ChapterPDF Available

The Swiss Postal Voting Process and Its System and Security Analysis

Authors:

Abstract and Figures

The Swiss postal voting system builds on trust in governmental authorities and external suppliers. The federal structure of Switzerland of cantons and municipalities leads to a distributed architecture. Detailed information on the current postal voting procedure are manifested as implicit knowledge within fragmented institutions and are not easily accessible. This work serves (i) as an overview of the Swiss remote postal voting system, (ii) a detailed insight into the process flow, and (iii) a respective risk assessment.
Content may be subject to copyright.
The Swiss Postal Voting Process
and its System and Security Analysis
Christian Killer and Burkhard Stiller
Communication Systems Group CSG, Department of Informatics IfI,
Universit¨at Z¨urich UZH, Binzm¨uhlestrasse 14, CH-8050 Z¨urich
{killer,stiller}@ifi.uzh.ch
Abstract. The Swiss postal voting system builds on trust in governmen-
tal authorities and external suppliers. The federal structure of Switzer-
land of cantons and municipalities leads to a distributed architecture.
Detailed information on the current postal voting procedure are mani-
fested as implicit knowledge within fragmented institutions and are not
easily accessible. This work serves (i) as an overview of the Swiss remote
postal voting system, (ii) a detailed insight into the process flow, and
(iii) a respective risk assessment.
Keywords: Remote postal voting, risk assessment.
1 Introduction
Around the globe, government services are becoming increasingly digitized [1].
Naturally, these efforts include electoral processes. In Switzerland, the federal
government defined strategies enabling digitization for public authorities and
processes, including Electronic Voting (EV) [32,11]. Private companies collab-
orate with Swiss authorities to actively define standards across e-Government
processes [35]. The Swiss EV typically refers to Remote EV (REV) carried out
over the internet, which is also often referred to as Internet Voting (I-Voting) [19].
According to recent studies [32], 47% of Swiss citizens would be more likely
to vote if EV were available, and almost 70% of Swiss citizens welcome an EV
system [21]. Despite the positive sentiment surrounding EV, a current political
position proposes a moratorium on EV in Switzerland [15]. According to their
initiative [15], a REV system has to be “at least as secure as the current remote
postal voting (RPV) system”. Thus, the key question is: what exactly does such
a minimal level of security involve? Which security metrics and mechanisms
are mandatory? In the general public perception, EV often provokes a fear of
change, presuming the current RPV system to be mostly analog and tamper-
proof. However, it can be argued that the current Swiss RPV system is already
partially EV, since many steps already involve distributed electronic systems.
Thus, defining and comparing the security properties of a REV also requires an
analysis of the current RPV system in Switzerland.
Reducing cost and increasing the voter turnout by providing a convenient
way to vote are important considerations for Swiss authorities [20]. By 1994,
2 C. Killer, B. Stiller
all cantons accepted votes by postal mail. As of today, RPV is the dominant
voting channel, used by approximately 90% of the voters in Switzerland [16].
Most eligible voters in Switzerland show trust in the authorities on the federal,
cantonal, and municipal level to handle electoral processes and protect voter
privacy [32]. The trust placed in authorities encompasses state-owned companies,
which are important stakeholders in the current RPV system.
Due to the federal and decentralized structure of Switzerland, each canton
and municipality autonomously manages their respective jurisdictional electoral
procedures. Cantons and municipalities execute a degree of independence in deci-
sions on how to handle certain parts of the voting process. Therefore, the current
RPV system in Switzerland is neither universally documented or specified, nor
homogeneous across entities.
This paper, therefore, summarizes major related work and terminology to
formalize the Postal Voting Process Flow (PVPF) in Switzerland. The approach
taken formalizes the PVPF in a step-based model, for which major assumptions
made, such as trust, people involved, and technology applied, are made explicit,
if known. The dedicated interpretation of social trust assumptions is discussed
within Sec. 3, along with the risk analysis, weaknesses and strengths of a person-
based RPV approach. Finally, the paper performs an overall risk assessment in
Sec. 4, providing the basis for discussions of security-relevant comparisons to
REV or I-Voting, while Sec. 5 draws main conclusions.
2 Legal Background and Related Work
Switzerland is organized as a decentralized system of municipal and cantonal
entities, working together under the umbrella of the Federal Government. The
federal structure is also mirrored in the legal framework (cf. Figure 1). At the
root rests the Federal Constitution of the Swiss Confederation, wherein Art. 39
[6] forms the basis for the Federal Act on Political Rights (BPR) [8]. In turn,
Art. 91 BPR [8] is the foundation for the Federal Decree for Political Rights
(VPR) [9]. On a cantonal level, the VPR builds the foundation for the Cantonal
Decrees (e.g., for the canton of Aargau [7]). Every canton is an independent legal
entity and defines its own constitution on the basis of the Federal Constitution.
The political system is under the authority of the cantons, i.e., cantonal laws
and ordinances regarding political rights define elements for these processes.
Various aspects of those elements are relevant for the RPV system in Switzer-
land, and each canton has its own decrees regarding political rights. The federal
structure is mirrored down to the municipal level: each municipality decides
on certain processes, again, aligned to cantonal laws and decrees. For instance,
keeping record of the electoral register is under the authority of municipalities,
leading to different approaches.
A direct comparison of the Swiss RPV system to REV was performed in [29].
Other countries discuss the usage of RPV critically because the secrecy of the
ballot cannot be fully ensured [29]. From a practical standpoint, thorough docu-
mentation is the easiest way to achieve verifiability for RPV. Supervisory bodies
The Swiss Postal Voting Process and its System and Security Analysis 3
Fig. 1. Swiss Legal Framework
and authorities should check the documentation and verify it [29]. Also, trust
is crucial for all voting methods. And the relationship between verifiability and
trust is neither linear nor one-dimensional. Technical measures are not sufficient
to create trust, sociopsychological aspects also have to be considered carefully.
An extended literature review is provided in [18] with a focus on Switzerland,
but also outlining the work done in Canada [22], Estonia [19], and Australia [30].
In order to analyze the RPV system with a focus on security aspects, the US
National Institute for Standards Technology (NIST) serves as reference, outlin-
ing and standardizing terminology on the “Effort, Detection and Impact levels
of Threat Events” [26,33,34]. Past work applied such principles to a RPV system
used in the United States [25]. To consistently apply terminology, Table 1 de-
fines the corresponding terminology used in Swiss legislation and their English
translations.
Tab. 1. Official German Terminology with Corresponding English Translations
German English
Zwei-Weg Abstimmungskuvert Two-Way Voting Envelope (VE)
Abstimmungsresultat Voting Result (VR)
Erwahrung des Abstimmungsresultates Legally valid determination of VR
Die Schweizerische Post The Swiss Post (SP)
Stimmkuverts Paper Ballot Envelope
Stimmrechtsausweis Voting Signature Card (VSC)
Stimmregister Electoral Register (ER)
Stimmzettel Paper Ballot (PB)
Vertrauensw¨urdiger Dritter Trusted Third Party (TTP)
3 Postal Voting Process Flow (PVPF) in Switzerland
This section details the illustrated Postal Voting Process Flow (PVPF) in-depth
(cf. Figure 3), containing an end-to-end process as it is currently implemented
in Switzerland. The detailed sub-steps are formalized and vary between cantons
and municipalities. However, the general process adheres to the federal laws and
4 C. Killer, B. Stiller
ordinances. The PVPF is divided into six main phases from Ato F, each phase
containing one or multiple sub-stages from 1 to N(cf. Figure 2). The ensuing
subsections are structured according to the PVPF within Figure 3 and describe
all the different steps in detail.
Fig. 2. Paper Voting Process Phases
3.1 Setup Phase
The Setup phase Acontains four sub-stages 1-4 describing the production and
assembly for dispatch of all necessary ballots and envelopes(cf. Figure 4). The
two-way Voting Envelopes (VE), the Voting Signature Cards (VSC), the Paper
Ballot Envelopes (PBE), and the Paper Ballots (PB) are the physical artifacts
produced in the Setup phase. The secure execution of the Setup phase is crucial,
since all following phases rely on the sound production and assembly of those
artifacts. The main stakeholders of this phase are the municipal and cantonal
authorities supervising the process. Due to cost, time, and capability constraints,
Trusted Third Parties (TTP) support the authorities during the Setup phase as
External Suppliers (ES).
Production of voting envelopes: In Step 1, the certified two-way VEs
are produced by an ES. In the canton of Aargau, the municipality secretaries
place a centralized buying order [28,17] for the two-way VEs at least a year in
advance. After production, the VEs are distributed among the municipalities. In
municipalities where Step 3 is outsourced, the VEs are directly delivered to the
corresponding ES. The exact process steps are under municipal authority and
can differ accordingly. Some cantons contract a single ES to handle the complete
Setup phase A, mainly due to the special requirements of EV systems [10].
Production of Paper Ballots and Voting Signature Card: Step 2 con-
sists of the production of the PBs and the VSCs. The printing of VSCs and
PBs is predominantly commissioned to an ES. Each political layer in Switzer-
land (Federal, Cantonal and Municipal) commissions the PBs within their legal
responsibility, i.e., the production of federal referendum PBs are commissioned
by the federal government, cantonal PBs are commissioned by the cantonal au-
thorities, and municipal PBs are commissioned by the municipal authorities.
A VSC contains the name and address of the eligible voter, embedded within
a template customized by the municipality. It is essential that the printed creden-
tials are valid, since the assembled VE is delivered to the credentials printed on
the VSC. The voter has to sign the VSC for the ballot to be valid. A substantial
amount of ballots are not counted because many VSCs remain unsigned.
The Swiss Postal Voting Process and its System and Security Analysis 5
Fig. 3. Paper Voting Process Flow (PVPF)
6 C. Killer, B. Stiller
TWO-WAY VOTING
ENVELOPE (VE)
VOTING SIGNATURE
CARD (VSC)
PAPER BALLOT
ENVELOPE (PBE)
PAPER
BALLOTS (PB)
Fig. 4. Abstract representation of the necessary paper artifacts
The individual VSCs are printed according to an electoral register (ER). The
ER is a centralized register containing all eligible voters. The ER is under the
authority of each municipality. Neither the cantonal, nor the federal authorities
have copies of the ER [28].
Since most municipalities contract an ES for the printing of the VSCs, a
dataset containing the eligible voters needs to be transferred to the correspond-
ing ES. Most municipalities export a file from the ER (e.g., XLS, CSV) and send
the snapshot to the ES directly via email [31,27]. The transmission of an unen-
crypted, unsigned dataset via standard email is critical, since the dataset could
be tampered with (e.g., the creation of fake identities, entries being removed),
either after the export, during transmission, or when the export file reached the
ES. Most ERs are administered by proprietary software systems provided by
companies targeting Swiss public administrations. Some cantons also provide
process checklists to municipalities. For instance, the election office (EO) of the
Canton of Aargau provides such checklists [17]. These checklists state that the
number of VSCs should be identical to the number of eligible voters present in
the ER. If issues arise, an in-depth manual control should take place. Whether or
not to adhere to these checklists is under the authority of the municipality. Also,
since printing and assembly of the VE is mostly done by ESs, the ES should
verify the integrity of the printed artifacts (e.g., content and amount).
Prepare voting envelope for dispatch: Step 3 concerns the final assembly
of the VEs. For each eligible voter (i.e., each VSC), a VE containing the VSC,
PBs, and the PBE is assembled (cf. Figure 4). The assembly is a monotonous
task, often outsourced to ESs or social institutions and foundations [18]. Receiv-
ing an incomplete VE increases the possibility of the voter’s abstention. Accord-
ing to cantonal checklists [17], handing out new PBs is only allowed if the voter
can make the loss credible. Then, the voter’s credentials should be recorded to
check for attempted dual voting in the Tallying phase E[17].
Dispatch of voting envelopes through postal service: The final step
involves the dispatch of the assembled VEs. In some municipalities, the VEs are
directly dispatched by the ES commissioned with the assembly of the VEs. The
Swiss Post (SP) offers a special service [14] for the dispatch and delivery of VEs.
The Swiss Postal Voting Process and its System and Security Analysis 7
3.2 Delivery
In Switzerland, the postal market was partially deregulated in 2009 [13]. Still,
the SP maintains a monopoly on postal letters below 50 gr. Therefore, the SP
is a crucial TTP, since the secure delivery to the municipality falls under the
responsibility of the SP. When using the special service provided by the SP, VEs
can be dispatched on a work day in the week prior to the specified delivery week
[14]. Then, the SP guarantees the delivery of the VEs will take place during the
specified delivery week [14].
3.3 Casting
Phase Coutlines the three different options to cast a vote. The vast majority of
ballots are not cast at the urn [16]. Statistics do not indicate whether VEs arrive
through postal services (I) or were delivered to the letterbox by the voter (II ).
I: The most popular way to cast the vote is to send the VE by postal mail.
Some cantons pre-stamp the VSCs, which can then be used to return VEs free
of charge [3]. For the voter, it is impossible to verify whether the ballot was
successfully delivered to the municipal office. The SP offers the ability to track
deliveries for an additional cost.
II: Thus, a favoured alternative is to deliver the VE into the letterbox of
the municipality, which is then emptied by municipal officials and safely stored.
According to [12], this option is still a favoured option by many voters.
III : The third option is to personally cast the ballot at the urn, which guar-
antees ballot secrecy. Casting ballots at the urn remains the most secure option
to cast a vote, since the PBE (containing the ballot) is directly cast into the urn
and separated from the VSC (containing the voter’s credentials and signature).
3.4 Storage
Phase Ddeals with the storage of VEs that were delivered via postal service (I),
or directly cast to the municipal letterbox (II ). Often, an employee is tasked to
fetch the postal mail addressed to the municipal office. During votes, the VEs
are collected from the SP and municipal letterbox, and then carried to the safe
storage location. Past incidents describe where municipal employees misused
that trust [4]. The storage safety varies heavily, depending on the municipality.
The Federal Act for Political Rights (BPR) [8] does not specify any security
requirements. Additional considerations include the exact definition of an access
control for the VE storage, (e.g., Who should have physical access to the VEs?).
Also, the definition of a process for incoming VEs can increase process security
(e.g., How many ballots arrived at which date and time? Who got the ballots
from the letterbox or postal office and transported them where?).
Thus, stricter access control and a secured ballot arrival process can maximize
the physical storage security. In practice, physical storage security is not priori-
tized, since the municipal infrastructure is often not sufficiently equipped [27,31].
8 C. Killer, B. Stiller
3.5 Tallying
Phase Especifies the process of tallying. The main stakeholders of phase Eare
the municipality and the local EO. The tallying is not regulated on a federal
level and is heterogeneous among cantons and municipalities [29].
Art. 14 No. 1 BPR [8] states that every polling station should create a report
containing the total number of eligible voters, the total eligible voters living
abroad, the total of blank, invalid, and valid ballot papers, and the number of
votes in favour and against the proposal [8]. Thus, the BPR serves as a federal
guideline, without specific requirements regarding the tallying process.
Approximately 10% [23] of the ballots cast are counted with the help of Elec-
tronic Counting (e-Counting) tools, provided by ESs. The parliamentary control
of the administration investigated e-Counting and concluded that the federal
requirements are neither functional, nor practical, and the control mechanisms
of the federal government are not sufficient [23].
Tallying of all ballots: The local EO usually hires paid and elected helpers
to assist with the manual counting. In large cities, thousands of helpers are
engaged to count the paper ballots [24]. The EO defines the details of the tallying
process. Some municipalities use e-Counting solutions or deploy high-precision
scales to weigh PBs and derive the tally from averaging the weight of (sometimes
pre-counted) batches.
Transmission of results from bottom to top: According to Art. 14
No. 2 BPR [8], the cantonal government is responsible for compiling provisional
results from the entire canton and notifying the Federal Chancellery (FC) of the
results, and publishing the same result in the Cantonal Gazette (or a special
issue thereof) within 13 days of the polling day. As soon as the EO concludes
tallying, the result is transmitted from the municipal EO to the cantonal EO, and
from there to the FC. Some cantonal EOs deploy dedicated software systems to
verify results using statistical methods. Also, most cantons make use of software
provided by ESs to transmit the results. Thus, this phase also includes the use
of web-based assistance tools [18].
Publication of results: The tallying phase is finalized with the publication
of all results on the municipal, cantonal, and federal level. Generally, the FC
publishes the collected results from the Cantonal Chancelleries in the Federal
Gazette. Cantons publish the results and protocols in their Cantonal Gazettes.
Each municipality publishes a final tally and tallying protocol with respect to
the cantonal law. Mostly, the publishing process is performed by uploading doc-
uments to a public web-server and displaying print-outs outside the municipal
offices.
Validation: Art. 15 BPR [8] defines the validation and publication of the
results. The official results can only be declared when no valid appeals are in
process at the Swiss Federal Court. After that, the official result is published by
the FC in the Federal Gazette and can not be appealed anymore.
Storage of paper ballots: Before the results are ascertained, the counted
PBs and VSCs have to be safely stored in the municipalities. It is important that
the ballots remain unaltered because a recount could be triggered before the
The Swiss Postal Voting Process and its System and Security Analysis 9
official result is determined. Past cases have shown that premature destruction
of the PBs and VSCs made a full re-count impossible [2].
Set official voting result: As soon as no more valid appeals are with the
Federal Supreme Court, or as soon as a decision has been made on such an
appeal, the legally valid voting result can be determined. According to Art. 15
No. 2 of the BPR [8], the validation decree shall be published in the Federal
Gazette. The officiation by the Swiss Federal chancellery finalizes the Validation
phase. Since a recount is no longer possible and the result is untouchable, the
final phase can be started.
3.6 Destruction
The final phase, G, involves the destruction of the stored VSCs and PBs. Ac-
cording to Art. 14 No. 3 BPR [8], “following validation of the result of the vote,
the ballot papers shall be destroyed.” In practise, the destruction is usually done
by physically shredding all PBs and VSCs [27].
4 PVPF Risk Assessment
A risk signifies the level of impact on the operation of an information system’s
task, given the potential impact of a threat and the likelihood of that threat
occurring [5,34]. Therefore, a risk assessment (RA) serves as the identification
and determination of the impact of vulnerabilities that an adversary can exploit.
A threat covers any event with a potentially adverse impact on the assessed pro-
cess [26]. With respect to the RPV in Switzerland, threat sources are groups or
individuals who could feasibly attack the RPV system. Threat sources can stem
from insiders or external adversaries. All Threat Events (TE) in the following
RA are general in nature and require multiple co-conspiring hostile individuals
or groups to achieve a large-scale effect. The effort for each threat defines the
relative level of difficulty of performing a successful attack based on a threat
[25]. Three relative levels of effort are defined:
Low (): An attack requires little / no resources or detailed knowledge of
the system.
– Moderate (): An attack requires significant resources (or the ability to
obtain these resources) or knowledge of the system. Inside attacks involving
a small number of co-conspirators fall into this category.
– High (+): An attack requires excessive resources, in-depth knowledge of
the system, or even access to the systems. It also requires specific tactics,
techniques, and procedures [26]. Insider attacks involving a large number of
co-conspirators fall into this category [25]
Detection describes the relative level of difficulty to notice whether a partic-
ular threat has been executed in an attack [25]. Thus, attacks are more severe
when they remain undetected. Three estimated levels of likelihood of detection
exist [25]:
10 C. Killer, B. Stiller
– Low (): An attack is unlikely to be detected without extraordinary re-
sources.
– Moderate (): An attack may be detectable, but could require a large
amount of resources and time. Such attacks are unlikely to be detected during
the election.
High (+): An attack would most likely be detected, given proper monitoring.
The impact on PVPF was analyzed according to [26] with the focus on con-
fidentiality (C), integrity (I), and availability (A) as defined in [33]. Some TEs
- all are shown in Tab. 2 - are interdependent or can be combined as depicted in
Figure 5. The following RA serves as a major discussion of potential TEs that
can lead to a loss of voter confidence. The mitigation of identified TEs concerns
the actions of establishing trust and confidence in a system.
Tab. 2. Threat Events on the Swiss Postal Voting Process Flow (PVPF)
Phase TE Description Effort Detection Impact
ATE1 Delay production of physical artifacts A+
ATE2 ER master records + I+
ATE3 ER data snapshot I+
ATE4 Forge physical artifacts + I
ATE5 Steal assembled VEs before dispatch +I+
BTE6 Re-route VEs unknown unknown A+
BTE7 Steal VEs from voter letterboxes +A
CTE8 Steal VEs from municipal letterbox I
CTE9 Re-route VEs + C+
CTE10 Cast stolen or forged VEs I
DTE11 Access stored VEs I+
ETE12 Manipulate tallying + I+
ETE13 Manipulate final tally + I+
FTE14 Initiate premature destruction +I+
Gno major threat events identified
4.1 Risk Assessment Phase A
The Setup Phase Aproduces all the necessary artifacts for the secure execution
of the whole PVPF.
TE1 describes the malicious delay of A1 and A2 in the PVPF. For instance,
delaying the production can be achieved by targeting contracted ESs or
directly attacking the municipal information systems.
TE2 describes the tampering of the ER master records. Often, the ER is pro-
vided and deployed by an ES. A targeted attack of an ES provider or munici-
pal information systems with access to the ERs creates the ability to tamper
with ER master records. The modification of master records can damage the
integrity of ER data and the exported subset of eligible voters.
The Swiss Postal Voting Process and its System and Security Analysis 11
TE3 describes the tampering of ER snapshot data. Instead of modifying the ER
master records, the snapshot used to print the VSCs can be modified. When
the snapshot is neither digitally signed nor encrypted, an adversary could
modify the data before, during, or after transmission to the ES.
TE4 describes the forgery of physical artifacts with (stolen) digital templates.
If an adversary gains access to digital templates used to produce the physi-
cal artifacts, the adversary can forge VSCs and PBs. Additional information
may be necessary to obtain (e.g., weight and type of paper used). PBEs and
VEs may also be forged, stolen, or even ordered from an ES. Since most
municipalities do not perform a validation of incoming VSCs (by compar-
ing the list of eligible voters with incoming VSCs), the attack can remain
undetected. Practically executing TE4 requires a high effort and specific
knowledge of the PVPF down to a municipal level.
TE5 describes the physical theft of the assembled VEs. By stealing assembled
VEs, the adversary can either destroy or cast ballots. The detection of this
threat event relies on individual voters noticing that they did not receive
their VEs, i.e., the detection probability increases with every voter notifying
municipal authorities.
The integrity and availability of ER is crucial for the Swiss RPV. By targeting
ERs, substantial damage can be inflicted on data integrity, but also on trust in
local authorities and can undermine voters’ confidence. Requirements for EV
systems can serve as a reference for process improvements [10].
4.2 Risk Assessment Phase B
The Delivery phase Bis a black-box. The internal processes of the Swiss Post
(SP) are not publicly available. When using the dedicated SP service to dispatch
and deliver VEs, the VSC design must adhere to special layout rules to facili-
tate automatic batch processing [14]. The special layout of VEs could simplify
identification of VEs, but requires an adversary to achieve partial control of the
SP routing system. To achieve such control, a hostile individual or group can
create an Advanced Persistent Threat (APT) within the SP and from there,
e.g., identify VEs according to specific attributes and re-route identified VEs, or
attempt to delay the delivery deliberately.
TE6 describes the re-routing of VEs. This TE requires adversarial access to
internal SP systems and the capability to covertly manipulate the postal
routing. A re-routing may require a co-conspiring postal employee because
re-routing a large number of VEs could raise suspicion. Assuming a successful
re-routing of VEs, the adversary is offered multiple options: Either to destroy
the VEs, or open, modify, and re-cast them (cf. Figure 5).
TE7 describes the theft of VEs from voter letterboxes before successful retrieval
by the recipient voter. In contrast to TE5, TE7 requires the adversary to
steal from individual letterboxes, not only at a single location. Similar to
TE5, detection increases with every voter noticing the absence of VEs.
12 C. Killer, B. Stiller
Phase Bis characterized by the trust placed in one large entity, the SP.
Thus, the effort and detection probability of TE6 can only be analyzed with ad-
ditional information or access to internal SP systems, operations, and processes.
Generally, however, an insider can achieve a low detection with moderate effort.
4.3 Risk Assessment Casting Phase C
TE8 describes the theft of VEs from the municipal letterbox. As shown in Figure
5, stolen VEs can either be destroyed or opened and modified.
TE9 describes the re-routing of VEs (before delivery to the municipality). Sim-
ilar to TE8, re-routing offers two different options: either the adversary can
decide to destroy the VEs, or open, modify, and re-cast. Similar to TE6, a co-
conspiring postal employee is crucial, since delivering a large amount of VEs
to a different location than the authorities may alarm an honest employee.
TE10 describes the casting of stolen or forged VEs. An adversary can attempt
to cast stolen and modified or forged artifacts to influence the voting result.
The interdependence among TEs is visualized in Figure 5.
In official logs provided by the municipalities, there is no differentiation be-
tween channels Iand II , both count as delivered by the SP. Even though 90%
of votes are cast through Iand II , keeping I I I remains crucial: Multiple chan-
nels strengthen confidence in results because it enables cross-channel comparison
with statistical methods.
4.4 Risk Assessment Storage Phase D
TE11 describes TEs originating from physical storage security. Depending on
the municipality, one or Nemployees have access to the cast VEs. The
access to VEs offers similar options as presented in Figure 5. Since most
municipalities do not log the amount of incoming VEs, the destruction of
VEs can remain undetected.
The physical access to the ballots stored allows an adversary to either destroy
VEs, modify them, or open VEs and break ballot secrecy. As past incidents show
[4], access control to the stored VEs is a again a question of trust.
4.5 Risk Assessment Tallying Phase E
TE12 describes the risk of manipulation during tallying. According to [23], over
10% of ballots cast in Switzerland are electronically counted. In 2014, sample
checks identified errors in these counting mechanisms and concluded that e-
Counting is neither more exact nor more secure than manual counting [23].
The manipulation of e-Counting requires an adversary to write targeted
malware to influence the counting mechanism in his favor.
The Swiss Postal Voting Process and its System and Security Analysis 13
Fig. 5. Threat Event Interdependencies
TE13 describes the possible manipulation of the final tally. Some cantons use
proprietary software to handle vote transmission from municipalities to the
cantonal EO [18].An adversary with access to these tools can tamper with
the final tally. Since the manual tallying process produces logs published on
a municipal level, large discrepancies can be detected by attentive observers.
However, a sophisticated adversary can anticipate that and tamper with all
digital traces to further obfuscate detection. Hence, the risk increases when
PBs were exclusively counted electronically, without any redundancy from
manual counting.
The tallying phase Ebuilds on the integrity of each and every individual
member of the municipal Election Offices (EO). The distribution of trust builds
the cornerstone of the Swiss RPV system.
4.6 Risk Assessment Validation Phase F and Phase G
TE14 describes the prematurely initiated destruction of the PBs and VSCs.
The destruction of PBs and VSCs before validation by the FC makes full
recounts impossible, which already occurred in 2011 [2].
Since the validation finalizes and validates the official result within Phase
G, a recount is no longer an option. Also, PBs are now irrelevant, since legal
appeals are impossible at this point.
14 C. Killer, B. Stiller
5 Conclusions
The Swiss postal voting system is highly successful, because substantial trust is
placed in third parties, which includes a wide range of governmental authorities,
state-owned companies, and various private companies and suppliers, and the
individual voter. The current Remote Postal Voting (RPV) system is inherently
built on external suppliers and trusted relationships among all parties involved.
For a regular citizen, the current process is hard to decipher. Thus, this paper
provides a coherent insight into the Postal Voting Process Flow (PVPF) and
identifies its weaknesses as well as strengths with practical examples.
The main advantage of the current RPV system is its physical decentraliza-
tion, which is undercut by using centralized information systems to administer
or transfer crucial data (e.g., Electoral Registers (ER) or Web-based assistance
tools to transmit votes). Many aspects regarding the ER, assistance tools, the
Voter Signature Card (VSC), or the physical storage of voting envelopes offer
room for improvements from a security perspective.
The deployment of a Remote Electronic Voting (REV) system potentially
decreases the necessary amount of trust placed in institutions and people, shifting
trust to verifiable processes instead [29]. As this work showed, assessing the risks
of the Swiss RPV system is reliant on the specific process across governmental
entities. This work identified crucial Threat Events (TE) and showed that the
system cannot serve as a suitable reference for electoral processes [15].
Furthermore, the Swiss federal structure leads to fragmented processes across
jurisdictional barriers, from federal to cantonal, down to municipal authorities.
The real-world deployment of the threat events identified requires a group of hos-
tile individuals with specific knowledge. In small municipalities, authorities and
citizens are intertwined and manipulations would either be not widely effective
or detected rather swiftly. In large municipalities or large cities, processes are
secured. Releasing an attack would require substantial effort from an attacker.
Hence, an attack on the RPV is most likely to be successful in medium-sized
municipalities, e.g., where processes have not yet adapted to the larger size of
the formerly smaller municipality.
Apparently, federal laws are not complete yet in guiding the deployment of
secure e-Counting tools [23]. Thus, the compilation of an open and transparent
list of all the electronic tools in use in the current PV flow can help to identify
further threat events and enable the design of mitigation measures to handle risks
better. Further, the Risk Assessment (RA) needs to be extended and ultimately
applied to full real-world processes of cantons. In turn, TEs identified can be
assessed in more detail and improvements can be provided to act as a comparison
to EV systems.
Acknowledgements
The authors would like to thank Anina Sax, Annina Zimmerli, Dr. Christian
Folini, Marco Sandmeier, and Dr. Benedikt van Spyk for their valuable input.
The Swiss Postal Voting Process and its System and Security Analysis 15
This paper was supported partially by (a) the University of Zurich UZH, Switzer-
land and (b) the European Union’s Horizon 2020 Research and Innovation Pro-
gram under Grant Agreement No. 830927, the Concordia project.
References
1. Andersen, K., Medaglia, R., Vatrapu, R., Zinner Henriksen, H., Gauld, R.: The
Forgotten Promise of E-Government Maturity: Assessing Responsiveness in the
Digital Public Sector. Government Information Quarterly, Vol. 28, Issue 4 , October
2011, pp. 439–445
2. Berner Zeitung: Stimmzettel fehlen, Nachz¨ahlung ¨uber Motorfahrzeugsteuern ist
gef¨ahrdet, August 2011, [Online] http://pvpf.ch/bz-pb, last visit July 9, 2019
3. uhlmann, M.: Schweiz am Sonntag, Aargauer Zeitung: Das Stimmcouvert per Post
verschicken - ein Gratisangebot, das viele Aargauer ausschlagen . [Online] http:
//pvpf.ch/az-ve, February 2016, last visit July 9, 2019
4. Bumbacher, B.: Neue Z¨urcher Zeitung: Hauswart f¨alscht aus Frust Stimmzettel bei
Gemeindewahl, October 2005, [Online] http://pvpf.ch/nzzfraud, last visit July
9, 2019
5. Computer Security Division, Information Technology Laboratory: Minimum Secu-
rity Requirements for Federal Information and Information Systems. (FIPS PUB
200), US Department of Commerce, NIST, March 2006
6. Der Regierungsrat des Kantons Aargau: 101 -Bundesverfassung der Schweiz-
erischen Eidgenossenschaft vom 18. April 1999 (Stand am 23. September 2018).
[Online] http://pvpf.ch/bv, last visit July 9, 2019
7. Der Regierungsrat des Kantons Aargau: 131.111 - Verordnung zum Gesetz ¨uber die
politischen Rechte (VGPR) in Kraft seit 01.01.2013, Beschlussdatum: 30.05.2012).
[Online] http://pvpf.ch/vgpr, last visit July 9, 2019
8. Der Schweizerische Bundesrat: 161.1 Bundesgesetz ¨uber die politischen Rechte
(BPR) (Stand am 1. November 2015). [Online] http://pvpf.ch/bpr, last visit
July 9, 2019
9. Der Schweizerische Bundesrat: 161.11 Verordnung ¨uber die politischen Rechte
(VPR) (Stand am 15. Januar 2014), [Online] http://pvpf.ch/vpr, last visit July
9, 2019
10. Die Schweizerische Bundeskanzlei: Anforderungskatalog Druckereien fur Vote
´electronique, [Online] http://pvpf.ch/bkreq, last visit July 9, 2019
11. Die Schweizerische Bundeskanzlei: Vote ´electronique, [Online] http://pvpf.ch/ve,
last visit July 9, 2019
12. Die Schweizerische Bundeskanzlei: ¨
Anderung des Bundesgesetzes ¨uber die politis-
chen Recht). Erl¨auternder Bericht zur Vernehmlassung, December 2018
13. Die Schweizerische Post AG: Das Briefrestmonopol - Finanzierungspfeiler f¨ur die
Grundversorgung, [Online] http://pvpf.ch/spmon, last visit July 9, 2019
14. Die Schweizerische Post AG: Factsheet, Briefe Wahl- und Abstimmungssendung,
[Online] http://pvpf.ch/spfs, last visit July 9, 2019
15. E-Voting-Moratorium: Initiativtext, [Online] http://pvpf.ch/evmor, last visit July
9, 2019
16. Gr¨uhnenfelder, P.: Neue Z¨urcher Zeitung: Digitale Demokratie verlangt Pio-
niergeist, September 2015, [Online] http://pvpf.ch/nzzpio, last visit July 9, 2019
17. Kantonales Wahlb¨uro Aargau: Wahlen und Abstimmungen, Checkliste Allgemeine
Arbeiten (Rahmenorganisation), November 2017
16 C. Killer, B. Stiller
18. Killer, C., Stiller, B.: A Flow Analysis of Today’s Swiss Postal Voting Process and
a Respective Security Scrutiny. IfI Technical Report No. 2019-02, Department of
Informatics IfI, University of Zurich, April 2019
19. Krimmer, R., Triessnig, S., Volkamer, M.: The Development of Remote E-Voting
Around the World: A Review of Roads and Directions. First International Joint
Conference on Electronic Voting and Identity (E-VOTE ID 2007). Bochum, Ger-
many, October 2008 , pp. 1–15
20. Luechinger, S., Rosinger, M., Stutzer, A.: The Impact of Postal Voting on Partic-
ipation: Evidence for Switzerland. Swiss Political Science Review , January 2007,
pp. 167–202
21. Milic, T., McArdle, M., Serd¨ult, U.: Haltungen und Bed¨urfnisse der Schweizer
Bev¨olkerung zu E-Voting = Attitudes of Swiss Citizens Towards the Generalisation
of E-Voting. Studienbericht, Aarau, September 2016
22. Pammet, H. Jon and Goodman, Nicole: Consultation and Evaluation Practices
in the Implementation in the Implementation of Internet Voting in Canada and
Europe. Research Study, November, 2013
23. Parlamentarische Verwaltungskontrolle (PVK): Elektronische Ausz¨ahlung von
Stimmen (E-Counting) Bericht der PVK zuhanden der Gesch¨aftspr¨ufungs-
kommission des Nationalrates, Februar 2017
24. Pauchard O. - Swissinfo: Tausende beim Z¨ahlen der Wahlzettel. [Online] http:
//pvpf.ch/swi, October 2003, last visit July 9, 2019
25. Regenscheid, A., Hastings, N.: A Threat Analysis on VOCAVA Voting Systems.
Threat Analysis, US Department of Commerce, NIST, December 2008
26. Regenscheid, A., Hastings, N.: Guide for Conducting Risk Assessments. NIST Spe-
cial Publication 800-30, US Department of Commerce, NIST, September 2012
27. Sandmeier, M.: Stadtschreiber und Leiter Stadtkanzlei Baden, February 22, 2019.
Personal Conversation, Stadtkanzlei, Baden
28. Sax, A.: Leiterin Wahlen und Abstimmungen, February 20, 2019. Personal Con-
versation, Staatskanzlei, Generalsekretariat, Kanton Aargau, Regierungsgeb¨aude,
Aarau
29. Serd¨ult, U., Dubuis, E., Glaser, A.: Elektronischer versus brieflicher Stimmkanal
im Vergleich. ¨
Uberpr¨ufbarkeit, Sicherheit und Qualit¨at der Stimmabgabe. Jusletter
IT, September, 2017
30. Smith, R.: Implications of Changes to Voting Channels in Australia. Research
Report comm. by the Electoral Regulation Research Network, December, 2018
31. van Spyk, B.: Vizestaatssekret¨ar Kanton St. Gallen, February 27, 2019. Personal
Conversation, Staatskanzlei, Recht und Legistik, Regierungsgeb¨aude St. Gallen
32. Staatssekretariat f¨ur Wirtschaft SECO, Schweiz: Nationale E-Government Studie
2019 . [Online] http://pvpf.ch/egov19, last visit July 9, 2019, March 2019
33. Stine, K., Kissel, R., Barker, W.C., Fahlsing, J., Gulick, J.: Volume I: Guide for
Mapping Types of Information and Information Systems to Security Categories.
NIST Special Publication 800-60, Vol. I, Rev. 1, US Department of Commerce,
NIST, August 2008
34. Stine, K., Kissel, R., Barker, W.C., Lee, A., Fahlsing, J.: Volume II: Appendices
to Guide for Mapping Types of Information and Information Systems to Security
Categories. NIST Special Publication 800-60, Vol. II, Rev. 1, US Department of
Commerce, NIST, August 2008
35. Verein eCH: eCH-Standards. [Online] http://pvpf.ch/ech, last visit July 9, 2019

Supplementary resource (1)

... For eligible voters, the current process is hard to decipher and impossible to verify. [12] provides (i) an overview of the Swiss RPV system, (ii) a detailed insight into the process flow, and (iii) a respective risk assessment. This analysis and risk assessment in [12] provides critical Threat Events (TE) emerging during various stages of the voting process. ...
... [12] provides (i) an overview of the Swiss RPV system, (ii) a detailed insight into the process flow, and (iii) a respective risk assessment. This analysis and risk assessment in [12] provides critical Threat Events (TE) emerging during various stages of the voting process. The current RPV system offers benefits, too, such as its physical decentralization and the distribution of trust. ...
... Since PV as of Definition 6 includes the resistance against conspiracies of any size, such an assumption is practically infeasible in many practical cases e.g., Proverum applied to the currently deployed Swiss RPV, since many External Suppliers are trusted, as well as municipal authorities, employees and citizens counting paper ballots [12]. On the one hand, the underlying trust model of DLs and BCs depends on the consensus of these networks i.e., on the architectural choices made. ...
Preprint
Full-text available
Trust in electoral processes is fundamental for democracies. Further, the identity management of citizen data is crucial, because final tallies cannot be guaranteed without the assurance that every final vote was cast by an eligible voter. In order to establish a basis for a hybrid public verifiability of voting, this work (1) introduces Proverum, an approach combining a private environment based on private permissioned Distributed Ledgers with a public environment based on public Blockchains, (2) describes the application of the Proverum architecture to the Swiss Remote Postal Voting system, mitigating threats present in the current system, and (3) addresses successfully the decentralized identity management in a federalistic state.
... REV requires less physical labor at different stages of the voting process and infrastructure [33], if designed securely and correctly, it allows for tallying results faster, in contrast to traditional, paper-based voting systems. REV reduces the threat events of traditional remote voting systems (e.g., potential threat events emerging during tallying, or the manipulation of votes sent by postal mail [31]). However, the complexity and technical operations of a distributed high-security system introduce new challenges, e.g., trade-offs between verifiability, transparency of the process, and voters' privacy. ...
... A crucial aspect of REV systems is considering the voter devices and the property of Cast-as-Intended Verifiability [31]. A compromised voter client presents various threat events caused by malware (i) blocking vote broadcasting (and not posting to the DPBB), (ii) changing the vote selection (iii) breaking ballot secrecy by leaking the voter's choice and identity. ...
Conference Paper
Full-text available
Remote Electronic Voting (REV) systems allow voters to cast their votes in an uncontrolled, distributed environment. At the same time, the REV system must provide ballot privacy and verifiability of the final tally. Research has proposed REV schemes offering ballot privacy based on computational intractability assumptions, but only a few provide Unconditional Privacy (UP). Therefore, this work proposes Æternum, a REV system with a voting scheme providing UP. Æternum does not require trust in a central authority, nor does it assume computational intractability of an underlying mathematical problem to provide UP. To satisfy UP's minimal trust assumptions, AEternum uses a permissioned Distributed Ledger (DL), that forms a decentralized network of permissioned nodes, which serve as a transparent, tamper-proof Decentralized Public Bulletin Board (DPBB).
Article
Full-text available
Many countries are forging ahead with convenient balloting methods, in particular electronic and postal voting, in order to re-engage voters. In this paper, we test whether thencost reductions with postal voting increase turnout. The empirical analysis is based on annewly collected data set on the introduction of postal voting in Swiss cantons. We take advantage of the unique fact that voting by mail was introduced at different times across cantons. This allows identifying the impact of postal voting on turnout, independent of time,nissue and canton specific effects. The estimated average effect on turnout is roughly 4.1npercentage points for an average turnout of 43 percent between 1970 and 2005.
Conference Paper
Full-text available
Democracy and elections have more than 2.500 years of tradition. Technology has always influenced and shaped the ways elections are held. Since the emergence of the Internet there has been the idea of conducting remote electronic elections. In this paper we reviewed 104 elections with a remote e-voting possibility based on research articles, working papers and also on press releases. We analyzed the cases with respect to the level where they take place, technology, using multiple channels, the size of the election and the provider of the system. Our findings show that while remote e-voting has arrived on the regional level and in organizations for binding elections, on the national level it is a very rare phenomenon. Further paper based elections are here to stay; most binding elections used remote e-voting in addition to the paper channel. Interestingly, providers of e-voting systems are usually only operating in their own territory, as out-of-country operations are very rare. In the long run, for remote e-voting to become a reality of the masses, a lot has to be done. The high number of excluded cases shows that not only documentation is scarce but also the knowledge of the effects of e-voting is rare as most cases are not following simple experimental designs used elsewhere.
Article
Assessing e-government responsiveness is one of the major gaps in the currently dominant e-government maturity models. While we have a relatively large pool of models focusing on technological and organizational integration from a supply side perspective, measures of responsiveness of e-government systems from a user perspective are still lacking. Replicating a study from New Zealand and Australia, this study explores the response time and quality of e-mail response in Danish local and central governments (N=175). Despite that Denmark is high ranking in international benchmark studies, we find that one third of central government agencies did not respond at all, and close to 80% of the ministries provided none or incomplete answers. Local government responds faster and provides answers that are more complete and accurate than those provided by central government. Implications for e-government are discussed.
Article
Many countries are forging ahead with convenient balloting methods, in particular electronic and postal voting, in order to re-engage voters. In this paper, we test whether the cost reductions with postal voting increase turnout. The empirical analysis is based on a newly collected data set on the introduction of postal voting in Swiss cantons. We take advantage of the unique fact that voting by mail was introduced at different times across cantons. This allows identifying the impact of postal voting on turnout, independent of time, issue and canton specific effects. The estimated average effect on turnout is roughly 4.1 percentage points for an average turnout of 43 percent between 1970 and 2005.
Consultation and evaluation practices in the implementation in the implementation of Internet Voting in Canada and Europe
  • H J Pammet
  • N Goodman
Pammet, H. Jon and Goodman, Nicole: Consultation and Evaluation Practices in the Implementation in the Implementation of Internet Voting in Canada and Europe. Research Study, November, 2013
Guide for conducting risk assessments. NIST Special Publication 800-30
  • A Regenscheid
  • N Hastings
Regenscheid, A., Hastings, N.: Guide for Conducting Risk Assessments. NIST Special Publication 800-30, US Department of Commerce, NIST, September 2012
Neue Zürcher Zeitung: Digitale Demokratie verlangt Pioniergeist
  • P Grühnenfelder
Grühnenfelder, P.: Neue Zürcher Zeitung: Digitale Demokratie verlangt Pioniergeist, September 2015, [Online] http://pvpf.ch/nzzpio, last visit July 9, 2019