Manajemen | Fakultas Ekonomi Universitas Maritim Raja Ali Haji 00074910152669163

Bulletin of Indonesian Economic Studies

ISSN: 0007-4918 (Print) 1472-7234 (Online) Journal homepage: http://www.tandfonline.com/loi/cbie20

THE NEW INDONESIAN EQUALISATION TRANSFER
Blane D. Lewis
To cite this article: Blane D. Lewis (2001) THE NEW INDONESIAN EQUALISATION TRANSFER,
Bulletin of Indonesian Economic Studies, 37:3, 325-343, DOI: 10.1080/00074910152669163
To link to this article: http://dx.doi.org/10.1080/00074910152669163

Published online: 17 Jun 2010.

Submit your article to this journal

Article views: 70

View related articles

Citing articles: 8 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=cbie20
Download by: [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RA JA ALI HA JI
TANJUNGPINANG, KEPULAUAN RIAU]

Date: 19 January 2016, At: 22:00

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

Bulletin of Indonesian Economic Studies, Vol. 37, No. 3, 2001: 325–43

THE NEW INDONES IAN EQUALIS ATION TRANS FER
Blane D. Lewis*
Research Triangle Institute, North Carolina
This paper examines the intergovernmental general purpose equalisation transfer
(DAU) that began operating in 2001. First, a description of the mechanism highlights the significant distributional role played by the ‘balancing factor’ relative to
the ‘formula’, and the key differences in distribution procedures for districts/mu nicipalities and provinces. Second, the approach and methods are empirically analysed and evaluated. The sums available to the DAU appear more than sufficient to
meet regional governments’ net expenditure requirements, at least in aggregate, and
given some caveats. The allocation mechanism has some design flaws in both its
structure and the methods it uses to operationalise expenditure needs and fiscal
capacity. Third, an examination of the ‘equalising’ effects of the transfer scheme

shows that conclusions about equalisation depend very much on how one defines
and operationalises the concept; policy makers need to be clear about what they
want the DAU to accomplish in respect of equalisation.

INTROD UCTION
Indonesia has embarked upon an ambitious program of fiscal decentralisation.
The effort has its genesis in two laws,
both promulgated in May 1999, one on
administrative matters and the other on
fiscal and finance issues (Republik Indonesia 1999a, 1999b).1 These two laws
have been supplemented by a large
number of implementing regulations.
Beginning in fiscal year 2001, provincial
and local governments assume major
new expenditure responsibilities. Substantial functions for provinces have
been outlined in a 1999 government
regulation (Republik Indonesia 1999c).
Local government responsibilities, regrettably, have been only rather vaguely
defined via a ‘negative list’, but are nonetheless expected to be considerable.
Kabupaten and kota (districts and municipalities) essentially become responsible


for all public services that the central and
provincial governments do not deliver,
at least in 11 important areas: public
works, health, education and culture,
agriculture, communications, industry
and trade, capital investment, environment, land, cooperatives and labour.
Regional governments have not, unfortunately, been awarded new authority over any major tax bases.2 Districts
and municipalities will instead be allowed to create their own taxes through
local by-laws, provided they satisfy a
number of ‘good tax’ criteria (Republik
Indonesia 2000d). Regional governments also now gain access to significant
amounts of natural resource revenues
(from fisheries, forestry, mining, gas and
oil) 3 and, in addition, receive a share of
personal income tax.4 Furthermore, two
new and important intergovernmental

ISSN 0007-4918 print/ISSN 1472-7234 online/01/030325-19


© 2001 Indonesia Project ANU

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

326

grants have been created: the Dana
Alokasi Umum (DAU, General Purpose
Fund) and Dana Alokasi Khusus (DAK,
Specific Purpose Fund). The latter has
not yet been made fully operational, and
the 2001 fiscal year’s allocations are relatively insignificant. These two transfers
together replace the old system of
Subsidi Daerah Otonom (SDO, Autonomous Region Subsidy) and Instruksi
Presiden (Inpres, Presidential Instruction) grants.5
The main goal of the DAU, according
to existing laws and regulations, is to
‘make even the fiscal capacities of regional governments to finance their expenditure needs’.6 It has therefore been
widely referred to as an ‘equalisation’
grant. The DAU will be a significant

source of finance for regional governments in 2001 and beyond; it will fund
approximately one-third of provincial
and between two-thirds and three-quarters of local government expenditures in
this first year. And it will clearly offer
regional governments more flexibility in
its use than existed under the old system. For the first time, regional governments will have complete authority over
how they spend the grant.
It may be helpful to recognise at the
outset that allocation procedures differ
significantly between districts/municipalities and provinces. The distribution
mechanism for districts and municipalities was developed first, and it was subsequently discovered that the same
mechanism could not be applied to
provinces. The particular reasons for
and consequences of this will become
clearer below. While some descriptive
references are made to provinces, the
empirical focus of the paper is on the
DAU allocation mechanisms for districts
and municipalities.
It is perhaps best to think of the DAU

allocation mechanism as being developed in two stages. The first consists of

Blane D. Lewis

the procedures generated by the Ministry of Finance (MOF) during the period
up to 15 November 2000, when the allocations were first approved by the
Dewan Pertimbangan Otonomi Daerah
(DPOD, Regional Autonomy Advisory
Co uncil) and (sem i-officially) announced to districts/municipalities and
provinces. 7 Soon after the original allocations were announced, certain assumptions in the draft state budget
(RAPBN, Rancangan Anggaran Pendapatan dan Belanja Negara) were altered.
These changes (relating most importantly to the price of oil and the foreign
exchange rate) resulted in an increase in
the total amount of money available to
the DAU, and this aggregate increase,
of course, necessitated a change in individual district/municipality and provincial allocations. The MOF took this
opportunity to ma ke so me other
changes to the allocation system as well.
The second stage, therefore, comprises
the adjustments made to the original allocation procedures that resulted in the

new distributions. These latter allocations were approved by the DPOD and
ratified in the form of a Presidential Decision (R epublik Indo nesia 2001 a)
signed by the then president on 20 December 2000.
THE DAU
ALLOCATION M ECHANIS M
D istrict/M unicipality Allocations
District/municipality DAU allocations
may be written:
DAU i = BFAi * + FAi + LSAi

(1)

where BFA is the balancin g factor
amount (and the asterisk denotes adjustments to it as described below), FA
is the formula amount, and LSA is the
lump sum amount. The subscript i refers to district/municipality governments.

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

The New Indonesian Equalisation Transfer


Balancing Factor Amounts. The balancing factor was, in general, conceived of
as a mechanism to ‘prevent a downturn
in the capacity of regional governments
to finance their new responsibilities’
(Republik Indonesia 2000b: article 17).
In actuality, it attempts to do two distinct things at once. First, it aims to
operationalise a ‘hold harmless’ condition: to assure each district/municipality that it receives this year (2001) at least
as much (in nominal terms) as it did in
the last fiscal year (2000) in annualised
SDO and Inpres grants.8 Second, it seeks
to account for possible new expenditure
requirements that result from the transfer of staff and functions (i.e. salaries and
other associated costs) from the central
government and its deconcentrated
agencies (kantor wilayah or kanwil and
kantor departemen or kandep) to districts/
municipalities. These new requirements
are assumed to be covered by a 30% increase in SDO and a 10% increase in
Inpres grants from the previous year.9

The basic balancing factor amount
can be written as:
BFAi = 1.3 · SDOi + 1.1 · INPRESi

(2)

where SDO and Inpres are annualised
nominal amounts from FY 2000.
The balancin g factor was subsequently adjusted for some local governments after the aggregate amount of the
DAU increased as a result of the altered
assum ptio ns in the dra ft budget.
Clearly, the allocations themselves had
to be adjusted to reflect the new bottom
line, although the particular adjustments made to the mechanism were not,
of course, required by the rise in total
DAU. Adjustments made to the original distributions include those to increase the funding for newly created
districts/municipalities, for example,
and were implemented by arbitrarily
increasing the coefficients of SDO and/


327

or Inpres in equation (2) above. 10 Details
of these adjustments are not generally
available and they are not further discussed here.
Formula Amounts. In equation form,
the formula amount (FA) for districts/
municipalities can be written as:
ö
æ
FGi
FAi = ç DAU T1 - å BFAi ÷ ·
ø å FGi
è
i

(3)

i


where DA U T1 repres ents the tota l
amount of DAU budgeted for districts/
municipalities before basic assumptions
in the draft budget were changed; BFA
is as defined above; and FG is the fiscal
gap.
The fiscal gap is defined as the difference between ‘expenditure needs’ (EN)
and ‘fiscal capacity’ (FC). That is:
FGi = ENi - FCi

(4)

The DAU formula defines expenditure needs as the product of average local government expenditure and an
expenditure needs index. Expenditures
are actual amounts from 1999/2000 local government budgets (Anggaran Pendapatan dan Belanja Daerah, APBD).
The needs index is a function of population, area, poor population, and a cost
element. Data for these variables are
from the most recent years available.
Variables, data years and data sources
are listed in the appendix.
More specifically, expenditure needs,
as defined in the pertinent government
regulation (Republik Indonesia 2000b)
can be expressed by the equation:
æ APBDEXPT ö
ENi = ç
÷
è
ø
n
·

Areai
Povi
Cost i ö
1 æ Popi
+
+
+
÷
ç
4 è PopT n Area T n PovT n 100 ø

(5)

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

328

Blane D. Lewis

where APBDEXP is local government
expenditure, Pop is the population, Area
is surface area, Pov is the number of poor
people, and Cost is a cost index (defined
more precisely below). The subscript T
indicates the total of the particular variable in question across all districts/mu nicipalities, and n is the number of local
governments.11
Fiscal capacity is defined in the government regulation as the average across all
local governments of the sum of actual
local own-source revenues, property tax,
and change in property title tax, multiplied
by an index of fiscal capacity. Government
revenues are actual amounts from the
1999/2000 regional budget. The fiscal capacity index is a function of natural
resources sector output, non-natural resources sector output, and the working age
population. Again, variables, years and
sources are detailed in the appendix.
In equation form, fiscal capacity can
be written:
æ OSRi + SPTi ö
FCi = ç
÷
è
ø
n
æ NROi
1 ç NROT
· ç
3 ç LFi
ç+
è LFT

GRDPi
GRDPT
Popi
PopT

+

NNROi GRDPi ö
NNROT GRDPT ÷
÷
÷
÷
ø

(6)
where, OSR is local own-source tax and
non-tax revenue, SPT is shared property
tax and shared change of property title
tax revenues, NRO is natural resource
sector output, NNRO is non-natural resource sector output, GRDP is gross regional development product, LF is the
working age population, and Pop is the
population as before. Subscripts are as
defined above.12
Lump Sum Amounts. The difference
between the total second stage DAU and
the sum of the adjusted balancing fac-

tor and formula amounts might be
termed the ‘residual’:
æ
ö
Residual = DAUT 2 - ç å BFAi * + å FAi ÷ (7)
è i
ø
i

where DAU T2 is the total DAU after
changes to the draft budget. The lump
sum amount received by each local government is given by:
LSAi =

Residual
n

(8)

where, again, n is the number of local
governments. Table 1 summarises aggregate district/municipality DAU transfer
amounts by balancing, formula, and
lump sum elements before and after
draft budget assum ptions were
changed.
Provincial Allocations
Total annualised SDO and Inpres grants
to provinces in FY 2000 (Rp 7.7 trillion)
exceeded the total amount of DAU available to provinces in FY 2001 (Rp 6.1 trillion). This meant that it was not possible
to im plement the same distribution
mechanism as that developed for districts/municipalities and described
above. A different set of procedures for
allocating the DAU among provinces
was therefore derived.
The adopted procedure is straightforward and begins with the recogniti o n t ha t, i n th e f i r s t r o u n d o f
estimations for districts/municipalities, the balancing factor and formula
amounts made up approximately 80%
and 20% of the total district/municipality DAU, respectively. With a view
to being consistent, it was decided to
use those percentage figures in the allocation of provincial DAU. It was
therefore resolved to distribute 80% of
total provincial DAU across provinces
as a function of FY 2000 SDO and

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

The New Indonesian Equalisation Transfer

329

TABLE 1 Summary of District/Municipality DAU Amounts Before (DAU T1)
and After (DAUT2) Draft Budget Changes

DAU
Component

DAUT1
(Rp trillion)

(%)

DAUT2
(Rp trillion)

(%)

40.4
10.1
n.a.
50.4

80.1
19.9
n.a.
100.0

44.0
10.1
0.4
54.5

80.9
18.5
0.6
100.0

Balancing factor
Formula element
Lump sum
Total

Source: State budget (APBN) and author ’s own calculations.

Inpres grants and 20% as a function
of the fiscal gap, as described above.
In equation form:
DAUi = 0.8 · DAU T 2
æ
ö
SDOi + INPRESi ÷
ç
·
ç å SDO + INPRES ÷
i

ç
è i
ø
æ
ö
FGi ÷
+0.2 · DAU T 2 · ç
ç å FG ÷

ç
è i
ø

(9)

where all variables are as defined above,
and the subscript i and DAU T2 now apply to provinces instead of districts/
municipalities.
D AU POOL OF FINANCE AND
ALLOCATION M ECHANIS M
D ES IGN
Pool of Finance
By law, the total pool of finance available to the DAU is, at a minimum, 25%
of domestic revenues net of the portions
distributed to provincial and local governments through the various revenue
sharing schemes.13 In 2001, total DAU
amounts to Rp 60.5 trillion. In the aggregate, this sum appears to be sufficient
to cover the net expenditure requirements of regional governments.

This assertion requires a little more
explanation. Strictly speaking, what is
actually required by regional governments, in the form of DAU, is a function
of their specific service responsibilities,
the legitimate costs related to carrying
out those responsibilities, and the other
fiscal resources (i.e. revenues from own
sources and other transfers) that regional
governments might reasonably be expected to have at their disposal to finance the delivery of such services.
Unfortunately, these three things are not
known with any degree of certainty. The
statement above that the DAU is, in total, sufficient to cover net expenditure
needs is based merely on a comparison
of past expenditures (both autonomous
regional government and deconcentrated agency expenditures), past revenues (own-source and shared property
tax revenues), and projected new shared
revenues and other grants (from natural resources, income tax, and the specific-purpose transfer), and not on real
needs and capacities. Table 2 summarises the pertinent data.
Also by law, provinces receive 10%
and districts/municipalities receive 90%
of the total amount of the DAU.14 In 2001,
these sharing arrangements mean that
provinces and districts/municipalities
take in Rp 6.1 trillion and Rp 54.5 tril-

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

330

Blane D. Lewis
TABLE 2 Estimated Regional Government Finances, 2001 a
(Rp trillion)

Expenditure
Routine
Development
Ex-kanwil/kandep
Total
R evenue
OSR
SPT
SNRR
PIT
DAK
Total
Revenue – expenditure
DAU
S urplus

All

Provinces

Districts/Municipalities

31.6
15.0
18.2
64.8

8.4
5.5
3.2
17.1

23.2
9.4
15.0
47.6

10.5
5.6
11.5
3.1
0.7
31.7
–33.1
60.5
27.5

6.5
2.3
2.7
1.2
n.a.
12.7
–4.4
6.1
1.6

4.1
3.4
8.8
1.9
0.9
19.0
–28.7
54.5
25.8

a

‘Routine’ and ‘Development’ are based on actual expenditures by regional governments in
1999/2000. ‘Ex-kanwil/kandep ‘is from planned routine expenditures by the deconcentrated
agencies in FY 2000. ‘OSR’ is own-source revenue and ‘SPT’ is shared revenues from property tax and transfer of title on property tax, both derived from 1999/2000. ‘SNRR’ and
‘PIT’ are estimated shared revenues from natural resources and personal income tax, respectively, both for 2001. ‘DAK’ is the 2001 budgeted amount for the specific purpose grant.
All figures have been annualised and adjusted for expected increases in salary and inflation
where appropriate.
Sources: APBN; APBD; author ’s own calculations.

lion, respectively, in DAU distributions.
Again, these amounts appear to be sufficient to cover both provincial and
district/municipality expenditure requirements, in the aggregate, and by the
standards adopted above. This statement is made with somewhat less confidence for provinces than it is for
districts/municipalities.
As table 2 shows, the DAU, together
with other sources of provincial revenue, appears to be just adequate to
cover total expected expenditures for
provinces. Given the uncertainty surrounding these estimates, it would not
be very surprising if provinces were to

experience an aggregate shortfall in financing. On the other hand, districts/
municipalities seem to have significantly more than enough in the aggregate to cover their net expenditure
requirements.15 Even the uncertainty of
the data should not throw this conclusion into doubt. In any case, these conc lu si o ns abo ut th e a deq ua c y o f
provincial and district/municipality
DAU allocations are made with the
same caveat about the nature of the
data as is made above for overall adequacy of DAU funding.
It is important to stress that the new
spending requirements of provinces and

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

The New Indonesian Equalisation Transfer

districts/municipalities include only
ro utin e-side expenditure s o f the
(form er) deconcentrated agencies
(kanwil and kandep). That is, only kanwil
and kandep staff costs and associated routine expenditures (Daftar Isian Kegiatan,
DIK) have officially been turned over to
provinces and districts/municipalities.
Regional development spending (Daftar
Isian Proyek, DIP) has not explicitly
been decentralised yet. This regional
DIP spending in FY 2000 amounted to
approxima tely Rp 2 5 trillion (in
annualised amounts). Under the assumption that responsibility for development activities formerly carried out
by deconcentrated agencies has not been
devolved, one reasonable conclusion
here is that the central government has
transferred too much revenue to local
governments, given other sources of finance available. Among other things,
this implies an added and unfortunate
pressure on an already overburdened
central government budget.16 In addition, the analysis suggests that the arrangem ents for s ha ring the DAU
between provinces and districts/municipalities may not be appropriate,
given the extent to which their respective net expenditure requirements are
covered by general transfers.
Allocation M echanism D esign
There are a number of design and operational flaws in the DAU allocation
mechanism, and some of the more significant are discussed briefly here.
Hold Harmless Provisions. Typically, a
hold harmless provision would be structured into an allocation mechanism with
a view to assuring local governments
that they will receive no less than a certain amount. In the current context, that
guaranteed amount is intended to be the
sum of last year’s SDO and Inpres
grants. But the particular hold harmless
mechanism employed here assures all

331

regions (at both provincial and district
level) that they receive that pledged
amount plus some additional allocation
(derived from the formula). This suggests the strong possibility of significant
‘overfunding’ for some regions, above
the level a typical hold harmless provision would imply.
New and Old Expenditure Requirements.
The DAU allocation mechanism treats
new and old expenditure requirements
quite differently. As explained above,
new expenditure requirements, that is,
those that derive from the decentralisa tion of new service responsibilities, are
estimated as the sum of 30% and 10%,
respectively, of last year ’s SDO and
Inpres payments for all regions. The distribution mechanism guarantees this
amount to each region, regardless of fiscal capacity. On the other hand, old expenditure requirements are accounted
for in the formula element of the mechanism; that is, they are first offset by potentially available revenues and then
that net amount relative to the sum of
all such amounts is used to allocate residual DAU. The more consistent way
to have accounted for new expenditure
needs would have been to include these
estimated requirements in the formula
term.
In addition, the accuracy of these partial SDO and Inpres distributions as an
estimate of new expenditure requirements can at least be questioned. An alternative proxy that might have been
used for estimating those needs was
deconcentrated agency (kandep) routine
(DIK) and development (DIP) expenditure on those functions to be transferred
to local governments. While it must be
admitted that region-specific information on deconcentrated development
costs to be transferred was often somewhat suspect, reliable deconcentrated
expenditure data on the routine side did
exist at the district/municipality level at

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

332

the time the allocation tools were being
designed and operationalised.
Specification of Cost Index. The main
problem with the cost index concerns its
specification, more particularly the
method by which it influences average
expenditures in the determination of
expenditure requirements. Usually a
cost index would be applied to the measure of expenditure directly (that is,
multiplicatively). So, for example, if the
cost factor in one region were 10% higher
than the base (typically set equal to 100),
expenditure requirements would be adjusted by 10% in the more expensive region, all other things remaining equal.
In the current case, the cost factor influences average expenditures after its effects have been averaged with three
other index variables. So, to continue
with the example, if a particular region’s
costs are 10% higher than the base, expenditure requirements are adjusted
upward by only 2.5%, other things being equal.
Weights of Various Factors Affecting the
Expenditure Needs Index. The overall expenditure needs index is calculated as
the unweighted average of the individual index variables. That is, each index variable is assumed to be equally
important in determining expenditure
requirements. This is a questionable notion. International experience and common sense suggest that population is
probably more important than area or
poverty, for example, in determining
overall spending needs. If this is the case,
then the population variable should
have received a larger weight than the
other index variables in determining the
overall index.
Specification of Average Revenue Potential. As can be seen in equation (6), fiscal
capacity is defined as average (ownsource plus property-related) revenues
multiplied by a capacity index. The definition posits that both own-source and

Blane D. Lewis

shared revenue potential are determined
by conditions in the particular jurisdiction and, moreover, suggests that local
governments have some control over the
extent to which collection matches potential. But, in reality, local governments
have much influence only over their
own-source revenues. Shared revenues,
including those based on the property
tax, are determined by the central government. This fact suggests that a more
appropriate formulation would be one
that posits fiscal capacity as a separate
function of potential own-source revenues, on the one hand, and shared revenues, on the other. Tha t is, fiscal
capacity would be set equal to ownsource revenue capacity plus shared revenues, where the latter are the exact
amounts as determined by central government.
Specification of Fiscal Capacity Index
Factors. Average revenue as defined
above is adjusted for region-specific capacity variables in order to come up with
an estimate of each region’s fiscal capacity. There are two criticisms that might
be made about the specification of these
fiscal capacity index variables. The first
concerns the selection of variables.
While GRDP is clearly important in determining potential local revenues, the
rationale for usin g both NRO and
NNRO as separate factors is not clear.17
The use of labour force size as a variable that influences fiscal capacity is also
somewhat unusual. Typically, one
thinks of labour as an input in the production process, that is, as a determinant of output and income. From this
perspective, the DAU fiscal capacity formulation employs both input (labour
force) and output (GRDP) measures to
determine potential local government
revenues.
The second problem concerns the actual structure of the sub-indexes themselves. On the expenditure needs side,

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

The New Indonesian Equalisation Transfer

each individual index was formed by
taking the value of the variable and dividing by the average value across all
regions. This has some logic to it, of
course. On the fiscal capacity side, however, the indexes are formed through the
use of ‘location quotients’. Here, what
becomes important is a region’s share of
NRO (and NNRO) in total GRDP as
compared to national NRO (and NNRO)
as a share of total GRDP across all regions. This is at least inconsistent with
methods used for defining the expenditure needs index.
Natural Resource Revenues in Fiscal
Capacity. Perhaps the greatest problem
in the estimation of fiscal capacity concerns the issue of natural resource revenues. Put sim ply, natural resource
revenues are basically ignored in the determination of capacity. In essence, DAU
distributions are awarded on top of
these shared resources. This is a significant windfall gain for those local governments that receive large amounts of
shared revenues from natural resources.
And, of course, it diminishes the DAU
allocations that all other regions would
otherwise have received. Clearly such
shared revenues should have been
added to own-source revenue potential
along with all shared revenues in the determination of fiscal capacity.18 Indeed,
reasonable estimates of these shared revenues existed at the time the DAU allocation methods were being discussed.
D AU EQUALISATION
PERFORM ANCE
The above critique of the allocation
mechanism has suggested the existence
of design and operational flaws. Some
of these shortcomings have possible
negative implications for distributional
equity. For example, to the extent that
the (augmented) hold harmless condition assures some local governments of
levels of funding that exceed their needs,

333

then other places might well receive less
than they require. As another example,
recall the argument that physical area
has been unduly stressed in determining expenditure requirements, and natural resource transfers inadequately
accounted for in measuring fiscal capacity. As it turns out, many regions with
large areas also tend to be rich in natural resource revenues;19 their local governm ents have therefore had their
expenditure needs overestimated and
their fiscal capacities underestimated.
Th e negative impact of this on the
equity of distributions should be clear.
Below, two separate standards are used
to examine the impact of such potential
problems on achievement of the goal of
equalisation. 20
Since the equalisation goal of the
DAU transfer scheme, as stated in the
law and relevant government regulation, is to ‘make even the fiscal capacities of regional governments to finance
their expenditure needs’, it is clear that
a proper test of the mechanism’s equalisation effects would require, in the first
instance, good measures of expenditure
needs and fiscal capacities. But the above
examination of the DAU formula suggests that current methods of estimating needs and capacities are somewhat
flawed.
The analysis proceeds therefore by
first providing an alternative measure of
fiscal capacity. The argument is that this
revised method is a reasonably sensible
one, in general, and clearly better than
that currently employed. Deriving a
sound measure of expenditure needs is
more problematic, given our lack of
knowledge about exact local level expenditure assignments and legitimate
costs of service delivery across various
sectors. Complications in the derivation
of practicable estimates for expenditure
needs are ignored for the time being. The
question is revisited under the second

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

334

Blane D. Lewis

method of examining equalisation carried out below.
Revised fiscal capacity (FC*) is defined by the following.
ˆ + SPT + SNRR
FCi * = OSR
i
i
i

(10)

ˆ is a proxy for potential ownwhere OSR
source revenues and is calculated as the
predicted value that obtains from regressing actual OSR against GRDP21 and
a dummy variable for urban status
(equal to 0 for kabupaten and 1 for kota);22
SPT is shared property-based taxes as
before; and SNRR is an estimate of actual natural resource revenues to be
transferred to local governments in
2001.23
The first approach to examining the
equalisation question focuses on the
variation in potential revenues. If DAU
transfers are to ‘equalise’ it might be
reasonable to expect, at a minimum, that
the variation in revenue capacity among
local governments would be less after
transfers are made than before. Table 3
provides so me info rm ation on the
regional variation in revenues from
different sources using three standard
measures.
The table shows the coefficient of variation (i.e. the standard deviation divided
by the average) of per capita potential

revenues, the standard deviation of the
natural logarithm of per capita potential revenues, and the gini coefficient of
per capita potential revenues across all
local governments in the sample. The
base case relates to potential own-source
revenues as estimated above. To these
ˆ ),
potential own-source revenues ( OSR
are added, in succession, shared property-related tax revenues (SPT), shared
natural resource revenues (SNRR), DAU
balancing factor am ounts, D AU
formula-derived amounts, and finally
DAU lump sum amounts. (Recall that
the latter three together comprise total
DAU.) Each measure in the table shows
the variation-cum-equality (smaller
variations in the distribution are indicative of greater equality) of the sum of
per capita potential revenues (up to and
including that revenue source found on
the particular line of the table) across the
sample of all local governments.
The table shows that, regardless of the
particular measure used, the variation
in potential revenues is lower after total
DAU allocations are incorporated than
before. (That is, the coefficient of variation, the standard deviation of logarithms, and the gini coefficient decline
from 2.229, 0.966 and 0.602 before DAU
transfers are added, to 0.905, 0.575 and
0.308, respectively, after these transfers

TABLE 3 Variation in Potential Revenues across Local Governments

Potential
Revenues

ˆ
OSR
+ SPT
+ SNRR
+ DAU balancing amounts
+ DAU formula amounts
+ DAU lump sums

Coefficient of
Variation
1.734
1.178
2.229
0.991
0.905
0.905

Source: Author ’s own calculations.

Standard Deviation of
Logarithms
0.723
0.673
0.966
0.588
0.574
0.575

Gini
Coefficient
0.408
0.364
0.602
0.320
0.308
0.308

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

The New Indonesian Equalisation Transfer

are incorporated.) In other words, the
distribution of per capita revenues is
more equal after the transfers than before. The general conclusion that can be
drawn from this analysis is that DAU
transfers appear to be equalising, at least
under the admittedly somewhat weak
standard considered here.
Furthermore, the table suggests that
the formula component of the transfers
is somewhat more equalising than the
balancing amount. (That, is the relevant
measures decline after formula amounts
are added to balancing factor distributions.) The table provides other interesting results. It shows, for example, that
the transfer of property-related taxes
tends to equalise the distribution of
own-source revenues. And it demonstrates the rather extreme unequalising
effect of the natural resource transfers.
There are at least two possible criticisms of this method of assessing the
equalisation effects of the DAU allocation mechanism. The first is that it has
not explicitly incorporated the notion of
expenditure needs. One way to get
around this, of course, would be to examine the variation in the distribution
of the ratio of potential revenues to expenditure needs across all local governments before and after transfers. The
above mentioned difficulties associated
with deriving a single measure of expenditure needs would seem to prohibit
such an approach, at least for the time
being.
The second and related concern is that
the method ignores the important simultaneous relationship among expenditure
needs, fiscal capacities, and transfers.
That is, in examining the equalisation
performance of the DAU alloc ation
scheme, it would be useful to know how
transfers vary in amount with respect to
variations in expenditure needs and fiscal capacities. More particularly, from an
equalisation point of view, it might be

335

expected that as expenditure needs rise,
transfers should increase, fiscal capacities remaining the same. And, as fiscal
capacities increase, it might be argued
that transfers should become smaller,
holding expenditure needs constant.
This is perhaps a slightly stricter standard of equalisation than the one employed just above.
The difficulty, again, concerns estimating expenditure needs. However,
while it may not be possible to derive a
plausible single measure of expenditure
needs, some of the factors that are important in determining needs are in fact
well known in general terms. And some
of these determinants are, of course,
those that were used in the current DAU
methodology for estimating needs. That
is, population, area, poverty, and relative cost factors, in some weighted combination, all clearly help to determine
real requirements, and therefore transfers. The approach used directly below
assumes that these four variables, along
with ‘urbanisation’, help determine expenditure requirements and influence
allocations. But rather than specifying
exactly how they do this a priori, as the
current DAU formula mechanism does,
the method employed here will ‘let the
data decide’.24
It is posited that transfers are a function of fiscal capacity—as defined in
equation (10) above—and expenditure
needs, and a simple linear regression
technique is used to operationalise the
relationship. M ore specifically, per
capita transfers are regressed against per
capita fiscal capacity (FISKPC); ‘cost-index adjusted’25 population (POP•CST),
area (AREA•CST), and poverty rate
(POV•CST);26 and a dummy variable for
urbanisation (KOTA, set equal to 0 for
kabupaten and 1 for kota).27 The multiplicative specification of the basic fiscal
needs variables is suggested by the
standard employment of a cost index as

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

336

Blane D. Lewis

argued in the previous section of the
paper. All variables (except the dummy,
of course) are entered into the equation
in logarithmic form.
Table 4 provides the results of the ordinary least squares (OLS) regression. 28
Panels A, B and C of the table show the
standard regression output for three different dependent variables: total per
capita DAU transfers (DAUPC), per

ca pita ba la ncin g factor a mo unts
(BALPC ), and per ca pita for mula
amounts (FORMPC), respectively. In
each case, the same independent variables, as described above, are used.
The results are somewhat mixed. In
all cases, as (cost-adjusted) area and poverty increase, per capita transfers (in total and by component part) increase as
well, all other things being constant;

TABLE 4 Equalisation Performance of the DAU Model: Regression Results

Unstandardised
Coefficients
B

Panel A: Dependent Variable DAUPC
(Constant)
7.168
FISKPC
0.130
POP•CST
–0.448
AREA•CST
0.075
POV•CST
0.119
KOTA
0.002

Standardised
Coefficients
Beta

t

0.256
–0.680
0.289
0.165
0.001

35.955
6.042
–18.722
5.497
4.815
0.023

*
*
*
*
*

0.273
–0.684
0.118
0.130
–0.108

33.589
5.896
–17.241
2.051
3.470
–1.825

*
*
*
*
*

0.064
–0.458
0.815
0.283
0.388

13.032
1.379
–11.570
14.236
7.599
6.594

*
Ê
*
*
*
*

Adjusted R Squared: 0.752
Panel B : Dependent Variable BALPC
(Constant)
7.401
FISKPC
0.140
POP•CST
–0.456
AREA•CST
0.031
POV•CST
0.095
KOTA
–0.138
Adjusted R Squared: 0.703
Panel C: Dependent Variable FO RM PC
(Constant)
3.330
FISKPC
0.038
POP•CST
–0.355
AREA•CST
0.250
POV•CST
0.241
KOTA
0.579
Adjusted R Squared: 0.705

* Denotes that the estimated unstandardised coefficient is significantly different from zero
at the 0.05 level.
Source: Author ’s own calculations.

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

The New Indonesian Equalisation Transfer

such a result is generally supportive of
equalisation goals. Interestingly, the
dummy variable for urbanisation is significant only for formula-driven allocations; here, municipalities receive larger
per capita transfers, all other things remaining the same. Municipal status is
not important for balancing factor distributions or for DAU transfers overall.
It is typically argued that cities have
greater per capita service delivery responsibilities than rural areas; if this is
true, then the results here suggest that
urban regions have not, in general, been
alloca ted tra nsfers in s ufficient
amounts—their expenditure needs are
probably greater but they do not receive
larger amounts of DAU. Note that there
are no a priori expectations about the significance or sign of the coefficient of the
(cost-adjusted) population variable. As
it turns out, population is the most important variable overall in explaining
variation in per capita transfers (as
judged by the values of the standardised
beta coefficients), and the results here are
indicative of an assumption of economies of scale in the provision of services
financed by transfers. That is, as population increases, per capita transfers decrease. Of course, much more analysis
would need to be done to confirm the
actual existence of such economies of
scale.
Attention is now directed to the influence of the fiscal capacity variable in
determining transfers. Panel A of table
4 shows that, as per capita fiscal capacity increases, per capita transfers increase as well, all other things (that is,
expenditure needs) remaining the same.
This suggests that current DAU allocations, in total, are unequalising with regard to fiscal capacity. The results in
panel B demonstrate that the balancing
factor appears to be unequalising by this
standard as well. And, as panel C shows,
the formula-driven amounts appear to

337

be neutral with regard to fiscal capacity
equalisation. That is, whereas the relevant coefficient is positive (suggesting
unequalising effects), it is not statistically
significant.
S UMM ARY AND CONCLUS IONS
The paper demonstrates the significant
role in distributions played by the ‘balancing factor ’ relative to the ‘formula’
that was so much discussed in the runup to the 2001 fiscal year. The balancing
factor allocates at least 80% of the total
grant for regional governments. This
may surprise some who have assumed
that the formula as it appears in the government regulation drives the allocation
results. In addition, the paper highlights
key differences in distribution procedures for districts/municipalities and
provinces.
A rather detailed analysis of the new
transfer scheme concludes that the sums
available to the DAU appear to be more
than sufficient to meet the net expenditure requirements of regional governments, at least in the aggregate, and
given some basic caveats. The aggregate
overfunding of regional governments
has obvious negative implications for
the central government budget.
This conclusion on the adequacy of
funds can be made with a good deal of
confidence for districts/municipalities,
but it is somewhat more tenuous for
provinces. The surplus of total district/
municipality revenues over expenditure
requirements is very large indeed, while
that for provinces is much smaller. Thus
the current arrangements for sharing the
DAU between districts/municipalities
(90%) and provinces (10%) may be inappropriate.
The examination of the allocation
mechanism itself reveals a number of
design and operational flaws, in both its
basic structure and the methods it uses
to measure expenditure needs and fis-

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

338

cal capacity. It is argued that some of
these shortcomings potentially have
rather negative consequences for the
equitable distribution of transfers.
The paper shows that conclusions
about the equalising effects of the current transfer scheme depend very much
on how one defines and operationalises
the concept of ‘equalisation’. It considers two basic standards of equalisation.
The first focuses on variation among local governments in potential per capita
revenues before and after the allocation
of transfers. This variation is shown to
be smaller after the distribution of transfers than before the allocations are made,
suggesting that the DAU mechanism is
equalising.
The second standard concentrates on
the effect on per capita transfers of
changes in expenditure needs and per
capita fiscal capacities. The analysis
shows that, for the most part, as expenditure needs (as measured by a set
of determinants) increase, per capita
transfers also increase, as would be expected. Unfortunately, as per capita fiscal capacities increase, per capita grants

Blane D. Lewis

rise too, everything else remaining the
same. Thus, by this latter and perhaps
stricter standard, DAU allocations appear to be unequalising, at least with
regard to fiscal capacity.
Government decision makers have
committed themselves to improving the
DAU distribution mechanism in the
very near term, with revisions in place
perhaps by as early as the 2002 fiscal
year. The analysis here suggests that
policy makers and modellers need to be
more clear about what it is they are trying to accomplish through the DAU, especially in respect of its equalising
effects. In addition, decision makers and
their staff need to examine the possible
consequences of alternative allocation
specifications more rigorously than was
the case this past year. Without both an
explicit recognition of policy preferences
in respect of equalisation and other goals
and a thorough investigation of how
well various alternatives under consideration will deliver preferred outcomes,
the allocation system will probably contin ue to perform in a substandard
manner.

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

The New Indonesian Equalisation Transfer

339

NOTES
*

1
2

3

4

The author currently serves as senior adviser to the Ministry of Finance under the
PERFORM project, which is financed by
the United States Agency for International Development (USAID). The views
expressed here are those of the author
and should not be attributed either to the
Ministry of Finance or to USAID. An earlier version of the paper was presented
at the International Regional Science
Conference in Jakarta on 20–21 March
2001 (Lewis 2001). The author would like
to thank Bert Hofman, Jorge Martinez,
Stefano Piperno and Bob Searle for useful discussions on the DAU, and two
anonymous referees for helpful comments on an earlier draft of this paper.
See Booth (1999) for a description of the
main provisions of the two laws.
The potential benefits of decentralising
appropriate taxing and charging authority to lower level governments are well
known: they include mobilisation of increased local revenues and enhancement
of local government accountability to
constituents through the establishment of
strong linkages between service provision and payments for such. See Alm,
Aten and Bahl (2001) for a standard discussion of the latter point in the Indonesian context.
Although all regional governments gain
to a certain extent from natural resource
revenue sharing, the benefits are, of
course, very unevenly distributed, with
the bulk of transfers going to provincial
and local governments in Aceh, Riau,
East Kalimantan and Papua (Irian Jaya).
Beginning in 2001, 20% of personal income tax is to be allocated to regional
governments, with provinces and districts/municipalities receiving 40% and
60% respectively of total shared amounts.
Distributions to provinces are to be based
on the residence of registered taxpayers.
Allocations to districts/municipalities
are determined by the relevant province,
and may or may not be shared on a derivation basis (Republik Indonesia 2000c).
Clearly these shared taxes are also distributed rather unevenly among regional
governments.

5

There are several good examinations of
the SDO and Inpres system. For a standard reference and partial review of other
analyses see Shah, Qureshi and Binder
(1994). On Inpres grants see also Silver et
al. (2001), in this issue.
6 See Republik Indonesia (2000b), article
15, for the precise language.
7 The DPOD provides advice to the president on the formation, abolition and annexation of regional governments; the
financial balance between the centre and
regional governments; and regional government capacity to carry out designated
functions. The DPOD comprises, most
importantly, the Minister of Home Affairs
and Regional Autonom y (chair), the
Minister of Finance (vice chair), and representatives from associations of provincial, district and municipal governments,
as well as representatives from provincial, district and municipal parliaments.
The DPOD has two Secretariats, one for
Regional Autonomy and one for Financial Balance. The latter approves the DAU
distributions and forwards them in the
form of a recom me ndation to the
president.
8 FY 2000 was a nine-month fiscal year, a
bridge between the previous system of
April–March financial years and the new
system in which the calendar and financial years coincide.
9 These estimates are based on similar increases in aggregate SDO and Inpres allocations between FY 1999/2000 and FY
2000. The increases were simply assumed
to have resulted from the transfer to local governments of staff and associated
costs from the six national level departments and agencies that were liquidated
during that fiscal year. The liquidated
institutions include the Departments of
Transmigration, Tourism and Culture,
Cooperatives, Social Affairs, Information,
and Public Works.
10 New districts/municipalities do not receive a formula amount as described immediately below, because the required
data are unavailable. To compensate for
this, the coefficient on the SDO factor in
equation (2) was increased to 2.0

nloaded by [Universitas Maritim Raja Ali Haji], [UNIVERSITAS MARITIM RAJA ALI HAJI TANJUNGPINANG, KEPULAUAN RIAU] at 22:00 19 January 20

340
11 The index variables are suggested by
Law 25/1999. It is assumed that the total
number of people, the size of the geographic area, and the number of poor
people are all positively related to expenditure needs. The cost index is intende d to m easure unavoid able
differences in costs faced by local governments, and these costs are assumed
to be positively related to expenditure
requirements. Internationally, such