forex laymen guide

8/14/2019 Forex Laymen Guide

1/16

F ED ERA L R E S E R V E B A N K O F S T. L O U I S

23

S EP TEM B ER/ O CTO B ER 1 9 9 7

Te ch n i ca lAn a l ysi s i nthe Fo r e i g nEx ch a n g eM a r k et : ALay m an s Gu id e

Chr i stop her J. Nee ly

Technical anal ysis suggests that a long-term ral ly

frequently is interrupted by a short-lived decline.

Such a dip, according to this view, reinforces the

original uptrend. Should the dollar fall below

1.5750 marks, dealers said, t echnical signals would

point to a correction that could pull the dollar back

as far as 1.55 marks before it rebounded.

Gregory L. White

Wall Street Journal

November 12, 1992

Technical analysis, which dates back acentury to the writings ofWall StreetJournaleditor Charles Dow, is the use of

past price behavior to guide trading deci-sions in asset markets. For example, atrading rule might suggest buying a currency ifits price has risen more than 1 percent fromits value five days earlier. Such rules arewidely used in stock, commodity, and (sincethe early 1970s) foreign exchange markets.More than 90 percent of surveyed foreign

exchange dealers in London report usingsome form of technical analysis to informtheir trading decisions (Taylor and Allen,1992). In fact, at short horizonsless thana weektechnical analysis predominatesover fundamental analysis, the use of othereconomic variables like interest rates, andprices in influencing trading decisions.

Investors and economists are interestedin technical analysis for different reasons.

Investors are concerned with beating themarket, earning the best return on theirmoney. Economists study technicalanalysis in foreign exchange marketsbecause its success casts doubt on the effi-cient markets hypothesis, which holds thatpublicly available information, like pastprices, should not help traders earnunusually high returns. Instead, thesuccess of technical analysis suggests thatexchange rates are not always determinedby economic fundamentals like prices andinterest rates, but rather are driven away

from their fundamental values for longperiods by traders irrational expectationsof future exchange rate changes. Theseswings away from fundamental values maydiscourage international trade and invest-ment by making the relative price of U.S.and foreign goods and investments veryvolatile. For example, when BMW decideswhere to build an automobile factory, itmay choose poorly if fluctuating exchangerates make it difficult or impossible to pre-dict costs of production in the United

States relative to those in Germany.Despite the widespread use oftechnical analysis in foreign exchange(and other) markets, economists have tra-ditionally been very skeptical of its value.Technical analysis has been dismissed bysome as astrology. In turn, technicaltraders have frequently misunderstoodwhat economists have to say about assetprice behavior. What can the two learnfrom each other? This article providesan accessible treatment of recent researchon technical analysis in the foreign

exchange market.

A PRIM ER ON TECHNICALANALYSIS IN FOREIGNEX CHA N GE M A RKETS

Technical analysis is a short-horizontrading method; positions last a few hoursor days. Technical traders will not hold

Christopher J. Neely i s an economist at the Federal Reserve Bank of St. Louis. Kent A. Koch provided research assistance.


2/16

positions for months or years, waitingfor exchange rates to return to where fun-damentals are pushing them. In contrast,fundamental investors study the economicdeterminants of exchange rates as a basisfor positions that typically last muchlonger, for months or years. Some traders,however, use technical analysis inconjunction with fundamental analysis,doubling their positions when technicaland fundamental indicators agree on thedirection of exchange rate movements.

Three principles guide the behaviorof technical analysts.1 The first is thatmarket action (prices and transactionsvolume) discounts everything. In otherwords, all relevant information about anasset is incorporated into its price history,so there is no need to forecast thefundamental determinants of an assetsvalue. In fact, Murphy (1986) claims that

asset price changes often precede observedchanges in fundamentals. The secondprinciple is that asset prices move intrends. Predictable trends are essential tothe success of technical analysis becausethey enable traders to profit by buying(selling) assets when the price is rising(falling), or as techn icians counsel, thetrend is your friend. Practitioners appealto Newtons law of motion to explain the

existence of trends: Trends in motion tendto remain in motion unless acted upon byanother force. The third principle of tech-nical analysis is that history repeats itself.

Asset traders will tend to react the sameway when confronted by the sameconditions. Technical analysts do notclaim their methods are magical; rather,they take advantage of market psychology.

Following from these principles, themethods of technical analysis attempt toidentify trends and reversals of trends.These methods are explicitly extrapolative;that is, they infer future price changesfrom those of the recent past. Formalmethods of detecting trends are necessarybecause prices move up and down around

the primary (or longer-run) trend. Anexample of this movement is shown inFigure 1, where the dollar/deutsche mark($/DM) exchange rate fluctuates around anapparent uptrend.2

To distinguish trends from shorter-runfluctuations, technicians employ two typesof analysis: chartingand mechanical rules.Charting, the older of the two methods,involves graphing the history of pricesover some perioddetermined by thepractitionerto predict future patternsin the data from the existence of pastpatterns. Its advocates admit that thissubjective system requires the analyst touse judgement and skill in finding andinterpreting patterns. The second type ofmethod, mechanical rules, imposes consis-tency and discipline on the technician byrequiring him to use rules based on mathe-matical functions of present and pastexchange rates.

Charting

To identify trends through the use ofcharts, practitioners must first find peaksand troughsin the price series. A peak isthe highest value of the exchange ratewithin a specified period of time (a localmaximum), while a trough is the lowestvalue the price has taken on within thesame period (a local minimum) . A seriesof peaks and troughs establishes downtrendsand uptrends, respectively. For example, as

1 These principles and a muchmore comprehensive t reatmentof technical analysis are provid-ed by Murphy (1986) and

Pring (1991 ) . Rosenberg andShatz (1995) advocate theuse of technical analysis withmore economic explanation.

2 Figure 1 shows only closingprices. In this, it differs frommost charts employed by tech-nical traders, which might showthe opening, closing, and dailytrading range.



24

Figure 1

Peak s, Tro ug hs, Tre nd s, Resis tanceand Suppo r t Leve l s I l l u st r a t ed f o r

t he $ / DM

Sell signal from a0.5% filter rule

Local t roughs

Resistance level

Trendline

Buy signal from a 0.5% filt er rule

Support level

Local peak

NOTES: Not all buy and sell signals from the filter rule are identified.

0.72

0.62

SeptAugJulyJuneMay0.60

0.66

0.70

0.64

0.68

1992

$ per DM


3/16

shown in Figure 1, an analyst may establishan uptrend visually by connecting twolocal troughs in the data. A trendline isdrawn below an apparent up trend or

above an apparent downtrend. As moretroughs touch the trendline withoutviolating it, the technician may place moreconfidence in the validity of the trendline.The angle of the trendline indicates thespeed of the trend, with steeper lines indi-cating faster appreciation (or depreciation)of the foreign currency.

After a trendline has been established,the technician trades with the trend,buying the foreign currency if an uptrend issignaled and selling the foreign currency ifa downtrend seems likely. When a market

participant buys a foreign currency in thehope that it will go up in price, that partici-pant is said to be longin the currency. Theopposite strategy, called shortingor sellingshort, enables the participant to makemoney if the foreign currency falls in price.A short seller borrows foreign currencytoday and sells it, hoping the price will fallso that it can be bought back more cheaplyin the future.

Spotting the reversal of a trend is justas important as detecting trends. Peaksand troughs are important in identifyingreversals too. Local peaks are called resis-tance levels, and local troughs are calledsupport levels(see Figure 1). If the pricefails to break a resistance level (a localpeak) during an uptrend, that may be anearly indication that the trend may soonreverse. If the exchange rate significantlypenetrates the trendline, that is considereda more serious signal of a possible reversal.

Technicians identify several patternsthat are said to foretell a shift from a trendin one direction to a trend in the opposite

direction. An example of the best-knowntype of reversal formation, called headand shoulders, is shown in Figure 2. Thehead and shoulders reversal following anuptrend is characterized by three localpeaks with the middle peak being thelargest of the three. The line between thetroughs of the shoulders is known as theneckline. When the exchange ratepenetrates the neckline of a head and

shoulders, the technician confirms areversal of the previous uptrend andbegins to sell the foreign currency. Thereare several other similar reversal patterns,including the V (single peak), the doubletop (two similar peaks) and the tripletop (three similar peaks). The reversalpatterns of a downtrend are essentiallythe mirrors of the reversal patterns forthe uptrend.

Mechanical Rules

Charting is very dependent on theinterpretation of the technician who isdrawing the charts and interpreting thepatterns. Subjectivity can permit emotionslike fear or greed to affect the tradingstrategy. The class of mechanical tradingrules avoids this subjectivity and so ismore consistent and disciplined, but,according to some technicians, it sacrificessome information that a skilled chartist

might discern from the data. Mechanicaltrading rules are even more explicitlyextrapolative than charting; they look fortrends and follow those trends. A well-known type of mechanical trading rule isthe filter rule, or trading range breakrule which counsels buying (selling) a cur-rency when it rises (falls) xpercent above(below) its previous local minimum (max-imum). The size of the filter, x, which is



25

Figu re 2

The Head and Shou lders Reversa l Pa t te rnI l l ust r a t e d f or t h e $ / D M

Right shoulder

Neckline

Left shoulder

Head0.66

0.56

MayDec AprFeb MarJanNovOctSept

0.60

0.64

0.58

0.62

1991-92

$ per DM

Exchange ratepenetrates t heneckline sellsignal


4/16

chosen by the technician from past experi-ence, is generally between 0.5 percent and

3 percent. Figure 1 illustrates some of thebuy and sell signals generated by a filterrule with filter size of 0.5 percent.

A second variety of mechanical tradingrule is the moving average class. Liketrendlines and filter rules, moving averagesbypass the short-run zigs and zags of theexchange rate to permit the technician toexamine trends in the series. A movingaverage is the average closing price of the

exchange rate over a given number ofprevious trading days. The length of themoving average windowthe number ofdays in the moving averagegoverns

whether the moving average reflects long- orshort-run trends.3 Any moving average willbe smoother than the original exchange-rate series, and long moving averages willbe smoother than short moving averages.Figure 3 illustrates the behavior of a 5-dayand a 20-day moving average of the exchangerate in relation to the exchange rate itself.A typical moving average trading rule pre-scribes a buy (sell) signal when a shortmoving average crosses a longer movingaverage from below (above)that is,when the exchange rate is rising ( falling)

relatively fast. Of course, the lengths ofthe moving averages must be chosen bythe technician. The length of the shortmoving average rule is sometimes chosento equal one, the exchange rate itself.

A final type of mechanical trading ruleis the class of oscillators, which are saidto be useful in non-trendingmarkets, whenthe exchange rate is not trending up ordown strongly. A simple type of oscillatorindex, an example of which is shown inFigure 4, is given by the difference betweentwo moving averages: the 5-day movingaverage minus the 20-day moving average.Oscillator rules suggest buying (selling) theforeign currency when the oscillator indextakes an extremely low (high) value. Notethat the oscillator index, as a differencebetween moving averages, also generatesbuy/sell signals from a moving average rulewhen the index crosses zero. That is,when the short moving average becomeslarger than the long moving average, themoving average rule will generate a buysignal. By definition, this will happen

when the oscillator index goes from nega-tive to positive. Therefore, an oscillatorchart is also useful for generating movingaverage rule signals.

Other Kinds of Technical Analysis

Technical analysis is more complexand contains many more techniques thanthose described in this article. For

3 For example, the five-day mov-ing average of an exchangerate series is given by:

where St denotes the closingprice of the spot exchange rateat day t.

M(5 ) = St- it

1

5 i= 0

4



26

Figu re 3

5 - and 2 0 - Day M ov i ng Ave r a ges

Sell signal, movingaverage rule

5-Day moving average

Buy signal, moving average rule

20- Day moving average

Exchange rate

NOTES: These moving averages smooth the exchange rate and can be used togenerate buy and sell signals in the foreign exchange market.

0.65

0.60

JunMayAprMarFeb0.59

0.62

0.64

0.61

0.63

1992

$ per DM

Figure 4

The Osc i l la to r Index

Oscillator rule buy signal

Diff erence inmoving averages

Movingaveragerule buysignal

Movingaverage rulesell signal

Oscillator rule sell signals

NOTES: The 5-day moving aver age minus the 2 0-day moving aver age can also beused to generate buy and sell signals.

1.0

0.2

JunMayAprMarFeb

0.4

0.4

0.8

0

0.2

0.6

1992

Normalized difference in moving averages

0.8

1.0

0.6


5/16

example, many techn ical analysts assign aspecial role to round numbers in supportor resistance levels. When the exchangerate significantly crosses the level of 100

yen to the dollar, that is seen as an indica-tion that further movement in the samedirection is likely.4 Other prominent typesof technical analysis use exotic mathemat-ical concepts such as Elliot wave theoryand/or Fibonacci numbers.5 Finally,traders sometimes use technical analysis ofone markets price history to take positionsin another market, a practice called inter-market technical analysis.

EFFICIENT M ARKETS AND

TECHN ICA L A N A LYSISTechnical analysts believe that their

methods will permit them to beat themarket. Economists have traditionally beenskeptical of the value of technical analysis,affirming the theory of efficient marketsthat holds that no strategy should allowinvestors and traders to make unusualreturns except by taking excessive risk.6

Investing in the ForeignExchange Market

To understand the efficient marketshypothesis in the context of foreignexchange trading, consider the optionsopen to an American bank (or firm) thattemporarily has excess funds to be investedovernight. The bank could lend that moneyin the overnight bank money market,known as the federal funds market. Thesimple net return on each dollar investedthis way would be the overnight interestrate on dollar deposits. The bank has otherinvestment options, though. It could

instead convert its money to a foreign cur-rency (e.g., the deutsche mark), lend itsmoney in the overnight German moneymarket (at the German interest rate) andthen convert it back to dollars tomorrow.This return is the sum of the Germanovernight interest rate and the change inthe value of the DM. Which investmentshould the bank choose? If the bank werenot concerned about risk, it would choose

the investment with the higher expectedreturn. While the U.S. and Germaninterest rates are known, the bank mustbase its decision on its forecast of the rate

of appreciation of the DM. If market par-ticipants expect the return to investing inthe German money market to be higherthan that of investing in the U.S. moneymarket, they will all try to invest in theGerman market, and none will invest inthe U.S. money market. Such a situationwould tend to drive down the Germanreturn and raise the U.S. return until thetwo were equalized. The excess returnon aGerman investment over an investment inthe U.S. money market (RtDM), at date t,from the point of view of a U.S. investor is

defined as

(1) RtDM itDM+ St it$,

where itDM is the German overnight interest

rate, St is the percentage rate of apprecia-tion of the DM against the dollar overnight,and it$ is the U.S. overnight interest rate.7

If market participants cared only about theexpected return on their investments, andif their expectations about the change in theexchange rate were not systematically wrong,the expected excess return on foreignexchange should equal zero, every day.

The assumption that market partic-ipants care only about the expected returnis too strong, of course. Surely, participantsalso care about the riskof their investment.8

Risk can come from either the risk of defaulton the loan or the risk of sharp changes inthe exchange rate, or both. If investing inthe German market is significantly riskierthan investing in the U.S. market, investorsmust be compensated with a higher expectedreturn in the German market, or they will

not invest there. In that case, the expectedexcess return would be positive and equalto a risk premium. The expected risk-adjusted excess return would be equalto zero. That is,

(2) E[RtDM] RPt=0,

where E[*] is a function that takes theexpected value of the term inside the

4 The 100 yen level for the dol-lar is still a very big psychologi-cal barrier and it will take a fewtests before it breaks. But onceyou break 100 yen, its notgoing to remain there for long.Youll probably see it tradebetween 102 and 106 for awhile, said Jorge Rodriguez,director of North AmericanSales at Credit Suisse, asreported by Creswell ( 199 5) .

5 Murphy (1986) discusses Elliotwave theory, Fibonacci num-bers, and many other technicalconcepts.

6 Samuelson (1 965 ) did seminatheoretical work on the moderntheory of efficient markets.

7 The excess return may also beconsidered the return to some-

one borrowing in dollars andinvesting those dollars inGerman investments.

8 Market participants may beconcerned about the liquidityof their position as well asthe expected return and risk.Liquidity is the ease with whichassets can be convertedinto cash.



27


6/16

brackets [*] and RPt is the risk premiumassociated with the higher risk of lendingin the German market.

Efficient Markets

The idea that the expected risk-adjustedexcess return on foreign exchange is zeroimplies a sensible statement of the efficientmarkets hypothesis in the foreign exchangecontext: Exchange rates reflect informationto the point where the potential excess returnsdo not exceed the transactions costs of acting(trading) on that information.9 In otherwords, you cant profit in asset markets(like the foreign exchange market) bytrading on publicly available information.

This description of the efficient marketshypothesis appears to be a restatement ofthe first principle of technical analysis:Market action (price and transactionsvolume) discounts all information aboutthe assets value. There is, however, a subtlebut important distinction between the effi-cient markets hypothesis and technicalanalysis: The efficient markets hypothesisposits that the current exchange rate adjuststo all information to prevent traders fromreaping excess returns, while technicalanalysis holds that current and past pricemovements contain just the informationneeded to allow profitable trading.

What does this version of the efficientmarkets hypothesis imply for technicalanalysis? Under the efficient marketshypothesis, only current interest rates andrisk factors help predict exchange ratechanges, so past exchange rates are of nohelp in forecasting excess foreign exchangereturnsi.e., if the hypothesis holds, tech-nical analysis will not work. Malkielssummary of the attitude of many economists

toward technical analysis in the stockmarket is based on similar reasoning:

The past history of stock prices cannotbe used to predict the future in anymeaningful way. Technical strategiesare usually amusing, often comforting,but of no real value. (Malkiel, 1990,p. 154.)

How do prices move in the hypotheticalefficient market? In an efficient market,profit seekers trade in a way that causesprices to move instantly in response to new

information, because any information thatmakes an asset appear likely to becomemore valuable in the future causes animmediate price rise today. If prices domove instantly in response to all newinformation, past information, like prices,does not help anyone make money. Ifthere were a way to make money with littlerisk from past prices, speculators wouldemploy it until they bid away the money tobe made. For example, if the price of anasset rose 10 percent every Wednesday,speculators would buy strongly on Tuesday,

driving prices past the point where anyonewould think they could rise much further,and so a fall would be likely. This situationcould not lead to a predictable pattern ofrises on Tuesday, though, because specula-tors would buy on Monday. Any pattern inprices would be quickly bid away by marketparticipants seeking profits. Indeed, thereis considerable evidence that markets oftendowork this way. Moorthy (1995) findsthat foreign exchange rates react veryquickly and efficiently to news of changesin U.S. employment figures, for example.

Because the efficient markets hypothesisis frequently misinterpreted, it is importantto clarify what the idea does notmean. Itdoes not mean that asset prices are unrelatedto economic fundamentals.10 Asset pricesmay be based on fundamentals like the pur-chasing power of the U.S. dollar or Germanmark. Similarly, the hypothesis does notmean that an asset price fluctuates randomlyaround its intrinsic (fundamental) value.If this were the case, a trader could makemoney by buying the asset when the price

was relatively low and selling it when it wasrelatively high. Rather, efficient marketsmeans that at any point in time, asset pricesrepresent the markets best guess, based onall currently available information, as to thefundamental value of the asset. Future pricechanges, adjusted for risk, will be close tounpredictable.

But if any pattern in pr ices is quicklybid away, how does one explain the

9 There are a number of versionsof the efficient markets hypoth-esis. This version is close tothat put forward by Jensen(1978).

10

For an example of an incorrectinterpretation of the efficientmarkets hypothesis, seeMurphy (1986, p. 20-21) whooffers, The theory is based onthe efficient markets hypothesis,which holds that prices fluctuaterandomly about their intrinsicvalue. . . . its just unrealistic tobelieve that allprice movementis random.



28


7/16

apparent trends seen in charts of assetprices like those in Figure 1? Believers inefficient markets point out that completelyrandom price changeslike those generated

by flipping a coinwill produce priceseries that seem to have trends (Malkiel,1990, or Paulos, 1995). Under efficientmarkets, however, traders cannot exploitthose trends to make money, since thetrends occur by chance and are as likely toreverse as to continue at any point. (Forexample, some families havepurely bychancestrings of either boys or girls, yeta family that already has four girls and isexpecting a fifth child still has only a 50percent chance of having another girl.)

EVALUATING TECHNICALANALYSIS

The efficient markets hypothesisrequires that past prices cannot be usedto predict exchange rate changes. If thehypothesis is true, technical analysisshould not enable a trader to earn profitswithout accepting unusual risk. This sec-tion examines how two common types oftrading rules are formulated and how thereturns generated by these rules aremeasured. Problems inherent in testingthe rules, measuring risk, and drawingconclusions about the degree of marketefficiency are discussed.11

Finding a Trading Rule

A basic problem in evaluatingtechnical trading strategies is that rulesrequiring judgement and skill are impossibleto quantify and therefore unsuitable fortesting. A fair test requires fixed, objective,commonly used trading rules to evaluate.

An objective rule does not rely on indi-vidual skill or judgement to determine buyor sell decisions. The rule should be com-monly used to reduce the problem ofdrawing false conclusions from datamining a practice in which manydifferent rules are tested unt il, purely bychance, some are found to be profitableon the data set. Negative test results areignored, while positive results are

published and taken to indicate thattrading ru le strategies can yield profits.For example, there is a vast literature onpricing anomalies in the equity markets,

summarized by Ball (1995) and Fortune(1991), but Roll (1994) has found thatthese aberrations are difficult to exploit inpractice; he suggests that they may be par-tially the result of data mining.

Trading Rules

With these considerations, two kindsof trading rules have been commonlytested: fil ter rulesand moving average rules.As a preceding section of this articleexplained, filter rules give a buy signal

when the exchange rate rises xpercentover the previous recent minimum.The analyst must make two choices toconstruct a filter rule: First, how muchdoes the exchange rate have to rise, orwhat is the size of the filter? Second, howfar back should the rule go in finding arecent minimum? The filter rules studiedhere will use filters from 0.5 percent to 3percent and go back five business days tofind the extrema.12 A moving average rulegives a buy signal when a short movingaverage is greater than the long movingaverage; otherwise it gives a sell signal.This rule requires the researcher to choosethe lengths of the moving averages. Themoving average rules to be tested will useshort moving averages of 1 day and 5 daysand long moving averages of 10 days and50 days. Both the filter rules and themoving average rules are extrapolative, inthat they indicate that the trader shouldbuy when the exchange rate has beenrising and sell when it has been falling.

Profits

The trading ru les switch between longand shortpositions in the foreign currency.Recall that a long position is a purchaseof foreign currencya bet that it will goupwhile a short position is the reverse,selling borrowed foreign currency now inthe hope that its value will fall. Denotingthe percentage change in the exchange rate

11 A number of previous studieshave documented evidence ofprofitable technical trading rules

in the foreign exchange market:Sweeney (1 986 ); Levich andThomas (1993); Neely, Weller,and Dittmar (199 7).

12 As with most aspects of techni-cal analysis, the choice of filtersize and window lengths hasbeen determined by practition-ers through a process of trialand error.



29


8/16

($ per unit of foreign currency) from date tto t+1 by S

t, and the domestic (foreign)

overnight interest rate by it$(itDM), then theovernight return from a long position isapproximately given by Equation 1:

(1) RtDM itDM+St it$.

The return to a short position is the nega-tive of the return to a long position. Thereturn to a trading rule over a period oftime is approximately the sum of dailyreturns, minus transactions costs for eachtrade. Transactions costs are set at 5 basis

points (0.05 percent) for each round tripinthe currency. A round trip is a move froma long position to a short position andback or vice-versa.13

Evidence from Ten Simple TechnicalTrading Rules

Six filter rules and four moving averagerules were tested on data consisting of the

average of daily U.S. dollar bid and askquotes for the DM, yen, pound sterling, andSwiss franc.14 All exchange rate data beginon 3/1/74 and end on 4/10/97. These fourseries are called $/DM, $/, $/, and $/SF.Because the results for the four exchangerates were similar, full results from onlythe $/DM will be reported in the tables.

Table 1 shows the annualizedpercentage return, monthly standarddeviation (a measure of the volatility ofreturns), number of trades per year, andtwo measures of risk, the Sharpe rati oandthe CAPM beta, for each of the 10 trading

strategies for the $/DM. The Sharpe ratioand CAPM betas are discussed in somedetail in the shaded insert. The meanannual return to the 10 rules was 4.4 per-cent, and 38 of the 40 trading rules wereprofitable (had positive excess return) overthe whole sample. These results cast doubton the efficient markets hypothesis, whichholds that no trading strategy should beable to consistently earn positive excess

13 The estimate of transactionscosts used here is consistentwith recent figures. Levich andThomas (1993) consider around-trip cost of 0.05 percentrealistic, as do Osler and Chang(1995).

14 The exchange rate data were

obtained from DRI and werecollected at 4:00 p.m. localtime in London from NatwestMarkets and S&P Comstock.Daily overnight interest ratesare collected by BIS at 9:00a.m. London time. Interestrates for Japan were unavail-able before 3/ 1/ 82, so theinterest rates before this datewere set to 0 for the $/ case.



30

Techni ca l Tra di ng Rule Resul ts fo r the $ / DMMoving Average Rule Results

Monthly Est imated StandardAnnual Standard Number of Sharpe CAPM Error ofShort MA Long MA Return Deviat ion Trades Rat io Beta Est . Beta

1 10 6.016 2.979 928 0.583 0.022 0.091

1 50 7.546 3.155 268 0.690 0.135 0.085

5 10 6.718 3.064 576 0.633 0.144 0.084

5 50 6.671 3.236 146 0.595 0.134 0.080

Filter Rule Results

Monthly Est imated StandardAnnual Standard Number of Sharpe CAPM Error of

Filt er Return Deviat ion Trades Rat io Beta Est . Beta

0.005 5.739 3.057 1070 0.542 0.071 0.089

0.010 6.438 2.951 584 0.630 0.092 0.093

0.015 3.323 3.255 382 0.295 0.037 0.085

0.020 1.934 3.348 234 0.167 0.128 0.087

0.025 0.839 3.236 142 0.075 0.118 0.082

0.030 1.541 3.578 92 0.124 0.086 0.077

NOTES: The first two columns of the top panel characterize the length of the short and long moving averages used in the moving-averagetrading rule. The third column is the annualized asset return to the rule, while the fourth column is the monthly standard deviation ofthe return. The fifth column is the number of trades over the 23-year sample. The sixth column is the Sharpe ratio, and the last twocolumns provide the CAPM beta with the S&P 500 and the standard er ror of that estimate. The lower panel has a similar structure,except that the first column characterizes the size of the fil ter used in the rule. All extrema for fi lter rules were measured over theprevious five business days.

Tab le 1


9/16

returns. The number of trades over the23-year sample varied substantially over the10 rules, ranging from 4 trades per yearto almost 50 trades per year. The moving

average rules were somewhat more profitablethan the filter rules.

There is little evidence that theseexcess returns are compensation forbearing excessive risk. The first measureof risk, the Sharpe ratio, is the meanannual return divided by the mean annualstandard deviation. The moving averagerules had higher Sharpe ratios (0.6 vs.0.25) than the filter rules. Six of the 10Sharpe ratios are better than the 0.3obtained by a buy-and-hold strategy inthe S&P 500 over approximately the same

period. This result indicates that theaverage return to the rules is very goodcompared to the risk involved infollowing the rules.

The second measure of risk, the CAPMbetas, reflects the correlation between themonthly trading rule returns and themonthly returns to a broad portfolio ofrisky assets (the S&P 500) . Significantlypositive betas indicate that the ru le isbearing undiversifiable risk. These CAPMbetas estimated from the 10 rules generallyindicate negative correlation with the S&P500 monthly returns. None of them issignificantly positive, statistically or econom-ically. In other words, there is no systematicrisk in these rules that could explain thepositive excess returns.

For Whom is Technical TradingAppropriate?

The discussion of risk and returns sug-gests that technical analysis may be veryuseful for banks and large financial firms

that can borrow and lend freely at theovernight interbank interest rate and buyand sell in the wholesale market for foreignexchange, where transactions sizes are inthe millions of dollars. Technical trading ismuch less useful for individuals, whowould face much higher transactions costsand must consider the opportunity cost ofthe t ime necessary to become an expert onforeign exchange speculating and to keep

up with the market on a daily basis. Howlarge would transactions costs have to be toeliminate the excess return to the technicalrules? If we assume a 6 percent annualexcess return to the rule and 230 trades (10trades a year), round-trip transactions costswould have to be greater than 0.6 percentto produce zero excess returns.

In addition to higher transactionscosts, individual investors following tech-nical rules also must accept the risk thatsuch a strategy entails. Figure 5 illustratesthe risk by depicting, at monthly intervals,the one-year-ahead excess return from1974 through 1996 for the (1,10) movingaverage rule on the $/DM and, for compar-ison, the total excess return on buying andholding the S&P 500 index, a popularmeasure of returns to a stock portfolio.The figure shows that the excess returns toboth portfolios vary considerably at the

annual horizon, often turning negative.While the technical trading rule excessreturn is less variable than the S&P excessreturn, it can still lead to significant lossesfor some subperiods. Two ways tomeasure losses over subperiods are themaximal single-period loss (maximumdrawdown) and maximum loss in acalendar year. Over the period from March1974 through March 1997, the maximum



31

Figu re 5

O n e - Ye a r M o v i n g A v e r a g e , Fo r w a r d -Look ing Ex cess Re t u r ns t o t he ( 1 , 1 0 )

M o v i n g A v e r a g e Tr a d i n g Ru l e

Excess return to $/ DM(1,10 ) moving average rule

Excess return t oS&P 500 index

50

30

10

20

9690 92 9480 8884 868278761974

20

10

40

0

30

Moving Average of Annual Return

March 1974-March 1996


10/16

that an investor could have lost by using themoving average trading rule was 28.2 per-cent; this loss, which would have occurredbetween March 7, 1995, and August 2, 1995

(a period of 149 days), translates into anannual rate of 69.2 percent. In otherwords, an investor using this rule wouldhave lost almost 30 percent of his capitalover this five-month period. Similarly, themaximum loss for this technical tradingrule in a complete calendar year was 9.8percent in 1995, but 17.8 percent for theS&P 500 in 1981.15

Perhaps the biggest obstacle toexploiting technical rules is that while thereturns to stocks depend ultimately on theprofitability of the firms in which the stock

is held, the sourceof returns to technicalanalysis is not well understood; therefore,the investor does not know if the returnswill persist into the future or even if theycontinue to exist at the present. Indeed,Figure 5 shows that the post-1992 returnto the (1,10) moving average rule for the$/DM has been negative.

Do These Results Measure theDegree of Market Efficiency?

There are a number of problems asso-ciated with inferring the degree of marketefficiency from the apparent profitability ofthese trading rules. The first problem isthe data. To test the profitability of atrading rule, the researcher needs actualprices and interest rates from a series ofsimultaneous market transactions. Unfor-tunately, simultaneous quotes for dailyexchange rates and interest rates are notgenerally available for a long time span.For example, these exchange-rate datawere collected late in the afternoon , while

the interest rates were collected in themorning. Although most economists

judge this problem to be very minor, someargue that the trading rule decisions couldnot have been executed at the exchangerates and interest rates used.

The second problem is that without agood model of how to pr ice risk, positiveexcess returns resulting from the use oftrading rules cannot be used to measure

the degree of inefficiency. Risk is notoriouslydifficult to measure. In fact, a major areaof study for macro and financial economistsfor the last 10 years has been to explain

why the return on stocks is so much higherthan that on bonds, a phenomenon calledthe equity premium puzzle. Of course, atleast part of the answer is that stocks aremuch riskier than bonds, but there is nogenerally accepted model of risk that willexplain the sizeof the return difference.16

Defenders of the efficient markets hypoth-esis maintain that the discovery of anapparently successful t rading strategy maynot indicate market inefficiency but, rather,that r isk is not measured properly.

Another problem is that of data

mining: If enough rules are tested, somepurely by chancewill produce excessreturns on the data. These rules may nothave been obvious to traders at the begin-ning of the sample. In fact, the rules testedhere are certainly subject to a data-miningbias, since many of them had been shownto be profitable on these exchange ratesover at least some of the subsample. Closelyrelated to the data-mining problem is thetendency to publish research that overturnsthe conventional wisdom on efficientmarkets, rather than research that showstechnical analysis to be ineffective. Onesolution to the data-mining problem issuggested by Neely, Weller, and Dittmar(1997), who apply genetic programmingtechniques to the foreign-exchange market.Genetic programming is a method by whicha computer searches through the space ofpossible technical trading rules to find agroup of good rules (i.e., rules that generatepositive excess return) . These good rulesare then tested on out-of-sample data tosee if they continue to generate positive

excess returns.

RETHINKING THE EFFICIENTM A RKETS HYPOTHESIS

Early research in finance on theefficient markets hypothesis was verysupportive; little evidence was found ofprofitable trading rules after transactionscosts were accounted for (Fama, 1970).

15 The returns for complete calen-dar years were available from1975 through 1995.

16 Kocherlakota (1996) andSiegel and Thaler (1997) dis-cuss the equity premium puzzleextensively.



32


11/16

The success of technical trading rulesshown in the previous section is typicalof a number of later studies showing thatthe simple efficient markets hypothesis

fails in important ways to describe howthe foreign exchange market actually func-tions. While these results did not surpr isemarket practitioners, they have helpedpersuade economists to examine featuresof the market like sequential trading,asymmetric information, and the role ofrisk that might explain the profitability oftechnical analysis.

The Paradox of Efficient Markets

Grossman and Stiglitz (1980) identified

a major theoretical problem with thehypothesis termed the paradox of efficientmarkets, which they developed in the con-text of equity markets. As applied to theforeign exchange market, the argumentstarts by noting that exchange rate returnsare determined by fundamentals likenational price levels, interest rates, andpublic debt levels, and that informationabout these variables is costly for tradersto gather and analyze. The traders mustbe able to make some excess returns bytrading on this analysis, or they will not doit. But if markets were perfectly efficient,the traders would not be able to makeexcess returns on any available information.Therefore, markets cannot be perfectly effi-cient in the sense of exchange rates alwaysbeing exactlywhere fundamentals suggestthey should be. Of course, one resolutionto this paradox is to recognize that marketanalysts can recover the costs of some fun-damental research by profiting from havingmarginally better information than the restof the market on where the exchange rate

should be. In this case, the exchange rateremains close enough to its fundamentalvalue to prevent less informed people fromprofiting from the difference. Partly forthese reasons, Campbell, Lo, and MacKinlay(1997) suggest that the debate about per-fect efficiency is pointless and that it ismore sensible to evaluate the degree ofinefficiency than to test for absoluteefficiency.

Empirical Reasons to Suspect Failureof Efficient Markets

The miserable empirical performance

of standard exchange rate models is anotherreason to suspect the failure of the efficientmarkets hypothesis. In an important paper,Meese and Rogoff (1983) persuasivelyshowed that no existing exchange rate modelcould forecast exchange rate changes betterthan a no-change guess at forecast horizonsof up to one year. This was true even whenthe exchange rate models were given truevalues of future fundamentals like outputand money. Although Mark (1995) andothers have demonstrated some forecastingability for these models at forecasting hori-

zons greater than three years, no one hasbeen able to convincingly overturn theMeese and Rogoff (1983) result despite 14years of research. The efficient marketshypothesis is frequently misinterpreted asimplying that exchange rate changes shouldbe unpredictable; that is, exchange ratesshould follow a random walk. This is incor-rect. Equation 2 shows that interest ratedifferentials should have forecasting powerfor exchange rate changes, leaving excessreturnsunpredictable. There is, however,convincing evidence that interest ratesare notgood forecasters of exchange ratechanges.17 According to Frankel (1996),this failure of exchange rate forecastingleaves two possibilities:

Fundamentals are not observed wellenough to allow forecasting ofexchange rates.

Exchange rates are detached fromfundamentals by (possibly irrational)swings in expectations about future

values of the exchange rate. Thesefluctuations in exchange rates areknown as bubbles.18

Which of these possibilities ismore likely? One clue is given by therelationship between exchange rates andfundamentals when expectations about thevalue of the exchange rate are very stable,as they are under a fixed exchange rate

17 Engel (1995) reviews thefailure of this theory, calleduncovered interest parity.

18 Swings in expectations thatare subsequently justified bychanges in the exchange rateare known as rational bubbles.Swings that are not consistentwith the future path of exchangerates are irrational bubbles.



33


12/16

regime. A fixed exchange rate regime isa situation in which a government is com-mitted to maintaining the value of itscurrency by manipulating monetary policyand trading foreign exchange reserves.Fixed exchange rate regimes are contrastedto floatingregimes, in which the governmenthas no such obligation. For example, mostcountries in the European Union had atype of fixed exchange rate regime, knownas a target zone, from 1979 through theearly 1990s. Fixed exchange rates anchorinvestor sentiment about the future valueof a currency because of the governmentscommitment to stabilize its value. Iffundamentals, like goods prices, or expec-tations based on fundamentals, rather thanirrationally changing expectations, drivethe exchange rate, the relationshipbetween fundamentals and exchange ratesshould be the same under a fixed exchangerate regime as it is under a floating regime.This is not the case. Countries that move

from floating exchange rates to fixedexchange rates experience a dramaticchange in the relationship between pricesand exchange rates. Specifically, realexchange rates (exchange rates adjustedfor inflation in both countries) are muchmore volatile under floating exchange rateregimes, where expectations are not t ieddown by promises of government interven-tion. Figure 6 illustrates a typical case:

When Germany and the United States ceasedto fix their currencies in March 1973, thevariability in the real $/DM exchange rateincreased dramatically. This result suggests

that, contrary to the efficient marketshypothesis, swings in investor expectationsmay detach exchange rates from fundamentalvalues in the short run.

Why Do Bubbles Arise?

If traders might profit by anticipatingswings in investor expectations, then theefficient markets hypothesis needs signifi-cant adjustment. The structure of theforeign exchange market has severalfeatures that might help drive these swings

in expectations that produce bubbles.Most foreign exchange transactions areconducted by large commercial banks infinancial centers like London, New York,Tokyo, and Singapore. These large banksmake a market in a currency by offeringto buy or sell large quantities (generallymore than $1 million) of currencies for aspecific price in another currency (e.g.,the dollar) on request. The exchange ratesat which they are willing to buy or sell dol-lars are known as the bid and ask pr ices,respectively. The market is highly compet-itive, and transactions occur 24 hours a dayover the telephone and automated tradingsystems. The first feature of this market thatmight influence technical trading is that spe-cific transactions quantities and prices arenot public information; the market is non-transparent. But the bid and ask exchangerates are easy to track, as banks freely quotethem to any participant. Second, the tradestake place sequentiallyi.e., there is time tolearn from previous trades. Third, the partic-ipants in this market differ from one another

in the information they have and their will-ingness to tolerate risk.19 In other words, theparticipants are heterogeneous.

How might these features combine toproduce bubbles? To the extent that someparticipants are better informed about cer-tain fundamentals than other agents (forinstance, they will know more about theirown and their customers demand forforeign exchange), the trading behavior of

19 It has long been assumed thatthere is little or no private infor-mation in foreign exchangemarkets, but this view hasbeen forcefully challenged withrespect to intraday trading byIto, Lyons, and Melvin (1997).



34

Figu re 6

NOTES: These changes become much more volatile after the end of the Bretton Woods system of fixed exchange rates in March 1973. The vertical line denotes this break

date in the series. Data cover January 1960February 1997.

16

82787470661962

4

4

12

0

8

8

12

M on t h l y Pe r cen t ag e Chang es i n t he$ / D M R ea l Ex ch a n g e Ra t e

86 90 94 98


13/16

the informed participants will revealsome of their private information to theun informed agents. For example, if theinformed agents know of fundamental

forces that are likely to make the exchangerate rise in the future, they are likely to buythe foreign currency and thereby bid up thepublicly observed bid and ask prices. Theuninformed agents might infer from the risethat the rate will continue to rise and, asa result, they might buy more foreignexchange, pushing the rate up themselvesin a self-fulfilling prophecy.20 This inferencefrom past price behavior is extrapolative tech-nical analysis: It assumes that the exchangerate will continue moving as it has in therecent past. The uninformed traders may

continue to buy foreign exchange past thepoint where it is supported by fundamentals.Although this story is most plausible for veryhigh-frequency (intraday) trading, it mightalso generate longer-term swings in theexchange rate.

There are other explanations forextrapolative trading that jettison theassumption of rational behavior in favor ofthe study of how people really make deci-sions. This field, called behavioral finance,has concentrated on examples of seemingirrationality in decision making. Two find-ings of this field are that (1) experimentalparticipants seem unusually optimisticabout their chances for success in gamesand (2) the behavior and opinions ofmembers of a group tend to reinforcecommon ideas or beliefs.21 For example,members of a jury may become more con-fident about their individual verdicts if theother members of the group agree.

Either explanation for extrapolativetrading implies that bubbles may be producedby slow dissemination of private information

into the market, coupled with extrapolativetrading rules. There is some evidence tosupport this explanation. Eichenbaum andEvans (1995) found that foreign exchangemarkets reacted gradually to money supplyshocks, over a period of many months,instead of instant ly incorporating the newinformation. Surveys revealed that foreignexchange market participants expectationsare extrapolative at horizons up to six

months. That is, if the exchange rate hasrisen recently, market part icipants expect itto continue to rise in the near future(Frankel and Froot, 1987). Also, the suc-

cess of extrapolative traders tends to feedon itself. Frankel and Froot (1990) arguethat extrapolative traders success duringthe early part of the large dollar apprecia-tion of 1981-1985 convinced many othertraders to follow extrapolative rules,driving the dollar up even further.

Central Bank Intervention

The other popular explanation for theapparent profitability of technical tradingrules is that technical traders are able to

profit consistently from central bank inter-vention. Some central banks frequentlyintervene (buy and sell currency) in theforeign exchange market to move theexchange rate to help influence other vari-ables like employment or inflation.22

Because these actions are designed to con-trol macroeconomic variables rather thanto make money, central banks may bewilling to take a loss on their trading.Trading ru le profits may represent atransfer from central banks to technicaltraders. Lebaron (1996) found that mosttrading rule profits were generated on theday before a U.S. intervention. Neely andWeller (1997) find that intelligenttrading rules tend to trade against the Fed;that is, they tend to buy dollars when theyfind out the Fed is selling dollars. Thistantalizing story does not fit all the facts,however. For example, Leahy (1995) findsthat U.S. foreign exchange operationsmake positive profits, on average.23 Thisfinding is inconsistent with the idea thatcentral banks are giving money away to

technical traders.

Why Are the Profits NotArbitraged Away?

Whether the trends or inefficiencies inexchange rates are created by swings inexpectations or by central bank intervention,efficient market advocates would ask whyany predictable returns in exchange rates

20 Treynor and Ferguson (1985),Brown and Jennings (1 989 ),Banerjee (1 992 ), and Kirman(1993) construct models ofbehavior in which informationis inferred from the actions ofothers. One easily understoodexample is the problem of con-sumers who must choosebetween two restaurants. Oneseemingly sensible strategy forchoosing would be to go to themore crowded restaurant on

the theory that it is likely to becrowded because it has betterfood. This phenomenondepends on asymmetricinformation.

21 Shiller (1988) and Shleifer andSummers (1990 ) discussbehavioral finance in moredetail. Ohanian (1996 ) consid-ers the reasons for the collapseof bubbles.

22 In the United States, theFederal Reserve and the U.S.

Treasury generally collaborateon foreign exchange interven-tion decisions, and operationsare conducted by the FederalReserve Bank of New York onbehalf of both.

23 See Szakmary and Mathur(1996) for more on centralbank intervention and tradingrule profits.



35


14/16

should not be arbitraged away. One answerto this question is that speculators haveshort horizons and are deterred from spec-ulating against the trends by the risk thatsuch a strategy would incur. There are sev-eral reasons for this: First, traders typicallyoperate on margin, borrowing some of the

money with which they trade. With a lim-ited line of credit, the borrowing costswould add up if traders were not able toturn a quick profit. Second, a tradersperformance is typically evaluated onrelatively short horizons (less than a year).Third, there may be institutional or legalrestrictions that prevent some types ofenterprises from taking on excessiveexchange risk. And finally, traders do not

know the equilibrium value of the exchangerate with any certainty, so they cannot dis-tinguish bubbles from movements infundamentals. Investors who bet on long-run reversion to fundamental values inexchange rates may be wiped out by short-run deviations away from those values.24

Explaining the success of technicaltrading rules with bubbles begs one morequestion: Why do destabilizing extrapolativetraders not lose their money? Friedman(1953) showed that destabilizing specula-tion is doomed to lose money and so drivethe speculators out of the market. Friedmanargued that speculation can only destabilizeasset prices if the speculators consistentlybuy when the asset price is above its equi-

24 Both Shleifer and Summers(1 990 ) and Shleifer andVishny (1 997 ) discuss theimportance of risk in speculat-ing against bubbles.

25 Essentially t he same argumentis presented more simply inShleifer and Summers (1 990 ).



36

HO W TO M EA SURE RISK ?

The simplest widely used measure of risk is the Sharpe ratioor the ratio of the

average annual excess return to a measure of excess return volatility called the stan-dard deviation. Higher Sharpe ratios are desirable because they indicate either higheraverage excess returns or less volatility. A commonly used benchmark of a goodSharpe ratio is that of the S&P 500, which Osler and Chang (1995) estimated to beabout 0.32 from March 1973 to March 1994.

A major drawback to Sharpe ratios is that they ignore an important idea infinance: An investment is risky only to the extent that its return is correlated withthe return to a broad measure of the investments available. To see this, consider therisk associated with holding a portfolio of assets whose returns are each individuallyvolatile but completely independent of each other. Each year, the assets in the portfo-lio that do unusually well will tend to offset those that do unusually poorly. Theportfolio as a whole will be much less risky than any of the individual assets. Themore assets in the portfolio, the less risky it will be. In fact, if enough of these inde-

pendent assets are grouped together into a portfolio, the return on this portfoliobecomes certain. This means that investors do not need to be compensated for hold-ing risky assets that are not correlated with all the other assets they can buy (the mar-ket portfolio), because the risk of each uncorrelated asset can be reduced to zero ifthe portfolio contains a large enough variety of these assets. On the other hand,assets for which returns are positively correlated with those of the other assets onthe market need a higher expected return to convince investors to hold them.

This idea motivates the second measure of riskiness, the CAPM beta: the coeffi-cient from the linear regression of an assets (or trading rules) excess return on theexcess return of a proxy for the market portfolio, the return to a broad equity indexlike the S&P 500. An estimated beta equal to zero means that the trading rule isbearing no systematic risk, while significantly positive betas indicate that a tradingstrategy is bearing some risk, and a beta equal to one means that the trading rulemoves closely with the market, so that following it requires the investor to acceptsignificant risk.


15/16

librium value (driving the price up further)and sell when the asset price is below itsequilibrium value; as the destabilizingspeculators lose their money, he

maintained, they will have less effect on themarket. The corollary to this argument isthat all successful speculation is stabilizing.Delong, Schleifer, Summers, and Waldman(1989) constructed a noise trader modelthat questioned this logic, however.25 Theyshowed that irrational (noise) traderscould create so much risk in asset marketsthat the returns to those assets would haveto be unusually high for rational traders totrade in them at all. In other words, theirrational traders make unusually highreturns (on average) by foolishly pursuing

risky strategies. Some go out of business,but, on average, this group increases itsmarket position.

CONCLUSION

Technical analysis is the most widelyused trading strategy in the foreign exchangemarket. Traders stake large positions ontheir interpretations of patterns in the data.Economists have traditionally rejected theclaims of technical analysts because of theappealing logic of the efficient marketshypothesis. More recently, however, the dis-covery of profitable technical trading rulesand other evidence against efficient marketshave led to a rethinking about the importanceof institutional features that might justifyextrapolative technical analysis such as pri-vate information, sequential trading, andcentral bank intervention, as well as therole of risk.

The weight of the evidence nowsuggests that excess returns have beenavailable to technical foreign exchange

traders over long periods. Risk is hard todefine and measure, however, and thisdifficulty has obscured the degree of ineffi-ciency in the foreign exchange market.There is no guarantee, of course, that tech-nical rules will continue to generate excessreturns in the future; the excess returnsmay be bid away by market participants.Indeed, this may already be occurring.Continued research on high-frequency

transactions data or experimental workon expectations formation may provide abetter understanding of market behavior.

REFERENCES

Ball, Ray. The Theory of Stock Market Efficiency: Accomplishments andLimitations, Journal of Applied Corporate Finance(Spring 1995) ,pp. 4-17.

Banerjee, Abhijit V. A Simple Model of Herd Behavior, QuarterlyJournal of Economics (August 19 92) , pp. 797-817.

Brown, David P., and Robert H. Jennings. On Technical Analysis, TheReview of Financial Studies(1989 ), pp. 527-51.

Campbell, John Y., Andrew W. Lo, and Archie Craig MacKinlay. TheEconometrics of Financial Markets, Princeton University Press, 1997.

Creswell, Juli. Currency Market Expects Rate Cut By Bank of Japan,

Wall Street Journal, September 5, 1995 , p. C16.

DeLong, J. Bradford, Andrei Shleifer, Lawrence H. Summers, and RobertJ. Waldmann. Noise Trader Risk in Financial Markets, Journal ofPolitical Economy(August 19 90) , pp. 703 -38.

Eichenbaum, Martin, and Charles L. Evans. Some Empirical Evidenceon the Effects of Shocks to Monetary Policy on Exchange Rates,Quarterly Journal of Economics(November 1995 ), pp. 97 5-1009.

Engel, Charles. Why is the Forward Exchange Rate Forecast Biased? ASurvey of Recent Evidence, Federal Reserve Bank of Kansas CityWorking Paper 95-06, September 1995.

Fama, Eugene F. Efficient Capital Markets: A Review of Theory andEmpirical Work, Journal of Finance(May 1970), pp. 383-417.

Fortune, Peter. Stock Market Efficiency: An Autopsy?, New EnglandEconomic Review, Federal Reserve Bank of Boston (March/ April1991), pp. 18-40.

Frankel, Jeffrey. How Well Do Foreign Exchange Markets Function:Might a Tobin Tax Help?, National Bureau of Economic ResearchWorking Paper 5422 , January 19 96.

_ _ _ _ _ _ _ and Kenneth A. Froot. Using Survey Data to Test StandardPropositions Regarding Exchange Rate Expectations, The AmericanEconomic Review(March 1987), pp. 133-53.

__ _ _ _ _ _ and __ _ _ _ _ _ . Chartists, Fundamentalists and theDemand for Dollars, Private Behavior and Government Policy inInterdependent Economies, Anthony S. Courakis and Mark Taylor,

eds., Clarendon Press, 1990.

Friedman, Milton. The Case for Flexible Exchange Rates, Essays inPositive Economics, University of Chicago Press, 1953.

Grossman, Sanford J., and Joseph E. Stigli tz. On the Impossibility ofInformationally Efficient Markets, The American Economic Review(June 1980), pp. 393-408.

Ito, Takatoshi, Richard K. Lyons, and Michael T. Melvin. Is There PrivateInformat ion in the FX Market? The Tokyo Experiment, NationalBureau of Economic Research Working Paper 5936, February 1997.



37


16/16

Jensen, Michael C. Some Anomalous Evidence Regarding MarketEfficiency, Journal of Financial Economics(June/ September 1978 ),pp. 95-101.

Kirman, Alan. Ants, Rationalit y and Recruitment, Quarterly Journal of

Economics(February 199 3), pp. 1 37-56.

Kocherlakota, Narayana R. The Equity Premium: It s Still a Puzzle,Journal of Economic Literature(March 1996), pp. 42-71.

Leahy, Michael P. The profitability of US intervention in the foreignexchange markets, Journal of International Money and Finance(December 1995), pp. 823-44.

Lebaron, Blake. Technical Trading Rule Profit abili ty and Central BankIntervention, National Bureau of Economic Research Working Paper5505, March 1996.

Levich, Richard M., and Lee R. Thomas. The Significance of TechnicalTrading-Rule Profits in the Foreign Exchange Market: A BootstrapApproach, Journal of International Money and Finance(October1993) , pp. 451-74.

Malkiel, Burton G. A Random Walk Down Wall Street, Fifth Edition,W. W. Norton & Company, 1990.

Mark, Nelson C. Exchange Rates and Fundamentals: Evidence onLong-Horizon Predictability, The American Economic Review(March 1995), pp. 201-18.

Meese, Richard A., and Kenneth Rogoff . Empirical Exchange RateModels of the Seventies: Do They Fit out of Sample?, Journal ofInternational Economics(February 1983), pp. 3-24.

Moorthy, Vivek. Efficiency Aspects of Exchange Rate Response toNews: Evidence from U.S. Employment Data, Journal of International

Financial Markets, Institutions and Money(1995) , pp. 1-18.Murphy, John J. Technical Analysis of the Futures Markets, New York

Institute of Finance, Prentice-Hall, New York, 1986.

Neely, Chris, and Paul Weller. Technical Analysis and Central BankIntervention, Federal Reserve Bank of St. Louis Working Paper 97-002A, January 1997.

_ _ _ _ _ _ _ , __ _ _ _ _ _ , and Robert Dittmar. Is Technical AnalysisProfitable in the Foreign Exchange Market? A Genetic ProgrammingApproach, Forthcoming in Journal of Financial and QuantitativeAnalysis(December 1997).

Ohanian, Lee E. When the Bubble Bursts: Psychology orFundamentals?, Business Review, Federal Reserve Bank of

Philadelphia (January/ February 1996) , pp. 3-13.

Osler, Carol L., and P. H. Kevin Chang. Head and Shoulders: Not Just aFlaky Pattern, Federal Reserve Bank of New York Staff Paper 4,August 1995.

Paulos, John Allen. A Mathematician Reads the Newspaper,Basic Books, 1995.

Pring, Martin J. Technical Analysis Explained, Third Edition,McGraw-Hill, 199 1.

Roll, Richard. What Every CFO Should Know About Scientific Progressin Financial Economics: What is Known, and What Remains to beResolved, Financial Management (Summer 1994) , pp. 69-75.

Rosenberg, Michael R., and Eric A. Shatz. The Merrill Lynch Guide toUnderstanding and Using Technical Analysis, Merrill Lynch and Co.,Global Securities Research & Economics Group, 1995.

Samuelson, Paul A. Proof that Properly Anticipated Prices FluctuateRandomly, Industrial Management Review(1965 ), pp. 41-49.

Shiller, Robert J. Fashions, Fads and Bubbles in Financial Markets,Knights, Raiders and Targets: The Impact of the Hostile Takeover,

J. Coffee, S. Ackerman, and L. Lowenstein, eds., Oxford UniversityPress, 1988 . Reprinted in Market Volatility, by Robert J. Shiller,MIT Press, 1989.

Shleifer, Andrei, and Lawrence Summers. The Noise Trader Approachto Finance, Journal of Economic Perspectives(Spring 1990) ,pp. 19-33.

_ _ _ _ _ _ _ and Robert W. Vishny. The Limits of Arbitrage, Journal ofFinance(March 1997), pp. 35-55.

Siegel, Jeremy J., and Richard H. Thaler. Anomalies: The EquityPremium Puzzle, Journal of Economic Perspectives(Winter 1997),pp. 191-200.

Sweeney, Richard J. Beating the foreign exchange market, Journal ofFinance(March 1986), pp. 163-82.

Szakmary, Andrew C., and Ike Mathur. Central Bank Intervention andTrading Rule Profits in Foreign Exchange Markets, Journal ofInternational Money and Finance(August 1 997) , pp. 513 -35.

Taylor, Mark P., and Helen Allen. The use of technical analysis in theforeign exchange market, Journal of International Money andFinance(June 1992), pp. 304-14.

Treynor, Jack L., and Robert Ferguson. In Defense of TechnicalAnalysis, Journal of Finance(July 1985), pp. 757-73.



forex laymen guide

Documents