How to cite:

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David

Aritonang. (2022). Analysis of Factors Affecting Regional Financial

Independence in the Framework of A Policy Strategy for Increasing

Regional Original Income. Journal Eduvest. Vol 2(4): 777-796

E-ISSN:

2775-3727

Published by:

https://greenpublisher.id/

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April, 2022

p- ISSN 2775-3735- e-ISSN 2775-3727

ANALYSIS OF FACTORS AFFECTING REGIONAL

FINANCIAL INDEPENDENCE IN THE FRAMEWORK OF A

POLICY STRATEGY FOR INCREASING REGIONAL

ORIGINAL INCOME

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

PKN STAN, Pusat Penelitian, Badan Keahlian, Setjen DPR RI, Badan Pusat

Statistik, Bengkulu, Indonesia

Email: vissiadewih@pknstan.ac.id, [email protected].id,

ferdinand.david@bps.go.id

ARTICLE INFO ABSTRACT

Received:

March, 26

2022

Revised:

April, 16

2022

Approved:

April, 18

2022

The regional autonomy policy, which is accompanied by the

provision of balancing funds, has a goal, one of which is to have

an impact on the financial independence of regency/municipal

governments in Indonesia. The balancing funds provided

should have an effect or have an impact on regional financial

independence. This research is directed to see to what extent

the relationship of fiscal decentralization can provide an

increase in regional financial independence. Based on the

results of the panel data regression analysis of the district/city

data clusters, the results showed various results. In clusters I

and IV, it shows that there is a positive and significant effect of

the provision of balancing funds on regional financial

independence. Meanwhile, based on the results of panel

regression analysis in regencies/cities in cluster II, it shows that

the provision of balancing funds has no significant effect on

regional financial independence. The districts/cities in cluster

III show that the balancing fund variable has a negative and

significant influence on the financial independence of local

governments. This implies that the provision of balancing

funds for districts/cities in cluster III actually makes local

governments dependent on transfer funds from the central

government.

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

Analysis of Factors Affecting Regional Financial Independence in the Framework of A

Policy Strategy for Increasing Regional Original Income 778

KEYWORDS

Fiscal Decentralization, Balance Fund, Regional

Independence

This work is licensed under a Creative Commons

Attribution-ShareAlike 4.0 International

INTRODUCTION

Law (UU) No. 22 of 1999 concerning Regional Government became the forerunner

and became a new milestone for the era of decentralized governance in Indonesia. If explored

more comprehensively philosophically, Law no. 22 of 1999 is diversity in unity. There are four

considerations that provide a philosophical foundation in Law no. 22 of 1999, but one of the

important points in the consideration is the paradigm of granting flexibility to the Regions to

carry out Regional Autonomy. The law also provides for the transfer of governmental authority

and the implementation of the duties and responsibilities of several central government

responsibilities to local governments. The delegation of authority is of course accompanied by

the provision of fiscal decentralization funds as stated in Law no. 33 of 2004(Siswanto, 2013).

Further developments Law no. 22 of 1999 concerning Regional Government Law no. 25

of 1999 concerning Fiscal Balance was revised several times, until in the end it became Law

no. 23 of 2014 concerning Regional Government and Law no. 33 of 2004 concerning Fiscal

Balance is the legal basis for the implementation of fiscal autonomy and decentralization in

Indonesia (Halim, 2004). The impact of the implementation of fiscal decentralization is marked

by the process of transferring authority from the central government to local governments. The

process of delegation of authority is accompanied by greater financial transfer assistance to

regional governments in the form of balancing funds (Otsuka & Kalirajan, 2012).

The balancing fund itself according to Law no. 33 of 2004 concerning Financial Balance,

consisting of 3 parts of balancing funds, namely profit-sharing funds (DBH), general allocation

funds (DAU) and special allocation funds (DAK). The distribution of the balancing funds is in

accordance with the formulations and calculations that have been determined by the central

government. For example, the total amount of DAU is set to be at least 26% of the Net Domestic

Revenue set in the APBN (Widarjono, 2013). The amount of DAU received by each region is

calculated independently by taking into account and taking into account the population, area,

construction cost index, regional gross domestic product (GRDP) per capita and the Human

Development Index (IPM). DBH is given to regions that have natural resources as a

consequence of having to get a share of the revenue in these natural resources. Meanwhile,

DAK is allocated specifically to certain regions to fund special activities which are regional

affairs(Asyifa, n.d.).

The provision of fiscal transfer funds, both DAU, DBH, DAK in principle is to reduce

the fiscal gap between the central government and regional governments (vertical fiscal

imbalance) and between regions (horizontal fiscal imbalance) in the administration of

government. This inequality can be measured by the deviation between the regional gross

domestic product, both nominal and per capita, between regions compared to the national

average as well as between regions or regions. Of course, the smaller the difference or gap, the

more successful the implementation of fiscal decentralization (Syahril, Parinsi, & Togas, 2021)

However, in the development of fiscal decentralization and regional autonomy, several

obstacles have emerged. One of them is the dependence of local governments on fiscal

decentralization funds provided by the central government to local governments. The provision

of fiscal decentralization funds does not provide a stimulus for local governments to increase

local revenue (PAD). This is because local governments have a proportionate dependence on

decentralized funds as the main source of revenue from the Regional Budget (APBD). The

dependence of local governments on the value of central government transfers shows that the

ability of the regions to finance their budgets is relatively small. The impact is that the regional

government demands a larger transfer from the central government than exploring the source

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April 2022

779 http://eduvest.greenvest.co.id

of its original regional revenue (PAD). This is indicated by the low contribution of local

government PAD in financing regional expenditures which is no more than 20 percent (Waluyo,

2007).

Whereas in the era of regional autonomy it was implemented with the ultimate goal of

increasing community welfare. The process of achieving community welfare improvement

requires a source of regional income and regional development. Regional taxes and regional

levies are one of the sources of regional revenue (Mustanir, Yasin, Irwan, & Rusdi, 2019). The

subsequent impact is the obligation for regions to take care of their own households and

continue to strive to strengthen their independence through efforts to maximize the collection

of regional taxes and regional levies. Law No. 28 of 2009 has provided policy points regarding

regional taxes and regional levies. The regulated provisions include the mechanism for the open

list to become a close list, the expansion of the authority of taxation and retribution by

expanding the regional tax base and improving the system management of regional taxes and

regional levies through a policy of sharing the results of provincial taxes to regencies/cities.

Related to tax collection, one of the indicators to measure the amount of tax revenue is

to use the tax ratio (tax ratio). The tax ratio is the ratio between the amount of tax revenue and

the income of an economy. In the context of regional finance, the tax ratio is the ratio between

the regional taxes of the region's economy and the Regional Gross Regional Income (GRDP).

Therefore, by knowing the tax ratio of a region, we can understand the amount of regional

revenue as 2019 data shows the tax ratio in the aggregate of provinces throughout Indonesia.

Based on the figure, it can be seen that the Province that has the highest tax ratio is Bali Province

with a ratio of 8.8%. The high tax ratio in the Province of Bali is due to the very high local tax

of the Province of Bali (Fafurida & Pratiwi, 2017). The average tax ratio across Indonesia is

2.9%. The existing condition is that most of the provinces in Indonesia are still below the

national average tax ratio throughout Indonesia, including the provinces of East Java, Central

Java to Papua and West Papua (ANTIKA, 2018).

Another factor in regional financial independence is the allocation of balancing funds

that pay more attention to the expenditure aspect but is not supported by the accuracy of

calculating the ability of the region to increase its PAD, so that local governments do not

optimize regional capabilities and demand a larger allocation from the central government. This

has become a phenomenon of the flypaper effect, where local governments allocate more

regional spending than the source of balancing funds, (Syukridan and Halim, 2004);

(Maimunah, 2006). In line with the results of research (Mulya Rahmatul, 2016) which states

that the impact of the flypaper effect can cause fiscal gaps, not optimal potential for extracting

PAD, dependence on the central government, making regional finance less independent.

The role of the funds allocated by the central government is not optimal in increasing

regional capacity, especially regional financial independence in increasing their fiscal capacity.

Regional financial independence, which is indicated by the size of the Regional Original

Income (PAD) compared to income from other sources, such as central government assistance

and regional loans, makes PAD an important contribution. PAD plays an important role in the

contribution of local government expenditure financing (Berwulo, Luigi LD et al., 2017) that

the effectiveness of regional finance is shown by the increase in regional revenues in the city

of Jayapura, although the regional independence is still small, this is due to the large number of

financing that must be allocated . In addition, PAD has a contribution to the Regional Revenue

and Expenditure Budget (APBD) which is allocated to finance the administration of

administrative processes, government services and regional development. Optimizing PAD

requires a strategy and exploring the potential and competence of regional financial managers.

(Nilawati, 2019). Regional taxes, regional levies, the results of separated state wealth

management and other legitimate regional revenues can increase PAD in the province of

Central Java (Nuzulistyan, Supriyanto, & Paramita, 2017).

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

Analysis of Factors Affecting Regional Financial Independence in the Framework of A

Policy Strategy for Increasing Regional Original Income 780

Based on the explanation above, previous empirical research has used panel regression

analysis on districts/cities without any attempt to use research data clusters. This research is

important because it is based on the results of studies by various academics and the Ministry of

Finance that reformulation of balancing funds is needed, especially in relation to this research

focusing on regional financial independence. This study tries to cluster districts/cities as the

research locus. The use of cluster data in this study is based on the results of a study conducted

by the Directorate General of Fiscal Balance with the Australian Indonesia Partnership for

Decentralization (AIPD) which explains that one of the criticisms of the balancing fund policy

so far is "one size fits all" which means that the fund formula the current balance tends to apply

equally between city and district governments. The implementation of this policy should be

improved with a solution of regional grouping (clustering) to avoid the provision of balancing

funds that are not in accordance with the characteristics of city and district governments.

The balancing funds provided should have an effect or have an impact on regional

financial independence, such as previous research conducted by Fafurida and Pratiwi (2017);

Simanjuntak and Mukhlis (2016) and Suprantiningrum (2015), Yannis and Zoi (2015). Within

the framework of this understanding, this research is directed to see as far as where the

relationship of fiscal decentralization can provide an increase in regional financial

independence. Based on the background explanation, this research is focused on examining

what factors encourage regional financial independence (2) How does the provision of

balancing funds affect regional financial independence (3) What policies can encourage

regional financial independence.

RESEARCH METHOD

The scope of the research includes all districts and cities in Indonesia that receive

balancing funds in the form of General Allocation Funds (DAU), Special Allocation Funds

(DAK), Revenue Sharing Funds (DBH). The total number of regencies/cities that became the

unit of analysis in this study were 508 regencies, except for the urban regencies in the DKI

Jakarta province by dividing the regencies/cities into 4 research data clusters. The variables

used in this study include the dependent variable and the independent variable. The dependent

variable is the ratio of regional financial independence, while the independent variables include

tax growth, regional retribution growth, GRDP growth, central government transfers, direct

spending, and indirect spending.

The analytical method in this research is divided into two, namely descriptive analysis

method and inferential analysis method. The second method of analysis is the method of

inferential analysis. The use of the inferential analysis method has the aim of making

conclusions from a problem based on statistical rules in proving the truth of a hypothesis. This

study uses panel data regression analysis as a method of inference analysis. The use of panel

data analysis in this study aims to determine the factors that influence regional financial

independence as the dependent variable. The independent variables in this study include tax

growth, regional retribution growth, GRDP growth, central government transfers, direct

spending, and indirect spending.

RESULT AND DISCUSSION

The central government distributes balancing funds to every district/city in Indonesia

to reduce the fiscal gap between regions. The balancing funds include general allocation funds,

special allocation funds, and special autonomy funds. These funds become one of the drivers

of the economy in each district/city. This study divides each district into 4 clusters based on the

criteria of each region's financial independence ratio. In addition to balancing funds, there are

several other indicators to determine the financial independence of each region. These

indicators are direct spending, indirect spending, GRDP growth, taxes and levies. The research

period for 7 years from 2013 to 2019 consisted of 415 districts/cities in Indonesia. In this study,

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April 2022

781 http://eduvest.greenvest.co.id

not all districts/cities were included in the research object because of the unavailability of

research data in these districts/cities.

The quadrant formation method refers to clusters based on data on average balancing

funds and PAD for each district/city during 2013 to 2019. The general picture per quadrant is a

picture of balancing funds and PAD from each district/city that shows a general picture of

regional financial independence. So the quadrant method will map each cluster into 4 quadrants.

The quadrant with the following details:

a) Quadrant I: regencies/cities that have the characteristics of having large/high balancing

funds but have low PAD.

b) Quadrant II: ie districts/cities that have the characteristics of having large/high balancing

funds with high/large PAD.

c) Quadrant III: ie districts/cities that have the characteristics of having low balancing funds

but having low PAD.

d) Quadrant IV: ie districts/cities that have the characteristics of having low balancing funds

and high/large PAD.

1. Overview of Financial Independence in Regencies/Cities in Cluster I Period 2013-2020

Cluster I are districts/cities that have additional funds other than the provision of

balancing funds by the central government to local governments, for example districts/cities in

Papua/West Papua Province, Aceh which receive funds in the framework of Special Autonomy

and the Province of the Special Region of Yogyakarta which receives privileged funds. . The

number of regencies/cities in cluster 1 is 47. Regencies/cities included in cluster 1 are

regencies/cities that receive additional funds in addition to balancing funds. In cluster 1 for 7

years from 2013 to 2019, the average balancing fund provided by the central government was

1.13 trillion rupiah with a standard deviation of 1.65 trillion rupiah. The largest balancing fund

value was 9.62 trillion rupiahs given to the Asmat district in 2014, while the minimum value

was 100 billion rupiahs given to the city of Jayapura in 2016.

In 2013-2019 the average value of PAD generated was 93.17 billion rupiah. PAD is a

source of revenue in a certain area which is collected based on certain laws. The standard

deviation value of district/city PAD in cluster 1 is 122.19 billion rupiah. The maximum PAD

value in cluster 1 is 867 billion, namely in Sleman district in 2019, while the smallest PAD

value is Puncak Jaya district in 2013.

GRDP growth is an economic indicator to calculate the amount of economic activity in

a region. Positive and increasing economic growth indicates that economic activity in the region

is running smoothly and well. On average from 2013 to 2019, economic growth in cluster 1

was 5.13 percent with a standard deviation of 2.65 percent. The highest economic growth was

in Kulon Progo district in 2019, while the lowest economic growth was in Lhokseumawe City

in 2015.

Taxes and levies are a source of income in an area. Through funds collected from taxes

and levies, regions can improve development and infrastructure. During 2013 to 2019, the

average amount of taxes and levies collected reached 34.25 billion and 12.1 billion with a

standard deviation of 86.13 billion and 13.54 billion, respectively. The largest taxes collected

in districts/cities located in cluster 1 reached 650 billion, namely Sleman district in 2019, while

Statistik Deskriptif Variabel Dana Perimbangan, PAD, Pertumbuhan PDRB

Tahun 2013-2019 di Kluster 1

No.

Indikator

Dana Perimbangan

PAD

Pertumbuhan

PDRB

(1)

(2)

(3)

(4)

(5)

Rata-Rata

1,131,880,173,945

93,178,864,825

5.13

Standar

deviasi

1,648,140,453,915

122,196,133,619

2.65

Nilai

Maksimum

9,620,279,299,740

867,643,469,527

13.49

Nilai

Minimum

100,241,952,200

3,509,940,885

-20.34

Sumber: Data Diolah (2020).

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

Analysis of Factors Affecting Regional Financial Independence in the Framework of A

Policy Strategy for Increasing Regional Original Income 782

the smallest was in Supiori district in 2013. The largest levy was in Aceh Tamiang district in

2014, while the smallest was 6. 04 million in Mamberamo Raya district in 2019.

Direct spending and indirect spending are government spending that aims as a stimulus

in moving the economy. From 2013 to 2019, the average direct and indirect expenditure of

districts/cities in cluster 1 was 566 billion and 600 billion with standard deviations of 236.23

billion and 277.91 billion, respectively. The largest direct expenditure was in Bintuni Bay in

2019, while the smallest was 135.55 billion in the city of Sabang in 2013. The largest indirect

expenditure was in Sleman district in 2016 of 1.56 trillion, while the smallest was in

Mamberamo Raya district in 2013.

The formation of quadrant in cluster 1 is based on data on average balancing funds and

PAD for each district/city during 2013 to 2019. Quadrant I in the figure indicates districts/cities

that have a high average balancing fund, but low PAD. The districts included in quadrant 1 are

Jayawijaya district, Yahukimo district, Boven Digul district, and Teluk

Bintuni district. The four regencies are located in Papua and West Papua. Furthermore,

in quadrant II are districts/cities that have high PAD and high balancing funds. Several

regencies/cities that fall into this quadrant are the City of Yogyakarta, Merauke Regency, and

Aceh Tamiang Regency. Quadrant III is the quadrant for districts/cities that have a low average

balancing fund and PAD. Several districts that fall into quadrant III are Gayo Lues and

Simeulue districts. Whereas in quadrant IV, the districts/cities that are included are

districts/cities that have low balancing funds, but have high PAD. Several regencies/cities that

fall into quadrant IV are West Aceh, Southwest Aceh and Central Aceh districts.

Quadrant III is the quadrant for districts/cities that have a low average balancing fund

and PAD. Several districts that fall into quadrant III are Gayo Lues and Simeulue districts.

Whereas in quadrant IV, the districts/cities that are included are districts/cities that have low

balancing funds, but have high PAD. Several regencies/cities that fall into quadrant IV are West

Aceh, Southwest Aceh and Central Aceh districts.

Quadrant Based on PAD and District/City Balancing Funds in Cluster 1 2013 – 2019

Statistik Deskriptif Variabel Pajak, Retribusi, Belanja Langsung dan Belanja Tidak Langsung

Tahun 2013-2019 di Kluster 1

No.

Indikator

Pajak

Retribusi

Belanja Langsung

Belanja Tidak Langsung

(1)

(2)

(3)

(4)

(5)

(6)

Rata-Rata

34,251,011,114

12,100,085,481

566,209,385,262

600,446,660,700

Standar deviasi

86,138,410,862

13,548,729,606

236,235,838,379

277,915,787,846

Nilai Maksimum

650,084,598,498

71,737,834,712

1,697,549,559,042

1,567,946,127,944

Nilai Minimum

42,630,640

6,040,000

135,556,540,511

77,239,982,651

Sumber: Data Diolah (2020).

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April 2022

783 http://eduvest.greenvest.co.id

2. Overview of Financial Independence in Regencies/Cities in Cluster II Period 2013-2020

Districts/cities classified into cluster II are districts/cities that have a ratio of PAD to

balancing funds of more than 50 percent. In cluster II, there are 23 districts/cities. During the

period 2013 to 2019, the average balancing fund sent by the central government to

districts/cities in cluster II was 1.2 trillion annually with a standard deviation of 530.72 billion

rupiah. The largest balancing fund value was given to the city of Bandung in 2016, while the

smallest balancing fund value was given to the city of Kendari in 2016.

PAD as a source of regional original income is the main driving force and motor of the

economy in a region. The average amount of district/city PAD during the period 2013 to 2019

in cluster II is 1.26 trillion rupiah with a standard deviation of 1.08 trillion rupiah. The largest

PAD during this period was in Bandung district in 2019, while the smallest PAD was in

Morowali district in 2013.

The average amount of tax collected by districts/cities in cluster II is 959.5 billion

rupiah per year with a standard deviation of 842.16 billion. The largest amount of taxes

collected during the 2013 – 2019 period was in Bandung district in 2019, while the smallest

was in Morowali district in 2014. In terms of expenditure, the city of Surabaya is the city that

has the largest direct expenditure in cluster II, which is 6.99 trillion rupiah in 2019, while the

smallest was Morowali district in 2014. During the 2013-2019 period the average district/city

direct expenditure in cluster II was 1.7 trillion rupiah with a standard deviation of 1.09 trillion

rupiah.

Based on the quadrant analysis in cluster II where the Y axis is the average district/city

balancing fund during the period 2013 to 2019 and the X axis is the average district/city PAD

during the period 2013 to 2019. Quadrant I is the quadrant for districts with equal balance funds.

high but low PAD funds. Only the city of Palembang is included in quadrant I. Furthermore,

quadrant II is the division of territory for districts/cities that have high balancing funds and high

PAD funds. Several regencies/cities that are included in quadrant II are Tangerang Regency,

Bekasi City, and Surabaya City. The city of Surabaya has the largest average PAD compared

to other districts/cities in cluster II.

Quadrant III is an area for districts/cities that have PAD and balancing funds that are

smaller than the average balancing funds and PAD for districts/cities in cluster II. Some areas

that fall into cluster II are Morowali district, Cirebon city, and Denpasar city. While quadrant

Analisis Deskriptif Dana Perimbangan, PAD, dan Pertumbuhan PDRB

Tahun 2013-2019 di Kluster II

No.

Indikator

Dana

Perimbangan

PAD

Pertumbuhan

PDRB

(1)

(2)

(3)

(4)

(5)

Rata-Rata

1,295,126,479,120

1,264,011,791,837

6.90

Standar deviasi

530,727,465,332

1,084,479,360,675

5.22

Nilai

Maksimum

2,802,754,414,240

6,791,520,731,810

67.82

Nilai Minimum

104,832,631,500

36,673,796,750

0.09

Sumber: Data Diolah (2020).

Analisis Deskriptif Pajak, Retribusi, Belanja Langsung, dan Belanja Tidak Langsung

Tahun 2013-2019 di Kluster II

No.

Indikator

Pajak

Retribusi

Belanja Langsung

Belanja Tidak

Langsung

(1)

(2)

(3)

(4)

(5)

(6)

Rata-Rata

959,509,762,673

87,480,057,277

1,703,674,379,480

1,426,347,143,806

Standar

deviasi

842,162,040,655

76,257,076,254

1,094,899,702,696

666,787,297,370

Nilai

Maksimum

4,217,319,393,186

557,966,574,670

6,993,376,613,816

3,407,308,688,191

Nilai

Minimum

5,478,277,179

10,925,787,599

215,955,286,716

258,188,282,984

Sumber: Data Diolah (2020).

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

Analysis of Factors Affecting Regional Financial Independence in the Framework of A

Policy Strategy for Increasing Regional Original Income 784

IV is a district/city that has a large PAD, but has a small balancing fund. Several counties. Cities

that fall into quadrant IV are Badung Regency and South Tangerang City.

The regencies/cities in cluster III that generate the largest taxes are Deli Serdang

Regency in 2019, which is 4.2 trillion, while the smallest is Padang Panjang City in 2013. The

average amount of taxes obtained by regencies/cities in cluster III is 959, 5 billion rupiah per

year with a standard deviation of 842.16 billion rupiah. For direct expenditures, Bandung

district in 2019 had the largest direct expenditure which reached 6.99 trillion rupiah, while the

smallest was Padang Panjang city in 2013. The average value of direct spending made by

districts/cities in cluster III was 1.7 trillion rupiah per year..

Quadrant I is an area for regencies/cities that have high balancing funds and low PAD.

One of the regencies/cities in quadrant I is West Lombok regency. Next is quadrant II, where

there is the city of Pekanbaru and the city of Padang in it. Quadrant II is an area with large PAD

and large balancing funds. Quadrant III is an area where PAD and balancing funds are relatively

small compared to the average for all districts/cities in cluster III. Some of these cities/districts

are the city of Salatiga and the city of Kediri. Finally, districts/cities in quadrant IV are

districts/cities that have large PAD, but small balancing funds. The regencies/cities that fall into

quadrant IV are Karimun regencies and the city of Pontianak.

4. Overview of Financial Independence in Regencies/Cities in Cluster IV Period 2013-2020

There are 292 regencies/cities in cluster IV, of which regencies/cities classified into

cluster IV are regencies/cities that have a proportion of PAD to balancing funds of less than 25

percent.

The average amount of balancing funds received by districts/cities in cluster III is 1.18

trillion with a standard deviation of 2.1 trillion. The largest balancing fund amount was 71.5

trillion rupiah, namely Tanggamus district in 2013, while the smallest amounted to 100 billion

was in Landak district in 2017. For PAD funds, the average district/city in cluster III was 97.5

billion rupiah with standard deviation of 88.23 billion. The largest PAD generated, which

amounted to 643 billion in Pasuruan district in 2019, while the smallest was in South Buru

district in 2013.

Analisis Deskriptif Pajak, Retribusi, Belanja Langsung, dan Belanja Tidak Langsung

Tahun 2013-2019 di Kluster III

Indikator

Pajak

Retribusi

Belanja Langsung

Belanja Tidak

Langsung

(1)

(2)

(3)

(4)

(5)

(6)

Rata-Rata

959,509,762,673

87,480,057,277

1,703,674,379,480

1,426,347,143,806

Standar deviasi

842,162,040,655

76,257,076,254

1,094,899,702,696

666,787,297,370

Nilai

Maksimum

4,217,319,393,186

557,966,574,670

6,993,376,613,816

3,407,308,688,191

Nilai Minimum

5,478,277,179

10,925,787,599

215,955,286,716

258,188,282,984

Sumber: Data Diolah (2020).

Analisis Kuadran di Kluster III

Sumber: Data Diolah (2020).

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April 2022

785 http://eduvest.greenvest.co.id

3. Overview of Financial Independence in Regencies/Cities in Cluster III 2013-2020

Period

Districts/cities classified into cluster III are districts/cities that have a proportion of

PAD compared to the Balancing Fund between 25 percent – 50 percent. In cluster III there are

53 districts/cities. The average amount of balancing funds received by districts/cities in cluster

III is 1.26 trillion rupiah with a standard deviation of 1.23 trillion rupiah. Meanwhile, the

average PAD generated by districts/cities in cluster III is 291.93 billion rupiahs with a standard

deviation of 188.37 billion rupiahs. Another indicator is GRDP growth, where the average

GRDP growth for districts/cities in cluster III is 5.56 percent with a standard deviation of 1.12

percent.

The largest balancing fund in cluster III of 9.7 trillion was in Buleleng district in 2015,

while the smallest was in Pontianak City in 2016. For PAD funds generated during the period

2013 to 2019, Pekanbaru City was the city with the largest PAD in cluster II in in 2016, while

the smallest was North Lombok Regency in 2013.

Analisis Kuadran Dana Perimbangan dan PAD Kabupaten/Kota di Kluster II

Sumber: Data Diolah (2020).

Analisis Deskriptif Dana Perimbangan, PAD, dan Pertumbuhan PDRB

Tahun 2013-2019 di Kluster III

No.

Indikator

Dana Perimbangan

PAD

Pertumbuhan

PDRB

(1)

(2)

(3)

(4)

(5)

Rata-Rata

1,262,950,445,970.32

291,936,813,464.31

5.56

Standar deviasi

1,230,397,365,354.28

188,373,365,755.51

1.12

Nilai Maksimum

9,770,182,860,000.00

1,243,438,534,336.00

9.30

Nilai Minimum

102,243,025,535.00

35,285,969,506.00

-0.86

Sumber: Data Diolah (2020).

Analisis Deskriptif Dana Perimbangan, PAD, dan Pertumbuhan PDRB

Tahun 2013-2019 di Kluster IV

No.

Indikator

Dana Perimbangan

PAD

Pertumbuhan

PDRB

(1)

(2)

(3)

(4)

(5)

Rata-Rata

1,186,221,913,009.08

97,504,884,344.94

5.46

Standar deviasi

2,104,071,966,592.22

88,231,996,091.18

2.11

Nilai Maksimum

71,567,021,624,736.00

643,350,343,365.00

38.22

Nilai Minimum

100,235,786,200.00

1,490,176,000.00

-9.66

Sumber: Data Diolah (2020).

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

Analysis of Factors Affecting Regional Financial Independence in the Framework of A

Policy Strategy for Increasing Regional Original Income 786

In addition to the indicators of balancing funds, PAD, and economic growth, there are

other indicators in measuring the financial independence of a region, namely the value of taxes,

user charges, direct spending, and indirect spending. The average amount of tax generated is

31.16 billion rupiah with a standard deviation of 41.73 billion. The largest amount of tax

generated was 383.74 billion rupiah, namely Pasuruan district in 2017, while the district/city

that had the smallest tax was Merangin district in 2016. From the expenditure side, it can be

seen that the average amount of direct expenditure of districts/cities in clusters IV is 552.77

billion rupiah with a standard deviation of 325.81 billion. The largest direct expenditure value

of 4.5 trillion was in the Kutai Kartanegara district in 2013, while the smallest direct

expenditure was in the Buru district in 2013.

The division of quadrants in regencies/cities in cluster IV is based on the average value

of balancing funds and PAD. Quadrant I is an area with high balancing funds but relatively

small PAD generated. There are several regencies/cities in cluster I, namely Sanggau Regency

and Balangan Regency. While quadrant II is a district/city that has a high balance of funds and

PAD, some districts/cities in quadrant II are Kutai Kartanegara Regency and Brebes Regency.

Furthermore, regencies/cities in quadrant III are regencies/cities that have balancing

funds and relatively small PAD. Several districts/cities in quadrant III are Samosir district and

Pariaman city. Quadrant IV is a district/city with relatively large PAD funds and small

balancing funds, some of which are Humbang Hasundutan Regency and Binjai City.

5. Panel Data Regression Analysis and Model Interpretation per Cluster

In conducting the analysis, the panel data regression and interpretation of the resulting

model went through several test stages. The panel data regression model consists of 3, namely

the common effect model, the fixed effect model, and the random effect model. In determining

the best model, it is necessary to do some statistical tests on the regression models that are

formed.

Analisis Deskriptif Pajak, Retribusi, Belanja Langsung, dan Belanja Tidak Langsung

Tahun 2013-2019 di Kluster IV

No.

Indikator

Pajak

Retribusi

Belanja Langsung

Belanja Tidak

Langsung

(1)

(2)

(3)

(4)

(5)

(6)

Rata-Rata

31,164,127,774.83

11,519,956,880.68

552,775,756,782.80

694,907,752,579.50

Standar deviasi

41,735,964,478.43

12,216,031,655.96

325,817,763,832.60

402,016,621,736.01

Nilai Maksimum

383,743,763,642.93

162,923,495,725.00

4,539,531,892,452.00

2,843,063,170,398.20

Nilai Minimum

184,700,111.00

175,625,100.00

94,772,518,311.00

91,740,905,568.00

Sumber: Data Diolah (2020).

Analisis Kuadran di Kluster IV

Sumber: Data Diolah (2020).

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April 2022

787 http://eduvest.greenvest.co.id

In Cluster I, the first test is to choose the best model between the common effect model

and the fixed effect model using the Chow test. The hypothesis is as follows:

H0 : 1 = 2 = . . . = i = (common effect model)

H1 : i j (fixed effect model)

Where is the resulting residual. Based on the chow test, the following results were

obtained:

Prob value. The p-value of the test in cross-section F and cross-section chi square of

0.00 gives the decision that H0 is rejected, so it is concluded that the fixed effect model is the

best model. Furthermore, testing is carried out through the Hausman test to choose the best

model between the fixed effect model and the common effect model.

The hypothesis in the Hausman test is as follows:

H0 : E(uit | Xit ) = 0 (random effect model)

H1 : E(uit |Xit ) 0 (fixed effect model)

Hausman test results are displayed as follows:

The p-value of the Hausman test is 0.00, this indicates that the decision is to reject H0,

so it can be concluded that the best model is the fixed effect model. Based on the results of the

Chow test and Hausman test, the best model is the fixed effect model.

After selecting the fixed effect model as the model that will estimate the parameters.

The next step is to test the classical assumptions to ensure that the resulting estimate is BLUE

(Best Linear Unbiased Estimator).

The p-value of the Hausman test is 0.00, this indicates that the decision is to reject H0,

so it can be concluded that the best model is the fixed effect model. Based on the results of the

Chow test and Hausman test, the best model is the fixed effect model. After selecting the fixed

effect model as the model that will estimate the parameters. Next is to test the classical

assumptions to ensure that the resulting estimate is BLUE (Best Linear Unbiased Estimator)

which includes normality test and multicollinearity test.

The estimation method in the fixed effect model uses the Generalized Least Square

(GLS) method. The use of the GLS method causes the resulting model to be robust to the

problems of heteroscedasticity and autocorrelation. The first test is to test the normality of the

resulting residuals. The results of the normality test using the jarque fallow test show a p-value

of 0.032. This value indicates that at a significance level of 10 percent, the resulting residuals

are normally distributed. The test results are as shown in the following table:

Tabel Hasil Uji Chow pada Kluster 1

Effects Test

Statistic

d.f.

Prob.

(1)

(2)

(3)

(4)

Cross-section F

11,55386721

(46,276)

0.00

Cross-section Chi-

square

353,1863725

0.00

Sumber: Data Diolah (2020).

Tabel Hasil Uji Hausman pada Kluster 1

Test Summary

Chi-Sq. Statistic

Chi-Sq. d.f.

Prob.

(1)

(2)

(3)

(4)

Cross-section

random

32.85056741

0.00

Sumber: Data Diolah (2020).

Uji Normalitas pada Kluster 1

Sumber: Data Diolah (2020).

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

Analysis of Factors Affecting Regional Financial Independence in the Framework of A

Policy Strategy for Increasing Regional Original Income 788

Next is to do multicollinearity testing. The test results show that the VIF value for each

independent variable is less than 5, so it can be concluded that the model is free from

multicollinearity violations.

Panel Data Regression Model Interpretation in Cluster I

In Cluster 1, the fixed effect model is the model chosen to estimate the population. The

classical assumptions have also been met, so the next step is to interpret the results in the panel

data regression model. The results of the panel data regression model can be seen in the table

as follows:

The results of panel regression analysis on data in cluster 1 show that the variables of

direct expenditure, indirect expenditure, taxes and levies as well as balancing funds have a

positive and significant influence on the regional original income variable with a significance

level of 10%. Meanwhile the regional economic growth variable does not have a significant

effect on regional independence which is proxied by the regional original income growth

variable. The results of the panel data regression analysis conclude that the variables of direct

expenditure, indirect expenditure, taxes, levies and balancing funds have a positive and

significant influence on increasing regional financial independence for districts/cities located

in Cluster I.

The effect of the provision of positive balancing funds on regional financial

independence as proxied by regional original growth indicates that the provision of balancing

funds actually has a positive impact on regional financial independence even though the

coefficient is relatively small. The results of this study are in line with research conducted by

several previous studies, including Sulistyo (2017) who conducted research on the effect of

balancing funds on financial independence. The results of the study found that the provision of

balancing funds had a positive and significant effect on increasing regional financial

independence.

In Cluster II, the first test in choosing the best model in panel data regression is to

perform the Chow test to choose between the common effect model and the fixed effect model.

The following is the hypothesis in the Chow test

H0 : 1 = 2 = . . . = i = (common effect model)

H1 : i j (fixed effect model)

Hasil Uji Multikolinearitas pada Kluster 1

Coefficient

Centered

Variable

Variance

VIF

(1)

(2)

(3)

3.510741862

LOG(BL)

0.002697325

2.029957162

LOG(BTL)

0.005149733

1.787063299

LOG(DAPER)

0.001015084

1.07552948

LOG(PAJAK)

0.001372237

2.047077856

PDRB

5.26E-05

1.050277013

LOG(RETRIBUSI)

0.000431242

1.084637505

Sumber: Data Diolah (2020).

Model Regresi Data Panel di Kluster 1

Variable

Coefficient

Std. Error

t-Statistic

Prob.

(1)

(2)

(3)

(4)

(5)

-19.470

1.874

-10.391

0.000

LOG(BL)

0.500

0.052

9.635

0.000

LOG(BTL)

0.886

0.072

12.346

0.000

LOG(DAPER)

0.058

0.032

1.833

0.068

LOG(PAJAK)

0.188

0.037

5.063

0.000

PDRB

0.005

0.007

0.698

0.486

LOG(RETRIBUSI)

0.035

0.021

1.697

0.091

Uji F

244.038

Prob. Uji F

0.000

R-Squared

0.979

Sumber: Data Diolah (2020).

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April 2022

789 http://eduvest.greenvest.co.id

Where is the resulting residual. Based on the chow test, the following results were

obtained:

Prob value. the p-value of the test in cross-section F and cross-section Chi-square is

0.00. The p-value which is less than 5 percent alpha indicates that at the 5 percent significance

level, the resulting decision is to reject H0 so that it can be concluded that the fixed effect model

is the best model. Next, perform a test to choose between the fixed effect model and the random

effect model through the Hausman test. The hypothesis in the Hausman test is as follows:

H0 : E(uit | Xit ) = 0 (random effect model)

H1 : E(uit |Xit ) 0 (fixed effect model)

Hausman test results are displayed as follows:

The p-value of the Hausman test is 0.00, this indicates that the decision is to reject H0,

so it can be concluded that the best model is the fixed effect model. Based on the results of the

Chow test and Hausman test, the best model is the fixed effect model. After selecting the fixed

effect model as the model that will estimate the parameters. Next is to test the classical

assumptions to ensure that the resulting estimate is BLUE (Best Linear Unbiased Estimator)

which includes normality test and multicollinearity test.

The estimation method in the fixed effect model uses the Generalized Least Square

(GLS) method. The use of the GLS method causes the resulting model to be robust to the

problems of heteroscedasticity and autocorrelation. The first test is to test the normality of the

resulting residuals. The results of the normality test using the fallow jarque test showed a p-

value of 0.49. This value indicates that at a significance level of 0.05, the resulting residuals are

normally distributed. The test results are as follows:

The next step is to perform multicollinearity testing to ensure that the independent

variables have no relationship or are independent. The multicollinearity test results show that

all variables have a VIF value of less than 5, so the multicollinearity assumption is met.

Interpretation of Harvest Data Regression Model in Cluster II

Hasil Uji Chow pada Kluster 2

Effects Test

Statistic

d.f.

Prob.

(1)

(2)

(3)

(4)

Cross-section F

11.73645867

(22,132)

0.00

Cross-section Chi-

square

174.5019211

0.00

Sumber: Data Diolah (2020).

Uji Hausman pada Klaster II

Test Summary

Chi-Sq. Statistic

Chi-Sq. d.f.

Prob.

(1)

(2)

(3)

(4)

Cross-section

random

105.461270

0.00

Sumber: Data Diolah (2020).

Uji Normalitas pada Kluster II

Sumber: Data Diolah (2020).

Uji Multikolinearitas pada Kluster II

Coefficient

Centered

Variable

Variance

VIF

(1)

(2)

(3)

3.2264617

LOG(BL)

0.0019274

2.859731544

LOG(BTL)

0.0048470

1.833430785

LOG(DAPER)

0.0019739

1.592257329

LOG(PAJAK)

0.0031565

4.513246917

PDRB

0.0000599

1.073575578

LOG(RETRIBUSI)

0.0009212

1.057615318

Sumber: Data Diolah (2020).

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

Analysis of Factors Affecting Regional Financial Independence in the Framework of A

Policy Strategy for Increasing Regional Original Income 790

After selecting the best model, the fixed effect model was chosen as the best model

using the GLS (Generalized Least Square) estimation method. From the model obtained, it is

known that the coefficient of determination (R2) is 0.9869, meaning that all independent

variables in the model are able to explain the dependent variable by 98.69 percent. The F test

(Simultaneous Test) to determine whether the model is suitable or not indicates that the

resulting model is suitable, this is based on the p-value of the F test of 0.00 which is smaller

than the 5 percent alpha value.

Based on the selected fixed effect model, there are 3 variables that do not have a

significant effect on PAD growth, namely the growth of balancing funds, GRDP growth and

retribution growth. Meanwhile, the variables of direct expenditure growth, indirect expenditure

growth, and tax growth have a significant effect on PAD growth at a significance level of 5

percent.

If there is an increase in direct expenditure growth of 1 percent, it will cause an increase

in PAD growth of 0.116 percent. In cluster 2, the biggest cause of the increase in financial

independence (growth in PAD) is the significant tax growth. If there is an increase in tax growth

of 1 percent, it will increase PAD growth by 1.029 percent. These results indicate that the

variables that affect the growth of district/city PAD in Cluster II are influenced by the growth

of the direct expenditure variables, indirect spending and tax growth.

The results of panel regression analysis which show that the growth of balancing funds

do not significantly affect the growth of PAD is in line with research conducted by Stone (2015)

which examines the effect of balancing funds on the financial condition of the government. The

results of the study found that the provision of balancing funds or fiscal decentralization had no

effect on increasing government finances (Stone, 2015). The provision of insignificant

balancing funds encourages PAD, this is also in line with research conducted by Budianto and

Alexander in 2016 (Budianto & Sos, 2016).

In Cluster III, the first test in choosing the best model in panel data regression is to

perform the Chow test to choose between the common effect model and the fixed effect model.

The following is the hypothesis in the Chow test

H0 : 1 = 2 = . . . = i = (common effect model)

H1 : i j (fixed effect model)

Where is the resulting residual. Based on the chow test, the following results were

obtained:

Prob value. the p-value of the test in cross-section F and cross-section Chi-square is

0.00. The p-value which is less than 5 percent alpha indicates that at the 5 percent significance

level, the resulting decision is to reject H0 so that it can be concluded that the fixed effect model

is the best model. Next, perform a test to choose between the fixed effect model and the random

effect model through the Hausman test. The hypothesis in the Hausman test is as follows:

H0 : E(uit | Xit ) = 0 (random effect model)

H1 : E(uit |Xit ) 0 (fixed effect model)

Model Regresi Data Panel pada Kluster II

Variable

Coefficient

Std. Error

t-Statistic

Prob.

(1)

(2)

(3)

(4)

(5)

-8.965

1.796

-4.991

0.000

LOG(BL)

0.116

0.044

2.653

0.009

LOG(BTL)

0.212

0.070

3.050

0.003

LOG(DAPER)

-0.030

0.044

-0.674

0.502

LOG(PAJAK)

1.029

0.056

18.313

0.000

PDRB

-0.004

0.008

-0.530

0.597

LOG(RETRIBUSI)

0.010

0.030

0.345

0.730

Uji F

355.294

Prob. Uji F

0.000

R-Squared

0.9869

Sumber: Data Diolah (2020).

Tabel Hasil Uji Chow pada Kluster III

Effects Test

Statistic

d.f.

Prob.

(1)

(2)

(3)

(4)

Cross-section F

122.698732

(52,312)

0.00

Cross-section Chi-

square

421.714305

0.00

Sumber: Data Diolah (2020).

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April 2022

791 http://eduvest.greenvest.co.id

Hausman test results are displayed as follows:

The p-value of the Hausman test is 0.00, less than 5 percent, this indicates that the

decision is to reject H0, so it can be concluded that the best model is the fixed effect model.

Based on the results of the Chow test and Hausman test, the best model is the fixed effect model.

After selecting the fixed effect model as the model that will estimate the parameters. Next is to

test the classical assumptions to ensure that the resulting estimate is BLUE (Best Linear

Unbiased Estimator) which includes normality test and multicollinearity test.

The estimation method in the fixed effect model uses the Generalized Least Square

(GLS) method. The use of the GLS method causes the resulting model to be robust to the

problems of heteroscedasticity and autocorrelation. The first test is to test the normality of the

resulting residuals. The results of the normality test using the fallow jarque show that the

resulting residuals are normally distributed at a significance level of 10 percent, where the

resulting p-value is 0.0106.

The estimation method used in the fixed effect model is Generalized Least Square

(GLS) which is robust against heteroscedasticity and autocorrelation problems, so that the

classical assumption tests carried out are normality tests and multicollinearity tests. The

following are the results of the multicollinearity test on the independent variables in cluster III.

All independent variables in the research model have a VIF value of less than 5, so it

can be concluded that there is no relationship between independent variables in the model, so

the non-multicollinearity assumption is fulfilled for the panel data regression model in cluster

III.

IInterpretation of Panel Data Regression Model in Cluster III

To ensure that the selected model is a suitable model or fit, it is necessary to carry out

simultaneous testing (F test) in the selected model. The magnitude of the F test is 159,405 with

a p-value of 0.00. The p-value of less than 5 percent indicates that at the 5 percent significance

Tabel Uji Hausman pada Kluster III

Test Summary

Chi-Sq. Statistic

Chi-Sq. d.f.

Prob.

(1)

(2)

(3)

(4)

Cross-section

random

81.767680

0.00

Sumber: Data Diolah (2020).

Tabel Hasil Uji Normalitas di Kluster III

Sumber: Data Diolah (2020).

Tabel Uji Multikolinearitas pada Kluster III

Coefficient

Centered

Variable

Variance

VIF

(1)

(2)

(3)

2.319944

LOG(BL)

0.001171

3.599559

LOG(BTL)

0.002838

1.536114

LOG(DAPER)

0.000333

1.112213

LOG(PAJAK)

0.00128

3.972653

PDRB

7.41E-05

1.0931

LOG(RETRIBUSI)

0.000701

1.024803

Sumber: Data Diolah (2020).

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

Analysis of Factors Affecting Regional Financial Independence in the Framework of A

Policy Strategy for Increasing Regional Original Income 792

level, the resulting regression model is fit. The value of the coefficient of determination

(R2) of 0.9673 indicates that all independent variables in the model are able to explain the

diversity of the dependent variable by 96.73 percent.

In cluster III, all independent variables have a significant effect on the dependent

variable (PAD growth) at a significance level of 5 percent. If there is a 1 percent increase in all

independent variables in the model, then the largest increase in PAD growth is caused by tax

growth, namely increasing PAD growth by 0.716 percent.

However, there are 3 variables that have a negative impact on the growth of PAD as a

proxy for financial independence, namely the growth of balancing funds, GRDP growth, and

retribution growth. If there is an increase of 1 percent in each of these independent variables, it

will cause a decrease in PAD growth in districts/cities that fall into cluster III.

The sign of the coefficient of the effect of balancing funds on regional financial

independence as a proxy for negative PAD growth indicates that the provision of balancing

funds actually has an impact on decreasing regional financial independence. In other words, the

provision of balancing funds encourages dependence on districts/cities in Cluster III. This is in

line with the research conducted by Said (2019), the results of this study indicate that the

provision of balancing funds has an impact on the dependence of local governments on

balancing funds from the central government.

Furthermore, Arbani (2020) emphasized that every year the provision of balancing

funds to regional governments should provide assistance so that regions with shortages can be

helped. However, this actually has a negative impact, namely the dependence of the region on

balancing funds in meeting regional revenues. This actually has the effect of deviating from the

original goal of providing balancing funds so that local governments can be financially

independent. The high degree of centralization in the taxation sector, and very low taxes

received by the regions as well as the lack of role for regionally-owned enterprises have made

the regions highly dependent on fiscal transfer funds from the central government every year.

In Cluster IV, the first test in choosing the best model in panel data regression is to

perform the Chow test to choose between the common effect model and the fixed effect model.

The following is the hypothesis in the Chow test

H0 : 1 = 2 = . . . = i = (common effect model)

H1 : i j (fixed effect model)

Where is the resulting residual. Based on the chow test, the following results were

obtained:

Prob value. the p-value of the test in cross-section F and cross-section Chi-square is

0.00. The p-value which is less than 5 percent alpha indicates that at the 5 percent significance

level, the resulting decision is to reject H0 so that it can be concluded that the fixed effect model

Model Regresi Data Panel pada Kluster III

Variable

Coefficient

Std. Error

t-Statistic

Prob.

(1)

(2)

(3)

(4)

(5)

-7.189

1.523

-4.720

0.000

LOG(BL)

0.301

0.034

8.791

0.000

LOG(BTL)

0.360

0.053

6.758

0.000

LOG(DAPER)

-0.038

0.018

-2.070

0.039

LOG(PAJAK)

0.716

0.036

20.017

0.000

PDRB

-0.019

0.009

-2.211

0.028

LOG(RETRIBUSI)

-0.070

0.026

-2.628

0.009

Uji F

159.405

Prob. Uji F

0.000

R-Squared

0.9673

Sumber: Data Diolah (2020).

Tabel Hasil Uji Chow pada Kluster IV

Effects Test

Statistic

d.f.

Prob.

(1)

(2)

(3)

(4)

Cross-section F

7.972643

(291,1746)

0.00

Cross-section Chi-

square

1727.878752

291

0.00

Sumber: Data Diolah (2020).

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April 2022

793 http://eduvest.greenvest.co.id

is the best model. Next, perform a test to choose between the fixed effect model and the random

effect model through the Hausman test. The hypothesis in the Hausman test is as follows:

H0 : E(uit | Xit ) = 0 (random effect model)

H1 : E(uit |Xit ) 0 (fixed effect model)

Hausman test results are displayed as follows:

The p-value of the Hausman test is 0.00, less than 5 percent, this indicates that the

decision is to reject H0, so it can be concluded that the best model is the fixed effect model.

Based on the results of the Chow test and Hausman test, the best model is the fixed effect model.

After selecting the fixed effect model as the model that will estimate the parameters. Next is to

test the classical assumptions to ensure that the resulting estimate is BLUE (Best Linear

Unbiased Estimator) which includes normality test and multicollinearity test.

The estimation method in the fixed effect model uses the Generalized Least Square

(GLS) method. The use of the GLS method causes the resulting model to be robust against

heteroscedasticity and autocorrelation problems. The first test is to test the normality of the

resulting residuals.

Normality test

The central limit theorem states that the sampling distribution curve for a sample size

of 30 or more will have all the properties of a normal distribution. Based on the explanation in

the theorem, the number of samples in cluster IV is 292 samples. Therefore, it can be said that

the number of samples in cluster IV has met the requirements to meet the normality test.

Multicollinearity Test

The assumption of non-multicollinearity is one of the assumptions that requires that

there is no relationship between independent variables in the research model. The measurement

method is through comparison of the obtained VIF values. If the VIF value is less than 5, it can

be said that there is no relationship between the independent variables in the model so that the

multicollinearity assumption is met. The VIF value in table 23, shows that all VIF values for

the independent variables in the model are less than 5, so it can be concluded that there is no

multicollinearity in the research model.

In estimating the parameters in the research model, it is necessary to do an F test to

find out whether the resulting model is suitable or not. The F test value of 237,132 with a p-

value of 0.00 decided to reject H0 thus concluding that the resulting regression model was

appropriate. The R2 value of 0.9758 indicates that all independent variables in the model are

able to explain the diversity of the dependent variable by 97.58 percent.

Tabel Uji Hausman pada Kluster IV

Test Summary

Chi-Sq. Statistic

Chi-Sq. d.f.

Prob.

(1)

(2)

(3)

(4)

Cross-section

random

72.512756

0.00

Sumber: Data Diolah (2020).

Tabel Uji Multikolinearitas pada Kluster IV

Coefficient

Centered

Variable

Variance

VIF

(1)

(2)

(3)

0.473688

LOG(BL)

0.000303

2.519807

LOG(BTL)

0.000816

2.590688

LOG(DAPER)

7.54E-05

1.053739

LOG(PAJAK)

0.000153

2.896851

PDRB

5.04E-06

1.043912

LOG(RETRIBUSI)

5.85E-05

1.088574

Sumber: Data Diolah (2020).

Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang

Analysis of Factors Affecting Regional Financial Independence in the Framework of A

Policy Strategy for Increasing Regional Original Income 794

The levy growth variable has no significant effect at the 5 percent significance level,

while other variables such as direct spending growth, indirect spending growth, balancing fund

growth, and tax growth have a significant effect on PAD growth at a 5 percent significance

level. If there is a 1 percent increase in direct growth, it will have an impact on an increase in

PAD by 0.749 percent. The direct expenditure growth variable is the variable that has the

greatest influence on the increase in PAD growth. In cluster IV, the growth of the balancing

fund has a positive and significant effect on the growth of PAD, namely if there is a 1 percent

increase in the balancing fund, it will cause a 0.028 percent increase in the growth of PAD.

The results of this study found that the effect of balancing funds on regional financial

independence as a proxy for PAD growth was positive and significant for districts/cities in

cluster IV. This means that every time there is an increase in the distribution of balancing funds,

it can increase regional financial independence. The results of this study are in line with research

conducted by several previous studies such as Falurida and Pratiwi (2017), Simanjuntak and

Mukhlis (2016) and Suprantiningrum (2015) which state that the provision of balancing funds

encourages regional financial independence.

Other variables that determine or affect regional financial independence based on panel

data regression analysis for clusters I, II, III and IV are direct and indirect expenditure variables.

Based on this, the management of district/city government expenditures in all data clusters must

be directed to quality direct and indirect expenditures. Changes in the pattern of APBD

management that are more rational and lead to productive investment can encourage regional

financial independence and increase economic growth.

Comparative Analysis of Financial Independence between Clusters

The fiscal decentralization policy aims to minimize the fiscal gap and achieve equity

among regencies/cities in Indonesia. During the implementation of the decentralization policy,

inequality between districts/cities still persists. This problem is caused by the low source of

district/city revenue for other areas. This happens in areas that do not have a massive economic

center. The analysis of the effect of balancing funds on regional financial independence cannot

be carried out simultaneously for all districts/cities. Therefore, it is necessary to divide each

district/city in each cluster based on certain criteria, namely cluster I for areas receiving

additional funds other than balancing funds, cluster II for areas with a ratio of PAD to balancing

funds of more than 50%, cluster III for areas with the ratio of PAD to balancing funds is 25%-

50%, and cluster IV is for areas with a ratio of PAD to balancing funds less than 25%.

The research model produced in each cluster gives different results in determining the

factors that affect regional financial independence. The factors that encourage regional

independence in cluster 1 are direct spending, indirect spending, taxes and levies and balancing

funds. Direct and indirect spending is one of the government's responsibilities in regional

autonomy. The effect of direct and indirect spending on financial independence, which is

positive and significant, indicates that district/city spending priorities in cluster 1 are oriented

towards the development of facilities and infrastructure so as to increase PAD as a proxy for

regional independence. Local revenue in the form of taxes and levies has a positive impact on

regional independence. This indicates that districts/cities located in cluster 1 have fiscal

capacity that is able to support their regional needs, so that central government transfer funds

Tabel Model Regresi Data Panel pada Kluster IV

Variable

Coefficient

Std. Error

t-Statistic

Prob.

(1)

(2)

(3)

(4)

(5)

-13.815

0.688

-20.073

0.000

LOG(BL)

0.401

0.017

23.014

0.000

LOG(BTL)

0.749

0.029

26.220

0.000

LOG(DAPER)

0.028

0.009

3.257

0.001

LOG(PAJAK)

0.303

0.012

24.523

0.000

PDRB

-0.009

0.002

-4.195

0.000

LOG(RETRIBUSI)

-0.009

0.008

-1.230

0.219

Uji F

237.132

Prob. Uji F

0.000

R-Squared

0.9758

Sumber: Data Diolah (2020).

Eduvest – Journal of Universal Studies

Volume 2 Number 4, April 2022

795 http://eduvest.greenvest.co.id

can be allocated to other regions that do not have district/city revenue potential.

Meanwhile in cluster II, the factors that influence regional financial independence are

direct expenditures, indirect expenditures, and taxes. Cluster II is a district/city that has a ratio

of PAD to balancing funds of more than 50%. Tax revenue is the main source of income in

realizing development in the region, so that positive tax growth will increase regional

independence in the cluster.

The factors that influence financial independence in clusters III and IV are the same,

namely direct spending, indirect spending, balancing funds, taxes, GRDP, and user fees.

Regencies/cities located in clusters III and IV have something in common, namely the ratio of

PAD to balancing funds which is less than 50%. Regencies/cities in cluster III and IV areas

must increase regional revenues through taxes and levies. In addition, an increase in regional

spending, both directly and indirectly, can also stimulate the pace of the economy which has an

impact on increasing regional income, thereby increasing the independence of the region.

The difference in the results found in each cluster is due to the different sources of

regional revenue in each cluster. For clusters with regions with relatively high PAD, the

provision of balancing funds does not affect financial independence, so they can be relocated

to regions with low PAD, thereby spurring financial independence in the region. The results of

this study are in line with research conducted by the Ministry of Finance Assistance Team for

Fiscal Decentralization in 2012, which found that there is a need for reformulation of balancing

funds to local governments. Local governments with large fiscal potential but small fiscal needs

will receive relatively small balancing funds. On the other hand, local governments with small

fiscal potential but large fiscal needs will receive relatively large balancing funds.

From the four clusters that have been formed in this study, it can be seen that the factors

that influence financial independence are direct spending, indirect spending, and taxes. The

increase in these three variables can be a stimulus in increasing regional financial independence.

Improving the quality of local government spending, both direct and indirect, can encourage

local government economic growth which can stimulate an increase in regional income. The

increase in regional income can ultimately encourage regional financial independence.

CONCLUSION

The regional autonomy policy, which is accompanied by the provision of balancing

funds, has a goal, one of which is to have an impact on the financial independence of

regency/municipal governments in Indonesia. The results of this study found that the factors

that significantly affect financial independence in Clusters I, III and IV are direct spending,

indirect spending, taxes and balancing funds. Factors that affect financial independence in

Cluster II are almost the same as in Clusters I, III, and IV, namely direct spending, indirect

spending, and taxes. However, for district/city governments in cluster II, the balancing fund

variable does not significantly affect regional financial independence.

The results of panel regression analysis on each research data cluster for districts/cities

show differences in the effect of balancing funds on regional financial independence. Based on

this, the policy strategy that must be issued is to reformulate the calculation of balancing funds

given to local governments. At the same time, local governments must carry out a policy

strategy for relocating local government spending to productive expenditures in the regional

government budget.

Based on the results of this study, the central government together with the House of

Representatives of the Republic of Indonesia (DPR RI) need to sit down together to have serious

talks related to the reformulation of the calculation of the balancing fund. The reformulation of

the calculation of the balancing fund is intended so that the role of the balancing fund as an

equalization grant can play a more significant role in increasing regional financial

independence.

Furthermore, district/city governments in all clusters need to formulate and allocate

direct and indirect expenditures in the APBD. Expenditure relocation, either directly or

 
Vissia Dewi Haptari, Ari Mulianta Ginting, Ferdinand David Aritonang 
 
Analysis of Factors Affecting Regional Financial Independence in the Framework of A 
Policy Strategy for Increasing Regional Original Income   796 
indirectly, must be directed to expenditures that are productive in nature in order to create an 
increase in the economy which in turn can spur increased regional financial independence in 
each district/city in all clusters. 
 
REFERENCES 
 
Antika, I. K. E. Juni. (2018). Analisis Kontribusi Bumd Terhadap Peningkatan Pad Di Kota Bandar 
Lampung  Dalam  Perspektif  Ekonomi  Islam  Pada  Tahun  2006-2015.  Uin  Raden  Intan 
Lampung. 
Asyifa, Laila Siti Nur. (N.D.). Peramalan Realisasi Penerimaan Pemerintah Provinsi Di Indonesia 
Menggunakan Model Grey-Markov (1, 1). Fakultas Sains Dan Teknologi Universitas Islam 
Negeri Syarif Hidayatullah …. 
Budianto,  Heri,  &  Sos,  S.  (2016).  Ilmu  Komunikasi:  Sekarang  Dan  Tantangan  Masa  Depan. 
Prenada Media. 
Fafurida, Fafurida, & Pratiwi, Erliz Nindi. (2017). Financial Independence Of Regencies And Cities 
In Central Java. Economic Journal Of Emerging Markets, 9(2), 199. 
Halim, Abdul. (2004). Akuntansi Keuangan Daerah. 
Maimunah, Mutiara. (2006). Flypaper Effect Pada Dana Alokasi Umum (Dau) Dan Pendapatan 
Asli  Daerah  (Pad) Terhadap  Belanja  Daerah  Pada Kabupaten/Kota  Di Pulau  Sumatera. 
Universitas Gadjah Mada. 
Mustanir, Ahmad, Yasin, Akhmad, Irwan, Irwan, & Rusdi, Muhammad. (2019). Potret Irisan Bumi 
Desa Tonrong Rijang Dalam Transect Pada Perencanaan Pembangunan Partisipatif. Moderat: 
Jurnal Ilmiah Ilmu Pemerintahan, 4(4), 1–14. 
Nilawati, Evi. (2019). Kecenderungan Isi Sistem Informasi Desa Kabupaten Gunungkidul (Village 
Information System Content Trends In Gunungkidul Regency). Jurnal Iptekkom (Jurnal Ilmu 
Pengetahuan & Teknologi Informasi), 21(2), 169–184. 
Nuzulistyan, Kendy Rama, Supriyanto, Agus, & Paramita, Patricia Dhiana. (2017). Pengaruh Pajak 
Daerah, Retribusi Daerah, Hasil Pengelolaan Kekayaan Daerah Yang Dipisahkan Dan Lain–
Lain Pendapatan Daerah Yang Sah Terhadap Pendapatan Asli Daerah Provinsi Jawa Tengah 
(Studi  Kasus  Pada  Dppad  Provinsi  Jawa  Tengah  Periode Tahun  2008-2015).  Journal  Of 
Accounting, 3(3). 
Otsuka, Keijiro, & Kalirajan, Kaliappa. (2012). Fiscal Decentralization And Development Outcomes 
In India: An Exploratory Analysis. 
Siswanto,  A.  (2013).  Identifikasi  Sumber-Sumber  Risiko  Fiskal  Daerah.  Jurnal  Ekonomi 
Pembangunan Vol, 5, 231–240. 
Syahril,  Muhammad,  Parinsi,  Mario  T.,  &  Togas,  Peggy  V.  (2021).  Media  Simulasi  Untuk 
Meningkatkan Hasil Belajar Komputer Dan Jaringan Dasar Siswa Di Smk. Ismart Edu: Jurnal 
Pendidikan Teknologi Informasi, 2(1), 27–31. 
Walpole, Barbara J., Walton, Rowena, Yusen, Merriam, Fearon, David Spencer, & Massaro, John. 
(N.D.). Nursing (Cont.). 
Waluyo,  Joko.  (2007).  Analisis  Crowding  Out  Effect  Penerbitan  Obligasinegara:  Studikasus 
Perekonomian  Indonesia  Periode  1970-2003.  Jurnal  Ekonomi  Pembangunan:  Kajian 
Masalah Ekonomi Dan Pembangunan, 7(1), 91–113. 
Widarjono, Agus. (2013). Ekonometrika Pengantar Dan Aplikasi Disertai Panduan Eviews Buku 
Edisi Keempat. Terjemahan. Jogjakarta: Upp Stim Ykpn. 
 
 
 
 