Professional Documents
Culture Documents
CONTRIBUTION EVALUATION
Student ID First name Parts contributed Contribution % Signature
S3893543 Chuong Part A1, A4A, A4B 100% Chuong
S3877976 Hieu Part A2, A3, A4B 100% Hieu
S3915093 Tram Part B1, B2, B3.4, B4 100% Tram
S3916005 Thu Part B2, B3, B4 100% Thu
2
Table of Contents
Abbreviations.............................................................................................................................5
Section A...................................................................................................................................6
Part 1A:.....................................................................................................................................6
Step 1: Identify the needed data columns based on description....................................................6
Step 2: Import CSV file to R....................................................................................................6
Step 3: Row-eliminate with “DON’T KNOW” and missing values..............................................6
Step 4: Recode for appropriate format.......................................................................................6
Part 2A......................................................................................................................................7
1. Descriptive Analysis:.......................................................................................................7
2. Model (1)........................................................................................................................8
3. Multicollinearity testing....................................................................................................8
4. Heteroscedasticity testing:................................................................................................8
Part 3A....................................................................................................................................10
1. Goodness-of-Fit: interpretation........................................................................................10
2. Interpretation:................................................................................................................11
Part 4A:...................................................................................................................................12
1. Choice of probability model:...........................................................................................12
2. Estimation results:..........................................................................................................13
3. Economic theory............................................................................................................15
Part 4B:...................................................................................................................................16
1. Key Findings:................................................................................................................16
2. Policy Recommendation:................................................................................................16
3. Limitation:....................................................................................................................16
Section B.................................................................................................................................18
Part 1: Data assembly and cleaning.............................................................................................18
Part 2: Descriptive statistics and relevant tests.............................................................................18
1. Line charts....................................................................................................................18
a. Bitcoin......................................................................................................................18
b. FTSE.........................................................................................................................20
c. Crude oil...................................................................................................................21
d. Gold..........................................................................................................................23
2. Descriptive Statistics......................................................................................................24
3. Multiple regression........................................................................................................25
3
4. Serial correlation............................................................................................................25
Part 3: Discussion results...........................................................................................................26
1. Goodness-of-fit..............................................................................................................26
2. Interpretation of coefficients:..........................................................................................27
3. Adding trend variable.....................................................................................................28
4. Coefficient expectation...................................................................................................30
Part 4: Conclusion....................................................................................................................31
1. Main finding..................................................................................................................31
2. Policy...........................................................................................................................32
3. Limitation.....................................................................................................................32
Industry talk.............................................................................................................................33
Reference.................................................................................................................................34
Appendices..............................................................................................................................39
4
Abbreviations
FTSE100 - Financial Times Stock Exchange 100 Index
SD - standard deviation
5
Section A
Part 1A:
Step 1: Identify the needed data columns based on description.
Needed data columns: n5a, n5b, d2, b5, l1, b1, b7 and b7a (see Appendix 1).
All missing values are from n5a and n5b, 1259 rows were dropped using drop_na().
“DON’T KNOW” responses are coded “-9”, subset() is utilized to drop all rows with “-9”.
N5a and n5b are responsible for most eliminations (152 rows); l1, b7 and b7a get 1 row
eliminated each (Appendix 2). After cleaning, the data size went from 1664 to 251.
Firm Investment is fixed assets investment including equipment, land, and buildings; thus,
the data is obtained as n5a+n5b.
Firm Age is obtained as current year 2019 minus beginning year b5.
Legal Status Dummy interested in whether the firm is “Sole proprietorship”, b1 define “Sole
proprietorship”=3, thus we use ifelse(), returning “1” if b1=3, otherwise “0”.
Similarly, Gender of Top Manager and b7a, ifelse() is used to code “1” =Female and “0”
=Male.
6
Part 2A.
1. Descriptive Analysis:
No. of
observation 251
Table_1:_Descriptive_statistics_of_all_the_variables
For Invest, Sales, Age and Size, their mean is significantly greater than Median meaning they
are heavily right-skewed (Appendix4-7). Furthermore, their Median is close to Minimum
value implying that the distribution has almost no left-tail. Their ranges remained large,
showing a wide-spread distribution, or in this case very long right tail. Except for Age, these
variables tend to have large disparities showcased by high standard deviations compared to
Mean. Hence, Invest, Sales and Size may contain a lot of extreme-large outliers.
7
For dummy variables, Legal’s mean is 0.1912 meaning 19.12% of observed firms are Sole
Proprietorships which is unusual because Sole Proprietorships is usually dominant in quantity
compared to other legalities (OECD 2022). ManaGender’s mean is 0.9363 meaning 93.63%
of firms have a female top manager.
2. Model (1)
log ( Invest ) =B 0+ B1 log ( Sales )+ B2 Age+B 3 ¿ B ¿4 Legal + B5 exp +B 6 ManaGender + μ
3. Multicollinearity testing
VIF test has been conducted to examine the existence of multicollinearity between the
explanatory variables (CFI Team 2022). The rule of thumb suggests that the VIF should not
exceed 10 (Wooldridge 2016). In this case, multicollinearity is insignificant as the VIF results
are very small, indicating there is low multicollinearity.
Sales 1.173675
Age 1.576587
Size 1.260594
Legal 1.093570
Exp 1.349093
ManaGender 1.040037
Table_2:_VIF_of_explanatory_variables
4. Heteroscedasticity testing:
From Figure 4, the unequal scatter of residuals at different levels of the dependent variable
can indicate a high possibility of Heteroscedasticity. To examine the existence of
heteroskedasticity, Breusch-Pagan and White tests will be conducted.
8
H 1: H 0is wrong: Presence of Heteroscedasticity (no equal variance distribution of the
residuals)
Figure_1:_Model_(1)’s_scatterplot
9
Breusch-Pagan test
BP 19.08
DF 6
p-value 0.00403
White Test
Statistic 35.7
p-value 0.000358
Table_3:_Breusch-Pagan_and_White_Test_results
Part 3A.
1. Goodness-of-Fit: interpretation
Model 1 Equation:
The goodness-of-fit will be interpreted to determine how the observed data correspond to the
fitted model (Great Learning Team 2022). For model 1, R2 = 0.1763, it denotes that 17.63%
of Investment’s variation is explained by the independent variables.
−6
(Intercept) 6.53427503 1.32480801 4.9322 1.504 × 10 ***
10
ManaGender 0.26150103 0.31535560 0.8292 0.407787
F-statistics 8.701
−8
P-value 1.363 ×10 =0.00000001363
Table_4:_Model_1’s_Estimation_Result_(after_correcting_heteroskedasticity)1
^
log (Invest )=6.534+ 0.344 log ( Sales)+0.017 Age+0.0007 ¿ egal−0.009 exp+0.261 ManaGender
2. Interpretation:
As the p-value of the Intercept, log(Sales) and Legal is lower than 0.10, null hypothesis is
rejected and relationship between these variables and log(Invest) exists. On the other hand,
Exp and ManaGender are statistically insignificant for this model.
11
Table_5:_Model_1’s_interpretation.
Part 4A:
Probit and Logit models can remedy LPM problems thanks to their nonlinear-forms and y-hat
interval within (0, 1). Both are estimated using MLE unlike linear models using OLS.
Theoretically, Probit is based on the probit-function and Logit is-based-on the logistic
function, but their graphs share similar shape-and-properties (mentioned above).
Consequently, to choose between them, the pseudo-R 2 of McFadden (1974) and the number
of correct predictions were adopted. McFadden’s-Pseudo-R 2 is a fitting-goodness
measurement for binary-responses which is basically a comparison between maximized log-
likelihood-of-the fitted-model versus intercept-only model. Table 1 shows that Logit is higher
both in No.-of-correct-predictions and pseudo-R2. Ultimately, Logit-model is chosen for
Model (2).
12
Logit Probit
Table_6:_Comparison_of_Logit_and_Probit_model
2. Estimation results:
13
% of correct 64.54%
prediction
Pseudo-R2 0.1221
Table_7:_Estimate_results_of_Model_(2)
Model (2) has 3 statistically-significant variables namely Age, Size and Legal, the p-value are
0.0701, 0.0201 and 0.0256 respectively, so we can reject H0 that β j=0 at 10%-significance for
these variables.
Interpretation:
Age Given an additional year in firm age, the rough probability that the
firm has high investment increases by 0.5%.
Size If firm size increases by 1 employee, the rough probability that the
firm has high investment increases by 0.087%.
Legal The rough probability of firm having high investment is 21.81% less
when the firm is a sole proprietorship rather than other legal statuses.
Table_8:_Model_2’s_Interpretation.
14
3. Economic theory
Older-and larger-scaled firms will likely have a higher investment to maintain their
competitiveness as a part of their-business-cycle (Graebner-and-Eisenhardt, 2004; Villalonga
and McGahan, 2005; Yildiz et al. 2013). Younger-and-smaller firms suffer-more difficulties
in investment-process due to lacking-experience; hence, lower-investment (García-Quevedo
et al. 2014). Thus, positive-relationships are expected between investment and firm age/size.
Overall, Model-(1) is-not quite aligned with-expectation as Age and Size are not found
significant. Model-(2) is consistent with-expectation except for the insignificance of Sales.
Mismatch with expectation may-be due-to problems with data collection that reduced the
significance of these variables especially for Sales in Model (2); for Model (1) mismatch
occurs after adoption of robust-SD. Both models found Legal significant and negative
relationships with Investment/HI which is expected.
15
Part 4B:
1. Key Findings:
After adjustment of Model (1), we found that firms with higher Sales tend to have higher
Firm Investment. Besides, sole proprietorships tend to have 77.6% less of firm investment
versus other legalities. Besides, the effect of Age, Size, Exp and ManaGender are
insignificant.
From Model (2), we found that older and bigger firms will have a higher chance of having
high investment. We also found that the legal status has a strong impact: sole proprietors will
have roughly 21.8% lower chance of having high investment versus other legal statuses.
Also, Sales, Exp and ManaGender have no impact on probability of “high investment”.
2. Policy Recommendation:
Sole proprietors tend to have lower investments according to both Model (1) and (2)
reflecting their lacking access to funds. However, firms usually choose to remain as “sole
proprietorships” to enjoy the advantages of full profit, lower tax and regulation. On a national
scale, this decreases investment in the private sector and decreases economic output.
Consequently, for policy-makers, encouraging sole proprietorships to become higher
legalities may force them out of their comfort-zone, and get more investment ongoing; hence,
boost national investment expenditure and outputs. A detailed and long-run agenda of ODA
investment plans more toward non-sole-proprietorships is recommended as a potential
motivation to encourage more sole proprietors to become higher statuses.
3. Limitation:
First, the initial sample is 1664, but drastically decreased to 251 due to missing values.
Although sample size remained sufficient, unusual eliminations make our models less
reliable.
Second, since fixed assets are usually long-term investments, we may have miscaptured firm
investment by observing only in 1 year. For instance, if a firm purchases machinery in 2018,
16
and won’t replace or expand until 2023, we will completely omit this investment observing
only 2019.
For better estimation of Model (1) and (2), firstly insignificant variables need to be dropped
to prevent overfitting and collinearity. While the first limitation is out-of-hand, the second of
miscapturing long-term investment can be remedied by examining a longer period, e.g., last-
5-years, last-10-years, or year-to-date.
17
Section B
1. Line charts
a. Bitcoin
The price had trend and no evidence for cyclical or seasonal patterns. The dramatic uptrend in
2020 was stimulated by inflationary pressures and Tesla’s acceptance of Bitcoin payments.
Afterwards, the price entered a downtrend as Tesla stopped accepting Bitcoin payments and
interest rate hikes by FED.
18
Figure_2:_Line_chart_of_Bitcoin_price_(USD_2012-2022.
Figure_3:_Components_of_Bitcoin_price_(USD)_2012-2022.
19
b. FTSE
Generally, the FTSE100 steadily accelerated until 2020, when the Covid-19 strike.
Nevertheless, FTSE100 rapidly recovered in late 2020 and reached the pre-pandemic level.
There is no seasonal or cyclical pattern.
Figure_4:_Line_chart_of_FTSE_100_index_(points)_2012-2022.
20
Figure_5:_Components_of_FTSE_index_(points)_2012-2022.
c. Crude oil
The price had trend and no evidence for cyclical or seasonal patterns. The downtrend until
2020 was owed to excessive supply (WorldBank 2018). The following uptrend was owed to
supply shortage caused by labour cut and sanctions against Russia (World Bank 2022).
21
Figure_6:_Line_chart_of_crude_oil_price_(USD/barrel)_2012-2022.
Figure_7:_Components_of_crude_oil_price_(USD/barrel)_2012-2022.
22
d. Gold
Following eased inflationary pressure, gold price followed a downward trend in the first half
period but since 2016, the data seems to be uptrend following. There is no seasonality or
cyclical pattern detected.
Figure_8:_Line_chart_of_Gold_price_(USD/troy_ounce)_2012-2022.
23
Figure_9:_Components_of_Gold_price_(USD/troy_ounce)_2012-2022.
2. Descriptive Statistics
Table_9:_Descriptive_Statistics_of_variables.
Most noticeably, only Bitcoin had outliers and observations widely spread from the mean
(standard deviation>mean). The remaining variables had no outliers and were distributed
24
close to the mean (standard deviation<mean ). Bitcoin’s maximum price is almost 12,500
times the minimum. The maximum values of FTSE 100, Crude oil and Gold were 1.46, 6.09
and 1.86 times the minimum values, respectively. Accordingly, differences between the
maximum and minimum values of these variables were significantly less than that of Bitcoin.
3. Multiple regression
Log-return transformation makes data become less skewed and improves linearity. The
equation to measure the impact of return of Bitcoin, crude oil and gold on return of FTSE 100
will be (Appendix 9-13):
log return ( FTSE 100 )=B 0 +B 1 log return ( bitcoin )+ B2 log return ( oil ) + B3 log return ( Gold )+ μ
4. Serial correlation
The Durbin-Watson and Breusch-Godfrey tests were conducted to examine the existence of
model 3’s serial correlation.
Figure_10:_PACF_for_model_3’s_residual.
The residual appears to have correlation with its first lag as lag 1 exceeds the threshold line.
With both tests having p-value<5%, null hypothesis was rejected and hence, there is serial
correlation.
Breusch-Godfrey test
LM test p-value
9.5674 0.001981
Durbin-Watson test
Autocorrelation p-value
-0.2618608 0.02
Table_10:_Results_of_Breusch-Godfrey_and_Durbin-
Watson_test_for_first_order_lag_model_3.
Although the variables’ coefficients remain unbiased, serial correlation makes standard error
and significance tests of explanatory factors become imprecise. To be specific,
positive/negative correlation might deflate/inflate the standard error causing Type 1/Type 2
error (Kacapyr 2014). Variables’ functional form should be reviewed carefully to specify as
taking the wrong function form might be the reason to drive impure serial correlation.
Otherwise, the Newey-West method could be used as a remedy for unreliable standard errors
owing to pure serial correlation.
1. Goodness-of-fit
26
Estimate Std. Error t value Pr(>|t|)
R2 0.2893
Adjusted R- 0.2724
squared
Table_11:_Regression_result_of_Model_3.
^
log FTSE=0.0011+0.0212 log (bit)+0.133 log(oil)+0.0890 log(gold )
The R-squared of 0.2893 suggested that 28.9% of the variation of log(return of FTSE100) can
be explained by the variation of independent variables. The remaining 71.1% was explained
by the unobserved variables.
2. Interpretation of coefficients:
The p-value<0.05 illustrates log-return bitcoin and log-return crude oil’s statistical
significance to log-return FTSE. Meanwhile, the effect of log-return gold is insignificant due
to higher-than-0.05-p-value.
B1=0.0212 means that an additional percentage of Bitcoin’s return causes FTSE100’s return
increase 0.02%, ceteris paribus.
27
B2=0.1336684 means that as Crude oil’s return increases by 1%, make FTSE100’s return
increase by 0.13%, ceteris paribus.
log return ( FTSE 100 )=B 0 +B 1 log return ( bitcoin )+ B2 log return ( oil ) + B3 log return ( Gold )+ B 4 Trend + μ
R2 0.2911
Adjusted R- 0.2684
squared
Table_13:_Regression_result_of_Model_4.
Figure_11:_PACF_Residuals_of_Model_4
Breusch-Godfrey test
28
LM test p-value
9.8749 0.001675
Durbin-Watson test
Autocorrelation p-value
-0.2653717 0.02
Table14:Results_of_Breusch-Godfrey_and_Durbin-Watson_test_for_first-
order_lag_model_4.
In model 4, serial correlation is detected with first-order lag due to smaller than 0.05 p-value. 2
The remedy is to apply the Newey-West method to re-estimate the standard errors.
Table_15:_Regression_results_of_Model_4_after_Newey-West.
^
log FTSE=0.003845+0.02032 log bit +0.09356 log gold + 0.1346 log oil −0.0000411trend
Hypothesis test for the impact of trend variable on the log-return of FTSE100:
With p-value=0.51864(>0.05), null hypothesis is not rejected and hence, FTSE100’s return
has no significant trend. Noticeably, after adjusting for serial correlation, the log-return of
bitcoin variables no longer appears to have significant impact on the return of FTSE100.
29
4. Coefficient expectation
Trend -4.110×10−5=0.0000411
(6.3504×10−5=0.000063504 )
Table_16:_Comparison_result_of_Model_3_and_4_after_the_Newey-West_test.
Overall in Model 3, two independent variables were significant, including the Bitcon and
Crude oil’s return, whereas only Crude oil’s return was significant in Model 4.
Regarding Bitcoin, positive impacts on the stock index were expected, because Bitcoin
price’s determinants were similar to equities’, such as inflationary pressure and interest rate
hikes (Chan et al. 2019). Therefore, the Bitcoin’s coefficient was more reasonable in Model
3.
Both models confirmed the insignificance of Gold’s return. However, gold’s return is
expected to negatively impact FTSE100’s return, since Fakhfekh et al. (2021) found that
commodities, especially gold, has been a safe haven for investors, when the financial market
is under pressure of economic slowdown or political uncertainties.
30
Both models confirmed Crude oil return’s significant positive impact on FTSE100’s return.
However, an immediate negative impact was expected, because oil price hikes would
increase input costs and lower firm’s profitability, especially those in the service, consumer
goods and financial sectors, which are currently taking approximately 70% of the FTSE100
index (Jiang and Liu 2021; Chang et al. 2013; Soucek and Todorova 2013; Giselda 2022).
Lastly, the insignificance of the Trend variable on FTSE100’s return was reasonable, due to
stock index’s non-constant volatility (Alexandar 2000; Kasibhatla et al. 2006). Stock index’
return is determined by the comprised stocks’ price fluctuation, which is volatile to the
unanticipated and inconsistently altering factors, such as economical and political
uncertainties (Kajurova 2015) and the firm’s performance itself (Abraham and Cox 2007).
Part 4: Conclusion
1. Main finding
Section B aimed to discover whether the return of FTSE100 is driven by the return of bitcoin,
gold and crude oil. We found that Bitcoin’s return significantly impacts the FTSE100’s
return; nevertheless, it has an insignificant effect in model 4 when trend variable is added.
Furthermore, crude oil’s return appears to have a positive relationship with FTSE100’s return
which is unrealistic, since 70% of FTSE100 index firms are in the consumer G&S sector
whose profitability would decrease given increasing oil price. There is no significant
correlation between gold’s return and FTSE100’s return. Additionally, the FTSE100’s return
has no trend.
2. Policy
With the discovered positive relation, whenever recognizing an uptrend in the Bitcoin’s
return, investors should buy FTSE100’s stocks or buy both stocks and Bitcoin to diversify
investment portfolio and reduce risk. Noticeably, with the dataset’s monthly interval,
investors should hold the long position in the long-term to maximise the benefits of the
discovered relationship. Moreover, accelerating Crude oil’s return is a good buying signal for
FTSE100’s stocks due to discovered positive impacts. However investors should only invest
31
in energy-sector firms, since this sector is most beneficial from oil price hikes (Jiang and Liu
2021).
3. Limitation
The low Adjusted-R2 in both models denied the significant independent variables’ explanatory ability
and hence, reduced the models’ reliability. To improve, stock market factors such trading volume and
market liquidity should be included, for their significant impacts on investor’s investing demand
(Afego 2017). Other macroeconomic factors including inflation and economic recession should be
considered for being the major stimulator of investors’ uncertainties regarding the stock market
(Samaha and Abdala 2012; Khalfaoui, Gozgor and Goodell 2022). Furthermore, the model could
have considered the lagged of Crude oil’ return, because Xiao and Wang (2020) argued that the
negative impact of oil price on stock index would have some delayed spillover effects,
because higher oil prices stimulate inflationary pressure and the implied tightened monetary
policies would negatively hit stock market.
32
Industry talk
The first main point covered the technical aspects of problem-solving relative to the usage of
data. Mr. Gia covered the involvements of bringing the theory to life, through a four-phase
process. The first phase consisted of basic steps such as identifying issues, results and
determination of required resources. Then, the results and insights can be found which then
requires interpretation and communication of the insight. Lastly, the insight is translated into
action to solve the issue.
With the theory mentioned, Mr. An had backed the claim with the realistic application of
quantitative analysis. He has well walked through the audience how the issue has been
identified and the solution from conducting counterfactual analysis and priositing which type
crash to handle first. Overall, the delivery has allowed the audience to have a better view of
how the theory is applied into real life situations.It is highly beneficial for approaching issues,
especially for fresh graduates that are new to the industry.
Besides the mentioned technical factors, according to Mr.Gia, there are beyond data factors
that should be taken into consideration. From the experience of the speaker, translation of
insights into action is also one of the biggest challenges that he has faced. To assist the
interpretation of the data which drives decision making, effective communication and
analysis of results is necessary. Factors should be taken in consideration to increase the
effectiveness of communication includes role identification, provides empathy to the
audience and understands personalities. When conducting data analysis, finding the solution
is not the only objective but also how the results are translated to be easy to comprehend is
crucial. Therefore, the sharing from Mr. Gia on the communication approach is highly useful
for individuals that have careers in the industry to foresee upcoming challenges and be
equipped with the necessary skills.
33
Reference
Abraham S and Cox P (2007) ‘Analysing the determinants of narrative risk information in
UK FTSE 100 annual reports’, The British Accounting Review, 39(3): 227–248.
doi:10.1016/j.bar.2007.06.002
Alexander C (2000) ‘Principal Component Analysis of Volatility Smiles and Skews’, SSRN
Electronic Journal. doi:10.2139/ssrn.248128
CFI Team (2022) Variance Inflation Factor (VIF), CFI, accessed 4 January 2023,
https://corporatefinanceinstitute.com/resources/data-science/variance-inflation-factor-vif/.
Chan WH, Le M and Wu YW (2019) ‘Holding Bitcoin longer: The dynamic hedging abilities
of Bitcoin’, The Quarterly Review of Economics and Finance, 71:107–113.
doiL10.1016/j.qref.2018.07.004
Chang CL, McAleer M and Tansuchat R (2013) ‘Conditional correlations and volatility
spillovers between crude oil and stock index returns’, The North American Journal of
Economics and Finance, 25:116–138. doi:10.1016/j.najef.2012.06.002
Dezsö C and Ross D (2012) 'Does female representation in top management improve firm
performance? A panel data investigation', Strategic Management Journal, 27 January 2012,
33(9) : 1072-1089, doi: 10.1002/smj.1955, accessed 10 January 2023,
http://dx.doi.org/10.1002/smj.1955.
Fakhfekh M, Jeribi A, Ghorbel A and Hachicha N (2021) ‘Hedging stock market prices with
WTI, Gold, VIX and cryptocurrencies: a comparison between DCC, ADCC and GO-GARCH
models’, International Journal of Emerging Markets. doi:10.1108/IJOEM-03-2020-0264
34
Finkelstein S (1992) 'Power in Top Management Teams: Dimensions, Measurement, and
Validation', Academy of Management Journal, August 1992, 35(3) : 505-538, doi:
10.5465/256485, accessed 10 January 2023, http://dx.doi.org/10.5465/256485.
García-Quevedo J, Pellegrino G and Vivarelli M (2014) 'R&D drivers and age: Are young
firms different?', Research Policy, November 2014, 43(9) : 1544-1556, doi:
10.1016/j.respol.2014.04.003, accessed 10 January 2023,
http://dx.doi.org/10.1016/j.respol.2014.04.003.
Giselda R (2022) FTSE 100 Index Sector Weightings, Siblisreserach website, accessed 10
January 2023. https://siblisresearch.com/data/ftse-100-sector-weights/
Graebner M and Eisenhardt K (2004) 'The Seller's Side of the Story: Acquisition as Courtship
and Governance as Syndicate in Entrepreneurial Firms', Administrative Science Quarterly,
September 2004, 49(3) : 366-403, doi: 10.2307/4131440, accessed 10 January 2023,
http://dx.doi.org/10.2307/4131440.
Great Learning Team (2022) Understanding Goodness of Fit Test, Definition | What is
Goodness of Fit?, Great Learning, accessed 8 January 2023,
https://www.mygreatlearning.com/blog/understanding-goodness-of-fit-test/.
Griliches, Z. (2007) R&D and productivity: The econometric evidence, The University of
Chicago Press, United States of America.
Groves J (2022) What Is The FTSE Index?, Forbes website, accessed 10 January 2023,
<https://www.forbes.com/uk/advisor/investing/what-is-the-ftse-index/>.
Hambrick D and Mason P (1984) 'Upper Echelons: The Organization as a Reflection of Its
Top Managers', Academy of Management Review, April 1984, 9(2) : 193-206, doi:
10.5465/amr.1984.4277628, accessed 10 January 2023,
http://dx.doi.org/10.5465/amr.1984.4277628.
35
Heiss A (2021) Robust and clustered standard errors with R, Program Evaluation, accessed 8
January 2023, https://evalf21.classes.andrewheiss.com/example/standard-errors/.
Jiang W and Liu Y (2021) ‘The asymmetric effect of crude oil prices on stock prices in major
international financial markets’, The North American Journal of Economics and Finance,
56:101357–. doi:10.1016/j.najef.2020.101357
Kacapyr E (2014) Serial Correlation. In A Guide to Basic Econometric Techniques (pp. 132–
145), Routledge. doi:10.4324/9781315706856-17
Kasibhatla, K. M., Stewart, D., Sen, S., & Malindretos, J. (2006). Are Daily Stock Price
Indices in the Major European Equity Markets Cointegrated? Tests and Evidence. The
American Economist (New York, N.Y. 1960), 50(2), 47–57.
doi:10.1177/056943450605000205
Kor Y (2006) 'Direct and interaction effects of top management team and board compositions
on R&D investment strategy', Strategic Management Journal, 2006, 27(11) : 1081-1099, doi:
10.1002/smj.554, accessed 9 January 2023, http://dx.doi.org/10.1002/smj.554.
36
OECD (2022) Financing SMEs and Entrepreneurs 2022, OECD Publishing, doi:
10.1787/e9073a0f-en, Paris, France.
Smith N, Smith V and Verner M (2006) 'Do women in top management affect firm
performance?A panel study of 2,500 Danish firms', International Journal of Productivity and
Performance Management, 1 October 2006, doi: 10.1108/17410400610702160, accessed 10
January 2023, http://dx.doi.org/10.1108/17410400610702160.
Souček M and Todorova N (2013) ‘Realized volatility transmission between crude oil and
equity futures markets: A multivariate HAR approach’, Energy Economics, 40:586–597.
doi:10.1016/j.eneco.2013.08.011
Villalonga B and McGahan A (2005) 'The choice among acquisitions, alliances, and
divestitures', Strategic Management Journal, June 2005, 26(13) : 1183-1208, doi:
10.1002/smj.493, accessed 10 January 2023, http://dx.doi.org/10.1002/smj.493.
1, Cengage Australia.
37
Wooldridge, J.M. (2015) Introductory econometrics: A modern approach, 6th edn, Cengage
Learning, Boston City, USA.
World Bank (2018) What triggered the oil price plunge of 2014-2016 and why it failed to
deliver an economic impetus in eight charts, World Bank website, accessed 8 January 2023.
https://blogs.worldbank.org/developmenttalk/what-triggered-oil-price-plunge-2014-2016-
and-why-it-failed-deliver-economic-impetus-eight-charts
Xiao D and Wang J (2020) ‘Dynamic complexity and causality of crude oil and major stock
markets’, Energy (Oxford), 193:116791–. doi:10.1016/j.energy.2019.116791
Yildiz O, Bozkurt Ö, Kalkan A and Ayci A (2013) 'The Relationships between Technological
Investment, Firm Size, Firm Age and the Growth Rate of Innovational Performance',
Procedia - Social and Behavioral Sciences, November 2013, 99 : 590-599, doi:
10.1016/j.sbspro.2013.10.529, accessed 10 January 2023,
http://dx.doi.org/10.1016/j.sbspro.2013.10.529.
38
Appendices
39
Label No. of rows having “-9” eliminated.
n5a 152
n5b
d2 0
b5 0
l1 1
b1 0
b7 1
b7a 1
40
Abbreviation
Sales Firm sales: total annual sales for all products and
services (in LCUs)
41
Appendix 4: Histogram of Firm Investment ( x = (0, 8,000,000))
42
Appendix 6: Histogram of Firm Age
43
Appendix 7: Histogram of Firm Size
44
Appendix 9:
The return (%) of FTSE 100 index, Bitcoin, crude oil and gold will be calculated for further
regression model as observations are positive, using formula:
Log-return (x) = Log(x)-log(x-1)
(a) (b)
Figure: (a) Scatter plot of Bitcoin price on FTSE
(b) Scatter plot of Bitcoin’s return on FTSE’s return
(a) (b)
Appendix9: (a) Scatter plot of Crude oil price on FTSE
(b) Scatter plot of Crude oil’s return on FTSE’s return
(a) (b)
Appendix 10: (a) Scatter plot of Gold price on FTSE
(b) Scatter plot of Gold’s return on FTSE’s return
45
Appendix 11: Components of Bitcoin return (%)
46
Appendix 13: Components of crude oil return (%)
47