STOCK PRICES TEHNICAL ANALYSIS

Transcription

1 STOCK PRICES TEHNICAL ANALYSIS Josp Arnerć, Elza Jurun, Snježana Pvac Unversy of Spl, Faculy of Economcs Mace hrvaske 3 2 Spl, Croaa jarnerc@efs.hr, elza@efs.hr, spvac@efs.hr Absrac Ths paper esablshes echncal analyss of sock prces based on average radng prces. Analyss procedure begns wh defnng average prces on daly bass whch are nvolved n sock marke nvesmen decsons for he frs me n fnancal heory as well as n pracce. Namely, all heorecal saemens are confrmed by movemens of Podravka socks, as componen of CROBEX ndex on Zagreb Sock Exchange. Usng exponenal smoohng mehodology dfference beween shor-erm and long-erm nvesmen sraegy s defned accordng o bull and bear sgnals. Keywords: average radng prces, echncal analyss, exponenal weghed movng average mehod, rollng sandard devaon, Bollnger's range, bull and bear sgnals. INTRODUCTION The approaches used o analyze socks and make nvesmen decsons are dvded no wo caegores: fundamenal analyss and echncal analyss. Fundamenal analyss nvolves analyzng he characerscs of a company n order o esmae s "value". Techncal analyss akes a compleely dfferen approach; doesn' care abou he "value" of a company. Techncans, somemes called charss, are only neresed n he prce movemens n he marke. Techncal analyss sudes supply and demand n a marke n an aemp o deermne wha drecon, or rend, wll connue n he fuure. Techncal analyss s a mehod of evaluang socks by analyzng he sascs of he pas prces movemens and volume. Therefore, uses chars and oher ools such as ndcaors and oscllaors o denfy paerns ha can sugges fuure movemens. Techncal analyss relays on hree basc assumpons: a any gven me, a sock's prce reflecs everyhng ha has or could affec he company - ncludng fundamenal facors; he repeve naure of prce movemens s arbued o marke psychology,.e. marke parcpans end o provde a conssen reacon o smlar marke suaons. I means ha hsory ends o repea self; prce movemens are beleved o follow rends. 2. PRICE MOVEMENTS IN TRENDS In echncal analyss has been shown ha afer a rend of prce movemens has been esablshed, he fuure prce movemen s more lkely o be n he same drecon. Mos echncal radng sraeges are based on hs assumpon. Emprcal researches dscover wo ypes of rend dsngushng accordng o: me srucure and general drecon. Accordng o me srucure here are long-erm rends, nermedae rends or shor-erm rends. These are conneced wh nvesmen sraeges. Namely, here s a sgnfcan dfference beween an nvesor and a rader. I means ha an nvesor expecs prof only n long-erm perod, whle raders prefer o prof n shor-erm perod. So, can be defned ha

2 long-erm nvesmen sraegy s assocaed whn me frame of 5 radng days; nermedae sraegy s adaped wh 2 radng days, whle shor-erm nvesmen sraegy s assocaed whn radng days. Furhermore, hese ranges are suggesed by John Bollnger. Accordng o general drecon prces could rend up, rend down, or rend sdeways. In fnancal leraure synonym for uprend marke s bull marke. A bull marke ends o be assocaed wh ncreasng nvesor confdence, movang nvesors o buy n ancpaon of furher capal earnngs. Techncal erm for downrend marke s bear marke. A bear marke ends o be accompaned by wdespread pessmsm. Invesors ancpang furher losses are movaed o sell. 3. SMOOTHING TECHNIQUES AND ROLLING ESTIMATES Smoohng echnques are very ofen used n fnancal leraure n general. In hs cones wll be used prmarly as an ndcaor of "bullsh" or "bearsh" sgns n he sock marke. Precsely speakng, n hs paper smoohng echnques wll be used o declne sochasc varaons. The smples mehod of smoohng me seres s smple movng average (SMA) mehod, whch s gven by: SMA ( k) = ( P + P + P P k+ ), () k where k s number of prevous perod for whch prces are observed. In oher words, forecas value for a one perod ahead s he smple average of curren sock prce and prevous ( k ) prces. Choce of perod k depends on he parcular purpose of he research. Namely, has o be noced ha he larger he choce of k, he smooher he seres wll be. The man dsadvanage of SMA s all observaons are equally weghed. To mgae he effecs of exreme observaons on movng average esmaes can be weghed dfferenly. Therefore, a common procedure ha pus more wegh on he mos recen observaons s based on exponenally declnng weghs and he resulng weghed movng average s called exponenal weghed movng average (EWMA). Accordng o exponenal smoohng mehod forecas values could be calculaed recursvely: Pˆ = ( λ) P + λ Pˆ (2) where Pˆ s presen perod forecas and Pˆ s prevous perod forecas. By connuous subsuon equaon n (2) becomes: 2 Pˆ = ( λ) P + λ ( λ) P + λ ( λ) P (3) + λ( λ) P +... =,, 2,...,k In equaon (3) parameer lambda < λ< s called smoohng consan. When k converges o nfny, relaon (3) can be noced as: ( λ) = ( λ) Pˆ λp. (4) From relaon (4) follows ha w when λ, accordng o: = k λ Pˆ ( k) = w P, w =. (5) k = λ = 2

3 From equaons (2), (3) and (4) follows ha he closer lambda s o uny he more wegh s pu on he prevous perod's esmae relave o he curren perod's observaon. Therefore weghs are decreasng n exponenal manner. However, weghs can decrease slowly or faser. Classcal approach of forecasng me seres suggess lambda wh mnmal sum of squared errors. Because he neres of hs paper s o descrbe paern of a me seres, hs suggeson wll no be srongly consdered. Namely, for he smoohng purpose, parameer lambda can be esmaed accordng o chosen perod k. Ths s n a close connecon wh me srucures of nvesmen sraeges. Tha's why parameer lambda can be esmaed n follow way: λ ˆ k =. (6) k + Accordng o formula (6) lambda s real number from nerval,. The larges s perod k he closer lambda s o uny. Therefore, weghs are decreasng slowly and me seres s smooher. 4. BOLLINGER BANDS Bollnger Bands were creaed by John Bollnger. Bollnger Bands are ploed a sandard devaon levels above and below a movng average (EWMA), where he sandard devaon s a measure of volaly. In hs paper he rollng sandard devaon of relevan prces s used: k SD = ˆ σ, = k,..., n, (7) k where k s number of perods whn he rollng sandard devaons are compued. In k equaon (7) facor ensures unbased esmaon. k Durng perods of exreme prce changes (.e., hgh volaly), he bands ndcae o dvergence. Durng perods of sagnan prcng (.e., low volaly), he bands narrow o conan prces. The longer prces reman whn he narrow bands he more lkely a prce breakou. They are one of he mos powerful conceps avalable o he echncally based nvesor, bu hey do no, as s commonly beleved, gve absolue buy and sell sgnals. Wha hey do s answer he queson of wheher prces are hgh or low on a relave bass. Usng hs nformaon, an nvesor can make buy and sell decsons, confrmng prce acon. Closng prces are mos ofen used o compue Bollnger Bands, whle oher varaons can also be used. For example he ypcal prce (TP) s: hgh + low + close TP =. (8) 3 The weghed prce (WP) s defned: hgh + low + 2 close WP =. (9) 4 In hs paper for he frs me we sugges so called he real weghed prce (RWP): 3

4 m pq = urnover RWP = =,, () m q volume = where m s number of ransacons n curren radng day, whle p s execuve prce of -h ransacon, and q s radng quany. Accordng o (8) and (9) equaons he advanage of he real weghed prce s n he fac ha execuve prces are weghed by radng quany,.e. execuve prce whch s raded more has greaer wegh and vce versa. Bollnger recommends usng a 2-day smple movng average for he cenre band and 2 sandard devaons for he ouer bands. The lengh of he movng average and number of devaons can be adjused. In hs paper he lenghs of he movng average are and 5 respecvely and numbers of devaons are,5 and 2,5 accordng o comparson of he shorerm and long-erm nvesmen sraeges. I can be concluded ha Bollnger Bands serve wo prmary funcons: o denfy perods of hgh and low volaly, and o denfy perods when prces are a exreme levels. Even so, a secury can become overbough or oversold for an exended perod of me. Knowng wheher or no prces are hgh or low on a relave bass can enhance he nerpreaon of oher ndcaors and oscllaors. Therefore, he relave srengh ndex (RSI) s used. The RSI s a echncal analyss oscllaor showng prce srengh by comparng upward and downward movemens. RSI =, () U + D where U s an absolue value of he movng average of upward execuve prce change, and D s an absolue value of he movng average of downward execuve prce change. The RSI s oscllaor ha ranges beween and. In suaon when RSI reach he 7% a secury s consdered o be overbough, or oversold a he level of 3%. Generally, f he RSI rses above 3% s consdered bullsh for he underlyng sock. Conversely, f he RSI falls below 7%, s a bearsh sgnal. The cenrelne for RSI s 5%. Levels of 8% and 2% are also used. 5. TEHNICAL ANALYSIS IN CROATIA The complee procedure of presened echncal analyss s esablshed usng observaons of Podravka socks as he mos frequenly raded sock from CROBEX ndex a Zagreb Sock Exchange. Fgure shows by he one lne rollng sandard devaon movemens from he pon of nvesor's vew (long-erm radng perods) and he oher lne are he movemens from he pon of rader's vew (shor-erm radng perods). Represenave for long erm perods s rollng sandard devaon for me frame of 5 radng days and represenave for shor erm perods s rollng sandard devaon of radng days. Esmaed accordng o real weghed prces shor erm Bollnger Bands are presened a Fgure 2. 4

5 Fgure. Rollng sandard devaon esmaes for shor-erm and long-erm radng perods Rollng sandard devaon, k= Rollng sandard devaon, k= Source: Accordng o daa on Fgure 2. Shor-erm Bollnger Bands esmaes accordng o real weghed prces Real weghed prce Lower band Upper band Source: Accordng o daa on Long-erm Bollnger Bands esmaed accordng o real weghed prces are llusraed a Fgure 3. Fgure 3. Long-erm Boolnger Bands esmaes accordng o real weghed prces Real weghedprce Lower band Upper band Source: Accordng o daa on Apar from varous ndcaors echncal analyss requres measuremens usng oscllaors. 5

6 An example of oscllaor relave srengh ndex s used n hs analyss. Fgure 4. Relave srengh ndex oscllaor Source: Accordng o daa on Fnancal heory and pracce recommended a smoohng perod of 4. Ths s by reckonng of EWMA smoohng.e. λ = / 4 or k = 27. Ths choce of k a Fgure 4. s managed by he fac ha s reasonable o chose nermedae me perod. 6. CONCLUSION REMARKS One of he ams of hs research s o dscover relable sgnals o buy or o sell on he sock marke usng echncal analyss. Usng exponenal smoohng mehodology dfference beween shor-erm and long-erm nvesmen sraegy s defned accordng o bull and bear sgnals. Accordng o concree daa abou Podravka socks a Zagreb Sock Exchange from March 7h 23 o Sepember 7h 26 (93 radng days) he fnal echncal analyss resuls n Table. gves precse suggesons o buy or o sell on he Sock marke from he pon of vew of shor-erm and long-erm sraegy. Table. Invesmen decsons accordng o performed echncal analyss Shor-erm sraegy Long-erm sraegy buy sgnal no sgnal sell sgnal Tradng days buy sgnal no sgnal sell sgnal Tradng days Source: Accordng o daa on A he end he fnal decson o buy or o sell depends on he concree nvesor (rader) preference o rsk more n order o earn more. Apar from echncal analyss resuls for such a decson addonal capal marke nformaon wll be used. REFERENCES. Colby, R.W., The Encyclopeda of Techncal Indcaors, McGraw-Hll, New York, Enders, W., Appled Economerc Tme Seres, Second Edon, Alabama: Wley, The Zagreb Sock Exchange: hp:// 6