While much research has been devoted to studying election fraud, little attention is given to what exactly causes the non-transparent practices (including those other than election fraud) used by election commissions in their inner workings. This study uses Michael Lipsky's theory of street-level bureaucracy to answer the question of why the members of Saint Petersburg election commissions favor non-transparent practices in tallying election results without displaying any apparent fraudulent intent. Using qualitative and quantitative research methods, I analyze the performance of election commissions in 2013-2018 through question-based surveys and semi-structured interviews conducted with election observers and members of election commissions as well as secondary data. The results show that the more exacting observers are able to considerably influence the performance of election commissions. What is more, several such observers have greater chances of influencing the commission practices. Greater numbers also help them provide moral support to each other and make their own performance more coordinate. Despite the lack of these advantages, even a small number of exacting observers as well as several members of election commission are able to convince the rest of the commission to follow the procedure and, through conflict, "educate" the commission about regulatory practices. Temporary restrictions also have an impact, since there exists a threshold beyond which the members of the commission deem committing fraud justifiable.
The quality performance from precinct election commissions (PEC) may prove quite significant for political results, as can be seen from the 2011-2012 protests motivated by large-scale election fraud at the 2011 Russian legislative election, which directly involved PEC members. The observable activities of the PEC seem to be the motive behind the collective protest effort [6: 167-181]. The repercussions of election fraud in the former Soviet Union are typically far more serious as they hobble the development and transformation of a political system .
The fact of failing to observe the procedures alone, however, is not necessarily a consequence of somebody's intention to alter election results. Using the methods developed by Dmitry Kobak  to evaluate the extent of fraud committed at the 2016 and 2018 elections in Saint Petersburg  has revealed that falsifications were fare more numerous in 2018 (Figure 1) as compared to 2016 (Figure 2). Such is implied in the abnormal shift in the turnout and the resulting position of the front-runner. At the same time, the accounts of election observers in Saint Petersburg suggest that there were no significant changes in the number of violations committed by the PEC (Tables 1, 2) in both years [9; 19]. This indicates that violations in PEC procedure are there even when there is no fraudulent intent.
Figure 1. The turnout-based voting pattern at the polling stations during the 2016 Legislative Assembly of Saint Petersburg election under the proportional representation system. Based on the data provided by the State Automated System (SAS) "Vybory" (Rus. for "elections") (http://www.st-petersburg.vybory.izbirkom.ru/region/st-petersburg).
Figure 2. The turnout-based voting pattern at Saint Petersburg polling stations during the 2018 presidential election. Based on data provided by SAS "Vybory" (http://www.st-petersburg.vybory.izbirkom.ru/region/st-petersburg).
|Rewriting electoral commission statements/ballot-box stuffing||4 (2%)||7 (4%)|
|Outside the voting premises||3 (2%)||10 (6%)|
|Documentation||29 (16%)||32 (18%)|
|Election observer rights||36 (20%)||46 (26%)|
|Vote counting||49 (27%)||52 (30%)|
|Inside the voting premises: other||63 (34%)||28 (16%)|
|Total||184 (100%)||175 (100%)|
|Tallying procedure||36 (23%)||35 (16%)|
|Election observer rights||45 (29%)||61 (29%)|
|Absentee ballots||16 (10%)||31 (15%)|
|Processing||18 (12%)||18 (8%)|
|Election day: other||34 (22%)||68 (32%)|
|Criminal||5 (3%)||1 (1%)|
|Total||154 (100%)||254 (100%)|
The violations themselves and the best ways of identifying them aside , a fair amount of research is focused on the reasons behind the manipulations. Comparative political studies both general reasons behind the manipulations  and their certain aspects, such as electoral system manipulation [5; 14], media manipulation and denying opposition candidates participation . Researchers typically attribute electoral malpractice to the fear of losing legitimacy  or, on the contrary, the desire to project power . While some studies stress the importance of low-level decision-making when it comes to maintaining a stable connection between the incumbent and their agents of manipulation in the case of direct fraud at polling stations , there is much less coverage of routine practices, as the typical task is to identify violations that had a direct impact on election results in order to separate them from involuntary violations . In the light of a contradiction between the minimum amount of low-level violations and the violations at the Saint Petersburg PEC, I put forward a research issue: what makes the members of precinct election commissions (PECs) favor non-transparent practices in tallying election results, while they do not display any apparent fraudulent intent?
To answer this question, I use street-level bureaucracy theory, so the first part of the study briefly describes the basics of this theory, gives reasons for its pertinence for the case study and proposes hypotheses based on it. Next step of the study characterizes the data selected for checking the factors that explain the choice of non-transparent practices as well as describes the methods used for testing the hypotheses. The concluding part tests the proposed hypotheses, answers the question of why the identified factors affect the behavior of PEC members and draws a conclusion.
To study the practices of PEC members with a deciding vote ("PEC members" or "commission members"), I propose analyzing the process of implementing a political course while focusing on the officials who interact with clients directly, i.e. analyzing street-level bureaucracy. Although the Law and Society approach offers more common ways to studying law enforcement, the concept of street-level bureaucracy allows to identify those factors affecting PEC members that stem from them performing their bureaucratic duties without addressing the cases of administrative coercion to commit violations. What is more, legal sociologists [4; 17; 3; 26] typically study the practices of judicial and law enforcement agencies, while street-level bureaucracy theory helps find the common ground between these agencies and the more innocuous institutions. For that reason, this paper uses Law and Society studies to illustrate the hypotheses of the street-level bureaucracy theory.
The concept of street-level bureaucracy was created by Michael Lipsky. He describes street-level bureaucrats as public service workers who directly interact with clients of state agencies on a regular basis and hold considerable discretionary power [13: 3]. Discretion means freedom to exercise authority and often arises from the need to work with unique situations, which cannot be fully solved with a single formal procedure. Such public service workers include teachers, police officers, judges and others. Lipsky's theory uses the practices of low-level structures to explain why the results of any given political course defy expectations . According to Lipsky, a number of factors prevent street-level bureaucrats from acting the way the legislator intends them to.
The PEC members with a deciding vote may be conceptualized as street-level bureaucrats since they meet all the basic criteria for such. First, they are employed by the state as public service workers. They also meet the second criterion, which requires them to directly interact with the clients: according to Article 64 of Federal Law No. 67-FZ of 12 June 2002 "On Basic Guarantees of Suffrage and the Right to Participate in Referendums of Citizens of Russian Federation" ("67-FZ"), they issue ballot papers directly to citizens. In addition, PEC members communicate with election observers and are required to give them capacity to ensure that all the necessary procedures (including tallying the results) are followed as well as access to certain documents. The third criterion – the considerable discretion – is less evident in PEC members' affairs, since there is little room for discretion while issuing ballots upon providing a passport. However, the procedures performed upon closing the polling station leave plenty of room for discretion while tallying the results, even if it is not triggered by a legislative error. A closer look at the interaction between PEC members and election observers also reveals that the degree of formality between them (meaning interactions in written form) is mostly at the discretion of the PEC members. What is more, the positions of the PEC chair, deputy chair and secretary do not exclude interaction with clients (observers specifically), which is why their individual conceptualization through the theory of street-level bureaucracy would be redundant.
Aside from the PEC members with a deciding vote, the following individuals are allowed on the voting premises: PEC members with consultative vote, election observers, voters, candidates and/or their authorized representatives, media representatives, police officers. However, the voters alone go the voting premises to exercise their right to vote. Nevertheless, the rest of the actors that are not payed by the government for their work as part of the PEC cannot be defined as street-level bureaucrats, which is why we shall refer to the legislation. Item 5 of Article 5 of Federal Law No. 67-FZ the states that commissions shall act on principles of fairness and transparency. According to Article 2, election observers shall monitor the activities of the PEC members who are obligated to allow the observers to make sure all of the procedures (including tallying the results) are properly followed. In this case, the observer may be described as a client of the PEC, whose demands must be met. The list of activities not allowed to the PEC members with consultative vote during the PEC's work (Item 22 of Article 29 of 67-FZ) suggests that their activity is least connected with providing clients with services. It also suggests that such PEC members are not being paid by the government and are mostly recruited by parties, candidates and NGOs. This category of actors can therefore be conceptualized as clients as well. The same applies to candidates, their authorized representatives and media representatives who are eligible to a lesser number of services from the PEC members with a deciding vote. With that said, election observers, PEC members with consultative vote, candidates and their authorized representatives as well as media representatives (collectively "observers") can be conceptualized as clients of government institutions in keeping with the theory of street-level bureaucracy.
In broad terms, street-level bureaucracy theory describes a number of factors that determine the behavior of low-level public service workers and may explain the non-transparent practices of PEC.
The first factor that determines the behavior of street-level bureaucrats is the resource shortage that often prevents them from completing set objectives, especially while the demand for public services is ever-growing most of the time [13: 27-28]. Naturally, this affects the quality of said services. The upper-level public employees determine the performance indicators for low-level public employees in order to control resource consumption and performance quality of street-level bureaucrats. Since street-level bureaucrats often work with unique situations, the gross figures introduced for monitoring purposes discourage the low-level public employees, which leads to conflict between their motivation and the client's objectives. This is a representation of a conflict between the institution's own goals and objectives set for it specifically. The conflict may be further exacerbated by vagueness and inconsistency of said objectives, which are thus open to interpretation while developing the so-called "surrogate indicators" [13: 52]. The examples of such indicators are given by criminal intelligence investigators who artificially overstate the gravity of solved crimes [28: 64] and the negative effect of the "judicial error" indicator that reduces the fairness of court decisions [18: 22]. In any event, most street-level bureaucrats share one resource in shortage: time, which is determined by the workload. This is evidenced by the consequences that may follow if Russian judges [27: 57] and criminal investigators [25: 85] experience time shortages in their line of work. Although the studies cited do not conceptualize law enforcement officials and judges as street-level bureaucrats, they do show that time restrictions and increased demand for their services make it overwhelmingly difficult to properly meet the objectives connected with administering justice and investigating the circumstances of an alleged crime.
There are no time regulations for PEC proceedings after voter admission is over, although there are de facto limitations concerning the capacity of the voting premises and the physical stress of PEC members who typically have to go back to their regular workplace next day. In this case, the resource shortage meets quite inadequate objectives, which is tallying the results while maintaining an open and transparent setting. As a result, PEC members have to give up strictly following the procedure due to time constraints. This brings me to the first hypothesis of my study: the more time-consuming the tallying procedure is, the more likely the PEC members are to resort to non-transparent practices.
The second aspect that determines the behavior of street-level bureaucrats is related to the actions of clients. Street-level bureaucrats strive to limit the clients' access to resources in shortage and control their behavior to save said resources. This is represented through imposing formal requirements on clients , some of whom may not meet these requirements as well as through giving informal cues to a potential client about how a quick resolution of their issue may be unlikely. The work of the Secretariate of the Russian Constitutional Court described by Arina Dmitrieva is a prime example, as the Court filters out filings written in simple language at an early stage to reduce the workload on the judges and catalogue clients with no knowledge of the legal language .
At the same time, street-level bureaucrats often have to deal with clients who are not easily "routinized," meaning their cases cannot be solved through a previously established method. Street-level bureaucrats use different means to minimize the effect of such "non-routinizeable" cases. A good example is found in the study by Richard Weatherley and Michael Lipsky , where they describe the way special-education reform in the US aimed at establishing a focused approach enabled teachers to get rid of "problem" children by transferring them to special-education classes. The client's behavior that indicates they require significant attention and resources meets bureaucratic resistance. Lipsky claims that one of the aspects of client control is isolating them from each other: street-level bureaucrats resist any organized effort coming from their clients, deeming it destructive and unnecessary [13: 119]. The capability to resist, however, should decrease as the number of demanding clients increases. It should be understood that priority clients of the PEC are voters. The role expectations of the PEC members towards clients rarely contradict reality as the services they perform are simple and apparent to the client. The same, however, cannot be said about election observers, whose requirements for the PEC members may be more surprising. This brings me to the second hypothesis of the study: the more demanding observers there are on the voting premises, the more transparent the PEC is in tallying procedures.
Lipsky points out that some of those who fall under the term of street-level bureaucrats do not act in the same vein described by the theory [13: 204] due to their occupational and moral characteristics. Sure enough, selecting PEC members through parties and NGOs allows people concerned with PEC's transparency to become members. Peter Hupe and Michael Hill claim that the behavior of street-level bureaucrats is influenced by who holds them accountable and how, among other things. For example, the behavior of the doctors is largely influenced by their colleagues' views on them . One can assume that PEC members dispatched by political parties and organizations will be least dependent on the rest of the members. We will refer to such members as independent PEC members from now on. As a consequence, commission members unaffiliated with organizations which concern themselves with objective reflection of the expressed will of the people are likely to be more dependent on their colleagues' opinion. This brings us to the third hypothesis of the study: an independent member on the PEC makes its proceedings more transparent.
I would therefore like to propose the following hypotheses based on the theory of street-level bureaucracy: 1) the more time-consuming the tallying procedure is, the more likely the PEC members are to resort to non-transparent practices; 2) he more demanding observers there are on the voting premises, the more transparent the PEC is in tallying procedures; 3) an independent member on the PEC makes its proceedings more transparent.
To test the influence of the time and observer factors, I have conducted a survey of two respondent groups. The first group consisted of Saint Petersburg PEC members active in 2013-2018: the questions concerned their recent experiences in electoral work. The survey took place between October 2017 and May 2019. The adopted unit of analysis was the PEC at the specific election, which means that the overall count of the specific PECs studied is 5577 (meaning one and the same PEC working at different elections is considered an isolated case). This means that a single respondent from a single PEC is enough to describe a single case. The survey covered representatives of 619 PECs. The respondents were randomly selected from all the districts based on the data provided by the Saint Petersburg election commission . Out of these 619 PECs, representatives of 385 (61.2%) were found on social media. Representatives from 216 PECs agreed to take part in the survey, meaning representatives of 43.9% of PECs refused to do so. This allowed me to describe the election-day situation inside 95% of PECs with an error of 6.7%. The search for respondents was conducted on social media sites, where 2699 (39.5%) PEC members were found out of 6827 selected (20730 overall).
The second survey covered 212 election observers. The overall number of this respondent category is unknown, as the official record of election observers is not publicized. For this reason, this survey uses the PEC at the specific election as the adopted unit of analysis as well. In this case, the error amounts to 6.7%. The search for respondents was conducted on social media as well: the selection process covered people who were members of at least two communities of any political party represented in Saint Petersburg Legislative Assembly. One of these communities had to represent a cell below the district level. Our assumption was that the more local the cell is in the party hierarchy, the more chance we have in finding the party activists. The survey also covered members of communities for the movements "Petersburg Observers" and "Golos".
The survey responses were analyzed through logistic regression, as the dependent variable is dummy (the question of whether there were non-transparent practices had to be answered with either "yes" or "no"). This type of regression analysis estimates the probability of an event taking place – non-transparent practices in our case – depending on determined factors. To test the influence of independent PEC members, we use secondary data analysis by means of linear regression. Secondary data is taken from two sources: the data from Saint Petersburg election commission contains the list of the PEC members and their characteristics , while the data from the Petersburg Observers Movement contains the amount and nature of violations .
Ten people agreed to take part in the semi-structured interview. All of them indicated their will to be interviewed in the surveys (Table 3). Overall, there were three blocks of questions on every working phase of the PEC (before 8AM, between 8AM and 8PM, after 8PM on election day). In total, around 19 questions were asked. The responses are analyzed through narrative structuring technique based on the following pattern: characteristics of PEC members; characteristics of observers and the respondent; conflicts during election day, their origin and resolution; presence of non-transparent practices during tallying, observers' and PEC members' reaction. The results of the analysis are summarized.
|No. in the text||Date of interview||Respondent status during elections||PEC district||Election date|
|1||21 January 2019||Observer/PEC member with consultative vote||Krasnogvardeisky||18 September 2016|
|2||3 February 2019||Observer/PEC member with consultative vote||Admiralteisky||14 September 2014|
|3||4 February 2019||Observer/PEC member with consultative vote||Kalininsky||18 September 2016|
|4||5 February 2019||PEC member with a deciding vote||Nevsky||18 March 2018|
|5||10 February 2019||PEC member with a deciding vote||Krasnoselsky||18 September 2016|
|6||10 February 2019||Observer/PEC member with consultative vote||Kalininsky||18 September 2016|
|7||18 February 2019||Observer/PEC member with consultative vote||Kronstadt||18 September 2016|
|8||26 February 2019||PEC member with a deciding vote||Krasnoselsky||18 March 2018|
|9||27 February 2019||Observer/PEC member with consultative vote||Petrogradsky||18 March 2018|
|10||8 March 2019||PEC member with a deciding vote||Tsentralny||18 March 2018|
Time constraints. The results from the analysis of PEC members' responses indicate that the time factor influences the discontinuity of procedure: there is a threshold beyond which the PEC members are more likely to commit violations.
In this case, the practices (dependent variable) are operationalized as a violation of the legislated (Item 1 of Article 68 of 67-FZ) tallying procedure registered by the respondent (сlosed-ended question: "During the recent election, was the work with electoral rolls going on at the same time as any other procedure (destruction of unused ballots, for example) after the polling station had closed?": 1 – Yes; 0 – No).
Since time as a resource decreases as morning draws closer, it makes sense to estimate these restrictions as a proxy variable by means of PEC's closing hours. As shown through analysis of variance (Table 4), the increased number of campaigns conducted on the same day (which is related to the PEC's workload, but not limited to it) is linked to how late the PEC's closing hours will be ("When did the PEC finish its work during the recent election?": before 12AM; between 12AM and 2AM; between 2:01AM and 6AM; after 6AM). This allows us to consider a connection between the PEC's closing hours and workload.
|PEC closing hours||N||Mean||Standard deviation||Standard error|
F-test equals 20.9, null hypothesis probability equals 0.
Table 5 shows that closing hour variables of before 2AM and between 2AM and 6AM have a more significant effect on procedure violations compared to the reference category (which was established as closing hour after 6AM). This indicates that the PECs that finish their work at any time before 6AM are equally likely to commit the violations under discussion. The negative estimates of parameters indicate that such PECs are less likely to commit violations than the PECs that finish their work after 6AM. Solving the logistic regression equation indicates a 16% probability of procedure violations occurring in the PECs that finish their work at any given time before 6AM, while finishing work after 6AM raises this probability to 50%.
|Finished before 2AM||-1.66**||0.6|
|Finished between 2:01AM and 6AM||-1.57*||0.64|
NB: * – 0.05 significance level, ** – 0.01; reference category – finished after 6AM; Nagelkerke pseudo R2 is 0.07.
It cannot therefore be said that there is a linear dependence between the time resource and the PEC's tendency towards violation. The results of the analysis indicate that there is a certain threshold in workload within which the probability of violations occurring is low and equal, but once it is crossed the probability to commit violations against formalized procedures increases. In this case, this threshold can be identified as the amount of workload that can be handled before 6AM.
Observers. An analysis of the surveys from both groups of respondents has shown that the number of demanding observers affects the way unused (unissued) ballots are counted: non-transparent counting is less likely when there most observers are demanding; besides, their chances increase with the overall number of observers at the PEC.
In this case, the practices are operationalized as non-manual counting of unused ballots, registered by the respondent (closed-ended question: "During the recent election, did PEC members count every unused ballot manually or did they rely on other counting methods after the polling station had closed?": 1 – Not all counting was manual; 0 – All counting was manual). The legislation gives a vague description of this practice, which enables PECs to act in a non-transparent manner without breaking the law, as the specific counting method for unused ballots is undisclosed: "Once the voting time runs out, the precinct election commission members with a deciding vote... count and destroy the unused ballots by cutting off the lower left corner in the presence of election observers..." (Item 3 of Article 68 of 67-FZ). In this case, the demanding observers are operationalized as observers who at least looked through the electoral rolls (Tables 6,7). This action requires effort from the observer, meaning an observer with no intention to control the PEC proceedings is unlikely to make this effort, although the action itself does not require any specific effort (closed-ended question: "During the recent election, roughly how many observers and PEC members with consultative vote looked through the electoral rolls before the voting started?": 0 – none or few; 1 – all or most).
|Few observers looked through the electoral rolls||0.84*||0.41|
NB: * – 0.05 significance level, ** – 0.001; reference category – all or most observers looked through electoral rolls; Nagelkerke pseudo R2 is 0.04.
|Number of observers in PEC||-0.13*||0.06|
|None or few observers looked through the electoral rolls||1.01**||0.38|
NB: * – 0.05 significance level, ** – 0.01; reference category – all or most observers looked through electoral rolls; Nagelkerke pseudo R2 is 0.1.
The coefficients for the variables are significant for both selections, meaning it is safe to conclude that the relative number of observers who look though electoral rolls has a valid connection to the way unissued ballots are counted. The parameter used as the reference category for estimating the impact of observers who look through electoral rolls is "most or all observers looked through electoral rolls." It can be thus argued that non-transparent practices are more likely when few observers look through electoral rolls then when most observers do so. A variable for the total number of observers was also added to Table 7 (Open-ended question: "Roughly how many observers and PEC members with consultative vote worked alongside you on the voting premises?"). The coefficient for this variable illustrates that a non-transparent practice becomes less likely as the number of observers of any level of training increases. As can be seen in the model from Table 7, the significance of both variables demonstrates that the absolute number of observers who look through electoral rolls is just as relevant for transparent practices as their relative number. Solving the logistic regression equation indicates that as the number of observers per person increases, the odds ratio for a non-transparent practice decreases by 12%. For example, if demanding observers dominate in a group of ten, the probability of non-transparent counting will be 10%. However, if there is a single observer at a polling station, but a demanding one, the probability of non-transparent counting will rise to 26%. By contrast, if only the minority is demanding in a group of ten observers, the probability of a non-transparent practice occurring will be 23%. The same may be said for a single undemanding observer at a polling station: if they are undemanding, the probability of non-transparent counting will rise to 49%.
Considering the data from these two surveys, it is safe to say that the majority of demanding observers is capable of convincing the PEC to satisfy their demands even if the PEC members are not required to do so (as is the case of unissued ballots). Besides, as can be seen from survey analysis, the observers' chances of successfully influencing the PEC increase with the number of observers.
Independent members of the commission. The secondary data used to test the hypothesis that independent PEC members increase the transparency of practices can neither confirm nor disprove the proposed hypothesis.
The dependent PEC members are operationalized as PEC members who changed their official party identity after transferring to a new composition of PEC or who were nominated by "voters' assemblies at their place of work, residence, study or service." It turned out that at least 49.6% of PEC members from the previous election in Saint Petersburg could be more dependent on their colleagues than on organizations that officially nominated them (Table 8).
|An estimate of the minimum number of dependent PEC members (% of the total number of PEC members)||The number of all registered violations in 2013-2018||Number of violations during tallying||TEC, district|
|267 (34.9)||148||45||1 Admiralteysky|
|402 (47.1)||26||3||2 Vasileostrovsky|
|215 (32.3)||99||13||3 Kirovsky|
|422 (58.6)||18||0||4 Krasnogvardeysky|
|738 (51.6)||149||4||5 Nevsky|
|460 (52.5)||7||0||6 Krasnoselsky|
|383 (44.4)||48||4||7 Kirovsky|
|224 (50.7)||6||0||8 Petrodvortsovy|
|179 (35.4)||18||2||9 Primorsky|
|285 (52.4)||9||0||10 Vyborgsky|
|412 (45.9)||46||1||11 Kalininsky|
|301 (49.5)||2||0||12 Primorsky|
|193 (57.8)||1||0||13 Kurortny|
|324 (46.6)||18||0||14 Vyborgsky|
|106 (56.4)||2||0||15 Kronshtadtsky|
|194 (38.6)||17||3||16 Tsentralny|
|167 (39.6)||31||4||17 Kalininsky|
|374 (56.7)||33||7||18 Petrogradsky|
|336 (44.4)||17||0||19 Moskovsky|
The table is not fully displayed Show table
Sources: http://www.st-petersburg.vybory.izbirkom.ru/region/st-petersburg/?action=ik and https://blacklist.spbelect.org.
It should be noted that this is a non-exhaustive method of identification, as PEC members may be dependent without changing the dispatching organization. To solve this problem, we grouped the data on identified percentage of dependent PEC members by territorial election commissions (TEC), which allowed us to see the difference in ratios between the districts. The violations are operationalized through the percentage of commissions under the jurisdiction of a single TEC (territory), where any violations occurred between 2013 and 2018 (according to Petersburg Observers Movement).
|Minimum % of dependent PEC members within the territory||-0.86*||0.32|
NB: * – 0.05 significance level, ** – 0.001; No autocorrelation and heteroscedasticity was detected.
Regression analysis findings (Table 9) contradict the proposed theory: less independent PEC members make for fewer violations. Such findings may have been caused by the nature of the selection: the data from Petersburg Observers only indicates the cases where observers and independent members who recorded the violations were present. It is thus possible to claim that this data likely indicates the relation between the number of independent PEC members and the chance to register violations instead of the relation between violations and said members. For this reason, it is likely that the greater the number of such members in the commission, the higher the odds to discover a violation. It is thus impossible to affirm if the mutual dependence of the PEC members has any influence on compliance with procedure with help of available data, so this issue requires additional examination.
Why are observers and independent members of the commission (in)capable of convincing the commission to act in a transparent manner? Different sets of circumstances allow us to describe different causes for significant or insignificant observer influence on practices carried out by PEC members. When it comes to commissions without independent members, it is easier for most demanding observers to influence the commission when they have the support of other observers. Coordinated effort helps them track all the actions performed by the commissions, and its members find it harder to resist such coordinated effort, which cannot be found in different circumstances. This becomes more difficult for a single demanding observer, especially with time constraints.
In the case of PECs with independent members on board, the root of significant influence may lie in previous negative working experience in the PEC shaped by violations. This is where the so-called "learning curve" may occur: through conflict, the PEC not only learns about the procedure, but also about the manner in which observers and independent PEC members will react in future. Still, regardless of the combination of these factors, the incompetence of most PEC members is typically considered one of reasons for non-compliance.
The data collected through surveys suggests a relation between higher demand from observers and PEC workload on the one hand, and non-transparent practices on the other. At the same time, the surveys do not explain how PEC members manage to avoid following the rules in the presence of at least one demanding observer, or, on the contrary, how a large number of observers manages to successfully influence the PEC. Explanations can be found in interviews conducted with PEC members and observers. These interviews help to infer how observers manage to convince PEC members to follow procedure (the number in the brackets that follows each quotation corresponds with the respondent's number in Table 3).
As expected, the respondents who worked at polling stations without demanding observers describe the atmosphere as less tense. Despite non-transparent practices, including violations ("...nobody counted them [unused ballots], they were just lying around" (01); "At the same time, meaning the things that must not be done this way: counting the ballots and working with electoral rolls, this was done" (04)) the parties involved are polite to each other in such a situation ("...the PEC members I worked with were quite pleasant..." (01)) and they are similar in their understanding what the good practice is: "...she [the chair] was waving them [observers] off, and they weren't especially persistent. Which was good, because otherwise it would've slowed down the commission even more" (04).
The respondents from commissions with at least one demanding observer give an opposite perspective. The analysis above has shown that in this case the observer finds it most difficult to influence the PEC, which is likely explained by inability to control the all the actions of a large commission ("Going alone is no use at all, because if you try to, I don't know, go for lunch, to the restroom, and they [PEC members] will do anything..." (03)) as well as by lack of support, which puts an emotional strain on the observer should they try to interfere. Because of this, observers may ignore certain violations to preserve their strength, or whatever is left of it ("I noticed that [violation], but, honestly, after a whole day of bickering with the commission I didn't have much fighting spirit left in me to prevent it (09)). The heavy workload and the resulting time constraints may also be the reason for the commission to ignore small numbers of demanding observers ("...they [PEC members] all were ranting about how no one shows understanding for them, how they have to spend 24 hours at the polling station when they have classes at 8AM...said it didn't matter what kind of election we're having, we have classes tomorrow and all, so get moving" (02)). Or, at least, the PEC members find it justified to appeal to time constraints and exhaustion while seeking understanding for their situation.
In cases with small number of demanding observers, the commission tends to ignore their demands ("...she [the chair] just flat-out ignored any objections..."(02)). Still, even in conditions as difficult, the observers may convince the commission to work in a more transparent manner. This may happen when the PEC members have to face the negative consequences of non-transparent work themselves ("I kind of turned a blind eye on it. Then, when they started drawing up a statement, it turned out they're missing two thousand ballots, or fifteen hundred or so. So I went ahead and forced them to count it all manually" (07)). Then the observer may successfully stand their ground ("...they realized that the law is on my side and if I file a complaint about how they blew fifteen hundred ballots, that would be a major screw-up. And the statement didn't add up. They all wanted to go home" (07)).
Finally, as can be seen earlier in the study, large numbers of demanding observers gain more traction with PEC members. In this case, observers can support each other, and even the slightest amount of support makes it easier to influence the commission ("Other observers supported me, but they were still too anemic, I think. But there was this guy, very spunky, he did have a good say, too" (03)). Emotional support aside, coordinated effort of observers makes it much easier to control the commission ("If they [PEC members] see that there is a serious team effort coming from observers, they don't try to do anything unlawful (03)). This effect gets a significant boost when one of the independent PEC members (if any such present) acts as a coordinator: "I told those [observers] I could rely on about the things they should turn their attention to," "told them what to keep an eye on before everything else" (10).
Being able to organize coordinated effort from demanding observers aside, continuous experience with the same PEC throughout multiple election campaigns may produce what can be called a "learning curve." The point is, if the PEC ever came into conflict with demanding observers or independent members, this painstaking process could help its members to adopt norms that would allow them to meet the expectations of observers and independent members in future, especially if they had worked together before: "...[in 2014] they first encountered a PEC member who demanded compliance with the procedure" (08), "I feel ... in 2018 the commission was already prepared to follow the procedure during election" (08), "he [a PEC member dispatched by the party "Yabloko"] is an educated young man, has been part of that commission for a long time, or at least worked with it before the election, so he commands some respect, you know. I suppose we got the reliable results thanks to him, in no small part" (06).
This interview summary provides an assessment of both how observers manage to convince the PEC to follow the procedure and why they fail to do so despite all their effort. It also suggests that there are other factors that influence PEC proceedings. Speaking of how PEC members interact with observers, it is worth mentioning that it varies depending on the level of demand coming from observers. As it turned out, the main reason for the failure of one demanding observer or a small number of such is their limited physical (inability to monitor everything that is happening in the PEC) and emotional (it is difficult to put forward new demands when you expect a previous conflict to repeat itself) capacity to control the PEC. In this regard, both limitations are removed if the number of demanding observers increases, especially when they are the majority and they get help from the workers already familiar with PEC practices. Besides, violation-related negative experience from interacting with independent PEC members and demanding observers gained during previous election may help to successfully influence the PEC.
Many respondents described a factor that deserves special mention. In their opinion, the incompetence of PEC members is a likely cause of violations: "lack of professionalism" (02), "...nearly all PEC members are incompetent..." (04), "...these violations were caused by incompetence, not the desire to falsify anything" (07). It is indeed difficult to follow the procedure if the individual has no idea how. It is possible to assume that the competence level of PEC members may influence the way they perform the practices. The negative experience from interacting with those familiar with the procedure may result in the "learning curve." Nevertheless, this "learning curve" phenomenon and the related level of competence of PEC members requires additional testing, as it is unclear whether this effect persists without demanding observers (or independent PEC members) or with limited temporal resources. Table 10 demonstrates a summary of interview analysis.
|Demanding observers||Independent PEC members||Reason for success||Reason for failure|
|None||Unwillingness to interfere|
|Minority (severe time constraints)||None||1) emotional strain|
|2) inability to monitor all the actions of PEC|
|3) time constraints are more important to PEC members than observer demands|
|Minority (minor time constraints)||Serious mistakes of PEC that make it impossible to continue working without eliminating them||1) emotional strain|
|2) inability to monitor all the actions of PEC|
|Minority||Present||Early negative violation-related interaction experience|
|Majority||None/Present||1) support from other observers|
|2) coordinated effort|
It can be argued that there are factors that make the PEC members more inclined to use non-transparent practices. First, we can affirm that observers definitely wield influence. The analysis showed that a large number of demanding observers is much more likely to succeed in convincing the commission to do something it is not even required to. This is explained by the fact that a large group of demanding observers is capable of making a coordinated effort, and an individual observer may count on support in case conflict arises. Although the analysis indicated that single observers have little chance of successfully convincing the commission, they may still count on it if the PEC needs their help in eliminating a serious mistake. Second, the PEC is influenced by the time constraints that depend on its workload. Although this connection is non-linear, there is a certain threshold beyond which the PEC members are more likely to commit violations.
Despite failing to reveal the influence of mutual dependence between PEC members, the analysis indicated the conditions under which the influence of this possible factor may be neutralized. First, if there are independent members in the commission, they are able to coordinate the actions of demanding observers, as they know how the commission usually operates. Second, independent PEC members may "educate" the rest of the members about regulatory practices: this likely happens at a subconscious level through violation-related conflict. As a result, the rest of the members may predict their independent colleague's reaction to non-regulatory practices and will pay greater attention to following the procedure in future to avoid conflict and save resources.
The analysis expands theoretical assumptions about the behavioral factors of street-level bureaucrats. The study has revealed that not only PEC members seek to organize the clients, but also the clients themselves may respond by successfully "structuring" street-level bureaucrats through transforming their practices in the way that benefits them. However, Lipsky's thesis on how the clients' capability to control bureaucrats prevent public employees from going back to their old practices [13: 208] has to be expanded: this condition is met when the clients are demanding and numerous.
The findings suggest that, unlike the street-level bureaucrats in Lipsky's classic model, the PEC members do not always stay within the discretionary framework established by the legislator. By rationalizing the choice between fully satisfying their clients' demands and effectively using the available time resource, PEC members may not realize they are breaking the law. The fact that public organizations gradually "educate" PEC members about regulatory procedures through the complex way of conflict indicates a basis for future studies. Further research in this area should be dedicated to what kind of influence independent PEC members have over non-transparent practices; what the role of professional training is for the members of commission and whether it has a negative effect over non-transparent practices in case public control is lacking.
I am grateful to Svetlana Tulayeva, associate professor at NWIM RANEPA for her invaluable advice and support.
Received 31.03.2020, revision received 12.04.2020.