Data analysis

The data analysis tests used in the current study using SPSS are descriptive statistics, normality analysis
using Kurtosis and Skewness, correlation analysis, and, finally, multiple linear regression to examine the
causal relationship of all understudied explanatory variables on dependent variables. Besides these
analyses, different classical assumptions like heteroskedastic, autocorrelation and absence of
multicollinearity have been tested. The explanatory variables used in the current study are per capita
income, exchange rate, Gross savings, inflation, and interest rate. The following econometric model has
been developed for the recent research.
Stock price = 𝛽0 +𝛽1 (interest rate)i,t +𝛽2 (Exchange rate) i,t + β 3(inflation rate) i,t + β 4 (savings) i,t + β 5(income per
capita) i,t + e
Descriptive statistics
Table 1: Descriptive Statistics

Number Min Max Average STD

Interest rate 51 1.30 15.50 8.77 3.04
Inflation 51 -0.50 5.38 2.07 1.30
84.6
Exchange rate 51 2 107.51 95.85 6.78
Gross savings 51 -1.00 0.35 0.09 0.18
Per capita
income 51 4.78 8.62 7.16 0.86
Stock price 51 8.03 13.44 10.63 1.58
Concise helpful coefficients mainly central tendency measures, known as descriptive statistics, are
utilized to sum up whole specific data set, which may be a subset of a population or a depiction of the
entire population. Table 1 shows the descriptive statistics, including minimum, maximum, average and
standard derivation values. The term “standard deviation” indicates to measuring the data point
scattering as of simple average value. Low STD entails that data is clustered across simple average
values, while high STD shows that data has dispersion. In Table 2, the minimum value for interest rate is
1.3 percent with a maximum of 15.5 percent having an average value of 8.78 percent with a standard
derivation of 3.04 percent. It shows that interest rates can fluctuate up to 3.04 percent from 8.78 percent,
either in a positive or negative direction in China during the sampled period.
Similarly, the minimum value for the inflation rate is -0.5 percent, with a maximum of 5.38 percent
having an average value of 2.07 percent with a standard derivation of 1.3. It shows that the inflation rate
can fluctuate up to 1.03 percent from the mean either in a positive or negative direction in China during
the sampled period. Likewise, the minimum value for the stock price of Xiaopeng Automobile China
incorporation is positive 8.03, with a maximum of 13.44 having an average value of 10.63 with a
standard derivation of 1.58. It shows that the stock price of Xiaopeng Automobile China incorporation
can fluctuate up to 1.58 chines yeans from the mean either in a positive or negative direction in China
during the sampled period. Similarly, the minimum value for the inflation rate is 4.78 percent, with a
maximum of 8.62 percent having an average value of 7.16 percent with a standard derivation of 0.86. It
shows that the inflation rate can fluctuate up to 0.86 percent from the mean either in a positive or
negative direction in China during the sampled period. On comparative bases, the highest maximum
value is 107.51 of the exchange rates, with a minimum value of 84.62 having a mean of 95.85 and with
a standard deviation of 6.78. And the variable with minimum value considering maximum value is 0.35
in the case of gross savings having a minimum score of -1 with a mean value of 0.09 and STD of 0.18 in
China during the sampled period.
Normality analysis
Table 2: Kurtosis and Skewness analysis
Skewness Kurtosis
Statisti
Statistic Standard error c Standard error VIF
Interest rate -0.34 0.33 0.26 0.65 2.87
Inflation 0.42 0.33 0.43 0.65 2.23
Exchange rate -0.20 0.33 -1.11 0.65 1.37
Gross savings -4.93 0.33 31.70 0.65 1.47
Per capita
income -0.40 0.33 0.22 0.65 2.27
Stock price 0.46 0.33 -1.14 0.65
The skewness and Kurtosis test is employed in the current paper to examine the data set normality, which
is depicted in Table 2. There are two different schools of thought for identifying whether data is normal
or not based on skewness and Kurtosis. According to the one school of thought, the threshold values are
between plus and minus for both. At the same time, another school of thought urges researchers to use a threshold
or rule of thumb of plus minus ten. As in our case, the recommendations of the second school of thought are
fulfilled, and since skewness and Kurtosis statistic values are between plus and minus ten (Griffin &
Steinbrecher, 2013), therefore, we can reliably confirm that normality is assured for all variables.
Classical assumptions testing
In the current study, three different classical assumptions, namely, multicollinearity, heteroskedastic
issue (normality of residuals) and autocorrelation, are tested before performing multiple linear
regression.
Multicollinearity
Besides assuring the normality of variables, furthermore, it is also necessary to ensure that there is not
any multicollinearity in the explanatory variables of the study. In the case of multicollinearity, our
results depicting the actual effect of IVs on DV cannot be reliable. There are three methods to identify
multicollinearity absence/presence in the dataset. These are VIF, graphical presentation and correlation
between IVs. But most of the researchers recommend using the VIF approach and then testing
correlation among IVs as it is the most reliable. So, based on that recommendation, in this study, firstly,
VIF values are calculated. The rule of thumb for it is that if the VIF value is less than five, it will be no
multicollinearity. In the case of 5 to 9, it is considered as a mild multicollinearity which is also
acceptable in the research field. However, in case of more than ten, it will be severe multicollinearity
that has to be taken seriously. Since, as we see in Table 2, all variable VIF values are less than 5.
Further, correlation values are also considered, for which the rule of thumb is that these self-explanatory
variable correlation values should be less than 0.80. The depicted results in Table 3 show no value
greater than 0.8. So, no multicollinearity issue is found in the understudy econometric model based on
both recommended approaches.
Residual’s normality/ Heteroskedastic issue
Furthermore, the Shapiro-Wilk W test for residual normality was performed to check heteroskedastic
issues in the regression model. The fundamental residuals obtained on the base of an established
econometric model for study are assumed to be normally scattered, or nearly so, under the concept of
normality. While a residual plot or normal plot of the residuals can show non-normality, the Shapiro-
Wilk test or a comparable test can rigorously evaluate the assumption. Of which null hypothesis is errors
are not normal, and the alternate is errors are normal. The decision is taken based on the P-value; its
significance value leads to acceptance of the alternate hypothesis and ensures no heteroskedastic issue,
while an insignificant P-value shows the problem of heteroscedastic in the regression model. The results
revealed values as W= 0.9186, V= 3.887, Z = 2.899 and P = 0.00187 which is less than five percent.
Therefore, based on the significant P-value, we go towards the acceptance of the alternate hypothesis,
which is that there is a normality in residuals. It confirms no heteroskedastic issue in the established
regression model for the current study.
Autocorrelation
Before running the regression, along with testing multicollinearity, heteroskedastic issue, and variable
normality, it is also necessary to consider whether there is autocorrelation in the dataset or not. For this,
in the current study, the Durbin-Watson statistic test is employed. A number between zero and four will
permanently be designated to Durbin-Watson statistics according to scholars of one school of thought.
The numerical value of two shows autocorrelation absence, more and less than two shows positive,
whereas between 2 and 4 indicates no autocorrelation (Hill & Flack, 1987). In the existing scenario, the
Durbin statistic value is 0.655, less than two, indicating a positive autocorrelation in the data set. Since
macroeconomic variables are used in the current study to examine the effect on stock price, therefore, a
chance of sluggishness and inertia may be a significant cause for this occurring autocorrelation.
Correlation Matrix
A correlation is an association in the middle of two variables. A correlation is weak in a case of less than
or equal to .29, moderate if it ranges from 0.30 to 0.69 and above value shows a high and perfect
correlation (Schober et al., 2018). If a value is near to one too much, it is more acceptable. In the current
scenario, the two-tailed Pearson correlation is conducted to find the correlation coefficient among
variables. As our main concern is to check the relation of the stock price with all independent variables,
namely per capita income, exchange rate, Gross savings, inflation, and interest rate. Therefore, only
these correlations will be considered.
The correlation value between interest rate and stock price is 0.383, which shows a moderate positive
correlation but is significant at 0.01 level (two-tailed). While the correlation value between the inflation
rate and stock price is -0.502, which indicates a moderate negative correlation but is significant at 0.01
level (two-tailed). The correlation value between stock price and exchange rate is -0.883, which offers a
strong negative correlation but is significant at 0.01 level (two-tailed). In the case of both gross savings
and the per capita income with the stock price, a positive correlation has been found and is significant at
0.05 level (two-tailed). However, on comparative bases, the gross savings correlation is moderate, and
per capita income is weakly correlated with the stock price.
Table 3: Correlation Matrix
Stock price Interest rate Inflatio Exchange rate Gross savings Per capita income
n
Stock price 1
Interest rate .383** 1
Inflation -.502** -.530** 1
Exchange rate -.883** -.338* .499** 1
Gross savings .338* 0.196 -.549** -0.267 1
Per capita .282* .739** -.341* -0.163 0.166 1
income
 Significance level of correlation *** p<0.01, ** p<0.05, * p<0.1
Regression analysis plus interpretations
Table 4 shows the model summary of the regression model having values of mainly R square, adjusted
R square and Durbin Watson statistic. R square is related explicitly to the degree of explained variability
in the econometric model. A greater r-squared in general means the model is in good health to explain
the degree of variability. In general, higher variability is explained by higher R square value by and in
the model. In the current study, Table 4 is showing the R square value as 0.808, which is 80.80 percent
means that the regression model accounts for 80.80 percent of the variability which is seen in the target
variable.
The difference between R squared and adjusted R squared value is that the former considers all
independent variables. In contrast, the latter only considers those independent factors that genuinely
have an effect on authenticity and .reliability of model. And a large number of scholars preferred to use
adjusted R squared. Because modified R-squared can generate a more than precise image of the
relationship between two variables, it may be preferred over R-square. Adjusted R-squared does this by
accounting for the number of explanatory variables that are included in a particular model that is used to
compare the stock index. The adjusted R square of the current model is 0.787, which shows that all the
explanatory variables, namely, inflation, interest, exchange rate, per capita income, and individual gross
savings, combine to create an effect of 78.7 % in stock price. At the same time, the remaining 21.3 %
variation is because of some unknown/unobserved variables.
Results of multiple linear regression
Table Unstandardized coefficients Standard error Standardized coefficients T Sig

Coefficients (Constant) 27.58 1.88 14.66 0.00

6 Interest rate -0.02 0.06 -0.03 -0.26 0.80
Inflation 0.02 0.12 0.01 0.14 0.89
Exchange rate -0.20 0.02 -0.85 -11.13 0.00
Gross savings 0.90 0.71 0.10 1.26 0.21
Per capita income 0.28 0.18 0.15 1.56 0.13

Anova Sum of Squares Dif. Mean Square F Sig.

5 Regression 101.03 5.00 20.21 37.91 .000b
Residual 23.99 45.00 0.53
Total 125.02 50.00

Model R R Square Adjusted R Standard error of the Estimate Durbin-Watson

Summary Square
4 .899a 0.808 0.787 0.73 0.66

a Predictors: (Constant), per capita income, exchange rate, Gross savings, inflation, interest rate
Dependent variable: price
The extent to which each predictor variable contributes to the overall model is revealed in the ANOVA
table, which also indicates whether the comprehensive model is a significant predictor of the result
variable. Table 5 shows the F statistic = 37.907 with a P-value of 0.000, declaring that model is
statistically fit. Table 6 presents the beta coefficients of individual variables along with their Sig value.
Two different coefficients, namely standardized and unstandardized, are reported in this table. The main
concern of the researcher is always to see and consider only unstandardized beta coefficients. This table
presents a standardized coefficient that may be used to compare the size of effects of predictors
measured in various units and can be thought of as a “unit-free” measurement of effect size. The
negative sign of the beta coefficient shows that if those variable increases (decrease), the stock price will
decrease (increase). Correspondingly, the positive sign of the beta coefficient shows that if those
variable increases (decrease), the stock price will increase (decrease). The table of Coefficients
tells us the values of beta, which implies that change in 1-unit will contribute in a
specific percentage to a dependent variable by keeping constant to all others.
The beta coefficient of interest rate is -0.015, which shows that stock price and interest rate are inversely
related, and this relationship is also insignificant. It shows that an increase in interest rates in China will
1.5 percent inversely affect the stock price of Xiaopeng Automobile China incorporation. So, we can
conclude that increase (decrease) in interest rate will decrease (increase) the stock price. The results are
in line with Mok (1993), who contends that an increase in the interest rate lowers the present-day value
of potential dividend income, which could lower stock prices, a negative beta with causality extending
from stock prices to interest rates is frequently anticipated. Similar to it, the beta coefficient of the
exchange rate is -0.198, which shows that stock price and exchange rate are inversely related. Still,
surprisingly, this relationship, unlike other variables, is not insignificant. It shows that an increase in
exchange rates in China will inversely affect the stock prices of Xiaopeng Automobile China’s
incorporation. So, we can conclude that increase (decrease) in the exchange rate will decrease (increase)
the stock price of XPEV. Comparative to these two variables, others, namely per capita income, Gross
savings, and inflation, are positively related to the stock price of Xiaopeng Automobile China
incorporation.
The beta coefficient of the inflation rate is 0.016, which shows that stock price and inflation rate are
positively correlated. However, this relationship is also insignificant, like interest and exchange rate. It
shows that an increase in inflation rates in China will 1.6 percent positively affect the stock price of
Xiaopeng Automobile China incorporation. So, we can conclude that increase (decrease) in interest rate
will increase (decrease) the stock price of Xiaopeng Automobile China incorporation. This study’s result
is in line with Jaffe and Mandelker (1976). They account for an inverse relationship between concurrent
inflation rates and annual stock returns over a little sample period. However, a positive relationship over
a much more prolonged period was observed. The beta coefficient of gross savings is 0.899, which
shows that stock price and gross savings are positively related. However, this relationship is also
insignificant, like interest and exchange rate. It shows that an increase in gross savings in China will
89.9 percent positively affect the stock price of Xiaopeng Automobile China incorporation. So, we can
conclude that increase (decrease) in interest rate will increase (decrease) the stock price of Xiaopeng
Automobile China incorporation. The results are in line with the argument of Pal and Mittal (2011), who
said that higher saving rate might indefinitely boost trade and industry growth, according to some
contemporary assumptions of evolution, ultimately leading to a positive change in stock prices of
different companies in the marketplace, either maybe for an extended period or perhaps a short period.
The beta coefficient of per capita income is 0.283, which shows that stock price and per capita income
are positively related; however, this relationship is also insignificant, like interest and exchange rate and
others except for exchange rate. It shows that an increase in per capita income in China will 28.3 percent
will positively affect the stock price of Xiaopeng Automobile China incorporation. So, we can conclude
that increase (decrease) in interest rate will increase (decrease) the stock price of Xiaopeng Automobile
China incorporation. The result of the current study for this variable is also supported by previous
research. Higher share prices could boost consumer spending and aggregate demand, where consumer
spending is determined by the present value of future income (Mehra, 2001). Gains in the stock market
can stimulate consumption growth because equities are a significant portion of the overall wealth of
individuals.
Table 7: Summary of results
Path Coefficients Decision
Interest rate → Stock price -0.02 Not significant
Standard errors -0.06
Inflation → Stock price 0.01 Not significant
Standard errors -0.12
Exchange rate → Stock price -0.198*** Significant
Standard errors -0.02
Gross savings → Stock price 0.88 Not significant
Standard error -0.72
Per capita income Not significant
0.28
→ Stock price
 Standard errors -0.18
Constant 27.60***
Standard errors -1.88
*** p<0.01, ** p<0.05, * p<0.1
P-value < 0.01 shows the causal relationship is highly significant with three stars. The final econometric
equation, along with the beta coefficients and signs, is given below:
Stock price = 27.60 -0.02 (interest rate) i,t – 0.198 (Exchange rate) i,t + 0.01 (inflation rate) i,t + 0.88 (savings) i,t +
0.28 (income per capita) i,t + e
Main findings
Expected Obtained result
Stock price and interest rate are inversely related. Supported
Stock price and exchange rate are positively related. Not supported
Inflation rate has an inverse effect on the stock price. Not supported
Savings and stock prices are positively related. Supported
Income per capita and stock price are positively related. Supported
 9

References
Griffin, M. M., & Steinbrecher, T. D. (2013). Large-scale datasets in special education research.
In International Review of Research in Developmental Disabilities (Vol. 45, pp. 155-
183). Elsevier.
Hill, R., & Flack, H. (1987). The use of the Durbin–Watson d statistic in Rietveld analysis.
Journal of Applied Crystallography, 20(5), 356-361.
Jaffe, J. F., & Mandelker, G. (1976). The" Fisher effect" for risky assets: An empirical
investigation. The Journal of finance, 31(2), 447-458.
Mehra, Y. P. (2001). The wealth effect in empirical life-cycle aggregate consumption equations.
FRB Richmond Economic Quarterly, 87(2), 45-68.
Mok, H. M. (1993). Causality of interest rate, exchange rate and stock prices at stock market
open and close in Hong Kong. Asia Pacific Journal of Management, 10(2), 123-143.
Pal, K., & Mittal, R. (2011). Impact of macroeconomic indicators on Indian capital markets. The
journal of risk finance.
Schober, P., Boer, C., & Schwarte, L. A. (2018). Correlation coefficients: appropriate use and
interpretation. Anesthesia & Analgesia, 126(5), 1763-1768.