculative Prices as Random
Neely, Christopher J., and Paul A. Weller. 1999. Walks: An Analysis of Ten Time Series of Flexible
“Technical Trading Rules in the European Monetary Exchange Rates.” Southern Economic Journal, 33(3):
System.” Journal of International Money and 468–78.
Finance, 18(3): 429–58. Poole, William. 1967b. “The Stability of the Canadian
Neely, Christopher J., and Paul A. Weller. 2001. Flexible Exchange Rate, 1950–62.” Canadian
“Technical Analysis and Central Bank Intervention.” Journal of Economics and Political Science, 33(2):
Journal of International Money and Finance, 20(7): 205–17.
949–70. Qi, Min, and Yangru Wu. Forthcoming. “Technical
Neely, Christopher J., and Paul A. Weller. 2003. Trading-Rule Profitability, Data Snooping, and
“Intraday Technical Trading in the Foreign Reality Check: Evidence from the Foreign Exchange
Exchange Market.” Journal of International Money Market.” Journal of Money, Credit, and Banking.
and Finance, 22(2): 223–37. Reitz, Stefan, and Mark P. Taylor. Forthcoming. “The
Neely, Christopher J., Paul A. Weller, and Robert Coordination Channel of Foreign Exchange
Dittmar. 1997. “Is Technical Analysis in the Foreign Intervention: A Nonlinear Microstructural Analysis.”
Exchange Market Profitable? A Genetic European Economic Review.
Programming Approach.” Journal of Financial and Saacke, Peter. 2002. “Technical Analysis and the
Quantitative Analysis, 32(4): 405–26. Effectiveness of Central Bank Intervention.” Journal
Neely, Christopher J., Paul A. Weller, and Joshua M. of International Money and Finance, 21(4): 459–79.
Ulrich. Forthcoming. “The Adaptive Markets Sager, Michael J., and Mark P. Taylor. 2004. “The
Impact of European Central Bank Governing Exchange Committee 2005 Annual Report. New
Council Announcements on the Foreign Exchange York: Federal Reserve Bank of New York.
Market: A Microstructural Analysis.” Journal of Surajaras, Patchara, and Richard James Sweeney. 1992.
International Money and Finance, 23(7–8): 1043–51. Profit-Making Speculation in Foreign Exchange
Sager, Michael J., and Mark P. Taylor. 2006. “Under the Markets. Boulder: Westview Press.
Microscope: The Structure of the Foreign Exchange Sweeney, Richard James, 1986. “Beating the Foreign
Market.” International Journal of Finance and Exchange Market.” Journal of Finance, 41(1):
Economics, 11(1): 81–95. 163–82.
Sapp, Stephen G. 2002. “Price Leadership in the Spot Sweeney, Richard James, 1997. “Do Central Banks
Foreign Exchange Market.” Journal of Financial and Lose on Foreign-Exchange Intervention? A Review
Quantitative Analysis, 37(3): 425–48. Article.” Journal of Banking and Finance, 21(11–12):
Sapp, Stephen G. 2004. “Are All Central Bank 1667–84.
Interventions Created Equal? An Empirical Sylla, Richard. 1992. “Chartist Analysis.” In The New
Investigation.” Journal of Banking and Finance, Palgrave Dictionary of Money and Finance, ed.
28(3): 443–74. Peter Newman, Murray Milgate, and John Eatwell.
Sarno, Lucio, and Mark P. Taylor. 2001a. “Official London: Macmillan, 342–44.
Intervention in the Foreign Exchange Market: Is It Szakmary, Andrew C., and Ike Mathur. 1997. “Central
Effective and, If So, How Does It Work?” Journal of Bank Intervention and Trading Rule Profits in
Economic Literature, 39(3): 839–68. Foreign Exchange Markets.” Journal of
Sarno, Lucio, and Mark P. Taylor. 2001b. “The International Money and Finance, 16(4): 513–35.
Microstructure of the Foreign Exchange Market: A Taylor, Alan M., and Mark P. Taylor. 2004. “The
Selective Survey of the Literature.” Princeton Studies Purchasing Power Parity Debate.” Journal of
in International Economics, no. 89, International Economic Perspectives, 18(4): 135–58.
Economics Section. Princeton: Princeton University Taylor, Mark P. 1995a. “Exchange-Rate Behavior under
Press. Alternative Exchange-Rate Arrangements.” In
Sarno, Lucio, and Mark P. Taylor. 2002. The Economics Understanding Interdependence: The Macroeconomics
of Exchange Rates. Cambridge and New York: of the Open Economy, ed. P. B. Kenen. Princeton:
Cambridge University Press. Princeton University Press, 34–83.
Schulmeister, Stephan. 1987. “An Essay on Exchange Taylor, Mark P. 1995b. “The Economics of Exchange
Rate Dynamics.” WZB Discussion Paper, no. Rates.” Journal of Economic Literature, 33(1):
IIM/LMP 87-8. 13–47.
Schulmeister, Stephan. 2005. “Components of the Taylor, Mark P. 2004. “Is Official Exchange Rate
Profitability of Technical Currency Trading.” Austrian Intervention Effective?” Economica, 71(281): 1–11.
Institute of Economic Research Working Paper, no. Taylor, Mark P., and Helen L. Allen. 1992. “The Use of
263. Technical Analysis in the Foreign Exchange
Schulmeister, Stephan. 2006. “The Interaction Market.” Journal of International Money and
between Technical Currency Trading and Exchange Finance, 11(3): 304–14.
Rate Fluctuations.” Finance Research Letters, 3(3): Taylor, Stephen J. 1992. “Rewards Available to
212–33. Currency Futures Speculators: Compensation for
Schwert, G. William. 2002. “Anomalies and Market Risk or Evidence of Inefficient Pricing?” Economic
Efficiency.” NBER Working Papers, no. 9277. Record, 68 (Supplement): 105–16.
Securities and Exchange Commission. 2003. Treynor, Jack L., and Robert Ferguson. 1985. “In
“Implications of the Growth of Hedge Funds.” Defense of Technical Analysis.” Journal of Finance,
Washington, D.C.: Securities and Exchange 40(3): 757–73.
Commission. http://www.sec.gov/news/studies/ Vigfusson, Robert. 1997. “Switching between Chartists
hedgefunds0903.pdf. and Fundamentalists: A Markov Regime-Switching
Sharpe, William F. 1966. “Mutual Fund Performance.” Approach.” International Journal of Finance and
Journal of Business, 39(1): 119–38. Economics, 2(4): 291–305.
Shiller, Robert J. 1984. “Stock Prices and Social Westerhoff, Frank H. 2003. “Expectations Driven
Dynamics.” Brookings Papers on Economic Activity, Distortions in the Foreign Exchange Market.”
2: 457–98. Journal of Economic Behavior and Organization,
Shleifer, Andrei, and Robert W. Vishny. 1997. “The 51(3): 389–412.
Limits of Arbitrage.” Journal of Finance, 52(1): 35–55. White, Halbert. 2000. “A Reality Check for Data
Silber, William L. 1994. “Technical Trading: When It Snooping.” Econometrica, 68(5): 1097–1126.
Works and When It Doesn’t.” Journal of Derivatives, Wilder, J. W., Jr. 1978. New Concepts in Technical
1(3): 39–44. Trading Systems. Greensboro, N.C.: Hunter
Snyder, Mark. 2005. “Chair’s Letter.” In Foreign Publishing Company.

Taylor Menkhoff Technical Analysis

Încărcat de

Informații document

Descriere originală:

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Taylor Menkhoff Technical Analysis

Încărcat de

Drepturi de autor:

Formate disponibile

dec07_Article2 11/29/07 10:20 AM Page 936

Journal of Economic Literature

The Obstinate Passion of

1. Introduction professional economists, the widespread,

938 Journal of Economic Literature, Vol. XLV (December 2007)

Obviously, the variety of both qualitative

940 Journal of Economic Literature, Vol. XLV (December 2007)

Study time financial target group number of response

942 Journal of Economic Literature, Vol. XLV (December 2007)

Study form of some use share of share of the relation

or as advisers to traders (Goodhart 1988; (1992). In particular, the weight given to

Menkhoff (1997, 1998) Gehrig and Menkhoff (2006)

944 Journal of Economic Literature, Vol. XLV (December 2007)

Taylor and Allen (1992; Table 4)

A. Studies considering technical and fundamental analysis only

B. Studies considering also further influences Cheung and Chinn

946 Journal of Economic Literature, Vol. XLV (December 2007)

13 It may also allow for tax payments, where tax treat-

Study period number form of consideration of excess

948 Journal of Economic Literature, Vol. XLV (December 2007)

19 An anonymous referee has pointed out a number of

950 Journal of Economic Literature, Vol. XLV (December 2007)

Study period number standard of risk adjustment risk-adjusted

risk premium for investing in one currency be a reasonable constellation of parameters

952 Journal of Economic Literature, Vol. XLV (December 2007)

954 Journal of Economic Literature, Vol. XLV (December 2007)

as an indication of not- users underestimate

956 Journal of Economic Literature, Vol. XLV (December 2007)

Hypothesis being tested aggregated figures Pearson χ2 probability

958 Journal of Economic Literature, Vol. XLV (December 2007)

cautiously questioned by Neely (1998). He exchange intervention data from the

960 Journal of Economic Literature, Vol. XLV (December 2007)

962 Journal of Economic Literature, Vol. XLV (December 2007)

964 Journal of Economic Literature, Vol. XLV (December 2007)

966 Journal of Economic Literature, Vol. XLV (December 2007)

968 Journal of Economic Literature, Vol. XLV (December 2007)

970 Journal of Economic Literature, Vol. XLV (December 2007)

972 Journal of Economic Literature, Vol. XLV (December 2007)

S-ar putea să vă placă și