Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Information flow among stocks, bonds, and convertible bonds

  • Kihwan Jo,

    Roles Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft

    Affiliation Yonsei University, Seoul, South Korea

    ⨯
  • Gahyun Choi,

    Roles Data curation, Formal analysis, Methodology, Project administration, Software, Validation, Visualization, Writing – original draft

    Affiliation Yonsei University, Seoul, South Korea

    ⨯
  • Jongwook Jeong,

    Roles Conceptualization, Formal analysis, Investigation, Methodology, Supervision, Validation, Writing – original draft, Writing – review & editing

    Affiliation SK Square, Seoul, South Korea

    ⨯
  • Kwangwon Ahn

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Resources, Supervision, Validation, Writing – original draft, Writing – review & editing

    k.ahn@yonsei.ac.kr

    Affiliation Yonsei University, Seoul, South Korea

    ⨯

Abstract

This study examines the information flow between convertible bonds (CBs) and other investment assets, such as stocks and bonds. In particular, we employ transfer entropy (TE) as a proxy for the causal effect between the two assets considering that one of the most widely used methods, Granger causality, requires strict assumptions. When adopting TE, we find that asymmetric information flow arising between assets depends on macroeconomic phases. The stock and bond markets affected the CB market prior to and during the global financial crisis, respectively. In the post-crisis period, we find no meaningful information exchange between CBs and other investment assets concerning their return series. However, we observe a significant cause–effect relationship between CBs and stocks in the rise–fall patterns of their price series. The findings suggest that the appearance of one-directional information flow depends on macroeconomic conditions and the level of data, for example, return series or price fluctuations. Accordingly, investors could exploit this pattern predictability in their portfolio management. In addition, policymakers must closely monitor the information flow among the three markets. When any two markets exchange information in a state of strong market integration, unbalanced regulation between them could lead to market distortions and regulatory arbitrage.

Citation: Jo K, Choi G, Jeong J, Ahn K (2023) Information flow among stocks, bonds, and convertible bonds. PLoS ONE 18(3): e0282964. https://doi.org/10.1371/journal.pone.0282964

Editor: Hung Do, Massey University - Albany Campus: Massey University - Auckland Campus, NEW ZEALAND

Received: July 10, 2022; Accepted: February 28, 2023; Published: March 23, 2023

Copyright: © 2023 Jo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The data underlying the results presented in the study are available from Bloomberg. One of authors works at SK Square in which subscribes the Bloomberg. All relevant data are within the manuscript and its Supporting Information files.

Funding: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (No. 2022R1A2C1004258; Kwangwon Ahn).

Competing interests: The authors have declared that no competing interests exist.

1. Introduction

Rapidly growing railway companies first issued convertible bonds (CBs) in the United States during the 19th century in the form of mortgage securities with conversion terms to attract investors. Companies in growing industries could finance their growth by issuing CBs that were less costly than traditional corporate bonds. Moreover, CBs offer investors the opportunity to protection against downside risks, such as those offered provided by bonds, while enjoying an upside return similar to that of a stock. When the private placement market was first introduced in the 1990s, CBs became a member of a global asset class providing investors with opportunities to select tailored securities. From 1994 to 2018, CBs exhibited performance between stocks and bonds in terms of annual average return and volatility. CBs have also exhibited a lower downside risk than stocks during stock market downturns and an upside return when interest rates have risen, and bond prices have fallen. These patterns suggest that CBs can be used to better manage risk than stock and bond portfolios and thus can be an attractive investment vehicle depending on market conditions.

This study analyzes nonlinear causality by examining transfer entropy (TE) and effective transfer entropy (ETE) between CBs and other investment assets, i.e., stocks and bonds, to determine the predictability of return series, or at least the price fluctuation patterns among these assets. Our results show how an asset could be used as a predictor for the other depending on different macroeconomic phases using information flow. Before the financial crisis, significant information flows were observed from the stock market to the CB market, and corresponding information flows were not observed from the bond market to the CB market. This finding implies that stocks can be used as a predictor to forecast CB performance. By contrast, significant information flows were observed only from the bond market to the CB market during the financial crisis, implying that bonds can be used as a predictor for CB evolution. We fail to find any causal link between the return distributions of CBs and stocks or bonds during the post-crisis period. However, the results indicate that fluctuations in the CB market can be used to predict those in the stock market. Therefore, investors could have been able to predict stock market fluctuations by monitoring CB market fluctuations, at least for the period after the crisis ended.

The remainder of this paper is structured as follows. Section 2 reviews the related literature; Section 3 describes the data and explains the methodology. Section 4 discusses the results, and Section 5 concludes the study.

2. Literature review

CBs have been extensively studied from various angles. However, the most common are pricing models with their issuance terms, including term to maturity [1–3], conversion provisions [4], redemption clause at maturity [5, 6], seniority of debt [7], investor puttability and issuer callability [8], interest rate [9, 10], and the relationship between CB and underlying stock [11]. In addition, several recent studies tested valuation models with historical firm data in the real world [12, 13]. Accordingly, CB valuation has become a common exercise using structural and reduced-form models that use different input variables. These studies collaborated in determining the adequacy of CB valuation. However, they are still insufficient for investors to use in constructing investment strategies, particularly in terms of sensitivity to different macroeconomic phases.

Other streams of research have been conducted from the portfolio management perspective, including CB effects on portfolio returns. For instance, [14] compared the performance of CB funds and low-grade corporate bonds. [15] adjusted the proportion of stocks and bonds in CB portfolios by decomposing the value of CBs into bond and option values. Furthermore, [16] inferred investment details through the performance of CB funds. Previous studies intensively reviewed investment strategies, such as hedging and arbitrage strategies, for improving portfolio performance [17, 18]. In particular, researchers studied asset allocation extensively by (1) employing regression models and correlation coefficients of CBs with stocks and bonds, which indicates that partial replacement of stocks with CBs could create a portfolio with an optimized risk-adjusted rate of return [19]; (2) characterizing the CB funds managed in the United States and Europe, which demonstrates that CB funds in the United States have been strongly correlated with stock returns, whereas those in Europe have been strongly correlated with bond yields [20]. However, with the increasing pace of information exchange and growing interdependency of financial markets, understanding and predicting markets to effectively manage portfolios by solely relying on correlation coefficients, regression analysis, and even the Granger causality test have become increasingly difficult [21–29]. Given that when any two markets have strong market integration with intensive information exchange, regulation risk arising from the unbalanced policy could negatively affect the market and cause regulatory arbitrage, as shown in previous studies regarding the CB market [30, 31].

In particular, Granger causality has been widely employed to identify the cause–effect relationships between financial assets, but in a very limited sense. Although the Granger causality test is commonly used to identify couplings between two systems, it is not free from criticism owing to its structural restrictions. Moreover, its use is always accompanied by a caution that its results should be carefully considered and interpreted [32]. In particular, (i) “both variables must be stationary” and (ii) “the residuals of the vector autoregression model should be Gaussian white noise” to ensure the reliability of the Granger causality test [21, 28, 29]. Unlike the Granger causality test, TE is a measure that estimates directly from data, does not require the input of any structural insights, and does not suffer from any noisy measurements of coupled dynamic systems, for example, ETE [32, 33]. Therefore, TE and ETE could be the right choice for analyzing nonlinear dependencies between dynamically coupled systems [32]: information flows between the European and US stock markets with overlapping trading hours [25]; causal linkages among virtual currencies, gold, and the dollar [28]; information flow among metals market [34]; ETE between the VIX and stock markets [26]; predicting US stock price movement using ETE with machine learning [35]; clarifying relationship among epu, investor sentiment, and stock market using ETE [36]; the information exchange between stock markets in emerging economies and foreign exchange rates [27]; group TE between cryptocurrency market [37]; information flow between new and old forks after hard forks in cryptocurrency markets [29]; the relationship among flight-to-quality, flight-from-quality occurrences, and stock-bond nexus during the pandemic era [38]; and information flow between investor sentiment and stock market [36].

3. Data and methodology

3.1 Data

The data were obtained from Bloomberg. The TR Global Convertible Index, MSCI World Stock Index, and BB Global Aggregate Bond Index were used as proxies for the CB, stock, and bond markets, respectively. Following [39, 40], we set the global financial crisis as the period from July 2007 to September 2009. A total of 1,627 trading days, covering 6 years and 3 months, comprise the global financial crisis period and 2 years before and after the crisis. Then, the data were converted into log returns because the price series has a unit root. Table 1 provides descriptive statistics for the entire sample, pre-crisis, crisis, and post-crisis periods. The whole sample period spans from July 2005 to September 2011. The three subperiods are defined as the pre-crisis (July 2005–June 2007), crisis (July 2007–September 2009), and post-crisis (October 2009–September 2011) periods.

thumbnail
Download:
Table 1. Descriptive statistics.

https://doi.org/10.1371/journal.pone.0282964.t001

Regardless of the sample period, stocks have the most volatile return distribution, followed in order by CBs and bonds. During the pre-crisis period, stocks had the highest average daily returns, again followed in order by CBs and bonds. However, bonds have the highest average daily returns during and after the global financial crisis, followed in order by CBs and stocks. For all subperiods, CBs exhibit average return and volatility between those of stocks and bonds. Stocks have a fat tail over the whole sample period [41–43], as do CBs and bonds to a lesser extent. As shown in Fig 1, the three asset classes had high excess kurtosis during the crisis period because of investor herding behavior and consequent volatility clustering [44].

thumbnail
Download:
Fig 1. Cumulative density function of the return of CB, stock, and bond in four subsample periods: Whole sample, pre-crisis, crisis, and post-crisis periods.

https://doi.org/10.1371/journal.pone.0282964.g001

3.2 Transfer entropy

[23] proposed the concept of TE to measure asymmetric interactions between two systems—an alternative nonlinear causality measure. TE is in an asymmetric form, which makes it possible to measure cause–effect relationships. However, the drawbacks of TE originated from computational complexity [45, 46], resulting in identification issues [47]. Xi and Yi are considered two discrete random variables. The length of each, that is, k and l, defines and , respectively. Then, TE can be defined as follows [23]: where stands for the degree of uncertainty in predicting Xi+1 for a given , and is the degree of uncertainty considering and in predicting Xi+1 Thus, TEY→X indicates the effect of on predicting Xi+1.

TE is likely to be biased because of finite sample effects, and thus, we employ ETE as a remedy. When estimating the ETE, the average TE, calculated over a finite number of independently shuffled time series, is subtracted from the observed TE. This process disrupts the temporal order in the time series and would therefore remove any possible causal relationship arising from the bias of a finite sample size. The ETE is calculated following [29, 48–50]: where Y(i) is an independently shuffled series of Y in the i-th trial; thus, ETE is demeaned TE, which is calculated by subtracting the arithmetic average of the randomized TE values from the estimated TE. Specifically, we shuffled 1,000 times and produced 1,000 TE values from independently shuffled time series over the same sample space.

We calculate the TE through histogram analysis, the most common method for dealing with discrete random variables. We use the mean squared error criterion to estimate the number of bins—an equally spaced interval of the sample range [51–53]. In addition, we use symbolic time series analysis (STSA) as an alternative method for estimating the TE. Owing to its robustness to noise, STSA is widely used in various fields, including physics, information theory, and finance [54, 55]. We use the log return data to convert each value to 0 or 1, reflecting the rise–fall pattern of the price series. Given the consecutive binary sequence, we convert the data into decimal numbers with a specific window size [56].

We then derive the significance level for TE by bootstrapping the underlying Markov process 300 times [26, 57]. The number of bootstrap replications for each direction of the estimated TE allows us to assess the statistical significance of TE estimates. Here we rely on a Markov block bootstrap proposed by [26]. In contrast to shuffling, the Markov block bootstrap preserves the dependencies within each time series. This process generates the distribution of TE estimates under the null hypothesis that no information transfer exists between two time series. Finally, we standardize the TE to a Z-score. As the mean of the Z-score is approximately zero according to the law of large numbers, the Z-score spread approximately indicates the statistical significance of TE [36, 49, 50]. In sum, we can loosely say that (i) a |Z-score| > 2.58 is in the top 1% of the critical region and hence is comparable to a p-value of 0.01; and (ii) a |Z-score| > 1.96 is in the top 5% of the critical region.

4. Results and discussion

4.1 Results

Fig 2(A) shows a summary of the information flow between CBs and the other investment asset classes—stocks and bonds—during the pre-crisis period. Our result indicates that statistically significant information transfer exists from the stock market to the CB market. ETE, the value in parentheses, further confirms that our results are not produced by random noise. This cause–effect relationship can be explained by the descriptive statistics reported in Panel B of Table 1. In particular, skewness and excess kurtosis—CBs are closely related to stocks. Prior studies documented that CBs illuminate the characteristics of stocks during times of continuous market rally [19] and depict the relationship between CB and the stock market [58, 59]. Therefore, we conclude that the CB market reflects the core features of the stock market along with the continued growth in the stock market since the early 2000s, during times of continuous market rally.

thumbnail
Download:
Fig 2. Information flow between CBs and other investment assets before and during the global financial crisis.

The arrow indicates the direction of the causal link, the color represents the significance level (red and blue lines represent statistically significant information flows with p-value lower than 0.01 and 0.10, respectively, and the dashed line indicates insignificant TE with a p-value higher than 0.10), and p means p-value. The number denotes the estimated value of TE, and the value in parentheses stands for ETE. The significance level was evaluated by bootstrapping the underlying Markov process [26, 57]. We further apply the quantile-based TE for robustness check. In particular, cutoff points of 0.05 and 0.95 are used for the estimation (see S1 Appendix).

https://doi.org/10.1371/journal.pone.0282964.g002

As shown in Fig 2(B), significant information flow occurs only from bonds to CBs during the global financial crisis, and our result is free from bias caused by finite sample effects. This finding is in line with [60], indicating that CBs showed strong bond-like characteristics in the US market during the global financial crisis. Moreover, this result is further confirmed by the conversion ratio, which was 0.543 on average during the global financial crisis [60], implying that the value of the conversion option had nearly disappeared—that is, it was out of the money. In Fig 3, the index performance of the three asset classes indicates that when the stock market struggles and falls sharply, the CB market exhibits downward rigidity and even rises with the bond market. This movement of the CB market is aligned with [61]. Thus, information flow from bonds to CBs during the financial crisis is well supported.

thumbnail
Download:
Fig 3. Index performance for the CB, stock, and bond markets during the sample period.

Each index is normalized at 100 starting from June 30, 2005.

https://doi.org/10.1371/journal.pone.0282964.g003

As shown in Fig 4(A), we cannot find any significant information exchange between the CB market and the two other markets during the post-crisis period—at least determined by employing histogram analysis between any two return distributions. Accordingly, we further investigate information flow at the level of rise–fall patterns among the different price series—that is, through STSA—and summarize the results, as shown in Fig 4(B). Specifically, TE presents two unidirectional information flows: the CB market receives information from the bond market, processes the details in its own manner, and then transmits information to the stock market. We then investigate ETE to overcome sample bias and conclude that our findings are quite robust. This result implies that the CB market is not only a transmitter of information but also a producer—the information outflow to the stock market is greater than the information inflow from the bond market.

thumbnail
Download:
Fig 4. Information flow among the CB, stock, and bond markets during the post-crisis period.

The arrow indicates the direction of the causal link, the color represents the significance level (red and blue lines represent statistically significant information flows with a p-value lower than 0.01 and 0.10, respectively, and the dashed line indicates insignificant TE with a p-value higher than 0.10), and p means p-value. The number denotes the estimated value of TE, and the value in parentheses stands for ETE. The significance level was evaluated by bootstrapping the underlying Markov process [26, 57]. We further apply the quantile-based TE for robustness check. In particular, cutoff points of 0.05 and 0.95 are used for the estimation (see S1 Appendix).

https://doi.org/10.1371/journal.pone.0282964.g004

4.2 Discussion

4.2.1 Near-zero interest rates and quantitative easing.

Information flows from bonds to other investment assets, at least about post-crisis price fluctuation patterns, are the results of interest rate cuts implemented by financial authorities as a risk-mitigation measure in times of financial turmoil [62]. A policy of monetary easing increased bond prices—bonds paying a fixed coupon become more attractive as interest rates fall, thereby driving up demand and bond prices. Near-zero interest rates and quantitative easing have stimulated economic activity, resulting in booming asset prices, for example, stocks and CBs [63–65]. In particular, the information flow from bonds to CBs suggests that an increase in CB prices reflects the bond floor well, i.e., the sum of the present value of coupon and par resulting from an increase in bond price [66].

4.2.2 Lifting of regulations on short selling.

Moreover, lifting regulations on short selling facilitated information flow from the CB market to the stock market during the post-crisis period. Restrictions on short selling were imposed in various countries during the global financial crisis [67]. Accordingly, arbitrage opportunities disappeared between the CB and stock markets—that is, through a combination of purchasing CBs and short-selling stocks. However, bans on short selling were lifted along with economic recovery. Thus, arbitrageurs once again provided liquidity [68, 69] to the CB market and helped reduce volatility in the stock market. Our findings are generally in line with those of prior studies, which documented that derivative markets (i) increase liquidity in spot markets [70–72] and (ii) exhibit price leadership over the corresponding spot market [73].

4.2.3 Return predictions on growth stocks.

TE provides details about the correlation between two time series. In particular, TE measures information transport that explicitly incorporates the directional and dynamic structure of information flow [23]. Thus, this step is essential en route to assessing causal relations between systems [74]. Accordingly, during the post-crisis period, the significant information flow—that is, at the level of rise–fall patterns in price series—from CBs to stocks implies that the price fluctuation patterns of CBs could be used to forecast the price fluctuation patterns of stocks, particularly those for growth stocks. Firms that issue CBs mostly share small and growing features, and CB funds comprise 60% small-cap growth stocks and only 40% AAA-rated corporate bonds [16]. Therefore, CB characteristics are by nature more like those of growth stocks than those of value stocks. When historically low interest rates weakened the value premium, the price fluctuation patterns of CBs led to price fluctuations in stocks. Moreover, changes in the stock market, particularly those for growth stocks, can be predicted by monitoring changes in the performance of CBs.

5. Conclusions

This study investigates the information transfer between CBs and other investment assets, particularly during different macroeconomic phases. We find that the stock market affects the CB market before the global financial crisis, whereas the bond market influences the CB market during the crisis. However, none of the significant information exchange between the CB market and the two other markets is found during the post-crisis period. The post-crisis period is a turmoil period with unstable and unconventional price movement. Moreover, our findings are well supported by prior studies: among cryptocurrency markets [75] and between stock and CDS markets [76]. However, we observe a significant cause–effect relationship from CBs to stocks at the level of rise–fall patterns in their price series.

The implications of our findings for investors and policymakers are as follows. First, similar predictability may be observed in the future if one-directional information flow arises depending on (i) different macroeconomic phases and (ii) different levels of data, such as return series and price fluctuations. Investors could take advantage of this pattern predictability in their portfolio management. Moreover, policymakers (regulators) must closely monitor information flow among the three markets, namely, CBs, bonds, and stocks, depending on the macroeconomic phase. When any two markets exchange information in a state of strong market integration, the unsuitable regulation between them leads to market distortions and regulatory arbitrage [30, 31, 77].

Future studies can investigate information flows between CBs and other financialized assets, such as commodities and REITs, to examine the predictability in their return series and the association of rise–fall patterns in their price series. Moreover, we can extend the data range, including more turmoil periods, and investigate the detail of the information network structure regarding different macroeconomic phases, such as the European debt crisis and cash crunch in China.

References

  1. 1. Merton R. C. (1974). On the pricing of corporate debt: The risk structure of interest rates. J Finance, 29(2), 449–470.
  2. 2. Brennan M. J., & Schwartz E. S. (1977). Convertible bonds: Valuation and optimal strategies for call and conversion. J Finance, 32(5), 1699–1715.
  3. 3. Kao D. L. (2000). Estimating and pricing credit risk: An overview. Financial Anal. J., 56(4), 50–66.
  4. 4. Ingersoll J. E. Jr (1977). A contingent-claims valuation of convertible securities. J Financ Econ., 4(3), 289–321.
  5. 5. Cox J. C., Ross S. A., & Rubinstein M. (1979). Option pricing: A simplified approach. J Financ. Econ., 7(3), 229–263.
  6. 6. Tan X., Zhang Z., Zhao X., & Wang S. (2022). DeepPricing: Pricing convertible bonds based on financial time-series generative adversarial networks. Financ. Innov., 8(1), 1–38.
  7. 7. Jarrow R. A., & Turnbull S. M. (1995). Pricing derivatives on financial securities subject to credit risk. J Finance, 50(1), 53–85.
  8. 8. Tsiveriotis K., & Fernandes C. (1998). Valuing convertible bonds with credit risk. J. Fixed Income, 8(2), 95.
  9. 9. Duffie D., & Singleton K. J. (1999). Modeling term structures of defaultable bonds. Rev. Financ. Stud., 12(4), 687–720.
  10. 10. Hung M. W., & Wang J. Y. (2002). Pricing convertible bonds subject to default risk. J. Deriv., 10(2), 75–87.
  11. 11. Zhang Z., Wang Z., & Chen X. (2021). Pricing convertible bond in uncertain financial market. J Uncertain Syst., 14(01), 2150007.
  12. 12. Finnerty J. D., & Tu M. (2017). Valuing convertible bonds: A new approach. Bus. Valuat. Rev., 36(3), 85–102.
  13. 13. Wang P., Lau W. Y., & Goh L. T. (2021). Valuation analysis of convertible bonds in China: Does theoretical value deviate from actual market price? Contemp. Chinese Political Econ. Strateg. Relat, 7(3), 1735–XXII.
  14. 14. Kihn J. (1996). The effect of embedded options on the financial performance of convertible bond funds. Financial Anal. J., 52(1), 15–26.
  15. 15. Ho T. S., & Pfeffer D. M. (1996). Convertible bonds: Model, value attribution, and analytics. Financial Anal. J., 52(5), 35–44.
  16. 16. Domian D. L., & Reichenstein W. R. (2008). Returns-based style analysis of convertible bond funds. J. Fixed Income, 18(3), 52–64.
  17. 17. Kim H. J., & Han S. H. (2019). Convertible bond announcement returns, capital expenditures, and investment opportunities: Evidence from Korea. Pac. Basin Finance J., 53, 331–348.
  18. 18. Xie L., & Li Y. (2022). Analysis of investment opportunities and strategies of convertible bonds. Front. Econ. Manage, 3(6), 151–155.
  19. 19. Ranaldo A., & Eckmann A. (2004). Convertible bonds: Characteristics of an asset class. UBS Research Pap.
  20. 20. Van Campenhout G., & Vanpée R. (2017). My global fund portfolio is not yours: The effect of home bias on European-and US-managed convertible bond fund exposures. Eur. J. Finance, 23(14), 1335–1361.
  21. 21. Granger C. W. (1969). Investigating causal relations by econometric models and cross–spectral methods. Econometrica, 37(3), 424–438.
  22. 22. Theiler J., Eubank S., Longtin A., Galdrikian B., & Farmer J. D. (1992). Testing for nonlinearity in time series: The method of surrogate data. Physica D, 58(1–4), 77–94.
  23. 23. Schreiber T. (2000). Measuring information transfer. Phys. Rev. Lett., 85(2), 461–464. pmid:10991308
  24. 24. Marschinski R., & Kantz H. (2002). Analysing the information flow between financial time series. Eur Phys J B, 30(2), 275–281.
  25. 25. Jizba P., Kleinert H., & Shefaat M. (2012). Rényi’s information transfer between financial time series. Phys. A: Stat. Mech. Appl., 391(10), 2971–2989.
  26. 26. Dimpfl T., & Peter F. J. (2013). Using transfer entropy to measure information flows between financial markets. Stud. Nonlinear Dyn. Econom., 17(1), 85–102.
  27. 27. Sensoy A., Sobaci C., Sensoy S., & Alali F. (2014). Effective transfer entropy approach to information flow between exchange rates and stock markets. Chaos, solitons & fractals, 68, 180–185.
  28. 28. Jang S. M., Yi E., Kim W. C., & Ahn K. (2019). Information flow between Bitcoin and other investment assets. Entropy, 21(11), 1116.
  29. 29. Yi E., Cho Y., Sohn S., & Ahn K. (2021). After the splits: Information flow between Bitcoin and Bitcoin family. Chaos, Solitons & Fractals, 142, 110464.
  30. 30. Shah A. K. (1997). Regulatory arbitrage through financial innovation. Account. Audit. Account. J., 10(1), 85–104.
  31. 31. Li H., Liu H., & Siganos A. (2016). A comparison of the stock market reactions of convertible bond offerings between financial and non-financial institutions: Do they differ?. Int. Rev. Financial Anal., 45, 356–366.
  32. 32. Gençağa D. (2018). Transfer entropy. Entropy, 20(4), 288. pmid:33265379
  33. 33. Friston K. J., Bastos A. M., Oswal A., van Wijk B., Richter C., & Litvak V. (2014). Granger causality revisited. Neuroimage, 101, 796–808. pmid:25003817
  34. 34. Huynh T. L. D. (2020). The effect of uncertainty on the precious metals market: New insights from Transfer Entropy and Neural Network VAR. Resour. Policy, 66, 101623.
  35. 35. Kim S., Ku S., Chang W., & Song J. W. (2020). Predicting the direction of US stock prices using effective transfer entropy and machine learning techniques. IEEE Access, 8, 111660–111682.
  36. 36. Yao C. Z., & Li H. Y. (2020). Effective transfer entropy approach to information flow among epu, investor sentiment and stock market. Front. Phys., 8, 206.
  37. 37. Dimpfl T., & Peter F. J. (2019). Group transfer entropy with an application to cryptocurrencies. Phys. A: Stat. Mech. Appl., 516, 543–551.
  38. 38. Bossman A., Agyei S. K., Owusu Junior P., Agyei E. A., Akorsu P. K., Marfo-Yiadom E., et al. (2022). Flights-to-and-from-quality with Islamic and conventional bonds in the COVID-19 pandemic era: ICEEMDAN-based transfer entropy. Complexity, 2022.
  39. 39. Filardo A., George J., Loretan M., Ma G., Munro A., Shim I., et al. (2010). The international financial crisis: Timeline, impact and policy responses in Asia and the Pacific. BIS Papers, 52.
  40. 40. Bai J., & Collin-Dufresne P. (2011). The determinants of the CDS-bond basis during the financial crisis of 2007–2009. Financ Manage., 48(2), 417–39.
  41. 41. Mandelbrot B. (1963). New methods in statistical economics. J Polit Econ, 71(5), 421–440.
  42. 42. Fama E. F. (1965). The behavior of stock-market prices. J. Bus., 38(1), 34–105.
  43. 43. Jondeau E., & Rockinger M. (2003). Testing for differences in the tails of stock-market returns. J Empir Finance, 10(5), 559–581.
  44. 44. Schmitt N., & Westerhoff F. (2017). Herding behaviour and volatility clustering in financial markets. Quant Finance, 17(8), 1187–1203.
  45. 45. Papana A., Papana-Dagiasis A., & Siggiridou E. (2020). Shortcomings of transfer entropy and partial transfer entropy: Extending them to escape the curse of dimensionality. Chaos, 30(16), 2050250.
  46. 46. Zhang X., Hu W., & Yang F. (2022). Detection of cause-effect relations based on information granulation and transfer entropy. Entropy, 24(2), 212. pmid:35205507
  47. 47. Runge J., Heitzig J., Marwan N., & Kurths J. (2012). Quantifying causal coupling strength: A lag-specific measure for multivariate time series related to transfer entropy. Phys. Rev. E, 86(6), 061121. pmid:23367907
  48. 48. Sandoval L. Jr, (2014). Structure of a global network of financial companies based on transfer entropy. Entropy, 16(8), 4443–4482.
  49. 49. Joo K., Jeong M., Seo Y., Suh J. H., & Ahn K. (2021). Shanghai crude oil futures: Flagship or burst?. Energy Rep., 7, 4197–4204.
  50. 50. Lee S., Yi E., Cho Y., & Ahn K. (2022). The path to a sustainable palm oil futures market. Energy Rep., 8, 6543–6550.
  51. 51. Larson H. J. (1982). Introduction to probability and statistical inference. John Wiley & Sons, Inc., 605 third Ave., New York, NY 10158, 1982, 480.
    • 52. Scott D. W. (1979). On optimal and data-based histograms. Biometrika, 66(3), 605–610.
    • 53. Freedman D., & Diaconis P. (1981). On the histogram as a density estimator: L 2 theory. Prob. Theory Related Fields, 57(4), 453–476.
    • 54. Ruiz M. D. C., Guillamón A., & Gabaldón A. (2012). A new approach to measure volatility in energy markets. Entropy, 14(1), 74–91.
    • 55. Risso W. A. (2018). Symbolic time series analysis and its application in social sciences. Intech.
    • 56. Ahn K., Lee D., Sohn S., & Yang B. (2019). Stock market uncertainty and economic fundamentals: An entropy-based approach. Quant Finance, 19(7), 1151–1163.
    • 57. Horowitz J. L. (2003). Bootstrap methods for Markov processes. Econometrica, 71(4), 1049–1082.
    • 58. Brupbacher R., Daragon M., Eckmann A., Sperl U., & Tranier C. The benefits of convertible bonds. (2021). UBS Research Pap.
    • 59. Ling Y. X., Xie C., & Wang G. J. (2022). Interconnectedness between convertible bonds and underlying stocks in the Chinese capital market: A multilayer network perspective. Emerg. Mark. Rev., 100912.
    • 60. Batten J. A., Khaw K. L. H., & Young M. R. (2018). Pricing convertible bonds. J Bank Financ., 92, 216–236.
    • 61. He Z., & Krishnamurthy A. (2018). Intermediary asset pricing and the financial crisis. Annu. Rev. Financ. Econ., 10, 173–197.
    • 62. Ramey V. A. (2019). Ten years after the financial crisis: What have we learned from the renaissance in fiscal research?. J. Econ. Perspect., 33(2), 89–114.
    • 63. Gambacorta L., Hofmann B., & Peersman G. (2014). The effectiveness of unconventional monetary policy at the zero lower bound: A cross‐country analysis. J. Money Credit Bank., 46(4), 615–642.
    • 64. Lima L., Vasconcelos C. F., Simão J., & de Mendonça H. F. (2016). The quantitative easing effect on the stock market of the USA, the UK and Japan: An ARDL approach for the crisis period. J Econ Stud., 43(6), 1006–1021.
    • 65. Brunnermeier M., Farhi E., Koijen R. S., Krishnamurthy A., Ludvigson S. C., Lustig H., et al. (2021). Perspectives on the future of asset pricing. Rev. Financ. Stud., 34(4), 2126–2160.
    • 66. Hundt S., Sprungk B., & Horsch A. (2017). The information content of credit ratings: Evidence from European convertible bond markets. Eur. J. Finance, 23(14), 1414–1445.
    • 67. Siciliano G., & Ventoruzzo M. (2020). Banning Cassandra from the market? An empirical analysis of short-selling bans during the Covid-19 crisis. Eur. Co. Financ. Law Rev., 17(3–4), 386–418.
    • 68. Choi D., Getmansky M., & Tookes H. (2009). Convertible bond arbitrage, liquidity externalities, and stock prices. J Financ. Econ., 91(2), 227–251.
    • 69. Ferreruela S., & Martín D. (2022). Market Quality and Short-Selling Ban during the COVID-19 Pandemic: A High-Frequency Data Approach. J. Risk Financ. Manag., 15(7), 308.
    • 70. Detemple J., & Jorion P. (1990). Option listing and stock returns: An empirical analysis. J Bank Financ., 14(4), 781–801.
    • 71. Kumar R., Sarin A., & Shastri K. (1998). The impact of options trading on the market quality of the underlying security: An empirical analysis. J Finance, 53(2), 717–732.
    • 72. Umar Z., Gubareva M., Yousaf I., & Ali S. (2021). A tale of company fundamentals vs sentiment driven pricing: The case of GameStop. J. Behav. Exp. Finance, 30, 100501.
    • 73. Ahn K., Bi Y., & Sohn S. (2019). Price discovery among SSE 50 Index‐based spot, futures, and options markets. J Futures Mark., 39(2), 238–259.
    • 74. Murari A., Lungaroni M., Peluso E., Gaudio P., Lerche E., Garzotti L., et al. (2018). On the use of transfer entropy to investigate the time horizon of causal influences between signals. Entropy, 20(9), 627. pmid:33265716
    • 75. Foroutan P., & Lahmiri S. (2022). The effect of COVID-19 pandemic on return-volume and return-volatility relationships in cryptocurrency markets. Chaos, Solitons & Fractals, 162, 112443. pmid:36068915
    • 76. Norden L., & Weber M. (2009). The co‐movement of credit default swap, bond and stock markets: An empirical analysis. Eur. Financ. Manag., 15(3), 529–562.
    • 77. Oster P. (2020). Contingent Convertible bond literature review: Making everything and nothing possible?. J. Bank. Regul., 21(4), 343–381.