REGIONAL DYNAMICS AND DECENTRALIZATION

1

• REGIONAL DYNAMICS AND DECENTRALIZATION
• Subnational disparities and dynamics increasingly
  important.
• Policy most developing countries decentralizing
  in some form, whether as part of plan to market
  in wake of deep crisis to forestall secessionist
  pressures.
• Analytics fusion of economics and geography,
  taking space seriously.
• The regions best connected to the global economy
  generally grow faster.
• By-passed regions cant be ignored, for
  geo-strategic and developmental reasons.
• Paradoxically, as countries become more open to
  the global economy, they sometimes become less
  integrated domestically.
• As national-level policy instruments converge
  (eg, macro, trade), local level governance and
  institutions increasingly important determinant
  of national progress.
• Indonesia and Philippines well suited two
  largest archipelagic nations both embarked on
  major decentralization program in wake of deep
  crises.

2

• New economic geography has come of age (Neary,
  2001).
• ... one of the best ways to understand how the
  international economy works is to start by
  looking at what happens inside nations. If we
  want to understand differences in national growth
  rates, a good place to start is by examining
  differences in regional growth if we want to
  understand international specialization, a good
  place to start is with local specialization.
  (Krugman, 1991, p. 3)
• Regional development policies have failed in all
  countries federal and unitary alike to reduce
  regional inequalities. (Shankar and Shah, 2003,
  p. 1437)
• Decentralization is neither a plague or a
  panacea (Bird and Villancourt, 1998, p.1).
• What is happening now to the Philippines is one
  of the best arguments why countries which face
  chronic fiscal deficits should go slow in the
  decentralization process. (Diokno, 2003, p. 11)

3

• ISSUES
• Regional economics now at the forefront of
  development economics.
• Why? To do with analytics policy transitions
  from crises or command economies uneven spatial
  impacts of rapid global integration.
• Some key questions/motivations in thinking about
  decentralization in the context of
  liberalization/globalization
• Why do some regions grow faster than others are
  the factors similar to those explaining
  inter-country differences?
• What determines trends in sub-national
  inequality why does regional inequality vary so
  greatly, within countries and over time is it a
  cause for concern what sorts of policies for
  by-passed regions?
• Why is the spatial location of economic activity
  so uneven in many countries?
• What is regional policy what are the key
  issues and lessons in decentralization reforms
  why are some programs more successful than
  others?
• As national boundaries become less important,
  will cross-border natural economic zones become
  more important, perhaps in an EU-type world?
• Is regional (sub-national) competition likely to
  improve local-level (and hence national)
  governance quality?

4

• The local-level impacts of these changes are
  uneven, especially combined with weakening CG
  commitment to fiscal equalization.
• The capacity to connect to the global economy is
  generally the key.
• If accompanied by declining infrastructure
  investment (eg, as in Indonesia and the
  Philippines), may result in a series of
  internationally connected enclaves. Particularly
  where decentralization transfers tax authority
  and results in over-grazing.
• Also, mixed evidence on impact of death of
  distance. Location still matters, illustrated by
  importance of clusters.
• Can the international growth literature explain
  inter-regional differences?
• With unrestricted factor mobility and homogeneous
  institutions nation-wide, presumably not.
  Conversely, if openness refers to connected
  to the international economy, and if institutions
  vary, there are lessons.
• And convergence framework is applicable. Growth
  theory predicts that convergence more likely
  within than between countries.
• As with the inter-country case, two concepts ß
  and ? convergence.
• That is, whether poorer regions within a country
  are catching up to richer ones, and an overall
  measure of inter-regional inequality.

5

• The standard growth model implies that per capita
  income growth in the long run will be
• ln ytln y e-?t ln y0 e-?t ln y e-?t
  (ln y0 ln y)
• yt is the current income per capita,
• y0 is the initial income per capita and
• y is the steady state income per capita.
• This trajectory is estimated in the reduced form
  as
• ln (yt / y0)/t a (e? 1) ln y0 ut.
• A negative value of the coefficient of (e? 1)
  and hence ß means that higher
• growth is experienced by an initially lower
  income economy.
• An alternative concept is ? convergence,
  employing a measure of standard
• deviation eg coefficient of variation, variance
  of log of initial income.
• Mixed evidence on convergence. Williamson (1965)
  hypothesized a
• Kuznets-type relationship. Generally occurring in
  OECD countries (Sala-I-
• Martin, 1996).
• China (Song (2007), etc absolute ß convergence
  1978-92, then divergence.
• India (Ahluwalia (2002), etc slowly rising
  inequality since independence.
• Latin America (Serra et al (2006), etc high
  inter-regional inequality, at most
• very slow convergence.

6

• Explaining differences in inter-regional
  inequality
• 1. Whether there exists a national commitment to
  inter-regional equity.
• Especially fiscal equalization measures. Seems to
  be declining.
• Political systems dont seem to matter, eg,
  whether democratic, federal, etc.
• 2. Trade regime More open economies more
  dispersed location?
• Perhaps, though regional inequality sometimes
  increases as economies
• open up. Sometimes owing to partial, selective
  liberalization (eg, China).
• Location of infra investments critical.
• Why do some regions grow faster (again)?
• 1. Location and openness though difficult to
  model.
• 2. Clustering and scale increasing returns,
  reputation effects, etc.
• 3. Governance and institutions though also
  difficult to measure.
• 4. Natural resource endowments though some
  evidence of resource
• curses.

7

• COMPARISONS
• China A key regional issue is significant
  increase in inequality since around 1990. No sign
  of convergence. Drivers
• (a) Large differences in international
  orientation and connections.
• (b) Restrictions on labour mobility, so factor
  markets could not adjust.
• (c) Centres retreat from inequality.
• Note recent change in policy direction.
• See comparative patterns of inter-regional
  inequality, China, Indonesia, Philippines.
• Though cant directly compare the numbers across
  countries.
• The ß coefficient of GDP per capita during
  1975-2002 is higher than the that of the central
  planning era in China, 1952-1965 (0.6) but less
  than its market
• reform period 1978-1993 of 1.7 (Jian, Warner and
  Sachs 1996).
• Vietnams market reform process resulted in a
  very low convergence rate at 0.3
• during the short period 1995-2000 (Klump and
  Nguyen 2004).
• Mexico experienced 0.2 rate of convergence
  during 1970-2003, while Brazil
• was slightly higher, 0.6 (Serra et. al. 2005).
• Trade liberalization in Brazil had a divergent
  impact on the regional development,
• in contrast to Indonesias reforms of 1981-86.
• The highest regional convergence reported has
  been the Philippines during 1988-

8
Figure 2 Philippines in East Asia
Provinces in China and Indonesia while regions
in the Philippines
9
(No Transcript)
10

• (3) Indonesia
• Worlds largest archipelagic state (see map).
  Enormous regional diversity.
• Unitary state, but big bang decentralization
  commencing in 2001, in wake
• of deep crisis still work in progress.
• Rapid fragmentation of boundaries since 1999
  complicates analysis.
• Provinces dont correspond to economic zones.
• Data here refer to pre-1999 boundaries with 26
  provinces.
• Owing to presence of natural resource enclaves,
  results are sensitive to
• the selection of economic activity indicators.
  Non-mining GRP the preferred
• indicator.
• Java dominates, with 60-66 of GDP vast area of
  Eastern Indonesia lt3.
• No major changes in shares of island groupings,
  except for Java increase,
• explained entirely by Jakarta. See Table 1. For
  GRP series, larger variations
• owing to resource cycle. Effects of conflict
  evident (eg, Maluku, Aceh).
• Gap between richest and poorest very large. See
  Table 2. Around 151 for
• JakartaMaluku.

11
(No Transcript)
12
(No Transcript)
13

• (2.2) Provincial economic rankings
• Table 2 3 measures, 3 points, normalized to
  national average.
• Large gaps regardless of series 15.9, 14.7,
  11.31
• See how enclaves affect GRPs.
• Focus mainly on non-mining GRP.
• Groups of provinces and characteristics
• Consistently wealthy - Jakarta, East
  Kalimantan, Riau.
• Consistently non-poor - N W Sumatra, C
  Kalimantan, E W Java,
• Bali.
• Very poor - E W Nusa Tenggara, Maluku, SE
  Sulawesi.
• Slipping behind - most of southern Sumatra, W
  S Kalimantan
• Papua, C Java/Yogya in some respects.
• Key points
• a) rise of Java, once associated with Malthusian
  scenarios, and
• seen as Indonesias major devt challenge
• b) Mixed record of resource-rich provinces
  greatly diverse circumstances

14
(No Transcript)
15

• Regional inequality
• 4-quadrant story relating 1975 pci and growth
  1975-2004, Figure 1.
• For non-mining GRP. Observations mainly located
  in top left quadrant,
• suggesting poorer provinces growing faster. No
  case of provinces falling
• far behind. Alternative welfare series show
  somewhat different results, but
• main conclusion holds.
• ß convergence Vidyattama (2008) estimates
  coefficient at about 1.5
• since 1975 for GRP series. But this reflects
  declining importance of oil gas
• over this period. But for non-mining series, ß
  convergence falls to 0.4 and
• is insignificant.
• For ? convergence, the CVs similarly depend on
  choice of variables. See
• Figure 2. For the non-mining and expenditure
  series, fairly stable. The GRP
• series declines as expected.
• For further insights, can also disaggregate by
  sector, and between/within
• region. On latter, see Akita and Alisjabhana
  (2002) largest source of
• inequality is within province.

16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19

• (4) Regional Social Progress and Demographic
  Change
• Demographic dynamics - see Table 9.
• Drivers historically highly uneven spatial
  patterns of settlement
• migration patterns in turn responding to uneven
  patterns of employment
• expansion, education opportunity official
  migration policy uneven demographic
• transition across regions.
• Percent of pop born outside province a good proxy
  for migration (Table 7).
• Major recipients richer provinces (esp Jakarta)
  frontier regions (most of
• Kalimantan, Papua, parts of Sulawesi) much of
  southern Sumatra (traditional
• networks, proximity to Jakarta, but declining)
  Yogya (education).
• Decomposition of pop growth to follow.
• Correlation between economic and social
  indicators levels, rates of change?
• See Table 10, Figure 4.
• Generally closely correlated.
• Note special cases of below average income/above
  average social North

20
(No Transcript)
21

• Huge diversity, especially for economic
  indicators less so for social.
• But remarkably even growth no major cases of
  provinces falling
• consistently behind. Eg, the East is much poorer,
  but its growth has
• been fairly similar to the West.
• Indonesia quite unusual in this respect.
• Inferences about regional dynamics
• Fast growers both small and large provinces,
  remote and central location,
• initially high and low pci.
• General observations
• 1. Globally connected regions grow faster Bali
  (1), Jakarta, Riau (since
• Singapore connection developed).
• 2. Clustering and increasing returns to scale
  West Javas industrialization.
• 3. Institutions and governance mixed evidence,
  decentralization too recent.
• 4. Resource rich regions diverse outcomes,
  depends on local management.
• 5. Conflict is obviously harmful. Eg, Maluku,
  Aceh.
• 6. Too early to judge impact of decentralization.

22

• (4) The Philippines
• Similarities with Indonesia archipelago (see
  map), great diversity,
• decentralization after a deep crisis, unrest in
  the frontiers.
• Also differences absence of extreme demographic
  concentration, less
• pronounced resource enclaves, slower growth.
• Major decentralization as part of 1987
  Constitution, but still a unitary state.
• Analysis conducted at level of regions, 17 in
  total, although these mainly
• administrative units.
• Manila dominates with two surrounding regions,
  about 55 of GDP.
• Also by far the richest 2.75 x national average,
  double 2, 12 x the poorest.
• Only two other regions above the national
  average both unusual.
• A third group above the ex-Manila national
  average 6 in total, diverse.
• A further six very poor, spread throughout the
  country.
• ARMM extremely poor, less than 1/4 national
  average pci very low HDI.
• See summary stats, Tables 3 4.

23
(No Transcript)
24
Current local government structure
17 regions
79 provinces 32 independent cities
1,501 municipalities 85 component cities
41,980 barangays
25
Table 3 Economic performance and structure, 2003
Note Average per capita GDP for the
Philippines in 2003 prices is Php 52,470. Source
National Income Accounts
26
Table 4 Social indicators, 2003
Source Estimates are based on the FIES 2003.
27

• Factors shaping regional dynamics
• 1. Trade liberalization removed the centralizing
  bias associated with IS
• regime, but introduced a new dynamic, depending
  on location of export-
• oriented infra and facilities led to
  reconcentration around greater Manila.
• 2. Dire consequences of prolonged conflict
  evident in ARMM.
• 3. Difficult to detect any systematic
  relationship, let alone direction of
• causality, between local level governance and
  growth.
• 4. CG grants have little overall impact on
  regional dynamics and equity. In
• any case, sub-national tiers of govt only about
  1/4 of public sector
• expenditure.
• 5. Migration, both internal and international, a
  key instrument of regional
• adjustment. Although individually beneficial, it
  probably widens inter-
• regional inequality, even allowing for
  remittances. International migrants
• disproportionately drawn from better off regions.
  Internal migration

28

• Regional inequality
• See 4-quadrant story, Figure 2.
• No clear trend, also in CVs (Figure 3).
• Heavy concentration in top left quadrant,
  suggesting catch-
• up. But results sensitive to outliers, and
  whether for regions or provinces.
• Compare various estimates of ß convergence, eg
  Balisacan (2007) and
• Manasan and Chatterjee (2003).
• Balisacan (2007) finds large dispersion around
  fitted line for provinces.
• Infra, human capital and supportive agricultural
  policies major drivers
• of regional growth.

29
(No Transcript)
30

• (5) Conclusions
• Large inter-regional differences in pci, of 101
  and more.
• No clear trend in inter-regional inequality in
  either country. Therefore differ
• from several other developing countries. Though
  results sensitive to
• measures.
• Both capital cities dominant, 3-4 x national pci.
  Jakarta rising. Capital city
• bias v/s natural economic advantages?
• The internationally connected regions grow more
  rapidly.
• No clear natural resource story.
• Lagging and by-passed regions pose particular
  development challenges.
• Very mixed record of past programs. Note also
  that inequality within regions
• typically greater than between them.

31

• Philippines lessons from an early decentralizer
• Assignment of functions broadly consistent with
  public finance theory, ie, that power should be
  with jurisdictions best able to internalize the
  costs and benefits of service provision. Except
  for special case of education.
• LGU expenditure as of GDP approx doubled since
  LGC, but still small 1.6 1985-91, to 3.3
  1992-2003. LGU expenditure share of total public
  sector below the (unweighted) average of the
  East Asian six 26 v/s 33.
• Clarity of expenditure assignments clear in
  principle, but in practice subverted by
  loopholes, unfunded mandates passed on from the
  CG.

32

• So system performs less well on autonomy
  criterion CG retains major taxes tightly
  controls rates which LGUs may set.
• Hence revenue and expenditure misalignments
• i) Growing imbalance between revenues and
  expenditures of LGUs resulting in increasing
  dependence on CG transfers from 38 of LGU
  revenue (net of borrowings) 1985-91 to 65
  1992-2003.
• ii) Very slow progress in downsizing CG agencies
• iii) Perverse incentives among LGU tiers
  pressure to become cities.
• CG grants in two forms formula-driven (IRA), and
  ad hoc categorical grants. But even IRA highly
  unpredictable, especially during (frequent)
  fiscal near-crises. Also doesnt meet
  horizontal equity objectives per capita grants
  positively related to pcys.
• Equally, LGUs have poor record on revenue
  raising. See above. Hence substitutive in
  nature. Local elites unwilling to tax themselves?
  Especially, major potential but untapped revenue
  from property tax. Also, poorly maintained and
  audited financial records. Little attempt to
  clean up proliferation of local nuisance taxes.
• Note also
• i) Considerable variation in performance among
  LGUs.

33

• ii) Still major inter-LGU coordination issues,
  especially on inter-jurisdictional issues (eg,
  infrastructure, environment).
• iii) Is there a link between national fiscal
  balance issues and IRA commitments? Might be a
  case for linking increased IRA allocations to
  national revenue effort, so LGUs add to pressure
  for fiscal reform.
• (j) Local Institutions and Governance Quality
• Have reforms delivered better quality governance
  (as literature generally hypothesizes if bring
  govt closer to the stakeholders), and is there
  a reform dividend for better governed regions?
• Mixed evidence
• No simple matter to define governance quality.
  Many attempts. Evidence on service quality (eg,
  health, roads) very mixed.
• No obvious correlates of well-governed regions
  Makati v/s Naga! Mixed evidence on
  neighbourhood demonstration effects.
• Are governance standards endogenous to pcy and
  growth?
• LGUs anyway have limited capacity to influence
  local development.
• Local politics remains highly personalistic, so
  difficult to measure quality of institutions
  independently.
• CG has not supplied the requisite framework,
  inputs, growth.

34

• Voice and exit
• Local democracy seems to have been more effective
  at rewarding good performance than at
  disciplining poor leadership.
• Exit - ie, migration - dominated by move to
  higher income regions.
• 5. SUMMARY POLICY IMPLICATIONS
• Key conclusions arguably (ii), (v), (ix).
• (a) Regional Dynamics
• i) The global economy increasingly shapes the
  economic geography of a now largely open
  Philippine economy trade remittances in
  particular. More internationally-oriented
  enclaves?
• ii) Investments in key public goods - especially
  infrastructure and education - are key drivers of
  a regions capacity to connect to the global
  economy, and to grow faster. Infrastructure
  decisions perhaps the most important (if
  unintended) tool of regional policy

35

• iii) Slower changes in regional socio-economic
  rankings (eg, cf China) reflect slow overall
  growth rates. Though more fluidity in rankings if
  measured at finer levels of disaggregation.
• (b) Decentralization an Evaluation
• iv) Philippine decentralization neither a notable
  success nor a disappointing failure.
• v) Decentralization program quite well
  formulated, sound principles. But implementation
  record very mixed division of responsibilities
  unclear widespread CG interference recurring
  fiscal near-crises LGU buck-passing on revenue
  raising.
• vi) Mixed record on vertical and horizontal
  balances. Growing vertical imbalances horizontal
  equity but small LGU shares (3.3).
• vii) Coordination between and among tiers of govt
  still inadequate. Also mismatches among LGUs
  (cities v/s provinces) some staff transfer very
  slow compounded by problems in national and
  local bureaucracies.

36

• viii) Difficult to measure LG quality, but it is
  clearly highly variable no obvious and
  systematic correlates.
• ix) Has decentralization ameliorated regional
  discontent with the centre? Not obvious,
  especially in case of Mindanao.
• x) Towards a national and regional policy reform
  agenda. Some common elements but no nation-wide,
  uniform regional policy program.