Behavioural models of electoral management: A world survey

Democratisation has continued in the world, but many newly emerged democracies are still struggling with a lack of electoral integrity. Intending to dealing with the lacking, traditional studies, have focused on Electoral Management Bodies (EMBs), such as Bangladesh Election Commission, while the election time government or, in other words, electoral government as a unique political institution has so far been unresearched. This paper sheds new light on this tradition by introducing the concept of electoral government and explaining how an electoral government shapes the electoral management and inﬂuence the electoral outcome. It commissions and uses a world survey and comes up with a set of behavioural models of electoral management.


Introduction
An election, which determines that the citizen indirectly runs the state affairs, is rooted in the citizen ownership of the state (see Azad, 2020).A classic example of such ownership practice is found in the ancient Greek city-state of Athens (see Blackwell, 2003;Jones, 1958;Landauer, 2014).In ancient Athens, citizens generally used to get whatever the majority of them wanted from the state as they ran state affairs on their own.In the large-scale present-day democracies, citizens themselves cannot run state affairs.Rather, they get the state governed by their representatives in a system termed representative democracy.Their representation is determined in an election and all the countries with a democratic constitution these days have a legalinstitutional framework for an election that shapes the electoral management.The electoral management varies by country.Catt et al. (2014) in the International IDEA's Electoral Management Design categorise the electoral management worldwide into three models2 -Independent Model, Governmental Model and Mixed Model.The categorisation is based on whether or how much an EMB is independent of the government.However, in a representative democracy, it is not certain that citizens get whatever they want from the state because representative individuals or political parties have emerged as an intermediary factor that can interfere in what citizens get.This uncertainty deepens further if the representative election procedure is questioned.
In the global arena, democratisation has continued, but elections in many newly emerged democracies are still facing questions.In advanced democracies 3 like Australia, Canada, the UK, and the USA, elections 4 are held under an election time government or, in other words, electoral government led by the current governing party or alliance, and the elections are usually acceptable.Contrary to this practice, results of elections held under a governing party or alliance administration in struggling democracies 5 are not acceptable in general.This kind of unacceptability points to an allegation that government mechanisms in struggling democracies help governing parties win elections (Akram andShadhan, 2007, 2006;Birch, 2007;Carothers, 2002;Collier and Pedro, 2012;Fireman, 1994;Karp, Nai and Norris, 2018;Lehoucq and Ivan, 2002;Lehoucq and Atenea, 2007;Long, 2012;López-Pintor, 2011;Norris, 2015;Norris, Frank and Coma, 2015;Norris, 2019;Zavadskaya, Grömping and Coma, 2017;Schedler, 2012).The governing party or alliance as the leader of the electoral government during elections uses government mechanisms and state organisations, including the Electoral Management Body (EMB) and personnel as tools to conduct fraudulent elections.The party or alliance in power politicises, among others, the EMB, gets a fake voter roll prepared, creates intimidation, spends money on electioneering beyond permitted limits, and get manipulated election results announced in the state-owned media.Noticing such incidences, a few struggling democracies have opted for a non-party election time government as a temporary fix, such as Caretaker 3 This paper considers a country an advanced democracy if elections of government in the country show a trend of acceptability.Examples of advanced democracies are Australia, Canada, the UK and the USA.'Fully Free' countries in the Freedom House database of Political Rights (Freedom House database, ND) include advanced democracies. 4An election in this paper means parliamentary or presidential election that determines who will run the state affairs of the country. 5If a country after embracing a democratic transition suffers from an electoral integrity crisis for a considerable period of time, the country is considered struggling democracy in this paper.Such democracies are usually seen struggling to consolidate and even to survive in Asia, Africa and Latin America (see e.g.Haynes, 2001, Freedom House database, ND;Polity, ND).Examples of these countries are Bangladesh, Kenya, Mexico and Pakistan.'Partly Free' countries in the Freedom House database of Political Rights (Freedom House database, ND) include such democracies.
Government in 1991 in Bangladesh, to have general elections held in free, fair and acceptable manners (Azad and Crothers, 2012).Besides, there has been a practice of a type of elctoral management aided by the United Nations (UN) in some crisis-torn countries.
Therefore, a legal-institutional framework seems not to be enough for an acceptable election.
The same way, the Catt et al. (2014) categorisation seems to be futile in practice.Many democracies with a non-independent EMB (Governmental Model) like the United Kingdom and the United States and less independent EMB (Mixed Model) like Japan produce acceptable elections while many others with an independent EMB (Independent Model) like Bangladesh produce unacceptable elections.Besides, with the same EMB models, different democracies generate different levels of election results ranging from unacceptable to acceptable.For example, Australia and Bangladesh both have the Independent Model of EMB, but electoral acceptability in Australia has been an exemplary while most of the elections held in Bangladesh suffered from a lack of acceptability (for an explanation, see Przeworski et al., 2000, Catt et al., 2014; Freedom House database, ND; Polity, ND).These facts suggest that the legal-institutional framework of a model of electoral managements may have different types of behaviour which are the key to an electoral outcome.Studies of such behaviours are abundant (e.g Buckley and Reidy, 2018;Bowler, Brunell, Donovan and Gronke;2015;Donovan and Karp, 2017;Garnett and Zavadskaya, 2017;Garnett, 2017;Groemping, 2018;Gauja, 2015;James, 2020;Norris, 2017) that use different samples and methods but do not use the concept of electoral government, which is introduced in this paper.Hence, it is still crucial to know about the behavioural pattern of different categories of electoral government and behavioural models of electoral management.
The remainder of this paper is organised as follows: First, what is the theory behind the concept of electoral government and what is the measurement theory for the performance of such government are described in the Theory section.Method section says how the study is carried out and the way the paper is organised.Then, the Findings section reveals the impacts of different categories of electoral government on elections and their relevance.The paper, in Recommendation, introduces a set of Behavioural Models of Electoral Management as an alternative to the International IDEA Models and, then, concludes with some claims of the paper's contribution, further evaluation of findings and remarks about future research.

Theory: The electoral government
The concept of electoral government is based on the theory of citizen ownership of the state (see Azad, 2020).According to the theory, the world trend of election status may be divided into three categories: acceptable election, unacceptable election, and no election.As a result of an acceptable election, a representative government is set up.Azad (2020) models the authority of a representative government to a power of attorney, where election promises of the winning party, alliance or individuals become the only written part of the power of attorney and winning party, alliance or individuals becomes the agent of the citizen, the owner of the state.Then, how much the ownership exists is calculated based on the implementation status of the election promises made by the governing party, alliance or individuals during their election.On the other hand, citizen ownership of the state is occupied by the governing party, alliance or individuals where there is no acceptable election or no election at all.These two latter trends are sometimes interrupted whenever an acceptable election takes place.However, if an election is held -no matter where and when the election is held and whether the election is acceptable or unacceptable -the existence of an electoral government is certain.This paper considers the electoral government as a government of an independent state that runs usual operations of the state affairs during the period of an election that determines who will govern the state affairs for the next term.Based on the theory of citizen ownership of the state, this paper considers that a compliant agent cooperates with the conduct of election by refraining themselves from interfering in the election.Soon after the election, this type of agent hands over the power to the newly elected party, alliance or individuals or, in other words, the newly appointed agent and then exits immediately.Even if the same party, alliance or individuals are elected again, still a formal changeover takes place.Conversely, a non-compliant agent interferes with the conduct of election and clings to the state power unlawfully and, thus, they occupy the citizen ownership of the state.Therefore, an incumbent electoral government could be either Outgoer or Occupier by nature.However, an Occupier incumbent can become Outgoer by cooperating with the conduct of an acceptable election.Apart from the incumbent electoral government, some other types of electoral government have also been seen.When there are no acceptable elections or no elections at all and parties get involved in confrontations, then the power of attorney is withdrawn.Under such circumstances, the citizen of a by-and-large national consensus takes over and conduct elections, which are generally acceptable.Furthermore, in a worst case scenario in a country where a confrontation goes out of intra-societal control and no power of attorney exists -the state runs on an anarchic basis, then outside societies that is the United Nations (UN) extend their cooperation in regard to conducting an election.This paper considers these are the possible backgrounds of different categories of electoral government.Accordingly, this paper classifies the electoral government into three primary categories: Incumbent Electoral Government (Incumbent EG), National Electoral Government (National EG), and International-aided Electoral Government (International-aided EG).The incumbent government leads the Incumbent EG.It is of two further categories: Outgoer and Occupier.By nature, the Outgoer does not interfere in the election.The Outgoer tends to produce acceptable elections and exits as soon as the newly elected government is installed.It is generally observed in advanced democracies.The Occupier tends to produce unacceptable elections and, based on those elections, occupies the citizen ownership of the state by clinging to the state power.Occupiers are generally seen in struggling democracies.The National EG tends to produce acceptable elections.Non-party citizenry leads the National EG.They generally come from the civil society and have, by and large, nationwide acceptability.So far, the National EG operated in Bangladesh (see Ahmed, 2011;Azad and Crothers, 2012) and Pakistan (see Malik, 2013;Sikander, 2018).Electoral outcomes of the International-aided EG tend to be situational.This is because the behaviours of different local and international parties involved in the electoral government vary by situation.
Thus, each category of electoral governments has an impact -strong or weak -on the electoral outcome.This paper considers that those impacts are generated from two types of behaviour of the electoral government -complying behaviour and intruding behaviour.A complying behaviour lets the EMB work independently while an intruding behaviour controls the EMB.
However, there might be some behaviour of EMBs which are independent of the electoral government.In such cases, this paper considers that those are some stray incidents which are not good enough to alter an election result at the national level.Therefore, this paper does not cover EMB behaviours and observes that the electoral government has an overall impact that generally defines the electoral outcome, the empirical status of EMBs as well as the whole electoral management.
A researcher is able to measure the performance of an electoral government.Traditionally proven, the best way to measure how a student performs is their exam results.Here, the election could be considered an exam for the electoral government.So, whether an election is acceptable or not is the obvious measurement of the status of an electoral government.A minimalist dichotomous democracy index, such as Prezworski et al. 2000, could be used to measure an election where democracy can mean acceptable election and non-democracy the other.

Method
This paper picks the International IDEA's handbook entitled "Electoral Management Design" by Catt et al., 2014 as a case that develops three models of electoral management.These models are based on the legal-institutional status of EMBs.Against this backdrop, the objective of this paper is to develop real world election management models, using empirical data.To achieve the objective, this paper hypothesises that there is a number of electoral governments whose behaviours shape the electoral management their own way and generally determine the electoral outcome -acceptable or unacceptable.
To explain how the electoral government shape the electoral management, a qualitative evaluation of relevant data (Appendix 1) is carried out.This paper considers that electoral management is primarily composed of two institutions -the electoral government and the EMB.Among these two, the electoral government is able to influence the EMB.The influences ranging from no influence to neutral influence are considered complying or positive, while all other influences are considered intruding or negative.If an election is found acceptable, the EMB is meant to be influenced positively by the electoral government.Conversely, if an election is found unacceptable, the EMB is meant to be influenced negatively by the electoral government.
Thus, in both cases, the electoral government by influencing the EMB shapes the electoral management.
How the electoral government determines an electoral outcome is tested using logistic regression.In the logistic regression, electoral government and International IDEA's electoral management variables are used as fixed factors, year as a covariate when election (acceptable and unacceptable) is used as dependent variable.Since elections were held in different years in different countries, it is an unbalanced panel data.Therefore, the unbalanced panel data are made a balanced one by filling the gaps in the years from 1990 to 2018, using 'NA' as missing value.
Besides, robust standard errors are used to obtain unbiased standard errors of OLS coefficients under heteroscedasticity as both the independent variables are categorical and the independent is binary where the structure of variations is unknown.

Variables and data
Sample -the electoral world: Out of 217 the countries and territories surveyed and divided into three categories of electoral management bodies by International IDEA (Catt et al, 2014), a total of 131 countries are chosen for this paper (see Appendix 1).These countries have to have at least one acceptable election held any time from the year 1990 through 2018.All electoral governments that worked during those elections are covered.The 1990-2018 time range is chosen as a series of democratic transition started taking place in the late 1980s, and 2018 is considered one of the latest years when this paper is prepared.
Where citizens cast their direct vote to elect the country's president, the presidential election and where there is no provision for their direct vote to elect a president, the parliamentary election, are chosen for this paper.Data about the type of election are collected mostly from the Database of Political Institutions (Cruz, Keefer and Scartascini, 2018), IPU Parline (ND) and BBC Country Profiles (ND).About the latest elections such as the ones held in 2018, data are collected by conducting web searches, using keywords like 'Tuvalu elections' or 'Tuvalu election 2018.' ELECTION/ Election: 0 = Unacceptable, 1 = Acceptable The word 'election' means either presidential or parliamentary election in this paper.For coding the acceptability and unacceptability of elections data, this paper uses Przeworski et al.'s (2000) theory of democracy and non-democracy.The elections which meet the criteria Przeworski et al. (2000) use for democracy are considered acceptable in this paper.On the other hand, the elections that do not meet the criteria are considered unacceptable.However, this paper does not use the Przeworski et al. (2000) criteria for democracy that incumbents do not hold their offices continuously for more than two terms.This paper uses Przeworski et al. (2000) data extended by Cheibub et al. (2010) and then by Bjørnskov and Rode (2018).
ELECTORAL_GOVT/ Electoral government: 0 = National EG, 1 = International EG, 2 = Incumbent EG As per the discussion in the Theory section, the electoral government is of three categories: Incumbent EG, National EG, and International-aided EG.Incumbent EG is of two types: Outgoer and Occupier.These two types are used unseparated in the regression model while separated in the Behavioural Models of Electoral Management.The variable is coded 1 only when the UN electoral assistance to the electoral government takes place along with the presence of a peace agreement or a related UN mission/ operation in the country.If an election is observed, organised, conducted, supervised or the election process is verified by the UN, then it is considered that the UN electoral assistance takes place.The data about International-aided EG are collected from UN Peacekeeping (ND) and UN Reports (ND).
EMB/ International IDEA's Electoral management models: 0 = Government, 1 = Mixed, 2 = Independent International IDEA's three models of electoral management are actually based on EMBs, which could be called EMB Models.They are Government Model, Mixed Model and Independent Model.According to the world survey of EMBs conducted by Catt et al. (2014, pp. 374-395), the EMB models are used in this paper for the whole 1990-2018 period of study.
ELECTION_YEAR/ Election year: An election year is the year when an election took place.The election year data are collected mostly from the Database of Political Institutions (Cruz, Keefer and Scartascini, 2018), IPU Parline (ND) and BBC Country Profiles (ND).About the latest elections, such as the ones held in 2018, data are collected by conducting web searches, using keywords like 'Singapore elections' or 'Singapore election 2018.' < .001N = 866 Note: The dependent variable in this analysis is election coded 0 as unacceptable and 1 as acceptable.

Findings
The findings of the statistical analysis are interesting.The electoral government in the logistic regression model (Table 1 and Figure 1) is predicted to have a strong influence on election.The regression coefficient for the incumbent electoral government (level 2) is around -17.084.It suggests that the incumbent electoral government will produce 17.084 points less acceptable elections than the national electoral government's (the reference level).At the same time, the international-aided electoral government (level 1) will attain -17.008 points acceptable election compared to that the national electoral government does.On the other hand, in the same model, the regression coefficient for the government model of EMBs (level 2) is 17.175.It reveals that the government model of EMBs will generate 17.175 points more acceptable elections than the so-called independent model of EMBs produces.Besides, the parliamentary election (FROM_ELECTION 1) is predicted to account for a 0.816 point more acceptable the elections than the presidential election does.Table 2).The incumbent electoral government accounts for 74 unacceptable elections, which is found the biggest contributor to the total unacceptable elections in the sample.It leads to the lowest regression coefficient for the incumbent electoral government in the model.International IDEA's Models of Electoral Management apparently ignores the practicality of electoral regimes.These models' practical-impractical anomaly has not been an issue so far in advanced democracies as electoral governments in those democracies provide a complying behaviour with a view to conducting acceptable elections and keeps EMBs independent in practice no matter the EMB model is Government, Mixed or Independent in theory.However, the electoral government in struggling democracies has had a behaviour that occupies the citizen ownership of the state and kept EMBs under their control in practice with a view to producing election results in favour of them or the ones they like most no matter the EMB model is Government, Mixed or Independent in theory.That is why International IDEAS's 'independent' EMB produces acceptable elections in Australia but not in Bangladesh, for example.

Recommendation: Behavioural Models of Electoral Management
In accordance with the findings above, this paper recommends a set of new models of electoral management, called Behavioural Models of Electoral Management (see Figure 2

Conclusion
Democratisation has continued in the world, but many newly emerged democracies are still struggling with a lack of electoral integrity.Intending to dealing with the lacking, traditional studies, have focused on EMBs, such as Bangladesh Election Commission, while the electoral government as a unique political institution has so far been unresearched.This paper is the first to define, test and introduce the concept of electoral government, which is based on the power of attorney model of the representative government (see Azad, 2020).This paper explains how an electoral government is associated with an election.It reveals that there are three categories of electoral government and the behaviour of each category shapes their respective electoral management their own way.The paper argues that the behaviour of electoral government determines the behaviour of the EMB, and then the EMB behaviour determines the acceptability or unacceptability of an election.There might be some other reasons, such as corruption in the EMB itself, but the electoral government has an impact that generally defines the electoral outcome, the empirical status of EMBs as well as the whole electoral management.Based on the behavioural pattern of electoral governments, this paper develops a set of Behavioural Models of Electoral Management (see Figure 2), using a world survey commissioned in this paper (see Sources: This appendix is developed based on the data collected from different sources.Out of 217 the countries and territories surveyed and divided into three categories of electoral management bodies by International IDEA (Catt et al, 2014), 131 countries are chosen for this study.The election year data are collected mostly from Cruz, Keefer and Scartascini, 2018), IPU Parline (ND) and BBC Country Profiles (ND).About the latest elections such as the ones held in 2018, data are collected by conducting web searches, using keywords like 'Singapore elections' or 'Singapore election 2018.' Based on the definition described in the 'Theory' and 'Variables and Data' sections of this paper, the data about Outgoer and Occupier electoral government and whether an election is acceptable or unacceptable are collected from Przeworski et al. (2000) data exetended by Cheibub et al. (2010) and then by Bjørnskov and Rode (2018).National electoral government data are collected from Ahmed, 2011;Azad and Crothers, 2012;Malik, 2013;Sikander, 2018.The data about International-aided EG are collected from UN Peacekeeping (ND) and Previous Reports of the UN Secretary-General (UN Reports, ND).The EMB data are according to the world survey of EMBs conducted by Catt et al. (2014, pp. 374-395).

Figure 1 .
Figure 1.Estimates plot: Influences of electoral governments on election ), as an alternative to the International IDEA models.The Behavioural Models developed based on the Appendix 1 data are Complier Model, Model, Citizen Model and International-aided Model.Each of all four models has two components -Electoral government and the EMB, where the electoral government is identified as the key component of the electoral management.In the Complier Model, the electoral government lets the other component, the EMB, work independently that produces an acceptable election.This model is generally observed in the advanced democracies.In the Intruder Model, the electoral government controls the other component, the EMB, and gets the electoral work done by the EMB as the governing party or alliance wishes.This process produces an unacceptable election, which is generally observed in non-Western countries.In the Citizen Model, the electoral government (CEG) lets the other component, the EMB, work independently that produces an acceptable election.This model has been observed in a few non-Western elections, such as the one held in 2018 in Pakistan.In the International-aided Model, the electoral government is not always in a position to let the other component, the EMB, work independently.As a result, there is a chance of both acceptable and unacceptable elections, such as the ones in East Timor and Central African Republic (see Appendix 1).

Figure
Figure 2. Characteristics of Behavioural Models of Electoral Management

Table 1 .
Logistic regression analysis of elections from 1990 to 2018 worldwide