Advanes in Eonomis, Business and Management Researh (AEBMR), volume 26 International Conferene on Eonomis, Finane and Statistis (ICEFS 2017) The Impat of Personal and Institutional Investor Sentiment on Stok Returns under the Chinese Stok Market Crash Kexuan Wang Shanghai University, China *wang1ke2xuan@126.om Abstrat. Under Chinese stok market rash, personal and institutional investor sentiment is out of ontrol, whih leads to the stok pries fall further. Therefore this artile uses the onept of behavioural finane to study the problem. This essay uses multiple regressions to study the influene of the urrent and lagged emotions of individuals and institutions on the stok returns. The onlusion is helpful to solve the stok market rash and improve the eonomy in China. Keywords: stok market rash, individual investor sentiment, institutional investor sentiment, multi-fators priing model 1 Introdution The stok market rash happened in 2015 made the investors suffer a lot. Stok pries dereased sharply and thousands of stoks fell in lok step. Individual and institutional investor sentiment was out of ontrol, and they sold the shares, whih leaded to the stok pries fall further. However, the measures taken by government had no effet. Thus, fousing on stok market rash ourred in 2008, this essay studies the impat of personal and institutional investor sentiment on stok yields. Foreign sholars began to explore this area in the 1990s, and mainly foused on the existene of internal mehanism between investor sentiment and stok yields. They reated different theoretial models to explain the internal mehanism, inluding DSSW model [1]. BSV model [2], DHS model [3] and HS model [4]. Both Chinese and foreign sholars onduted empirial exploration based on the theoretial mehanisms. Wang et al. (2006) examined the effetive use of investor sentiment to predit stok returns volatility with GARCH model [5]. Wang and Liu (2011) paid lose attention to the Bull and Bear, and found that at the bear market, investor sentiment affeted stok returns more deeply than that at bull market [6]. When seleting sentiment proxies, there are mainly two ways. Foreign sholars usually uses diret way to express personal and institutional investor sentiment [7]. By ontrast, domesti sholars both use diret and indiret ways [8, 9]. 2 Experimental 2.1 Theoretial Derivation Ross (1976) believed that fators that affet apital asset priing are not only the market portfolio return, but also a variety of fators. After, Ross proposed arbitrage priing model [10]. expeted return β β β (1) Chinese stok market is an imperfet market, so the fluidity of information is very slow, whih means investors annot always remain rational as well as hoose effiient portfolio. Thene, the fators that affet investor sentiment should be added into asset priing model. Moreover, it is Copyright 2017, the Authors. Published by Atlantis Press. This is an open aess artile under the CC BY-NC liense (http://reativeommons.org/lienses/by-n/4.0/). 162
Advanes in Eonomis, Business and Management Researh (AEBMR), volume 26 neessary to divide sentiment into 2 parts. Furthermore, aademis believe that in China, the government prefer to arry out human intervention in the stok market whih leads to the volatility of stok market prie. Therefore, maro variables are added into equation. R (2) 2.2 Proposing Hypotheses Fousing on and R, the study wants to onern the impat of individual and institutional investor sentiment on stok market returns before and after the rash. Meanwhile, the paper plans to divide time dimension into urrent and lag time. Then this artile puts forward 5 hypotheses. Assumption 1: Before rash, individual and institutional investor sentiment has a positive impat on the urrent market returns. After rash, the influene will be weakened. Assumption 2: In the urrent relationship, the degree of the individual and institutional investors sentiment affet the overall rate of return is asymmetri. Assumption 3: lag individual and institutional investor sentiment negatively affet the stok returns and after stok market rash, the ability to predit market yields will be weakened. Assumption4: the asymmetry still exists in the lag relationship. Assumption 5: the maroeonomi variables are positively orrelated with stok market returns insignifiantly under every situation. 2.3 Variables Seletion For the seletion of investor sentiment, this essay hooses to use new aounts of individual and institutional investors beause this data is more aurate and reliable. Figure 1 Seletion & soure of variables variables symbols soures Stok market returns Shr GTA & Hong Kong University database Individual investor sentiment sent China seurities learing firm Institutional investor sentiment sent China seurities learing firm Growth rate of industrial added-value Indgdp GTA Money supply M1-M2 GTA rate GTA Maroeonomi prosperity index State Statistis Bureau 2.4 Experimental Design Based on the five assumptions, the paper made an experimental design. To test the assumption 1, 2&5, this paper uses OLS regression to link Individual and institutional investor sentiment and maroeonomi variables to the stok returns before and after stok rash. The lag stok returns are also onsidered here. Shr 1 2 1 ε (3) Shr 1 2 1 ε (4) To test the assumption 3, 4&5, the lag time is divided into short term, middle term and long term. Shr 1 2 1 ε (5) 163
Advanes in Eonomis, Business and Management Researh (AEBMR), volume 26 Shr 1 1 2 (6) ε N=1, 2 (short term); N=6 (middle term); N=10 (long term) 4 Results and Disussion The paper intends to use OLS model to explore the urrent relationship as shown before and the regression results are presented as follows. Figure 2 OLS results of the stok rash for urrent relations Before stok rash 2006.01~2007.12 After stok rash 2008.01~2013.12 Personal investor Institutional investor Personal investor Institutional investor sentiment 1 sentiment 2 sentiment 3 sentiment 4 0.083 0.008 0.027 0.003 2.485** 0.118 1.368 0.179 0.043 1.714-0.529-0.360 2.974 4.814-0.257 36.57 3.498 11.07-0.152 12.16 0.648 1.152 2.246-0.201 16.93 0.086 0.047 2.45 1.32-0.17-1.38-1.68* 0.722 1.25 1.223 0.958 2.02** 0.598-0.505 0.893-0.011-1.06 1.135 0.806-0.995 1.029 1.91* 0.764 0.488 0.455 0.764 0.321-0.013-1.289 0.333-0.185 1.54-0.599 0.046 14.58 To test assumption 1, though observing (1) (2) (3) (4), the oeffiients of sentiment are positive, whih proves the first half of hypothesis 1 is orret. At the same time, ompared to (1) (2), the value of oeffiients in (3) & (4) are dereased, whih onfirmed the seond half of hypothesis 1. About testing assumption 2, the t value and of personal investor sentiment are higher than those of institutional sentiment. Thus, assumption 2 is right and personal investor sentiment is dominant. The oeffiients of maro variables are positive, whih is same as the results of assumption 5. While after the stok rash, growth rate of industrial added-value is signifiantly related to stok prie. As a result, assumption 5 is proved inompletely. Moreover, the paper studies the lagged relations to test assumption 3, 4 & 5. 164
Advanes in Eonomis, Business and Management Researh (AEBMR), volume 26 Figure 3 Effet of lagged personal investor sentiment on returns before the stok rash 9 10 11 12 0.035 0.024 0.875 (0.402) -0.077 (-1.16) (-1.44) 0.0186 0.052 0.37 (0.81) -0.0397-0.59 (-0.441) 0.0457 1.175 (1.97*) 1.02 1.19 d(indg 0.484 1.079 0.8999 dp) 0.505 (0.752) d(m1-1.8126 1.253 M2) 0.796 (0.548) Inflatio 5.3795 2.393 n 1.08 (0.471) 0.26 0.22 10.202 15.737 16.263 0.924 (1.374) 0.85 0.48-0.284-0.182-0.2468-0.137-0.865 (-0.60) (-0.54) -0.26 10.01 7.61 16.10 20.17 0.071 0.719 0.46 0.25 1.826 0.513 0.54 0.14 1.616 (16) -0.0397 0.063 14 15-0.038 0.036 C 13 1.6999 12.433 0.067 0.0415 0.074 (2.12*) (1.01) 1.14 d(indg 0.421 1.730 1.155 0.051 dp) (0.424) (1.14) (0.591) 0.02 d(m1-1.164 1.835 1.385 0.423 M2) (0.536) (0.83) (0.495) 0.11 Inflati 4.307 2.164 1.323 0.653 on (0.864) (0.44) (0.22) 0.08 10.28 16.12 12.892 10.90 (0.866) (1.44) (0.78) (0.422) Shr -0.154-0.227-0.144-0.122-1 (-0.515) (-0.8) (-0.35) (-0.229) 12.96 23.62 14.01 5.70 Form figure 3, the artile an onlude short-term personal investor sentiment is onneted to stok yields ambiguously. Middle-term personal investor sentiment negatively predits returns while long-term predits returns positively. The lagged institutional investor sentiment is not always negatively linked to stok returns. To onlude, assumption 3 is rejeted. Lagged institutional investor sentiment has a stronger apability to reflet stok returns before the stok market rash by omparing oeffiients and of personal and institutional investor sentiment. Assumption 4 is partial proved. Then, the essay study the effet of lagged personal and institutional investor sentiment on returns after the rash. The empirial results show that hypothesis 3 is wrong by observing the sign before the oeffiients of investor sentiment. Furthermore, and oeffiients value prove that after the stok rash, personal investor sentiment is more onvined to 165
Advanes in Eonomis, Business and Management Researh (AEBMR), volume 26 predit market yields than institutional investor sentiment. Assumption 4 is verified. Similarly, assumption 5 is found to be inorret. 5 Conlusions and Suggestions In onlusion, the urrent investor sentiment is positively related to stok returns, and after the rash, this positive orrelation weakens. Besides, investor sentiment and expeted stok returns have a different link in short, medium and long term both before and after the rash. Thirdly, the growth rate of industrial added value is linked to stok returns signifiantly after the stok rash. Then omparing two kinds of investor sentiment, individual and institutional investor sentiment responds to stok market return asymmetrially before and after the rash. Three suggestions are raised to manage sentiment to save the rash for Chinese government. The government should guide both personal and individual investors to invest reasonably. Furthermore, adjusting the ratio of personal and individual investors is also important. Lastly, government should take good use of maroeonomi poliies. With these suggestions, the government an solve the stok rash easily and the eonomy will therefore be improved. Referenes 1. J. B. Delong, A. Shleifer, L. Summers, etal. Noise trader risk in finanial markets, Journal of Positive Eonomy, vol. 98, pp.703 748, (1990). 2. N. Barberis, A. Shleifer, R. Vishny, A model of investor sentiment, Journal of Finanial Eonomis, vol.49, pp.307-343, (1998). 3. K. DanieL, D. Hirshleifer, A. Subrahmanyam, Investor psyhology and seurity market under- and over- reation, Journal of Finane, vol.53, pp.1839-1885, (1998). 4. H. Hong, J. C. Stein, A unified theory of under-reation, momentum trading, and overreation in asset markets, Journal of Finane, vol.54, pp.2143-2184, (1999). 5. Y. Wang, A. Keswani, S. J. Taylor, The relationship between sentiment, returns and volatility, International Journal of Foreasting. vol.22, pp.109-123, (2006). 6. Y. R. Wang, S.C. Liu, Investor sentiment and stok returns: bull and bear market & US omparison and ontrast, Journal of Beijing University of Aeronautis and Astronautis (Soial Sienes Edition). vol.24, pp.74-80, (2011). 7. R. Verma, H. Baklai, G. Soydemir, The impat of rational and irrational sentiments of individual and institutional investors on DJIA and S&P500 index returns, Applied Finanial Eonomis. vol.18, pp.1303-1317, (2008). 8. Q. Zhang, S. E. Yang, H. Yang, An empirial study of Chinese stok market investor sentiment and stok Returns, Systems Engineering, vol.25, pp.13-17, (2007). 166
Advanes in Eonomis, Business and Management Researh (AEBMR), volume 26 9. W. Q. Liu, X. X. Liu, Individual and institutional investor sentiment and stok returns - based on the researh of Shanghai A-share market, Journal of Management Sienes in China, vol. 17, pp.70-87, (2014). 10. S. A. Ross, The arbitrage theory of apital asset priing, Journal of Eonomis Theory, vol.13, pp.341-360, (1976). 167