Convergence of Caribbean stock exchanges/Convergencia en las Bolsas de Valores del Caribe/Convergence des Bourses Caribeennes.
This paper uses daily data for market returns as well as for the cross-listed securities on the three main stock exchanges in the Caribbean--the Barbados Stock Exchange, the Trinidad and Tobago Stock Exchange and the Jamaica Stock Exchange--to examine the beta-convergence and sigma-convergence of the markets. The results suggest that with respect to sigma-convergence, while the markets are becoming increasingly integrated, the convergence of the returns of the cross-listed securities is debatable, indicating some degree of information asymmetry. The findings of the beta-convergence models imply that the speed of convergence is still rather slow in comparison to other markets throughout the world.
Este trabajo se basa en datos diarios sobre los retornos de inversiones en el mercado y sobre las garantias comparadas de las tres bolsas de valores principales en el Caribe--la Bolsa de Valores de Barbados, la Bolsa de Valores de Trinidad y Tobago y la Bolsa de Valores de Jamaica--a fin de analizar la betaconvergencia y la sigmaconvergencia de los mercados. Los resultados sugieren que con respecto a la sigmaconvergencia, toda vez que la integracion de los mercados se torna cada mas creciente, la convergencia de los retornos de inversiones de las garantias comparadas es discutible y se indica cierto grado de asimetria en la informacion. Los hallazgos a partir de los modelos de betaconvergencia implican que la velocidad de la convergencia es aun lenta en comparacion con otros mercados a traves del mundo.
Cet article se sert de donnees quotidiennes de rendement sur les marches boursiers et sur les titres listes sur les trois bourses de la Caraibe--la Bourse de la Barbade, la Bourse de Trinite et Tobago et la Bourse de la Jamaique--afin d'examiner la beta-convergence et la sigma-convergence des marches. Les resultats suggerent qu'en matiere de sigma-convergence, si les marches sont en phase acceleree d'integration, la convergence des rendements des titres inscrits a plusieurs bourses est contestable, ce qui indique un certain degre d'asymetrie des informations. Les conclusions du modele de beta-convergence laissent a penser que la convergence est encore assez lente comparativement a d'autres marches dans le monde.
The three main stock exchanges in the Caribbean--the Barbados Stock Exchage (BSE), the Jamaica Stock Exchange (JSE) and the Trinidad and Tobago Stock Exchange (TTSE)--have been seriously considering combining their markets into one Caribbean stock market in order to surmount some of the challenges they have faced in their 30-odd years of existence. As extremely small markets by internatonal standards, the first major challenge has been their struggle to attract new listings. The largest market of the three is the JSE, with 39 ordinary shares registered at the end of 2008, while the smallest market, the BSE, had only 24 ordinary shares listed.
There are a number of reasons why there are so few companies opting to register on the exchanges. One that has been debated publicly is the difficulty some firms face when trying to meet the listing criteria. Since the vast majority of Caribbean companies are small, and the total number of firms operating in a given country is limited, there is a dearth of eligible companies. Moreover, within an industry there is sometimes only one qualified corporation, which makes analysis of the firm difficult because there is no industry benchmark and no comparison company. Another reason is the culture of Caribbean businesses. In general, many firms are reluctant to go public, as they perceive it as a loss of control of the enterprise, as well as potentially risky from a takeover standpoint. As such, most companies still source the bulk of their financing from banks and, to a limited extent, bond issuance. While the exchanges have been trying to address this perception, the number of new securities listing each year continues to be minuscule, if any at all.
A limited investor base is yet another concern in Caribbean stock markets, a hindrance for firms considering registering on the exchanges. The populations in these Caribbean countries under review are very small--Jamaica has the largest at 2.8 million, while Barbados has a mere 276,000 people. They have lower income levels compared to more established markets; the GDP per capita of Jamaica and Barbados are US$7,400 and US$19,300, respectively, relative to US$47,000 in the United States. (1) The combination of these factors significantly limits the investor pool of these stock markets. The JSE has made strides in addressing this constraint by aggressive marketing in developed markets and has succeeded in attracting a substantial number of foreign investors.
It is believed that creating one stock market will help to address these challenges. In addition to technical harmonization, one tangible step in this direction was the decision to allow securities listed on one exchange to also register on one or both of the other exchanges. The commencement of cross-border trading in April 1991 allowed investors in these countries to have access to cross-listed securities without having to enroll with all three exchanges, while the cross-listed securities would have the right to use funds from an expanded investor pool. The intention was to create a larger market without necessitating the formal combination of the three exchanges, even though the ultimate goal remains one stock market instead of three.
This policy has been in place for more than a decade and to date, to the authors' knowledge, no work has been done to study whether it has indeed fostered greater convergence. Research undertaken so far has focused on investigating market efficiency (see Craigwell and Grandbois 1999, Robinson 2001, and Alleyne and Craigwell 2007 who examined the BSE; Koot et al. 1989, Agbeyegbe 1994, and Robinson 2005, who looked at the JSE; Sergeant 1995, Singh 1995 and Bourne 1998, who investigated the TTSE; and Greenidge and Arjoon 2007, and Watson 2009 who analyzed all three exchanges). Some of these are quite dated but, except for Robinson (2001) and Watson (2009), all conclude that these markets show signs of weak form efficiency. No formal work has been found, however, that evaluates the convergence of these markets and this paper is the first in a series that attempts to shed some light in this area.
Research on stock market convergence in the literature usually approaches the topic by examining the betas for each market at the country level, i.e. the focus is on market betas and not individual stocks (as pioneered by Blume 1971). For example, Jayasuriya and Shambora (2008) in a recent paper investigated the relationship between eleven emerging stock markets and the U.S. stock market using country level betas estimated from a GARCH model, while Gangemi, Brooks and Faff (1999) considered the convergence of eighteen developed markets utilising cross-sectional, mean reversion-based estimations of OLS-obtained country betas. Studies on the convergence of stock betas across countries have been very limited.
The preference in the literature for country-level analysis is perhaps a result of little or no stocks that appear on all the markets under investigation. In the case of the sample of Caribbean countries used here, however, where the cross-listed stocks represent as much as 40% of the total number of listed securities, (2) it would be interesting to assess convergence through the behaviour of these common securities, which would in turn shed some light on the convergence of the three markets involved. Conceptually, it is expected that while the absolute returns earned for a given security on the three exchanges may differ, the direction should be the same (i.e. the return is anticipated to be positive/negative, regardless of the exchange) and the magnitude should be similar. Theore-tically, therefore, it is irrelevant which exchange the security is listed on. In reality, differences in the returns of the cross-listed securities are observed, but a simple mapping of the percentage changes in the market indices suggests some level of co-movement. Thus, it is yet unclear if the markets are converging.
This study attempts to develop the work on Caribbean stock market convergence by examining whether there is [beta]-convergence and [sigma]-convergence amongst the cross-listed securities on the three exchanges in question. The remainder of the paper will therefore present some stylised facts on the three markets under investigation; explain the methodology and data to be used; present the results of the study; and conclude with some suggestions for policymakers based on the findings.
Stylised Facts on the BSE, JSE and TTSE
Established in 1987, the BSE was the last of the three stock exchanges in the region to be formed. It followed the JSE (in February 1969) and the TTSE (in October 1981). The types of securities traded on the three exchanges include ordinary shares, preferred shares and fixed-income securities. However, the BSE and TTSE allow the trading of government securities while the JSE does not. Trades were initially done using a manual, open auction, outcry method two days per week--on Tuesdays and Fridays--and then three days a week when Wednesday was added. Effective March 2007, the BSE extended its number of trading days to a full working week, while the TTSE made the change in April 2008.
The number of companies listed on the stock market and the market capitalization-to-GDP ratio are two measures of stock market development. The number of companies listed on the BSE increased from 13 to 24 between 1989 and 2008. The growth of the other two exchanges was not quite as spectacular, with the number of firms on the TTSE rising from 36 in 1984 to 39 in 2008, while those on the JSE totalled 39 in 2008, slightly down from 40 in 1986. With respect to the market capitalization-to-GDP measure, the BSE also outperforms the other two exchanges, with a market capitalization-to-GDP ratio of 185.6% in 2008 versus 98.6% for the JSE and 49.2% for the TTSE. Furthermore, the rate of growth in market capitalisation of the BSE has surpassed that of the JSE and the TTSE.
Despite the faster growth of the BSE in terms of the number of companies and the market capitalisation ratio, Table 1 and Figure 1 show that on average the indices and market-capitalization ratios of the JSE and the TTSE grew at quicker annual rates, than those of the BSE. Furthermore, the JSE and the TTSE had higher average turnover ratios than the BSE. In contrast, the average rate of growth in the number of companies listed on the BSE exceeded those of the JSE and the TTSE. The above results indicate that the JSE and the TTSE are the more active exchanges and have experienced more rapid expansion. Recently, Alleyne and Craigwell (2007) also noted that the volatility on the BSE is relatively low, possibly due to the share ownership.
Methodology and Data Description
The methodology to be utilized in this study considers two measures of stock market convergence--[beta]-convergence and [sigma]convergence--which stem from the economic growth literature.
The use of [sigma]--convergence will allow the measurement of the degree of integration of the three stock markets and involves computing the cross-sectional dispersion in the return spread of the cross-listed securities. When the cross-sectional standard deviation of a variable is trending downwards, the degree of financial integration is increasing. If the standard deviation converges to zero, it implies that full integration has been achieved. The calculation to be employed is as follows:
where [g.sub.it] is the yield on asset i at time t and [g.sub.t] is the cross-section
[[sigma].sub.t] = [square root of ((1/N-1) [N.summation over (i=1)] [[log([y.sub.it]) - log([y.sub.t])].sup.2])]
mean yield at time t. For the purpose of this paper, N can equal either two or three, depending on the number of exchanges that the security is listed on. Note the measure of sigma-convergence is not a statistical measure and therefore provides no strong evidence against or in favor of integration between or among stock markets and cross-listed securities. In other words, a [[sigma].sub.t] of 0.8 may not be statistically different from zero. As a result this approach is complemented with the [beta]-convergence measure.
[FIGURE 1 OMITTED]
Once the degree of integration is established using [sigma]-convergence, [beta]-convergence will be used mainly to determine the speed of convergence of returns of the cross-listed securities by estimating the following regression:
where [R.sub.i,t] represents the return spread of specific assets (cross-listed
[DELTA][R.sub.i,t] = [[alpha].sub.i] + [beta][R.sub.i,t-1] + [L.summation over (l=1)] [[gamma].sub.l] [DELTA][R.sub.i,t-1] + [[epsilon].sub.i,t]
securities or market indices) between market i and the benchmark market at time t (the regression is run using each market as the benchmark market, given that there is no clear justification for the choice of one market over another); 1 is the lag order of the autoregressive component determined by the Schwartz model selection criterion and added to the standard growth model to avoid issues of serial correlation; D is the first difference operator; [a.sub.i] is the country-specific constant; and [e.sub.i,t] is the white-noise disturbance.
The dataset to be utilized is based on daily prices for each cross-listed stock for each market as well as the daily index for each market from 1998-2008, for which daily returns are calculated and then summarized into a monthly series. The start date for a particular series may differ from January 1, 1998, depending on when a stock is first listed on a given market (see Appendix 1). For example, a security may not have been registered on more than one exchange until May 2005. In that case, the commencement date of the series would be May 1, 2005. Figure I shows the development of the returns for the markets as well as the cross-listed securities. Note, as a result of the high frequency nature of the data set used in this paper, control variables could not be added to the models; thus the parameter estimates of the models could suffer from specification bias.
The company acronyms discussed in this section are defined in Appendix 1. Table 2 presents the findings for each combination of market con-vergence, while Figure 2 illustrates them graphically. The results suggest that there are signs of convergence, as evidenced by the reduction in the standard deviation for all three markets from an average of 0.43 in 1998 to 0.10 in 2008. There are also some differences with respect to the convergence of certain markets. The evidence implies that the level of convergence between the BSE and the TTSE was the highest from the beginning of the dataset and has steadily expanded over the sample period, while the JSE has increasingly converged with both the TTSE and the BSE over time.
The results for the cross-listed securities, presented in Tables 3 to 5 and Figures 3 to 5, suggest that the degree of convergence is not as strong as at the market level. Nonetheless, there are signs of convergence as most of the securities have experienced declining standard deviations over time across the markets that they participate in. The only exceptions are TCL and SFC (as listed on the BSE and the TTSE), whose standard deviations started to rise again in the latter half of the the last decade, and CCMB and DBG, where no clear pattern can be observed.
[FIGURE 2 OMITTED]
The findings for the [beta]-convergence models are presented in Tables 6 to 8. These models were subjected to several standard diagnostic tests, includ-ing no serial correlation, homoskedasticity and normal residuals, before being adopted. The beta regressions indicate that all combinations of benchmark markets and securities are significant. All of the values are negative, which implies that there is convergence of stock market returns. In general, the coefficients are close to one, suggesting that the speed with which the differences in return differentials are levelled is relatively fast. The only exceptions are DBG, RBTT and SFC, whose results have values above 2, indicating a much slower response rate than GKC and OCM, which are between 0.7 and 0.9.
When compared to studies that considered international markets, the results imply that the speed of convergence in the Caribbean region is relatively slow in comparison. The study conducted by Babetskii et al. (2007) found that the convergence between the new EU member states and the euro-zone countries was fast, averaging roughly one period (which in their case was one week). It would therefore be useful to re-run the regressions using weekly data, rather than monthly data, to ascertain whether there would be any notable differences in the results and to make the results more comparable to international studies.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
The results of this paper suggest that while the BSE, the JSE and the TTSE are showing signs of convergence at the market level, the evidence is less conclusive with respect to the cross-listed securities. This implies that it could be a situation where the economic development of the countries in question follow similar patterns, as the stock markets should be tied to the macroeconomy, but the stock markets themselves are not converging to a sufficient enough degree that the returns of the cross-listed securities convincingly suggest integration. This may be due to the weak efficiency found in previous studies as well as information asymmetries across the various markets, as posited by Leon, Nicholls and Noel (2001) in their study which, amongst other aims, attempted to determine whether volume is a sufficient proxy for information flow on the TTSE. To foster further integration, therefore, policymakers should endeavour to improve the efficiency of the markets as well as promote faster information dissemina-tion. Further work should be aimed at revising the dataset to reflect weekly values, as opposed to monthly values, given that stock markets have short memories.
Appendix 1: Cross-Listed Securities Code Name of Company BST Barbados Shipping and Trading Co. Ltd. CCMB Capital and Credit Merchant Bank Limited DBG Dehring Bunting & Golding Limited FCIB FirstCaribbean International Bank GHL Guardian Holdings Limited GKC Grace, Kennedy & Company Limited JMMB Jamaica Money Market Brokers Limited NCBJ National Commercial Bank Jamaica Limited NML Neal and Massy Holdings Ltd. OCM One Caribbean Media RBTT RBTT Financial Holdings Limited SFC Sagicor Financial Corporation TCL Trinidad Cement Limited Date First Appearing in Sample Code BSE JSE TTSE BST 01-Jan-98 Not listed 03-May-05 CCMB Not listed 26-May-03 21-Mar-07 DBG Not listed 01-Jan-98 02-Jan-07 FCIB 01-Jan-98 01-Jan-98 03-May-05 GHL Not listed 20-Sep-00 03-May-05 GKC 05-Oct-99 01-Jan-98 03-May-05 JMMB 29-Nov-05 02-Mar-03 03-May-05 NCBJ Not listed 01-Jan-98 03-May-05 NML 14-Dec-99 Not listed 01-Jan-98 OCM 24-Nov-06 Not listed 03-May-05 RBTT Not listed 28-Nov-01 02-Jul-07 SFC 04-Feb-03 Not listed 03-May-05 TCL 10-Nov-98 08-May-99 01-Jan-98
Agbeyegbe, T. 1994. "Some Stylized Facts About the Jamaican Stock Market." Social and Economic Studies 43 (4): 143-156.
Alleyne, K. and R.C. Craigwell. 2007. "Information Efficiency and Volatility in the Barbados Stock Exchange." Central Bank of Barbados (mimeo).
Babetskii, I., L. Komarek, and Z. Komarkova.Z. 2007. "Financial Integration of Stock Markets Among New EU Member States and the Euro Area." Czech National Bank Working Paper Series 7.
Blume, M. 1971. "On the Assessment of Risk." Journal of Finance 26 (1): 1-10.
Bourne, C. 1998. "Economic Aspects of the Trinidad & Tobago Stock Market." In Money and Finance in Trinidad & Tobago, edited by C. Bourne and Ramesh Ramsaran. Mona: ISER.
Craigwell, R.C. and C. Grandbois. 1999. "The Performance of the Securities Exchange of Barbados." Savings and Development 23 (4): 434-455.
Gangemi, M., R. Brooks and R.W. Faff. 1999a. "Country Betas, Their Time Sensitivity and The International Crash of October 1987." International Journal of Finance 11: 1390-1399.
--. 1999b. "Mean Reversion and the Forecasting of Country Betas: A Note." Global Finance Journal 10: 231-245.
Greenidge, K. and V. Arjoon. 2007. "The Efficiency of Caribbean Stock Markets." Central Bank of Barbados Working Paper.
Jayasuriya, S. and W. Shambora. 2008. "The World is Shrinking: Evidence for Stock Market Convergence." Economics Bulletin 7 (14): 1-12.
Koot, R., J. Miles and G. Heitmann. 1989. "Security Risk and Market Efficiency in the Jamaican Stock Exchange." Caribbean Financial Management 5 (2): 18-33.
Leon, H., S. Nicholls and D. Noel. 2001. "Non-linear Behaviour of Returns in an Emerging Stock Market." Central Bank of Barbados Working Paper.
Robinson, J. 2001. "Stock Price Behaviour in a Small Emerging Market: Tests for Predictability and sseasonality on the Barbados Stock Exchange." Savings and Development 1: 103-113.
--. 2005. "Stock Price Behaviour in Emerging Markets: Tests for Weak Form Market Efficiency in the Jamaican Stock Exchange." Social and Economic Studies 54 (2): 51-69.
Sergeant, K. 1995. "The Trinidad & Tobago Stock Exchange." In Insights Into an Emerging Financial Structure: The Experience of Trinidad & Tobago, edited by R. Ramsaran, 189-226. St Augustine: Caribbean Centre for Monetary Studies..
Singh, R.O. 1995. "An Examination of Return Predictability on the Trinidad & Tobago Stock Exchange." In Insights Into an Emerging Financial Structure: The Experience of Trinidad & Tobago, edited by R. Ramsaran, 227-244. St Augustine: Caribbean Centre for Monetary Studies..
Watson, P. K. .2009. "The Efficiency of the Stock Market in the CARICOM Sub-Region: An Empirical Study." Applied Financial Economics 19 (23): 1915-1924.
(1) CIA World Factbook 2008.
(2) The proportions of cross-listed securities on the BSE, the TTSE and the JSE are 33%, 40% and 26%, respectively.
Table 1: Average Annual Growth Rate for the Performance Indicators of the Regional Stock Exchanges: 1989-2008 Number of Stock Listed Turnover Exchange Companies Ratio MCAP Index BSE 4.9% 1.7% 25.5% 7.3% JSE 0.8% 5.9% 37.1% 30.9% TTSE 0.8% 5.9% 36.4% 13.8% Notes: MCAP is defined as market capitalization, that is, the value of all outstanding shares on the stock exchange. Turnover ratio is defined as turnover divided by market capitalization; turnover being the market value of all shares traded during the year. Table 2: Summary of Sigma Convergence Results for Market Indices BSE, JSE & TTSE BSE & TTSE JSE & TTSE BSE & JSE 1998 0.85 0.43 0.75 0.75 1999 0.40 0.30 0.27 0.24 2000 0.51 0.25 0.42 0.44 2001 0.44 0.23 0.44 0.28 2002 0.91 0.70 0.23 1.00 2003 0.40 0.24 0.21 0.45 2004 0.52 0.23 0.42 0.49 2005 0.30 0.14 0.24 0.21 2006 0.29 0.16 0.16 0.18 2007 0.20 0.10 0.12 0.13 2008 0.19 0.10 0.14 0.08 Table 3: Summary of [sigma]-Convergence Results for Securities Cross-Listed on the BSE & TTSE BST SFC OCM NML FCIB JMMB GKC TCL 1998 1999 0.23 1.29 2000 0.42 1.22 2001 0.43 0.99 2002 0.35 1.32 2003 0.45 0.48 2004 0.42 0.57 2005 0.20 0.34 0.39 0.05 1.10 0.49 0.66 2006 0.29 0.26 0.08 0.21 0.18 0.53 0.72 0.66 2007 0.36 0.11 0.24 0.14 0.22 0.48 0.23 0.22 2008 0.01 0.20 0.14 0.14 0.20 0.33 0.33 0.42 Table 4: Summary of v-Convergence Results for Securities Cross-Listed on the JSE & TTSE CCMB DBG GHL NCBJ RBTT FCIB JMMB GKC TCL 1998 1999 0.8 2000 0.9 2001 0.5 2002 0.5 2003 0.3 2004 0.6 2005 0.6 0.6 0.0 0.8 0.4 0.3 2006 0.7 0.6 0.4 0.5 0.4 0.5 2007 0.4 0.3 0.6 0.4 0.3 0.3 0.6 0.2 0.3 2008 0.4 0.5 0.5 0.1 0.3 0.3 0.3 0.3 Table 5: Summary of [sigma]-Convergence Results for Securities Cross-Listed on the BSE, JSE & TTSE FCIB JMMB GKC TCL 1998 1999 1.08 2000 1.90 2001 1.32 2002 1.96 2003 0.73 2004 0.76 2005 0.10 1.52 0.80 1.00 2006 0.57 0.94 1.09 1.36 2007 0.37 0.81 0.37 0.53 2008 0.36 0.58 0.58 0.79 Table 6: [beta]-Convergence Results Using JSE as Benchmark Market SECURITY TTSE BSE Markets Coefficient -0.956433 -1.023610 P-value 0.0000 *** 0.0000 *** FCIB Coefficient -1.163495 -1.447331 P-value 0.0005 *** 0.0000 *** JMMB Coefficient -1.329895 -0.869715 P-value 0.0003 *** 0.0044 *** GKC Coefficient -1.137638 -0.905985 P-value 0.0023 *** 0.0000 *** TCL Coefficient -0.996746 -1.037490 P-value 0.0000 *** 0.0000 *** CCMB Coefficient -0.997638 P-value 0.0378 ** DBG Coefficient -2.056688 P-value 0.0021 *** GHL Coefficient -1.326226 P-value 0.0001 *** NCBJ Coefficient -0.482185 P-value 0.1516 RBTT Coefficient -2.073634 P-value 0.0465 ** Notes: * Significant at the 10% level; ** significant at the 5% level; *** significant at the 1% level Table 7: [beta]-Convergence Results Using BSE as Benchmark Market SECURITY TTSE JSE Markets Coefficient -1.097069 -1.023610 P-value 0.0000 *** 0.0000 *** FCIB Coefficient -1.027376 -1.447331 P-value 0.0023 *** 0.0000 *** JMMB Coefficient -1.063046 -0.869715 P-value 0.0015 *** 0.0044 *** GKC Coefficient -0.822799 -0.905985 P-value 0.0041 *** 0.0000 *** TCL Coefficient -1.255710 -1.037490 P-value 0.0000 *** 0.0000 *** NML Coefficient -0.940669 P-value 0.0000 *** SFC Coefficient -2.514112 P-value 0.0000 *** OCM Coefficient -0.939452 P-value 0.0270 ** BST Coefficient -1.409618 P-value 0.0002 *** Notes: * Significant at the 10% level; ** significant at the 5% level; *** significant at the 1% level Table 8: [beta]-Convergence Results Using TTSE as Benchmark Market SECURITY BSE JSE Markets Coefficient -1.097069 -0.956433 P-value 0.0000 *** 0.0000 *** FCIB Coefficient -1.027376 -1.163495 P-value 0.0023 *** 0.0005 *** JMMB Coefficient -1.063046 -1.329895 P-value 0.0015 *** 0.0003 *** GKC Coefficient -0.822799 -1.137638 P-value 0.0041 *** 0.0023 *** TCL Coefficient -1.255710 -0.996746 P-value 0.0000 *** 0.0000 *** CCMB Coefficient N.a. -0.997638 P-value 0.0378 ** DBG Coefficient N.a. -2.056688 P-value 0.0021 *** GHL Coefficient N.a. -1.326226 P-value 0.0001 *** NCBJ Coefficient N.a. -0.482185 P-value 0.1516 RBTT Coefficient N.a. -2.051857 P-value 0.0061 *** NML Coefficient -0.940669 P-value 0.0000 *** SFC Coefficient -2.514112 P-value 0.0000 *** OCM Coefficient -0.939452 P-value 0.0270 ** BST Coefficient -1.409618 P-value 0.0002 *** Notes: * Significant at the 10% level; ** significant at the 5% level; *** significant at the 1% level