The Empirical Implication for Theoretical Modelling (EITM) Practices

This article is part of the final exam for the Empirical Political Theory class. The article divided into three parts. The First part is a definition of theory and theory building. The second part is analyzing the exemplary article which similar to my research proposal. The last part is developing my own EITM steps according to my research proposal

Part 1

  1. My own statement of the theory

My own definition of theory is a tentative conjecture of definition, assumption, and a simplification of a complex phenomenon which conducted by actor or institution that causally related that can be used to forecast events in the future.

My own statement of theory building

According to (Fiorina 1975) the benefits of theory building in a formal way. The process of building a statement of concepts and their interrelationships that shows how and/or why a phenomenon occurs. He mentioned the benefit of formal models as:

  • First, formulating a model forces precision in the terms of one’s argument. In everyday discourse, we use words with multiple or ambiguous meanings
  • A second level on which model building contributes to the clarity of thought is on the level of assumption
  • A third way in which formal models contribute to clarity of thought is closely related to the clarity and completeness of definitions and assumption. A formal argument is by far the easiest to check for logical validity (provided one understands whatever abstract reasoning which may be involved).
  • The precision and clarity of thought which these models require, and the depth of argument which they allow

According to (Kellstedt and Whitten 2009) there is no magical formula or cookbook for developing good theories about politics. But there are strategies for developing theories that will help you to develop good theories p.22. Furthermore, Kellstedt and Whitten mentioned that theory building is an art and scholars can develop a good theory if they have good strategies that consist many things such as variation in time series and cross-sectional, variation independent and independent variables that causally link together and have a testable hypothesis. According to (Neff 1998) grounded theory can work for formal theory building as well as for substantive theory building. A researcher can apply grounded theory to several case studies and use those studies as the empirical basis for a meta-analysis across cases p.131. Barbara Hanson (Hanson 2008) argued that the qualitative method should be combined with the quantitative method as an effort the project of exploratory, theory building or a part of triangulation p.104.

In order to bridge the gap between theory and policy Joseph S. Nye (Nye 2008) suggested theoretical trends have always strongly influenced by the outside world, that is why the changes from the outside world should drive the theory building, but often the swing in academic fashion are excessive and lack balance. According to the scholars that I have mentioned, I will describe my own statement of theory building as an effort or process to combined between a formal theory with empirical analysis and the interaction between formal theory and empirical analysis which then find a new additional concept/definition/factor that can develop the previous. Theory building in the illustration is like building a Lego, from the original form you can add or reduce some blocks without changing its fundamental form. If the fundamental form is changing, then the old theory will be replaced by the new theory.

My own statement of empirical modeling and analysis

According to (Granato and Scioli 2004) formal models, case studies, applied statistical models has their own strength and weaknesses. In an attempt to bridging the gap between empiricist and theorist Granato and Scioli created a concept called Empirical Implication of Theoretical Models (EITM). EITM aim is to encourage political scientists to build formal models that are connected to an empirical test. With the merge of formal and empirical analysis means that concepts must be clarified, causal linkages must be specified, theories must meet the challenge of these tests, and empirical work must link to a theory p.314. Based on this definition my own definition of empirical modeling and analysis is an attempt to represent the empirical world in a model that combine multiple methods or approaches which is expected to provide the most similar representation of the phenomena or real world that can explain and explore a better understanding about the phenomena or the real world.

The key components of empirical political research in general and why they are essential

The key components of empirical political research can refer to several scholars such as Fiorina, Kellstedt and Whitten, Diermeier and Krehbiel and Granato and Scioli




Why they are essential

1. (Fiorina 1975) -        Primitive, concept and assumptionThere are three classifications in formal models:

-        rational vs. behavioral/psychological model:

  • Decision-theoretic models, game theoretic models, spatial model

-        micro-macro

  • The actors are individuals the basic unit of analysis or are groups, classes, nations, etc. the basic analytic unit

-        Static and dynamic

The very basic component or the minimum requirement of theory, if a theory does not consist the minimum element it is not a good theory.
2. (Kellstedt and Whitten 2009) -        Independent variable and dependent variable at a theoretical level-        Operationalization of independent variable and dependent variable from the theoretical level to empirical level

-        Measuring the causality between the independent variable and dependent variable at the empirical level

The components can explain and measure the causal relation of theory.With the explanation from causal theory→ hypothesis→ empirical test→ evaluation of hypothesis→ evaluation of causal theory →scientific knowledge, it can make a good theory building.
3. (Diermeier and Krehbiel 2003) -        Preference-        Institution The components can provide a better understanding for the researcher to understand the contextThis is the basic equation to measure the outcome by multiple the preference and institution.
4. (Granato, Lo, and Wong 2010) Three steps in EITM:-        Identify a theoretical concept

-        Develop formal and statistical analogues

-        Unite theoretical and statistical analogues in testable theory

Connected the theoretical world with the empirical world in an attempt to develop a theory building and reduce the gap between theorist and empiricist.

Part 2

  1. Exemplary journal article

The article entitled “From decentralized to centralized irrigation management” written by Steven M. Smith (Smith 2018). Division of Economics and Business, Payne Institute for Earth Resources, Colorado School of Mines, United States. Journal of Economic Behavior and Organization 151 (2018) 62–87.

Critical assessment of the article according to Part 1




1. -        Primitive, concept and assumptionThere are three classifications in formal models:

-        rational vs. behavioral/psychological model:

  • Decision-theoretic models, game theoretic models, spatial model

-        micro-macro

  • The actors are individuals the basic unit of analysis or are groups, classes, nations, etc. the basic analytic unit

-        Static and dynamic

Concept: centralized, decentralized, public goods, irrigation system.Assumption: In theory, the decision should be driven by the expected net gains of internalizing decisions compared to the current transaction costs of decentralized management (Coase, 1937; Libecap, 1993).

-        Formal model: rational decision making (decision to change the irrigation system from the original small decentralized communal Spanish irrigation systems (acequias) to centralized quasi-public irrigation districts (ID) altered agricultural development and production, a drastic shift in resource governance structure, bilateral and multilateral contracts between ditches and can reallocate irrigation water through administrative decisions, reducing the transaction costs associated with market-type transactions. Why rational because with the acequias system the district can not collect the tax, issue debt. The acequias depend on savings and individual contributions. The decision to change from acequias to irrigation district is rational because, with ID, districts ID will serve larger water area, districts will have more access to water and share the infrastructure costs this decision can solve the public goods issue and gain revenue from ID system.

-        Actor: The districts in The State of New Mexico.

-        Unit of analysis: local government (micro level)

-        Years: 1910, 1920, 1930, 1940, 1950, 1959, 1969, 1978 (the years of irrigation management system changes)

-        Dynamic: why I choose dynamic even though there is only 1 actor because there are three preferences to manage the irrigation system from 1910 to 1978 which decided to organize by commercial, district or cooperative. The empirical finding evidence shows that there are significant changes decision from the district who had chosen decentralized system (cooperative and commercial) to centralized system (organized by district)

2. -        Independent variable and dependent variable at a theoretical level-        Operationalization of independent variable and dependent variable from the theoretical level to empirical level

-        Measuring the causality between the independent variable and dependent variable at the empirical level

-        The x-factor is the irrigation system-        X1 is the decentralized system (acequias) and S2 is the centralized system (irrigation districts or ID)

-        Y is public goods in water distribution

3. -        Preference-        Institution There are three preferences in the article:-        The irrigation management conducted by districts (centralized)

-        The irrigation management conducted commercial (private enterprises) (decentralized)

-        The irrigation management conducted by cooperation between districts and private enterprises (decentralized)

Institution: public goods issue, common pool resource issue.

4. Three steps in EITM:-        Identify a theoretical concept

-        Develop formal and statistical analogues

-        Unite theoretical and statistical analogues in testable theory

In the first step, the author explains the problem in the common pool resource issues. The more develop the irrigation system the water division become more problematic. At first, many districts select to organize the water are by them self (decentralized) but later because of the common pool resource issue the cost of water management is too high, limited water access with no profit. Later the district decides to collaborate with other districts and hand over the management to one district, they will share the infrastructure cost and share the profit.Step 1: relating governance structure (decentralized and centralized) with a decision to maintain irrigation system (acequias) or to change it to irrigation district (ID) system.

Step 2: developing the analogues relating the governance structure with irrigation management system the first analogues in the decentralized system

Second analogues are adding the centralized system, which adds the number of districts which select to cooperate in the irrigation district system.

Step 3: Unifying and evaluating the analogue

The third steps, the author apply the analogue with the empirical data and the result of empirical findings.

5. Empirical Findings -        Larger centralized IDs (irrigation districts) tended to form in counties where the common-pool losses were likely to be larger, primarily where more irrigators share a common source of water.-        IDs were particularly equipped to overcome free-riding issues to construct and maintain irrigation infrastructure through their quasi-government status, ability to tax, and ability to issue tax-exempt bonds. And to this end, IDs were successful in New Mexico, particularly in Dona Ana County

-        Centralized IDs primarily altered the governance structure of existing decentralized communal irrigation systems and did not seek to substantially increase irrigated acreage, there is no evidence of improved irrigation and related agricultural outcomes. The results support the fact that not only are acequias among the types of common-arrangements that can avoid the falling prey to the tragedy of the commons over long time frames, but also that centralized governmental control does not necessarily improve upon the commons-arrangements.

6. Conclusions -        Centralized organizations worked well to overcome the public good issues; financing and constructing large shared infrastructure needed to expand economic productivity.
7. Theory building  -        The centralized IDs yielded better outcomes addressing the public good issues in contrast to the CPR (common pool resource) issues.-        Perhaps it is simply the particulars of the institutional context explored: the federal government required an ID to contract with, meaning substantial expansion required, and was not a product of, centralized IDs. But perhaps any need for external finance would necessitate a centralized entity and the government is not alone in preferring a single contract to multiple bilateral contracts to finance large investments. In contrast, the CPR issue is primarily internal and thus the parties can generally address problems through cooperative measures without forming a new organization.

-        The decision to use centralized or decentralized system usually related with political factor and desire to make independent policy, but this article give evidence that economic factor has more influence in the decision to maintain decentralize system or change it to a centralized system. Which prove by the empirical finding that more and more districts select to switch to the centralized system. This argument strengthens the theory of Coase and Libecap used in this article that the decision should be driven by the expected net gains of internalizing decisions compared to the current transaction costs of decentralized management (Coase, 1937; Libecap, 1993).

Part 3

Research project according to EITM

My research proposal entitled

“Government structure in public goods delivery in Banten Province Indonesia”

The outline of the proposal will be divided into six parts, which is a simple paragraph version of part one of the key components on empirical political research. The first part is a brief introduction, a general condition of decentralization in Indonesia and a short context of Banten. The second part is a brief literature review to accommodate Fiorina components of theory, then the classification of formal models. The third part is the variables. The fourth part consists of preference and institution. The fifth part will elaborate the EITM steps and the last part is my expectation from my research since there are no conclusions, empirical findings and theoretical building.

1. A Brief Introduction

Decentralization in Indonesia started in 2001. Indonesia’s democratic progress was unpredicted when Soeharto regime in 1997-1998 came suddenly (Aspinal, 2018). The demand of democratic forces in Indonesia arises because people outside of Java believe that the power not distributed fairly (Rabasa and Chalk, 2001). Law Number 22 and 25 enacted in 1999 in Habibie administrative era. The law devolves power and revenue to the local government. Within this new system, provincial and municipalities split is possible. By this, it means that some potential areas may propose to separate themselves from former areas.

The law that regulates decentralization after the reformation in 1998 has changes for three-time, Law number 22 in the year 1999, then revised to law number 32 in the year 2004 and last the law number 23 in the year 2014. Consequently, the local government should continuously adapt to the law, because certain power is re-centralized by the central government like mining. So far, there are six authorities that centralized and cannot distributed to local government which are defense, security, religion, law, foreign affairs, fiscal and monetary. While there are around 34 authorities distributed to provinces, districts, and collaboration among central government and local government such as maritime, forestry, energy and so forth. The changes will affect public goods delivery in local districts. Hence the law will influence public funding.

The first province proliferated in Indonesia after the reformation was North Maluku. North Maluku separated from Maluku in 1999. The second followed by West Papua separated from Papua in 1999. Banten from West Java in 2000. Bangka Belitung from South Sumatera in 2001. Gorontalo from North Sulawesi 2001. Riau islands, from Riau in 2002. and West Sulawesi from South Sulawesi in 2002. The new local politics challenged the central government over the distribution of power and resources (Hamid, 2014). Decentralization was urged because of distrust among state apparatuses (Klinken and Berenchot, 2018). Province splitting is an effort to suppress separatism as shown by Aceh, Papua, Maluku and Kalimantan

Banten was one of the newly autonomous provinces, which was separated from former province West Java in 2000 just a year before decentralization in Indonesia was formally begun. During the 18 years of being an autonomous province, Banten has not shown any significant improvement in the deliverance of public goods. The idea of splitting from West Java emerged since 1953. Disparities and Sultanate history become the main factors Banten wants to separate from West Java. Chasan Sohib (Atut father’s, Banten first vice governor), even financing some long march from Serang municipality to Jakarta to succeeded Banten efforts to separate from West Java. After Atut arrested due to corruption cases, and the large scale of political dynasty of Atut’s family in Banten showed the expansion of Banten more to the division of powers to local elites.

Even after 18 years of decentralization in Banten, the poverty rate is still high, infrastructures such as roads, bridges, schools, hospitals, health centers are still very few, poorly managed and not feasible. Some schools and hospitals need to be repaired as they are not adequate to provide the services. Based on the evaluation of The Ministry of Home Affairs in 2011 for public services performances in the regional expansion, Banten is in second positions from below.

Table 1

Evaluation of Ministry of Home Affairs to Performances of Proliferated Province in 2011



North Maluku 55.88
Gorontalo 51.31
Bangka Belitung 49.64
West Sulawesi 46.73
Riau Islands 46.64
Banten 44.57
Papua 24.99

Banten which located in Java Island is not able to compete with northern Maluku which is far from the island of Java. While the Java Island is in the center of development and governance in Indonesia and the central government located. Banten is only able to compete from Papua which is infrastructure and accessibility and human resources are still under other provinces. In the year 2000 Banten, initially consists of 3 regencies and 1 municipality which are Serang regency, Tangerang regency, Pandeglang regency, and Cilegon municipality. Later, Banten split into 4 regencies and four municipalities, Serang regency, Tangerang regency, Pandeglang regency, and Cilegon municipality plus Serang municipality, Tangerang municipality, South Tangerang Municipality, and Lebak regency.

2. A Brief Literature Review and Formal Model

  1. Some concepts that will be used in this study are:


  • a centralized system, the government chooses a uniform level of public spending for each district (Oates, 1972) in (Wei, Yabin, and Shaobo 2018)
  • a centralized system forms policy in a legislature comprising of elected representatives from each district (Lockwood, 2002) in (Besley and Coate 2003)
  • Seabright (1996) in (Gradstein 2017) develops an incomplete contracts model in which centralization improves coordination but has costs in terms of diminished accountability, the latter being defined as the probability that the welfare of the region determines the re-election of the government.
  • This basic trade-off is also at the heart of Tommasi and Weinschelbaum (1999) in (Holzhacker, Wittek, and Woltjer 2015) who emphasize how principal-agent problems between citizens and governments can worsen under centralization.
  • Bardhan and Mookherjee (2000) in (Aspinall Edward 2003) emphasize how differences in political awareness and political competition can affect the likelihood of capture by special interests under centralized or decentralized decision making
  • a centralized system creates a conflict of interest between citizens in different districts (Besley and Coate, 2003)


  • Decentralization in Indonesia: The first was the public finance both revenues and expenditure and how budgeting affects equity. The second area was the political aspects of decentralization policy to understand the political-institutional structure of decentralization and its related consequences on the Indonesia political environment (Mokoginta, 2013)
  • Rondinelli and Cheema (2007) there are four factors that can affect the success and failure of decentralization which is: environmental conditions: inter-organizational relationship; available resources; and characteristic of implementing agencies
  • Since different region will have different preferences and pattern of demands for public services, locating the decision-making process in the hand of local governments will likely improve public services provisions (Smith, 2002)
  • Robbins (2000), decentralization as a managerial and organizational phenomenon and it refer to the degree to which decision making is concentrated to a single point in the organization
  • A method of internal organization of the nation-state. It does not take place against its interest but in cooperation in it. It is defined through the relations established between the central government and the regional and the local institutions (Savy et al, 2017).

2. Formal model: rational decision making

  1. Actor: The local government. In the centralized system, the actor is the central government while in the decentralized system the actor is regional government (province or district)
  2. Unit of analysis: local government (micro level)
  3. Years: 1978-2018 (centralized 1978-1997, decentralized 1998-2018)
  4. Dynamic: because there is a shift in the government structure, the actor, time series and the law that rules centralized, and decentralized authority has revised in 1999, 2004 and 2014. The alteration in the law will affect the decision making in term of the producing and applying the local policy.

3. The Variables

The dependent variable which known as the X factor in the research is government structure. The dependent variable consists of two X, X1 is a centralized system and X2 is a decentralized system. The independent variable which known as Y is public goods delivery as I mentioned before the law has changed three times, because of that reason I will select education and health. Education and health are the basic public goods and the authority belong to the local government, this authority does not re-centralized by the central government.

4. The Preferences and Institutions

There are three preferences in the article:

-        Health system and education system delivered in the centralization period conducted by the central government is better than the local government.

-        Health system and education system delivered in the decentralization period conducted by local government is better than the central government.

-        Health system and education system should be delivered with collaboration between central government and local government.

Institution: health system, education system, decentralization laws, culture, bureaucracy performance, political dynasty, corruption, patron-client.

5. The EITM Steps

Step 1: relating governance structure decision (decentralized and centralized) with public goods delivery

In step 1 I will collect analyze the theoretical framework. Which concept offers a more effective and efficient way of delivering the public goods for local government. And what factor theoretically influence the decision of central government whether to distribute or re-centralized the authorities to the local government.

Step 2: developing the analogues relating the governance structure with public goods delivery.

In the second step, I will compare the law in centralized and decentralized period, which authorities distribute to local government and which authorities are re-centralized. The law will affect the policy outcome because it was related with public fund. Consequently, the law will affect public goods delivery. Therefore, I will merge the theoretical concept in step 1 with the empirical data (the law and data, years, and other institution that might influence the public goods delivery) to build the analogue. For the time being because I have no ability (not yet) to develop my own analogue, I will be adopting Steven M. Smith analogues. But in my own research, I hope I can create my own analogue.

Second analogues by adding the centralized system

Step 3: Unifying and evaluating the analogue

In the last steps, I will create hypotheses and the apply the analogue with the empirical data. And test my hypotheses. The hypotheses probably (I’m not so sure) sound like this.

Hypothesis 1, centralized system delivered better public goods for local government

Hypothesis 2, decentralized system delivered better public goods for local government

Hypothesis 3, quasi-system delivered better public goods for local government

6. My Expectation for the Research

In this research I have two expectation, the first one is my expectation of empirical findings. And the second expectation regards to theoretical building. The empirical finding, I hope for four significant findings in:

  • Comparing the public goods delivery which one better in a centralized or decentralized system
  • Comparing among 8 districts (4 municipalities and 4 regencies) which districts can deliver more public goods in the decentralized system because it has big disparities between the area in South Banten and North Banten. The North Banten were more developed than The South Banten and The South Banten considered as the poorest area in Banten, one district in North Banten included in top 10 poorest districts in Indonesia
  • I can find the factors that influence public goods delivery
  • I can predict which government structure is preferred for the effectiveness and efficiency in public goods delivery by the local government.

Regarding the theoretical building, I wish I can contribute for a better understanding of centralization, decentralization and public goods. The government does not have to always decide on one structure whether centralized or decentralized but other alternatives can be collaboration between the two structure.


Aspinall, E. 2018. Chapter Title: Democratization: travails and achievements. Book Title: Routledge Contemporary Book of Indonesia. Page 83-94. New York: Routledge

Aspinall Edward, Fealy Greg. 2003. Local Power and Politics in Indonesia: Decentralisation and Democratisation. Pasir Panjang Singapore: Institute of Southeast Asian Studies.

Besley, Timothy, and Stephen Coate. 2003. “Centralized versus Decentralized Provision of Local Public Goods : A Political Economy Approach.” 87: 2611–37.

Diermeier, Daniel, and Keith Krehbiel. 2003. “Institutionalism as a Methodology.” Journal of Theoretical Politics 15(2): 123–44.

Fiorina, Morris P. 1975. “Formal Models in Political Science.” Americal Journal of Political Science 19(1): 133–59.

G. Shabbir Cheema and Dennis A. Rondinelli (eds), (2007) Decentralizing Governance: Emerging Concepts and Practices. Washington, DC: Brookings Institution Press.

Gradstein, Mark. 2017. “Government Decentralization as a Commitment.” Journal of Comparative Economics 45(1): 110–18.

Granato, Jim and, and Frank Scioli. 2004. “Puzzles, Proverbs, and Omega Matrices: The Scientific and Social Significance of Empirical Implications of Theoretical Models (EITM).” Perspectives on Politics 2(2): 313–23.

Granato, Jim, Melody Lo, and MC Sunny Wong. 2010. “A Framework for Unifying Formal and Empirical Analysis.” American Journal of Political Science 54(3): 783–97.

Hamid, Abdul. 2014. A Family Matter: Political Corruption in Banten Indonesia. Asian Politics & Policy—Volume 6, Number 4—Pages 577–593

Hanson, Barbara. 2008. “Wither Qualitative / Quantitative ?: Grounds for Methodological Convergence.” Quality&Quantity 42 (1): 97–111.

Holzhacker, Ronald L., Rafael Wittek, and Johan Woltjer. 2015. Decentralization and Governance in Indonesia Decentralization and Governance for Sustainable Society in Indonesia.

Kellstedt, Paul M, and Guy D Whitten. 2009. The Fundamentals of Political Science Research. New York: Cambridge University Press.

Klinken, Van Gerry and Ward Berenschot. 2018. Chapter Title: Everyday Citizenship in Democratizing Indonesia. Book Title: Routledge Contemporary Book of Indonesia. Page 151-162. New York: Routledge.

Neff, Joyce Magnotto. 1998. “Grounded Theory A Critical Research Methodology.” In Under Construction, ed. Chris M. Anson Christine Farris. Utah, The United States: University Press of Colorado, Utah State University Press, 124–35.

Nye, Joseph S. 2008. “Bridging the Gap between Theory and Policy.” 29(4): 593–603.

Mokoginta, Ivantia S. 2012. Leviathan Government Behaviour: The Case of Local Government in Indonesia. Dissertation for Flinders Institute for Public Policy and Management. Australia

Rabasa, Angel and Peter Chalk. 2001. Book Chapter: Reinventing Indonesia: The Challenge of Decentralization. Book Title: Indonesia’s Transformation and the Stability of Southeast Asia. RAND Cooperation

Savy, Robert, Hélène Pauliat, and Michel Senimon. 2017. Edited by Jose Manuel Ruano and Marius Profiroiu. The Process of Decentralisation in Europe in The Palgrave Handbook of Decentralisation in Europe. Palgrave Macmillan.

Smith, Benjamin. 2008. The Origins of Regional Autonomy in Indonesia: Expert and the Marketing of Political Interest. Journal of East Asian Studies 8. May-Aug 2008; 8, 2; Research Library 211-234

Smith, Steven M. 2018. “Journal of Economic Behavior and Organization From Decentralized to Centralized Irrigation Management R.” Journal of Economic Behavior and Organization 151: 62–87.

Stephen, P. Robbins, (2000). “Managing Today” Prentice Hall, Upper
Saddle River, New Jersey.

Wei, Que, Zhang Yabin, and Liu Shaobo. 2018. “The Spatial Spillover Effect of Fiscal Decentralization on Local Public Provision : Mathematical Application and Empirical.” 331: 416–29.

Category: Thoughts

Leave a Reply

Your email address will not be published. Required fields are marked *



You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>