FINANCIAL ANALYTICS USING R GROUP PROJECT EFFECT OF LOKSABHA ELECTIONS ON VOLATILITY OF INDIAN STOCK MARKET Submitted by 2016PGP054 2016PGP261 ANIMESH KUMAR KRUPAL PATEL 1|Page Contents PROBLEM STATEMENT .................................................................................................................................. 3 BRIEF LITERATURE REVIEW ........................................................................................................................... 4 INTRODUCTION ............................................................................................................................................. 5 HISTORICAL ESTIMATES OF VOLATILITY.................................................................................................... 5 Close to Close volatility ......................................................................................................................... 5 OHLC Volatility: Garman and Klass ....................................................................................................... 5 High-Low Volatility: Parkinson .............................................................................................................. 5 OHLC Volatility: Rogers and Satchell ..................................................................................................... 6 OHLC Volatility: Garman and Klass - Yang and Zhang ........................................................................... 6 OHLC Volatility: Yang and Zhang ........................................................................................................... 6 PREDICTIVE ESTIMATE OF VOLATILITY .................................................................................................. 7 GARCH MODEL ...................................................................................................................................... 7 INDIAN LOK SABHA ELECTIONS ................................................................................................................. 7 The Tenth Lok Sabha (1991-96): ........................................................................................................... 7 The Eleventh Lok Sabha (1996-98): ...................................................................................................... 7 The Twelfth Lok Sabha (1998-99): ........................................................................................................ 8 The Thirteenth Lok Sabha (1999-2004): ............................................................................................... 8 The Fourteenth Lok Sabha (2004-09): .................................................................................................. 8 The Fifteenth Lok Sabha (2009-14): ...................................................................................................... 8 The Sixteenth Lok Sabha (2014-19): ..................................................................................................... 9 METHODOLOGY ............................................................................................................................................ 9 RESULTS AND ANALYSIS .............................................................................................................................. 11 Results from Historical Volatility Estimates ............................................................................................ 11 GRAPHICAL ESTIMATES OF HISTORICAL VOLATILITY (2014) ................................................................... 12 RESULTS FROM GARCH MODEL .............................................................................................................. 14 GRAPHICAL ESTIMATES FROM GARCH MODEL ...................................................................................... 14 ANALYSIS ................................................................................................................................................. 14 BIBLIOGRAPHY ............................................................................................................................................ 15 2|Page PROBLEM STATEMENT Indian General Elections have great impact on the economy and business of the country. General election sets the direction for most of the economic activities for next five years by selecting the majority political party. Each party has different agendas and ideologies. So, people also have different expectations from different parties. Some parties are considered better for development of country than others based on past performance, expectations, leaders, their manifestos and many other factors. Investors also evaluate parties on different basis such as business friendly environment, foreign policies, fiscal policies, budget expenditures, focus on different sectors, expected GDP growth and any other financial and economic reforms. The event of election and its outcome creates political uncertainty for the country as well as financial markets which in turn expected to increase market volatility. The aim of our study is to identify the link between Indian general election and Indian equity market performance. The focus of the project is to find if there are abnormal returns (which can be linked to volatility based on various literatures) and abnormal volatilities around the election period compared to other periods without any major events. We will compare across various measures of volatility during event window with those measures during pre-event window and post-event window. Comparison across various measures will be done to see whether results are consistent throughout. 3|Page BRIEF LITERATURE REVIEW It can be empirically verified that fear in stock market is at peak during major macro-economic events like elections, announcement of policies etc. Although there hasn’t been significant research related to effect of Lok-Sabha elections, we have taken up this problem with inspiration from various similar literature studies. The summary of the above-mentioned literature reviews which studies effect of elections dates vis-à-vis stock market is provided below: Political uncertainty surrounding elections can affect how corporate investment responds to stock prices. In a large panel of elections around the world, investment is 40% less sensitive to stock prices during election years compared to non-election years. The decrease in investment-to-price sensitivity appears to be due to stock prices becoming less informative during election years making them noisier signals for managers to follow. Further, the drop-in investment-to-price sensitivity is larger when election results are less certain, in countries with higher corruption, large state ownership, and weak standards of disclosure by politicians. It is found that the country-specific component of index return variance can easily double during the week around an election, which shows that investors are surprised by the election outcome. Several factors, such as a narrow margin of victory, lack of compulsory voting laws, change in the political orientation of the government, or the failure to form a government with parliamentary majority significantly contribute to the magnitude of the election shock. Furthermore, some evidence is found that markets with short trading history exhibit stronger reaction. Our findings have important implications for the optimal strategies of institutional and individual investors who have direct or indirect exposure to volatility risk. The excess return in the stock market is higher under Democratic than Republican presidencies: 9 percent for the value-weighted and 16 percent for the equal-weighted portfolio. The difference comes from higher real stock returns and lower real interest rates, is statistically significant, and is robust in subsamples. The difference in returns is not explained by business-cycle variables related to expected returns, and is not concentrated around election dates. There is no difference in the riskiness of the stock market across presidencies that could justify a risk premium. The difference in returns through the political cycle is therefore a puzzle. An image of Greece Stock Market movement over the years highlighting election dates 4|Page INTRODUCTION HISTORICAL ESTIMATES OF VOLATILITY The implied volatility for a certain strike and expiry has a fixed value. There is, however, no single calculation for historical volatility. The number of historical days for the historical volatility calculation changes the calculation, in addition to the estimate of the drift (or average amount stocks are assumed to rise). There should, however, be no difference between the average daily or weekly historical volatility. We also examine different methods of historical volatility calculation, including close-to-close volatility and exponentially weighted volatility, in addition to advanced volatility measures such as Parkinson, Garman-Klass (including Yang-Zhang extension), Rogers and Satchell and Yang-Zhang. Close to Close volatility πππ = π πππ‘(( π ) ∗ ∑(ππ − πΜ )2 ) π−2 π€βπππ ππ = log( πππ πΜ = πΆπ ) πΆπ−1 ∑π−1 π=1 π π−1 OHLC Volatility: Garman and Klass The Garman and Klass estimator for estimating historical volatility assumes Brownian motion with zero drift and no opening jumps (i.e. the opening = close of the previous period). This estimator is 7.4 times more efficient than the close-to-close estimator. π π»π 2 πΆπ 2 (2πππ2 σ = √( ∗ ∑ [0.5 ∗ (πππ ) − − 1) (πππ ) ]) π πΏπ ππ High-Low Volatility: Parkinson The Parkinson formula for estimating the historical volatility of an underlying based on high and low prices. π π π»π 2 π=√ ∗ π πππ‘(∑ (πππ ) ) 4π ∗ πππ2 πΏπ π=1 5|Page OHLC Volatility: Rogers and Satchell The Roger and Satchell historical volatility estimator allows for non-zero drift, but assumed no opening jump. π΅ π―π π―π π³π π³π π = √( ∗ ∑ [πππ ∗ πππ + πππ ∗ πππ ] π πͺπ πΆπ πͺπ πΆπ OHLC Volatility: Garman and Klass - Yang and Zhang This estimator is a modified version of the Garman and Klass estimator that allows for opening gaps. π΅ πΆπ π π―π π πͺπ π ) + π. π ∗ (πππ ) − (π ∗ πππ(π) − π) ∗ (πππ ) ] π = √( ∗ ∑ [(πππ π πͺπ−π π³π πΆπ OHLC Volatility: Yang and Zhang The Yang and Zhang historical volatility estimator has minimum estimation error, and is independent of drift and opening gaps. It can be interpreted as a weighted average of the Rogers and Satchell estimator, the close-open volatility, and the open-close volatility. σ2 = σ2o + k ∗ σ2c + (1 − k)σ2rs where, σ2o 2 π ππ ) − µπ ) = ∗ ∑ (log ( π−1 πΆπ−1 1 ππ π πΆπ−1 µπ = ∗ ∑ log( π= σ2c = ) πΌ π+1 1+ π−1 2 π πΆπ ∗ ∑ (log ( ) − µπΆ ) π−1 ππ 1 πΆ π ππ µπΆ = ∗ ∑ log( π ) 2 πππ =Rogers and Satchell Volatility Estimate 6|Page PREDICTIVE ESTIMATE OF VOLATILITY GARCH MODEL The Generalized Autoregressive Conditional Heteroskedasticity model is another popular model for estimating stochastic volatility. It assumes that the randomness of the variance process varies with the variance, as opposed to the square root of the variance as in the Heston model. The standard GARCH (1,1) model has the following form for the variance differential. 2 2 ππ‘2 = πΌ0 + πΌ1 ∗ ππ‘−1 + π½1 ∗ ππ‘−1 π€βπππ, πΌ0 > 0, πΌ1 > 0, π½1 > 0, π½1 + πΌ1 < 1 The GARCH model has been extended via numerous variants, including the NGARCH, TGARCH, IGARCH, LGARCH, EGARCH, GJR-GARCH, etc. Strictly, however, the conditional volatilities from GARCH models are not stochastic since at time t the volatility is completely pre-determined (deterministic) given previous values. INDIAN LOK SABHA ELECTIONS The Indian parliament follows a bicameral system. It has two houses, namely the Rajya Sabha (Upper House) & the Lok Sabha (Lower House). The party (or a coalition) that gets a majority in the Lok Sabha gets to form the central government. The term of office is for a maximum period of 5 years or until such time the party (or a coalition) enjoys a majority in the Lok Sabha, whichever is earlier. Here is a look at the history of Indian Elections (Lok Sabha) since independence. The data is sourced from the statistical reports of the Election Commission of India. The Tenth Lok Sabha (1991-96): Rajiv Gandhi was assassinated in the run upto the 1991 general elections by the LTTE. These elections were also termed as the ‘Mandal-Mandir’ elections after the two most important poll issues; the Mandal Commission fallout and the Ram Janmabhoomi-Babri Masjid issue. While the Mandal Commission report implemented by the VP Singh government gave 27 per cent reservation to the Other Backward Castes (OBCs) in government jobs, the Mandir issue referred to the debate over the disputed Babri Masjid structure at Ajodhya, which the Bhartiya Janata Party was using as its major electoral issue. The Mandir issue led to numerous riots in many parts of the country and the electorate was polarized on caste and religious lines. No party could get a majority. Congress emerged as the single largest party with 232 seats while the BJP won 120 seats out of 521 seats. P V Narasimha Rao headed a minority government and was the first person from South India to occupy the Prime Minister’s chair. He is credited with ushering in economic reforms and also identifying Manmohan Singh who went onto become the Prime Minister. Dates: May–June 1991 The Eleventh Lok Sabha (1996-98): The Indian National Congress came into the election on the back of several government scandals and accusations of mishandling. There were various factions within the congress. The BJP grew from strength 7|Page to strength and emerged as the single largest party in a hung house. The BJP won 161 seats, Congress 140 and the Janata Dal 46. The rise of regional parties started with this election. The regional parties won 129 seats. Prominent among them were TDP, Shiv-Sena & the DMK. As per the prevailing custom, the President invited BJP to form the government. The BJP attempted to build a coalition, but could not go far and Atal Bihari Vajpayee had to resign as the PM in 13 days. His resignation address in the Lok Sabha is well known. The Congress Party declined to form the government but chose to extend outside support to Janata Dal and other smaller parties that formed into the ‘United Front’. Out of nowhere, H D Devegowda became the Prime Minister and he lasted for 18 months before he had to step down and make way for I K Gujral. He also could not last long following differences within the Janata Dal. Dates: April-May 1996 The Twelfth Lok Sabha (1998-99): The BJP emerged as the single largest party with 182 seats out of 543. Congress won 101 and the other regional parties won 101 seats. The BJP formed the National Democratic Alliance (NDA) with other regional parties. Atal Bihari Vajpayee was sworn in as the Prime Minister for the second time. His government could not last long and he had to resign after 13 months in office after the AIADMK withdrew support. The NDA lost by just one vote when Dr. Giridhar Gamang, the then Chief Minister of Odisha and also a MP, voted against the NDA. The nuclear tests at Pokhran, The Kargil war were some of the important incidents in this term. Dates: February–March 1998 The Thirteenth Lok Sabha (1999-2004): These elections were held in the backdrop of the Kargil war. The BJP again emerged as the single largest party with 182 seats while the congress could win only 114. This time the regional parties won 158 seats. The BJP was able to form a more stable NDA this time around and this was the first time that a noncongress alliance lasted a full five-year term. Atal Bihari Vajpayee was sworn in as the Prime Minister for the third time. Dates: September–October 1999 The Fourteenth Lok Sabha (2004-09): The BJP went in for early elections alongside launching an ‘India Shining’ campaign. Though it could win the middle-class vote, the poorer sections voted for the Congress and other regional parties resulting in the defeat of the NDA. The BJP could win only 138 seats while the Congress improved its tally to 145. The regional parties again ruled the roost with 159 seats. The BJP conceded defeat and the Congress then formed the United Progressive Alliance (UPA) with support from other parties and outside support from the left parties. Sonia Gandhi refused to become the Prime Minister amidst the controversy about her foreign origin. Manmohan Singh was chosen as the Prime Minister. Dates: Between 20 April and 10 May 2004 The Fifteenth Lok Sabha (2009-14): The Congress led UPA implemented a lot of its promises including the enactment of Right to Information (RTI) & the National Rural Employment Guarantee Scheme (NREGS). It also waived off farm loans in 2008. 8|Page Against this background, it went into the polls in 2009. The NDA on the other hand was led by L K Advani. The Congress won 206 seats, a huge improvement from 2004. The BJP could win only 116. The regional parties won 146 seats. The UPA came to power for the second term in a row. Dr. Manmohan Singh was sworn in as the Prime Minister for the second time. Dates: Between 16 April 2009 and 13 May 2009 The Sixteenth Lok Sabha (2014-19): The second term of the UPA proved to be a disaster with numerous allegations of corruption & scams. 2G, Coal Block, Adarsh, Commonwealth Games to name a few. The silence of the Prime Minister and the perception that he had no real power made matters worse. The BJP was successfully able to project Narendra Modi as the man of the hour and also as its Prime Ministerial candidate. Rahul Gandhi could not match Narendra Modi. The BJP won majority on its own with 282 seats while the Congress recorded its worst ever performance with just 44 seats. This was the first time since 1984 that a party won a majority on its own. Dates: 7 Apr 2014 – 12 May 2014 METHODOLOGY • • • • • • • • Data was collected for BSE SENSEX 30 from BSE Website Data for Election dates is collected from Election Commission of India Website For each election, the dataset has been divided in three parts: Pre-Election Window, Election Window, Post-Election Window. The window for Pre-Election and Post-Election period was given 4 months each. Election window is the window which is of prime interest to study volatility Election starts on the onset of first phase of voting and ends after 2 weeks of results. Period of 2 weeks after result is included in election window because market takes a bit time to absorb the results of election and so we have given 2 weeks’ time to get the results absorbed. Different types of volatility measures have been used to measure volatility. Volatilities have been calculated using these measures for all three windows for 10th to 16th Lok Sabha elections. 9|Page Figure: Event Study Time Horizon Pre Election Window Election & Result Window Post Election Window Loksabha Election Start End Start Result End Start End 16 05-12-2013 04-04-2014 07-04-2014 13-05-2014 27-05-2014 28-05-2014 25-09-2014 15 16-12-2008 15-04-2009 16-04-2009 14-05-2009 28-05-2009 29-05-2009 25-09-2009 14 22-12-2003 19-04-2004 20-04-2004 11-05-2004 25-05-2004 26-05-2004 23-09-2004 13 06-05-1999 03-09-1999 06-09-1999 04-10-1999 18-10-1999 20-10-1999 17-02-2000 12 16-10-1997 13-02-1998 16-02-1998 01-03-1998 16-03-1998 17-03-1998 15-07-1998 11 27-01-1996 26-05-1996 27-05-1996 08-05-1996 22-05-1996 23-05-1996 20-09-1996 10 23-01-1991 17-05-1991 20-05-1991 16-06-1991 01-07-1991 02-07-1991 30-10-1991 9 25-07-1989 21-11-1989 22-11-1989 27-11-1989 12-12-1989 13-12-1989 12-04-1990 Table: Timelines of Pre-Election, Election and Post-Election Window for different Lok Sabha Elections 10 | P a g e RESULTS AND ANALYSIS Results from Historical Volatility Estimates $`Year 1991` (10th Lok Sabha Elections) yangzhang garman-yang election 0.2149117 0.2175837 pre-election 0.3004172 0.2997493 post-election 0.310672 0.2995287 th $`Year 1996` (11 Lok Sabha Elections) yangzhang garman-yang election 0.2035733 0.1883376 pre-election 0.1523228 0.1523505 post-election 0.1710462 0.1676076 th $`Year 1998` (12 Lok Sabha Elections) yangzhang garman-yang election 0.226833 0.2223152 pre-election 0.2175278 0.2151544 post-election 0.2920761 0.2844138 $`Year 1999` (13th Lok Sabha Elections) yangzhang garman-yang election 0.2541177 0.2599868 pre-election 0.2343988 0.243196 post-election 0.2840161 0.3038643 $`Year 2004` (14th Lok Sabha Elections) yangzhang garman-yang election 0.4049049 0.4003949 pre-election 0.2172794 0.2243678 post-election 0.162284 0.1651373 th $`Year 2009` (15 Lok Sabha Elections) yangzhang garman-yang election 0.4213615 0.4262412 pre-election 0.3114342 0.3124799 post-election 0.2326564 0.2360176 $`Year 2014` (16th Lok Sabha Elections) yangzhang garman-yang election 0.1632295 0.1680688 pre-election 0.1064092 0.1088041 post-election 0.111206 0.1155711 roger close garman Parkinson 0.1180532 0.2921654 0.2175837 0.1486945 0.1441235 0.3505269 0.2997493 0.168415 0.1558925 0.3563674 0.2995287 0.1644011 roger close garman Parkinson 0.1869059 0.2543973 0.1883376 0.1864029 0.1082619 0.2580587 0.1523505 0.1425396 0.1607409 0.2466402 0.1676076 0.177462 roger close garman Parkinson 0.1780277 0.3349548 0.2223152 0.194184 0.1902532 0.2630838 0.2151544 0.1902907 0.2514464 0.3680279 0.2844138 0.2550429 roger close garman Parkinson 0.1946972 0.238036 0.2599868 0.2101986 0.2062162 0.2743278 0.243196 0.2064538 0.2179761 0.3104698 0.3038643 0.2334431 roger close garman Parkinson 0.3598101 0.5507825 0.4003949 0.3820803 0.2052451 0.2694064 0.2243678 0.2199355 0.1509946 0.1895955 0.1651373 0.1594917 roger close garman Parkinson 0.2799461 0.5884254 0.4262412 0.3278535 0.2488334 0.4275505 0.3124799 0.2783891 0.2136825 0.2966484 0.2360176 0.2297163 roger close garman Parkinson 0.15599345 0.156249 0.1680688 0.15245293 0.09219181 0.1229157 0.1088041 0.09547076 0.10072714 0.1432441 0.1155711 0.10507912 11 | P a g e GRAPHICAL ESTIMATES OF HISTORICAL VOLATILITY (2014) For the sake of verification, we have posted a comparative volatility plots across various estimates for the year 2014. We can observe that the pattern is more or less the same. Estimates of other years have been shared in the link given along with the mail or can be found by running the source code provided. 12 | P a g e 13 | P a g e RESULTS FROM GARCH MODEL Mean of GARCH Volatilities Pre-Election Election Post-Election 2014 2009 2004 1999 1998 1996 1991 0.007887172 0.02443804 0.01566484 0.01714497 0.01710982 0.01454396 0.01920073 0.007880947 0.02925935 0.02490636 0.01727825 0.01878532 0.01450614 0.01862979 0.007877975 0.02262932 0.01357216 0.01739763 0.01953677 0.01446881 0.01838545 GRAPHICAL ESTIMATES FROM GARCH MODEL ANALYSIS • • From the above results, it can be inferred that volatility during election period is very high in all the election years taken into consideration except 1998 and 1999 elections. The anomaly in election year 1998 can be attributed to unstable government in the previous regime (1996 to 98). Some mishaps in the previous government are as follows: 14 | P a g e o • • • • Shri Atal Bihari Vajpayee had resigned in 13 days after he was sworn in as Prime Minister. This led to a short-lived government of 2 year (1996-98) o The government could not stand long because of lack of confidence in the house following which Prime Minister H.D. Devegowda and I.K. Gujral’s reign could not last long. o This led to untimely elections in 1998 also can be the reason attributed to inconsistent behaviour in the stock market during the same period. The anomaly in election year 1999 can be attributed to short-lived and unstable government of 1998-99 in other words it is nothing but repercussions of the previous unstable parliament. o The nuclear tests at Pokhran, The Kargil war were some of the important incidents in this term. Also, it can be seen that volatility is higher in post-election period than pre-election period as government is highly engaged in laying down various schemes in the house of parliament as promised in the pre-election period. It can be seen from the graphical estimates that each estimate follows more or less same pattern From GARCH model fitting it can be seen that it has predicted well in 2009, 2004, 1998, 1999 and not in other cases. This has been done to cross validate our analysis Hence it can be inferred that the stock market anticipation arises as a result of performance of the previous government. This creates expectation among people for the upcoming government and hence there is anticipation and fear in the stock the market which gives rise to volatility. Hence volatility is otherwise termed as fear factor. BIBLIOGRAPHY https://www.rdocumentation.org/packages/TTR/versions/0.23-2/topics/volatility http://eci.nic.in/eci_main1/ElectionStatistics.aspx https://www.newslaundry.com/2015/09/14/a-brief-history-of-the-lok-sabha-elections BiaΕkowski, Gottschalk, Wisniewski; “Stock market volatility around national elections” http://www.sciencedirect.com/science/article/pii/S0378426607004219 Siokis, Kapopoulos; “Parties, elections and stock market volatility: evidence from a small open economy” http://onlinelibrary.wiley.com/doi/10.1111/j.1468-0343.2007.00305.x/full Clara, Valkonov; “The Presidential Puzzle: Political Cycles and the Stock Market” http://onlinelibrary.wiley.com/doi/10.1111/1540-6261.00590/full https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1549714 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1827462 15 | P a g e