Fixed Effect Regression Model Based on STATA Analysis to Study the

Impact of R&D Expense Plus Deduction Policy on TFP

Wen Zhang

School of Economics, Anhui University, jiulong Road, Hefei, China

Keywords: Fixed Effect Regression Model, R&D Expense Plus Deduction, TFP, Firm Heterogeneity, Technological

Innovation.

Abstract: This paper selects the data of China's A-share listed companies from 2015 to 2021 as a sample to study the

impact of R&D expenditure plus deduction on total factor productivity of enterprises. Using STATA

measurement software, this paper constructs a fixed effect regression model and observes its linear regression

relationship with the R&D expenditure plus deduction preference intensity as the independent variable and

TFP as the dependent variable. The results show that the implementation of the R&D expenditure plus

deduction policy has significantly improved the TFP level, and the policy incentive effect of non-state-owned

enterprises and small-scale enterprises is more significant. Finally, based on the regression conclusion, some

suggestions are put forward to improve the preferential tax policy of R&D expense deduction.

1 INTRODUCTION

In recent years, with the rapid development of

computer technology, the use of computers for data

analysis has been widely used. Mining valuable

information from massive data to provide users with

more accurate data services has achieved very

significant results. STATA is a commonly used

computer data analysis software. This paper uses this

software technology to establish a fixed effect

regression model for regression analysis, test the

impact of the R&D expense plus deduction policy on

TFP, analyze the data of listed companies, and

provide data basis for the government to improve

policies.

Most scholars believe that the R&D expense plus

deduction policy has a positive impact on enterprise

development. The implementation of this policy has

reduced the tax expenditure of enterprises (Zhang

Wenchun, 2006), enhanced the enthusiasm of

enterprises for technology research and development,

and can improve the R&D investment of enterprises

in general (Zhou Keqing and Jingjiao, 2012).

However, the heterogeneity of enterprises, including

their life cycle, industry characteristics, domestic and

foreign market development conditions, will affect

https://orcid.org/0000-0002-2126-1643

the effect of policy (Ren Haiyun, Song Weichen,

2017). Policies can reduce enterprise operating costs

and improve enterprise performance by promoting

technological innovation (Hong Lianpu et al., 2019;

Wang Xi and Liu Meng, 2020). The improvement of

enterprise profitability guarantees investors'

investment income, creates favorable conditions for

enterprises to obtain external equity financing at a

lower cost, and improves enterprise market value

(Wang Ling et al., 2011). In terms of influencing TFP,

TFP can be improved by promoting enterprises' R&D

investment and improving technological innovation

(Ren Cancan et al., 2021). However, the improvement

effect is different due to different regions, enterprises'

scale, industry competition and life cycles of

enterprises (Liu Ye and Lin Chendan, 2021). In

addition, tax incentives such as additional deduction

of R&D expenses can also attract foreign capital, so

as to improve TFP (Luosha et al., 2014). However,

some scholars believe that the implementation of

policies does not necessarily improve the total factor

productivity of enterprises (Weinstein, 1996). Due to

the existence of market failure, if the low preferential

intensity is not enough to compensate for the R&D

risks and externalities spillovers caused by market

failure, it will have a reverse effect on the R&D

investment of enterprises, and it is difficult to

Zhang, W.

Fixed Effect Regression Model Based on STATA Analysis to Study the Impact of R&D Expense Plus Deduction Policy on TFP.

DOI: 10.5220/0012035900003620

In Proceedings of the 4th Inter national Conference on Economic Management and Model Engineering (ICEMME 2022), pages 517-521

ISBN: 978-989-758-636-1

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

517

promote the improvement of TFP (Feng Haihong et

al., 2015). To some extent, R&D incentive measures

will induce enterprises to implement R&D

manipulation, leading to a decline in the company's

scientific research performance and weakening TFP

(Yang Guochao et al., 2017).

2 RESEARCH HYPOTHESES

2.1 Impact of R&D Plus Deduction on

TFP

Based on the theory of technological innovation,

enterprises can develop new technologies and new

products through technological innovation, improve

production efficiency, enhance the core

competitiveness of enterprises, and further reduce

production costs, improve enterprise efficiency.

Therefore, good technological innovation activities

and achievements will bring benefits and overall

improvement to enterprises, thus promoting the

improvement of TFP. The R&D plus deduction policy

can reduce the input cost of technological innovation,

reduce the tax base, reduce the tax expenditure of

enterprises, ease the financial pressure, improve the

ability to resist risks, alleviate market failure,

stimulate the enthusiasm of enterprises for R&D

innovation, and improve TFP through technological

innovation. Secondly, the R&D expense plus

deduction policy is the transfer of national tax

benefits, which can reduce taxable income and cash

outflow. The increase of cash retention can allocate

funds in other aspects, increase factor input, expand

reproduction, and improve TFP. Therefore, the

following assumptions are made:

H1: The R&D expense super-deduction policy

can significantly increase TFP.

2.2 The Impact of the Nature of

Property Rights on Policy Effects

There are differences between state-owned

enterprises and non-state-owned enterprises in

domestic and foreign policy environment, economic

resource conditions, competition and internal

governance mechanisms. State owned enterprises are

heavily interfered by the government and have

performance requirements for the management.

Enterprise managers lack enthusiasm for reducing

costs and improving production quality through

technological innovation. They may give up projects

with high risks and long investment periods that are

conducive to the innovation and development of

enterprises and instead pursue short-term goals. This

greatly reduces the effect of improving TFP through

policies such as R&D expense addition and

deduction. For non-state-owned enterprises, the

market they are facing is not constrained and

supported by the government. Enterprises hope to

reduce production costs, improve product quality and

enhance market competitiveness through

technological innovation, so as to achieve the goal of

maximizing enterprise value. The policy of R&D

expense addition and deduction has stimulated

enterprises to carry out technological innovation

activities, thus improving TFP. The following

assumptions are proposed:

H2: The policy of R&D expense plus deduction

has a more significant effect on the TFP improvement

of non-state-owned enterprises.

2.3 The Impact if Enterprise Size on

Policy Effects

The R&D strength, anti risk ability and resource

acquisition ability of enterprises with different scales

are different. Large scale enterprises have strong

comprehensive strength, sufficient cash flow for

R&D activities, strong anti risk ability, obvious

advantages in resource acquisition, and low desire for

external R&D incentive policies. However, small-

scale enterprises are relatively short of R&D funds,

lack of R&D experience, imperfect operation and

management processes, and weak overall

technological innovation capability. In order to

enhance market competitiveness and expand

enterprise scale, they are more likely to be

encouraged by policies, and are willing to seize

various policy opportunities to increase innovation

resources, thus promoting enterprise technological

progress. Therefore, the R&D expense plus deduction

policy plays a more significant role in improving the

overall strength of small-scale enterprises. Therefore,

the following assumptions are made:

H3: The R&D expense plus deduction policy has

a more significant effect on the TFP improvement of

small-scale enterprises.

3 RESEARCH DESIGN

3.1 Sample Selection and Data Sources

The data of A-share listed companies from 2015 to

2021 were selected from CSMAR database, and

Stata17 was used for statistical analysis. After

ICEMME 2022 - The International Conference on Economic Management and Model Engineering

518

treatment, 3921 sample enterprises were screened,

with a total of 19051 observations.

3.2 Variable Selection and

Measurement

Interpreted variable: TFP. When LP method is used

to estimate TFP in this paper, the production function

is shown in formula (1) below.

lnY

＝β

＋β

lnK

＋β

lnL

＋β

lnM

＋ε

(1)

Explanatory variable: the specific calculation

method for the preferential intensity of R&D expense

plus deduction is:

Intensity=(R&D×Additional deduction

rate×Enterprise income tax rate)/total assets at the end

of the period

(2)

The selection of control variables is shown in

Table (1):

Table 1: Selection of control variables.

symbol

Variable

meaning

Variable definitions

Size Enterprise size ln (total assets).

Lev Gearing ratio

Total liabilities /

Total assets

ROE

Return on

equit

Net assets / ending

shareholders' equit

Cashflow Cashflow ratio

Net cash flow from

operating activities /

total assets

ListAge

The listing age

of the business

LN (Year of the Year

-Year of Launch +1).

TobinQ TobinQ value TobinQ value

Mfee

Management

expense rates

Administrative

expenses / operating

income

3.3 Empirical Model Design

With the development of computer technology and

econometrics, panel data models play an increasingly

important role in the research of social and economic

issues. Fixed effect models are widely used in

empirical tests.In order to test the impact of R&D

expense plus deduction policy on TFP, establish the

following fixed effect model:

TFP

=β

+β

Intensity

+αControls

+η

+μ

+δ

(3)

4 ANALYSIS OF EMPIRICAL

RESULTS

4.1 Benchmark Regression Analysis

In order to study the influence of R&D Expense Plus

Deduction Policy on TFP, by using the STATA

software, the experimental data are analyzed by fixed

effect model analysis. when other variables are

controlled. The results are as follows. Model (1) does

not control individual fixed effect and time fixed

effect, and model (2) is added. The regression

coefficients of Intensity are 17.29 and 12.12

respectively, which are significantly positive

correlation at the level of 1%. The regression results

show that the R&D plus deduction policy has a strong

incentive effect on TFP, which verifies H 1.

Table 2: Baseline regression results.

variable

TFP

(1) (2)

Intensity

17.290***

(0.951)

12.120***

(1.032)

Control variables Significant Significant

Enterprise fixed

effects

Uncontrolled control

Time fixation effect Uncontrolled control

Observations 19051 19051

0.637 0.964

4.2 Heterogeneity Analysis

In order to investigate whether the incentive effect of

R&D expense deduction on TFP with different

property rights is different, this paper divides the

sample enterprises into two categories, namely state-

owned enterprises and non-state-owned enterprises.

The regression results are shown in Table 4. Under

the control of individual fixed effect and time fixed

effect, the intensity regression coefficients of state-

owned enterprises and non-state-owned enterprises

are 7.724 and 13.13, The magnitude of both

coefficients indicates that the policy incentive effect

of non-state-owned enterprises is more significant,

which verifies H 2.

Fixed Effect Regression Model Based on STATA Analysis to Study the Impact of R&D Expense Plus Deduction Policy on TFP

519

Table 3: Heterogeneity of property rights.

variable

State-

owned

enterprises

Non-state-

owned

enterprises

Intensity

7.724***

(2.264)

13.130***

(1.169)

Control variables Significant Significant

Enterprise fixed

effects

control control

Time fixation

effect

control control

Observations 4729 14322

0.974 0.955

In order to study the difference of policy incentive

effect of heterogeneous enterprises of different sizes,

this paper chooses the median of the sample of

enterprise size as the dividing standard. The sample

with enterprise size smaller than the median is small

scale enterprises, and the others are large scale

enterprises. It can be seen from

Table 4 that the

regression coefficients of the intensities of large-

scale enterprises and small-scale enterprises are

10.98 and 14.20, which are significant at the 1%

level. The magnitude of the regression coefficients

indicates that the R&D expense plus deduction

policy has a more significant effect on the TFP of

small-scale enterprises, which verifies the H 3.

Table 4: Heterogeneity of enterprise scale.

variable

Large-

scale

enterprises

Small-scale

businesses

Intensity

10.980***

(1.527)

14.200***

(1.438)

Control variables Si

nificant Si

nificant

Enterprise fixed

effects

control control

Time fixation

effect

control control

Observations 9501 9550

R2 0.961 0.911

4.3 Robustness Tests

In order to ensure the reliability of the empirical

results, this paper conducts a robustness test, which

specifically includes: (1) replacing the measurement

method of the explained variable, and recalculating

TFP with OP method. (2) In order to alleviate

endogenous problems, the explanatory variables and

control variables will be lagged for a period of

robustness test. (3) The sample data from 2018 to

2021 were intercepted from the total sample and

regressed again. The regression results are shown in

Table 5. The sign of the correlation coefficient of

Intensity is still significantly positive, proving the

robustness of the regression results.

Table 5: Robustness tests.

variable TFP_OP

Lagging

one perio

Shrink the

sample

Intensity

3.258***

(1.072)

6.527***

(1.726)

23.500**

(1.158)

Control

variables

Significant Significant

Significan

Firm and

time fixed

effect

control control control

Observatio

19051 14352 12002

0.941 0.951 0.981

5 CONCLUSIONS AND

RECOMMENDATIONS

In the 21st century, computer technology has been

reasonably applied to the economic field and

skillfully introduced into data analysis, effectively

improving the efficiency and accuracy of data

analysis, realizing effective interpretation of data

information, understanding the information contained

in data, and finding problems through data analysis.

It can be seen that computer technology has a high

application value in the economic system. This paper

uses STATA measurement software to conduct

statistical analysis on the data of A-share listed

companies from 2015 to 2021, and empirically

examines the incentive effect of R&D expense plus

deduction policy on TFP. The research results show

that the policy of R&D expense plus deduction has

significantly improved TFP, and has more significant

effect on non-state-owned enterprises and small-scale

enterprises.

This paper puts forward the following policy

suggestions: (1) The government should actively play

the role of "visible hand", make use of the advantages

of computer technology to effectively analyze market

data, find out policy problems, constantly improve tax

preferential policies, maintain the coordination and

interaction between tax policies and market

mechanisms, stimulate the technological innovation

vitality of enterprises, and promote the promotion of

TFP. (2) The government should continue to pay

ICEMME 2022 - The International Conference on Economic Management and Model Engineering

520

attention to the incentive role of the R&D expense

plus deduction policy. It can consider expanding the

100% plus deduction proportion to all industries,

while strengthening the review and supervision

mechanism, standardizing the implementation of

policies, giving play to the guiding function of

government policies on enterprise innovation and

development, and improving TFP. (3) Due to the

heterogeneity of enterprises, diversified plans should

be adopted when formulating policies. Appropriate

preference should be made according to the

characteristics of enterprises, and targeted

preferential policies should be implemented, such as

appropriately expanding the pre tax plus deduction

preferential policies for non-state-owned enterprises

and small-scale enterprises, maximizing the

effectiveness of policies, and promoting the balanced

growth of the overall TFP.

REFERENCES

Beason R, Weinstein D E. Growth, Economies of Scale,

and Targeting in Japan (1955-1990) [J]. The Review of

Economics and Statistics,1996,78 (5):286-295.

Feng Haihong, Qu Wan, Li Minglu. R&D? [J]. Studies in

Science of Science,2015,33(5): 665–673

Hong Lianpu, Liu Yan, Zhang Xiang. Analysis of the Effect

of Corporate R&D Tax Preferential Policy—An

Empirical Study Based on China's Empirical Data[J].

Journal of Taxation and Economics, 2019, 24(1): 63–

Liu Ye, Lin Chenyan. and TFP[J]. Studies in Science of

Science, 2021, 39(10): 1790–1802

Luosha Du, Ann Harrison, Gary Jefferson. FDI Spillovers

and Industrial Policy: The Role of Tariffs and Tax

Holidays [J]. World Development, 2014, 64.

Ren Cancan, Guo Zeguang, Tian Zhiwen. and TFP[J]. East

China Economic Management, 2021, 35(5): 119–128

Ren Haiyun, Song Weichen Enterprise Heterogeneity

Factors, R&D Expense Plus Deduction and R&D

Investment [J] Science of Science Research, 2017, 35

(8): 1232 – 1239

Wang Xi, Liu Meng. ance: An Empirical Analysis Based

on Listed Companies in China[J]. Journal of Public

Finance, 2020(11): 101–114

Wang Ling, Zhu Zhanhong. thod[J]. Science of Science and

Management of Science and Technology, 2011, 32(9):

43–50

Yang Guochao, Liu Jing, Lian Peng, et al. ance[J].

Economic Research Journal, 2017, 52(8): 110–124

Zhang Wenchun. ysis[J]. Journal of Chinese Minmin

University, 2006(1): 59–64

Zhou Keqing, JING Jiao. fect of preferential tax policies on

R&D: A case study of ChiNext listed companies[J].

Tax Research, 2012(6): 20–24

Fixed Effect Regression Model Based on STATA Analysis to Study the Impact of R&D Expense Plus Deduction Policy on TFP

521