+ All Categories
Home > Documents > The Impact of Remi ance ASSESSMENTS OF THE RUSUNAWA...

The Impact of Remi ance ASSESSMENTS OF THE RUSUNAWA...

Date post: 25-Mar-2019
Category:
Upload: vannhu
View: 212 times
Download: 0 times
Share this document with a friend
11
52 | EDISI 02 • TAHUN XX • SEPTEMBER 2014 The Impact of Remi Ʃ ance on Labor Supply by Gender Case Study : Honduras Istasius A. Anindito 1 Abstract T ujuan dari makalah ini adalah melakukan invesƟgasi tentang bagaimana migrasi internasional dan remitansi mempengaruhi suplai tenaga kerja di negara pengirim mirgan. Penulis menggunakan data Living Standard Measurement Survey (LSMS) Honduras tahun 2006 yang diperoleh dari Bank Dunia dan melakukan metode two-step instrument variable (variabel instrumen dua tahap), yakni menggunakan data koneksi migrasi, serta menggunakan Heckman correcƟon model (model koreksi Heckman) untuk menghasilkan korelasi yang Ɵdak bias antara remitansi dan suplai tenaga kerja. Makalah ini menyimpulkan bahwa dampak dari rumah tangga penerima remitansi bervariasi tergantung dari jender kepala rumah tangga. Dalam hal pilihan sektor pekerjaan, kepala rumah tangga wanita penerima remitansi akan cenderung berpindah dari sektor formal dan pengangguran menuju sektor informal. Namun fenomena ini Ɵdak terjadi di rumah tangga pria penerima remitansi. Bagian kedua dari peneliƟan ini mempelajari dampak remitansi kepada jumlah total jam kerja. Setelah memperƟmbangkan dan memperbaiki sample selecƟon bias dan masalah endogenitas, model menunjukkan bahwa remitansi secara signikan mengurangi jumlah total jam kerja bagi kepala rumah tangga wanita. Namun, lagi-lagi, remitansi Ɵdak mempengaruhi jumlah total jam kerja bagi kepala rumah tangga pria. Kata Kunci : Remitansi, suplai tenaga kerja, pilihan lapangan pekerjaan, total jam kerja, mulƟnomial logit, model koreksi Heckman, variabel instrumen 1 Author is a stafrom directorate of industry, science-technology and tourism-creaƟve economy BAPPENAS and alumni of Graduate Program in Economic Development, Vanderbilt University INTRODUCTION What is remiƩance? RemiƩance is a transfer of money by internaƟonal migrants to individuals or households in their home country. In many of developing countries, remiƩance has played an increasingly important role and becoming an integral part of their economics. According to the World Bank (2010), internaƟonal remiƩance accounts for about one third of total global external nance. For some countries in LaƟn American and Caribbean (LAC) it accounts for more than one Ōh of their GDP. Increasing amount of remiƩance received by the developing countries aects the economy of the recipient countries in many areas, such as: labor supply, income inequality, and human capital accumulaƟon. The eect of remiƩance RemiƩance aects labor supply through increasing the lowest wage rate at which household member would be willing to accept a parƟcular job (reservaƟon wage). In some cases, remiƩance tends to shiŌ people from being employees to become a self-employed, while in other case it completely shiŌs people away from labor market. RemiƩance can also aect income inequality. A considerable amount of research has tried to examine the net impact of remiƩance on overall inequality map. Some believe that remiƩance improves income inequality because the recipients of remiƩance are not randomly drawn from the populaƟon; instead migrants are more likely coming from poorer households. Hence, remiƩance
Transcript

52 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

The Impact of Remi anceon Labor Supply by Gender

Case Study : HondurasIstasius A. Anindito1

Abstract

Tujuan dari makalah ini adalah melakukan inves gasi tentang bagaimana migrasi internasional dan remitansi mempengaruhi suplai tenaga kerja di negara pengirim mirgan. Penulis menggunakan data Living Standard

Measurement Survey (LSMS) Honduras tahun 2006 yang diperoleh dari Bank Dunia dan melakukan metode two-step instrument variable (variabel instrumen dua tahap), yakni menggunakan data koneksi migrasi, serta menggunakan Heckman correc on model (model koreksi Heckman) untuk menghasilkan korelasi yang dak bias antara remitansi dan suplai tenaga kerja. Makalah ini menyimpulkan bahwa dampak dari rumah tangga penerima remitansi bervariasi tergantung dari jender kepala rumah tangga. Dalam hal pilihan sektor pekerjaan, kepala rumah tangga wanita penerima remitansi akan cenderung berpindah dari sektor formal dan pengangguran menuju sektor informal. Namun fenomena ini dak terjadi di rumah tangga pria penerima remitansi. Bagian kedua dari peneli an ini mempelajari dampak remitansi kepada jumlah total jam kerja. Setelah memper mbangkan dan memperbaiki sample selec on bias dan masalah endogenitas, model menunjukkan bahwa remitansi secara signifi kan mengurangi jumlah total jam kerja bagi kepala rumah tangga wanita. Namun, lagi-lagi, remitansi dak mempengaruhi jumlah total jam kerja bagi kepala rumah tangga pria.

Kata Kunci : Remitansi, suplai tenaga kerja, pilihan lapangan pekerjaan, total jam kerja, mul nomial logit, model koreksi Heckman, variabel instrumen

1 Author is a staff from directorate of industry, science-technology and tourism-crea ve economy BAPPENAS and alumni of Graduate Program in Economic Development, Vanderbilt University

INTRODUCTION

What is remi ance?Remi ance is a transfer of money by interna onal

migrants to individuals or households in their home country. In many of developing countries, remi ance has played an increasingly important role and becoming an integral part of their economics. According to the World Bank (2010), interna onal remi ance accounts for about one third of total global external fi nance. For some countries in La n American and Caribbean (LAC) it accounts for more than one fi h of their GDP. Increasing amount of remi ance received by the developing countries aff ects the economy of the recipient countries in many areas, such as: labor supply, income inequality, and human capital accumula on.

The eff ect of remi anceRemi ance aff ects labor supply through increasing

the lowest wage rate at which household member would be willing to accept a par cular job (reserva on wage). In some cases, remi ance tends to shi people from being employees to become a self-employed, while in other case it completely shi s people away from labor market. Remi ance can also aff ect income inequality. A considerable amount of research has tried to examine the net impact of remi ance on overall inequality map. Some believe that remi ance improves income inequality because the recipients of remi ance are not randomly drawn from the popula on; instead migrants are more likely coming from poorer households. Hence, remi ance

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 49

Rp 10 -100 billion, approval from President is required. When the assets are not transferred, it creates opera on and maintenance lag because the ministries do not always allocate budget for opera on and maintenance and city/regency governments are not allowed to allocate budget if the assets are not owned by them. Consequently, rusunawa are easily deteriorated in the short me.

Second, even though local governments have commi ed to provide water supply and electricity in the agreement, some rusunawa are s ll not connected to water supply and electricity. For example, rusunawa in Surabaya and Jepara could not be u lized because the fl ats are not connected to water supply and electricity.

Third, rusunawa are poorly managed due to the absence of professional management unit and suffi cient fi nancial resource. City/regency governments usually impose low user charges to the occupants, which could not cover opera on and maintenance cost. Moreover, revenue from user charges could not be directly allocated for opera ng and maintaining rusunawa since all revenue must be transferred to local revenue offi ce (Picture 2). Meanwhile, the annual planning and budge ng process in city/regency governments might not always guarantee a suffi cient budget allocated for the rental fl ats. Without suffi cient fi nancial resource, the rusunwa management unit could not repair breakdowns in mely manner.

Picture - 2. Flows of User Charges and Opera on & Maintenance Budget

ASSESSMENTS OF THE RUSUNAWA PROGRAM

Expenditure AssignmentThe rusunawa program should be implemented by city/

regency governments rather than by central government, as regulated by the Government Regula on No. 38, 2007. There are fi ve reasons why city/regency governments are more suitable for this expenditure assignment.

First, the rusunawa program do not generate spillover eff ects across city/regency government’s jurisdic on. As argued by Robin Boadway and Anwar Shah, when the infrastructure generates benefi ts in local scale, the provision of this infrastructure should be decentralized to local level (Boadway, R., Shah, A, p. 147).

Second, local governments are likely be er informed on demand and need of rusunawa than the ministries; therefore, the local governments will be more capable in responding diff erent demands and needs (Boadway, R., Shah, A, p. 127). Even though the ministries have collected informa on and assessed local needs by calling proposals from local governments, the ministries s ll have limited me and resources for checking and verifying the validity

of informa on. Even more, the agreement among central, provincial and city/regency governments is not enough to guarantee that city/regency governments are able to acquire land and to provide u li es in mely manner as shown by many rusunawa have not been connected to water supply and electricity.

Third, city/regency governments might be more eff ec ve in implemen ng the rusunawa program since they will be able to integrate the program with other related programs, such as slum upgrading program. As argued by Boadway and Shah, “decentraliza on may lead to improvement and innova on in program design and program delivery because of the opportuni es and constraints faced by state-level decision makers.”

Fourth, if city/regency governments implemen ng the program, it will reduce administra ve costs of verifi ca on and transfer asset process since the assets of rusunawa will be directly own by city/regency governments.

48 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

Picture - 1b. Procedures and Steps of Implemen ng the Mul -Storey Rental Flats

Step1 - Preparing and Submi ng Proposal. Before construc ng the fl ats, the ministries need to iden fy city/regency governments that need and want to have mul -storey rental fl ats located in their region. The ministries call city/regency governments to prepare and submit proposal as part of the selec on process. The proposal has to refl ect local governments commitments to provide land, to allocat matching funds, to ensure availability of u lity and infrastructure, to ease building permit process and to ensure the u liza on, opera on and maintenance of constructed fl ats.

Step 2 – Verifying the Proposal. The ministries will verify the proposals by conduc ng fi eld visits. The verifi ca on aims to check the validity of proposal and the readiness of city/regency governments. During the verifi ca on, the Ministry of Public Works also involves provincial government.

Step 3 – Defi ning Program Recipients. Based on verifi ca on results, the ministries determine eligible city/regency governments as program recipients every year. In this step, the ministries, provincial and city/regency governments will sign an agreement that cover shared tasks and responsibili es in fi nancing, construc ng, opera on, maintaining and asset management.

Step 4 – Construc ng Rusunawa. Based on this agreement, the ministries will execute the procurement for Detailed Engineering Design (DED) and construct rusunawa.

Step 5 – Transferring Asset. A er construc on, the ministries will send requests to the Ministry of Finance to process assets transfer of rusunawa from central government to local governments.

Step 6 – Opera ng, Maintaining and Managing. A er the asset has been transferred, city/regency governments

are responsible to u lize, operate, maintain, manage and fi nance rusunawa. During this phase, city/regency governments need to establish a management unit that will be responsible on day-to-day opera on. The management units are usually under responsibility of local departments, such as Public Works Department or Dinas Pekerjaan Umum. City/regency governments could impose user charges to the fl at’s tenants.

Problems in Implemen ng the Rusunawa Program

Not all rusunawa are u lized a er the construc on phase. It is a classic problem that has happened in implementa on of the rusunawa program since 2003. According to data from the Ministry of Public Works, there are s ll 39% of unu lized rusunawa that were constructed between 2003-2009. Even more, roughly 85% of rusunawa constructed by the Ministry of Public Works in 2010-2011 were not u lized.While, according to data from the Ministry of Public Housing, 16% of rusunawa were not u lized.

Three causes of low u liza on of rusunawa are iden fi ed. First, not all rusunawa have been transferred to city/regency governments. The Ministry of Finance is processing the transfer of 159 out of 393 twin towers rusunawa constructed from 2003-2012 by the Ministry of Public Works. The transfer asset from central government to city/regency government is a lengthy process because the Ministry of Finance has to undertake several steps, such as asset valua on, administra ve checking and technical assessment. Moreover, if the value of rusunawa is between

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 53

could improve income inequality. Another thought is that remi ance aff ects human capital accumula on in terms of children’s schooling-enrollment and schooling-a endance. Remi ance received by households prevents child labor and provides more resources to fi nance children tui on fees and schooling supplies.

The aim of this paper is to focus on the impact of remi ance on labor supply. Focusing on labor supply, this paper tries to specify the possible route of remi ance eff ects, fi nds the correct labor supply variables, selects the correct control variables, solves the possible sample selec on and endogeniety problem, and eventually interprets the result and understanding the limita on of the model.

The Framework ModelAs explained, there at least three routes of remi ance

eff ects on the economy. Figure 1.1 helps to explain basic conceptual framework of the models that are used in this paper.

Figure 1.1. Basic Framework of the paper

This paper focuses on labor supply by using two main variables which are: decision to work and total hours worked. These two main variables are the main objec ve variables or Le Hand Side (LHS) variable, while remi ance and other control variables are the independent variables or Right Hand Side (RHS) variables.

LITERATURE REVIEWThere are considerable amount of research covering

interna onal migra on and remi ance. All studies selected here are the one which uses current microeconomic survey data on the period between 1990 and 2014. Since countries receiving remi ance are not only La n American and Caribbean (LAC), the literature review presented here is not limited to LAC (Mexico, El Salvador, Nicaragua, Dominican Republic), but several countries as well such as in Central Asia (China, Tajikistan), South Asia (India), South-East Asia (Philippines, Indonesia), and also from Middle East (Lebanon) and Africa (Ghana). In spite of many micro level data being used, none of them—except one—using panel data, which is a small drawback of this research.

Adams (2010) on his review of the economic impact of remi ance and interna onal migra on claimed that migra on and remi ance tend to reduce labor supply and par cipa on of migra ng sending household. He argued remi ance increases preserva on wage and provides start-up fund to create family businesses. In Mexico, Pfeiff er and Taylor (2008) found female migrant associate with less educa on spending than those with male migrant. Guzman et al. (2008) found in Ghana, household receiving

remi ance from female have higher expenditure on health, but lower expenditure in educa on and food. However these ar cles in Mexico and Ghana did not taking into account for possible endogeniety problem. Acosta (2011) tried to use panel data in rural El Salvador to be er explain and overcome reverse causality, sample selec on, and omi ed variable problem. He concluded that male migra on has null to slightly posi ve eff ect on children enrollment ra o while female migra on has the opposite eff ect. Cabeign (2006) using two stage probit model found in Philippine, remi ance reduces the probability to work full me for both married man and woman.

Nguyen and Purnamasari (2011) using Indonesia Family and Life Survey (IFLS) 2007, found that parallel with most of the literature reviews, interna onal migra on has a nega ve impact on average labor supply of remaining household members. In detail, this is not the case of household head withdraws from the labor force but

54 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

rather every member works fewer hours than before. Overall, migra on reduces working hour. The simple OLS regression, suggest the migrant sending household tend to work less than non-migrant sending household. The migrant sending household work 5.99 hour/week less than non-migrant sending household (mean 75 hours/week). When the authors included the IV, migrant sending household works even fewer (26.34 hours/week less) than non-migrant sending household. However, when the analysis considering only female migrant, the migrant sending household does not reduces the working hour.

In Nicaragua, Funkhouser (1992) argued that increase in remi ance would have nega ve impact on domes c labor force par cipa on but a posi ve impact on self-employment. In accordance with Funkhouser, Woodruff and Zateno (2004) argued in Mexico remi ance increase small enterprise growth investment. Cox-Edwards and Ureta (2003) concluded that remi ance reduce school dropout ra o in El Salvador. Hanson and Woodruff (2003) found posi ve rela onship between receiving remi ance and child educa on. In Philippines, Rodriguez and Tiongson (2001) concluded that remi ance reduces labor par cipa on of the recipient household. Labor par cipa on for men reduces by 9.4 percent while for women it reduces by 18.1 percent. However, when the migrant has ter ary educa on interes ng eff ects occurs. Labor par cipa on for men reduces by 2.9 percent, while for women increases by 12.4 percent. For women, having a ter ary educa on migrant somehow signifi cantly “inspire” them to par cipate in the working labor. While Acosta (2006) in El Salvador, provides more careful research which controlling possible selec on bias and household wealth, found that remi ance has nega ve eff ect on child labor and adult female labor supply, but not aff ects adult male labor par cipa on.

HypothesisHouseholds receiving remi ance have the luxury

of having more non-labor income. Households receiving remi ance also possess the capital required to start a family business and thus may shi people away from being formal employees. Remi ance provides the resources required by the household members and broadens their opportuni es. Thus, the fi rst hypothesis of this model is that remi ance do aff ects household working choice.

Second hypothesis, remi ance not only increases reserva on wage of the recipient, but also provides household member with constant monthly non-labor income. Even though li le is known about the con nuity of this remi ances—whether these remi ances are permanent or temporary shock—remi ance may aff ect people decision on how many hours they are willing to work. According to basic income and subs tu on eff ect theories, people receiving remi ance would trade some of their working hours with more leisure me. Hence, the second hypothesis is remi ance do reduces total working hours.

DATA DESCRIPTION

Honduras Socioeconomic Background Honduras is a poor country with a total popula on

8 million people, located in Central America, bordered by Caribbean Sea (north), Guatemala (northwest), El Salvador (southwest), pacifi c ocean (south), and Nicaragua (east). According to CIA World Factbook, in 2010, 60% of its popula on lives under poverty line. The poverty is more severe especially in the rural areas and among indigenous people in South and West part of Honduras. GDP per capita (PPP) is around USD 4,800. Its economic growth in 2010-2012 was steady at 3.0-4.0%, which was lower than many developing countries, but was much higher than the 1980’s lost decade. Honduras Human Development Index (HDI) is very low, even compared to their peers. In Central America region, Honduras has the second lowest HDI, outranking only Nicaragua.

The U.S. dominate its share in Honduras total export and Foreign Direct Investment (70%). Honduras export accounted for 30% of the GDP and remi ance accounted for more than 20% before it went down because of the 2008 U.S. fi nancial crisis. The remi ance/GDP ra o peaked at 21.5% in 2006 and is 16.1% to date. Historically, they heavily depended on export agriculture commodi es, such as banana and coff ee, with low-skilled labor work in agricultural sector. Since the introducing of CAFTA-DR (Central America-Dominican Republic Free Trade Agreement) in 2006, bilateral trade between the U.S. and Honduras has drama cally increased. Honduras has benefi ted from export apparel enhancement to the U.S since then.

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 47

This research paper aims to analyze current implementa on of the rusunawa program including sharing of responsibili es, procedures of implementa on and problems. Further more, this paper will assess the possibility to decentralize the implementa on of the rusunawa program. The assessment covers expenditure assignment, revenue assignment, intergovernmental transfer and capacity building.

THE MULTI STOREY RENTAL FLAT PROGRAM RUSUNAWA

Rusunawa are rented out temporarily to low-income families who cannot aff ord to buy home in urban area. A fl at usually consists of 96 units and area of unit is varied from 24 – 36 m2. For example, a rusunawa in Sleman, Yogyakarta is 5-storey fl at that has 96 units, in which each unit is 27 m2 (Picture 1). In the following, I will elaborate more about sharing of responsibili es, implementa on and problems of the rusunawa program.

Responsibili es of Rusunawa Program According to the Government Regula on No. 38,

2007, responsibili es in providing rusunawa are shared among central, provincial and city/regency government. Central government is responsible to allocate budget for construc ng fl ats, u li es and infrastructure. Provincial government is responsible to construct fl ats, u li es and infrastructure, and to assist city/regency government in managing and maintaining the fl ats. City/regency government is responsible to construct the fl ats, u li es and infrastructure.

In fact, not only alloca ng budget, central government is actually construc ng fl ats in selected loca on with the Ministry of Public Housing and the Ministry of Public Works as implemen ng agencies (Table 1). The ministries are also in charge in establishing regula on and standard as well as conduc ng monitoring and evalua on. Provincial government is involved in planning, especially selec ng eligible city/regency governments as well as monitoring

and evalua on. Though city/regency governments are involved in u lizing, opera ng and maintain the fl ats a er the assets are transferred from central government to city/regency governments, involvement in planning and budge ng process is limited to preparing and submi ng proposal to be eligible as program recipients.

Procedures in Implemen ng the Rusunawa Program

The Ministry of Public Housing and the Ministry of Public Works have their own procedures in implemen ng the rusunawa program. In general, the procedures consist of six steps from submi ng city/regency government

proposal to opera ng and maintaining the fl ats (Picture 1b).

Table - 1. Current Shared Responsibility among Levels of Government

ActorProvision Produc on U liza on Policy

Planning & Budge ng Financing Cons

truc onAsset

Transfer Occupancy Opera on & Maint-enance

Regula on & Standards

Monitoring & Evalua on

Ministries v v v v v vProvincial government v vCity/regency government V v v v v vUsers v

Source : several sources

Sour

ce :

h p

://r

usun

awa.

slem

anka

b.go

.id

Picture - 1a. A Mul -Storey Rental Flat in Sleman

46 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

Financing public housing in Indonesia : the role of central government and local governments

Case study : Mul -Storey Rental FlatsIra Lubis1

Abstrak

Pemerintah Indonesia melaksanakan program rumah susun sederhana sewa (rusunawa) untuk menyediakan hunian bagi rumah tangga berpenghasilan rendah. Dalam Rencana Pembangunan Jangka Menengah Nasional (RPJMN)

2010-2014, pemerintah menargetkan pembangunan 650 twin blok rusunawa. Pelaksanaan pembangunan rusunawa tersebut menghadapi kendala, baik dalam tahap perencanaan maupun pemanfaataan. Hal ini ditunjukan dari belum terhuninya 39% rusunawa yang dilaksanakan oleh Kementerian Pekerjaan Umum pada 2003-2009 dan 16% rusunawa yang dilaksanakan oleh Kementerian Perumahan Rakyat. Penyebab rendahnya keterhunian rusunawa mencakup lambatnya proses pengalihan kepemilikan aset rusunawa dari pemerintah pusat kepada pemerintah daerah, dak tersedianya infrastruktur pendukung rusunawa dan lemahnya pengelolaan rusunawa secara professional. Untuk mengatasi permasalahan tersebut, pembangunan rusunawa sebaiknya dilakukan oleh pemerintah kabupaten/kota dengan menggunakan mekanisme dana alokasi baru atau mendesain ulang Dana Alokasi Khusus Bidang Perumahan dan Kawasan Permukiman yang telah ada. Mekanisme dana alokasi khusus tersebut sebaiknya berdasarkan output-based matching transfer dan serta disertai dengan peningkatan kapasitas untuk transformasi unit pengelola rusunawa yang lebih profesional.

1 Staf Perencana di Direktorat Permukiman dan Perumahan, Bappenas, Jalan Taman Suropa 2, Jakarta 10310. Tel. (021) 319-34819, Email : [email protected].

INTRODUCTIONProviding aff ordable housings for all ci zens is one

of objec ves s pulated in the Indonesia Na onal Long-Term Development Plan 2005-2025. The Government of Indonesia has responsibility to provide housing for all. Most of housing for the upper-to-middle-income households is provided through housing market. However, the lower-to- middle-income households have more limited choices in the market since they have lower ability to pay. To address this issue, the government provides public housing for low-income households.

The mul -storey rental fl at or rumah susun sederhana sewa (rusunawa) is one of central government programs that aims to provide aff ordable housings for low-income families. In the Na onal Medium-Term Development Plan 2010-2014, the Ministry of Public Housing has responsibility to construct 380 twin towers the Ministry of Public Works

has responsibility to construct another 270 twin towers of rusunawa.

Implementa on of the rusunawa program faces several problems. In the planning phase, both ministries face diffi cul es in iden fying city/regency governments that are interested and ready to par cipate in the program. City/regency governments are required to prepare land for rusunawa and ensure the availability of water supply, electricity and access road. A er the construc on phase, some city/regency governments face diffi cul es in u lizing rusunawa since assets transfer from central government to city/regency government is lengthy. In addi on, some of city/regency governments do not allocate enough budgets for opera on and maintenance of rusunawa. Consequently, some rusunawa are deteriorated within a short period of me.

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 55

Honduras job market characteris c is dominated by informal sector. In 2006, share of informal sector workers was 71%, compared to 58% in Mexico (ISACC, 2009). Honduras is among the lowest rank in the World Bank’s Doing Business Index (DBI). In 2014, its DBI rank is 127th out of 189 countries. Registering and running a formal business en ty have become challenging tasks for Hondurans.

The U.S. is also a primary des na on for Honduras migrants. A er the Mitch Hurricane (1998), around 90% of Honduran migrants have been living in the U.S. The remaining 10% sca ered in Canada, Spain, Mexico, and other neighboring countries in Central America. Honduran migrants are dominated by young and low-educated workers. Around 65% of them are aged 15-29 years old and 59% of them have completed primary or lower educa on. Tradi onally, rural Hondurans accounted for about two third of the total migrant. Since 2006, the propor on of rural and urban Honduran migrants has become more equal 50-50 (CEMLA, 2008).

Living Standard Measurement Study (LSMS)This research uses micro-level data from Living

Standard Measurement Study (LSMS) of 2006 conducted by the World Bank. The 2006 LSMS dataset consist of basic household characteris cs such as age, marital status, educa on, total member of household, children and adult living in household, total income, total hours worked ability to read/write, region, city, village, zip-code, urbanness. This dataset also contain informa on on interna onal migra on (informa on on receiving remi ance or not, amount of remi ance received, and number of family living abroad) and few other informa on. The survey was randomly collected in April 2006 and amounted to 21,076 head of household. However, since this research aim to examine the impact of remi ance on labor supply, this observa on needs to be confi ned into working age only. Thus, all observa ons must age 15 to 65 years old. This process dropped a few observa ons and in the end this research used 18,140 observa on—of which 13,141 are male and 4,999 are female (Table 2.1). In terms of receiving remi ances, 3,271 respondents reported receiving remi ance and 14,869 respondents reported not receiving remi ance. In terms of schooling, the mean years of schooling in Honduras is 6.8 years, meaning is majority of head of the household for this sample only fi nished elementary school.

When comparing working sector profi le across gender, there is a clear dis nct diff erence between men and women. While the percent of not working and working in informal sector is more or less the same, the percentage of individual working in public and private sector is really diff erent. Out of all female head of household in Honduras, 12.68 percent of them work as a public employee, while for male head the fi gure is about one-half which is 6.43 percent. Most of the men, work in private sector (41 percent), of course a er working in informal sector (50.1 percent). Bo om line, it is unlikely that working sector choice is randomly drawn across gender. Women are more likely to be off ered a public worker or women prefer working in public sector rather than any other sectors.

The hours worked profi le in 2006 LSMS for men is 47.3 hours/week while for women is 39.6 hours/week. The mean of hours worked for the en re popula on is 45 hours/week.

THE MODEL

Labor Supply Measurement 1 : Working Sector Choice

The fi rst ques on :The following are four separate logit regressions to

predict the likelihood of an individual to work in a specifi c sector.

Logit regressions

Where : is the dummy variable, =1 if not working, =0

otherwise is the dummy variable, =1 if working in public, =0

otherwise is the dummy variable, =1 if working in private,

=0 otherwise

56 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

is the dummy variable, =1 if working in informal, =0 otherwise

is the n vector characteris cs of household i is the rela ve remi ance per-capita of household i

is the unobservable

The second equa on :MLE model of probability

Where : probability if not working probability if working in public sector probability if working in private sector probability if working in informal sector vector characteris c include female, age, marital

status, urban, read and write, schooling level, log income per capita, total family member, worker ra o, number of children

is rela ve remi ance per capita of household i is the unobservable component

The most important coeffi cient es mator is as it would represent the eff ect of remi ance to the possibility of individual move away to par cular sector, rela ve from their reference sector.

Labor Supply Measurement 2 : Total Weekly Hours Worked

Using Two Stage Heckman with IV to solve poten al endogeniety and sample selec on bias.

Where : is hours worked of head of household i

vector of household characteris cs include female, age, age-squared, marital status, urban,

read and write, years of schooling, log income per capita, total family member, number of family member living abroad, formal sector, worker ra o, number of children

is whether not a household i receiving interna onal remi ance.

is the unobservable component of household i is the instrument variable (network) of household i

The most important result here is as it would represent the impact of remi ance to the total working hours, assuming all other control are constant (ceteris paribus).

RESULT AND DISCUSSIONWorking Status Choice

In the fi rst model, I es mated the eff ect of remi ance to head of household working sector decision. One technique to test working sector decision for the head of household is by mul nomial logit model. This technique produces the coeffi cients of each objec ve and control variables to the possibility of an individual enter par cular sector. The objec ve variable in this model is rela ve remi ance per capita (r_remi ance_pc), which defi ne as total remi ance received in one household each month divides by income per capita.

Where is rela ve remi ance per capita of household i, is monthly remi ance received by household i, is total monthly income in household i, is total number of people living in household i.

Employing mul nomial logit, this model may generate biased es mator of the coeffi cient. There is poten al endogeniety and omi ed variable bias in the model. Thus, to solve these problems, I employed Instrumental Variable (IV). I an cipated to produce unbiased es mators of all betas a er the introducing of IV.

The IV which had been used in this model is interna onal networking. Networking or family, friends, or neighbors living abroad plays an important role on the decision to migrate. Having friends or rela ves abroad

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 45

Lockwood, B. (2006) “The poli cal economy of decentraliza on”. In Ahmad, E., and Brosio, G. (ed.) Handbook of fi scal federalism. United Kingdom : Edward Elgar.

Miller, M. A., and Bunnel, T. (2012) Introduc on : Asian Ci es in an era of decentraliza on. Space and Polity. 16 (1) : 1-6

Musgrave, R. A. (1959) The theory of public fi nance. New York : McGraw Hill.

Ners, K. (2007) “Infrastructure development in Poland : the issues at stake”. In Peterson, G. E., and Annez, P. C (ed. ). Financing Ci es : Fiscal Responsibility and Urban Infrastructure in Brazil, China, India, Poland and South Africa.Washington : The World Bank and Sage, p. 158-182

Peterson, G. E., and Annez, P. C. (2007) “Na onal frameworks to balance fi scal discipline and local investment needs.” In Peterson, G. E., and Annez, P. C (ed.). Financing Ci es : Fiscal Responsibility and Urban Infrastructure in Brazil, China, India, Poland and South Africa.Washington : The World Bank and Sage, p. 27-39

Poli , C., and Bouckaert, G. (2011) Public management reform : a compara ve analysis, new public management, governance and the Neo-Weberian state. 3rd ed. Oxford : Oxford University Press.

Qiao, B., and Shah, A. (2006) “Local government organiza on and fi nance : China”. In Shah, A. Local Governance in Developing Countries. Washington : The World Bank.

Robinson, M. (2007) Does decentraliza on improve equity and effi ciency in public service delivery provision? IDS Bulle n, 38 (1) : 7-17

Rondinelli, D. A., and Cheema, G. S. (2007) “ Implemen ng decentraliza on policies : an introduc on”. In Rondinelli, D. A., and Cheema, G. S. (ed). Decentralizing Governance : Emerging Concepts and Prac ces. Washington : Brooking Ins tu on Press, pp. 1-20

Salim, W., and Kombaitan, B. (2009) Jakarta : the rise and challenge of Capital. City. 13 (1) : 120-127

Sassen, A. (2005) The global city : introducing a concept. The Brown Journal of World Aff airs. XI (2) : 27-43

Shakarishvili, G., and Davey, K. (2005) “Introduc on : trends in reforming the provision and fi nancing of healthcare services in CEE/CIS regions during 1990s.” In Shakarishvili, G (ed) Decentraliza on in Healthcare : Analyses and Experiences in Central and Eastern Europe in the 1990s. Budapest : Local Government and Public Service Reform Ini a ve and Open Society Ins tute, pp. 1-44

Smith, B.C. (1985) Decentraliza on : the territorial dimension of the state. London : Allen and Unwin.

Strategic Asia. (2013) Decentraliza on assessment report. Jakarta : UNDP and Strategic Asia.

Texier, P. (2008) Floods in Jakarta : when the extreme reveals daily structural constraints and mismanagement. Disaster Preven on and Management. 17 (3) : 358-372

Treisman, D. (2007) The architecture of government : rethinking poli cal decentraliza on. Cambridge : Cambridge University Press.

UCLG (2009) Decentraliza on and local democracy in the World : First Global Report. Barcelona : United Ci es and Local Government and the World Bank.

UN-Habitat (2008) State’s of the world ci es 2008/2009 : harmonious ci es. London : Earthscan.

Vazquez, J. M., and Vaillancourt, F. (2013) “An overview of the main obstacles to decentraliza on”. In Vazquez, J. M., and Vaillancourt, F (ed.) Decentraliza on in Developing Countries : Global Perspec ves on The Obstacles to Fiscal Devolu on. The United Kingdom : Edward Elgar, p. 1-22

Wa s, R. L. (2008) Comparing federal systems. 3rd Ed. Canada : McGill Queen University Press.

Wong, C.P.W. (1991) Central – local rela ons in an era of fi scal discipline : the paradox of fi scal decentraliza on in Post Mao China. The China Quarterly. 128 (1991) : 619 -715

Xu, C. (2011) The fundamental ins tu ons of China’s reforms and development. Journal of Economic Literature. 49 (4) : 1076-1151

Zuo, X. (2009) “China decentraliza on and its impact on urbaniza on.” In Ichimura, S., and Bahl, R. Decentraliza on policies in Asian Development. Singapore : World Scien fi c p. 85-106

44 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

REFERENCE ADB (2008) Managing Asian Ci es : sustainable and inclusive urban solu ons. The Philipines : Asian Development Bank.

Badcock, B. (2002) Making sense of ci es : a geographical survey. London : Arnold Publisher.

Bagaeen, S. (2007) Brand Dubai : the instant city; or the instantly recognizable city. Interna onal Planning Studies. 12 (2) : 173-197

Bahl, R. (2009) “Promise and reality of fi scal decentraliza on”. In Ichimura, S., and Bahl, R. Decentraliza on Policies in Asian Development. Singapore : World Scien fi c, pp.1-26

Bardhan, P., and Mookherjee, D. (2006) “The rise of local governments : an overview.” In Bardhan, P., and Mookherjee, D. Decentraliza on and Local Governance in Developing Countries : A Compara ve Perspec ve. Massachuse s : MIT Press, p.1 –52

Blore, I., Devas, N., and Slater, R. (2004) Municipali es and fi nance : a sourcebook for capacity building. United Kingdom : Earthscan.

Bunnel, T., Miller, M. A., and Phelps, N. A. (et al., 2011). “Urban development in a decentralized Indonesia : two success stories?” In the Joint Conference of The Associa on of Asia Studies (AAS) and Interna onal Conven on of Asia Scholar. Hawaii, 2011.

Campbell, A., and Denezhkina, E. (2009). “Ver cal and horizontal coordina on in Russian Local Government”. In the European Poli cal Research Consor um Conference. University of Postdam, 11 September 2009.

Chalkley, B., and Essex, S. (1999). Urban development through hos ng interna onal events : a history of the Olympic Games. Planning Perspec ves. 14 (1999) : 369 -394

Cohen, J. M., and Peterson, S. B. (1999) Administra ve decentraliza on : strategies for developing countries. Connec cut : Kumarian Press.

Davey, K., and Devas, N. (1992) Urban Government Finance : The ins tu onal framework of Urban Management Working Paper No 4. Birmingham : Development Administra on Group, School of Public Policy, the University of Birmingham.

Devas, N., and Delay, S. (2008) “Local democracy and the challenges of decentralizing the state : an interna onal perspec ve”. In Coulson, A., and Campbell, A. (ed.) Local Government in Central and Eastern Europe : The Rebirth of Local Democracy. United Kingdom : Routledge, pp. 135-153

Fengler, W., and Hofman, B. (2009) “Managing Indonesia’s rapid decentraliza on : achievements and challenges”. In Ichimura, S., and Bahl, R. Decentraliza on Policies in Asian Development. Singapore : World Scien fi c, p. 245-262

Firman, T. (2003) Poten al impacts of Indonesia’s fi scal decentralisa on reform on urban and regional development : Towards a new pa ern of spa al disparity. Space and Polity. 7 (3) : 247-271

Firman, T. (2009) The con nuity and change in mega-urbaniza on in Indonesia : A survey of Jakarta–Bandung Region (JBR) development. Habitat Interna onal. 33 (4) : 327-339

Gaubatz, K. T. (2009) City-state redux : rethinking op mal state size in an age of globaliza on. New Global Studies. 3 (1) : 1 -211

Gold (2010) Local Government Finance : the challenges of 21th Century, second global report on decentraliza on and local democracy. Barcelona : United Ci es and Local Government and The World Bank.

Gruber, J. (2010) Public fi nance and public policy. 3rd ed. The US : Worth Publisher.

Hague, R., and Harrop, M. (2010) Compara ve government and poli cs : an introduc on. 8th ed. London : Palgrave Macmillan.

Hall, P. (2005) The world’s urban systems : a European perspec ve. Global Urban Development. 1 (1) : 1-11.

Homme, R.P. (1995) The dangers of decentraliza on. The World Bank Research Observer. 10 (2) : 201-210

Kazepov, Y. (2005) “Ci es of Europe : changing contexts, local arrangements, and the challenge to social cohesion”. In Kazepov, Y. (ed.) Ci es of Europe : Changing Contexts, Local Arrangements, and The Challenge to Urban Cohesion” The United Kingdom : Blackwell Publishing, pp. 1-42

Kotkin, J. (2005) The city : a global history. London : Weidenfeld & Nolson.

Kris ansen, S., and Santoso, P. (2006) Surviving decentraliza on? Impacts of regional autonomy on health service provision in Indonesia. Health Policy. 17(2006) : 247-259

Lewis, B.D. (2003) Minimum local public service delivery standards in Indonesia : fi scal implica ons and aff ordability concerns. Research Triangle Ins tute Interna onal. Available from : h p ://www1.worldbank.org/publicsector/decentraliza on/Feb2004Course/Background%20materials/Lewis.pdf [Accessed 6 December 2013]

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 57

is associated with having more probability of becoming a migrant. Hence, networking would serve as a great candidate for IV because it directly aff ects remi ances but not directly aff ects working sectors decision. Networking only aff ects working sector decision through remi ance received. Because the LSMS 2006 did not contain any ques on about number of neighbor, friends, rela ves working abroad, I had to construct variable networking based on the availability of the data. I defi ned network as the propor on of family in the same zip-code who receive remi ance.

Where is network of household of i, is number of households receiving remi ance in par cular neighborhood area of j, and is total number of households in par cular neighborhood. In dataset, there are four level of neighborhood area that can be u lized to construct this IV. Depto, munic, aldea, caserio represents state, city/town, village, and zip-code. For this model, I used level 4 neighborhoods (zip-code) and level 3 neighborhoods (village). Level 4 neighborhood is provides informa on from friends or rela ves living in the same zip-code which is crucial to provide informa on for future migra on. However, with the development of technology where people can easily access informa on from their friends or rela ves not too close from their home, level 3 (village) network might serve a good candidate for an IV as well.

The mul nomial logit model u lized in this research produced coeffi cient of the possibility of changing one’s dependent variables would change one’s working sector. Working sector for head of household being defi ne as (i) not working, (ii) working in public sector, (iii) working in private sector, and (iv) working in informal sector. For the purpose of this research, this model will use informal sector as a reference group.

The fi rst result (Table 4.4) compares the possibility of switching between not working and informal sector. The column (1) is when the model applied to overall sample, while column (2) and (3) are when the observa ons is confi ned to female head of household and male head of household, respec vely. The objec ve variable—which is the remi ance—shows no signifi cant correla on in the

overall sampling and male head. Remi ance, nonetheless, shows a signifi cant correla on with working sector choice when I confi ned my observa on to female head. Every 10 percent of increase in rela ve remi ance per capita would induce 5.45% shi ing from not working to informal sector. The other control variables show signifi cantly aff ects working sector decision between not working and informal such as : gender, age, log income per capita, total number of children, and worker ra o.

Tabel 4.4. Working Sector Choice between Not Working-Informal Sector

Dependent Variable :Working Sector Choice

(1)MLE_IV

(2)MLE_IV_F

(3)MLE_IV_M

Not working to Informal sector

Female 0.454**

(0.174)

Age 0.0148** -0.00298 0.0158*

(0.00553) (0.0124) (0.00674)

Married -0.00287 0.230 -0.0299

(0.126) (0.421) (0.134)

urban 0.0621 -0.616* 0.248

(0.132) (0.264) (0.150)

ReadWrite -0.367 -0.442 -0.362

(0.229) (0.521) (0.259)

Schoollevel -0.00230 -0.0203 0.00516

(0.0522) (0.118) (0.0586)

linc_pc 0.271*** 0.313 0.263***

(0.0640) (0.160) (0.0698)

TChildLT6 -0.616*** -0.550*** -0.620***

(0.0715) (0.164) (0.0814)

TChild7_14 -0.487*** -0.581*** -0.459***

(0.0570) (0.142) (0.0625)

workerra o -12.53*** -12.99*** -12.59***

(0.508) (1.041) (0.587)

r_remi ance_pc_IV -0.131 -0.545* -0.000976

(0.151) (0.250) (0.221)

_cons -0.0462 2.008 -0.210

(0.518) (1.248) (0.587)

N 15215 2992 12223

pseudo R-sq 0.165 0.260 0.138

Standard errors in parentheses

* p<0.05 ** p<0.01 *** p<0.001

58 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

The second result (Table 4.5) compares the possibility of switching between public sector and informal sector. Rela ve remi ance per capita shows signifi cantly aff ects in the overall and female head test group. While remi ance is not signifi cantly aff ect male head test group. Every 10 percent of increase in r_remi ance_pc would induce 3.56% shi ing from public sector to informal sector for overall sample and induce 4.43% for female head test group. These results are very strong fi ndings against the common believe of public sector worker rigidity. Public sector workers are associated

with educated, diffi cult to a ain, and for exis ng worker the job is very unlikely to forgo. In this model, even though male head would not sacrifi ce his public sector works when he receives remi ances, female head would forgo her works and join informal sectors instead. One possible explana on of this phenomenon is it could be a ributed to the role of husband and wife tradi on in Honduras. Tradi onally, wives are expected to stay at home and raise children while husband go to work to provide for the family. A wife or a mother, is preferably working at home in a micro-scale self-employment thus she would be able to take care of the children rather than going to work. However, a li le can be conclude from this model as the dataset did not contain any variable explaining self-employment.

The other control variables in Table 4.5 show interes ng correla on. Income per capita coeffi cient shows signifi cant result that more wealthy people are more likely to be work in the public sector. The higher educa on someone possesses would be more likely to work in the public sector as well. While age and total number of children is not signifi cantly aff ect working sector choice between public and informal sector.

The third result (Table 4.6) compares the possibility of switching between private sector and informal sector. Although remi ance insignifi cantly aff ects male head working sector choice, remi ance signifi cantly aff ects overall and female head working sector choice. Every 10 percent of increase in r_remi ance_pc would induce 2.1% shi ing from private sector to informal sector for general sample and induce 6.5% for female head of household. Again, remi ance aff ects women working sector choice, but not for men. Receiving remi ances, women tend to forgo her formal works, public and private, to work in informal sector. This is indeed a very strong fi ndings because it is not just against the public sector worker rigidity but all of the evidences persistently displayed a convergence towards informal workers.

The other control variable in Table 4.6 shows very compelling story about the dis nct characteris c of private sector worker about being young, unmarried, educated and living in urban because variable Age, Married, Schoollevel, and urban signifi cantly aff ect probability to pick between private and informal sector worker.

Tabel 4.5. Working Sector Choice between Public Sector-Informal Sector

Dependent Variable :Working Sector Choice

(1)MLE_IV

(2)MLE_IV_F

(3)MLE_IV_M

Public sector to Informal sector

Female 0.412***

(0.113)

Age 0.00141 -0.00708 0.00545

(0.00332) (0.00698) (0.00430)

Married -0.203* 0.0220 -0.211*

(0.0820) (0.246) (0.0878)

urban 0.660*** 0.506* 0.779***

(0.0976) (0.202) (0.113)

ReadWrite 2.787*** 4.302*** 2.311***

(0.214) (0.499) (0.240)

Schoollevel 0.540*** 0.828*** 0.453***

(0.0283) (0.0607) (0.0327)

linc_pc 0.655*** 0.888*** 0.595***

(0.0481) (0.104) (0.0543)

TChildLT6 0.0629 0.0357 0.0968

(0.0472) (0.0991) (0.0557)

TChild7_14 0.0320 0.200* -0.00936

(0.0407) (0.0901) (0.0459)

workerra o -1.324*** -1.787*** -1.005***

(0.205) (0.443) (0.224)

r_remi ance_pc_IV -0.356** -0.443* -0.377

(0.133) (0.197) (0.203)

_cons -10.39*** -13.29*** -9.657***

(0.393) (0.888) (0.452)

N 15215 2992 12223

pseudo R-sq 0.165 0.260 0.138

Standard errors in parentheses

* p<0.05 ** p<0.01 *** p<0.001

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 43

commitment and clear division authori es. It is formally s pulated in cons tu on/legisla on and prac cally set through procedures/ norms which regulate actors who will be responsible for par cular ac ons/func ons. Cohen and Peterson (1999) no ced in developing Asian Countries that poli cal accountability covers state commitments, local leaders and actors as well as clear task divisions; it guarantees that administra ve opera ons can run, fi nancial resources can be eff ec vely and effi ciently u lized and stable central-local rela onships can be nurtured.

In a centralized system the responsibili es of the central government and mandates given to local governments should be clear. In decentralized system, as the interests of local governments are o en diff erent from the central government’s, it is much more important for clearly divided responsibili es between state and ci es and formal commitments s pulated in legisla on. Strategic Asia (2013) records that since the division of authori es are s ll unclear and a low poli cal commitment from both local and central governments in the provision of basic public services; cases of malnutri on, starving, a lack of a clean water supply, slums, squa ers in Jakarta, Surabaya and many big ci es in Indonesia s ll remain. Shakarishvili and Davey (2005) add that the decentraliza on of healthcare in Central and Eastern Europe (e.g. Georgia, Croa a, Bulgaria) has not been successful since there is less poli cal support from the state as well as an unclear division of health responsibili es between diff erent levels of government.

Another important element is coordina on. This is part of public management ensuring that communica on and interac on procedures between diff erent levels of government are set to achieve common goals (Polli and Bouckaert, 2011). Campbell and Denezhkina (2009) label it ver cal coordina on, which explains the rules of the interac ons between diff erent levels of government, dis nguishing it from horizontal coordina on. The failure to overcome the danger of fl ooding in Jakarta and its surrounding areas was caused by a lack of coordina on between the state and ci es and amongst the ci es. The central government blamed the fl ooding on the failures of Jakarta to manage their river, while Jakarta claimed it was the failure of the central government and other surrounding areas by allowing people to live in catchment areas which caused the water to overrun Jakarta (Texier, 2008).

In addi on, coordina on between state and city cannot be benefi cial for either party if the capacity of state and city employees is low. Blore, Devas and Slater (2004) based on experience of India and some Asian Countries consider that the capacity of local government offi cers in municipali es is essen al to coordinate the implementa on of decentraliza on policy, eff ec vely and effi ciently in u lizing money, achieves a high performance in delivering local services and maintains benefi cial rela ons between the state and municipali es.

CONCLUSIONIt can be concluded that fi nance is an essen al

component in state-city rela onships. It has a very important role in suppor ng both the state and city to achieve their developmental goals by facing the challenges of urbaniza on and globaliza on. Using a mul -level governance concept, two main types of state-city fi nancial rela onship can be iden fi ed : a centralized and decentralized system. In centralized systems, the state ghtly controls and scru nizes city spending and revenues and ci es have a high dependency on state grants. Contrary to this, in a decentralized system, the state transfers more funds to the city and to some extent gives discre onary power to the ci es to generate their own revenues.

Moreover, the dynamics and complexity of state-ci es rela ons can be clearly felt in a decentralized system, when local needs are taken into account with na onal interests in equal balance, poten ally benefi ng both par es. However, in decentralized system, state-city rela onships are o en tense since there is a confl ict of interests caused by the diff erent needs of both city and state. O en, the interests of state should be sacrifi ced to fulfi l ci es’ needs and vice versa.

However, the role of fi nance itself is not suffi cient to maintain mutually benefi cial and stable state and city rela onships. In this paper, we acknowledge poli cal accountability and coordina on as other equally important components in state-city rela onships, complemen ng fi nancial aspects in building and maintaining be er rela onships between state and city. While the fi rst component is related to the poli cal dimension, the second refl ects the administra ve or management aspect of state-city rela onships.

42 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

in the quality of and access to local educa on and health services in Indonesia as well as infrastructure supplied by the ci es to the preference of local people. On the other hand, Strategic Asia (2013) point out the existence of specifi c block grants (Dana Alokasi Khusus) as a main grant transferred by the central government to ci es/regions account, that can guarantee the fulfi lment of all na onal priori es like ensuring good public roads, the decrease of diseases such as HIV and dengue fever, and minimizing risks of pollu on and fl oods and s ll maintaining the needs of ci es to con nuously develop their industries based on their willingness. A similar experience also occurred in China. A er the implementa on of an ini ally limited decentraliza on system, Qiao and Shah (2006) claim that the state objec ves to con nually promote growth balanced with environmental protec ons can s ll be maintained while simultaneously giving some ci es the fl exibility to determine which kinds of industries/services they prefer to provide and the types of infrastructure suitable for the needs of the area.

As there are the opportuni es for mutually benefi cial rela ons and the rela onship becomes dynamic, the tension in city-state rela onships o en increase since the state’s interests o en contrast with local needs. The needs of Jakarta and Bandung, two big ci es in Indonesia, to build more department stores, shopping malls and villas are o en contrary to na onal interests to protect catchment and green areas, reduce levels of pollu on and prevent fl ooding that poten ally spreads beyond a city’s boundaries (Firman, 2009).

In a decentralized system, the state’s interest is some mes sacrifi ced by the needs of ci es to generate more of their own revenues which possibly crea ng unexpected outcomes. Firman (2003) explains that the increasing demand of ci es to gain more addi onal revenues in addi on to central government’s grants through a new formula of income sharing, reduces the disparity between rich ci es and the state but it increases horizontal inequality between one city an another city/region. In China, although the needs of ci es are be er accommodated by the state the disparity of wealth amongst ci es/regions a er the implementa on of fi scal decentraliza on policy, doubled shown by high rate of its a Gini Coeffi cient (Xu, 2011).

Furthermore, the high level of fi scal decentraliza on by giving the ability to ci es to borrow from commercial

sectors without any precise calcula on in the ability of ci es to pay can lead to overall macro-economic instability. Devas and Delay (2008) and Peterson and Annez, (2007) take an experience in the Asian and European countries (e.g. India, the Philippines and Hungary) show the accumula on of ci es’ loan to commercial sectors to run more massive development should be recovered by the state which causes the defi cit and instability to overall na onal budget and macro-economic condi on. These create unexpected outcomes which disrupts the rela onship between the state’s interests and local needs. Homme (1995) and Vazquez and Vaillancourt (2013) no ced these fi scal problems as dangers and obstacles to decentraliza on and par cularly to na onal goals.

POLITICAL ACCOUNTABILITY AND COORDINATION AS OTHER IMPORTANT COMPONENTS ON STATE CITY RELATIONS

It has been shown that fi nance is the main component in state-city rela ons and its roles become clearer and more essen al in decentralized system. Nevertheless, this understanding is insuffi cient to recognise the complex nature of state-city rela onships. We acknowledge poli cal accountability and coordina on as other complementary components which comprise state-city rela onships as shown in Figure 2 below. While poli cal accountability represents poli cal aspects, coordina on refl ects the administra ve dimension of state-city rela onships (Bardhan and Mookherjee, 2006).

Figure 2.The Rela onship between Finance, Poli cal Accountability and Coordina on on State-Ci es Rela ons

Poli cal accountability is another essen al factor which establishes mutually benefi cial rela onships between state and city. The poli cal accountability covers a wide range of sub-components including the poli cal

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 59

Tabel 4.6. Working Sector Choice between Private Sector-Informal Sector

Dependent Variable :Working Sector Choice

(1)MLE_IV

(2)MLE_IV_F

(3)MLE_IV_M

Private sector to informal sector

Female -0.787***

(0.0623)

Age -0.0497*** -0.0846*** -0.0437***

(0.00180) (0.00516) (0.00207)

Married -0.315*** -0.0981 -0.343***

(0.0420) (0.197) (0.0427)

urban 0.400*** 0.560*** 0.403***

(0.0444) (0.128) (0.0479)

ReadWrite -0.0288 1.042*** -0.211*

(0.0787) (0.250) (0.0838)

Schoollevel 0.0404* 0.194*** 0.00237

(0.0176) (0.0503) (0.0188)

linc_pc 0.364*** 0.613*** 0.340***

(0.0235) (0.0766) (0.0243)

TChildLT6 -0.0396 0.0170 -0.0293

(0.0238) (0.0718) (0.0257)

TChild7_14 -0.00865 0.0782 -0.0208

(0.0195) (0.0641) (0.0202)

workerra o -0.803*** -1.465*** -0.704***

(0.109) (0.321) (0.112)

r_remi ance_pc_IV -0.210*** -0.653*** -0.149

(0.0556) (0.136) (0.0758)

_cons -0.334 -2.502*** -0.224

(0.174) (0.543) (0.189)

N 15215 2992 12223

pseudo R-sq 0.165 0.260 0.138

Standard errors in parentheses

* p<0.05 ** p<0.01 *** p<0.001

In general, rela ve remi ance per capita aff ects female working decision and it tend to shi female head away from not working, public, and private sector to informal sector. Such a convergent movement is very strong and persistent when the observa on is confi ned to female head of household. For men, however, there is no evidence that remi ance aff ects men’s working sector decision. The es mator for remi ances is not signifi cant in any cases. For men, the public sector rigidity is hold. Male head of household working sector preference is not aff ected by remi ance.

It is diffi cult to jus fy the reason why female head prefers to become informal worker rather than formal worker in Honduras. One possible reason is she prefer to nurture her children and willing to forgo her income in the formal sector because she has enough non-labor income from remi ance. The higher rela ve remi ance she received the more it mo vates her to le her formal jobs to become informal worker where she can spend more me with her family. However, I doubt this reasoning and strongly believe the be er explana on is that female wanted to work informally. Here is why. The remi ances she received helped her as a required capital to establish a self-employment. Because of some unknown reasons—one of which probably me fl exibility—she preferred informal job to any formal job. According to ACDI/VOCA and USAID fi eld report (2012), the typical of women self-employment in Honduras would be working on farm, growing fresh fruit and vegetables, and becoming an informal broker for such commodi es. My hypothesis is supported by the fact that remi ance also shi s women away from not working to become informal sector worker. If their main aim is to spend more me nurturing her children, why would they becoming an informal sector if they already are not working.

Confi ning the model to female head of household, this model displays a very high pseudo R-squared (0.26) even compares to overall test group where number of observa on is much higher (pseudo R-squared 0.16). When I confi ne the observa on to male head only, the pseudo R-squared decreased to 0.13. This implies that for female head of household, the explanatory variables succeeded to explain the majority her working sector decision. For men, however, there may exist few le out explanatory variables that could explain men’s working decision. Remi ances and possibly other non-labor incomes highly aff ect female working sector preference because women decision to work is purely because she has insuffi cient income. While men, he knew what sector he preferred and his decision to work is not solely to provide income, it might be something else which these sets of explanatory variables failed to expose. That explains the lower level of pseudo R-squared and insignifi cant remi ance per capita to working sector choice for male head of household.

60 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

Hours WorkedIn the second model, I tried to generate the impact of

remi ance to total hours worked by the head of household. Total hours worked is arguably the best variable to measure labor supply. Total hours is con nuous variable, not binary variable as in the fi rst model, and hence provide more robust interac on in the model. In anywhere in the world, the magnitude of remi ance impact is measured by how many people works less a er people receive remi ances. Weekly total hours worked is a numeric variable thus provide an easy and be er methodology than working sector choice. For this second model, I use Ordinary Least Squared model to check the rela onship between total weekly hours worked and rela ve remi ance per capita. For control variables, I used the same set with the previous model with small adjustment.

The second model contains more control variable than the fi rst model. First, I add Age-squared to show the in linear rela onship between hour worked and worker age. The standard human capital theory o en use this 2 type of measurement for age, even though in the mul colinearity test, both age and age-squared highly correlate with each

other. But I tolerated the mul colinearity problem because it is more important to show the nonlinear rela onship between age and total working hours. The other new control variable that I used is formal sector. I defi ne formal sector as head of household who works in either public or private sectors. For the sake of simplifi ca on and the similarity in nature of formal sector, I combine these two sectors.

Table 4.8 column (4), (5), and (6) is the output of the regression using IV for overall, female head, and male head test group. The objec ve variables (remi ance a er correc ng with endogeniety and omi ed variable) behave not too far from the OLS technique. Rela ve remi ance per capita shows signifi cantly reduces in the overall, female head, and male head test group. In overall test group, every 10 point of increase in r_remi ance_pc would reduce 14.7 hours from its mean 45 hours/week. While when the model confi ne to the female and male head of household separately, every 10 point of increase in r_remi ance_pc would reduce 20.5 hours and 12.2 hours, respec vely. All of the control variables do not deviate compares to the OLS es ma on without IV. All the expected sign in the OLS and

Table 4.8. OLS+IV Regression Total Hours Worked

Dependent Variable Hours Worked

(1)OLS

(2)OLS_Female

(3)OLS_Male

(4)IV

(5)IV_Female

(6)IV_Male

Female -8.081*** -7.778*** (0.393) (0.458)

Age 0.235** 0.234 0.225* 0.222* 0.205 0.220*(0.0891) (0.244) (0.0937) (0.0915) (0.250) (0.0963)

Age2 -0.00376*** -0.00438 -0.00357** -0.00369*** -0.00425 -0.00351**(0.00105) (0.00283) (0.00110) (0.00107) (0.00288) (0.00114)

Married -1.818*** -3.427** -1.665*** -1.720*** -3.287* -1.678***(0.319) (1.266) (0.319) (0.324) (1.407) (0.320)

urban 2.508*** 4.237*** 2.226*** 2.509*** 4.355*** 2.222***(0.342) (0.931) (0.361) (0.347) (0.958) (0.365)

ReadWrite -2.142*** -3.209 -1.905** -2.176*** -3.362* -1.894**(0.582) (1.678) (0.607) (0.587) (1.692) (0.612)

Schoollevel -0.990*** -1.890*** -0.801*** -1.009*** -1.935*** -0.803***(0.127) (0.353) (0.133) (0.129) (0.361) (0.135)

linc_pc 2.733*** 4.477*** 2.491*** 2.784*** 4.528*** 2.505***(0.156) (0.471) (0.161) (0.172) (0.531) (0.173)

TotPerhh 0.592*** 0.877*** 0.496*** 0.603*** 0.950*** 0.500***(0.0799) (0.207) (0.0850) (0.0864) (0.238) (0.0918)

formalsec 2.359*** 2.380** 2.242*** 2.224*** 2.079* 2.206***(0.311) (0.920) (0.323) (0.328) (1.033) (0.332)

workerra o -3.363*** 0.533 -4.768*** -3.905*** -0.473 -4.851***(0.714) (1.869) (0.759) (0.799) (2.301) (0.787)

r_remi ance_pc -1.613*** -2.432*** -1.189***(0.171) (0.384) (0.194)

r_remi ance_pc_IV -1.468*** -2.053* -1.215*(0.420) (0.867) (0.571)

_cons 28.28*** 7.741 30.59*** 28.63*** 8.844 30.64***(1.964) (5.640) (2.057) (1.974) (5.707) (2.063)

N 14786 2879 11907 14786 2879 11907R-sq 0.089 0.096 0.060 0.084 0.085 0.058adj. R-sq 0.088 0.092 0.059 0.083 0.081 0.057Standard errors in parentheses* p<0.05 ** p<0.01 *** p<0.001

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 41

UCLG (2009) and Gold (2010) dis nguished two types of central-local rela onship with fi nance as the main component of the rela onship. These are shown in Figure 1 below.

Figure 1.State-Ci es Rela ons : Financial Perspec vesCentralized / Delega on /De-concentra on Decentraliza on /Devolu on

Spending • State controls spending of ci es/municipali es (e.g. India, Thailand, The UK, Pakistan, Romania)

• No/less share of inter-governmental transfer, mostly comes from direct central grants (e.g. China, Indonesia before decentralized system, Belarus)

• Moderate/Complete discre on on spending (e.g. Indonesia post reform, Vietnam)

• Higher share on central government transfer (e.g. Indonesia post reform, Japan a er trinity reform, Denmark, Germany, Sweden)

Revenues • States control/limit local taxing or other local revenues (e.g. Bangladesh, Nepal, China,, Bulgaria)

• Ci es can set moderate tax or sharing taxes with states (e.g. Indonesia post reform, the Philippines, France, Belgium,)

• Ci es have discre on to fi nd other alterna ve sources of income such as borrowing money, bond, public-private partnership (e.g. Korea)

Source : Adopted and Modifi ed from UCLG (2009) and Gold (2010)

ROLE OF FINANCE ON STATE CITIES RELATIONSFinance is required by all levels of governments to

do rou ne ac vi es (e.g. salaries and opera onal costs) and development. For ci es, fi nance is required to provide good public health and educa on services. Davey and Devas (1992), focusing on Southeast Asian Countries, show the minimum amount of money required for urban areas to provide suffi cient local health and educa on services to their residents. Looking at Indonesia, Lewis (2003), also no ces that local governments devote more than half of their local budget to supply just the minimum standards on health and educa on.

Money is also required by ci es for infrastructure (e.g. roads/transporta on, sewage, clean water, electricity and gas) to ensure their ci es are more compe ve for investment and more a rac ve to interna onal visitors, leading to local economic development and public welfare improvement. Dubai spends billions of dollars building ci es equipped with modern and luxurious facili es and suffi cient infrastructure to en ce foreign investors and to establish it as central place of business and leisure (Bagaeen, 2007). Warsaw also devotes a higher percentage of its budget to building infrastructure in promo ng their industries (Ners, 2007). Furthermore, extra money can be spent by ci es on building infrastructure for hos ng interna onal events such as the G20 as well as interna onal sports games like the Olympic Games and the World Cup. Chalkley and Essex (1999) noted that, as the hosts of the Olympic Games, Tokyo and London disbursed huge amounts of fi nance to pay for building facili es and infrastructure.

On the other hand, the state also needs money for their interests that cover inter-city and inter-regional interests such reducing income inequality and minimizing nega ve externali es. Musgrave (1959) and Gruber (2009)

discovered that fairness on resource alloca on and income distribu on, economic stabiliza on and overcoming the nega ve external problems over city boundaries are some goals of state when u lizing public money. This is also similar to Kazepov’s explana on (2005) asser ng that maintaining macro-economic indicators and reducing inequality among diff erent regions are the main goals of most states in European Countries. Badcock (2002) and UN-Habitat (2008) also iden fy other state concerns including overcoming nega ve externali es such as pollu on, contagious disease, traffi c jams and fl ooding which extend beyond a city’s boundaries and are o en ignored by ci es.

In a centralized system, the state-city rela onship is more constant with lower tension since money will be addressed according to the state’s interests. Like in China’s centraliza on system (Zuo, 2009) na onal goals such as reducing disparity and overcoming nega ve externali es will be priori zed rather than local needs. By contrast, in a decentralized system, regardless of the unitary or federalist system, there is space for local needs alongside na onal interests with equal balance and vice versa as men oned by Wa s (2008) which makes the rela onship more dynamic and poten ally benefi cial for both state and city. Some tension can exist since the state’s interests are o en extremely diff erent to what a city needs.

If we look at Indonesia’s implementa on of fi scal decentraliza on a balance between local needs and na onal interests can be observed. On one hand, Kris ansen and Santoso (2006) and Robinson (2007) report an improvement

40 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

This paper will begin by defi ning the concept of the state-city rela onship through the concept of mul -level governance. In this part, two main types of state-city rela onships, centralized and decentralized, will be iden fi ed. In the second part, the roles of fi nance in state-city rela onships will be defi ned. It also explains the complexity and tensions in state-city rela onship in a decentralized system. In the last part, other components of state-ci es rela ons will be acknowledged including their roles and eff ects.

STATE CITY RELATIONSHIPSOne way to understand the rela ons between state and

city is through a mul -level government concept. This concept illustrates how the government func ons are distributed and rules are set to produce mutual benefi ts for diff erent levels of government (Hague and Harrop, 2010). Based on this concept, two main rela onships can be iden fi ed. The fi rst is centralized, when the state takes central control over local governments including quasi centralized system such as delega on and de-concentra on. Based on centralized systems in China, Wong (1991) argues that the state manages the overall development with the ci es ac ng only as agents of state that implement the development based on the state’s interests/plan. Rondinelli and Cheema (2007) observe that some Southeast Asian Countries also use delega on where some power is given to ci es/local administra on to run specifi c func ons (e.g. Special Economic Zones) and de-concentra on when the state transfers power to local personnel outside the state’s agencies without followed by a distribu on of authority and fi nancial power.

In contrast, while the size of state can be too big and the span of service is too large, the state has the op on to transfer some powers to ci es to running its func ons through decentraliza on or/and devolu on system. Smith (1985) no ces that the delimita on of territory and transfer of power from state to ci es as two main elements of decentraliza on. Gaubatz (2009) asserts that the state should iden fy its op mal capacity to provide eff ec ve public services to communi es or the power should be decentralized to fi t the op mal range of services off ered. Also as noted by Bunnel (et al. 2011) that bringing government closer to communi es with expecta on of be er quality public service is the main jus fi ca on for the decentraliza on

process in most Southeast Asian Countries. According to Lockwood’s (2006) regardless of their governmental system, whether unitary or federal, shows that decentralized systems support the preferences/interests of ci es/local government is taken into account. This is in line with the claims by Salim and Kombaitan (2009) based on Indonesia’s prac ces that decentraliza on policy gives more space to local authori es and actors to decide what they want to achieve with their city and how to implement it.

From a fi nancial perspec ve, a centralized system refers to the kind of fi scal arrangement when money is s ll administered under the central government’s accountability, while ci es depend on direct funding from the state which is not acknowledged as city’s own revenue. Bahl (2009) points out that during the implementa on of centralized fi scal policy in Indonesia and Pakistan, money was directly disbursed from the state’s account to a city’s residents without any consulta on with local administra ons about their needs. Unfortunately, this grant became the only revenue stream for ci es so it created a dependency where the ci es were reliant on the state.

In contrast, decentraliza on is the conceptualiza on of fi scal distribu on from the central to local administra on/accountability and guaranteeing autonomy for ci es to spend/disburse on their needs. ADB (2008) iden fi ed that in some decentralized European and Asian Countries, the central government transfer comprises more than fi y percent of local government revenues and is even higher in some countries (e.g. Indonesia seventy percent and Pakistan ninety percent) and is accompanied by the decision making powers to spend money based on their interests. Treisman (2007), based on experiences in Sweden, China and Thailand points out that fi scal decentraliza on is set when ci es has the capacity to generate their own revenues through local taxes (property and adver sement taxes) and charges, with direct access to fi nancing from private markets, selling bonds, and public-private partnerships without any scru ny or need for approval from the state. Fengler and Hofman (2009) also add that in devolved fi scal arrangement, more funding is addressed to ci es/municipali es’ account through intergovernmental transfer but the level of state transfers decreases when more autonomy is given to ci es to raise their own incomes through local taxes and other poten al sources of income.

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 61

OLS+IV are the same with the hypothesis. This means, the possible endogeniety and omi ed variable biased which might aff ect the beta es mator and deviate it from the true beta is not very severe or does not exist at all.

Applying Heckman Correc on Model However, regression using OLS+IV would s ll contain

poten al sample selec on bias. In the data descrip on (table 2.2) we know that there is some people in the dataset who is not working. In the OLS, the model simply uses all the observa on regardless the observa on contains working hour or not. Not working people would undershoot the mean value which would cause the biased es mator. Hence, we need to correct this sample selec on bias by predic ng the value how many hours would the no working people had they work using his/her own characteris cs. This model employ Heckman correc on model (Heckman, 1979) to solve the sample selec on problem.

From the table 4.9, a er solving for sample selec on using Heckman method, shows the rela onship between hours worked and remi ance per capita along with other

control variables. Rela ve remi ance per capita shows signifi cantly reduces hours worked in the overall and female test groups. In overall test group, every 10 point increase in r_remi ance_pc would reduce 11.6 hours/week from its mean which is 45 hours/week. For female head of household, every 10 point increase in r_remi ance_pc would reduce 19.9 hours/week. When the model is confi ned to male head of household, I found there is no signifi cant impact of remi ance per capita for men working hours. The other control variables in column (7), (8), and (9) behave very consistent and persistent with in column (4), (5), and (6) when I haven’t include Heckman correc on mechanism.

For the second equa on in the Heckman model, when the model predict the hours worked for not working people, I used total number of children and number of family who migrates. Total number of children is signifi cant for both children under 6 year old and children between 7 and 14 year old. However, the total number of family abroad is not a signifi cant predictor of working decision. Mills lambda coeffi cient in Table 4.10 represents the degree of severity of sample selec on problem. The signifi cant

Table 4.9. Heckman Regression Total Hours Worked (main equa on)

Dependent VariableHours Worked

(4)IV

(5)IV_Female

(6)IV_Male

(7)IVHeckman

(8)IVHeckm_F

(9)IVHeckm_M

Main equa on Female -7.778*** -6.340***

(0.458) (0.546)Age 0.222* 0.205 0.220* 0.110 0.150 0.136

(0.0915) (0.250) (0.0963) (0.0949) (0.264) (0.101)Age2 -0.00369*** -0.00425 -0.00351** -0.00209 -0.00352 -0.00227

(0.00107) (0.00288) (0.00114) (0.00113) (0.00310) (0.00123)Married -1.720*** -3.287* -1.678*** -1.855*** -3.284* -1.824***

(0.324) (1.407) (0.320) (0.327) (1.404) (0.326)urban 2.509*** 4.355*** 2.222*** 2.957*** 4.388*** 2.771***

(0.347) (0.958) (0.365) (0.361) (0.957) (0.415)ReadWrite -2.176*** -3.362* -1.894** -2.167*** -3.347* -1.729**

(0.587) (1.692) (0.612) (0.590) (1.689) (0.619)Schoollevel -1.009*** -1.935*** -0.803*** -0.973*** -1.912*** -0.753***

(0.129) (0.361) (0.135) (0.130) (0.362) (0.137)linc_pc 2.784*** 4.528*** 2.505*** 2.828*** 4.531*** 2.548***

(0.172) (0.531) (0.173) (0.174) (0.530) (0.175)TotPerhh 0.603*** 0.950*** 0.500*** 0.445*** 0.894*** 0.375***

(0.0864) (0.238) (0.0918) (0.0923) (0.253) (0.102)formalsec 2.224*** 2.079* 2.206*** 0.805 1.318 1.207*

(0.328) (1.033) (0.332) (0.440) (1.596) (0.487)workerra o -3.905*** -0.473 -4.851*** -6.700*** -1.776 -6.821***

(0.799) (2.301) (0.787) (0.987) (3.104) (1.055)r_remi ance_pc_IV -1.468*** -2.053* -1.215* -1.155** -1.989* -0.945

(0.420) (0.867) (0.571) (0.427) (0.872) (0.583)_cons 28.63*** 8.844 30.64*** 33.19*** 11.41 33.76***

(1.974) (5.707) (2.063) (2.195) (7.028) (2.352)N 14786 2879 11907 17872 4933 12939R-sq 0.084 0.085 0.058adj. R-sq 0.083 0.081 0.057 Standard errors in parentheses* p<0.05 ** p<0.01 *** p<0.001

62 | EDISI 02 • TAHUN XX • SEPTEMBER 2014

es mator for mills lambda for overall test group and male head test group implies that there was sample selec on biased on the model that have been taken care by the Heckman model. While the insignifi cant mills lambda ra o for female head test group means that there was much less severe sample selec on problem when we confi ne the observa on just for female had.

Table 4.10. Heckman Regression Total Hours Worked (choice equa on)

(7)IVHeckman

(8)IVHeckman_F

(9)IVHeckman_M

workFemale -0.924***

(0.0378)Age 0.0814*** 0.0890*** 0.0735***

(0.00867) (0.0139) (0.0113)Age2 -0.00108*** -0.00117*** -0.000980***

(0.0000995) (0.000158) (0.000131)Married 0.0942** -0.00666 0.148***

(0.0342) (0.0774) (0.0396)urban -0.324*** -0.0421 -0.519***

(0.0344) (0.0526) (0.0462)ReadWrite -0.0707 0.0656 -0.186*

(0.0581) (0.0958) (0.0765)Schoollevel -0.0439*** -0.0302 -0.0565***

(0.0130) (0.0222) (0.0164)linc_pc 0.00210 0.00911 0.0123

(0.0163) (0.0289) (0.0203)formalsec 1.313*** 1.946*** 1.027***

(0.0440) (0.0953) (0.0491)workerra o 2.855*** 3.389*** 2.411***

(0.0807) (0.134) (0.102)N_Immigrants -0.00754 0.0590 -0.0390

(0.0401) (0.0563) (0.0565)TChildLT6 0.244*** 0.199*** 0.275***

(0.0170) (0.0264) (0.0237)TChild7_14 0.215*** 0.257*** 0.190***

(0.0142) (0.0224) (0.0192)r_remi ance_pc_IV -0.203*** -0.100* -0.249***

(0.0377) (0.0430) (0.0670)_cons -1.506*** -3.323*** -0.998***

(0.201) (0.348) (0.256)mills lambda -5.809*** -1.380 -5.853** (1.179) (2.210) (2.060)

CONCLUSIONUsing the large 2006 LSMS data, this paper tried to

carefully examine the impact of remi ance to labor supply in Honduras. Overall, the fi nding shows that the impact of remi ance to labor supply is likely to vary depend on the head of household gender. Remi ance aff ects female head working sector preference. Female head of household

preference signifi cantly converges from formal sector and no working to informal sector. Even in the case for public sector workers, when receiving remi ance, female head of households swi away to informal workers. For male head of households, the result is diff erent. Remi ance insignifi cantly aff ects working sector preference. For male, public sector worker rigidity theory is prevailed.

A er sample selec on problem were solved, remi ance seems to signifi cantly aff ect total hours worked for only female head of household. For every 10 point of increase in the rela ve remi ance per capita, total weekly hours worked for female head is reduced by 19.9 hours from their mean. Again, similar with the fi rst model, male head of household total working hours is insignifi cantly aff ected by remi ances.

More detailed data on informal sector breakdown are needed to be er understand the exact mechanism of female head convergence shi ing. The fi nding in this research is robust and persistent, but also it contrasts with the common belief of public sector rigidity. Public sector rigidity theory claims that public sector workers tend not to move away to other sector. Public sector is known as high-paying job and provides good working benefi ts. Hence, it is unlikely that people forgo their job because it is uneasy to a ain. There might be a country-specifi c phenomenon which explains female head of households in Honduras shi away from public workers to informal workers in the presence of remi ances. Panel data would be an ideal case to carefully examine the impact on labor supply. First diff erence would be the best technique to prevail the true es mator of the objec ve variables. Other limita on of this paper is that the dataset that I used only contain head of household data. It does not taking into account the other family members.

Despite the limita ons, this study provides strong evidence to enrich the understanding about migra on and remi ance pa erns. Remi ance aff ects the labor supply especially for women migra on. However, no policy recommenda on can be made to date. Further research accommoda ng not only the head of household, but all member of household to generate be er beta es mator is need to thoroughly examine the remi ance eff ects.

EDISI 02 • TAHUN XX • SEPTEMBER 2014 | 39

Finance is The Essen al Thing in Establishing the Rela onship between State and City but It is not

Suffi cient : Learning from Asian and European State-City Rela onships

By Mohammad RoudoPhD Program in Interna onal Development

The University of Birmingham-The United Kingdom

Abstract

Mengambil contoh dari pengalaman beberapa negara di Asia dan Eropa, kami berpendapat bahwa keuangan merupakan faktor yang sangat pen ng di dalam hubungan antara pusat dan daerah atau kota dan negara,

walapun keuangan bukan merupakan satu-satunya faktor yang melandasi hubungan tersebut. Kami menyadari bahwa komponen akuntabilitas poli k dan koordinasi yang baik antar aktor menjadi faktor lain yang juga pen ng di dalam hubungan kota dan negara. Selanjutnya, menyadari bahwa keuangan merupakan dasar dari hubungan kota dan negara atau pusat dan daerah, kami juga yakin bahwa hubungan tersebut menjadi lebih dinamis di dalam sistem yang terdesentralisasi ke ka kota mendapatkan kewenangan yang besar dari negara untuk beberapa urusan. Hanya saja di dalam sistem yang terdesentralisasi ini, hubungan pusat dan daerah menjadi lebih unik dan spesial serta seringkali hubungan ini menjadi’memanas’dan berada dalam tensi yang nggi, karena munculnya perbedaan kepen ngan nasional dan lokal yang sangat besar.

Keywords: State-Ci es Rela onships, Mul -level Governance, Financial Decentraliza on

INTRODUCTION

The urban popula on in the world has increased signifi cantly and is now mostly concentrated in urban

centres of Europe and Asia. In 2000, more than half of the global popula on inhabited urban areas in Asia and Europe (Miller and Bunnel, 2012) with this tendency predicted to increase in future (Kotkin, 2005). The rise of globaliza on has exacerbated the challenges faced by ci es in Asia and Europe, simultaneously opening up opportuni es for coopera on and crea ng networks, but also increasing compe on to appear more a rac ve for ci zen and foreigners, as well as for investment and interna onal events (Hall, 2005; Sassen 2005). These complex challenges necessitate urban development and greater budget, but budgets are o en insuffi cient to fulfi l a city’s unlimited needs, so alterna ve revenue genera on schemes are required from city administra ons, or more precariously

demanding more money from state or higher levels of government. As a result, the rela onships between state and city are grounded on the need for fi nance.

In this paper, we argue that fi nance is the essen al, but not the solitary component in establishing the rela onship between state and city. We acknowledge the importance of poli cal accountability and coordina on as the other complementary components in the state-city rela onship. Recognising fi nance as the founda on of the rela onship, we also believe state-city rela onship is more dynamic in a decentralized system where ci es receive devolved powers from state but in this system there are unique and increased tensions in the rela onship since the na onal and local interests are o en diff erent.


Recommended