Red Blood Cell Folate Levels in Canadian Inuit Women

This article discusses the study on Red Blood Cell Folate Levels in Canadian Inuit Women of childbearing years. It highlights the empirical strategy, challenges confronted by the researchers, and the credibility of the results. The article also suggests potential future frameworks to strengthen the empirical strategy.

Running head:
Name of the Student:
Name of the University:
Author Note:

1
Table of Contents
Answer to question 1..................................................................................................................3
Introduction............................................................................................................................3
Discussion..............................................................................................................................3
The literatures related to red-blood cell folate in human body..............................................4
Prevalence of anemia among Inuit women in Nunavik, Canada.......................................5
Prevalence and correlates of high red blood cell folate concentrations in the Canadian
population using 3 proposed cut-offs.................................................................................5
Empirical challenges confronted by the researchers..............................................................6
Results credible in terms of providing a causal interpretation of the estimates.....................8
Conclusion..............................................................................................................................9
Answer to question 2................................................................................................................13
Introduction..........................................................................................................................16
Data Description...................................................................................................................17
Methodology........................................................................................................................19
Robustness check.................................................................................................................22
Results..................................................................................................................................23
Conclusion............................................................................................................................25
Reference..............................................................................................................................27
Appendices...........................................................................................................................29
Appendix 1: Missing values and descriptive statistics of variables.................................29
Appendix 2: Graphical presentation of Variables............................................................30

2
Appendix 3: Analysis to find the relationship between Beta and quality of manager.....35
Appendix 4: Test of results validity.................................................................................36

3
Answer to question 1
Introduction
The current assignment focuses on brief study of article “Red blood cell folate levels
in Canadian Inuit women of childbearing years: Influence of food security, body mass index,
smoking, education and vitamin use”. This article was authored by K. Duncan, A. Erickson,
G. Weiler and L. Arbour in the year 2018. The article was published in Canadian Journal of
Public Health. The paper aims at discussing on the contribution made by the article. The
paper also highlights the previous literature of the paper in order to understand the literature
gap in the study. The main study of the paper also demonstrates the empirical strategy of the
paper and the empirical challenges are confronted in the paper by the researches addressing
the research question of interest. The study demonstrates strategies that have been applied by
the author in order to address the main empirical challenges. As per the tools Economics 326,
in order to verify whether the result of the paper is credible in terms of proving a casual
interpretation of the estimates. The other evidence are described and addressed that were used
by the author in order to build creditability of result of the paper. In addition at last the paper
suggests potential future framework in order to strengthen the empirical strategy.
Discussion
The paper basically on a research is to determine whether the red-blood cells folate
levels of Inuit women reached accepted target levels. The Vitamin-B, which is also known as
folic acids helps to regulate red-blood cells in the human body. When this folic acid content
is found to be low in the body then, the body is unable to make more red blood cells and
ultimately which causes anemia to the individual. It is evident that the major population in
Inuit of Canada, where the vitamin consumption is found to be low and access to folate-rich
foods is limited, then the fortification is likely to be the major source of intake. In order to

4
carry out the task the Inuit Health Survey in the year 2007-2008 did research on the following
issue which included the evaluation and test regarding the level of red blood cell folate in
women. For the study the sample size was 249, were selected by random sampling technique
specially the non-pregnant women of reproductive age (Duncan et al. 2018). Using the linear
regression analysis and descriptive statistics, the red-blood cells folate level were compared
and assessed across the various demographic variables in order to evaluate the various
characteristic associated with red-blood cells folate.
The contribution which was made by this paper can be analyzed by the result of the
study, it was found that the level of red-blood cell folate 935.5 nmol /L (± 192) which
reached to the proposed target level (> 906 nmol/L), but it has also been found that around
47% of women had lower level of red-blood cells folate than the proposed target level. It is
observed that increased level of smoking in women has resulted a negative impact on the
level of red-blood cell folate (− 5.8 nmol/L per cigarette smoked per day (p = 0.001). It was
estimated that around 6.8% of women reported that taking the vitamin supplements has
resulted a positive outcome on the level of red-blood cells in women (Duncan et al. 2018).
The study deferred that still there higher percentage of women who are dealing with low level
of red-blood cell folate and very snall number of women are taking vitamin supplements. It is
suggested that the folate status is found to be low in Inuit women of reproductive age women.
From the above research initiatives were taken by Inuit of Canada to improve the food
security, vitamin supplements, folate rich traditional food, and culturally relevant education
to the people of Canada and smoking cessation reduction program will improve and benefit
the women of Inuit to overcome health issues and have a healthy life (Jamieson et al. 2012).
The literatures related to red-blood cell folate in human body
The below literature demonstrates the previous researches done on this particular issue:

5
Prevalence of anemia among Inuit women in Nunavik, Canada
The related article on red-blood cell folate was published in the year 2011 by Céline Plante,
Carloe Blanchet, Lous Rochette & Huguette Turgeon O’Brien on “Prevalence of anemia
among Inuit women in Nunavik, Canada”. The main objective of above article was to assess
the prevalence and the types of anemia that are present among the non-pregnant Inuit women
of Nunavik, Canada. From the above research it was found that the prevalence of anemia in
Nanavik women is quiet similar to the level in non-industrial countries and also represented
severe public health issues. The women in Nunavik were found to have sources of heavy
metals such as iron, in their diet which is probably could explain the positive association that
is found between iron status and heavy metals (Colapinto et al. 2015). The study failed to
analyze the main cause of the anemia and at what level the women population in Inuit of
Canada is deprived of red-blood cells and the main cause of low percentage in red-blood
cells.
Prevalence and correlates of high red blood cell folate concentrations in the Canadian
population using 3 proposed cut-offs
The following research paper “Prevalence and correlates of high red blood cell folate
concentrations in the Canadian population using 3 proposed cut-offs” The research is done by
Cynthia K.Colapinto in the year 2015. The paper was a study on district shift towards the
higher level of folate level in the people of Canada. The higher level of folate is known to
have good benefits which involve prevention of neural tube defects; however there is certain
concern which has been raised regarding the certain health outcomes. The aim of the research
was to propose the limit for highlevel of red blood cells folate concentration. According to
the research the high red blood cells folate was found to be 16%, 6%, and 2% at thresholds of
1450 nmol/L, 1800 nmol/L, and 2150 nmol/L, respectively (Plante et al. 2011). Some
females aged between 60 to 79 years were found to be facing obese and had greatest

6
prevalence of higher level of red-blood cell folate. However since some years the level of red
blood cells were found to be low therefore, research has been done on the particular issue
“Red blood cell folate levels in Canadian Inuit women of childbearing years: Influence of
food security, body mass index, smoking, education and vitamin use”.
Empirical challenges confronted by the researchers
The strategy that has been taken into account by the Canada government in
identifying the level of mandatory folic acid fortification in grains and cereals so that the
level of the maternal blood folate during the periconceptional period. The main empirical
strategy of the paper is to identify the number of women who are having Red Blood Cell
Folate (RBCF) and the factors that are being helping in the development of this kind of
problems. Through the incorporation of the survey, the study has successfully identified the
number of women those who are consuming cigarette. Through the use of sampling method
and bivariate analysis, it has been seen that average age of the women who took participation
in the sample is around 29.6 years. Among them 82% of the women are current smokers and
about 35% of the women smoked more than 10 cigarettes per day (Ncbi.nlm.nih.gov, 2019).
Only 6.8% of the women took vitamins in the daily basis. However, one of the main insight
of this study is that the probability of developing RBCF is more for the women who are non-
vitamin consumers compared to vitamin consumers. Using the simple random sampling can
be concluded as one of the important strategy that has been taken by the study because of the
fact that Folate is an important component of DNA that is having huge impact on birth
defects and birth related outcomes (Ncbi.nlm.nih.gov, 2019). The correlation coefficient is
mainly showing the fact that Body Mass Index, level of education of the consumers and
income level of consumers are having a strong correlation and high impact on the food
security and RBCF levels.

7
The main empirical challenge that has been seen from the study is that most of the
women who are not taking vitamins on daily basis will be having problem in the form of
congenital heart disease during pregnancy. Most of the women in child bearing age were not
seen to take the vitamins and from the statistical study it can be seen that foods that are
fortified are low in price and rich in nutrition compared to other traditional foods. Though
intake of foods that are containing high levels of folate within them can be harmful for the
foetus. For example consumption of ring seal liver should be limited to 50g during the
pregnancy or when a woman is expecting pregnancy.
Figure 1: Table showing amount of folate in traditional foods
(Source: Ncbi.nlm.nih.gov, 2019)
The above table is showing the amount of folate acids that is being consumed by the woman
during pregnancy should be kept under strict control and monitoring in order to maintain
healthy RBCF. Due to cross-sectional study, the number of samples that has been collected
consists of women mainly of Inuit region.
In order to mitigate the empirical challenges, the study has identified the statistics that high
amount of direct and passive smoking is injurious during pregnancy and it is helping the
destruction of RBCF in an extensive manner that in turn is hampering the BMI of the women
that are having pregnancy (Ncbi.nlm.nih.gov, 2019). Even it has been seen that most of the

8
women living in Inuit region, is not even reaching the minimum level of RBCF and low
income, low education, food insecurity and smoking is having direct effects on the
development of foetus and family of these women should take care of them while they are in
the period of pregnancy.
Results credible in terms of providing a causal interpretation of the estimates
It has been seen from the bivariate analysis of the study that food insecurity, high rate
of smoking and food habit is making significant impact on the development of nutrition
within the woman body during the period of pregnancy. Most of the women are not having
required level of the BMI and RBCF and this is one of the main reason why most of the
foetus is not taking normal and healthy birth. This is one of the alarming issue in the sense
that it has been seen that women belonging from instable family incomes are more prone in
developing congenital heart diseases (Ncbi.nlm.nih.gov, 2019). Even in countries having
high growth of GDP is showing that presence of high level of inequality is mainly resisting
the equal spread of food items and nutritional level among all parts of the society. Women
within the age (19-50) having low income and lack of educational knowledge is having high
chance of getting issues regarding folate compared to aged women within the upper strata of
the society. Statistical positive association is being found within the RBCF and BMI and
similar correlation has been found with the variables circumference of waists and presence of
body fat (Ncbi.nlm.nih.gov, 2019). It has been seen that in order to increase the level of
nutrition level within the pregnant women is definitely having high correlation with the level
of education and level of the income level. The mean RBCF value was 935.5 ± 192 nmol/L
(range 373.7 to 1440.5 nmol/L), indicating considerable variability in values. The mean
RBCF level of non-vitamin users was significantly lower than that of the vitamin users (920.1
± 181.4 vs. 1146.1 ± 212.8 nmol/L, p<0.001). Not only this, the study has also found
significant association among RBCF and all factors of smoking.

9
Conclusion
This particular study was mainly done to assess the standard of living for the
Canada’s Inuit population so that government can frame policies that will be helping them in
gaining benefits that will be forcing the government to bring in changes in social policies that
will improve the standard of living of the people living in the region (Ncbi.nlm.nih.gov,
2019). The women with less number of BMI is more prone in getting tough disease that are
mainly related with development of pregnancy. It should be taken into consideration, that
they should provide foods having high nutrition so that they can fight with the problems that
can cause harm to the patients.

10
Reference list
(2019). Retrieved 7 August 2019, from
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6267647/pdf/41997_2018_Article_85.pdf
Colapinto, C.K., O’Connor, D.L., Dubois, L. and Tremblay, M.S., 2015. Prevalence and
correlates of high red blood cell folate concentrations in the Canadian population using 3
proposed cut-offs. Applied Physiology, Nutrition, and Metabolism, 40(10), pp.1025-1030.
Daly, L. E., Kirke, P. N., Molloy, A., Weir, D. G., & Scott, J. M. (1995). Folate levels and
neural tube defects. Implications for prevention. JAMA, 274(21), 1698–1702. Egeland,G.
(2010).InuitHealthSurvey2007–2008:Nunavut.Ste-Annede-Bellevue: McGill University.
Duncan, K., Erickson, A.C., Egeland, G.M., Weiler, H. and Arbour, L.T., 2018. Red blood
cell folate levels in Canadian Inuit women of childbearing years: influence of food security,
body mass index, smoking, education, and vitamin use. Canadian Journal of Public Health,
109(5-6), pp.684-691.
Egeland, G. M., Berti, P., Soueida, R., Arbour, L. T., Receveur, O., & Kuhnlein, H. V.
(2004). Age differences in vitamin A intake among Canadian Inuit. Canadian Journal of
Public Health, 95(6), 465–469.
Egeland, G. M.,Johnson-Down, L.,Cao,Z. R., Sheikh,N., & Weiler, H. (2011). Food
insecurity and nutrition transition combine to affect nutrient intakes in Canadian arctic
communities. The Journal of Nutrition, 141(9), 1746–1753.
Egeland, G.M., Duncan, K., Erickson, A.C., Weiler, H.A. and Arbour, L., 2018. Red blood
cell folate Levels in Canadian Inuit women of childbearing years: invluence of Food
insecurity, body mass index, smoking, education, and vitamin use.

11
Folic acid and neural tube defects. Ottawa: Health Canada, 2018. Available from:
https://www.canada.ca/en/public-health/services/ pregnancy/folic-acid.html (Accessed April
3, 2018).
Hidiroglou, N., Peace, R.W., Jee, P., Leggee, D. and Kuhnlein, H., 2008. Levels of folate,
pyridoxine, niacin and riboflavin in traditional foods of Canadian Arctic indigenous peoples.
Journal of Food Composition and Analysis, 21(6), pp.474-480.
Hidiroglou,N.,Peace,R. W.,Jee, P.,Leggee,D.,&Kuhnlein, H. (2008). Levels of folate,
pyridoxine, niacin and riboflavin in traditional foods of Canadian Arctic Indigenous People.
Journal of Food Composition and Analysis, 21(6), 474–480.
Huet, C., Rosol, R., & Egeland, G. M. (2012). The prevalence of food
insecurityishighandthedietqualitypoorinInuitcommunities.The Journal of Nutrition, 142(3),
541–547.
Jamieson, J.A., Weiler, H.A., Kuhnlein, H.V. and Egeland, G.M., 2016. Prevalence of
unexplained anaemia in Inuit men and Inuit post-menopausal women in northern Labrador:
international polar year Inuit health survey. Canadian Journal of Public Health, 107(1),
pp.e81-e
Jamieson, J.A., Weiler, H.A., Kuhnlein, H.V. and Egeland, G.M., 2012. Traditional food
intake is correlated with iron stores in Canadian Inuit men. The Journal of nutrition, 142(4),
pp.764-770.
Plante, C., Blanchet, C., Rochette, L. and O’Brien, H.T., 2011. Prevalence of anemia among
Inuit women in Nunavik, Canada. International journal of circumpolar health, 70(2), pp.154-
165.

12
Valberg, L.S., Birkett, N., Haist, J., Zamecnik, J. and Pelletier, O., 1979. Evaluation of the
body iron status of native Canadians. Canadian Medical Association Journal, 120(3), p.285.

13
Answer to question 2
Table of Contents
Introduction
Data Description
Methodology
Robustness check
Results
Conclusion
Reference
Appendices
Appendix 1: Missing values and descriptive statistics of variables.
Appendix 2: Graphical presentation of Variables.
Appendix 3: Analysis to find the relationship between Beta and quality of manager
Appendix 4: Test of results validity

14
List of Table
Table 1: Details of missing values of all the variables.............................................................29
Table 2: Frequency of categorical variable MBA....................................................................29
Table 3: Descriptive statistics of the variables.........................................................................29
Table 4: Correlation table showing relationship between Beta and other variables................35
Table 5: Regression analysis....................................................................................................35
Table 6: Robust regression analysis.........................................................................................35
Table 7: Breusch-pegan/Cook-Weisbberg test for heteroskedasticity.....................................37
Table 8: Multicollinearity test using pairwise correlation.......................................................38
Table 9: Multicollinearity test using VIF.................................................................................38

15
List of Figure
Figure 1: Pie chart of MBA showing area covered by each category......................................30
Figure 2: Histogram of Beta.....................................................................................................30
Figure 3: Histogram of SAT....................................................................................................31
Figure 4: Histogram of Age.....................................................................................................31
Figure 5: Histogram of Tenure.................................................................................................32
Figure 6: Matrix graph including all the variables...................................................................32
Figure 7: Scatter plot of Beta against SAT..............................................................................33
Figure 8: Scatter plot of Beta against MBA.............................................................................33
Figure 9: Scatter plot of Beta against Age...............................................................................34
Figure 10: Scatter plot of Beta against Tenure.........................................................................34
Figure 11: Normality check by k-density estimate..................................................................36
Figure 12: Normality check by p-norm plot of r......................................................................36
Figure 13: Normality check by q-norm plot of r......................................................................37

16
Introduction
Mutual fund is an investment which is collected from so many investors to invest in
government securities and bonds, other securities for example, stocks, bonds and instruments
of money market and other assets (Agarwal, Ma and Mullally 2018). This funds are operated
and allocated by business professionals like professional money managers. Managers
allocates the fund’s assets to generate capital gains and income for the investors of the fund
(Ma and Tang 2019). In order to generate the earning, these funds are invested in a large
number of securities. The managers decides the investing areas depending on market
conditions and nature of the securities and previous trend of the market (Jordan and Riley
2015). These types of decision taking and problem solving skills are gained by experience of
the manager that means the number of years the manager has been in charge of the same post
(Ibert 2017). Besides, for the post they earn degrees in simple words they have relevant
academic degrees and have perused for that. These academic records and their experiences
may have a significant impact on taking decisions while allocating the funds (Berk and Van
Binsbergen 2015.). The efficiency of allocating funds can be measured by the return of that
funds and also by the market risk for investing in that mutual funds. The market risk again
depends on the performance of money market and the economic conditions of the market.
In this report, the discussion is about the relationship between characteristics of the
manager and the risk of the mutual fund. The variables which are described briefly in the next
section of the report, are selected accordingly and assumed to be strong enough to predict the
risk of mutual fund. Now, to test the fact with statistical techniques, data on these variables
are collected. Each observation contains the beta value of the mutual fund and above
mentioned four characteristics of the manager of that mutual fund. The paper has generated
the descriptive statistics for each variable, discussed the normality of the variables, identified
the dependent and independent variables that are used to generate the correlation table and

17
scatter plot. Then a regression analysis is conducted in order to check the concerned
relationship. There are few test used in this paper to check whether the result of the
regression analysis is free from error like heteroskedasticity and multicollinearity.
This report is divided in to six parts and combining all the parts, it completes the
report by presenting the research step by step. First part is the introduction which provides a
brief description about the mutual funds and the mangers’ characteristics and the overview of
the research. Secondly, the data description which discusses about the variables that are
included in the analysis and the nature of the variables. Thirdly, methodology that describes
the statistical techniques that are used in order to find out statistical evidence of the
relationship. Fourth part of the report contains the robustness check of the empirical analysis.
Fifth section contains the results from the analysis and refers the table and figures generated
from the analysis. Finally, the conclusion part which summarizes the report with a brief
information about the relationship.
Data Description
There are few techniques to measure the risk of a mutual fund but the most reliable
measure is Beta as it considers the volatility of market at the time of calculating the Beta
value. Here, the relevant characteristics that may have the impact on the risk of the mutual
fund, are the age, tenure, university SAT score and the MBA degree of the manager (Hoberg,
Kumar and Prabhala 2017). These characteristics are used as the independent variables to
find the relationship with the risk. The variable, tenure of the manager indicates the number
of years the manager has the post which shows the level of experience of the manager in
dealing with the crucial decision makings in order to make a gain from the investment. Thus,
it can be expected that the more experienced manager will have more efficiency in making
income for the investor which have the potential to reduce the risk of the mutual fund
(Kempf, Manconi and Spalt 2017). Similarly, the university SAT score and the MBA degree

18
indicates the relevant academic knowledge of the manager (Jordan and Riley 2015). It can be
assumed that the higher university SAT score and having an MBA degree indicates the higher
level of academic knowledge relevant to the money market and mutual funds than those who
have lower university SAT score and don’t have the MBA degree. This indicates the negative
relation of risk of mutual funds with SAT score and having an MBA degree.
The data contains 2029 rows and 5 columns. In simple words, there are 2029
observations and each observation contains the data for total 5 variables. There is no missing
values in any row and column which means all the observations can be used. The risk of
mutual fund among these variable is dependent variable and the variables that describes the
characteristics of the manager are the independent variable. The risk of the mutual fund is
dependent on the decisions made by the manager and the relevant skill comes from the
experience and the academic knowledge of the manager (Nallareddy and Ogneva 2017). Thus
the dependent and independent variables are selected. Now, the dependent variable, Beta is a
continuous and takes the fractions too which lies between 0.23 and 1.76. The variable SAT is
the nominal and continuous variable that lies between 657 and 1662. The MBA is the
categorical and discrete variable that takes the value 0 or 1. The value 0 means that the
manager has no MBA degree and the value 1 indicates that the manager has the MBA degree.
The Age is the numerical and continuous variable that lies between 27 and 59. The Tenure is
the ordered and discrete variable that takes the value from 0 to 8 and does not take any
fractions (Appendix 1: Table 1). The value 0 indicates that the manager has no experience or
he/she is newly appointed on the post and thus 8 indicates the highest experience level.
There are more variables that may have impact on the risk of the mutual funds which
are not incorporated in the data set. For example, does the manager invests in mutual fund, if
he/she invests then does he/she invest in his own company’s’ stock, has he/she at other posts
which is related to investment or was he/she employed in any mutual fund company at a

19
different post than manager (Choi, Kahraman and Mukherjee 2016). Though the sample size
is large enough, it needs more data as there are various types of mutual funds depending on
the level of risk. The minimum recorded beta value is 0.29 and the highest is 1.76. This study
does not consider mutual funds beyond these level of risk. Moreover, there are managers
whose experience or the tenure is more than 8 years for which the impact might be higher on
the risk of the mutual funds. To extend the range of the tenure more observations are needed.
Methodology
Frist of all the basic descriptive statistics is measure using the data analytical software
called STATA. The descriptive statistics is calculated for all the variables except the
categorical variable MBA. This includes the mean, median, standard deviation, variation,
maximum and minimum value. This provides the location, shape and spread of the variables
from which it is clear that the sample is enough efficient to predict the population. For the
categorical variable, the percentage and frequency is calculated by which it is clear that how
many managers have the MBA degree and how many does not have the degree. This clearly
says about the biasedness of the variable in the data on which the analysis is conducted. The
pie chart is generated which visualizes the numerical stat of the variable MBA. The
histogram of the other variables is also generated which shows the distribution of the
variables and finally the study headed towards constructing the model for the relationship
between the risk of mutual funds and the characteristics of the managers of the mutual fund.
To do so, fist the matrix graph is prepared to visualise the trend among the variables. Then,
scatter plots are prepared for each independent variable against the risk of mutual funds
separately. The scatter plots show the tendency of positive or negative relationship depending
on the trend of the data. After that correlation table is prepared which finalises the model for
the research aim. The model that has been used in the analysis is mentioned below:
Y =α +β1 X1+ β2 X2 + β3 X3 + β4 X4 +ui

20
In the above model Y stands Beta which shows the risk of mutual funds, X1 stands for SAT
which shows the university SAT score, X2 stands for the categorical variable or the only
dummy variable in the model which is MBA that indicates weather the manager have the
MBA degree or not, X3 stands for Age of the manager and X4 stands for the Tenure that
shows the experience of the manager on that post. In the model, α stands for the constant term
which will show the risk of the mutual fund while the value of the other variables is zero. The
coefficient of the independent variables X1, X2, X3, and X4 are β1, β2, β3 and β4 respectively
that shows the marginal effect of the respective variable on the risk of the mutual fund.
Finally, ui presents the error term or the disturbance term of the model. As the dependent
variable is continuous and numeric, the ordinary least square method is applied. The
regression analysis first estimates the adjusted R2 and runs an ANOVA test at 0.05
significance level to check whether the coefficient of covariance is acceptable or not. The
hypothesis of the ANOVA test is mentioned below:
Null Hypothesis, H0: The value of coefficient of covariance is not significantly
different from 0.
Alternative Hypothesis, H1: The value of coefficient of covariance is significantly
different from 0.
The F-stat value and the p-value of the F-stat says which hypothesis will be accepted.
The decision rule is if the estimated F-stat is greater than the critical value of the F-stat at the
given level of confidence interval or the significance level then there will be enough evidence
to reject the null hypothesis and the alternative hypothesis will be accepted. The alternative
decision rule is if the p-value is less than the value of significance level (0.05) then the null
hypothesis will be rejected and the alternative hypothesis will be accepted. The alternative

21
hypothesis also implies that the incorporated independent variables are better than the
intercept model (there is no independent variable in the model).
After these, there is t-test at 0.05 significance level for each of the independent
variable to confirm whether the mean coefficient are zero or not. The hypothesis of the test is
mentioned below:
Null Hypothesis, H0: The mean value of the coefficient of the independent variable is
not significantly different from 0.
Alternative Hypothesis, H1: The mean value of the coefficient of the independent
variable is significantly different from 0.
The t-stat value says which hypothesis will be accepted. The decision rule is if the
estimated t-stat is greater than the critical value of the t-stat at the given level of confidence
interval or the significance level then there will be enough evidence to reject the null
hypothesis and the alternative hypothesis will be accepted. The alternative decision rule is if
the p-value is less than the value of significance level (0.05) then the null hypothesis will be
rejected and the alternative hypothesis will be accepted. The alternative hypothesis also
implies that the coefficient of the independent variable has an impact on the dependent
variable.
Finally there will be some special test to check whether it holds or does not hold the
assumptions of ordinary least square (OLS). In other words, it also can be said that the
validity test of the results. To check the normality of the residuals k-density estimate is
plotted against the normal plot. Moreover, p-plot and q-plot is also generated to check the
normality of the residuals which shows the fitted values and the observed values. The
heteroscedasticity is checked by conducting the Breusch-pegan/Cook-Weisbberg test while
considering the fitted values of Beta. The hypothesis for the test is mentioned below:

22
Null Hypothesis, H0: There is no significant heteroscedasticity or there exists a
constant variance.
Alternative Hypothesis, H1: There is significant heteroscedasticity or there does not
exist a constant variance.
The decision rule for the test is if the estimated χ2-stat is less than the critical value of
the χ2-stat at the given level of confidence interval or the significance level then there will not
be enough evidence to reject the null hypothesis and the null hypothesis will be accepted. The
alternative decision rule is if the p-value is greater than the value of significance level (0.05)
then the null hypothesis will not be rejected. The same test is conducted again but this time,
considering the independent variables which will provide the F-stat and the decision rule for
the F-stat is described above. Multicollinearity is also tested by computing the pairwise
correlation table and also by VIF. The higher the VIF value is higher the multicollinearilty
exists.
Robustness check
The robustness check is usually conducted while researcher wants to know the
behaviour of the estimated regression coefficient at the time of adding or omitting the
regressor to modify the regression specification. Behaviour of the coefficients of independent
variables of the model means that how the variables reacts to the dependent variable
(Dervilis, Worden and Cross 2015). The coefficients will remain significant or become
insignificant or a huge change in the standard error of the coefficients. However, here the
robust regression analysis is conducted and the result table is attached in the Appendix 3 and
table 6. The normal least square regression result is attached in the Appendix 3 and table 5.
Comparing both the regression results, it is found that the result of the robust regression
provides the robust standard error which is greater than the normal standard error. The t-stat

23
in the normal regression is greater than the t-stat estimated in the robust regression analysis
for which the p-value has changed. However, the significance level of the estimated
coefficients does not change. From, this it can be said that there is no need of any
modification in the specification of regression with the introduction of any regressor or
removal of any regression. This indicates that the regression model is better as there is no
behavioural change in the coefficients of the regressions.
Results
The descriptive statistics of the variables are described below one by one. The
descriptive statistics of all the variables is presented in the appendix 1 and table 3.
Beta: The mean value is 0.9711, standard deviation is 0.2403 and the variance is 0.0577. The
median of the variable is 0.97.
SAT: The mean value is 1142.003, standard deviation is 143.948 and the variance is
20721.08. The median of the variable is 1142.
Age: The mean value is 42.331, standard deviation is 4.842 and the variance is 23.445. The
median of the variable is 42.
Tenure: The mean value is 4.249, standard deviation is 1.17 and the variance is 1.369. The
median of the variable is 4.
MBA: The percentage of MBA degree holder is 59.64 and the 40.36% of the manager does
not have the MBA degree. (Appendix 1: Table)
The histogram of the variables shows that all the variables are normally distributed as
the histogram creates a bell shape (Chambers 2017). The generated scatter plots are not able
to show a clear relation between the risk and the dependent variables. However the
correlation table shows the stats with correlation coefficient value for the relationship

24
between two variables. The correlation between beta and SAT is 0.2981 which shows the
weak and positive correlation. The correlation between beta and MBA is 0.0822 which shows
the weak and positive correlation. The correlation between beta and Age is 0.1027 which
shows the weak and positive correlation. The correlation between beta and Tenure is -0.088
which shows the weak and negative correlation (Appendix 3: Table 4). The regression result
shows that the adjusted R2 is 0.1294 which indicates that the model can explain the 12.94%
variance in the dependent variable. The F-stat is 75.20 and the corresponding p-value is 0.000
which indicates that there is enough evidence to reject the null hypothesis and the alternative
hypothesis is accepted. Hence, it can be said that the model is better than the intercept model.
The p-value for the coefficients of the independent variables SAT, MBA, Age and Tenure are
0.000. This implies that all the coefficients are statistically significant at 5% significance
level. The mean coefficient of SAT, MBA, Age and Tenure are 0.00504, 0.0366078,
0.0087801 and -0.0352218. These mean coefficient values implies that one unit change in the
variable will change the Beta by the amount of mean coefficient (Fox 2015). One unit rise in
SAT will raise the value of Beta by 0.00504 unit. One unit rise in MBA will raise the value of
Beta by 0.0366078 unit. One unit rise in SAT will raise the value of Beta by 0.0087801 unit.
One unit rise in Tenure will reduce the value of Beta by 0.0352218 unit (Gunst 2018). The
intercept term is also significant as the p-value is less than 0.05. (Appendix 3: Table 5). The
plot in figure 11 shows the k-density estimate, figure 12 shows the p-plot and the figure 13
shows the q-plot. These three plots concludes that the normality of the residuals is valid
(Appendix 4: Figure 11, 12 & 13). The Breusch-Pagan/ Cook-Weisberg test for
heteroskedasticity considering fitted values of Beta gives the χ2 value which is 0.17 and the
corresponding p-value is 0.6794 which indicates that there is not enough evidence to reject
the null hypothesis and the null hypothesis is accepted. This indicates constant variance that
is there exists homoscedasticity. The Breusch-Pagan/ Cook-Weisberg test for

25
heteroskedasticity considering fitted values of Beta gives the F-stat value which is 0.29 and
the corresponding p-value is 0.8850 which indicates that there is not enough evidence to
reject the null hypothesis and the null hypothesis is accepted. This indicates constant variance
that is there exists homoscedasticity (Chatterjee and Hadi 2015) (Appendix 4: Table 7). The
test for multicollinearity using VIF presents the VIF values which hovers around 1. This
value is acceptable for non-existing multicollinearity (Appendix 4: Table 9).
Conclusion
The above analysis shows that there exist a relationship between the risks of
mutual fund and the characteristics of manager. The robustness check confirms the validity of
the variables, heteroskedasticy and multicollinearity tests confirms that there is no different
variance and no interdependency among the variables which indicates that the results are
valid and it is acceptable. Hence, it can be concluded that if the age of the manager is higher
than the average then the risk of the mutual fund will be higher and similarly for the
university SAT score of the manager. The MBA degree holder manager will raise the risk of
the mutual fund by 0.0366078 unit. However, the concept says that the higher university SAT
score and having an MBA degree indicates the higher level of academic knowledge relevant
to the money market and mutual funds than those who have lower university SAT score and
don’t have the MBA degree. This indicates the negative relation of risk of mutual funds with
SAT score and having an MBA degree. However, there are some important variables that are
not considered in the model may improve the results. There are more variables that may have
impact on the risk of the mutual funds which are not incorporated in the data set. For
example, does the manager invests in mutual fund, if he/she invests then does he/she invest in
his own company’s’ stock, has he/she at other posts which is related to investment or was
he/she employed in any mutual fund company at a different post than manager. Moreover,
the number of observation is high but the sample is not enough to establish a proper

26
relationship between the risk and the characteristics of the manager. Though the sample size
is large enough, it needs more data as there are various types of mutual funds depending on
the level of risk. The minimum recorded beta value is 0.29 and the highest is 1.76. This study
does not consider mutual funds beyond these level of risk. The correlation between beta and
Tenure is -0.088 which shows the weak and negative correlation. The mean coefficient of
Tenure is -0.0352218. These mean coefficient value implies that one unit rise in Tenure will
reduce the Beta by the amount of 0.0352218. This result also does not match the concept
discussed in the introduction which says that the tenure of the manager indicates the number
of years the manager has the post which shows the level of experience of the manager in
dealing with the crucial decision makings in order to make a gain from the investment. Thus,
it can be expected that the more experienced manager will have more efficiency in making
income for the investor which have the potential to reduce the risk of the mutual fund.
However, there are managers whose experience or the tenure is more than 8 years for which
the impact might be higher on the risk of the mutual funds. To extend the range of the tenure
more observations are needed.

27
Reference
Agarwal, V., Ma, L. and Mullally, K., 2018. Managerial multitasking in the mutual fund
industry. Available at SSRN 1910367.
Barber, B.M., Huang, X. and Odean, T., 2016. Which factors matter to investors? Evidence
from mutual fund flows. The Review of Financial Studies, 29(10), pp.2600-2642.
Berk, J.B. and Van Binsbergen, J.H., 2015. Measuring skill in the mutual fund industry.
Journal of Financial Economics, 118(1), pp.1-20.
Cao, C., Goldie, B.A., Liang, B. and Petrasek, L., 2016. What is the nature of hedge fund
manager skills? Evidence from the risk-arbitrage strategy. Journal of Financial and
Quantitative Analysis, 51(3), pp.929-957.
Chambers, J.M., 2017. Graphical methods for data analysis: 0. Chapman and Hall/CRC.
Chatterjee, S. and Hadi, A.S., 2015. Regression analysis by example. John Wiley & Sons.
Choi, D., Kahraman, B. and Mukherjee, A., 2016. Learning about mutual fund managers.
The Journal of Finance, 71(6), pp.2809-2860.
Dervilis, N., Worden, K. and Cross, E.J., 2015. On robust regression analysis as a means of
exploring environmental and operational conditions for SHM data. Journal of Sound and
Vibration, 347, pp.279-296.
Fang, Y. and Wang, H., 2015. Fund manager characteristics and performance. Investment
Analysts Journal, 44(1), pp.102-116.
Fox, J., 2015. Applied regression analysis and generalized linear models. Sage Publications.
Gunst, R.F., 2018. Regression analysis and its application: a data-oriented approach.
Routledge.

28
Hoberg, G., Kumar, N. and Prabhala, N., 2017. Mutual fund competition, managerial skill,
and alpha persistence. The Review of Financial Studies, 31(5), pp.1896-1929.
Ibert, M., Kaniel, R., Van Nieuwerburgh, S. and Vestman, R., 2017. Are mutual fund
managers paid for investment skill?. The Review of Financial Studies, 31(2), pp.715-772.
Jordan, B.D. and Riley, T.B., 2015. Volatility and mutual fund manager skill. Journal of
Financial Economics, 118(2), pp.289-298.
Jordan, B.D. and Riley, T.B., 2015. Volatility and mutual fund manager skill. Journal of
Financial Economics, 118(2), pp.289-298.
Kempf, E., Manconi, A. and Spalt, O.G., 2017. Learning by doing: The value of experience
and the origins of skill for mutual fund managers. Available at SSRN 2124896.
Ma, L. and Tang, Y., 2019. Portfolio manager ownership and mutual fund risk taking.
Management Science.
Ma, L., Tang, Y. and Gomez, J.P., 2016. Portfolio manager compensation and mutual fund
performance.
Ma, L., Tang, Y. and Gómez, J.P., 2019. Portfolio manager compensation in the US mutual
fund industry. The Journal of Finance, 74(2), pp.587-638.
Nallareddy, S. and Ogneva, M., 2017. Accrual quality, skill, and the cross-section of mutual
fund returns. Review of Accounting Studies, 22(2), pp.503-542.

29
Appendices
Appendix 1: Missing values and descriptive statistics of variables.
Table 1: Details of missing values of all the variables
Table 2: Frequency of categorical variable MBA
Table 3: Descriptive statistics of the variables

30
Appendix 2: Graphical presentation of Variables.
Figure 1: Pie chart of MBA showing area covered by each category
Figure 2: Histogram of Beta

31
Figure 3: Histogram of SAT
Figure 4: Histogram of Age

32
Figure 5: Histogram of Tenure
Figure 6: Matrix graph including all the variables

33
Figure 7: Scatter plot of Beta against SAT
Figure 8: Scatter plot of Beta against MBA

34
Figure 9: Scatter plot of Beta against Age
Figure 10: Scatter plot of Beta against Tenure

35
Appendix 3: Analysis to find the relationship between Beta and quality of manager
Table 4: Correlation table showing relationship between Beta and other variables
Table 5: Regression analysis
Table 6: Robust regression analysis

36
Appendix 4: Test of results validity
Figure 11: Normality check by k-density estimate
Figure 12: Normality check by p-norm plot of r

37
Figure 13: Normality check by q-norm plot of r
Table 7: Breusch-pegan/Cook-Weisbberg test for heteroskedasticity

38
Table 8: Multicollinearity test using pairwise correlation
Table 9: Multicollinearity test using VIF