How Government Size Optimization affect on European Economies?

JACEK RODZINKA

The Institute for Financial Research and Analyses (IFRA)

The University of Information Technology and Management in Rzeszow

Sucharskiego 2, 35-225 Rzeszów

POLAND

TOMASZ SKICA

The Institute for Financial Research and Analyses (IFRA)

The University of Information Technology and Management in Rzeszow

Sucharskiego 2, 35-225 Rzeszów

POLAND

TERESA MROCZEK

Department of Artificial Intelligence

The University of Information Technology and Management in Rzeszow

Sucharskiego 2, 35-225 Rzeszów

POLAND

ELŻBIETA OCIEPA-KICIŃSKA

Institute of Spatial Management and Socio-Economic Geography,

University of Szczecin

al. Papieża Jana Pawła II 22a 70-453

POLAND

Abstract: The issue of optimal size of the general government sector is analyzed by researchers using various

methods, most often through the prism of a specific goal. The article is an attempt to determine the optimal size

of the general government sector from the perspective of EU economies. To achieve this goal, the innovative

decision tree technique - the c5.0 method was used. The study covered data describing 28 EU member states in

the years 2000-2017 and 16,632 input data were analyzed.

The results of the conducted research showed that despite the fact that there is no single optimal and universal

solution, a series of dependencies can be observed. Knowing the impact of individual actions on the economy,

you can choose such instruments, as well as such a configuration that will help in a given area without harming

others. Thus, the technique used, combined with specific priorities in terms of impact on the economy, may

show which values of specific variables in the general government sector level should be pursued in order to

model the desired effect.

Key-Words: government size, economy, decision tree, public finance, data mining, European economies

Received: May 9, 2021. Revised: January 23, 2022. Accepted: February 2, 2022. Published: February 18, 2022.

1 Introduction

The impact of the general government sector (GGS),

in particular its size, on the economy remains

beyond discussion [1]. At the same time, it should

be noted that the relationship between government

size and economy is not clear and fully explained.

Authors usually use both: the size and structure of

government as a factor affecting the pattern, shape

of governance, and growth of an economy [2].

According to R. Ram [3], larger government size is

more likely to reduce economic growth. On the

other hand, J.M. Henreksonz-Paramo and D.

Martinez [4] proved that government spending

could improve the relationship between private and

social interests and commercial openness. As a

result, public investment can favor economic

growth. Just as the views on the impact of the

government size on the economic situation of a

country differ, so do the ways in which government

policies could affect on the country's economy.

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

647

Volume 19, 2022

S.A.Y. Lin [5] found that government can affect

positively on economic growth through provision of

public goods and infrastructure, social services and

targeted intervention. A. Fölster and M. Henrekson

[6], claims that at low levels of both: government

spending and taxation, the productive effects of

public goods are likely to exceed the social cost of

raising funds. V. Tanzi and H.H. Zee [7] notes that

the positive impact of the size of government

(measured by the size of spending) on the economy

only takes place up to a certain level. After it is

exceeded, the economic growth is likely to be

negatively affected by further increases in public

expenditure.

The analysis of the literature on the

optimization of the size of the government sector

with the use of expenditure measures allowed for

the formulation of several conclusions. A literature

review suggests that the government sector in

small countries is bigger than in a large ones

[8][9]. The results of empirical analyzes are subject

to differentiation depending on the period used for

the analysis, which countries (or groups of

countries) are studied, as well as the sources of

data used in the estimates [10]. Finally, based on

the share of government expenditure in GDP, it is

possible to divide studied countries on three

groups: countries with small government sector -

about 30% of GDP, countries with medium

government sector - 40% of GDP and countries

with big government sector more than 50% of GDP

[11].

The obtained results allow us to conclude that

with the discrepancies in the optimal size of the

government sector for the same countries

demonstrated by various authors (sometimes very

large), it is justified to look for alternative

solutions that would allow not only to measure the

size of the government sector, but also its impact

on the economies of the analyzed countries.

Considering the above, the main goal of the article

is to fill the diagnosed research gap by examining

how the optimal size of GGS influences the

development of the economies of the European

Union (EU) countries and analyze how the

selection of an appropriate structure and size of the

GGS, should depend on the adopted priorities for

economic development. For achieving the research

objective it is extremely important to determine the

answers on two questions. The first one concerns

the identification of the most frequent measures

describing the size of the government sector. While

the second refers to indication of economy

measures, allowing to identify the influence of

GGS on the economies of studied countries. To

achieve this goal, the decision tree technique - the

c5.0 method was used. The study covered data

from 28 EU member states in the years 2000-2017.

The stages of the analysis are presented in Figure

1. In the first stage, it comprised 16 632 input data.

Applied analysis made it possible to indicate

recommendations for economic practice. Such a

large and complex structure as the EU countries,

characterized by a different level of economic

development, is a good sample for analysis. Thus,

this knowledge can be used by policy makers,

helping them to make decisions about the size of

the public finance sector and apply a policy in this

respect that meets the adopted economic goals.

Fig. 1: The stages of the analysis.

Source: own study

The paper has four parts. First, it reviews previous

studies on the quantifying the GGS and economic

growth. Then in the next section we discuss

decision tree technique. In the third part, we

present a description of research methodology.

Findings discussion and a summary are in section

four. The article concludes with discuss of the

theoretical and practical implications of the study

results.

2 Literature Review: The Indicators

of Quantifying the Government Sector

Size and Economic Development

L. Di Matteo points out, in contemporary economic

theory there is no single, universal measure of

government size that would fully reflect all

relations resulting from both, the function and role

of government [12]. As shown in section 2.1. the

most common approach to expressing government

size is measures based on public spending. At the

same time, these measures reflect a relatively

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

648

Volume 19, 2022

simple view of how government affects the

economy. In particular, if this impact is not

reduced only to the amounts of funds spent but

taking the influence of the state from the regulatory

side, such as income redistribution, and indirect

spending via tax expenditures [13]. For this reason,

inter alia, Pathirane and Blades [14] argue that

measuring government size should also consider a

number of measures, including final public sector

demand, generated value added, public sector

employment, and even net lending. Such a

differentiated approach is important primarily

because different measures of government size can

lead to diametrically opposed conclusions. A

similar view is expressed by R. Hjerppe [15]. This

approach is also supported by H. Handler, B.

Koebel, J.P. Reiss & M. Schratzenstaller [16].

Authors arguing that the public sector is difficult to

measure with a single indicator and recommends

using several measures to express government size.

They recommend using public employment (as a

proxy for the production of public services by

government), the ratio of government expenditure

to GDP (as a measure of the volume of transactions

that involve the public sector), as well as the ratio

of total taxes to GDP (to reflect the financing side

of the government size).

L. Peters & J. Verrinder [17] believe that the

measures used to quantify the size of government

should depend on the analyzed role of the

government in the economy. The government plays

many roles in the economy and fulfills a number of

functions. These include the production of goods

and provision of services, consumption,

management of public funds (in terms of

expenditure and income), or being an employer.

The role of government as producer in the

economy should be expressed as a percentage

relation between the total value added produced by

the public sector of the total value added produced

by the country [18]. In contrast, the measurement

of government as a consumer is made by referring

the percentage of public spending on consumption

to GDP [17][19]. Government is also a spender. In

this role, the government's expenditure activity is

not limited only to consumer spending (as in the

case of: Landau [20][21]; ; Hsieh & Lai [22];

Chiou-Wei, et. al. [23]. In a wider meaning, it

considers public expenditure on investments,

interests, public procurement, social transfers and

subsidies for the private sector. Therefore,

government size can be expressed as a percentage

share of total public expenditure of GDP. This

approach is suggested by L. Peters & J. Verrinder

[17], B. Fakin & A. De Crombrugghe [24], as well

as Y.V. Samusevych & A. Shamaelh [25].

The role of the government as a revenue-raiser

is connected with tax policy. In this case, there are

two possible approaches to measuring government

size. The first covers the cumulation of public tax

revenues and social security contributions and their

percentage of GDP [17]. While the second

approach considers the percentage of total public

revenues (without division divided into tax and

non-tax revenue sources), in GDP [26][27][28].

The role of government as a borrower relates to

borrowing money from the private sector to

finance its activities [29]. In this case, just like

before, two approaches are used to measure

government size. The first one is the percentage

ratio of the state budget deficit or surplus in GDP,

while the second covers the percentage of

government debt to GDP (see: [17][18].

Government as a re-distributor refers to the

activity in the field of social protection and covers

all interventions intended to relieve households and

individuals of the burden of a defined set of risks

and needs [30][31]. Social transfers do not require

any return actions from the recipient entities. It's

just redistribution [18]. Redistribution is measured

by social protection benefits expenditure expressed

as a proportion of GDP [32][17]. This measure is

also used in other configurations. For example, A.

Meltzer & S. Richard [33] use the share of income

redistributed by government as a relative measure

of it's size.

Government is also an employer [34][35]. To

measure this government activity, it is needed to

estimate the percentage of state employees in the

total number of employed people in the country

(see: [36][37][38][39][40]). According to F. M.

Häge [41], a measure of government size based on

public employment are also a wages and salaries of

public employees being a major part of

government consumption expenditure. B.J.

Clements, S. Gupta, I. Karpowicz & S. Tareq [42]

expand and catalog the measures of government

employment into three groups. The first group is

"compensation of employees". These measures

include following configurations: as a share of

GDP, as a share of total expenditure, as a share of

domestic revenue, and comparison to spending on

non-wage outlays. The second group includes

"government employment". These measures

include percentage of: private sector employment,

total employment, and population. Third group of

measures "wage level", includes: average

government wages relative comparator private

sector wages, average government wages relative

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

649

Volume 19, 2022

to GDP per capita, and ratio of the highest

government wage to the lowest (compression

ratio). The view of the necessity to use different

measures of government size depending on its role

in the economy is supported by L. Di Matteo [12].

The author distinguishes the following roles and

fields of governments' activity: goods production

and provision of services, consumption of

resources, employer, capital investor, provider of

social transfers and subsidies, regulator and the

beneficiary of funds. A similar position is

presented by N. Gemmell, D. Gill & L. Nguyen

[18].

When analyzing alternative approaches to

expressing the size of government in the economy,

it should be noted that nowadays measures based

on the impact of public sector regulation on

various macroeconomic variables are playing an

increasingly important role. I. A. Kahn [43] states

that the Index of Economic Freedom is the best

way to quantify government size in the context of

regulation. His research proves that countries with

smaller governments have higher GDP per capita.

F. L. Pryor [44], while examining the degree of

regulation of a given country, indicated two

variables as statistically significant. The first is the

size of the economy, and the second is the income

inequality of the population. According to its

findings, the degree of regulation is directly

proportional to the size of the economy and

indirectly proportional to the income inequality of

the population. J.S. Ferris [45], indicates three

other measures used in determining the size of

government. The first one is consumption

expenditure divided by GDP. Government

consumption is the sum of expenditure on wages

and salaries and other non-wage consumption

expenditure. The second measure takes into

account current disbursements defined as

government consumption plus subsidies, social

benefits, current transfers, and property income

paid by government. This measure adds subsidies

and transfers to the service dimension of

governments activities. The third measure of

government is total disbursements. This measure

adds government investment expenditures and

consists of current disbursements plus government

gross investment minus both consumption of fixed

capital and net capital transfers received.

The analyzes presented above demonstrate the

multitude of measures that can be used to express

the size of a government. Their study leads to the

conclusion that the variation in the ways

government influences the economy is not

adequately represented by measures based solely

on government spending or employment. The costs

and benefits of establishing indirect subsidies such

as tax credits, and the power of government to

contract as employer and consumer allow

government for a significant impact on economic

resources with relatively little reflection in

expenditure or employment data [46]. The

presented conclusion is not only an extremely

important summary of the above-presented

overview of approaches to measuring the size of

the government, but also unequivocally justifies

the use of not one, but several complementary

measures covering all the spheres of its activity

discussed above in research on the optimization of

the size of the government sector.

Economic development in the simplest terms, is

understood as a process of positive changes

comprising both quantitative growth and

qualitative progress [47]. Among the measures of

economic development, the most popular are those

based on the system of national accounts (GDP,

GNP, PNN), the value of GDP per capita is still the

basic and commonly used measure of socio-

economic development [48]. At the same time,

Pater, Harasym and Skica [49] indicate that

economic development measured by GDP per

capita does not take into account many aspects

other than economic growth, including structural,

social and ecological changes. Many attempts have

been made to construct a synthetic measure that

takes into account all identified aspects of

development, but due to the impossibility of

standardizing and identifying all determinants

influencing the level of socio-economic

development, these attempts were abandoned [50].

The literature review clearly shows that the authors

use these two different approaches based on

GDP/GDP per capita or integrated assessment of

development based on multi-criteria approaches

[51]. Remeikienė et al. [52] concludes that the

main reason that GDP per capita to be used to

measure countries’ economic development is its

complexity - it combines all economic

performance, both sectoral and territorial. From the

point of view of this publication, a long going

debate on the relationship between financial

development and economic growth is crucial. As

noted by Škare et al. [53], financial development

stimulates economic growth through five channels:

facilitating risk management, allocating resources,

exerting corporate control, mobilizing savings and

ease trading of goods and services, leading to

capital accumulation and technological innovation

and growth. In this context, when examining the

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

650

Volume 19, 2022

size of the government, we analyze variables

relating to the public finance sector in practice.

The literature is dominated by studies showing

a variety of approaches to determine the

government's size (see: [54][55], but despite of

this, a kind of monotony is noticeable in the ways

of describing the size of the GGS. The approaches

to the GGS optimization, concentrates (with some

exceptions) on measure expressed share of public

expenditure in GDP [56]. R. J. Barro [57], using

the data for the period 1970-1985, determined that

the average, optimal size of government for OECD

countries is 14% of GDP (+/- 4%). G. Karras [58]

established the optimal size of government for 20

EU countries. Based on data for the years 1950-

1990, he found that this size is 16% of GDP (+/-

3%). R. Vedder & L.E. Gallaway [59], focused

their research on the optimization of the

government sector on selected countries. Contrary

to other studies, the authors used a long time series

of data in the estimates. Among the European

countries for which they estimated optimal size of

the government were: Denmark (1854-1988)

26.14% of GDP, Italy (1862-1988) 22.23% of

GDP, Sweden (1881-1988) 19.43% of GDP as well

as Great Britain (1830-1988) 20.97% of GDP.

P. Pevcin [60] also dealt with optimization of

the government sector in the EU countries. Based

on the data for the years 1970-2007, he found that

the optimal GGS, in terms of expenditure, looks as

follow: for Italy 37.09% GDP, France 42.90%

GDP, Finland 38.98% GDP, Sweden 45.96% GDP,

Germany 38.45% GDP, Ireland 42.28% GDP, the

Netherlands 44.86% GDP and Belgium 41.91%

GDP. D. Chobanov & A. Mladenova [61] based

their research on the size of the government sector

on the period 1970-2007. Using the expenditure

measure, they found that the optimal government

size for OECD countries is 25.00% of GDP. The

studies related to selected European countries

showed that in the case of Austria it is 18.00% of

GDP, Belgium 23.00%, Denmark 25.90%, United

Kingdom 22.00% of GDP, and Sweden 27.00% of

GDP.

M. Mutaşcu & M. Milo [62] in their research

focused on the optimal size of the government

sector in the old and new EU member states. Using

statistics for 1999-2008, they found that the

optimal GGS (measured by the level of

expenditure) for new EU countries is 30.42% of

GDP, while for the old member states it is 27.46%

of GDP.

F. Forte & C. Magazzino [63] and C.

Magazzino [64] also analyzed the issue of the

optimal size of the government sector. The

research was based on data for the years 1970–

2009 and 1960–2008. The analysis revealed that,

for the EU27 member states, the optimal

expenditure level is 37.29% of GDP, while the

average effective ratio is 47.90%. Their

estimations for individual EU countries indicate

the following values of optimal public expenditure

levels: Belgium 35.39% of GDP, Netherlands

35.52% of GDP, UK 43.50% of GDP, Ireland

44.47% of GDP, Austria 38.21% of GDP,

Denmark 38.63% of GDP, Finland 40.38% of

GDP, France 39.49% of GDP, Germany 41.99% of

GDP, Greece 39.33% of GDP, Italy 37.68% of

GDP, and Portugal 42.28% of GDP. Their later

research [65] for 30 European countries found that

the optimal total expenditure level is 39.65% of

GDP, the optimal current expenditure level is

30.03% of GDP, while the optimal capital

expenditure is 10.50% of GDP. For the euro area

countries, these values were respectively 38.43%

of GDP, 29.41% of GDP and 10.96% of GDP,

while for non-euro area countries: 39.71% of GDP,

30.11% of GDP, and 10.69% of GDP.

M. Boór [66], conducted research for EU

countries in 1995-2013. According to the findings,

the optimal size of government spending was

within 45.49% of GDP and 52.06% of GDP, while

the average value of the government expenditures

for that time period was equal only to 45.65%

GDP. In his latest research [67] determined, based

on data for 1995-2017, that the optimal

government size as the ratio of total government

expenditure to GDP for EU countries is 51.11% of

GDP.

3 Decision Tree

In our research regarding general government

sector size optimization we used the decision tree

technique – the c5.0 method [68][69][70]. Decision

trees are a tool used by economists, including

representatives of the European Commission, who

use tree-based approaches for understanding

growth patterns in the European regions [71]. This

method is also used to analyze Structural

Similarities of Economies for Innovation and

Competitiveness [72] and to study various aspects

of economic development [73][74]. Generally, the

procedure of the c5.0 is as follows:

1. import of input data,

2. for each attribute a, calculate the normalized

information gain ratio from splitting on a,

3. find the attribute with the highest normalized

information gain best_a,

4. create a node that splits on best_a,

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

651

Volume 19, 2022

5. repeat the procedure on the sublists obtained

by splitting on best_a and add child nodes to

the node created in previous step.

Data from which we induced tree were

expressed in the form of decision table. Rows of

the decision table represent cases (values of the

selected countries variables), while columns

represent variables describing GGS (attributes). A

decision represents values of the economic

indicators of member countries respectively. The

set of all cases is denoted by U. The set of all cases

labeled by the same decision value is called a

concept. A simple example of the decision table is

presented as Table 1 in which attributes are: Total

tax revenue, Public administration employment and

General government gross fixed capital formation

and decision is GDP in current prices.

Table 1. Example of decision table

Source: own elaboration.

Attributes are independent variables while the

decision is a dependent variable. The set of all

cases is denoted by U. In Table 1, U = {1, 2, 3, 4,

5, 6}.

Applied algorithm c5.0 uses the concept of

information entropy to choose the most informative

variables. Let a be variable (attribute or decision)

with a domain a1,..,an. The entropy of variable a is

defined as follows:

󰇛󰇜 󰇛󰇜



 󰇛󰇜 (1)

where p(ai) is relative frequency of the value ai of

the attribute a. A conditional entropy for the

decision d given an attribute a is defined as

follows:

󰇛󰇜 󰇛󰇜



    󰇛󰇜





(2)

where is conditional probability of the

value dj of the decision d given the ai of the

attribute a, and all values of decision d are

d1,…,dm.

The information gain ratio is defined as:

Gain ratio(a) = Gain(a)/H(a)

(3)

where Gain(a) = H(d) – H(d|a). All logarithms are

binary.

The process of computing the conditional entropy

H(GDP in current prices|Total tax revenue) is

illustrated on Figure 2.

Fig. 2: Computing of the conditional entropy H

(GDP in current prices|Total tax revenue)

Source: own elaboration.

H(GDP in current prices|Total tax revenue) =

3/6*(-2/3*log(2/3)-1/3*log(2/3))+3/6*0 = 0.459.

Similarly, the two remaining conditional entropies

are computed as follows: H(GDP in current

prices|Public administration employment) = 0.459;

H(GDP in current prices|General government gross

fixed capital formation) = 0.874.

Next, the information gain ratio is determined:

Gain ratio(Total tax revenue) = Gain(Total tax

revenue)/H(Total tax revenue) = 0.459/0.918 = 0.5

where Gain(Total tax revenue) = H(GDP in current

prices) - H(GDP in current prices|Total tax

revenue)

Similarly, the two remaining gain ratios are

calculated: Gain ratio(Public administration

employment) = 0.5 and Gain ratio(General

government gross fixed capital formation) = 0.048.

The attribute with the highest value of normalized

information gain (gain ratio) is chosen as a node.

Gain ratio Total tax revenue and Public

administration employment have the same values.

The first one should be chosen. The attribute Total

tax revenue is placed in the root of the tree. All the

samples for value high of the attribute Total tax

revenue belong to the decision negative and thus

the leaf node is created (Figure 3).

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

652

Volume 19, 2022

Fig. 3: The root of decision tree

Source: own elaboration.

Next, the algorithm is repeated recursively on the

partitioned sub lists – for value low of the attribute

Total tax revenue. In result the attribute Public

administration employment with the highest value

of normalized information gain ratio is chosen as

the second node. The completed decision tree for

Table 2 is shown on Figure 4.

Fig. 4: The decision tree

Source: own elaboration.

4 Data and Methods

The research was conducted for EU Member

States. The timeframe covered the years from 2000

to 2017 (at the research stage, 2017 was the last

year for which a complete set of input data was

available). Sources of data for the purpose of this

research were publicly available including:

Eurostat, OECD, as well as the World Bank. In the

first stage of the research, 15 variables describing

public finance sector and 18 variables describing

the economy were selected. The correlation

analysis applied at this stage made it possible to

determine the relationship between variables

describing the public finance sector and the

economy, and to select variables that are of

significant importance. The variables expressed in

nominal sizes were also removed. As a result, 11

variables describing the economy and 10 variables

representing the GGS were qualified for further

research.

The variables describing the size of the general

government sector correspond to three dimensions

needed to identify the relationship between the

GGS and the economy: employment, the

government sector output and financial effects of

government activity (see Table 2 for a complete

summary of the data describing these measures).

Table 2. The variables describing the size of the

general government sector

Source: own study

The variables describing the EU economies

covered several complementary aspects: the

situation on the labor market, the EU economies

include indices expressing changes in prices and

the exchange rate, foreign economic contacts of

individual countries and their effects, (as well as

the accompanying financial flows) and the overall

mapping of the state of EU economies (a summary

of data describing these measures is included in

Table 3).

Table 3. The variables describing the economies

Source: own study

Due to the fact that the variables had a constant

character, they have been subjected to a

discretization process. The values of each variable

have been divided into four intervals. An interval

criterion was an equal number of occurrences in

every interval, so there were exactly the same

number of countries belonging to the EU in every

single interval. Every variable and every interval

were studied for impact on the variables that

describe a condition of the economy. In these

analyses’ decision tree were used. The analysis has

been implemented in the R language. A separate

program has been developed for the needs of data

preprocessing (ie data preparation for research,

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

653

Volume 19, 2022

cleaning, discretization), additional validation and

interpretation of the results. The C50 package

available in the R system was used in the process

of generating decision trees.

The decision trees were generated considering

each year separately (i.e. the first set was generated

on the basis of data from the year 2000, the second

from the year 2005, etc.). Next, decision trees were

analyzed in order to identify the most informative

attributes (variables GGS) describing economies of

the countries chosen. A detailed analysis of the

nodes and split of data according to the attributes

with the highest value of normalized information

gain ratio, provided relevant information on

intervals values of variables describing size of the

GGS which affect the economy and validate their

overall importance.

Due to the large number of variables and data

accepted for the study, six periods were selected

for the analysis. These were the years 2000, 2005,

2010, 2013, 2015 and 2017. The choice of years

was not accidental as the authors aimed to show

the impact of public finance variables on the

economy in various phases of the business cycle.

There were significant differences in the border

values of the individual variables adopted for the

study between the individual years, it was assumed

that in each research period (year) the values of the

variables would be divided into quartiles. In the

case of variables describing the GGS, if the values

of the variable fell within the first quartile, the

range was described with the letter "A", if in

quartile II it was described with the letter "B", if in

quartile III - with the letter "C", and in quartile IV

the letter "D".

In the case of variables describing the economy,

it is of great importance which direction of the

volatility of the variables will be adopted, because

for some variables, the higher its value

(stimulants), the better the situation should be

assessed, while in the case of others, the opposite is

true (destimulants). The table 4 shows which

direction of volatility has been adopted.

Table 4. Direction of volatility favourable for the

economy

Source: own study

In the case of seven variables, marking of

ranges was used exactly as in the case of the

variables describing public finance sector (I

quartile - A, II quartile - B, III quartile - C, IV

quartile - D), and in the case of four variables

describing the economy, which were considered

that the lower the value of the variable, the better

for the economy the reverse designation of ranges

was used: (1st quartile - D, 2nd quartile - C, 3rd

quartile - B, 4th quartile - A). Such provision

allowed for uniform interpretation of obtained

results, as it made it possible to uniformly name

the ranges positively indicating the condition of the

economy (table 5).

Table 5. Approved titles of variables’ ranges

describing the economy

Source: own elaboration.

In the further analysis, it was assumed that the

authors' area of interest will only include situations

where the variables describing the economy took

values from the optimal range. The idea was to find

an answer to the questions: Is there an optimal

GGS size for the development of EU economies

and which variables describing the GGS and in

which value ranges cause the variables describing

the economy to take values from the best range for

the economy, i.e. the range defined as "optimal"?

5 Results and Discussion

For the purposes of this study, it was assumed that

the analysis will concern the optimal impact on

variables describing the economy will be exerted

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

654

Volume 19, 2022

by variables describing the public finance sector.

The variables describing the economy and the

variables describing the public finance sector that

affect the optimal level of economic variables are

listed below. Ranges in which the variables

describing the public finance sector should be

included were also indicated, so that the variables

on the side of the economy took value from the

optimal range for the economy (range D).

A detailed summary of the results of the

analysis is presented in the Appendix, the analysis

of these data showed that seven variables

representing the GGS affect eleven variables on

economy's side with different strength and in

different configurations. The greatest positive

impact (calculated as the number of relations that

take values in the D range) on the economy of the

analyzed countries has total general government

revenue and public administration employment, the

variable total general government expenditure had

a slightly smaller impact on the economy. For the

remaining variables, the observed effect is visible,

although it is weaker. At the same time, it was

observed that the occurrence (or lack thereof) and

the strength of the relationship between particular

variables differed from year to year. Only the

relationship between the total general government

expenditure and GDP in current prices always

takes values from the expected range (D). Thus, it

is not possible to unequivocally indicate the

optimal data set with best value ranges for the

variables describing the GGS from the point of

view of the economy. However, dependencies

between them can be indicated, e.g., by

appropriately shaping the size of the GGS

described by the variable named the share of the

GGS sector in GDP, decision-makers can influence

the level of foreign direct investment. In order for

this type of investment to flow into the economy,

share of the public finance sector should be rather

high. The research shows that the share of general

government sector in GDP should be in the second

or fourth quartile. However, for the balance of

foreign direct investment to be the highest, the

share of general government sector in GDP should

be rather low. The research shows that it should be

in the first to the third quartile. Consolidated public

debt as % of GDP is important for the economy. Its

low level (in quartile I or II) has a positive effect

on the economic activity index, the outflow of

foreign direct investment and unemployment rate.

It is worth noting that higher debt levels (in

quartile III or IV) have a positive effect on the

current account balance, the real effective interest

rate and the dynamics of potential production.

Gross fixed capital formation describing the

public finance sector should be kept at a low level.

Then they influence the optimal levels of such

indicators describing the economy as: the current

account balance, the inflow of foreign direct

investment (in relation to GDP), the value of GDP

in current prices per capita, the real effective

interest rate and a lower level of unemployment.

The tax burden in relation to GDP have a

positive impact on the balance of foreign direct

investment, as well as the value of potential

production. High expenditures of the public

finance sector positively affect the activity rate, the

current account balance, GDP dynamics in current

prices and the level of unemployment. Low

spending positively affects the inflow of foreign

direct investment, the dynamics of the harmonized

index of consumer prices (expressing the level of

inflation), the dynamics of GDP in current prices

and the level of the real effective interest rate. The

high level of revenues of the public finance sector

has a positive effect on activity rate, the current

account balance, the inflow of foreign direct

investment and the unemployment rate. It has not

been observed to indicate the direction of the

impact on the variable GDP growth in current

prices per capita.

It is obvious that some recommendations may

seem contradictory. In some cases, it is proposed to

keep a given variable describing the economy at a

high level, in other at a low level. It is difficult to

indicate the universal level of the variables

describing GGS that optimizes the EU economies.

There are no simple solutions in the economy, and

it is not justified to jump to too radical conclusions.

Despite the fact that the result of the research

conducted negates the question posed in the title, it

highlights a number of dependencies. Knowing the

impact of individual actions on the economy, you

can choose such instruments, as well as such a

configuration that will help in a given area without

harming others. The decision about the choice of

instruments rests with the decision-makers who

know best the condition and needs of specific

economies.

6 Conclusion

This study gives policymakers two suggestions. On

the one hand, it suggests which variables on the

GGS side and in what direction should be “steered”

in order to positively influence a specific variable

on the side of the economy. The second suggestion

is to indicate the effects on the economy of

"controlling" the level of the GGS. Due to the

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

655

Volume 19, 2022

impact analysis of a specific variable on the GGS

side on the variables on the economy,

policymakers got a hint about the behavior in

specific configurations describing GGS. The used

data mining decision tree technique allows to

perform simulations for specific data sets and the

selection of an appropriate structure and size of the

GGS, that should depend on the adopted priorities

for economic development. The results of the study

showed that it is not possible to indicate the

preferred, best for the entire economy of the GGS

description variable set.

The authors of this study realize that, as

emphasized [75], governments differ substantially

not only in size, but also in priorities, moreover,

the role and size of governments around the world

has changed drastically in the last couple of

centuries. On the other hand, in line with the basic

principles of economics, prudent policies favor

economic growth in these economies, as confirmed

by research findings [76].

This study is unique, as it is an unambiguous

indication of economic success. It gives decision-

makers who have an impact on shaping the size of

the public finance sector a clear indication what

should be done, how to influence the variables

describing the public finance sector in order to lead

the economy through its individual components to

optimal level. An additional added value is the data

mining technique used in this context. In the next

stage, the study can be extended to more countries

and group them, e.g. in terms of the level of

economic development, thus looking for a

relationship.

References:

[1] Tsouhlou, A. and Mylonakis, J. (2011) Public

expenditure, public sector size and growth: The

European Union marked structural differences.

Review of European Studies, Vol. 3, No 2: 33-

51.

[2] Peace, O., Chimezie, P., Omankhanlen, A., &

Eriabie, S. (2020). Nexus Between Public

Finance and Economic Growth in Nigeria.

WSEAS Transactions on Business and

Economics, 17, 184–194.

https://doi.org/10.37394/23207.2020.17.20

[3] Ram, R. (1986) Government size and economic

growth: A new framework and some evidence

from cross-section and time-series data. The

American Economic Review, 76(1): 191-203.

[4] González-Páramo J M and Martínez D (2003)

‘Convergence across Spanish regions: New

evidence on the effects of public investment’,

The Review of Regional.

[5] Lin, S.A.Y. (1994) Government spending and

Economic Growth. Applied Economics, Vol.

26: 83-94.

[6] Fölster, S. and Henrekson, M. (1999) Growth

and the public sector: a critique of the critics.

European Journal of Political Economy, 15(2):

337-358.

[7] Tanzi, V. and Zee, H. H. (1997) Fiscal policy

and long-run growth. Staff Papers, 44(2): 179-

209.

[8] Alesina, A. and Wacziarg, R. (1998) Openness,

country size and government. Journal of public

Economics, 69(3): 305-321.

[9] Sabra, M. M. (2016) Government size, country

size, openness and economic growth in selected

MENA countries. International Journal of

Business and Economic Sciences Applied

Research, 9(1): 39-45.

[10] Jetter, M. and Parmeter, C.F. (2015) Trade

openness and bigger governments: The role of

country size revisited. European Journal of

Political Economy, (37): 49-63.

[11] Ministry of Finance of the Slovak Republic

(2019) Využitie verejných výdavkov

(http://www.rozpocet.sk/web/#/prehlad).

[12] Matteo, L. D. (2013) Measuring Government in

the Twenty-first Century: An International

Overview of the Size and Efficiency of Public

Spending, Fraser Institute.

[13] Peltzman, S. (1980) The Growth of

Government. Journal of Law and Economics,

23(2): 209–287.

[14] Pathirane, L. and Blades, D.W. (1982) Defining

and measuring the public sector: some

international comparisons. Review of Income

and Wealth, 28(3): 261-289.

[15] Hjerppe, R. (2004) Uncovering the dimensions

of the common good – problems of

measurement of the size of the public sector. In

European Commission & Eurostat (Eds.), 24th

CEIES seminar: the size of the government

sector: how to measure. (pp.49–63).

Luxembourg: Office for Official Publications

of the European Communities.

[16] Handler, H., Koebel, B., Reiss, J. P. and

Schratzenstaller, M. (2006) The size and

performance of public sector activities in

Europe: An overview. Acta Oeconomica,

56(4): 399-422.

[17] Peters, L. and Verrinder, J. (2004) The size of

the government sector from different

perspectives. [In:] European Commission &

Eurostat (Eds.), 24th CEIES seminar: the size

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

656

Volume 19, 2022

of the government sector: How to measure. (pp.

64–80). Luxembourg: Office for Official

Publications of the European Communities.

[18] Gemmell, N., Gill, D. and Nguyen, L. (2016)

The Changing Size of the State in New

Zealand, 1900-2015. Working Paper 04: 1-26.

[19] Lynch, R. (2008) Measuring Non-Market

Activity in the National Accounts–Who is the

Customer. [In:] The General Conference of the

International Association for Research in

Income and Wealth: 24-30.

[20] Landau, D. (1983) Government expenditure

and economic growth: a cross-country study.

Southern Economic Journal: 783-792.

[21] Landau, D. (1986) Government and economic

growth in the less developed countries: an

empirical study for 1960-1980. Economic

Development and Cultural Change, 35(1): 35-

75.

[22] Hsieh, E. and Lai, K.S. (1994) Government

spending and economic growth: The G-7

experience. Applied Economics, 26(5): 535-

542.

[23] Chiou-Wei, S.Z., Zhu, Z. and Kuo, Y.H. (2010)

Government size and economic growth: an

application of the smooth transition regression

model. Applied Economics Letters, 17(14):

1405-1415.

[24] Fakin, B. and De Crombrugghe, A. (1995)

Patterns of government expenditure and

taxation in transition vs. OECD economies.

Center for Social & Economic Research.

Warsaw: CASE Research Foundation.

[25] Samusevych, Y. V. and Shamaelh, A. (2017)

The Relationship Between the Tax Burden and

Financing Public Services: A Comparison of

Ukraine and European Countries. Financial

Markets, Institutions and Risks, Volume 1,

Issue 3: 55-64.

[26] Tureac, C. E. and Dragomir, G. (2014) The

Analysis of the Evolution of Tax Revenues in

EU Member States during 2009-2013. Acta

Universitatis Danubius. OEconomica. Issue No

5: 113-123.

[27] Afonso, A. and Furceri, D. (2010) Government

size, composition, volatility and economic

growth. European Journal of Political

Economy, 26(4): 517-532.

[28] Bergh, A. and Karlsson, M. (2010).

Government size and growth: Accounting for

economic freedom and globalization. Public

Choice, 142(1-2): 195-213.

[29] Chemla, G. and Hennessy, C. A. (2016)

Government as borrower of first resort. Journal

of Monetary Economics, 84: 1-16.

[30] Dion, M.L. and Birchfield, V. (2010) Economic

development, income inequality, and

preferences for redistribution. International

Studies Quarterly, 54(2): 315-334.

[31] Peracchi, F. (1998) Patterns of social protection

expenditure in the European Union. [In:] TMR-

SFB 504 Conference on Savings, Pensions and

Portfolio Choice, Deidesheim, Germany

(March 5–7).

[32] Adema, W., Fron, P. and Ladaique, M. (2014)

How much do OECD countries spend on social

protection and how redistributive are their

tax/benefit systems? International Social

Security Review, 67(1): 1-25.

[33] Meltzer, A. and Richard, S. (1981) A rational

theory of the size of government. Journal of

Political Economy, 89: 914–925.

[34] Masters, M. F., Gibney, R., Shevchuk, I., and

Zagenczyk, T. (2008) The state as employer.

The SAGE handbook of industrial relations.

London: Sage, 305-324.

[35] Areen, J. (2008) Government as educator: A

new understanding of first amendment

protection of academic freedom and

governance. Geo. LJ, 97: 945-1000.

[36] Pevcin, P. (2004) Cross-country differences in

government sector activities. Journal of

Economics and Business, 22(2): 41-59.

[37] Gupta, S., Leruth, L., De Mello, L. and

Chakravarti, S. (2003) Transition economies:

how appropriate is the size and scope of

government? Comparative Economic Studies,

45(4): 554-576.

[38] Garrett, T. A. and Rhine, R.M. (2006) On the

size and growth of government. Federal

Reserve Bank of St. Louis Review, 88(1): 13-

30.

[39] Katsimi, M. (1998) Explaining the size of the

public sector. Public Choice, 96: 117–144.

[40] Caponi, V. (2017) The effects of public sector

employment on the economy. IZA World of

Labor, 332: 1-10.

[41] Häge, F.M. (2003). Determinants of

government size: The capacity for Partisan

policy under political constraints. University of

Konstanz (https://d-nb.info/1079554270/34).

[42] Clements, B.J., Gupta, S., Karpowicz, I. and

Tareq, S. (2010) Evaluating government

employment and compensation (Vol. 10).

International Monetary Fund, Fiscal Affairs

Department.

[43] Kahn, J.A. (2011) Can We Determine the

Optimal Size of Government? Cato Institute

Development Policy Briefing (7): 1-10.

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

657

Volume 19, 2022

[44] Pryor, F. L. (2002) Quantitative notes on the

extent of governmental regulations in various

OECD nations. International Journal of

Industrial Organization, 20(5): 693–714.

[45] Ferris, J.S. (2003) Do alternative measures of

government result in alternative explanations

for government size? Economics Bulletin, 8(4):

1-11.

[46] Peters, B.G. and Heisler, M.O. (1983) Thinking

About Public Sector Growth: Conceptual,

Operational, Theoretical, and Policy

Considerations. [in:] Ch.L. Taylor (ed.). Why

Governments Grow. Measuring Public Sector

Size. (pp. 177-198). Beverly Hills/London/New

Delhi: Sage.

[47] Klóska, R., Ociepa-Kicińska, E., Czyżycki, R.,

& Szklarz, P. (2020). Regional Development in

Poland in Taxonomic Terms. Sustainability,

12(11), 4780.

https://doi.org/10.3390/su12114780.

[48] Obrębalski, M. (2013). Dysproporcje

rozwojowe krajów i regionów Unii

Europejskiej. Wiadomości Statystyczne. The

Polish Statistician, nr 10, 40–52.

[49] Pater, R., Harasym, R., & Skica, T. (2015).

Index of Regional Economic Development:

Some Considerations and the Case of Poland.

Studia Regionalne i Lokalne, nr 1, 54–85.

[50] Kubiczek, A. (2014). Jak mierzyć dziś rozwój

społeczno-gospodarczy krajów? Nierówności

Społeczne a Wzrost Gospodarczy, z. nr 38 W

kierunku zintegrowanego rozwoju, 40–56.

[51] Lisiński, M., Augustinaitis, A., Nazarko, L., &

Ratajczak, S. (2020). Evaluation of dynamics

of economic development in Polish and

Lithuanian regions. Journal of Business

Economics and Management, 21(4), 1093–

1110. https://doi.org/10.3846/jbem.2020.12671

[52] Remeikienė, R., Belas, J., Kliestik, T., &

Smrcka, L. (2020). Quantitative assessment of

dynamics of economic development in the

countries of the European Union.

Technological and Economic Development of

Economy, 26(4), 933–946.

https://doi.org/10.3846/tede.2020.12892.

[53] Škare, M., Sinković, D., & Porada-Rochoń, M.

(2019). Measuring credit structure impact on

economic growth in Croatia using (VECM)

1990-2018. Journal of Business Economics and

Management, 20(2), 294–310.

https://doi.org/10.3846/jbem.2019.8344

[54] Skica, T., Rodzinka, J., and Mroczek, T.

(2015a) Data mining approach to determine the

relationships between the economy and the

general government sector size, Finansowy

Kwartalnik Internetowy e-Finanse, Vol. 11, No

(3): 1-21.

[55] Skica, T., Rodzinka, J. and Fryc, B. (2015b)

Application of lem2 algorithm in identification

of relationships between the size of general

government sector and the economy. Hyperion

International Journal of Econophysics & New

Economy, 8(2): 361-401.

[56] Peková, J. (2008) Veřejné finance: úvod do

problematiky. Praha: ASPI.

[57] Barro, R. J. (1990) Government Spending in a

Simple Model of Endogeneous Growth. Journal

of Political Economy, 98(5): 103–125.

[58] Karras, G. (1997) On the optimal government

size in Europe: theory and empirical evidence.

The Manchester School, 65(3): 280-294.

[59] Vedder, R. and Gallaway, L.E. (1998)

Government size and economic growth. The

Joint Economic Committee.

(https://www.jec.senate.gov/public/_cache/files

/da82f715-a9d3-4625-8cf2-

264b11153962/government-size-and-

economic-growth.pdf).

[60] Pevcin, P. (2004) Does optimal size of

government spending exist? University of

Ljubljana, 10(1): 101-135.

[61] Chobanov, D. and Mladenova, A. (2009) What

is the optimum size of government? Bulgaria:

Institute for Market Economics, 1-47.

[62] Mutaşcu, M. and Milo, M. (2009) Optimal Size

of Government Spending. The Case of

European Union Member States. Annales

Universitatis Apulensis Series Oeconomica,

11(1): 447–456.

[63] Forte, F. and Magazzino, C. (2011) Optimal

size government and economic growth in EU

countries. Economia Politica, 28 (3): 295-322.

[64] Magazzino, C. (2012) Wagner versus Keynes:

Public spending and national income in Italy.

Journal of Policy Modeling, 34(6): 890-905.

[65] Forte, F. and Magazzino, C. (2014) Optimal

size of governments and the optimal ratio

between current and capital expenditure. [in:]

F. Forte, R. Mudambi and P.M. Navarra (eds.).

A Handbook of Alternative Theories of Public

Economics. Cheltenham-Northampton: Edward

Elgar Publishing.

[66] Boór, M. (2015) Optimal Government size in

European Union in interaction with economic

growth. Finance and Risk, Vol. 1: 170-180.

[67] Boór, M. (2019) Optimal government size in

Czech Republic and Slovakia. Journal of

Modern Economic Research, 1(1): 5-22.

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

658

Volume 19, 2022

[68] Quinlan, J.R. (1993) C4.5: Programs for

Machine Learning. Morgan Kaufmann

Publishers.

[69] Kuhn, M. and Johnson, K. (2013) Applied

predictive modeling (Vol. 26). New York:

Springer.

[70] Quinlan, J.R. (2014) C4.5: programs for

machine learning. Elsevier.

[71] Annoni, P. (2016). Tree-based approaches for

understanding growth patterns in the European

regions. REGION, 3(2), 23.

https://doi.org/10.18335/region.v3i2.115

[72] Erkut, B. (2016). Structural Similarities of

Economies for Innovation and

Competitiveness—A Decision Tree Based

Approach. Studia Oeconomica Posnaniensia,

4(nr 5), 85–104.

https://doi.org/10.18559/SOEP.2016.5.6

[73] Wu, W., Shuyan, J., Liu, R., Jin, W., & Ma, C.

(2019). Economic development, demographic

characteristics, road network and traffic

accidents in Zhongshan, China: Gradient

boosting decision tree model. Transportmetica

A: Transport Science, 16.

https://doi.org/10.1080/23249935.2020.171154

[74] Zeng, L., Guo, J., Wang, B., Lv, J., & Wang,

Q. (2019). Analyzing sustainability of Chinese

coal cities using a decision tree modeling

approach. Resources Policy, 64, 101501.

https://doi.org/10.1016/j.resourpol.2019.10150

[75] "Ortiz-Ospina, E., & Roser, M. (2016).

Government Spending. Our World in Data.

https://ourworldindata.org/government-

spending

[76] Ahmad, N., & Ahmed, F. (2005). Does

government size matter? A Case Study of D-8

Member Countries. Pakistan Economic and

Social Review, 43(2), 199–212.

Contribution of Individual Authors to the

Creation of a Scientific Article (Ghostwriting

Policy)

Jacek Rodzinka, Tomasz Skica, Teresa Mroczek and

Elżbieta Ociepa-Kicińska were responsible for

conceptualization, literature review, methodology,

investigation and writing an original draft.

Additionally: Jacek Rodzinka and Tomasz were

responsible for the project administration, selection

of variables, development of databases and their

initial processing; Teresa Mroczek was responsible

for the AI methodology, implementation and

conducting research, Elżbieta Ociepa-Kicińska was

responsible for writing - review & editing.

Sources of Funding for Research Presented

in a Scientific Article or Scientific Article

Itself

The project is financed within the framework of the

program of the Minister of Science and Higher

Education under the name "Regional Excellence

Initiative" in the years 2019 - 2022; project number

001/RID/2018/19; the amount of financing PLN

10,684,000.00.

Creative Commons Attribution License 4.0

(Attribution 4.0 International, CC BY 4.0)

This article is published under the terms of the

Creative Commons Attribution License 4.0

https://creativecommons.org/licenses/by/4.0/deed.en

_US

WSEAS TRANSACTIONS on BUSINESS and ECONOMICS

DOI: 10.37394/23207.2022.19.57

Jacek Rodzinka, Tomasz Skica,

Teresa Mroczek, Elżbieta Ociepa-Kicińska

E-ISSN: 2224-2899

659

Volume 19, 2022