 Research
 Open Access
 Published:
General election effect on the network topology of Pakistan’s stock market: networkbased study of a political event
Financial Innovation volume 6, Article number: 2 (2020)
Abstract
To examine the interdependency and evolution of Pakistan’s stock market, we consider the crosscorrelation coefficients of daily stock returns belonging to the blue chip Karachi stock exchange (KSE100) index. Using the minimum spanning tree networkbased method, we extend the financial network literature by examining the topological properties of the network and generating six minimum spanning tree networks around three general elections in Pakistan. Our results reveal a starlike structure after the general elections of 2018 and before those in 2008, and a treelike structure otherwise. We also highlight key nodes, the presence of different clusters, and compare the differences between the three elections. Additionally, the sectorial centrality measures reveal economic expansion in three industrial sectors—cement, oil and gas, and fertilizers. Moreover, a strong overall intermediary role of the fertilizer sector is observed. The results indicate a structural change in the stock market network due to general elections. Consequently, through this analysis, policy makers can focus on monitoring key nodes around general elections to estimate stock market stability, while local and international investors can form optimal diversification strategies.
Introduction
Political risk is a key factor affecting the performance of a country’s financial market. There are many ways political risk can take place, such as a change in legislation, general elections, or a political regime shift. According to (Pantzalis et al. 2000), positive stock returns are expected after the resolution of political uncertainty. Additionally, political events capture investor attention due to the possible revision of their investment strategies, depending on the outcome of the event (Pástor and Veronesi 2013). Previous research proposes various methods to study the impact of elections on stock markets [e.g., GARCH and Cumulative abnormal volatility (CAV) (Białkowski et al. 2008), Support vector regression (Chiu et al. 2012), Cumulative average abnormal returns (CAAR) and Abnormal returns AR (Oehler et al. 2013; Savita and Ramesh 2015), Regression analysis (Abidin et al. 2010; Liew and Rowland 2016) Granger cointegration and Johansen test (Goodell et al. 2015), and DCC bivariate GARCH (Sultonov and Jehan 2018)].
We extend the financial network literature by studying the impact of general elections on the network topology of Pakistan’s stock market. Networkbased methods are widely applied by researchers to study interdependency and the evolution of stock markets, such as minimum spanning tree (MST) (Mantegna 1999; Onnela et al. 2003b; Zhao et al. 2018; Yao and Memon 2019), threshold networks (CT) (Boginski et al. 2005; Lee and Nobi 2018), planar maximally filtered graphs (PMFG) (Tumminello et al. 2005; Yan et al. 2015; Musmeci et al. 2016), wavelet (Wang et al. 2017), and multiple criteria decision making (MCDM) (Kou et al. 2014). We have chosen MST, a main network mapping methodology extensively used to analyze various financial crisis (Wiliński et al. 2013; Majapa and Gossel 2016; Coletti and Murgia 2016; Xia et al. 2018; Memon and Yao 2019; Kou et al. 2019), currency crisis (Jang et al. 2011; Sultornsanee et al. 2013), sovereign debt crisis events (Dias 2012), as well as macroeconomic phases (Zhang et al. 2011). Therefore, this paper examines the interdependency and evolution of Pakistan’s stock market by using MST before and after the general elections of 2018, 2013, and 2008.
In the summary, our study has made two contributions. First, we fill a research gap by examining the stock market’s structural change and topological evolution around a political event: general elections. Unlike previous studies, that apply networkbased methods to examine crisis related events of stock markets (Onnela et al. 2003a; Yang et al. 2014; Radhakrishnan et al. 2016), assesses structural change of stock market around other political event (Yao and Memon 2019), and does not take into consideration structural change and topological evolution particularly across general elections. Second, we apply MST around the three general elections in an emerging stock market instead of a developed stock market. Thus, understanding the state and structures of stock market across general elections is important for local and institutional investors in forming diversified strategies and estimating stock market stability for policy makers.
The rest of this paper is organized as follows. In Section 2, we describe the data, the MST methodology, and network topology centrality measures. In Section 3, we present and discuss the results. Finally, Section 4 presents our concluding remarks.
Data and methodology
Data
We use crosscorrelations in the daily closing prices of 82 major stocks listed on the Pakistan stock exchange (PSX) for the blue chip companies’ index, which is also known as the Karachi stock exchange (KSE100) index, before and after three general elections in Pakistan to determine MST maps. The division of the data around the three general elections is shown in Table 1.
To study the role of every stock and the topological evolution, we construct six MSTs for before and after the general elections, where each subsample comprises 40 trading days. Additionally, companies are divided based on their economic activities for sectorial analysis. All 82 companies (acting as nodes in the network), individual sectors, and color for identification in the MST network are presented in Table 2. The data are obtained from the websites of (OpendoorsPK 2018), and (BusinessRecorder 2018).
Methodology
Let P_{i}(t) be the closing price of stock i at time t, while logarithm return r_{i}(t) of a stock is defined as:
We construct correlation matrix C = (c_{ij}) of size N × N among stocks i and j as follows:
where r_{i} and r_{j} are the logarithm returns of stocks i and j, respectively, and 〈…〉 is the statistical average for the period under investigation. Following the methodology of (Mantegna 1999), the correlation matrix of (82 × 82) is transformed into a distance matrix among each pair of stocks i and j, being given by:
The MST is a backbone network of correlation tree that uses N − 1 linkages from N(N − 1)/2 crosscorrelation coefficients by minimizing the sum of all edges. The 82 stocks on PSX are the nodes in the network joined by edges and characterized by weights W. We apply (Kruskal 1956) algorithm to an undirected graph (G) of form G = (N, E, W) into the MST.
To monitor the sectorial properties of Pakistan’s stock market network, we examine three centrality measures—degree, closeness, and betweenness centrality—in the periods before and after the general elections. The degree of centrality of stock i is given by:
where A_{ij} is the component in the ith row and jth column of adjacency matrix A. The higher the value of C_{D}(i) is, the more power stock i carries in influencing other stocks in the network (Marsden 2015).
Closeness centrality (Sabidussi 1966) is utilized to examine the amount of influence attached to an individual node in a network. This measure aims to examine the closeness of nodes with other nodes or sectors in the network and is defined as:
where d_{G}(k, h) represents the minimum distance between nodes k and h. Finally, to analyze the intermediary role of a node in the network, betweenness centrality is used to evaluate the frequency of a node acting as shortest link among two nodes and in transferring valuable information to the network:
where σ_{st} is the number of shortest paths between nodes s and t.
Results and discussion
Here, we present six MSTs of 82 stocks in KSE100 of Pakistan stock market before and after three general elections (2018, 2013, and 2008). The stocks (nodes) are colored based on their business sectors, as per Table 2.
Figure 1 shows the MST map before the general election of 2018. There is no single most dominant stock, but four principal nodes^{Footnote 1}: Nishat Mills (NML 6), D.G. Khan Cement (DGKC 5), Kohat Cement (KOHC 5), and Sui Northern Gas (SNGP 5). Therefore, MST highlights the strong dependency of Pakistan’s economy on cement, textile, and oil and gas sectors before the 2018 election. Additionally, the results show that the stocks in the cement (red) and oil and gas (yellow) sectors form clusters and connect with each other. However, companies in the commercial banks sector (blue) are split into two groups, one at the bottom left of the map and one on the middleright side. Furthermore, a strong intermediary role is being played by SNGP, which is carrying the highest betweenness centrality score of 2265.
The MST map after election 2018 of the KSE100 index companies is presented in Fig. 2. We observe the scattered pattern of the commercial banking sector, where banks are linked to companies from other sectors. Similarly, oil and gas companies are separated from one major cluster before elections 2018 to a disbursed distribution after elections period. The split of these two sectors possibly reflect different responses to the risks constituted by general elections. In terms of connectivity, a main hub node is Lucky Cement (LUCK 10), followed by three key nodes: D.G. Khan Cement (DGKC 7), ENGRO Corporation (ENGRO 5), and Kohat Cement (KOHC 5). Several emerging countries such as South Africa (Majapa and Gossel 2016), Brazil (Tabak et al. 2010), China (Zhuang et al. 2008), and India (Sinha and Pan 2007) have key nodes in the MST, which is not uncommon on Pakistan’s stock market. In sum, the MST shows a starlike structure after the general elections of 2018, possibly reflecting a structural transition.
Figure 3 represents the 2013 preelection MST map of the KSE100 network. The figure shows a treelike structure of the MST (Nobi et al. 2015) due to low value of centrality. The three key nodes of Mapple Leaf Cement (MLCF), Pak Suzuki Motors (PSMC), and Searle Company (SEARL) have a maximum of five connections each in the 2013 preelection period. In addition, there is no major cluster formation and reduced connectivity in the network. Only three cement sector companies—DGKC, MLCF, and PIOC—connect with each other, the rest linking with companies from different sectors. Further, companies from almost all the sectors are scattered, and around 77% of nodes have two or fewer connections.
The MST map of the KSE100 index after the general elections of 2013 is presented in Fig. 4. We can observe a slight increase in the degree of connectivity per node for Maple Leaf Cement (MLCF 6), while four nodes—Cherat Packaging (CPPL), Hub Power (HUBC), Kohinoor Textile Mills (KTML), and Nishat Mills (NML)—have five connections each. Additionally, Searle (SEARL) and Pak Suzuki (PSMC) nodes have now two and one connection, respectively. This reconfiguration possibly represents the diversification of these companies after the general elections. The MST in post the 2013 election has a treelike structure compared to the starlike structure in the post 2018 election period.
The pre2008 election MST map of the KSE100 index companies is presented in Fig. 5. A major node is Pakistan Petroleum (PPL 9), followed by two significant nodes—Nishat Mills (NML 7) and Muslim Commercial Bank (MCB 5). Hence, PPL is the dominating company in the network before the 2008 general elections, and companies surrounding this node will be directly influenced by any shock due to the general elections (Sharif et al. 2016). Additionally, the oil and gas sector (yellow) and commercial banking sector (blue) cluster, whereas cement sector nodes (FCCL, MLCF, CHCC, PIOC, ACPL, LUCK) are spread over the entire network.
Figure 6 shows the post 2008 election MST map of KSE100 index network. We observe a treelike structure and reduction in the degree of connections per node compared to the starlike structure before the election. D.G. Khan Cement (DGKC 7) is the most significant and central node, followed by Attock Refinery (ATRL 6) and Oil and Gas Development Company (OGDC 5). However, PPL has moved from nine connections before election to three. Additionally, companies from the cement sector form a cluster. Overall, the results show a structural change in the network topology from before to after the 2008 general elections.
Sectorial centrality measures
We here present the sectorial centrality properties of the KSE100 index MST network for all three general elections. As centrality measures are important in network analysis to examine the behavior of networks (Ghanbari et al. 2018), we present the average sectorial degree of centrality in Fig. 7. The figure shows the cement sector nodes have a mean of approximately three incident edges, followed closely by oil and gas and fertilizer sectors with means of two and half. We can observe a large increase in the average number of incident edges for the nodes in the cement sector after the 2018 general elections compared to any other sectors. The other sector that exhibits an increase in connectivity is the paper and board sector after the 2013 elections. This increase in the number of edges highlights an expansion in the economic activities of these sectors in the elections periods. Additionally, the results show a consistent decrease in the connectivity of commercial banks in all the three postelection periods, compared to the high average degree of connections during the preelections periods. The comparison between pre and postelections centrality metrics shows a consistent average connectivity in the textile sector during all periods, except for a substantial increase in the post 2013 election period.
Figure 8 presents the average sectorial closeness centrality for the periods before and after the elections. The cement sector has a large average closeness centrality of 0.26 after the 2018 election, followed by the oil and gas (0.22), pharmaceutical (0.22), technology and communications (0.21), and paper and board 0.21 sectors over the same period. This could be interpreted as the key role of these sectors in the MST after the 2018 election. Additionally, the cement sector is closer to other sectors’ nodes as the companies from this sector have the highest average closeness centrality. By contrast, the lowest average closeness centrality is registered by the paper and board sector (0.090 after the 2008 general elections. However, the overall average closeness centrality is 0.15 for all sectors during all periods.
Regarding average sectorial betweenness centrality, Fig. 9 shows that the fertilizer sector has the maximum number of shortest routes and has a strong mediating role, especially before the 2013 elections. Additionally, the average betweenness centralities of the oil and gas and cement sectors are also high. Similarly, an increase in the mediation effect of the cement sector is also registered after the 2018 election. Furthermore, the food and personal care products, tobacco, pharmacy, and automobile sectors present the lowest overall average betweenness centrality scores, reflecting their weak mediation effects.
Conclusions
We presented the topology of the network formed by the 82 main stocks on the Pakistan stock market using MST for the three general elections in 2018, 2013, and 2008. We observed the status of and changes in the MST network as a result of these political events. The MST had no dominant node, but had two major clusters for the cement and oil and gas sectors and a treelike structure before the general elections of 2018, compared to a starlike structure and a major node (LUCK) and broken clusters after the 2018 elections. Comparatively, for the 2013 elections, we found a decrease in the connectivity and scattered sectors before the elections and an improvement in connectivity due to diversification after the elections. A starlike structure of the MST is observed before the 2008 elections, with the dominating node being PPL compared to a significant drop in connections and in PPL’s influence along with a treelike structure after the 2008 general elections. In addition to the six MST maps, we also assessed our results by using degree centrality, sectorial closeness, and betweenness centrality. The average sectorial degree of centrality measure shows the economic expansion and highest number of incident edges for the cement sector nodes, followed by the oil and gas and fertilizer sectors. Similarly, the sectorial closeness and betweenness centrality measures revealed the importance of the cement sector nodes and strong overall intermediary role of the fertilizer sector.
Studying stock market topology around general elections has significant implications for local and international investors for portfolio management, such as forming optimal diversification strategies. Additionally, identification of core nodes and sectors, systematic and political risk assessment is useful for policy makers or regulators in assuring stock market stability. Our study examines stock market structural change and topological evolution around general elections through MST technique. Perhaps, we could compare networks formed through alternative methods of PMFG (Tumminello et al. 2005), and asset graphs (Onnela et al. 2003b, c). In addition, we can analyze the impact of various political and macroeconomic events on rolling windows of different sizes in a temporary dynamic stock market network. Further, we can apply MST method to analyze structural change and dynamics around political event of general elections into other stock markets (e.g., developed markets or other emerging markets). The future research may address these limitations.
Availability of data and materials
Data are available on http://www.opendoors.pk/ and https://markets.brecorder.com/.
Notes
The number against each stock symbol specifies the number of connections per node.
Abbreviations
 AR:

Abnormal returns
 CAAR:

Cumulative average abnormal returns
 KSE:

Karachi Stock Exchange
 MST:

Minimum Spanning Tree
 PMFG:

Planar maximally filtered graph
 PSX:

Pakistan Stock Exchange
References
Abidin SZ, Old C, Martin T (2010) Effects of New Zealand general elections on stock market returns. Int Rev Bus Res Papers 6:1–12
Białkowski J, Gottschalk K, Wisniewski TP (2008) Stock market volatility around national elections. J Bank Financ 32:1941–1953
Boginski V, Butenko S, Pardalos PM (2005) Statistical analysis of financial networks. Comput Stat Data Anal 48:431–443
BusinessRecorder (2018). Available: https://markets.brecorder.com/ [Accessed 06 Jan 2019]
Chiu D, Ying M, Yu P (2012) A study of Taiwan political elections and stock market dynamism. 2012 Sixth International Conference on Genetic and Evolutionary Computing, 25–28 Aug 2012. 562–565
Coletti P, Murgia M (2016) The network of the Italian stock market during the 2008–2011 financial crises. Algorithmic Finance 5:111–137
Dias J (2012) Sovereign debt crisis in the European Union: a minimum spanning tree approach. Phys A Stat Mech Appl 391:2046–2055
Ghanbari R, Jalili M, Yu X (2018) Correlation of cascade failures and centrality measures in complex networks. Futur Gener Comput Syst 83:390–400
Goodell JW, Mcgroarty F, Urquhart A (2015) Political uncertainty and the 2012 US presidential election: A cointegration study of prediction markets, polls and a standout expert. Int Rev Financ Anal 42:162–171
Jang W, Lee J, Chang W (2011) Currency crises and the evolution of foreign exchange market: evidence from minimum spanning tree. Phys A Stat Mech Appl 390:707–718
Kou G, Chao X, Peng Y, Alsaadi FE, HerreraViedma E (2019) Machine learning methods for systemic risk analysis in financial sectors. Technol Econ Dev Econ 25:716–742
Kou G, Peng Y, Wang G (2014) Evaluation of clustering algorithms for financial risk analysis using MCDM methods. Inf Sci 275:1–12
Kruskal JB (1956) On the shortest spanning subtree of a graph and the traveling salesman problem. Proc Am Math Soc 7(1):48–50
Lee JW, Nobi A (2018) State and network structures of stock markets around the global financial crisis. Comput Econ 51:195–210
Liew VKS, Rowland R (2016) The effect of Malaysia general election on stock market returns. SpringerPlus 5:1975
Majapa M, Gossel SJ (2016) Topology of the south African stock market network across the 2008 financial crisis. Phys A Stat Mech Appl 445:35–47
Mantegna RN (1999) Hierarchical structure in financial markets. Eur Phys J B 11(1):193–197
Marsden PV (2015) Network centrality, measures of. In: Wright JD (ed) International encyclopedia of the Social & Behavioral Sciences (second edition). Elsevier, Oxford
Memon BA, Yao H (2019) Structural change and dynamics of Pakistan stock market during crisis: A complex network perspective. Entropy 21:248
Musmeci N, Aste T, di Matteo T (2016) Interplay between past market correlation structure changes and future volatility outbursts. Sci Rep 6:36320
Nobi A, Maeng SE, Ha GG, Lee JW (2015) Structural changes in the minimal spanning tree and the hierarchical network in the Korean stock market around the global financial crisis. J Korean Phys Soc 66:1153–1159
Oehler A, Walker TJ, Wendt S (2013) Effects of election results on stock price performance: evidence from 1980 to 2008. Manag Financ 39:714–736
Onnela JP, Chakraborti A, Kaski K, Kertész J (2003a) Dynamic asset trees and black Monday. Phys A Stat Mech Appl 324:247–252
Onnela JP, Chakraborti A, Kaski K, Kertész J, Kanto A (2003b) Dynamics of market correlations: taxonomy and portfolio analysis. Phys Rev E 68:056110
Onnela JP, Chakraborti A, Kaski K, Kertsz J, Kanto A (2003c) Asset trees and asset graphs in financial markets. Phys Scr T106:48
OpendoorsPK (2018) Available: http://www.opendoors.pk/ [Accessed 04 Jan 2019]
Pantzalis C, Stangeland DA, Turtle HJ (2000) Political elections and the resolution of uncertainty: the international evidence. J Bank Financ 24:1575–1604
Pástor Ľ, Veronesi P (2013) Political uncertainty and risk premia. J Financ Econ 110:520–545
Radhakrishnan S, Duvvuru A, Sultornsanee S, Kamarthi S (2016) Phase synchronization based minimum spanning trees for analysis of financial time series with nonlinear correlations. Phys A Stat Mech Appl 444:259–270
Sabidussi G (1966) The centrality index of a graph. Psychometrika 31:581–603
Savita, Ramesh A (2015) Return volatility around National Elections: evidence from India. Procedia Soc Behav Sci 189:163–168
Sharif S, Ismail S, Zurni O, Theng LH (2016) Validation of global financial crisis on Bursa Malaysia stocks market companies via covariance structure. Am J Appl Sci 13:1091–1095
Sinha S, Pan RK (2007) Uncovering the internal structure of the Indian financial market: large crosscorrelation behavior in the NSE. In: Chatterjee A, Chakrabarti BK (eds) Econophysics of markets and business networks: proceedings of the EconophysKolkata III. Springer Milan, Milano
Sultonov M, Jehan SN (2018) Dynamic linkages between Japan’s foreign exchange and stock markets: response to the Brexit referendum and the 2016 U.S. Presidential election. J Risk Financ Manage 11:34
Sultornsanee S, Duvvuru A, Radhakrishnan S, Chowdhary H, Kamarthi S (2013) Phase synchronization based minimum spanning trees for the analysis and visualization of currency exchange markets. Proc Comput Sci 20:460–465
Tabak BM, Serra TR, Cajueiro DO (2010) Topological properties of stock market networks: the case of Brazil. Phys A Stat Mech Appl 389:3240–3249
Tumminello M, Aste T, di Matteo T, Mantegna RN (2005) A tool for filtering information in complex systems. Proc Natl Acad Sci 102:10421–10426
Wang GJ, Xie C, Chen S (2017) Multiscale correlation networks analysis of the US stock market: a wavelet analysis. J Econ Interac Coord 12:561–594
Wiliński M, Sienkiewicz A, Gubiec T, Kutner R, Struzik ZR (2013) Structural and topological phase transitions on the German stock exchange. Phys A Stat Mech Appl 392:5963–5973
Xia L, You D, Jiang X, Guo Q (2018) Comparison between global financial crisis and local stock disaster on top of Chinese stock network. Phys A Stat Mech Appl 490:222–230
Yan XG, Xie C, Wang GJ (2015) Stock market network’s topological stability: evidence from planar maximally filtered graph and minimal spanning tree. Int J Modern Phys B 29:1550161
Yang R, Li X, Zhang T (2014) Analysis of linkage effects among industry sectors in China’s stock market before and after the financial crisis. Phys A Stat Mech Appl 411:12–20
Yao H, Memon BA (2019) Network topology of FTSE 100 index companies: from the perspective of Brexit. Phys A Stat Mech Appl 523:1248–1262
Zhang Y, Lee GHT, Wong JC, Kok JL, Prusty M, Cheong SA (2011) Will the US economy recover in 2010? A minimal spanning tree study. Phys A Stat Mech Appl 390:2020–2050
Zhao L, Wang GJ, Wang M, Bao W, Li W, Stanley HE (2018) Stock market as temporal network. Phys A Stat Mech Appl 506:1104–1112
Zhuang R, Hu B, Ye Z (2008) Minimal spanning tree for ShanghaiShenzhen 300 stock index. 2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence), 1–6 June 2008. 1417–1424
Acknowledgements
We are grateful to the editor and anonymous referees for their insightful suggestions. We also thank the Southwestern University of Finance and Economics for providing open access to our article.
Funding
Not applicable.
Author information
Authors and Affiliations
Contributions
BAM (First & Corresponding Author) carried out the idea, designed the research, analyzed and interpreted data, applied methodology and visualization, discussed the results and wrote the article. HY (second author) supervised the research, provided necessary resources, and reviewed the article. RT (third author) helped in data collection, and revising the article. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare that they have no competing interests.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Cite this article
Memon, B.A., Yao, H. & Tahir, R. General election effect on the network topology of Pakistan’s stock market: networkbased study of a political event. Financ Innov 6, 2 (2020). https://doi.org/10.1186/s408540190165x
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s408540190165x
Keywords
 Minimum spanning tree
 Centrality measures
 General elections
 Emerging market
 Pakistan
 Stock market network