Volume 29, Issue 3 Profitability of the O-Balace Volume Idicator William Wai Him Tsag Departmet of Ecoomics, The Chiese Uiversity of Hog Kog Terece Tai Leug Chog Departmet of Ecoomics, The Chiese Uiversity of Hog Kog Abstract I the literature, there is a lack of empirical studies documetig the profitability of volume-based techical idicators. This paper evaluates the profitability of the O-Balace Volume (OBV) tradig rule. Our result shows that the OBV tradig rule is icreasigly profitable ad rewards ivestors with otable returs i the stock markets of Greater Chia. We thak Pak-Ho Leug for able research assistace. All errors are ours. Correspodig author: Terece Chog, Departmet of Ecoomics, The Chiese Uiversity of Hog Kog, Shati, N.T., Hog Kog. Email: chog2064@cuhk.edu.hk. Webpage: http://www.cuhk.edu.hk/eco/staff/tlchog/tlchog3.htm. Citatio: William Wai Him Tsag ad Terece Tai Leug Chog, (2009) ''Profitability of the O-Balace Volume Idicator'', Ecoomics Bulleti, Vol. 29 o.3 pp. 2424-2431. Submitted: Jul 20 2009. Published: September 24, 2009.
1. Itroductio The profitability of techical tradig rules has bee a popular research topic sice the pioeerig work of Ferguso ad Treyor (1985). Most of the previous studies examie the profitability of price-based tradig rules, such as the variable legth movig average (VMA) rule (Brock et al.; 1992; Hudso et al., 1996; Mills, 1997) 1, cadle sticks (Marshall et al., 2008) ad mometum (Chog ad Ip, 2009). Results for the performace of these price-based tradig rules are mixed. I some circumstaces, these tradig rules do reward ivestors with returs otably higher tha the buy-ad-hold bechmark. 2 A possible explaatio for the mixed results is that covetioal tradig strategies rarely icorporate volume as a key factor i geeratig tradig sigals. 3 Volume cotais useful iformatio of market setimet that caot be fully reflected by price iformatio. This paper bridges this gap by assessig the performace of the O-Balace Volume (OBV) idicator. 4 The OBV is calculated usig the iformatio of the closig price ad the tradig volume of a stock. Mathematically, the OBV of a stock at time t is defied as: OBV t OBV + V, C > C = { OBV V, C < C t 1 t t t 1 t 1 t t t 1, ad C t ad V t are the closig price ad tradig volume respectively at time t. By defiitio, the OBV icreases whe prices rise ad vice versa. Note that the OBV is defied recursively. To complete the defiitio, we defie the OBV at the referece day t=0 to be zero, i.e., we let OBV 0 = 0. At t=1, OBV 1 =V 1 or - V 1, depedig o whether the price at t=1 rises or falls relative to the price at t=0. A high value of OBV idicates good market setimet. The OBV ca also be used to predict market reversal. Whe OBV treds up ad price treds dow, or vice versa, it idicates a immiet reversal of market setimet. 2. Data ad Methodology The daily closig prices 5 ad the turover by volume 6 of ie major stock idices aroud the world are retrieved from DataStream. The details are reported i Table 1. 1 The VMA rule states that oe should take a log positio if the short-term movig average is above the log-term movig average ad stay short otherwise. 2 The existece of abormal returs, however, cotradicts the efficiet market hypothesis (EMH), which purports that oe caot profit by oly usig past iformatio (Fama, 1970). Fama (1991) has also agreed the extreme versio of EMH that prices reflect all available iformatio certaily does ot stad due to positive iformatio costs. 3 Blume et al. (1994) have discussed the importace of tradig volume. 4 The OBV was first ivestigated by Woods ad Vigolia i 1946. They called it "cumulative volume". Joseph Graville ames it "o-balace volume" i his book Graville's New Key to Stock Market Profits i 1963. 5 Ulike the closig price of a idividual stock, which is i local currecy uits, the closig price of a stock market idex is the daily closig value of the respective idex, which is a uit-free measure. 6 Accordig to the defiitio i DataStream, the tradig volume of a idex is computed by summig up the daily total umber of shares traded of the respective costituet stocks. 2
Table 1: The Sample Details Coutry/Regi Idex o From To CAC40 Frace 2/1/1992 3/4/2009 DAX 7 Germay 1/8/2003 3/4/2009 DowJoes Idustrials Uited States 3/4/1989 3/4/2009 FTSE100 Uited Kigdom 3/4/1989 3/4/2009 Hag Seg Hog Kog 3/4/1989 3/4/2009 Shaghai A Chia 4/1/1993 3/4/2009 Shaghai B Chia 25/11/1994 3/4/2009 ShezheA Chia 5/10/1992 3/4/2009 TWSE Taiwa 3/4/1989 3/4/2009 The tradig rule examied i this paper is the crossover of OBV ad its movig average (). The -day at time t is defied as: t = 1 i= 1 OBV i Tradig Rule Buy at day t: Sell at day t: OBVt 1 < t 1 ad OBVt 1 > t 1 ad OBV t > t OBV t < t Thus, a log positio is take whe OBV rises above its -day movig average, ad the positio is liquidated oce OBV drops below the -day. I this paper, the performace of the 10-, 20-, 50- ad are assessed. Short-sellig is prohibited ad cosecutive buyig actios are ot allowed. To compare the returs of OBV across differet markets, we calculate the aual rate of retur, which is defied as 250 [( 1+ r )( 1+ r )( 1+ r ) ( 1+ )] 1 R = T A 1 2 3 r m, where 1 + r i = S B ( j) ( j) S(j) ad B(j) are the sellig ad buyig prices of the trasactio; m is the umber of trasactios i the sample; T is the umber of tradig days i the sample. 7 The tradig volume series of DAX dates back oly to 2003 i DataStream. 3
3. Results ad Coclusio Table 2 reports the aualized rate of retur (i percetage). Figures i paretheses are umbers of trasactios. The last colum shows the retur rate of the buy-ad hold strategy. For each idex, the rule that geerates the highest rate of retur is bolded. For istace, the rule produces the highest retur for the Hag Seg Idex over the past 20 years. The OBV outperforms the buy-ad-hold strategy i six out of the ie idices. I particular, the 10- ad perform well i predictig price movemets. The performace of the OBV is pheomeal i the stock markets of Chia. For example, the highest aual retur for the Shaghai B market is 34.85%. The itegratio of the Greater Chia has also beefited the stock markets i the regio. For Hog Kog, the highest aual retur is 13.74%. For Taiwa, the OBV geerates a retur of 10% per aum. I cotrast, the OBV i geeral caot beat the buy-ad-hold strategy i the ad Europea markets. For the Dow Joes Idustrials, CAC40 ad FTSE100, the OBV geerates returs sigificatly lower tha the buy-ad-hold strategy. The oly exceptio is the rule, which beats the buy-ad-hold strategy i the Frakfurt DAX Idex by a slight margi. Thus, the OBV tradig rule is comparatively more effective i the markets of Greater Chia tha i U.S. ad Europea markets durig our sample period. Table 2: Aual Rate of Retur of the OBV (%) 10-day 20-day Idex Europe ad CAC40-1.00(468) -1.75(331) 0.96(117) 0.94(140) 3.05 DAX -6.27(162) -5.25(108) 4.88(49) 2.53(43) 4.29 Dow Joes 1.6(585) 2.35(390) 3.34(240) 2.86(180) 6.40 FTSE100-0.53(556) 0.29(390) 1.24(246) 0.22(173) 3.3 Greater Chia TWSE 10.08(494) 8.52(334) 5.44(189) 9.26(92) -1.36 Hag Seg 9.28(454) 10.76(301) 13.74(171) 9.15(115) 8.33 Shaghai A 16.11(390) 11.10(261) 7.58(147) 7.02(97) 7.11 Shaghai B 33.64(284) 34.85(182) 17.90(107) 15.19(76) 6.32 ShezheA 24.66(380) 19.46(249) 15.05(139) 9.00(84) 7.02 Buy-ad-Hold To icorporate the effect of trasactio costs ito our aalysis, we compute the aualized trasactio costs as follows: 250c c wm A T =, where c w represets the cost of each trasactio, ad m ad T are defied i Sectio 2. Tables 3a to 3c report the aualized trasactio costs for c w =0.1%, 0.25% ad 0.5%. 4
Table 3a: Aualized Trasactio Cost for c w =0.5% 10-day 20-day Idex Europe ad CAC40 13.41(%) 9.48 3.35 4.01 0.029 DAX 13.99 9.33 4.23 3.71 0.086 Dow Joes 14.55 9.70 5.97 4.48 0.025 FTSE100 13.75 9.64 6.08 4.23 0.025 Greater Chia TWSE 12.67 8.56 4.85 2.36 0.026 Hag Seg 11.48 7.61 4.32 2.91 0.025 Shaghai A 12.29 8.22 4.63 3.06 0.032 Shaghai B 10.23 6.56 3.85 2.74 0.036 ShezheA 11.83 7.75 4.33 2.62 0.031 Buy-ad-Hold Table 3b: Aualized Trasactio Cost for c w =0.25% 10-day 20-day Idex Europe ad CAC40 6.70(%) 4.47 1.68 2.01 0.014 DAX 7.00 4.66 2.12 1.86 0.043 Dow Joes 7.27 4.85 2.98 2.24 0.012 FTSE100 6.87 4.82 3.04 2.14 0.012 Greater Chia TWSE 6.33 4.28 2.42 1.18 0.013 Hag Seg 5.74 3.80 2.16 1.45 0.013 Shaghai A 6.14 4.11 2.32 1.53 0.016 Shaghai B 5.12 3.28 1.93 1.37 0.018 ShezheA 5.92 3.88 2.16 1.31 0.016 Buy-ad-Hold Table 3c: Aualized Trasactio Cost for c w =0.1% 10-day 20-day Idex Europe ad CAC40 2.68(%) 1.90 0.64 0.80 0.006 DAX 2.80 1.87 0.85 0.74 0.017 Dow Joes 2.91 1.94 1.19 0.90 0.005 FTSE100 2.75 1.93 1.22 0.86 0.005 Greater Chia TWSE 2.53 1.71 0.97 0.47 0.005 Hag Seg 2.30 1.52 0.86 0.58 0.005 Shaghai A 2.46 1.64 0.93 0.61 0.006 Shaghai B 2.05 1.31 0.77 0.55 0.007 ShezheA 2.37 1.55 0.87 0.52 0.006 Buy-ad-Hold 5
To gauge more clearly the efficacy of the OBV rule, Table 4 reports the et relative rate of retur of the OBV rule, which is obtaied by deductig the aual trasactio cost ad the relevat buy-adhold retur from the absolute OBV retur. Table 4: Net Rate of Retur (usig 0.25% cost) of the OBV et of the buy-ad hold retur(%) Idex 10-day 20-day Europe ad CAC40-10.75-9.54-3.77-4.12 DAX -17.56-14.2-1.53-3.62 Dow Joes -12.07-8.9-6.04-5.78 FTSE100-10.17-8.12-5.1-5.22 Greater Chia TWSE 5.11 5.6 4.38 9.44 Hag Seg -4.79-1.37 3.25-0.63 Shaghai A 2.86-0.12-1.85-1.62 Shaghai B 22.2 25.25 9.65 7.5 ShezheA 11.72 8.56 5.87 0.67 I geeral, the icorporatio of trasactio cost makes the rule uprofitable i all the Europea ad markets, but profitability is preserved i the Greater Chia area. The highest et retur for each rule is bolded. It is foud that the 10-, 20- day rules are still the most profitable for the Shaghai A ad the Shaghai B idices respectively. The rule is ow the most profitable i the Taiwa market. The ettig of trasactio cost favors the 100 as fewer tradig sigals are geerated. Thus, our coclusio regardig the effectiveess of the OBV tradig rule i Greater Chia still holds i the presece of tradig costs. To examie the tred of returs, Figures 1 to 3 plot the returs of the OBV over time. Returs for the ad Europea idices are plotted i Figure 1. Returs for the three mailad Chiese idices are plotted i Figure 2, while Figure 3 plots the cases for Taiwa ad Hog Kog. The dotted lies i the figures idicate the tred of returs i the correspodig stock markets. Note from Figure 1 that the retur of the OBV is geerally lower ad has gradually dimiished i developed markets. If the effectiveess of a tradig rule is a proxy for market efficiecy, the result implies that these markets have become icreasigly efficiet over time. Figure 2 shows that the OBV is icreasigly profitable i the Chiese market. The tred lie is fairly horizotal for Shaghai A ad eve slightly upward slopig for Shaghai B ad Shezhe A. This seemigly couter-ituitive observatio ca be accouted for by the fact that Chia has experieced a period of extremely rapid growth. The ecoomic boom traslates ito a soarig stock market i 2005-2007, pushig the rate of retur of the OBV to a level of 200%, which cotributes to the risig tred of the OBV returs. 8 From Figure 3, we ote that the treds of returs for the Hog Kog ad Taiwa markets are oly mildly dowward slopig over the last two decades compared to the ad Europe. Thus, the boom of the Chiese ecoomy has also fueled the retur of the OBV tradig rule for markets i the regio. 8 The clear spike of OBV retur i 2006-2008 admittedly cotributes to our coclusio that techical tradig rules are icreasigly profitable i Chia. However, sice the Chiese stock market has a relatively short history, we believe it is ot appropriate to further split the sample. 6
Figure 1: Returs of CAC, FTSE100 ad Dow Joes Figure 2: Returs of the Chiese Idices Figure 3: Returs of Hag Seg ad TWSE 7
Refereces 1. Blume, L., Easley, D. ad M. O'Hara (1994) Market Statistics ad Techical Aalysis: The Role of Volume Joural of Fiace 49, 153-181. 2. Brock, W., J. Lakoishok, ad B. LeBaro (1992) Simple Techical Tradig Rules ad the Stochastic Properties of Stock Returs Joural of Fiace 5, 1731-1764. 3. Chog, T. T. L. ad H. Ip (2009) Do Mometum-Based Strategies Work i Emergig Currecy Markets? Pacific-Basi Fiace Joural 17, 479-493. 4. E. F. Fama (1970) Efficiet Capital Markets: A Review of Theory ad Empirical Work, Joural of Fiace 25, 383-417. 5. E. F. Fama (1991) Efficiet Capital Markets: II Joural of Fiace 46, 1575-1617. 6. Ferguso, R. ad J.L. Treyor (1985) I Defese of Techical Aalysis Joural of Fiace 40, 757-773. 7. Hudso, R., M. Dempsey ad K. Keasey (1996) A Note o the Weak Form Efficiecy of Capital Markets: The Applicatio of Simple Techical Tradig Rules to UK Stock Prices 1935 to 1994 Joural of Bakig ad Fiace 20, 1121-1132. 8. Joseph E. Graville (1976), Graville s New Strategy of Daily Stock Market Timig for Maximum Profit, Pretice-Hall, Ic.., ISBN 0-13-363432-9. 9. Marshall, B., Youg, M. ad R. Caha (2008) Are Cadlestick Techical Tradig Strategies Profitable i the Japaese Equity Market? Review of Quatitative Fiace ad Accoutig 31, 191-207. 10. Mills, T. C. (1997) Techical Aalysis ad the Lodo Stock Exchage: Testig Tradig Rules usig the FT30 Iteratioal Joural of Fiacial Ecoomics 2, 319-331. 11. Shik, T. ad T. T. L. Chog (2007) A Compariso of MA ad RSI Returs with Exchage Rate Itervetio Applied Ecoomics Letters 14, 371-383. 8