Spectator Demand in Russian Premier League

The compilation and data processing are provided in Excel since this is the most common method of working with datasets and it is very convenient to process data for further analysis. As each match is considered as the “dictionary”, all the matches are firstly gathered to the array, then processed in the Jupyter Notebook using “Pandas” packages, with further extraction as Excel file. Finally, the dataset will be revised and the gaps in the data will be filled manually. “Pandas” package is used for the processing of the data.

The data collection and processing methods are the subject to the limitation, which is: some data are not gathered due to the process is pretty laborious and the usage of additional data sources could rapidly increase the longitude of this stage because chosen methodology does not provide the high speed of the data gathering and takes a lot of time for installation. The methodology of data collection is not limited since one of the most appropriate methodologies used for data collection.

3.3 Data analysis

Basing on the analysis of the prior studies, Ordinary least squares (OLS) regression is used to perform this thesis's empirical analysis, due to this method is the most commonly used in demand-based studies' analysis of professional sports (Villar & Guerrero, 2009). The choice of this method also stems from the appropriate usage of this method in the multivariate model with several independent variables. Occupancy at game-level, which is used as a dependent variable, is continuous and its distribution is normal, so it permits the linear models to be used. Moreover, the analysis of the literature showed that unbiased coefficients can be generated by controlling the unobserved effects. To follow this requirement, the panel data should be used and the home clubs fixed effects should be definitely considered.

OLS regression is a technique of analysis that is used to estimate the dependent variable's dependence with one or several independent variables. The basic principle is the minimization of the sum of the squares in the difference between the values, predicted and observed. Basically, the method provides an opportunity to model the relationship between X and Y, where X is an explanatory variable and Y is response variable, with the line of best fit, which is predicting Y by X.

The analysis itself is conducted in the Jupyter Notebook using the programming language Python, applying “StatsModel” package. Usage of Python to conduct the analysis of this research stems from several reasons. Firstly, Python is one of the most used programming languages in the world, aiming for the performance of different types of tasks. Secondly, Python is the most used languages in the topic of the research. Moreover, usage of the Jupyter Notebook software package provides the convenience and the speed of the analysis, it affords to decrease the number of mistakes during the analysis. Finally, this software package was used in the data collection and processing stage, so it is more convenient to continue using the same software.

The package “StatsModel” is used for developing the model and the interpretation of this model. The usage of this package stems from the convenience of this package. The most popular package for such topic is “SkLearn” but since it is based on the machine learning and it does not allow to investigate the direction of the correlation. Hence, the analysis and the visualization of the results is provided in Jupyter Notebook.

The model of attendance estimation is the following:

where occupancy is the number of people who came to watch the match divided to the capacity of the stadium, kotimeh is the hour of the match's kick-off, meaning the beginning of the first time, month is the month in which the game took place, weekday is the number of the day of the week (0 for Monday, 1 for Tuesday, 2 for Wednesday, 3 for Thursday, 4 for Friday 5 for Saturday and 6 for Sunday), htgsl5 is the average number of goals scored by the home team in last 5 matches of the league, atgsl5 is the average number of goals scored by the away team in last 5 matches, htgcl5 if the average number of goals conceded by the home team in last 5 matches, atgcl5 is the average number of goals conceded by the away team in last 5 matches, htsl5 is the average number of shots of the home team in last 5 matches, atsl5 is the average number of shots of the away team in last 5 matches, htadsl5 is average number of areal duel successes of the home team in last 5 matches, atadsl5 is average number of areal duel successes of the away team in last 5 matches, htdwl5 is average number of dribbles won of the home team in last 5 matches, atdwl5 is average number of dribbles won of the away team in last 5 matches, httl5 is the average number of tackles of the home team in last 5 matches, attl5 is the average number of tackles of the away team in last 5 matches, htfe is the fixed effect of the home team, which is estimated using one hot encoding . The hour of kick-off is used instead of the kick-off itself because of convenience in further analysis in Jupyter Notebook.

The average of last 5 matches in the league in general, not just home matches, stems from the supposition that people who attend matches usually are aware of the situation in league matches in general, so they either watch the away games broadcasts or just follow the results. Again, the decision of this variable is based on the previous studies (Coates, Naidenova & Parshakov, 2017; Villar & Guerrero, 2009) excluding the dependent variable occupancy. The majority of the studies use attendance as a dependent variable but using occupancy is considered more relevant since the capacity of the stadium is estimated. For example, Coates, Naidenova and Parshakov (2017) used capacity as the independent variable but it seems irrelevant since attendance is always dependent on the capacity and, finally, it showed the positive signs in the regression. So then, the occupancy is used as the dependent variable. The code of this stage is available at https://drive.google.com/open?id=1lzULxeePBdYgXWt9YyxIdi9SfkNuBFuH.

3.4 Conclusion

The chosen methodology provides the possibility to gather, process and analyze the data in the most appropriate way, since the choice is based on the prior studies, considering the alternative modern tools for fastening and increasing the accuracy of the processes. Generally, the research design is not contributing to the novelty of analysis methods but it is based on the strong foundation of the common methodology for the area of study. Also, the methodology is relevant to the research question because the regression analysis allows to directly investigate the significance of each determinant and to define the ones which have the most convincing impact.

In fact, the methodology is the subject to several limitations, but these limitations are not crucial and are not significantly influencing the results of the thesis. Finally, the proposed methodology should be considered relevant.

Results

The analysis of the current attendance situation is made based on the scraped data on attendance on a seasonal basis. These data contain the general seasonal and the average per match statistics of attendance for 17 years, form season 2002 to season 2018/ 2019. The data used are shown in Table 1.

Table 1 The data on seasonal attendance in 2002 - 2018/2019 seasons*

*Source: Russian Premier Liga Statistics, n.d.

For a better understanding of the attendance changing trends, the visualization is provided in Figure 1.

Figure 1. The seasonal and average attendance of the Russian Premier League matches in 2002 - 2018/2019 seasons

The peak of the seasonal attendance (4522405 attendees) falls on the 2011/2012 season and is explained with the expanded number of matches (44 instead of 30) so it should not be considered in the estimation of attendance situation. The average attendance shows more appropriate values. The figure shows that the trend of attendance increase was from the season 2002 to 2007's, then the stagnation period started, continuing with a small leap in 2011/2012 seasons, while the “spring-autumn” system of scheduling was introduced, and it had been finished with a significant decrease by 2014/2015 season, reaching the minimum value of 10151 attendees per match averagely. Then the great increase is observed, both in seasonal and average per match values. Attendance rates of the 2018/2019 are the peak of both values (3869843 seasonally and 16680 average per match).

Further, the data on a match-by-match basis was successfully scraped and processed. Generally, the data on 1431 matches were collected and then was extracted to the txt-file and then to xlsx-file for visibility. Then the data was processed to 1391 matches since the first 5 matches were used as a basis for introducing the variables on the on-field performance in last 5 matches. Then there were 31 matches excluded since there were gaps in data (e.g. 10 matches were played without fans in case the clubs were punished for several matches). Further the “htadsl5”, “htdwl5”, “htsl5”, “httl5”, “htgsl5”, “htgcl5”, “atadsl5”, “atdwl5”, “atsl5”, “attl5”, “atgsl5”, “atgcl5” variables were introduced basing on the mean value of the “htads”, “htdw”, “hts”, “htt”, “htgs”, “htgc”, “atads”, “atdw”, “ats”, “att”, “atgs”, “atgc” for previous 5 matches of a team respectively. The columns weekday and kotimeh were also extracted from the basic dataset for further analysis. The final dataset was not extracted to xlsx-file since it was no need for it. The data was farther been used in Jupyter Notebook software.

The basic descriptive statistics of all the used determinants are shown in Table 2. The mean value of the home and away team aerial duel success in last 5 matches 0.4994 and 0.5003 mean that averagely 49.94% and 50.03% of the aerial duels are won by home and away teams respectively in last 5 matches. Estimating the occupancy of the stadiums, there were several cases when the attendance was even higher than a capacity, which is described with the equipment of additional seats, that is why the maximum occupancy if more than 1, meaning more than 100% of the stadium. The attendance varies from 700 to 78360. Since the months are named with the numbers which they mean (e.g. December is 12), the minimum number of “month” determinant explains that no matches are played in January and February. The mean value of occupancy, which is 0.4648, demonstrates the base for potential growth, the minimum value of 0.036 is described as the 3% of occupancy of the least attendant match, showing the great necessity of revising and developing of such cases.

Table 2 Descriptive statistics

The developing of the ordinary least squares regression model provided the results, the presentation of which begins with the general estimation of the results, which is shown in Table 3.

R-squared coefficient describes which of the dispersion's percentage is actually explained with the current model. In the case of the proposed model, this coefficient is 45.7%, thus it could be estimated appropriate for such type of research.

In the categorical variables, one value was taken as the basic, in “month” it was the March (3), in “weekday” it was Monday (0). Thus, basing on the regression output, July (7), October (10), November (11) and December (12) statistically differ from the basic value (p-value is lower than 5%). Besides, the в-coefficients are negative for the last 3 months of the year which means that the occupancy of the stadiums is much lower compared to the basic value, whereas the effect is increasing directly with much more significant impact in December than in November, and in November than in October (-0,12 is more significant than -0.07, -0.07 is more significant than-0.04). July (7) is on contrary positively significant for attendance (0.07). Such a result was expected.

“Weekday” is also categorical variable and here the p-value of [T.5] (Saturday) and [T.6] (Sunday) is less than 5% and it means that в-coefficients are not equal to zero. Since it is positive - the dependence is direct. The matches scheduled on weekends are attended by more people. The result also was expected.

Table 3 Estimation results

Considering the results of on-field determinants analysis, the findings are not obvious. There are several results which were expected to be significant like the average goals scored in the last 5 matches of both home and away teams. Whereas, the model indicates that a number of goals scored by the home team are insignificant, while the number of away team's goals scored has positive significance. A number of tackles for both home and away teams is insignificant. Both home and away teams' number of shots are positive and significant, whereas the average number of conceded goals is negative, meaning the more goals a team conceded averagely in last 5 league matches, the fewer people come to watch the next game. Interestingly, the other determinants connected with away team's on-field performance (atadsl5, atdwl5) are statistically significant and positively influence the attendance. Besides, away team's aerial duel success has the most powerful influence on attendance (0.19). Since the variables “htdwl5”, “atdwl5”, “htsl5”, “atsl5”, “htgsl5“ and “atgsl5” are defining whether the game will be more or less spectacular, it was expected that these variables would be significant due to conclusions of Parshakov (2017) that more fans attend a match if the game is anticipated to be spectacular. Suchwise, only “htdwl5” and “htgsl5” are found insignificant (p-value equals to 77% and 5% respectively), the others have a positive impact on attendance.

These results can be interpreted as fans intend to partially base the decision whether to go or not to go to the stadium on the performance of the teams in last 5 matches, estimating the average number of away team's dribbles won, scored goals and aerial duel success, number of shots and number of conceded goals of both teams. It was supposed that these factors impact the potential spectacle and showmanship of a game. The insignificance of the home team's dribbles won, the home team's aerial duel success and goals scored variables could be interpreted with the expected quality of the away team.

Moreover, the results indicate that the statistical significance of the kick-off time hour variable is found positive. This factor is sure to be taken into consideration. Implementation of the home club's fixed effect provided the unbiased coefficients. The estimation of the significance of fixed effects provides interesting results. Positive statistical significance of CSKA Moscow, spartak Moscow, and Zenit could be explained with the brand of these teams, whereas the same or even higher significance of Arsenal Tula, Orenburg, Rostov, and Ural could be interpreted by the regional significance of these teams and some specific policy on the clubs' promotion. The negative significance of Anzhi, Krylya Sovetov, Torpedo Moscow, Tosno, Urozhay, Lokomotiv Moscow, Rubin, and Volga could be explained by the averagely poor performance of these teams in last seasons. The insignificance of the rest football clubs could be interpreted with the low brand strength of these clubs.

The H1, which stated that the matches, which are scheduled closely to mid-season break are attended in a lower case than the others (December, March), is not confirmed. The analysis's results showed that scheduling the matches in March has not statistically significant influence the attendance, while the same cannot be said for October, November, and December. Matches in these months are attended with a lower number of people.

The second hypothesis, which supposed that the day of the week and the kick-off time determinants are significantly influencing the attendance, is confirmed. Table 3 demonstrates that the attendance is dependent on the day of the week determinant, showing that scheduling the matches during the weekends provides the increase in attendance, compared to the weekdays. Also, the dependence of the kick-off time is presented, showing that the earlier the match is scheduled the fewer people tend to go to the stadium to watch the game.

The H3, which was predicting the significance of a team's on-field performance is partially confirmed, the visualized data analysis results show the importance of 7 out of 12 determinants of the on-field performance of a team in the last 5 home matches. The suppositions that the majority of fans usually decide whether to go or not to go to the stadium basing on the performance of the team in last 5 matches have been partially confirmed.

According to the literature of the field of study, the results confirm the findings of Parshakov (2017) and Coates, Naidenova and Parshakov (2017), which stated that the away team's brand is more important in estimating attendance than the home's one. Also, the findings on the on-field performance and the attendance's dependence on the month and kick-off time correspond to the findings of the majority of the foreign tournaments' researches (Villar & Guerrero, 2009).

A large amount of literature in the field of study provides great opportunities to get relevant observations basing on this literature. The first objective of the study, the understanding of the current situation of attendance in the Russian Premier League, basing on the data on attendance on a seasonal basis, has been achieved successfully. The visualization of the data provided the opportunity to estimate the nowadays case.

Further, the determinants, influencing the attendance the most significantly, are defined. First of all, the revised literature provided an understanding of which factors should be taken into account. Hence, the experience of the prior studies was combined with the availability of the data for the Russian Premier League's matches to define and collect the necessary data.

The stage of the collection and processing of the data has been finished further. The usage of Jupyter Notebook software with Python's packages “Request” and “BeautifulSoup” provided the speed and accuracy of data collection and “Pandas” package allowed to process all the scraped data to the single dataset, fix all the errors, introduce new variables. Filling in the gaps was done manually, using the data from the Russian Premier League official website. Also, the data was partially incorrect since there were 6 cases when the attendance was exceeding the capacity of the stadium more than 1.1 times. Such errors were also fixed manually. The matches with spaces in data were excluded.

Hence, the analysis of the data has taken place. Usage of the Python's package “StatsModel” let expanding the possibilities of the usual for such analysis methodology, Stata and R-Studio, providing the freedom in choosing the additional methodologies. Thus, the main findings from the data analysis are as follow:

- The matches in October, November, and December are less attendant, with the negative increasing trend.

- Kick-off time has positive statistical significance on the attendance

- Matches played in weekends are more attendant that weekday's

The H1 is rejected, the influence of scheduling in March is found insignificant, whereas October, November, and December are negatively significant. H2 is confirmed since the influencing of kick-off time and weekday is considered important. H3 is partially confirmed because the findings contradict the supposition that “htadsl5”, “htdwl5”, “httl5”, “attl5”, “htgsl5” influence attendance rates.

The significance of independent variables is demonstrated in the regression output which is the result of the OLS regression analysis. Also, after the analysis is conducted, the evaluation of the descriptive potential has become possible using R-squared coefficient, so it equals to 45.7%. Eventually, the provided results make it possible to conduct the recommendations to provide the management of the above-mentioned football institutions in Russia with up-to-date information on the determinants actually influencing demand in nowadays case of Russian Premier League and the ways of contributing to the improvement in such situation.

Recommendations

Since attendance is much higher during the spring and summer, which cannot be said about autumn and winter months, the scheduling should be reconsidered and discussed. The trend of attending matches less with an approximation of the January could be explained with the low temperatures outside. The scheduling of the matches more tightly during the spring and summer periods could provide the possibility to decrease the share of matches played in cold seasons. Since top 5-6 teams of the previous season take place in European Cups, it seems impossible to put additional matches in this period, last matches of the spring are usually played once a week. Due to the fact that Russian teams rarely reach the Ѕ finals and finals of the European tournaments, some of the matches could be added in this period. Also, the summer break should be reduced. For example, the break between seasons 2017/2018 and 2018/2019 was 76 days (Russian Premier Liga Statistics, n.d.). The shortening of this break just for 7 days gives the opportunity to exclude the December from the tournament schedule. Thus, it is expected that changing the tournament's calendar for the next season due to the findings of this study would increase attendance rates. Also, while forming the tournament calendar for the 2019/2020 season, the estimation of the day of the week is significant. The scheduling on weekends is preferable.

Basing on the estimated on-field performance variables' significance, the clubs of the Russian Premier League could pay more attention to the characteristics of the potential player, which are providing the opportunity to increase the rates of this index during the match. For example, the number of tackles insignificance and number of shots significance of both teams playing could indicate that fans prefer watching the games with fewer duels and more attacked reached the shot, so the decision between technical and powerful player could potentially be made towards the former. Same for the decision of a new head coach, the proposed game philosophy should be chosen based on these factors.

The critical analysis is provided and the recommendations for the football managers in different Russian institutions are given based on the findings from the empirical analysis of this thesis.

Conclusion

Analyzing the results of the empirical analysis, several important findings are provided. First of all, the current situation with the attendance of the RPL matches is discussed. The significant decrease in attendance started from 2016/ 2017 season and is explained with the commissioning of the new stadiums, built in Russia for the FIFA 2017 Confederations Cup and the FIFA 2018 World Cup. Generally, hosting of such events provides increases in the overall interest for football matches, and the development of the infrastructure and commissioning of the new stadiums boosted the attendance rates (PwC, 2018a; PwC, 2018b). Thus, the season 2018/2019 has become the record-breaking in attendance rates even since it has not finished yet. But, since no significant premises for further increases are observed in the previous periods, so it is expected that the pace of increase would slightly slow down. Hence, the situation of the Russian Premier League's matches attendance seems favorable for developing the new strategies to be implemented during the opportune period.

Further, the hypotheses are revised. Thus, the rejection of the H1 could be interpreted with the influence of the outside temperature on the fans intention to attend a match. Meanwhile, the confirmation of the H2 stems from the usual working days of the citizens, so it is more convenient for fans to have matches in weekends since they are able and are having the intention to attend them. The H3 is not completely confirmed but the findings show the important results which provide the understanding of whether a factor has significant influence.

Since the main objective of the study is to create the recommendations of changes in RPL clubs', the league's strategies, these recommendations are created based on the results of the analysis.

Suchwise, the main findings of this study are to be discussed. Surely, there was some obvious correlation found, like the dependence of the month in which the match is played with attendance or the significance of the day of the week and kick-off time. Also, the on-field performance estimation has provided several findings. For example, the importance of the average last 5 matches performance of the away teams. 5 out of 6 variables estimating the performance of the away team are found statistically significant in increasing the attendance of matches, meanwhile, the home team's effects controversially have a weak impact on attendance rates (2 out of 6 are significant). The significance of some teams' fixed effects could be explained with the brand strength of these clubs. The unexpected significance of Orenburg's and Rostov's fixed effects in line with top Russian clubs like Zenit provides an opportunity for further investigation of the potential regional effect.

The results of this study are to be beneficial for several different institutes. First of all, the organizational structure of management of Russian football named all the football clubs of the Russian Premier League and the management of RPL itself. This is true just if the recommendations, given in this research “Results” section will be implemented. Increase in attendance, which will be caused by changes based on this thesis's results, will impact football clubs, as commercial organizations, which will increase profits from matchday gate attendance. Football fans, being customers of a team's product, would possibly get better atmosphere stadium due to an increased number of attendees and development of entertainment zones in the stadium area. Citizens of Russian Premier League teams' home cities would possibly experience infrastructure developments since the administration of these cities will be forced to provide fans with a convenient way to attend football. Increased number of away team's supporters from other cities or even countries (due to prolonged participation in the European international contents, more abroad fans would visit Russia) would provide these developments.

Management of the league will also be the beneficiary in case the popularization of the league will increase its own revenue due to new sponsorship contracts. Moreover, since low attendance rates, one of the most topical problems to be solved in today's Russian football, also, one of the top the nowadays issues in Russian football will get the basis for solution, so the others will get more attention Globally, the potential attendance boost could influence a great number of areas, e.g. increasing the profits from ticket sales will influence changings in revenue structure of Russian clubs, making them less dependent on the donations and sponsorship contracts, finally making them profitable. Also, such an increase in peoples wish to attend the stadium will contribute to the popularization of football and sports in general in the country, making the nation healthier. Thereby, the development of clubs' and the league's policy on attendance control possibly could influence the great number of factors, which are interdependent.

From the scientific perspective, this study contributes with one of the first demand-based studies investigating the determinants of attendance in the Russian national championship, developing the new dependences discovered during the analysis. This thesis can serve a base for future researches in this field in line with the papers of Coates, Naidenova, and Parshakov (2017) and Parshakov (2017), providing some tested hypotheses to be confirmed or refused.

The research is subject to a few limitations.

- Perhaps, the results of this research could be considered not transferable to other abroad tournaments due to unique features of the Russian Premier League. This is a common limitation for studies conducted for developing markets due to differences in further development. Also, the results are hardly transferable to lower divisions in Russia due to the great development gap between RPL and lower divisions.

- The lack of periodical data lowers the relevance of key findings. Research, which is based on the minimum number of variables is not providing the significance, the enlargement of variables leads to a decrease in the period of available data. This study is the example of a balance between the number of variables to be investigated and the enlargement of attendance data to be considered. The limitation's core is the absence of collected data. Future research could overcome this limitation by getting insider information from the league's management directly to expand the dataset or through finding the other alternative sources of data for gathering and analyzing.

- A number of determinants are not investigated since the data is absent. Ticket prices and broadcasting data would have improved the relevance of results. Seems that the only possible way to reduce such limitation is to use the period when these data are available. Possibly, it could happen in near decades. Moreover, there are a great number of variables, which were considered by the prior studies, can potentially be considered due to the Russian Premier League's case.

- The questionnaire could be included in the research to define the unobvious determinants to be analyzed. Basing just on the literature conducted in the area of research limited this study with just already observed determinants. Using a questionnaire could expand the list of variables for analysis.

Using different methodologies most likely would not provide the difference in results since the area of study is being properly researched since the middle of the 20^th century and the researchers used a great number of methods to conduct demand-based studies, testing the different methodology and choosing the most appropriate. Probably, the usage of fundamentally new methods could make a difference, but the relevance of these findings is arguable. Finally, the determinants for analysis are defined through examining the literature in the field of study, the collected data is analyzed through the proposed methodology. The limitations of the data and methods used for this research has not negatively impacted on the study. Generally, all the tasks of the research are performed and all the objectives are achieved.

Since the topic of this research intends to analyze several determinants together, future researches can concentrate on the specific determinant to be deeply analyzed. This would improve the relevance of the research, confirming of refusing the significance of different variables. Further, based on the estimation of these variables specifically for the Russian Premier League's case, the current research's analysis could be repeated for testing the relevance of the results. Also, the same type of research could be conducted for a league in the developed sports market in order to predict what determinants theoretically should be paid extra attention at during the adjustment of the organizations' strategies in Russia in near years. This approach would contribute to expanding the recommendations and would let the clubs, the league and the RFU create the strategic plans for a longer period of time. Moreover, gaining the data lacked in this research would surely be a very strong base for contributing to the area of study, covering the gap in this research. Estimating ticket prices and TV-broadcasting variable would probably provide unexpected results, which are impossible to predict nowadays.

Reference list

1. Arnold, R., & Veth, K. M. (2018). Racism and Russian Football Supporters' Culture. Problems of Post-Communism, 65(2), 88-100. doi:10.1080/10758216.2017.1414613

2. Bird, P. J. (1982). The demand for league football. Applied Economics,14(6), 637-649. doi:10.1080/00036848200000038

3. Borland, J., & MacDonald, R. (2003). Demand for sport. Oxford Review of Economic Policy, 19(4), 478-502. doi: 10.1093/oxrep/19.4.478

4. Buraimo, B. (2008). Stadium attendance and television audience demand in English league football. Managerial and Decision Economics, 29(6), 513-523. doi: 10.1002/mde.1421

5. Buraimo, B., Forrest, D., & Simmons, R. (2009). Insights for clubs from modelling match attendance in football. Journal of the Operational Research Society, 60(2), 147-155. doi: 10.1057/palgrave.jors.2602549

6. Buraimo, B., Tena, J. D., & de la Piedra, J. D. (2018). Attendance demand in a developing football market: the case of the Peruvian first division. European Sport Management Quarterly, 18(5), 671-686. doi: 10.1080/16184742.2018.1481446

7. Charyev, G. (2014). Consequences of the limit on foreign players in Russian football. Soccer & Society,17(4), 571-587. doi:10.1080/14660970.2014.980727

8. Coates, D., Naidenova, I., & Parshakov, P. (2017). Determinants of Russian Football Club Brands. International Journal of Sport Finance, 12(4), 321-341.

9. Demmert, H. G. (1973). The economics of professional team sports. Lexington, Mass: Lexington Books.

10. Dobson, S., & Goddard, J. A. (2001). The economics of football. Cambridge: Cambridge University Press.

11. Downward, P., Dawson, A., & Dejonghe, T. (2009). Sports economics. Routledge.

12. FIFA.com. (n.d.). More than Half the World Watched Record-Breaking 2018 World Cup. Retrieved from https://www.fifa.com/worldcup/news/more-than-half-the-world-watched-record-breaking-2018-world-cup

13. Feehan, P., Forrest, D., & Simmons, R. (2003). Premier League Soccer: Normal or Inferior Good? European Sport Management Quarterly,3(1), 31-45. doi:10.1080/16184740308721938

14. Forbes (n.d.). Manchester United on the Forbes Soccer Team Valuations List. Retrieved May 5, 2019, from https://www.forbes.com/teams/manchester-united/#74e4fd0c13f9

15. Forrest, D., & Simmons, R. (2006). New issues in attendance demand: The case of the English football league. Journal of Sports Economics, 7(3), 247-266. doi: 10.1177/1527002504273392

16. Garcнa, J., & Rodrнguez, P. (2002). The determinants of football match attendance revisited: Empirical evidence from the Spanish football league. Journal of Sports Economics, 3(1), 18-38. doi: 10.1177/152700250200300103

17. Global attendances. (n.d.). Retrieved February 18, 2019, from http://www.sportingintelligence.com/finance-biz/business-intelligence/global-attendances/

18. Jere, M. G., & Mathidza, S. (2014). Investigating motivational factors that influence football match attendance in the South African Premier Soccer League post the 2010 FIFA World Cup. Mediterranean Journal of Social Sciences, 5(20), 563-569. doi: 10.5901/mjss.2014.v5n20p563

19. Martins, A. M., & Cro, S. (2016). The Demand for Football in Portugal. Journal of Sports Economics, 19(4), 473-497. doi:10.1177/1527002516661602

20. Medvedev, D. (2015). A new reality: Russia and global challenges. Russian Journal of Economics, 1(2), 109-129. doi: 10.1016/j.ruje.2015.11.004

21. Neale, W. C. (1964). The peculiar economics of professional sports. The Quarterly Journal of Economics, 78(1), 1-14. doi: 10.2307/1880543

22. Noll, R. G. (2002). The economics of promotion and relegation in sports leagues: The case of English football. Journal of Sports Economics, 3(2), 169-203. doi: 10.1177/152700250200300205

23. Parshakov, P. (2017). Brands or Uncertainty? An Empirical Test of the Uncertainty of Outcome Hypothesis in Russian Football. SSRN Electronic Journal. doi:10.2139/ssrn.2978587

24. PwC. (2018). Russian Football Premier League. A Comprehensive Study of the Economics of Russian Football. Retrieved from PwC Website: https://www.pwc.ru/ru/assets/pwc-rfpl-research-2018.pdf

25. PwC. (2018). Russian Premier League. A Comprehensive Study of the Economics of Russian Football. Season 2018/2019. Retrieved from PwC Website: https://www.pwc.ru/en/sports/assets/pwc-rpl-research-season-18-19-en.pdf

26. RPL (n.d.). About Russian Premier Liga. Retrieved March 18, 2019, from https://eng.premierliga.ru/rfpl/championat/

27. Ro?ca, V. I. (2018). The Influence of European Cup Performances on Domestic Stadium Attendances in Romanian Football, Proceedings of the International Conference on Business Excellence, 12(1), 875-884. doi: 10.2478/picbe-2018-0078

28. Rottenberg, S. (1956). On choice in labor markets. ILR Review, 9(2), 183-199. doi: 10.1177/001979395600900201

29. Russian Premier Liga Statistics. (n.d.). Retrieved May 3, 2019, from https://eng.premierliga.ru/tournaments/championship/

30. Salaga, S., & Tainsky, S. (2015). Betting lines and college football television ratings. Economics Letters, 132, 112-116. doi:10.1016/j.econlet.2015.04.032

31. Szymanski, S., & Smith, R. (1997). The English football industry: profit, performance and industrial structure. International Review of Applied Economics, 11(1), 135-153. doi: 10.1080/02692179700000008

32. Szymanski, S. (2003). The economic design of sporting contests. Journal of Economic Literature, 41(4), 1137-1187. doi: 10.1257/002205103771800004

33. UEFA.com. (2018). Club competition access list 2018/2019. Retrieved March 15, 2019, from https://www.uefa.com/MultimediaFiles/Download/competitions/General/02/49/98/51/2499851_DOWNLOAD.pdf

34. UEFA.com. (n.d.). Member Associations - UEFA Coefficients - Country Coefficients. Retrieved February 20, 2019, from https://www.uefa.com/memberassociations/uefarankings/country/#/yr/2019

35. Villar, J. G., & Guerrero, P. R. (2009). Sports attendance: A survey of the literature 1973-2007. Rivista di Diritto e di Economia dello Sport, 5(2), 112-151. Retrieved from https://www.researchgate.net/profile/Placido_Rodriguez/publication/46508178_Sports_Attendance_A_Survey_of_the_Literature_19732007/links/562d08e008ae518e3482476d/Sports-Attendance-A-Survey-of-the-Literature-19732007.pdf

36. WhoScored.com. (n.d.). Premier League tables, rankings & standings. Retrieved May 11, 2019, from https://www.whoscored.com/Regions/182/Tournaments/77/Russia-Premier-League

Размещено на Allbest.ru