Manna from heaven': does the presence of central banks make technical analysis profitable?

Trivedi, Smita Roy Article Manna from heaven': does the presence of central banks make technical analysis profitable? European Journal of Economics and Economic Policies: Intervention (EJEEP) Provided in Cooperation with: Edward Elgar Publishing Suggested Citation: Trivedi, Smita Roy (2021) : Manna from heaven': does the presence of central banks make technical analysis profitable?, European Journal of Economics and Economic Policies: Intervention (EJEEP), ISSN 2052-7772, Edward Elgar Publishing, Cheltenham, Vol. 18, Iss. 1, pp. 11-28, https://doi.org/10.4337/ejeep.2020.0072 This Version is available at: https://hdl.handle.net/10419/277497 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ ‘Manna from heaven’: does the presence of central banks make technical analysis profitable? Smita Roy Trivedi* Assistant Professor of Money, International Banking, and Finance Area Group, National Institute of Bank Management, Pune, India Profitability of technical-analysis strategies has been explained with reference to central-bank intervention in markets (Neely 1998; LeBaron 1999; Saacke 2002). I argue that central-bank intervention is a market shock which leads to a generation of trends, making technical analysis profitable. Looking at empirical evidence from the Indian foreign-exchange market, I find returns calculated for the entire period are consistently and substantially higher than when intervention periods are removed. Thirteen out of the 15 strategies demonstrate higher returns with intervention periods included, compared to without intervention periods. The Kolmogorov–Smirnov sample tests show statistically significant differences in the returns between the entire period and the without-intervention period for four strategies, which is confirmed by bootstrap estimation. The paper contributes first by including actual trading strategies in the empirical testing of profitability of technical analysis and second by emphasizing the efficacy of technical analysis rather than the action of the central bank itself in explaining profitability, in a departure from the existing literature. Keywords: central bank intervention, technical analysis, profitability of trading strategies, Reserve Bank of India JEL codes: E44, E58, F31 1 INTRODUCTION Traders concur on the profitability of technical analysis strategies, in spite of the deep academic distrust of it. If a random walk holds, the possibility of excess returns out of technical analysis strategies is ruled out (Fama 1998). However, a host of empirical studies have shown that technical analysis is profitable and generates excess returns, eschewing the academic negation of it (Pinches 1970; Surajaras/Sweeney 1992; Levich/Thomas 1993; Menkhoff/Schlumberger 1995; Neely et al. 1997; LeBaron 1999; Saacke 2002; Menkhoff/Taylor 2007). If technical analysis is indeed profitable, what explains its profitability? I contend that central-bank intervention plays a role in generating trends which are better interpreted by technical analysis strategies. I examine this for the Reserve Bank of India, India’s central bank, which consistently intervenes in the market to contain volatility. Using daily data from the Indian foreign-exchange market, it is seen that technical analysis strategies yield greater profitability in the presence of central-bank intervention. * Email: [email protected]; [email protected]. Received 25 May 2018, accepted 9 April 2019 European Journal of Economics and Economic Policies: Intervention, Vol. 18 No. 1, 2021, pp. 11–28 First published online: September 2020; doi: 10.4337/ejeep.2020.0072 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd The Lypiatts, 15 Lansdown Road, Cheltenham, Glos GL50 2JA, UK and The William Pratt House, 9 Dewey Court, Northampton MA 01060-3815, USA The presence of the central bank does lead to unexpected and (unexplained?) gains in the market, much like ‘manna from heaven.’ My findings are in conformity to an increasing literature in recent years showing that the profitability of technical analysis can stem from the presence of central-bank intervention in the market, put by LeBaron (1999: 137) as: ‘Federal Reserve activity has something to do with the observed predictability.’Identifying periods of intervention, I calculate the returns generated from 15 technical analysis strategies over a period of two years, 2015–2017, on a daily frequency. To address the unavailability of daily data on intervention in the Indian market, data on big deals by public-sector banks in the foreign-exchange market are employed, through which the central bank carries out its intervention. The returns calculated for the entire period are substantially higher than when intervention periods are removed. Thirteen out of the 15 strategies demonstrate higher returns in the presence of intervention compared to without intervention periods. Further, the Kolmogorov–Smirnov test is used to compare the mean returns between the full sample and the sample without intervention. I find that for four strategies there is a statistically significant difference in the returns between the full sample and the sample without intervention period for the four strategies, which is confirmed by bootstrap estimation. To answer why intervention should lead to the increased profitability of technical analysis strategies, I look at intervention by the central bank as a shock in the market. The shock in the market generates new information leading to the creation of trends in the market. Given the suitability of technical analysis indicators in understanding trend creation, brought about by the reaction of market participants to new information, technical analysis strategies work well during such random shocks. This can help explain the increasing empirical evidence on profitability in the presence of central-bank interventions. I differ from LeBaron (1999) in focusing on the efficiency of technical analysis strategies in understanding market psychology, and not on the central bank’s actions. Evidently, technical analysis can interpret the central bank’s actions and market reactions to it well enough to generate substantial profits. The profitability of technical analysis is empirically proven and supported by anecdotal evidence from traders. Why should technical analysis be profitable if markets do indeed follow a random walk? If efficient market hypothesis works, it means that the incorporation of information and prices rules out the possibility of abnormal profits. However, in the presence of shocks to the macroeconomic system, I contend that the move away from efficiency would make technical analysis more profitable. In cases where the trend-following indicators work well for technical analysis, the creation of trends during intervention would lead to profitability of technical analysis indicators. I add to the literature in two important ways. First, along with simple technical strategies usually used for empirical testing, I introduce complex strategies popular with traders to bring it close to the real-life trading scenarios. Second, I explain the profitability of the technical analysis strategies with reference to central-bank intervention as a random shock in the market. In a departure from earlier studies which placed the onus on the central bank for such profits being generated (LeBaron 1999), I argue that it may be the efficiency of technical analysis in reading signals from the central bank and the reaction to it, rather than the actions of the central bank itself which may be responsible for such a distinct increase in profitability when intervention periods are included. 2 THEORETICAL BACKGROUND Technical analysis is the ability to forecast price movements based on qualitative and quantitative study of historical price data. Strategies developed either from visual analysis 12 European Journal of Economics and Economic Policies: Intervention, Vol. 18 No. 1 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd of graphs or statistical analyses of price patterns are used to forecast future prices. This contradicts the very heart of the random walk hypothesis. If a random walk holds, consequential changes in prices are random and therefore the forecasting of prices cannot be done ‘in any meaningful way’(Fama/Blume 1966: 226). Fama/Blume (ibid.) point out that with the random walk holding in the financial markets, any technical analysis strategy cannot outperform a buy-and-hold strategy. If markets are efficient, prices at any given point in time correctly estimate its intrinsic value based on all information available until that point in time. If, with new information coming, the changes in prices behave in a random manner or are distributed independently as a random variable (Pinches 1970), the forecasting of future prices is largely ruled out. In its narrow variant, random-walk theory postulates that future price movements cannot be predicted on the basis of past price data alone. In the broader sense, the random walk points out that present prices already reflect all past public information so that there is no scope for predicting future prices on the basis of history (ibid.). Following the seminal work of Meese/Rogoff (1983), it is held that the random walk exists in forex markets. However, if the success of technical analysis comes from understanding under-reaction or over-reaction to information as sentiment sways the market, it does not violate market efficiency. For example, Fama (1998) points out that if over-reaction to the market is matched by under-reaction at some in time, it would suggest efficiency exists over the longer run. Menkhoff/Taylor (2007) point to the belief among traders that technical analysis can represent changes in market psychology. If fundamental factors cannot reflect changes or swings in sentiment, prices will not reflect all information. Prices may be over-reacting or under-reacting to new information from random shocks in the market. Over-reaction or under-reaction reflects market sentiment, interpreted well by technical analysis. Trend-following indicators can then be successful if market participants place a significant value on psychological influences (Menkhoff/Taylor 2007). Are technical analyses self-fulfilling? If traders are confident on the ability of technical analysis indicators in interpreting psychological biases, logically technical analysis will be self-fulfilling. For this to happen, however, the same kinds of signals must elicit the same responses from traders, so that herd behavior ensues. However, given the wide variety of technical analysis rules, it is hardly plausible that they would generate uniform signals. This rules out the possibility of self-fulfilling movements generated in the market as it is unlikely that all traders trading at the same time reach the same interpretation and therefore take similar long or short positions. The earliest empirical studies on the profitability oftechnical analysis indicators questioned the success of technical analysis strategies in the presence of the random walk. Brock et al. (1992) showed, using technical analysis based on filter techniques, that profit can be generated substantially in excess of buy-and-hold returns. Later studies confirmed that technical analysis served as an important tool in the hands of market practitioners in enabling effective trading decisions (Pinches 1970; Surajaras/Sweeney 1992; Menkhoff/Schlumberger 1995; Neely et al. 1997; LeBaron 1999; Saacke 2002; Menkhoff/Taylor 2007). Moreover, Silber (1994), Szakmary/Mathur (1997), and Neely (1998) have all extended evidence that the presence of intervention is strongly associated with profits from technical analysis indicators. Neely (1998) points out that official interventions are usually during periods of sharp market movement, that is, when markets are trending, which also make technical analysis profitable. The analysis of Bundesbank interventions with high-frequency data (Frenkel/Stadtmann 2004) and daily data (Neely/Weller 2001) show that these rules are most profitable on the day before interventions take place. LeBaron (1999) points out that central-bank intervention would introduce noticeable trends in exchange-rate movement making it possible for market participants to gain from trading. The strongest explanation of ‘Manna from heaven’: does the presence of central banks make technical analysis profitable? 13 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd profitability of technical analysis strategies in the presence of central-bank intervention comes as the generation of trends post-intervention (Saacke 2002). The presence of the central bank in the market leads to different reactions from market participants. As the central bank intervenes, the participants will adjust their trading positions, keeping in view the central bank’s actions. In this case, the central bank’s action can be identified as a market shock, which will lead to a change in the market movement. Again, central banks intervene to curb volatility following an exogenous shock and in this case too the exchange-rate adjustment would lead to trend creation. The success of technical analysis indicators in this environment comes from the ability to recognize trend creation. The basic premise of technical analysis is that the market moves in trends which can be recognized through suitable indicators. Technical analysis indicators are largely trend-following. Indicators based on moving averages try to understand the start of a new trend by comparing the present price movement to longer-term averages. Divergence of prices from the longer-term averages suggests the existence of trends in the market. Momentum technical indicators, by seeing the rate of price change, try to get an early indication of the trend change. If central-bank activity leads to trend generation, it is contended that technical analysis indicators would be able to recognize the same and thereby generate excess returns. I test the profitability of technical analysis indicators in the presence the Indian central bank, the Reserve Bank of India (RBI). The RBI intervenes in the Indian forex market, largely to keep the managed float stable (RBI 2013; Roy Trivedi 2019). The creation of a market shock when the central bank intervenes, I argue, should lead to greater profits from technical analysis in the presence of intervention. While a host of studies have looked at the profitability of technical analysis indicators in the presence of intervention, they have used only basic technical analysis strategies like moving-average indicators. I add to the literature by bringing, in addition, a host of strategies covering moving averages, momentum, and volatility indicators. While moving averages are helpful to identify the trend after it has been in motion for some time, momentum indicators by capturing the incremental changes in price movement lead to identification of trend earlier. 3 DATA AND METHODOLOGY Data on exchange rates has been taken from the Cogencis trading platform (Cogencis Information Services 2015) on a daily frequency. Data on interest rates are obtained from the RBI and Federal Reserve database. Intervention data by the RBI are not available on a daily basis. Intervention data would lose granularity if we took it on a monthly basis; therefore it is very important to take a proxy of intervention at the daily level. As a proxy for daily intervention, I use data on big deals by public-sector banks as the RBI intervenes through the big public-sector banks in the Indian market. Intervention in the Indian market is thus secret (Roy Trivedi 2019). This data has also been taken from the Cogencis trading platform (Cogencis Information Services 2015), which reports the ‘Key deals in the Indian Foreign Exchange market,’detailing which banks groups have bought or sold in the Indian foreign-exchange market. Table 1 shows a sample of information taken from Cogencis. It shows the key deals in the Indian foreign-exchange market on Tuesday 3 January 2017. It is reported that a large state-owned bank was buying dollars upwards of 68.05 rupees. I take this to be proxy for central-bank intervention. A criticism of this methodology is that the big deals reported by public-sector banks could also be bank’s own deals. There is no way to distinguish between the two. However, since intervention is secret, market 14 European Journal of Economics and Economic Policies: Intervention, Vol. 18 No. 1 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd participants would consider any big deal by a public-sector bank as a possible contender for intervention, building up speculation around it. For the purpose of our paper, it therefore serves as a proxy variable, since it also acts as a market shock. Even if market participants interpreted wrongly any deal as central-bank intervention, markets would be impacted in the same way as in a case of real intervention. Therefore, the data on big deals serve well as a proxy for intervention given the unavailability of daily intervention data. The daily data are mapped with intervention data (as proxied by the big deals of public-sector banks) in the foreign-exchange market. Fifteen technical analysis strategies, commonly used by traders, are chosen and profits from these strategies calculated on the historical data. Technical strategies are designed to generate buy-and-sell signals on the basis of qualitative or quantitative studies of the data. I detail below the technical analysis strategies used and methodology followed for calculations of returns in each case. The moving average is one of the simplest and yet most powerful technical analysis indicators. The basic moving-average signals used commonly in empirical literature stipulate a buy (or sell) signal if prices are greater (or smaller) than the moving average. Formally, for an exponential moving average (EMA) of ∝periods, if Pt>EMA∝ t:þ1ðbuyÞ(1) if Pt<EMA∝ t:−1ðsellÞ(2) EMA∝ t¼EMA∝ t−1ð1−wtÞþPtwt;(3) where wt¼2 ∝þ1and ∝¼number of periods. And the EMA for period 1 is taken as the simple moving average (SMA): if SMAt¼∑tþ∝−1 t¼1Pt ∝:(3a) A commonly used moving-average strategy is the double-crossover method using shorterand longer-period moving averages. Double-crossover strategies generate trading signals depending on the position of the shorterand longer-period moving average. Table 1 Sample of news in Cogencis used for intervention data (Cogencis, Tuesday 3 January 2017, Mumbai): the key deals in India’s foreign-exchange market by 1315 IST a Bank Dollar/rupee level Dollars bought in the market in millions Large state-owned bank 68.05 upwards multiple levels Large Bank Dollar/rupee level Dollars sold in the market in millions UK-based bank 68.13 upwards Moderate Another UK-based bank; two US-based banks 68.06 Moderate Notes: a. These deals are not officially available. The compilation is based on information provided by interbank dealers. Trading counterparties are withheld in some cases, unknown in most cases. Source: Compiled by (‘Congencis Journalist name’), Congencis Information Services Ltd. ‘Manna from heaven’: does the presence of central banks make technical analysis profitable? 15 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd The shorter-period average follows the market closely while the longer period is the smoother indicator. Thus when the trend changes, the shorter-period average (following market action closely) will turn earlier than the longer-period average. If that is so, a downtrend will be indicated by the shorter-run average ∝1crossing the longer-run average ∝2from above. An uptrend is denoted by the shorter-run average, ∝1crossing the longer-run average ∝2from below. EMA∝1 t−1<EMA∝2 t−1and EMA∝1 t>EMA∝2 t:þ1ðbuyÞ(4a) EMA∝1 t−1>EMA∝2 t−1and EMA∝1 t<EMA∝2 t:−1ðsellÞ(4b) The addition of one more moving average is frequently done to get a confirmation on the signals, called the triple-crossover strategy. In this case, three moving averages, ∝1 ;∝2 ;and ∝3, are used, where ∝1<∝2<∝3: EMA∝2 t−1<EMA∝3 t−1and EMA∝2 t>EMA∝3 tand EMA∝1 t>EMA∝2 t>EMA∝3 t:þ1ðbuyÞ(5a) EMA∝2 t−1>EMA∝3 t−1and EMA∝2 t<EMA∝3 tand EMA∝1 t<EMA∝2 t<EMA∝3 t:−1ðsellÞ: (5b) While using historical data for back-testing strategies it is important to avoid the lookahead bias (Chan 2009). The look-ahead bias comes from incorporating in the present time period any transaction requiring any future information. For example, if closing prices are used to calculate moving averages, signals generated by a double-crossover method in any period are based on the closing prices of the same period. The signal generated by the double-crossover strategy would necessarily be available after the end of the period. Transactions on the signal therefore can happen only after the period, that is, in the next period. For calculating the transaction prices, then, taking the close price of the same period would lead to look-ahead bias. Instead, the next period’s open, close, or average should be taken. The first four strategies used in the paper, R1, R2, and R3, are variants of the movingaverage strategy. R1 is the returns calculated from applying the double crossover strategy, with a 5and 10-period moving average. R2 is the returns calculated from applying the double-crossover strategy with a 10and 20-period moving average. R3 uses a double-crossover strategy (5and 10-period) with a 1 percent filter, commonly used by traders. Whipsaws occur when the crossover strategy generates false signals which are reversed within a few periods. This means that a buy (sell) signal is generated but the market turns down (up), swiftly leading to losses for the trader. Such whipsaws are common in sideways-moving markets or when there is high volatility in the market. These trades are not profitable for the traders as the trend is not correctly recognized. To address this problem, the filter technique is frequently used to generate better signals. In this case the trader will take the buy (sell) signal as confirmed if the price holds above (below) a certain percentage of the price given at the time the signal was generated. In each of the above cases mentioned, the transaction price (TP) is the open price in the next period as the strategy is assumed to materialize into a decision in the next period and returns are calculated accordingly. Returns are calculated as shown below, assuming a buy position is held until the next sell signal, and the position is short until the next buy signal. In other words, a trader will remain long on a strategy 16 European Journal of Economics and Economic Policies: Intervention, Vol. 18 No. 1 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd until s/he gets an opportunity to sell, and similarly s/he will remain short till s/he gets an opportunity to buy back the security. For a buy decision, Rt¼1−TP t TPt−1 :(6a) And for a sell decision, Rt¼TP t TP t−1 −1:(6b) R4 calculates excess or dynamic returns for the double-crossover strategy using 5and 10-period moving averages with a 5 percent filter. R5 calculates excess or dynamic returns for the double-crossover strategy using 10and 20-period moving averages. R6 calculates excess returns for the triple-crossover strategy. In calculating the returns, it does not take the transaction price but considers a ‘buy’position as long (þ1) and a sell position as short (−1). Excess returns for each period adjusted for interest-rate differential (Saacke 2002) is calculated as R¼log Pt Pt−1  −r−r 260 ;(7) where r−r* is the interest-rate differential between the foreign and the home country (*), and P t is the exchange rate at period t(closing price), defined as the number of units of home-country currency for one unit of foreign-country currency. R7 brings in another simple strategy, comparing prices with the EMA 20 (where the superscript denotes the number of periods) to generate trading signals and using normal returns for calculating the profits from the strategy. If there is a rise in price, the return will be positive, while a fall in price will generate negative returns. Therefore the ‘buy’ decision will be taken to be successful if it is followed by a price rise, and the ‘sell’decision will be taken to be successful if it is followed by a fall in price. As the dummy for ‘buy’or long position is +1, profits would be positive only ifthereturnsarepositivefor the given period. Similarly a ‘sell’decision or short position (−1) would generate positive profits when the return is negative.R8,R9,andR10comparepriceswithEMAs 5 , 10 ,and 20 respectively, using excess returns for calculating profits from the said strategies. R11 and R12 use the volatility-based indicator, Bollinger bands, to calculate the trading profits. Bollinger bands plot the upper and lower ranges for price movement based on the volatility of the series. The upper and lower bands are constructed at ±2 standard deviation around the mean as follows: BBUB ¼MA∝þ2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð∑∝ t¼1Pt−  PÞ2 ∝ s(8a) BBLB ¼MA∝−2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð∑∝ t¼1Pt−  PÞ2 ∝ s:(8b) ‘Manna from heaven’: does the presence of central banks make technical analysis profitable? 17 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd The above formulae underline the logic behind Bollinger bands: if markets follow a standard normal distribution, Bollinger bands will cover 95 percent of the market movement. The difference between the upper band and the lower band (bandwidth) is an indicator of the extent of volatility of the series. The Bollinger bands strategy requires determining the volatility of the series as it is expected that periods of high volatility, represented by higher bandwidth, are followed by periods of low volatility, with the bands becoming closer. In market parlance this is called as a ‘squeeze.’A squeeze is generally followed by a period of increased volatility. R11 uses a Bollinger band strategy whereby decisions are taken in periods of higher volatility. The strategy is framed in the following way. I differentiate periods of high and low volatility based on the bandwidth. A variable is assigned the dummy value, 1, when the difference between the upper band and lower band exceeds the 5-period average difference. Otherwise the dummy takes a value of 0. Once the price reaches the upper band during a period of high volatility, or as the Bollinger bands expand, we expect prices to go down. In this case a ‘sell’decision should be taken. Similarly, once prices reach the lower band during a period of expanding volatility, they are expected to go up; in this case a‘buy’decision is taken. The strategy is assumed to materialize into a decision the next day at open prices and returns are calculated accordingly as in the cases of first three strategies. R12 uses a squeeze for trading: in periods of low volatility (that is, volatility <average volatility), if the high of the day is greater than the upper band, a sell signal is generated as we expect prices to fall. If the low of the day is lower than the lower band we expect prices to go up from there, and a buy strategy is generated. Similar to the previous strategy, the transaction price is taken to the open price of the next period and returns are calculated accordingly. I therefore distinguish between high and low volatility for Bollinger bands trading strategy design. The next group of strategies use momentum-based indicators. The momentum indicator compares the price movement in the present period with reference to a certain period in the past. The most basic momentum indicator is the rate of change (ROC). It is given by the simple formula: ROC ¼P Pt 100;(9) where Pis the price today and P t is the price tdays earlier. I take ROC to be 10 periods, in conformity to the moving averages used. As I have used 5-, 10-, and 20-period averages, the 10-period ROC will help in comparisons between momentum and the moving-average indicator. Multiplying it by 100 converts it to a comparable figure. This strategy requires alongpositiontobetakeniftheROCvalueisabove100andashortpositiontobe taken if the value calculated is below 100. Normal returns are used to calculate profits from the strategy R13. The relative strength index (RSI) indicator forms the basis of the next two strategies. The RSI, like any other momentum indicator, gives the strength in the market. It is given by RSIα¼100 −100 1þRS ;(10) where RS ¼up closes/down closes over a period of αperiods. In this paper, using standard market convention, I take αto be 14 periods. The RS will show higher values if there are more up closes than down closes. Up closes are marked if prices in a said period are greater than in the previous period. The greater the 18 European Journal of Economics and Economic Policies: Intervention, Vol. 18 No. 1 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd findings from the Kolmogorov–Smirnov test, I calculate the achieved significance level (ASL) of the test statistic by using bootstrap estimations with 1000 resampling. The samples are drawn with replacement making it a Markov process (Schmuland 2017). The ASL calculated for the four strategies (8, 9, 10, and 14) for which the Kolmogorov–Smirnov test showed significant difference in means is presented in the last column of Table 8. I see the ASL is significant for all four strategies. The results are also in line with the graphical analysis of Figure 1, where the differences are highest for strategies 8, 9, 10, and 14. What explains the profitability of such strategies? It implies that as a central bank intervenes in the market, it leads to new movement emerging as well as a reorientation of trading Table 8 (continued) Strategy Group DP-value Exact Bootstrap ASL (p-value) 7 0 0.0820 0.173 N/A1 −0.0005 1.000 0.0820 0.345 0.323 8 0 0.1658 0.001 0.00 1−0.0557 0.446 0.1658 0.002 0.001 9 0 0.1565 0.002 0.00 1−0.0580 0.416 0.1565 0.003 0.003 10 0 0.1769 0.000 0.00 1−0.0373 0.696 0.1770 0.001 0.000 11 0 0.0144 0.947 N/A1 −0.0082 0.983 0.0144 1.000 1.000 12 0 0.0443 0.600 N/A1 −0.0274 0.822 0.0443 0.960 0.948 13 0 0.0714 0.265 N/A1 −0.0148 0.944 0.0714 0.519 0.492 14 0 0.1314 0.011 0.00 1−0.0851 0.151 0.1314 0.022 0.02 15 0 0.0568 0.431 N/A1 −0.0543 0.464 0.0568 0.794 0.766 Source: Author’s estimates. ‘Manna from heaven’: does the presence of central banks make technical analysis profitable? 25 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd strategies. The departure of markets from efficiency as seen in over-reaction or under-reaction to information is interpreted better with technical-analysis strategies. Over-reaction or underreaction thus reflect market sentiment, interpreted well by technical analysis (in which market participants place a significant value on psychological influences). The findings suggest that central-bank intervention is suggestively coincident with trend generation in the market, which can explain why more moving-average-based strategies are found to be doing well during interventions. Tables 5 and 6 both show that the simple strategies 8, 9, and 10 have significantly high returns during intervention periods. Moving-average strategies recognize trends that are already ongoing as compared to momentum indicators which try to pre-empt the change of trend. In market parlance, therefore, momentum indicators are popular as ‘leading’indicators compared to ‘lagging’moving-average-based indicators. The intervention periods must therefore be coinciding with the period where the trend has already begun. Thus, the presence of intervention and increased profitability during such a period from moving-average indicators require the trend to be already in motion. Table 7 compares the returns from different intervention periods. I see that periods characterized by interventions on both sides of the market lead to greater profitability as well as to volatility of returns. This means that there is a trend generation associated with these periods which can explain the increased profits. Interventions on the buy side also have seen sharp volatility in technical-analysis returns during the period under consideration. This result is similar to the findings in economic literature underlining the creation of sharp market movements with central-bank intervention, and higher profitability of technical-analysis indicators during these periods (Neely 1998; LeBaron 1999; Neely/Weller 2001). Are profits from using technical-analysis strategies in the presence of central-bank intervention ‘manna from heaven’: unexpected or windfall gains? The analogy is used to highlight the substantial difference in profits seen during periods of intervention. These profits arguably are crucial for many trading portfolios. However, what is the reason behind the significantly greater profitability during these periods? We emphasize the trend generation caused by the market shock. Trend is nothing but the movement of prices in one direction. From a sideways and indecisive market move, the trend generated shows either the bears winning over the bulls, or the contrary. This requires a market shock leading to a change in the trading positions of the market participants. Central-bank intervention qualifies well as a market shock, given that central banks take care to use the signaling impacts of intervention effectively. For the Indian market, this is all the more true given the secret nature of intervention. The choice of the proxy variable used in this paper for intervention lends support to this argument. As news of big deals in the market is taken as a proxy for intervention, it is likely that many of these deals reported may be simply banks’own deals and not interventions by the central bank. However, for other market participants, there is an element of uncertainty when they see the big deals by public-sector banks, as they are unsure whether or not the central bank has really intervened. The big deals therefore will be seen as market shocks by other participants, as they are speculating whether or not the central bank has intervened through the publicsector banks. Accordingly, market participants will reorient positions leading to a change in market movement or an expected trend generation. Technical analysis thus sees an increased profitability with intervention by the central bank in the Indian forex market. 5 CONCLUSION The paper finds that central-bank intervention leads to higher profitability of technicalanalysis indicators in the Indian foreign-exchangemarket.Theprofitsareunexpected, 26 European Journal of Economics and Economic Policies: Intervention, Vol. 18 No. 1 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd windfall gains: ‘manna from heaven’to the trading desks. We contend that the trend generation brought about by central-bank intervention or any other market shock is the reason behind this increased profitability. Analysing the behavior of technical-analysis profitability in the presence of market shocks would not only lead to a better understanding of foreign-exchange market dynamics but would also help in implementing appropriate policies. ACKNOWLEDGEMENTS The author would like to thank the market participants, discussions with whom helped me to gain insight on the impact of central-bank intervention on trading. The insightful discussions with faculty colleagues at NIBM are also sincerely acknowledged. The author would also like to thank the reviewer/s for the very useful comments and suggestions. No funding or grants were applied for or obtained for the study. REFERENCES Brock, W., Lakonishok, J., LeBaron, B. (1992): Simple technical trading rules and the stochastic properties of stock returns, in: Journal of Finance, 47, 1731–1764. Chan, E.P. (2009): Quantitative Trading: How to Build Your Own Algorithmic Trading Business, The Wiley Trading Series, Hoboken, NJ: John Wiley. Chari, A. (2007): Heterogeneous market-making in foreign exchange markets: evidence from individual bank responses to central bank interventions, in: Journal of Money, Credit and Banking, 39(5), 1131–1162. Cogencis Information Services (2015): Database, URL: https://www.cogencis.com (accessed 2015). Fama, E. (1998): Market efficiency, long-term returns, and behavioral finance, in: Journal of Financial Economics, 49(1998), 283–306. Fama, E., Blume, M. (1966): Filter rules and stock market trading, in: Journal of Business, 39, 226–241. Frenkel, M., Stadtmann, G. (2004): Trading rule profitability and central bank interventions in the dollar–deutsch mark market, in: Journal of Economics and Statistics, 224(6), 653–672. LeBaron, B. (1999): Technical trading rule profitability and foreign exchange intervention, in: Journal of International Economics, 49(1), 125–143. Levich, R., Thomas, L. (1993): The significance of technical trading rule profits in the foreign exchange market: a bootstrap approach, in: Journal of International Money and Finance,12, 451–474. Meese, R., Rogoff, K. (1983): Empirical exchange rate models of the seventies: do they fit out of sample?, in: Journal of International Economics, 14, 3–24. Menkhoff, L., Schlumberger, M. (1995): Persistent profitability of technical analysis on foreign exchange markets?, in: Banca Nazionale del Lavoro Quarterly Review, 48(193), 189–215. Menkhoff, L., Taylor, M.P. (2007): The obstinate passion of foreign exchange professionals: technical analysis, in: Journal of Economic Literature, 45(4), 936–972. Neely, C.J. (1998): Technical analysis and the profitability of U.S. foreign exchange intervention, in: Federal Reserve Bank of St. Louis Review, 80(4), 3–17. Neely, C., Weller, P.A. (2001): Technical analysis and central bank intervention, in: Journal of International Money and Finance, 20(7), 949–970. Neely, C.J., Weller, P.A., Dittmar, R. (1997): Is technical analysis in the foreign exchange market profitable? A genetic programming approach, in: Journal of Financial and Quantitative Analysis, 32, 405–426. Pinches, G.E. (1970): The random walk hypothesis and technical analysis, in: Financial Analysts Journal, 26(2), 104–110. ‘Manna from heaven’: does the presence of central banks make technical analysis profitable? 27 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd RBI (Reserve Bank of India) (2013): Intervention in foreign exchange markets: the approach of the Reserve Bank of India, in: BIS, Market volatility and foreign exchange intervention in EMEs: what has changed? An overview, in: BIS Papers, No 73, URL: https://www.bis.org/publ/ bppdf/bispap73.htm. Roy Trivedi, S. (2019): The Moses effect: can central banks really guide foreign exchange markets?, in: Empirical Economics, 58, 2837–2865. Roy Trivedi, S., Apte, P.G. (2016): Central bank intervention in USD/INR market: estimating its reaction function and impact on volatility, in: Asia-Pacific Financial Markets, 23(3), 263–279. Saacke, P. (2002): Technical analysis and the effectiveness of central bank intervention, in: Journal of International Money and Finance, 21, 459–479. Schmuland, B. (2017): Stochastic processes without measure theory, URL: www.stat.ualberta.ca/ people/schmu/preprints/without.pdf (accessed 4 April 2019). Silber, W.L. (1994): Technical trading: when it works and when it doesn’t, in: The Journal of Derivatives,1(3)39–44. Surajaras, P., Sweeney, R.J. (1992): Profit-Making Speculation in Foreign Exchange Markets, Boulder, CO: Westview Press. Szakmary, A.C., Mathur, I. (1997): Central bank intervention and trading rule profits in foreign exchange markets, in: Journal of International Money and Finance, 16, 513–535. 28 European Journal of Economics and Economic Policies: Intervention, Vol. 18 No. 1 © 2021 The Author Journal compilation © 2021 Edward Elgar Publishing Ltd