arXiv CSopen access

A Global Readiness and Sovereignty Capability Model for Post-Quantum Cryptography Migration

Research Article

M. A. Bouke

A Global Readiness and Sovereignty Capability Model for Post-Quantum Cryptography Migration Mohamed Aly Bouke

#1,*

1 Centre for Intelligent Cloud Computing, CoE for Advanced Cloud,

Faculty of Information Science and Technology, Multimedia University, Jalan Ayer Keroh Lama, Bukit Beruang, 75450, Melaka, Malaysia *[email protected] Research Article, September 17, 2026

arXiv:2609.18477v1 [cs.CR] 16 Sep 2026

Abstract Cryptographic dependence predates the quantum era, but the migration to post-quantum cryptography (PQC) opens a rare window to reshape it, because the algorithms, implementations, hardware, and standards adopted now can lock in dependence or sovereignty for decades. This paper introduces the Readiness– Sovereignty Capability Model (RSCM), a national measurement model that operationalizes PQC readiness together with cryptographic sovereignty, which current maturity models score only as readiness and the sovereignty literature defines without measuring. RSCM decomposes sovereignty into three distinct constructs, indigenous cryptographic capacity, indigenous post-quantum control, and external dependency, and certifies a post-quantum maker only through a gate requiring demonstrated, institutionally sustained creation in at least one core layer, whether design, implementation, or validation. Applying it to fiftyseven documented cryptographic actors coded from cited public evidence, and testing that coding with an independent second coder, a plausible-state bootstrap, and convergent-validity checks, we find that twenty countries clear the gate, fifteen as full-stack makers and five as research makers, eleven hold strong general capacity without post-quantum control, one is a ready adopter, and twenty-five are dependent. The gate cells show substantial weighted agreement, a quadratic-weighted kappa of 0.71, and the maker classification is stable in its core though uncertain at the threshold. Readiness tracks independent cyber indices at rank correlations up to 0.70, while post-quantum creation shows no significant correlation with the commitment index, a rank correlation of only 0.22 that separates control from readiness. The paper contributes the framework, the evidence-graded assessment, and policy directions for building indigenous quantum-safe capacity. Keywords: post-quantum cryptography; cryptographic sovereignty; quantum-safe migration; crypto-agility; digital sovereignty.

1

Introduction

The standardization of module-lattice and hash-based algorithms by the United States National Institute of Standards and Technology (NIST) in 2024 turned the migration to post-quantum cryptography (PQC) from a research concern into a national program [1–3]. The migration is urgent because encrypted traffic captured today can be stored and decrypted once a cryptographically relevant quantum computer exists, a threat model known as harvest now and decrypt later. Several governments have responded with binding timelines and inventory obligations [4–7]. Assessing national preparedness has so far meant measuring readiness, that is the presence of a strategy, a cryptographic inventory, standards adoption, deployment, and skills. A second question receives far less structured atten-

tion, namely whether the algorithms, vendors, and hardware that carry a national migration are controlled domestically or supplied from abroad. This second question is not new to the quantum era, since a country’s classical cryptographic stack already rests to varying degrees on foreign algorithms, vendors, and hardware, and cryptographic dependence long predates post-quantum cryptography. What the transition changes is not the existence of that dependence but the occasion to re-decide it, because a country can mandate and deploy PQC while depending entirely on foreign standardized algorithms, foreign vendors, and foreign hardware security modules. This distinction between readiness and sovereignty is invisible to a one-dimensional readiness score, yet it determines whether a migration produces durable autonomy or only relocates dependence into the quantum era.

Page 1 of 25

Research Article

M. A. Bouke

The stakes of getting this distinction right are high and long-lived, and they follow from path dependence. Cryptography underpins every other digital service, so a dependency at the cryptographic layer is a dependency of the entire digital country, and the confidentiality lifetimes at risk, which for government, health, and identity records run to decades, mean that a choice made during this transition will bind a country well beyond the tenure of those who make it. Because every country is rebuilding its cryptographic stack at once and against near deadlines, the algorithms, implementations, hardware, vendors, standards, and skills a country puts in place now are likely to remain the foundation of its quantum-safe stack for decades, and recovering sovereignty once that foundation has set is slower and more costly than building it into the migration. The transition is therefore not the origin of cryptographic sovereignty but a rare window in which countries reshape it for a generation, and measuring only readiness during that window can credit a rapid but externally supplied migration as progress while missing the dependence it entails, which is what the second axis is designed to detect. Three gaps motivate this work. First, published maturity models we reviewed assess the readiness of products or organizations rather than nations, and do not score sovereignty [8, 9]. Second, the idea of cryptographic sovereignty is now discussed in policy and industry writing, but in the sources we found it is defined rhetorically rather than operationalized into a measured, reproducible national instrument [10, 11]. Third, the assessment literature concentrates on a small set of Western and East Asian leaders, while Africa and the Arab world, whose constraints differ sharply, are covered by journalism and policy briefs rather than systematic analysis [12, 13]. We address these gaps with RSCM, which scores readiness alongside a decomposition of cryptographic sovereignty into three distinct constructs, general indigenous cryptographic capacity, indigenous post-quantum control, and external dependency, so that a country’s classical capability is never mistaken for control of the post-quantum layer. Sovereignty over the post-quantum layer is established through a maker gate that requires demonstrated indigenous post-quantum technical capability in at least one core layer of design, implementation, or validation, sustained institutionally rather than as an isolated academic result. RSCM is operationalized through a coded rubric anchored to cited public evidence, so that scores are checkable rather than assertive. We position RSCM against product and organizational maturity models, against the broad and non-reproducible sovereignty indices that have begun to appear, and against qualitative strategy surveys, and we apply it to a purposive sample of fifty-seven doc-

umented cryptographic actors, with deep coverage of the under-studied countries of the Global South, presented as a demonstration of the model rather than a definitive ranking of individual countries. The paper delivers the following contributions. • RSCM, a national model that scores post-quantum readiness alongside a decomposition of cryptographic sovereignty into indigenous capacity, indigenous post-quantum control, and external dependency, with a strict post-quantum maker gate that recognizes demonstrated indigenous creation in at least one core layer of the quantum-safe stack on an institutionally sustained basis, and a capacity-by-creation typology that keeps classical strength from being read as postquantum sovereignty. • A methodology for defensible, uncertainty-aware scoring, comprising a codebook anchored to cited public evidence with an audit trail, decision rules specified in the codebook before coding, per-cell evidence grades with a plausible-state bootstrap, an independent second-coder reliability study, robustness checks, and convergent validation of readiness against independent indices, alongside constructalignment checks for the sovereignty dimensions. • An evidence-graded assessment of fifty-seven documented cryptographic actors worldwide, spanning post-quantum makers, capable but dependent countries, and the under-studied Global South. • A set of policy directions for building indigenous quantum-safe capacity in dependent countries, with a focus on the Global South. The remainder of the paper is organized as follows. Section 2 reviews related work and positions the model. Section 3 gives the technical and policy background. Section 4 presents RSCM. Section 5 describes the methodology. Section 6 reports the assessment. Section 7 discusses the findings. Section 8 presents limitations and threats to validity. Section 9 gives recommendations, and Section 10 concludes.

2

Related Work

2.1

Migration and maturity models

Several maturity models describe how far an organization or a product has progressed toward quantum safety. The Post-Quantum Cryptography Maturity Model grades the readiness of products and services in a supply chain, which suits procurement decisions in which a buyer must judge a vendor [8], and related models grade the internal posture of an organization across people, process, and technol-

Page 2 of 25

Research Article

M. A. Bouke

ogy so that an enterprise can plan its own migration [9]. The technical migration literature complements these with guidance on crypto-agility and on the sequencing of discovery, prioritization, and replacement [14]. These instruments are valuable for the tasks they were built for, but three properties limit them for a national sovereignty assessment. They take the organization or the product as the unit of analysis rather than the country, so they cannot aggregate to a national position. They measure progress on a single readiness dimension, so a laggard and a leader differ only in how far each has advanced along one axis. And they are silent on whether the capability being graded is domestically controlled, so they cannot express the difference between a nation that authors its own algorithms and one that adopts them, which is the very difference that decides whether a national migration builds autonomy or merely relocates dependence. Recent comparative work has moved beyond organization-level maturity models toward institutional migration analysis. El Bizri et al. compare the transition approaches of major cybersecurity agencies and propose an institutional–sectoral convergence framework linking policy, standardization, certification, and risk-management pillars to sectoral readiness [15], and a systematic review of eighteen published lifecycle models scores them on governance, sequencing, risk, and resource management [16]. This work advances comparative migration assessment, but its object remains migration strategy, lifecycle governance, and sectoral readiness rather than national control of the underlying cryptographic capability. The present model keeps the ordinal, evidence-anchored spirit of a maturity model while changing the unit to the country and adding the missing axis, national control of the post-quantum layer.

the right second question, and it supplies the vocabulary of control, dependence, and autonomy that the present model formalizes. Most of it treats the concept rhetorically or narratively, while recent benchmarking work has begun to operationalize related notions of sovereignty [18]. A definition tells a reader what sovereignty is; it does not tell a minister whether a particular national migration has it, still less how one country compares with another. The broad sovereign indices, for their part, aggregate so many heterogeneous factors that a single score cannot be traced back to a checkable cryptographic fact, and they are not reproducible from public evidence. What is missing is a reproducible cross-national instrument that separates general cryptographic capacity, post-quantum-specific control, and external dependency, each grounded in cited public evidence.

2.3

Quantum-safe risk scoring

A third strand scores technical risk rather than national capability. Quantum-safe risk assessment frameworks estimate the exposure of an organization or a system to the quantum threat, weighing the sensitivity and lifetime of data against the maturity of its protection [19], and resilience scores grade the quantum-safety of concrete computing and Internet-of-things systems so that an operator can prioritize remediation. This work is methodologically close to ours in that it too builds an ordinal, criterion-based score, and it is complementary in substance, since a national capability and a system-level exposure are different and both matter. But it operates at the level of an asset or an organization and answers how exposed a system is, not how autonomous a nation is, and it says nothing about who controls the cryptography that a system depends on. A system can be fully patched with the standardized algorithms 2.2 National strategy and sovereignty and score as low-risk while the nation that operates it has A second body of work examines national strategy and the no control over those algorithms at all, which is exactly politics of the quantum transition, and it is where the idea the gap between exposure and sovereignty that the present of cryptographic sovereignty has taken shape. Policy briefs model is built to measure. warn that the migration can deepen international inequality 2.4 Digital sovereignty and supply chains and create new dependence on foreign vendors, placing the least-resourced countries at the greatest risk [12]. In- The present work also connects to a broader literature on dustry writing defines quantum-safe sovereignty as the digital sovereignty and on dependence in critical technolability of a nation to protect its critical systems with tech- ogy supply chains [20]. That literature argues that control nology it controls, standards it trusts, and infrastructure it over foundational digital technologies, from semiconducowns [10]. Strategy surveys narrate government roadmaps tors to cloud infrastructure to cryptography, has become a toward quantum autonomy and situate them in the wider dimension of national power, and that a country dependent contest over technological leadership [11], comparative on foreign suppliers for a foundational technology inherits policy analyses rank national quantum ambitions and repu- both a security risk and a loss of policy autonomy. Cryptations across countries [17], and broader indices that mix tography is a particularly acute case within this literature, cyber, artificial-intelligence, and digital-sovereignty capac- because it is the technology on which the confidentiality ity into a single sovereign score have begun to appear [18]. and integrity of every other digital service ultimately rest, This literature establishes convincingly that sovereignty is so dependence at the cryptographic layer is not confined

Page 3 of 25

Research Article

M. A. Bouke

Table 1: RSCM against the closest instruments along seven dimensions. It alone is national in unit, measures sovereignty, separates indigenous from adopted capability, cites evidence for every score, and is comparable across countries. Instrument

Unit

Sovereignty

Indig. vs adopted

Evidence traceable

Crossnational

Repro.

Maturity models [8, 9]

Product or organization

Not measured

No

Self-assessed

No

Partly

Sovereign indices [18]

Country, broad

Asserted

No

No

Broad

No

Risk scores [19]

System

Not measured

No

Technical

No

Varies

Sovereignty concept [10]

Nation, concept

Defined only

No

No

No

No

Strategy surveys [11]

National strategy

Narrated

No

Qualitative

Partly

No

RSCM

Nation

Decomposed

Yes

Yes, cited

Yes

Computational

to one service but propagates through the entire digital order. The quantum transition sharpens the case further, because it forces every country to rebuild its cryptographic supply chain at once and on a deadline. The contribution of the present model to this literature is to make the abstract concern concrete for one foundational technology, replacing a general argument about dependence with a measured, country-by-country account of who controls the cryptography and who does not.

2.5

Synthesis and positioning

The four strands leave a specific opening. Maturity models and the recent comparative migration frameworks measure readiness, lifecycle governance, or sectoral convergence but not national control of the cryptographic capability, and most are not country-level in unit. The strategy and sovereignty literature raises sovereignty but largely does not measure it. Risk-scoring frameworks measure exposure at the system level. To our knowledge, we found no existing instrument that jointly measures national postquantum readiness together with a decomposed, evidencetraceable account of cryptographic sovereignty, separating general cryptographic capacity, post-quantum-specific creation and control, and external dependency, and gating the maker label on demonstrated, traceable capability. RSCM occupies this opening. It is national, it decomposes sovereignty into capacity, post-quantum control, and external dependency, it gates post-quantum control on demonstrated capability, and it is built on a coded rubric with an audit trail, which together turn sovereignty from a slogan into a measurement. Table 1 sets the model against the closest existing instruments along the dimensions that matter for a national sovereignty assessment.

3

Background

3.1

The harvest-now quantum threat

A cryptographically relevant quantum computer would run Shor’s algorithm at a scale sufficient to compromise the widely deployed public-key schemes based on integer factorization and discrete logarithms, so that the keyestablishment and signature mechanisms built on them would fail together rather than one at a time. Estimates of when such a machine will exist vary widely, from about a decade to considerably longer, and that uncertainty is often read as a reason to wait. The harvest-now-and-decryptlater threat removes that comfort. Because an adversary can capture and store ciphertext today and decrypt it once the machine exists, the effective deadline for any given data item is not the arrival of a quantum computer but that arrival minus the length of time the item must remain confidential. This arithmetic moves the deadline into the present for exactly the data a country most needs to protect. Diplomatic and intelligence records, health and genomic data, and identity and financial records carry confidentiality lifetimes measured in decades, so even a distant quantum computer places their effective protection deadline in the near term, and for the longest-lived secrets it has passed already. The transition must therefore begin well before a quantum computer is demonstrated, which is the reasoning behind the binding national deadlines that several governments have set ahead of any demonstrated capability, and it is why a wait-and-see posture is itself a decision to expose long-lived data. The urgency is also unevenly distributed, since a country holding large archives of long-lived secrets faces a nearer effective deadline than one that does not, which is part of what the readiness axis is built to capture.

3.2

Standardized algorithms and hybrids

The standardized replacements are a module-lattice keyencapsulation mechanism, a module-lattice digital sig-

Page 4 of 25

Research Article

M. A. Bouke

nature algorithm, and a stateless hash-based signature scheme [1–3]. The first two rest on the presumed hardness of structured lattice problems, which are believed to resist both classical and quantum attack, while the third rests only on the security of a hash function and serves as a conservative fallback whose assumptions are the most thoroughly understood. The concentration of the two primary schemes on a single mathematical family is itself a source of systemic risk, because a future advance against structured lattices would affect key establishment and signing at once, and it is one reason several national agencies require the standardized schemes to be deployed in hybrid combination with a classical algorithm or an additional conservative post-quantum scheme, so that a weakness in any single new algorithm does not by itself expose the traffic [21, 22]. Migration is complicated further by the scale of embedded cryptography, which makes discovery and inventory a prerequisite for any transition, by the performance and message-size differences of the new schemes, which stress constrained protocols and devices, and by the long tail of protocols, libraries, and hardware security modules that must be updated in a coordinated sequence rather than all at once. A distinct response to the same threat should be set aside at the outset, since quantum key distribution establishes keys through a physical channel rather than by replacing an algorithm, requires specialized optical hardware, is limited in range and topology, and protects key exchange rather than the broad landscape of signatures and authenticated protocols the standardized algorithms cover. Several countries in this study operate quantum key distribution pilots while having no post-quantum algorithm deployment, so treating the two as equivalent would credit a narrow physical link as a national migration; the assessment therefore never scores a quantum key distribution deployment as post-quantum deployment, a rule applied uniformly alongside the separation of announced plans from operational capability.

3.3

The national mandate landscape

National responses vary in force, and the variation is the empirical ground on which the readiness axis is built. The United States couples binding deadlines for national security systems with annual inventory obligations, and it sets the migration horizon that many other countries echo [4–6]. The European Union frames the transition around technological sovereignty and prepares supporting legislation [23], while national agencies in Germany and France issue their own migration profiles and hybrid requirements that diverge from the pure adoption of the international standard [21, 22]. The United Kingdom, Canada, Australia, Singapore, and India each publish national timelines that track the leading agencies while adopting the

standardized algorithms [24–28]. China pursues a parallel track built on its own commercial cryptography suite and a national post-quantum call [29, 30], which raises the prospect that the world will migrate to more than one algorithm family rather than a single global suite. In the Arab region the United Arab Emirates has issued a binding national encryption policy that requires transition plans and cryptographic discovery [7], the only binding mandate in the regional core. Most other countries, by contrast, have general cybersecurity strategies with no post-quantum content, which is itself a finding rather than a gap in coverage.

3.4

Crypto-agility

Because the standardized algorithms may themselves be revised, and because divergent national suites may coexist, the durable objective of a migration is not the one-time installation of a particular algorithm but crypto-agility, the ability to replace cryptographic primitives repeatedly and at low cost [14]. Crypto-agility depends on an accurate cryptographic inventory, on abstractions that separate applications from the primitives they call, and on governance that can mandate a change across an estate. It also depends, at the national level, on the capacity to detect that a change is needed and to obtain a trusted replacement, which is precisely where sovereignty enters, since an agile country that must still source every replacement from abroad has automated its dependence rather than removed it. The transition is therefore best read as the acquisition of a standing national capability, and the model measures how much of that capability each country controls.

3.5

Standardization and its geography

The algorithms now being deployed emerged from a multiyear public process in which candidate schemes were submitted, analyzed, attacked, and winnowed before a small number were standardized [1–3]. The process was open in that anyone could submit and analyze a candidate, and its openness is the main reason the resulting standards are trusted, but the authority to select and to standardize rested with a single national institute, and the winning designs, while internationally authored, are promulgated as that institute’s standards. The geography of authorship therefore differs from the geography of standardization, and both differ from the geography of adoption. A European academic team may design a scheme, a United States agency may standardize it, and a ministry in the Global South may adopt it, and each of these is a different relationship to the technology. The model’s algorithmic and governance dimensions are built to capture this separation, crediting authorship and standard-setting where they occur rather than assuming that adoption of a good algorithm confers control over it. The parallel Korean and

Page 5 of 25

Research Article

M. A. Bouke

Chinese processes [29–31] show that the selection authority need not be singular, and they raise the prospect that the coming decade will see more than one standardized family in use, which makes the position of an authoring or standard-setting country materially different from that of a pure adopter.

3.6

The economics of the transition

The transition is expensive and its costs fall unevenly, which is part of why sovereignty and readiness diverge. Discovering and cataloging the cryptography embedded across a national estate, re-engineering protocols and applications to accommodate larger keys and signatures, replacing or reconfiguring hardware security modules, and retraining staff together constitute a multi-year program whose cost scales with the size and age of a country’s digital infrastructure. A wealthy country can fund this program and reach high readiness quickly by purchasing foreign products and services, which is precisely the path that produces high readiness with low sovereignty, since buying it can often be the faster and lower-cost short-term route to a completed migration. Building indigenous capacity, by contrast, tends to be slower and more expensive in the short run and pays off only over the longer horizon across which primitives will be replaced repeatedly. The economic logic therefore pushes almost every country toward adoption, and the external-dependency construct measures the long-run cost of that short-run choice, so that a country can weigh the cost of dependence against the cost of capacity rather than only the latter.

4

The RSCM Framework

RSCM scores each country on readiness and on a decomposed sovereignty. Readiness measures how far a national migration has progressed. Sovereignty is not one axis but three, because a single sovereignty score conflates a country’s general cryptographic capacity, its control of the post-quantum layer specifically, and its external dependency, and these are held separately. Readiness and indigenous cryptographic capacity are each the mean of five ordinal dimensions scored zero to four against the codebook of Table 2, in which every level is tied to a required public artifact; indigenous post-quantum control is read from five objective sub-variables; and a country is a post-quantum maker only when it clears a gate on the post-quantum layer, so classical or institutional strength alone never earns the label. Figure 1 places the countries on the plane of capacity against post-quantum creation that the typology uses.

4.1

The readiness dimensions

Readiness has five dimensions, ordered roughly from intention to execution, with the full level anchors given in Table 3. Strategy and mandate captures whether a country has taken a national position at all, and at the top level whether that position is a binding legal instrument with an enforced deadline rather than an aspirational document, because a mandate with teeth is what actually moves an estate. Inventory captures whether the country can discover and catalog its cryptography, which is the prerequisite that every migration guide places first, since an organization cannot replace what it cannot see. Standards adoption captures whether the standardized algorithms or national profiles have been taken up, in guidance, in pilots, or in a binding profile. Deployment captures movement from awareness through pilots to production in critical sectors, and it excludes quantum key distribution, which is a different response to the same threat. Workforce captures the training and skills base, from ad hoc awareness to a sustained national pipeline, because a migration that outruns the people who can execute it stalls. The five are scored separately rather than collapsed, so that a country with a strong strategy but no deployment is distinguished from one with quiet production and no strategy.

4.2

Defining cryptographic sovereignty

Cryptographic sovereignty over the post-quantum transition is the degree to which a country controls the quantumsafe cryptography on which its digital order will depend, meaning the capacity to design, implement, validate, govern, and if necessary repair or replace the post-quantum primitives, implementations, and hardware it deploys. Three facts about the construct organize the rest of the model. First, sovereignty over the transition is not the same as general cryptographic capacity: a country may operate a mature national public-key infrastructure, a domestic vendor base, and an indigenous classical algorithm, and still control none of the post-quantum layer it is migrating to. Second, sovereignty is not autarky: a country that adopts a widely scrutinized foreign standard such as the standardized module-lattice mechanism may retain substantial sovereignty if it controls the implementation, validation, deployment, key management, and replacement of that primitive, so using a foreign standard is not by itself a loss of control. Third, control is layered and the layers can be held separately, so a single sovereignty number conflates distinct capabilities. The model therefore decomposes sovereignty into three distinct constructs rather than one, and reserves the label of post-quantum maker for demonstrated indigenous creation in at least one core layer of the post-quantum stack.

Page 6 of 25

Research Article

M. A. Bouke

4 United States

Indigenous post-quantum creation (Cr)

Ukraine Germany South Korea Russia Japan France Netherlands UAE Canada Switzerland China Taiwan

United Kingdom

3

Brazil Australia Vietnam Austria

post-quantum creation region

Senegal

2

Argentina Mexico

1

0 0

Estonia Israel

India

Singapore Qatar Iran Pakistan Tunisia Lebanon South Africa Saudi Arabia Jordan Iraq New Zealand Cameroon Morocco Malaysia Indonesia Bahrain Ghana Egypt Cote d'Ivoire Namibia Ethiopia Kuwait Oman NigeriaAlgeria Kenya Botswana Tanzania Uganda Mauritius Rwanda

PQC maker

1

2

3

Indigenous cryptographic capacity (ICC, general) Capable, PQC-dependent

Turkey Belgium

Ready adopter

Dependent

4 boundary case

Figure 1: The capacity-by-creation plane, indigenous cryptographic capacity ICC (horizontal) against indigenous post-quantum creation Cr (vertical). The shaded band marks high creation, not the gate; makers sit in it, dependents at the origin, and boundary cases are ringed.

4.3

Decomposing sovereignty

The first construct is indigenous cryptographic capacity (ICC), the general and largely classical-inclusive ability to produce and control cryptography. It is the mean of five layers scored zero to four from cited evidence against the codebook of Table 4: algorithmic research and authorship, the vendor and supply chain, national infrastructure operation, governance and standards authority, and the research and talent base. A country with a national trustedalgorithm project, a domestic vendor base, and a national public-key infrastructure scores high here whether or not any of it is post-quantum, which is the correct reading of a real national capability rather than a claim about the quantum transition. The second construct is indigenous post-quantum control,

the distinctive quantity of the paper, and it credits only the post-quantum layer. It is read from five objective subvariables coded separately rather than from a single judgment, against the codebook of Table 5: whether the country has an indigenous post-quantum primitive, a nationally standardized or state-backed quantum-safe algorithm as distinct from an individual academic co-authorship; a domestic post-quantum implementation it can build and maintain; an independent post-quantum validation and evaluation capability; control of national post-quantum integration and migration; and the agility to modify or replace a deployed primitive on its own, coded as a standing capacity rather than a demonstrated swap. Coding the layers separately, rather than deriving a single indigenous, adopted, or mixed tag, keeps the most consequential judgment diagnosable and addresses the low reliability that a

Page 7 of 25

Research Article

M. A. Bouke

single control tag was found to carry. The third construct is external dependency (ED), coded zero to four, recording how much of the post-quantum stack a country deploys but does not control, from a selfsufficient national stack to a fully foreign one. A capable adopter is expected to score high here, and this dependency, restricted to the components a country deploys without controlling, is the borrowed post-quantum sovereignty that a one-dimensional sovereignty score hides.

4.4

The post-quantum maker gate

A country is a post-quantum maker only if it demonstrates indigenous post-quantum-specific capability in at least one core creation layer, primitive design, implementation, or validation, and that capability is institutionally sustained rather than an isolated academic activity. The gate is deliberately strict on two points where a one-dimensional sovereignty score fails. Classical or institutional strength alone never qualifies, so a national classical algorithm, a governance seat, or a national public-key infrastructure is credited as general capacity but not as post-quantum control, and a country whose only indigenous cryptography is classical and whose post-quantum primitives are adopted does not become a maker on that basis. And algorithm authorship is treated as strong evidence rather than the sole requirement, so a country that does not originate a primitive but can implement, harden, validate, and replace post-quantum cryptography domestically still clears the gate through those layers. Writing b1 , b2 , and b3 for the ordinal levels, each coded from zero to four, of the three creation layers, primitive design, implementation, and validation, and I for the set of countries whose only qualifying creation cell rests on isolated, non-sustained academic activity, a country c is a post-quantum maker exactly when

the gate asserts no more than that for any other type. A research maker that has designed a standardized primitive but builds nothing, and an implementation maker that builds schemes it did not originate, are each genuinely sovereign in one layer and dependent in the others, and the type records exactly which, so a single strong layer earns the maker label without being mistaken for control of the whole stack. Within the makers, the creation profile names one of four types. A full-stack maker both designs and implements post-quantum cryptography (b1 ≥ 3 and b2 ≥ 3); a research maker designs it without a domestic production base (b1 ≥ 3, b2 < 3); an implementation maker implements it without originating a primitive (b1 < 3, b2 ≥ 3); and a validation maker clears the gate through validation alone (b1 < 3, b2 < 3, so b3 ≥ 3). The distinction separates a country that authored a standardized primitive from one that can build post-quantum cryptography it did not design, different kinds of sovereignty that a single maker label would equate. The type is a qualifying threshold rather than a depth ranking, recording which creation layers clear the level-three anchor rather than how deep or broad the national base is, so two countries of the same type can differ widely in scale. Depth within a type is read from the ordinal scores, the indigenous capacity ICC and the creation score Cr, and from the plane of Figure 1, not from the type label.

4.5

Aggregation and measurement

Let ri (i = 1, . . . , 5) be the ordinal levels of a country’s five readiness dimensions, a j ( j = 1, . . . , 5) the levels of its five capacity layers, bk (k = 1, . . . , 5) the levels of its five postquantum-control sub-variables, and e the level of its single external-dependency cell, each coded in {0, 1, 2, 3, 4}. Among the control sub-variables, b1 , b2 , and b3 are the creation layers of primitive design, implementation, and validation, and b4 and b5 are national integration and remaker(c) ⇐⇒ max(b1 , b2 , b3 ) ≥ 3 ∧ c ∈ / I , (1) placeability. Readiness R and indigenous cryptographic capacity ICC are the means of their five levels, indigenous where max(b1 , b2 , b3 ) is the strongest of the three creation post-quantum creation Cr is the mean of the three creation layers, ∧ is logical conjunction, and maker(c) is the result- levels, and external dependency ED is the directly coded ing maker verdict, true or false. level, Requiring a single core layer at level three, rather than several, is deliberate and is the reason the maker label is always paired with a type. The gate marks one boundary, the line between a country that creates nothing in the postquantum layer and one that has demonstrably created in at least one of design, implementation, or validation, and crossing that line is a real change in kind that a country either has or has not made. It is not a claim of broad control across all creation layers; the full-stack maker type specifically denotes demonstrated design and implementation, its stricter condition being level three on both, and

R = 51 ∑5i=1 ri ,

(2)

ICC = 15 ∑5j=1 a j , Cr = 13 (b1 + b2 + b3 ),

(4)

ED = e,

(5)

(3)

each therefore lying in [0, 4]. Cr is a descriptive score used for placement and comparison; the maker gate of (1) does not operate on Cr, but on whether at least one of b1 , b2 , or b3 reaches level three on a sustained institutional basis. The post-quantum-control construct is coded across all five

Page 8 of 25

Research Article

M. A. Bouke

Table 2: Condensed codebook. Each dimension is scored zero to four against cited public evidence.

Construct

Dimension

Level-4 anchor

Readiness Readiness Readiness Readiness Readiness Capacity Capacity Capacity Capacity Capacity PQC control PQC control PQC control PQC control PQC control

R1 Strategy and mandate R2 Inventory R3 Standards adoption R4 Deployment R5 Workforce A1 Algorithmic authorship A2 Vendor and supply chain A3 Infrastructure operation A4 Governance and standards A5 Research and talent base B1 Indigenous PQC primitive B2 PQC implementation B3 PQC validation B4 National PQC integration B5 PQC replaceability

binding legal instrument with a named enforced deadline and critical-sector scope mandated continuous cryptographic inventory across critical infrastructure mandated national post-quantum profiles across government and critical systems broad production across critical sectors (quantum key distribution excluded) sustained pipeline of degree tracks, national certification, and active groups nationally standardized indigenous algorithm, classical or post-quantum domestic supply chain including primitives and hardware sovereign national stack including domestic hardware sets standards adopted by others or authors an international standard sustained national research ecosystem producing talent and IP nationally standardized indigenous post-quantum algorithm full domestic post-quantum implementation and toolchain authoritative national post-quantum evaluation program full sovereign control of national post-quantum integration agility plus authorship or standardization to design a replacement

Table 3: Readiness codebook. Each dimension is scored on the ordinal level whose indicator the cited evidence meets. Dim

0

1

2

3

4

R1 strategy and mandate

no public position

awareness ments only

state-

announced or draft strategy, no deadline

none

guidance recommends inventory

pilot inventory in some agencies

R3 standards adoption

none

standards referenced in guidance

binding legal instrument with an enforced deadline mandated continuous inventory across critical infrastructure mandated national profiles

R4 deployment

none

awareness only

pilots using standardized post-quantum schemes pilots or proofs of concept

adopted strategy with a sector deadline national discovery program or tool available government profiles issued

R2 inventory

production in some critical sectors

R5 workforce

none

ad hoc training or awareness

broad production across critical sectors sustained pipeline of tracks and certification

sub-variables B1 to B5, but the two beyond creation, national integration b4 and replaceability b5 , are reported per country rather than folded into a single control index, because the gate and the maker type read the creation layers and merging integration into the same axis would blur that distinction. Cr is therefore the gate-relevant creation score and not a summary of the whole control construct, and the integration and replaceability layers are read alongside it from the worksheet. The zero-to-four levels are ordinal anchors, and treating them as an equal-interval index for these means is a modeling choice, the same one that composite policy indicators such as the human development and national cyber indices make. Two features keep the choice from driving the results. The primary maker gate is a threshold on a single creation dimension reaching level three, which needs only the ordering of the levels and not their spacing, so it is invariant to any monotone rescaling.

university courses or some professionals

national skills program or agency teams

And for the aggregated scores we test sensitivity directly, re-scoring under alternative monotone spacings of the levels and reporting classification retention in Section 6. The model deliberately does not merge capacity, control, and dependency into a single number, because that merger is exactly the conflation the decomposition exists to undo, and a minimum across dimensions, which would define a country by its weakest layer, was likewise set aside as discarding information about countries strong on several layers and weak on one.

4.6

A capacity-by-creation typology

Each country is placed on the plane of indigenous cryptographic capacity against indigenous post-quantum creation, shown in Figure 1, and the two are not collapsed. A post-quantum maker clears the creation gate on a real capacity base, and its type, full-stack, research, implemen-

Page 9 of 25

Research Article

M. A. Bouke

Table 4: Indigenous cryptographic capacity codebook, the general and classical-inclusive base. Dim

0

1

2

A1 algorithmic authorship

foreign only, no research

general crypto research

active groups

A2 vendor and supply chain

foreign only

local products on foreign primitives

A3 infrastructure operation

foreign-controlled

A4 governance and standards

pure taker

foreign vendors with local integrators foreign infrastructure operated locally observer

A5 research and talent

none

isolated researchers

vendors

research

national infrastructure on foreign hardware active contributor university research groups

3

4

indigenous construction or national competition domestic product lines on foreign primitives national infrastructure with some domestic hardware issues national profiles a national research program or institute

nationally standardized indigenous algorithm domestic supply chain including hardware sovereign stack including domestic hardware authors a standard adopted by others a sustained national ecosystem producing talent and IP

Table 5: Indigenous post-quantum control codebook, crediting only the quantum-safe layer. The final row is the external-dependency construct, scored in the opposite direction. Dim

0

1

2

3

4

B1 indigenous PQC primitive

none

domestic research on foreign schemes

indigenous candidate or national call

nationally standardized indigenous algorithm

B2 PQC implementation

imports only

wraps a foreign library

B3 PQC validation

foreign assurance

academic analysis

B4 national PQC integration

follows a foreign schedule waits for a foreign fix

adopts a foreign profile as-is swaps only vendorshipped

domestic implementation of standard schemes a national evaluation activity tailors a national profile documented agility on a foreign stack

indigenous scheme published or internationally submitted hardened or production implementations national testing or certification issues and enforces a national profile agility with domestic implementation

self-sufficient national stack

largely domestic, minor foreign parts

foreign primitives offset by domestic implementation

foreign primitives and products, some domestic tooling

B5 PQC replaceability

ED external dependency

full domestic toolchain authoritative national program full sovereign control agility with authorship or standardization fully foreignsupplied stack

tation, or validation, is read from which creation layers it class and, for a maker, its type are holds. A capable, post-quantum-dependent country holds   maker (typed) if maker(c), high general capacity but has not built post-quantum con    trol; this is where a country such as Malaysia sits, high capable, PQC-dep. else if ICC ≥ 2, class(c) = (6) on capacity from a decade of national classical cryptog ready adopter else if R ≥ 2,    raphy yet near zero on post-quantum creation because its  dependent otherwise, quantum-safe primitives are adopted, so describing it as  capable but post-quantum-dependent states what the ev full-stack b1 ≥ 3 ∧ b2 ≥ 3,   idence supports and corrects a reading that would call it  research b1 ≥ 3 ∧ b2 < 3, a maker on its classical base. A ready adopter has adtype(c) = (7)  implementation b1 < 3 ∧ b2 ≥ 3,  vanced its migration on readiness without either capacity    or post-quantum control, and a dependent country is low validation b1 < 3 ∧ b2 < 3. across the board; the classes are listed in Table 6. The Here ∧ is logical conjunction, and the cases in (6) are rules reproduce from the published scores. A country’s read from top to bottom, so each “else if” applies only when every row above it does not; “maker (typed)” denotes a maker labeled by its type from (7), and “otherwise” covers a country that clears neither the capacity nor the

Page 10 of 25

Research Article

M. A. Bouke

Table 6: The capacity-by-creation classes and their counts. Only makers clear the post-quantum creation gate.

Class

Countries

Full-stack maker (designs and implements) Research maker (designs) Implementation maker (implements) Validation maker (validates) Capable, post-quantum-dependent Ready adopter Dependent

15 5 0 0 11 1 25

readiness threshold. The capacity and readiness cutoffs are the midpoint of the zero-to-four scale, so every class follows from the maker gate of (1) together with two stated thresholds rather than a discretionary judgment, and the gate verdict with the ICC and readiness scores reconstructs each country’s class. Readiness and external dependency are reported alongside as separate axes, so the typology answers what a country controls without conflating it with how far its migration has progressed or how much of the stack it has borrowed. The maker gate is a substantive threshold rather than a midpoint on a mean. It asks whether a country has reached the level-three creation anchor on at least one of design, implementation, or validation, and whether that capability rests on a sustained national institution rather than a single result, so it separates a demonstrated national post-quantum capability from an isolated academic one. Countries whose only qualifying evidence sits exactly at the anchor, or rests on a single low-confidence source, are reported as boundary cases rather than definitive makers, and the uncertainty analysis of Section 6 quantifies how often each classification survives plausible re-coding.

cell of the codebook, assigns an ordinal level from zero to four by matching the evidence to that level’s written anchor, recording alongside it an evidence-strength grade, a one-line justification, and a citation to the source. All of the judgment in the primary scoring pipeline enters here, at the level of one cell against a fixed rubric, where it is visible and contestable rather than buried inside a composite. How far this stage reproduces across coders is exactly what the independent second coding of Section 6 measures. The second stage is deterministic computation. A scripted pipeline takes the coded worksheet as its only input and applies fixed rules with no further judgment. It aggregates the cell levels into the readiness, capacity, creation, and dependency scores of (2) to (5), resolves the maker gate of (1), assigns each country a class and, for a maker, a type through (6) and (7), and runs the plausible-state bootstrap that yields the retention figures. The scoring and classification rules are a pure function of the coded levels, so identical inputs always yield identical scores and classes, and the plausible-state bootstrap is reproducible under a fixed random seed, so every table, figure, and class in the paper can be regenerated from the released worksheet by running the released scripts. The consequence is a clean division of labor. Subjectivity is localized in the coding, which is cited and audited and whose reproducibility is measured directly; the transformation from codes to results carries none, so a reader who disputes a classification can trace it to the exact cells that produced it and, on re-coding a cell, recompute the consequence deterministically. The companion explorer applies the same second-stage rules in the browser, so an added or edited country is scored identically to the paper.

5.2

5

Methodology

Country selection

To keep the sample from appearing arbitrary, the selection is defined by an explicit sampling frame and a set of inclusion and exclusion criteria specified in advance that are applied uniformly and that another researcher can reconstruct.

Because any capability index invites the objection that its scores are subjective, the methodology is designed so that each value is evidence-anchored, auditable, and open to challenge, and so that the results can be recomputed from the released coding. Figure 2 summarizes the workflow. The frame itself has two parts. The focus population is the set of countries of Africa and the Arab world, taken as the 5.1 The Scoring Pipeline The assessment is produced in two clearly separated stages, membership of the African Union together with the League which confines every subjective judgment to a transparent, of Arab States, which is the region the existing literature cited record and makes the rest of the pipeline exactly neglects and the study is designed to characterize. To this reproducible. The separation is deliberate and is what lets is added a calibration set drawn from a distinct, objectively the paper be at once evidence-based and computationally defined population, namely the countries that are documented post-quantum actors, identified by the external and reproducible. checkable criterion of being recorded as having a national The first stage is human coding. For each country a coder post-quantum program, mandate, or standardization effort runs a structured search of public evidence and, for every in public government-initiative trackers and the leading Page 11 of 25

Research Article

M. A. Bouke

Stage 1 · Human coding

Stage 2 · Deterministic computation Derive reported scores R, ICC, Cr, ED

Cited public evidence

Code against codebook and decision rules

Class from maker verdict, ICC, and R

Coded worksheet 16 dimensions

Checks on coding, scores, and classification

Maker gate on B1–B3 sustained-institution rule

Figure 2: Two-stage RSCM scoring workflow, read left to right. All human judgment in the primary scoring pipeline ends at the coded worksheet; stage two is deterministic, so classes are computed rather than judged and the result is auditable rather than asserted. Blue band, human coding; gray band, deterministic computation.

national-agency publications [4,7,13,21,22,24–28,30,31]. The calibration set provides documented post-quantum actors outside the focus region, so the model can be exercised across a wider range of capability configurations than a region-only study would reach. The frame is defined by these two roles rather than by geography, so the assessment is reported in Section 6 across the six continents, and the Arab-world members of the focus population fall within Africa for the countries of North Africa and within Asia for the countries of the Gulf and the Levant. Within each frame, a country is included if and only if it meets a minimum-evidence threshold, namely that public evidence is sufficient to score a majority of the coded dimensions from cited sources rather than from conservative defaults. The threshold is defined on the availability of evidence rather than on the score a country would receive, so it is not outcome-dependent, but it is not neutral either. Evidence availability is itself correlated with capability and geography, since countries with stronger institutions, more published output, and more English-indexed material are likelier to clear the threshold, and our search was principally in English. The inclusion gate therefore selects toward the more capable and better-documented countries, which narrows coverage of the least-resourced countries rather than tilts the scores of those included, and we read the sample accordingly. The focus population was swept in full against this gate, so every country of the African Union and the League of Arab States with a documented national cryptographic footprint, meaning a national public-key infrastructure or certification authority, a national cryptography or standards agency with published output, documented cryptographic or quantum research, or recorded participation in the international cryptography standards committee, is included. In the tables the countries are grouped by class and, within a class, ordered by indigenous post-quantum creation for readability rather than presented as a ranking of individual countries.

Conversely, a country in the frame that fails that same threshold, typically a low-capacity or conflict-affected country with no public cryptographic footprint, is excluded and assigned to a separate no-evidence category rather than scored, because scoring it would mean fabricating a profile from defaults. Mauritania is an instructive example, since it has a general national cybersecurity strategy but no public cryptographic or post-quantum footprint, so nearly all of its dimensions could be scored only by default; applying the threshold uniformly, it is placed in the no-evidence category rather than in the sample, even though it lies within the focus population. This is the honest reason the regional sample is a subset of its frame rather than a full census, and it is a transparent rule rather than a discretionary choice; the excluded countries are not hidden but named as a category, and extending coverage to them is identified as future work. Both criteria are applied evenhandedly. Every included country, whether in the focus population or the calibration set, is scored with the identical codebook, the identical source hierarchy, and the identical decision rules, so no country receives favorable or unfavorable treatment by virtue of its group. Because the frame and the threshold are stated explicitly, the sample can be reconstructed and audited by anyone with access to the same public record. Reproducibility here has a precise and limited meaning. Releasing the codebook, the coded worksheet, and the scripts makes the assessment computationally reproducible, in that every table and figure can be regenerated from the coded data, and auditable, in that every cell can be traced to its cited source. It does not establish that an independent coder would derive the same levels from the raw evidence, which the moderate exact inter-coder agreement of Section 6 shows is not guaranteed. We claim the first two senses and are explicit that the third, full reproducibility of the coding itself, is not demonstrated.

Page 12 of 25

Research Article

5.3

M. A. Bouke

Data sources and collection

The assessment is built entirely from public evidence, gathered through a structured search conducted for each country in turn and recorded in a per-country worksheet. For every country the search covered the same classes of source, in a fixed order of preference. Primary sources were sought first, namely national laws, regulations, decrees, and official strategy documents; the publications, standards, and tools issued by national cyber-security and cryptography agencies; the membership and contribution records of the international standards bodies; and the technical documentation of national public-key infrastructures and certification authorities. Where primary sources were silent, secondary sources were used and flagged as such, namely reputable technical and industry press, vendor and telecommunications announcements, and the academic output of national institutions and their researchers. Each country’s national cyber agency, national standards body, national public-key infrastructure operator, principal telecommunications operators, and leading universities were checked in this way, and international bodies such as the standards committees and regional consortia were consulted for cross-national signals. Three properties of the collection process bear on the results. Primary evidence was preferred over secondary throughout, and a score that rested on a single secondary source, or on a structural inference such as the likely foreign origin of an undisclosed hardware security module, was flagged in the worksheet rather than presented as established fact. The search was conducted principally in English, with Arabic and French government sources consulted where they were reachable, which is a limitation for exactly the regional core the paper focuses on, since some national material in those languages is not indexed by open search; this limitation biases the dependent countries toward lower scores and is revisited in Section 8. Finally, the evidence is a snapshot taken at a fixed date, because national mandates move quickly, and the worksheet records the date and the source for every cell so that the assessment can be reproduced and re-coded as the field advances. The complete worksheet records the level, its evidence-strength grade, the justification, and the citation for every scored cell.

5.4

Evidence and decision rules

Every dimension is scored against the codebook from at least one cited public source, recorded in the worksheet with a one-line justification, so that each cell can be traced to the artifact that produced it. The sources are the public record a diligent analyst can reach, namely national strategies and regulations, statements and tools published

by national cyber and cryptography agencies, the membership and contribution records of the international standards bodies, vendor and telecommunications announcements, and the research output of national institutions, and where a claim rests on a single source or on an inference it is flagged as such. Three rules are fixed before scoring and applied uniformly. Where no public evidence supports a higher level, the lower level is assigned and flagged as evidence-limited rather than inferred, which biases the assessment toward understating capability. Announcements, plans, and pilots are capped below the operational levels, so that a strategy document or a vendor pilot cannot be recorded as a completed migration. And quantum key distribution is never scored as post-quantum deployment, because it is a different response to the same threat. These rules do most of the work of keeping the scoring disciplined, because the majority of contestable judgments in a fast-moving field concern exactly the distinction between what has been announced and what is operational. Two further distinctions are recorded because they bear on how the scores should be read. The first separates the absence of evidence from evidence of absence, since no public document being found is not the same as a public document establishing that a capability is missing. The worksheet therefore tags each cell at one of three confidence bands, high for a level resting on primary evidence such as a national law, standard, or agency publication, medium for one resting on reputable secondary reporting, and low or evidence-limited for one resting on a single source, a structural inference such as the presumed foreign origin of an undisclosed hardware security module, or a confirmed absence in an incomplete record. The second distinction is that the model measures publicly evidenced capability rather than complete national capability. Where cryptographic work is classified, as it is for several countries with defense or signals-intelligence programs, the public score understates the true capability, and this is stated for the affected countries rather than concealed. The quantity the model estimates is therefore what a country can be shown to control from the public record, which is the quantity a comparative and reproducible assessment can defend, and the evidence-availability bias this introduces, which falls hardest on countries whose national material is not indexed by open search, is revisited among the threats to validity.

5.5

Reliability, robustness, and validity

Reliability rests on the released coding worksheet, which records for every cell the level, its own evidence-strength grade, a justification, and a citation, so that any reader can re-code a country against the same anchors. An indepen-

Page 13 of 25

Research Article

M. A. Bouke

dent second coder re-coded a stratified sample of two hundred and twenty cells across twenty countries, spanning the capacity, post-quantum-control, and external-dependency dimensions and including all of the B1 to B5 control cells that decide the maker gate. We report the agreement in full, the exact agreement, the agreement within one level, and the quadratic-weighted Cohen’s kappa, per construct rather than through a single figure, because within-one agreement is a lenient target on a five-level scale, counting a level-two against a level-three disagreement as agreement, and that is the disagreement the gate is sensitive to. Because the gate is a threshold that turns a one-level cell disagreement into a whole reclassification, we also report the agreement on the maker verdict itself, which is the quantity that bears on the headline count. Robustness is examined in three ways, by re-scoring the aggregated capacity under alternative monotone spacings of the ordinal levels, by testing the maker gate against a shift in its threshold, and by a plausible-state bootstrap over the per-cell evidence grades that perturbs the most judgment-dependent cells, since a single check can leave undisturbed a boundary that another would move. All of these reliability and robustness results are reported in Section 6.

5.6

Coding examples

Four coded examples show how the codebook turns evidence into a score. Each gives the dimension and country, the public evidence, the codebook level that evidence matches, and the resulting score. 1. Strategy and mandate, United Arab Emirates. The evidence is a binding national encryption policy that requires transition plans and cryptographic discovery [7]. It matches the level-four indicator, a binding legal instrument with an enforced deadline, so the score is four. 2. Standards adoption, Saudi Arabia. The evidence is a national cryptographic standard that is operational but classical and treats PQC only as an appendix [34]. It matches the pilot level rather than the mandated-profile level, so the score is two. 3. Infrastructure and dependency, Kenya. The evidence is a national public-key infrastructure whose platform and hardware are supplied by a foreign vendor [35]. It matches the national-infrastructure capacity level, a score of two, while the foreign supply of the platform raises the external-dependency construct.

Validity is addressed by comparison against independent 4. Indigenous post-quantum primitive, Senegal. measures. The readiness axis is compared with the Global The evidence is a code-based scheme its researchers Cybersecurity Index and the National Cyber Security Inauthored and submitted to an international compedex [32, 33], and the capacity and post-quantum-control tition [36, 37]. It matches the indigenous-primitive constructs are checked against objective markers such as level, a creation score of three, but as a single univerthe nationality of hardware security module manufacture, sity result it does not meet the sustained-institution participation status in the relevant international standards requirement of the maker gate. committee, and whether a country’s institutions have designed a standardized algorithm. Because some of these The examples also show the conservative rule at work, markers also inform the coding, the check is one of consince where a higher level would require evidence that struct alignment rather than fully independent validation, no public source provides, the lower level is assigned and and it is reported as such rather than as proof. flagged rather than inferred. Finally, the scores are treated as ordinal levels, and the classification foregrounds the maker gate and the resulting 6 Results classes rather than precise decimal rankings, because the The decomposition separates what a single sovereignty ordinal construction does not support arithmetic on small score conflates. Table 7 reports, for all fifty-seven coundifferences and because the policy conclusions depend on tries, readiness R, indigenous cryptographic capacity ICC, class membership rather than on the exact position within the indigenous post-quantum creation score, external dea class. The maker gate is a threshold on the ordering of a pendency ED, and the class, and Figure 1 shows the plane single creation level, so it depends only on that ordering of capacity against creation. The first result is that caand is invariant to any monotone rescaling of the levels. pacity and post-quantum creation are distinct axes: sevThe codebook, the evidence worksheet, and the scoring eral countries with high general capacity sit near zero on scripts are released as supplementary material so that every post-quantum creation, which is the configuration a onetable and figure can be reproduced and any country re- dimensional score hides and by which strong classical scored against the same anchors. capability could be read as post-quantum control.

Page 14 of 25

Research Article

M. A. Bouke

Table 7: Readiness R, capacity ICC, creation Cr, dependency ED, and class for all fifty-seven countries, grouped by class and, within a class, ordered by creation Cr for readability; the order is not a ranking of countries. Capable abbreviates capable but post-quantum-dependent; a dagger marks a boundary case; full per-cell scores and citations are in the released worksheet.

6.1

Country

R

ICC

Cr

ED Class

Country

R

ICC

Cr

ED Class

United States Germany Ukraine

3.8 3.0 2.0

4.0 3.6 3.4

3.7 3.3 3.3

1 2 1

Full-stack Full-stack Full-stack

Malaysia Ethiopia Singapore

2.2 0.8 2.6

2.8 2.4 1.6

0.3 0.0 1.0

4 4 4

Russia France South Korea Netherlands United Kingdom China Switzerland Taiwan Canada UAE Belgium Austria Japan Turkey† Vietnam† Brazil† Australia† Estonia Argentina Israel India Mexico Iran Qatar Pakistan Saudi Arabia

2.0 3.0 2.6 2.6 3.0

3.8 3.6 3.2 3.0 2.4

3.0 3.0 3.0 3.0 3.0

1 2 2 1 2

Full-stack Full-stack Full-stack Full-stack Full-stack

Senegal Iraq Tunisia Jordan Lebanon

0.4 0.6 1.0 1.2 0.8

1.6 1.8 1.6 1.6 1.6

2.0 0.7 0.7 0.7 0.7

3 4 4 4 4

Capable Capable Ready adopter Dependent Dependent Dependent Dependent Dependent

3.0 2.4 2.0 2.6 3.2 1.8 1.0 2.0 1.4 1.4 1.8 2.4 2.4 0.8 2.4 2.4 0.6 0.8 1.8 0.8 1.8

3.8 3.4 2.8 2.8 2.6 3.4 2.8 3.2 3.6 2.6 2.4 2.0 2.8 2.2 3.4 2.2 2.0 2.4 2.0 2.0 2.4

2.7 2.7 2.7 2.7 2.7 2.3 2.3 3.0 2.3 2.3 2.3 2.3 1.7 1.7 1.3 1.3 1.3 1.0 1.0 1.0 0.7

1 2 2 2 2 2 2 2 3 2 2 3 3 2 3 3 3 4 4 4 4

Full-stack Full-stack Full-stack Full-stack Full-stack Full-stack Full-stack Research Research Research Research Research Capable Capable Capable Capable Capable Capable Capable Capable Capable

South Africa New Zealand Indonesia Egypt Morocco Cameroon Algeria Kenya Bahrain Nigeria Rwanda Mauritius Uganda Cote d’Ivoire Oman Kuwait Ghana Namibia Tanzania Botswana

1.2 1.4 1.6 1.0 1.0 0.2 0.4 0.4 1.0 0.4 0.4 1.0 0.6 0.2 0.6 0.6 0.4 0.4 0.6 0.4

1.6 1.8 1.8 1.8 1.8 0.8 1.4 1.4 1.2 1.2 1.2 1.2 1.2 1.2 1.0 0.8 0.8 0.8 0.8 0.8

0.7 0.3 0.3 0.3 0.3 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent Dependent

The post-quantum makers

Twenty countries clear the post-quantum creation gate on a sustained institutional basis, and the typology does not equate them. Fifteen are full-stack makers that both design and implement post-quantum cryptography, among them the United States, China, Germany, France, South Korea, the United Kingdom, the Netherlands, Switzerland, Russia, Ukraine, Taiwan, Canada, Belgium, Austria, and the United Arab Emirates. The type is a qualifying threshold rather than a ranking of depth, and the full-stack makers span a wide range on both axes. The United States sits at the top with a creation score of 3.7 and a broad national ecosystem, while the United Arab Emirates enters the same type at 2.7 on a much narrower base, its design and implementation layers clearing the gate without matching that scale. The shared label certifies that design and

implementation both reach the anchor, not equal breadth, which the creation score and the capacity axis carry separately. Five are research makers that design or analyze post-quantum cryptography without a domestic production base, Japan together with the boundary cases Australia, Vietnam, Brazil, and Turkey. The remaining two types are empty here. No country is a pure implementation maker, which implements without originating a primitive, and none is a validation maker, which clears the gate through validation alone; both types are defined for completeness because each is a distinct route to control, and their absence, every maker in the sample holding at least a design or a design-and-build profile, is itself informative. Table 8 gives the type, the post-quantum creation score with its uncertainty interval, and the retention of maker status under re-coding.

Page 15 of 25

Research Article

M. A. Bouke

The corrected attribution matters here. The United States is credited for standardization authority and full-stack engineering rather than for authoring the primitives, which international academic teams designed. The United Arab Emirates enters as a full-stack maker on a genuine postquantum basis, on demonstrated design and implementation [38] rather than on its binding national encryption policy, which is credited separately under national integration. Boundary status has an explicit rule that a reader can reconstruct from the worksheet: a maker is a boundary case when it clears the gate on a single creation cell at exactly the level-three anchor, with no second qualifying layer above it. Four makers meet this rule, Turkey, Vietnam, Brazil, and Australia, each carried by an indigenous primitive design and nothing else at level three. This structural fragility is distinct from evidential fragility, which the retention column reports separately, and the two do not coincide. Australia clears on a single cell but that cell rests on high-grade primary evidence, so its bootstrap retention is complete, whereas Turkey, Brazil, and Vietnam clear on a single cell graded medium, so their retention is lower. A boundary flag therefore marks a country that a re-coding of one cell could move, and the retention figure says how likely the evidence makes that move. The gate also excludes a case that a looser reading would admit. Argentina operates a genuine domestic secure-communications system, but its only claimed post-quantum content is a proprietary, undisclosed scheme that has never been published or independently analyzed, so no publicly verifiable post-quantum-specific capability reaches the gate, and Argentina is classed capable but postquantum-dependent rather than a maker. The exclusion is deliberate, since crediting an undisclosed proprietary claim would reintroduce exactly the over-attribution the gate is built to prevent.

6.2

Capacity without post-quantum control

The correction is sharpest for the countries that hold high general cryptographic capacity but have not built postquantum control. Eleven countries are capable but postquantum-dependent: Malaysia, Saudi Arabia, Qatar, Israel, India, Iran, Estonia, Argentina, Mexico, Pakistan, and Ethiopia. Malaysia is the clearest case, with an indigenous cryptographic capacity above the midpoint from a decade of national classical cryptography but adopted post-quantum primitives and near-zero post-quantum creation. The same logic applies to the Gulf countries and to Israel and India, whose strong general and academic bases have not yet produced controlled national post-quantum capability.

Table 8: The twenty post-quantum makers with type, creation score, its ninety-percent interval, and the share of bootstrap draws that retain maker status. A dagger marks a boundary maker.

Country

Type

Cr

Interval

Retention

United States Germany Ukraine Russia France South Korea Netherlands United Kingdom China Switzerland Taiwan Canada UAE Belgium Austria Japan Turkey Vietnam Brazil Australia

Full-stack Full-stack Full-stack Full-stack Full-stack Full-stack Full-stack Full-stack

3.7 3.3 3.3 3.0 3.0 3.0 3.0 3.0

3.67–3.67 3.0–3.67 3.0–3.67 2.67–3.33 2.67–3.33 2.67–3.33 2.67–3.33 2.67–3.33

100% 100% 100% 100% 100% 100% 100% 100%

Full-stack Full-stack Full-stack Full-stack Full-stack Full-stack Full-stack Research Research Research Research Research

2.7 2.33–3.0 2.7 2.33–3.0 2.7 2.33–3.0 2.7 2.33–3.0 2.7 2.33–3.0 2.3 2.0–2.67 2.3 2.0–2.67 3.0 2.67–3.33 2.3 1.67–3.0 2.3 1.67–3.0 2.3 1.67–3.0 2.3 1.67–3.0

100% 100% 100% 100% 100% 100% 100% 100% 92% 87% 87% 100%

6.3

The dependent majority

Twenty-five countries are dependent, low on capacity and on post-quantum creation alike, and one, Singapore, is a ready adopter that has advanced its migration without either capacity or control. Figure 3 maps the classes across the assessed countries.

6.4

Regional distribution

Reporting the assessment across the six continents shows a steep capability gradient that the focus-and-comparator split of the sampling frame conceals. Table 9 gives the mean of each construct and the class mix by continent. Europe and North America lead, with the highest indigenous capacity and post-quantum creation and the lowest external dependency, and between them they hold twelve of the twenty makers. Asia is the widest-spread continent and the pivot of the picture, since its twenty-one countries include six makers in East Asia and the Gulf alongside seven capable but post-quantum-dependent countries and seven dependent ones, so a single continental mean averages over China, Japan, South Korea, and Taiwan at one end and the low-capacity countries of South and West Asia at the other. South America and Oceania are each represented by only two countries, so their per-continent means are reported for completeness and carry no interpretive weight; each pairs one maker at the two-to-three creation

Page 16 of 25

Research Article

M. A. Bouke

Post-quantum maker

Capable, PQC-dependent

Ready adopter

Dependent

Figure 3: Distribution of the classes across the assessed countries; makers concentrate in Europe, North America, and East Asia. Drawn on the equal-area Equal Earth projection [39]; the sample is purposive, not a world census.

boundary with one lower-capability country, and a twocountry average is too thin to place on the gradient. Africa is the floor of the distribution, with the lowest capacity and creation and the highest external dependency of any continent, seventeen of its eighteen countries dependent and none a maker, the one exception being a single capable but post-quantum-dependent country. Two features of the sample bound this reading. The assessed set is a purposive sample of documented cryptographic actors rather than a census, so each continent is represented by its better-documented members. Because evidence availability is associated with institutional capacity, the assessed sample may overrepresent betterdocumented and more capable countries, and the direction and magnitude of the resulting bias cannot be established for countries that were not scored. The gradient is therefore a statement about the assessed countries rather than an unbiased description of the world.

6.5

Uncertainty and robustness

The independent second coding measures how reliably the coding reproduces, and the post-quantum-control cells that decide the gate are its most reliable part. Across the two hundred and twenty re-coded cells, exact agreement was fifty-three percent, agreement within one level ninety percent, and the quadratic-weighted kappa 0.65; the five control cells reached 0.74 and the three gate cells B1 to B3 reached 0.71, with sixty-three percent exact and ninetythree percent within-one agreement, above the capacity

Table 9: Per-continent means of R, ICC, Cr, and ED with the class mix (makers / capable but post-quantum-dependent / ready adopters / dependent), for the fifty-seven countries. Russia and Turkey count in Europe, Mexico in North America.

Continent

R

ICC

Cr

ED

Europe 11 2.2 North America 3 2.3 South America 2 1.3 Asia 21 1.7 Oceania 2 1.9 Africa 18 0.6

3.3 2.9 2.3 2.2 1.9 1.3

2.7 2.6 2.0 1.3 1.3 0.2

1.9 10/1/0/0 2.0 2/1/0/0 2.0 1/1/0/0 3.3 6/7/1/7 3.5 1/0/0/1 3.9 0/1/0/17

n

Classes

cells at 0.39 and the external-dependency cells at 0.58, with Table 10 reporting the full per-construct agreement. Because the gate is a threshold, the decisive quantity is the agreement on the maker verdict itself, and here the two coders agreed for fourteen of the twenty countries and disagreed on six, every disagreement falling among countries whose creation cells sit at the two-to-three threshold rather than in the core. Two of the four formal boundary makers, Vietnam and Brazil, fell below the gate for the second coder while Turkey and Australia held, and four non-makers just under the gate, Malaysia, Israel, Argentina, and Singapore, crossed above it, while the nine makers that clear the gate on more than one cell or on highgrade evidence never moved. The maker classification is therefore reliable in its core and uncertain along the twoto-three threshold in both directions, so the twenty-maker

Page 17 of 25

Research Article

M. A. Bouke

Table 10: Inter-coder reliability from the second coding, by

applies.

construct. Exact is exact agreement, W1 agreement within one level, and κw the quadratic-weighted Cohen’s kappa.

7

Construct

n Exact

All coded cells 220 Cryptographic capacity (A1–A5) 100 Post-quantum control (B1–B5) 100 Maker-gate cells (B1–B3) 60 External dependency (ED) 20

53% 46% 60% 63% 50%

W1

κw

90% 84% 94% 93% 95%

0.65 0.39 0.74 0.71 0.58

count should be read as a robust core of roughly a dozen makers with a contested margin that the formal boundary flag marks only on the maker side. Each cell carries an evidence-strength grade, and its plausible-state set is bounded at one level, with per-grade perturbation probabilities that reproduce the ninety percent overall within-one-level agreement between the two coders. We are cautious about what this calibration can and cannot claim. Within-one agreement is a lenient target, since it counts a level-two against level-three disagreement as agreement, and that is exactly the disagreement the maker gate is sensitive to, so a bootstrap tuned to withinone understates the uncertainty at the gate. The bootstrap is therefore a lower bound on gate uncertainty, and we read its retention figures alongside the direct gate-sensitivity evidence, the boundary makers and the countries that sit just below the gate, rather than as a sufficient account on their own. Over a four-thousand-draw bootstrap across these countries, maker status is retained in ninety-six percent of draws on average, and eighteen of the twenty makers retain the label in at least ninety percent of draws. The full-stack core, the United States, China, Germany, France, South Korea, and the United Kingdom, retains it in every draw. The countries whose status is most sensitive are those sitting just below the creation gate, Estonia, Israel, Argentina, Mexico, and India, which cross into the makers only when a low-confidence creation cell is drawn upward, and the four boundary makers, whose intervals Table 8 reports. Because the maker gate is a threshold on the ordering of a single creation level rather than on the spacing of the scale, it is invariant to any monotone rescaling of the levels, and re-scoring the aggregated capacity under alternative monotone spacings leaves the class of every country unchanged.

Discussion

The assessment supports a single structural reading once sovereignty is decomposed. Control of the post-quantum layer is concentrated in the makers that design, implement, or validate quantum-safe cryptography on a sustained national basis, and it is a narrower set than the countries with strong general cryptographic capacity. The decomposition is the paper’s central correction: a onedimensional sovereignty score merges general capacity, post-quantum control, and external dependency, and it is that merger, not any single weight, that allowed classical capability and governance structure to be read as postquantum sovereignty. Separating the constructs shows that many capable countries have not yet built post-quantum control, that using a foreign standardized primitive is not by itself a loss of sovereignty when a country controls implementation and replacement, and that authorship is one strong route to control among several.

7.1

Case studies

Five cases make the typology concrete, one for each way a country meets or misses the gate. The United States is a full-stack maker that designs, implements, validates, and controls national post-quantum integration, and it standardized the primary algorithms through an open international competition. It is credited for standardization authority and full-stack engineering rather than for authoring the primitives, which international academic teams designed, and the correction sharpens rather than diminishes its position, since it holds every creation and control layer.

The United Arab Emirates is a full-stack maker on a genuine post-quantum basis. It enters through its national research institute, which co-designed a standardized mechanism and produced a post-quantum library with a corresponding hardware implementation integrated into several secure-communication products [38], so its design and implementation layers both reach the level-three anchor. The binding national encryption policy is credited separately under national integration rather than as the basis for the maker classification, which rests on demonstrated postquantum creation, not on the governance and infrastructure that a general sovereignty score would have credited, The bootstrap and the second coding agree that the maker which is the distinction the decomposition is built to draw. core is stable and locate all the movement at the gate Malaysia, by contrast, is capable but post-quantumthreshold. The bootstrap retention nonetheless understates dependent. It holds high general cryptographic capacity the margin, because inter-coder disagreement at the twofrom a decade of national classical cryptography, a nato-three line, which the second coding measures directly, tional trusted-algorithm project, and a domestic research is larger than the evidence-grade perturbation the bootstrap

Page 18 of 25

Research Article

M. A. Bouke

community, yet its post-quantum primitives are adopted and its post-quantum creation is near zero. It is capable but post-quantum-dependent, not a maker, and treating its classical capability as evidence of post-quantum sovereignty is exactly the conflation the decomposition prevents. Belgium and Austria are full-stack makers on a researchled base. Two of the strongest cryptographic-design nations, they both design and implement post-quantum cryptography, Belgium through the authorship of SABER and masked, side-channel-resistant implementations at COSIC, Austria through the co-authorship of the standardized SPHINCS+ and the KaLi lattice coprocessor at TU Graz. They clear the gate on both the design and the implementation layer and are full-stack makers, though their implementation base is research-led rather than a broad national industry, which is the difference between clearing the gate on both creation layers and doing so at the scale of the largest makers. Senegal, finally, is a dependent country with an isolated result. A single university group’s code-based submission is a real academic contribution but not a sustained national capability, so Senegal does not clear the gate and is classed dependent, an illustration of the sustained-institution requirement that keeps an isolated result from being read as national sovereignty.

7.2

Broader implications

The pattern connects the quantum transition to the wider debate on digital sovereignty and supply-chain security. A country that imports its algorithms, its hardware security modules, and its migration tooling places the foundation of its digital order in a supply chain it does not control, and the transition is an occasion to reconsider that dependency in the cryptographic supply chain. The concentration of post-quantum creation in a few makers also raises the prospect of a multi-standard world, in which a country must choose between algorithm families backed by different powers, and there the ability to design, or at least to independently implement and evaluate, a scheme becomes a strategic capability rather than a technical detail. Decomposing sovereignty is intended to make this dimension legible to the ministries that will make procurement and policy decisions over the next decade. The framing has limits that should be stated. It compresses a continuous and multi-layered reality into a small set of classes, and although the creation profile and the externaldependency axis recover much of the within-class magnitude, the class itself remains a discrete kind rather than a point on a continuum. It also treats the layers of control as separable when in practice they interact, and it says nothing about the quality or trustworthiness of a given maker’s

cryptography, only about who controls it. The model measures control and dependency, not cryptographic security, so domestic capability is not treated as more secure and foreign capability as less secure; a well-scrutinized foreign open standard may be more trustworthy than an unproven domestic one, and a country may rationally adopt it for interoperability, cost, assurance, or efficiency. The vocabulary is operational rather than normative, and none of the class labels implies political inferiority. What the model claims is narrow, that ignoring the distinction between general capacity, post-quantum control, and dependency leads a country to mistake a foreign-supplied migration for a sovereign one.

7.3

Scenario and historical parallel

The practical meaning of borrowed post-quantum sovereignty is clearest under a concrete scenario. Suppose that within the coming decade a cryptanalytic advance weakens the structured-lattice family on which the primary standardized algorithms rest. A full-stack maker would respond from a position of control, drawing on its own analysts to assess the advance, its design capacity to propose a revision, and its implementations to reissue software and hardware, and it would have a seat where the corrected standard is decided. A capable but postquantum-dependent country, however strong its general cryptography, would depend on a maker to detect the problem, design the fix, and schedule the correction, and its migration would proceed on the terms that others set. The dependent countries would be furthest back in the queue, reliant on foreign vendors to ship a remediation they could neither evaluate nor accelerate. The scenario is hypothetical, but it is the exact contingency for which post-quantum control is insurance, and it shows why measuring that control is not academic. The pattern has historical parallels in other foundational technologies. Countries that adopted foreign telecommunications, satellite, or industrial-control standards without domestic capacity found that their apparent modernization carried a durable dependence that surfaced whenever the supplying relationship changed. Cryptography differs in degree rather than kind, but the degree matters, because cryptography sits beneath everything else and because the confidentiality lifetimes at stake are unusually long. In this respect the quantum transition follows a familiar pattern, in which countries that use a foundational shift to build domestic capacity retain more autonomy afterward than countries that meet it mainly through procurement, and the resulting difference in dependence persists well beyond the transition.

Page 19 of 25

Research Article

M. A. Bouke

it measures, and by reporting the capacity, creation, and dependency scores separately so that no single number hides We state the principal limitations first, then take the threats a country’s true position. The framing still compresses to validity by type. a multi-layered reality into a small set of classes, which is a deliberate simplification whose cost is discussed in 8.1 Limitations Several limitations bound what the assessment can claim, Section 7. and each is analyzed in detail where it arises. The first A distinct construct-validity question concerns not whether and most consequential is the reliability of the maker clas- the three axes separate but whether each is measured by sification at its boundary. The gate is a threshold, and the right dimensions and levels, and here the framework although the creation cells that feed it are coded reliably, is operationalized and reliability-tested rather than shown an independent second coder moved the maker verdict for to be content-valid. The readiness dimensions are adapted six of twenty re-coded countries, every one of them sitting from the national post-quantum migration and maturity at the two-to-three gate threshold rather than in the core, literature, and the control dimensions follow the workso the twenty-maker count is best read as a robust core ing definition of sovereignty as the capacity to design, of roughly a dozen makers together with a contested mar- implement, validate, integrate, and replace the cryptoggin rather than an exact figure, as Sections 5 and 6 report. raphy a nation deploys, but neither set was established The second is that the central construct, indigenous post- through a formal content-validity procedure such as expert quantum control, has no independent external criterion, so elicitation, a Delphi study, or an empirical dimension rethe evidence for it is internal robustness and discriminant duction. We therefore do not claim that the five readiness separation from readiness rather than construct validity dimensions, the five capacity layers, or the five control in the strict sense. The third is that the capacity axis is sub-variables are exhaustive or non-overlapping, and a the least reliable component of the coding and its scores defensible alternative might promote procurement, budshould be read as indicative. The fourth is that the sam- geting, or crypto-agility to a dimension of its own. The ple is a purposive set of documented actors rather than zero-to-four scale follows the common maturity-model a census, and because evidence availability is correlated convention of five ordered levels, and although its anchors with capability and with English-language indexing, cov- are stated so a reader can judge each step, the number erage is thinner for the least-resourced countries. The of levels and the placement of the boundaries are design fifth, consequential for a framework that presents itself choices rather than empirically derived thresholds, as is the as a measurement instrument, is that its dimensions and equal-interval treatment of the levels in the reported means. ordinal anchors are conceptually grounded and reliably ap- The same holds for the two classification cutoffs beyond plied but not established as content-valid through a formal the gate, the scale midpoint for capacity and readiness, derivation, so the model is best read as a proposed and op- which are stated conventions; only the maker gate at level erationalized instrument rather than a validated one, as the three carries a substantive rationale, the transition from acconstruct-validity discussion below sets out. Two narrower tivity to demonstrated national creation. Establishing the limits complete the list, that the assessment scores publicly content validity of the dimensions and anchors, through exevidenced capability and so understates classified national pert elicitation or empirical dimension analysis on a larger programs, and that it is a snapshot of a fast-moving field. sample, is the most important measurement-development The remaining subsections take the corresponding threats step left open, and until it is done the framework should be to validity in turn. read as a proposed, operationalized, and reliability-tested measurement model rather than a validated instrument. 8.2 Construct validity

8 Limitations and Threats to Validity

The most consequential construct decision is the decomposition itself, the claim that general cryptographic capacity, post-quantum control, and external dependency are distinct constructs rather than one sovereignty axis. Conflating them would let classical capability and governance authority stand in for post-quantum control, and the maker gate exists to hold the post-quantum construct to demonstrated post-quantum capability. We mitigate the residual risk by grounding each dimension in an observable national artifact rather than a latent trait, so that a reader who disputes a dimension can inspect exactly what

8.3

Measurement validity

The assessment rests on public evidence, so an absent verdict indicates that no public source was found rather than that a capability does not exist, and government material in Arabic and French may hold detail that open search does not index. This risk is largest for the dependent countries, and it cuts against them, because a capability that exists but is not publicly documented is scored low and can make a country look more dependent than it is. We do not treat that missing evidence as support for the dependent finding,

Page 20 of 25

Research Article

M. A. Bouke

since absence of evidence is not evidence of dependence. What supports the finding is positive evidence rather than absent evidence: the dependent classifications rest on documented foreign provenance, such as foreign-manufactured hardware security modules, adopted foreign primitives, and nationally operated public-key infrastructure supplied by foreign vendors, so the reading is anchored in what the record shows a country uses rather than in what it fails to show a country builds. The residual risk is that a country with undocumented indigenous capability is placed too low, which we flag for the affected countries rather than resolve. The most judgment-dependent cells are deployment maturity and the post-quantum creation cells that decide the maker gate, and these carry the lowest evidence grades. The plausible-state bootstrap shows that class assignments survive perturbation of these cells, and the sustained-institution requirement of the gate constrains them, but they would firm up with per-sector production data that is not publicly available for most countries.

8.4

External validity, sampling, and time

Beyond the construct-alignment check of Section 6, the readiness axis shows convergent validity against two independent third-party indices that are used nowhere in the coding. Over the forty-nine assessed countries that the National Cyber Security Index [33] covers, Spearman’s rank correlation between readiness and the index is 0.70 at a significance below the one percent level; against the ITU Global Cybersecurity Index [32], a coarser five-tier commitment measure that also covers most of the countries the first index omits, readiness again correlates, at 0.48 and again below the one percent level. The readiness axis therefore tracks two independently constructed measures of national cyber posture. The post-quantum construct behaves differently and in the expected way. Indigenous post-quantum creation correlates with the fuller first index at 0.63, since capable countries tend to score well on both, but it is essentially uncorrelated with the commitment tiers of the second index, at 0.22 and not significant, so post-quantum control is not a restatement of a country’s cyber-security commitment. Its distinctness from readiness rests on this discriminant behavior together with the within-sample separation the model exhibits, where countries at equal readiness diverge sharply in post-quantum creation. Taiwan is absent from both indices, and a few countries are absent from one or the other, and are excluded from the respective correlation. This discriminant check is necessary but not sufficient, and we are explicit about what it does not do. It shows that post-quantum control is not readiness by another name, but it is not an independent confirmation of the post-quantum-

control construct itself, because no external index measures that construct, and the objective markers that would serve, algorithm authorship and standards participation, are the same markers that inform the coding, so comparing against them would be circular. A fully independent external criterion for post-quantum control, such as an audited registry of national quantum-safe production, does not yet exist, and building or obtaining one is the single most valuable next step for validating the central construct rather than only its separation from readiness. The sample is purposive rather than a census, and its inferential role should be read at three levels. It supports the model’s internal structure and discriminant behavior, since the post-quantum axis separates from readiness within the sample and the class structure survives the sensitivity analyses. It characterizes regional structure, since the per-continent means of Table 9 describe a gradient across the assessed countries, read only for the continents with enough countries to average. And it does not support a definitive global ranking of individual countries, with boundary countries in particular reported as boundary cases rather than ranked. Because the sample is purposive and evidence availability is associated with institutional capacity, the assessed regional picture may overrepresent the better-documented countries, and the direction and magnitude of the bias from countries excluded for want of public evidence cannot be established, since those countries are not scored; they are named as a no-evidence category rather than hidden. Finally, the data are a snapshot, and fast-moving national mandates will move some boundary cells over time, which is why the worksheet is released for periodic re-coding rather than presented as a fixed ranking.

9

Recommendations

The findings suggest five directions for dependent countries, and for the Global South in particular, where the readiness and sovereignty deficits are largest. The first four are domestic, and the fifth concerns the international support without which the others may be out of reach for the least-resourced countries. These are policy directions drawn from the assessment rather than results derived from it, and where they compare costs or timing they state expectations that the model does not itself quantify.

9.1

Continental coordination

The single most consequential gap in the regional core is the absence of any coordinating body that sets a migration clock and a cryptographic inventory obligation, which leaves most countries without even a starting point and without the shared urgency that a common deadline creates. A continental mechanism, plausibly built on an existing

Page 21 of 25

Research Article

M. A. Bouke

regional institution rather than a new one, could publish a staged timeline, require a cryptographic inventory in critical sectors, and maintain a common register of quantumvulnerable systems. Its value is not only technical but economic, since a shared inventory requirement written into public procurement would aggregate the demand of many small markets into a signal large enough to attract regional vendors, which no single ministry can generate alone. Coordination is also, we expect, the lowest-cost of the recommendations, because it requires convening and rule-making rather than industrial investment.

9.2

Governance sovereignty first

Because we expect governance sovereignty to be among the lower-cost layers to build and it is the one on which the region already has a foothold, it is the natural first target. Several countries run national certification authorities and issue national profiles, and a handful maintain active research groups, yet almost none contributes to the international standards committees where the primitives and their revisions are decided. Moving from observer to active contributor in those committees would give a country a voice in the decisions it currently only inherits, at a cost measured in expert time rather than capital. Governance sovereignty does not by itself remove dependence at the algorithm and hardware layers, but it converts a country from a silent taker into a participant that can at least contest a bad decision and prepare for the next revision.

9.3

Sovereignty-aware procurement

Procurement is where dependence is quietly locked in, because a national public-key infrastructure delivered and operated by a foreign vendor is easily recorded as a sovereign asset when it is not. Procurement rules should therefore separate the layers, asking not only whether a capability is national in name but whether the algorithm, the vendor, the hardware, and the key material are domestically controlled, and weighting bids accordingly. The external-dependency construct gives such rules a concrete and auditable criterion, and applying it at the point of purchase is expected to be far cheaper than retrofitting sovereignty after a foreign stack is entrenched for a generation, a policy expectation rather than a cost the model quantifies.

9.4

Indigenous research as a lever

funded mandates, procurement preferences for domestically analyzed or authored schemes, and sustained support for the university groups that host the talent, would convert academic output into national capacity. No dependent country will match the largest makers, but the capable but post-quantum-dependent cases demonstrate that even a modest indigenous base changes a country’s position from pure dependence toward partial control, which is the realistic ambition for the coming decade.

9.5

Support and burden sharing

The deadlines that make the transition urgent apply equally to countries that cannot fund an industrial response, and leaving them behind is a shared risk rather than a local one, because insecure links propagate through the interconnected systems that everyone uses. There is therefore a case for international support that is specific rather than rhetorical, including risk audits and inventory assistance for countries that lack the capacity to perform them, shared access to open and independently analyzed post-quantum implementations so that a dependent country is not forced to choose between an unvetted domestic option and a foreign proprietary one, and training programs that build the regional talent base the other recommendations rely on. Such support is most effective when it strengthens indigenous capacity rather than substituting for it, because assistance that arrives as another foreign-supplied stack reproduces the dependence it was meant to relieve. The external-dependency construct can serve donors and recipients alike as a criterion for distinguishing support that builds sovereignty from support that entrenches dependence.

10

Conclusion

We presented RSCM, a national capability model that measures post-quantum readiness alongside a decomposition of cryptographic sovereignty into indigenous capacity, post-quantum control, and external dependency, with a maker gate that recognizes demonstrated indigenous creation in at least one core post-quantum layer on an institutionally sustained basis. Applied to fifty-seven countries, it shows that post-quantum creation is concentrated among makers that are almost all wealthy or longindustrialized, that many capable countries hold strong general cryptography without having built post-quantum control, and that the countries of Africa and the Arab world combine low readiness with post-quantum layers resting on foreign foundations, while a research institute or a sustained national program can carry even a middle-income country through the gate.

The region’s stranded research capacity is its most direct path to genuine sovereignty, and it is being wasted. The authored work already present, such as the Senegalese code-based submission to an international competition and the Ethiopian administration mandated to produce cryptographic products, shows that indigenous capability exists but is disconnected from any national migration. The decomposition is the central correction, keeping classiConnecting these efforts to national programs, through Page 22 of 25

Research Article

M. A. Bouke

cal capability and governance authority from being read as post-quantum sovereignty. A country that measures only how far its migration has progressed can mistake speed for autonomy and emerge more dependent than it entered, and the evidence that post-quantum control is reachable by sustained institution-building, yet absent in several capable countries whose indigenous capacity is unconnected to the post-quantum layer, is what turns the diagnosis into an agenda. Two extensions matter most. The sample can be grown toward a full census, particularly of the lowcapacity countries omitted here for lack of public evidence, and because the data are a snapshot of a fast-moving field, a longitudinal track would show whether the transition entrenches the current concentration of post-quantum control or begins to redistribute it. By measuring sovereignty rather than only readiness, the model makes visible a structure that one-dimensional assessments conceal and offers the assessed countries a concrete basis for deciding how much of their cryptographic future they intend to control.

Data Availability

mercial National Security Algorithm Suite 2.0 (CNSA 2.0),” Cybersecurity Advisory, 9 2022. [Online]. Available: https://www.nsa.gov/Press-Roo m/News-Highlights/Article/Article/3148990/nsa-r eleases-future-quantum-resistant-qr-algorithm-req uirements-for-national-se/ [6] U.S. Office of Management and Budget, “Migrating to Post-Quantum Cryptography (M-23-02),” 2022. [Online]. Available: https://www.whitehouse.gov/w p-content/uploads/2022/11/M-23-02-M-Memo-o n-Migrating-to-Post-Quantum-Cryptography.pdf [7] UAE Cyber Security Council, “National encryption policy and national post-quantum migration programme,” Regulatory instrument, 2025. [Online]. Available: https://u.ae/en/about-the-uae/strategies-i nitiatives-and-awards/policies/cyber-activities/Nati onal-Encryption-Policy [8] PKI Consortium, “Post-quantum cryptography maturity model (pqcmm),” 2025. [Online]. Available: https://pkic.org/wg/pqc/pqcmm/

All project files, code, and the assessment dataset are at [9] CyberSecurity Non-Profit (CSNP), “Quantum readihttps://github.com/drbouke/rscm-explo ness assurance maturity model (qramm),” 2025. rer. An interactive explorer for previewing the tool and [Online]. Available: https://qramm.org/ following updates is at https://drbouke.github [10] QNu Labs, “Quantum-safe sovereignty in 2026,” .io/rscm-explorer/. Industry white paper, 2026. [Online]. Available: https://www.qnulabs.com/blog/quantum-safe-sover References eignty-2026 [1] National Institute of Standards and Technology, “Module-Lattice-Based Key-Encapsulation Mecha- [11] L. Paglieri, A. Bonomi Savignon, F. Scalabrini, and L. Costumato, “Navigating the quantum frontier: Exnism Standard,” NIST, Federal Information Processamining government strategy to the next technologiing Standards Publication NIST FIPS 203, 8 2024. cal revolution,” Transforming Government: People, [2] ——, “Module-Lattice-Based Digital Signature StanProcess and Policy, vol. 19, no. 4, pp. 700–724, 2025. dard,” NIST, Federal Information Processing Stan[12] United Nations University, “International inequality dards Publication NIST FIPS 204, 8 2024. and post-quantum cryptography,” Policy brief, 2026. [3] ——, “Stateless Hash-Based Digital Signature Stan[Online]. Available: https://unu.edu/cpr/policy-brief dard,” NIST, Federal Information Processing Stan/international-inequality-and-post-quantum-crypt dards Publication NIST FIPS 205, 8 2024. ography [4] The White House, “National Security Memorandum [13] GSMA, “Post-quantum government initiatives by on Promoting United States Leadership in Quantum country and region,” 2025. [Online]. Available: Computing While Mitigating Risks to Vulnerable https://www.gsma.com/newsroom/post-quantum-g Cryptographic Systems (NSM-10),” National Seovernment-initiatives-by-country-and-region/ curity Memorandum, 5 2022. [Online]. Available: https://bidenwhitehouse.archives.gov/briefing-roo [14] National Institute of Standards and Technology, “Considerations for Achieving Crypto Agility: Stratem/statements-releases/2022/05/04/national-securit gies and Practices,” 2025. y-memorandum-on-promoting-united-states-leade rship-in-quantum-computing-while-mitigating-ris [15] M. El Bizri, A. M. El-Hajj, L. Sliman, and A. M. ks-to-vulnerable-cryptographic-systems/ Haidar, “Institutional approaches to post-quantum cryptography: A comparative analysis of migration [5] National Security Agency, “Announcing the Com-

Page 23 of 25

Research Article

M. A. Bouke

frameworks,” IEEE Access, vol. 14, pp. 3259–3283, 2026.

post-quantum roadmap,” 2025. [Online]. Available: https://www.cyber.gc.ca/en/guidance/roadmap-mig ration-post-quantum-cryptography-government-c anada-itsm40001

[16] T. D. Williams, “Governance before discovery: A systematic comparative review of post-quantum cryptography (pqc) lifecycle models,” 2026, sSRN [26] Australian Signals Directorate, “Information security working paper. [Online]. Available: https://ssrn.com manual and post-quantum guidance,” 2025. [Online]. /abstract=6833360 Available: https://www.cyber.gov.au/business-gov ernment/asds-cyber-security-frameworks/ism/cybe [17] V. Weber and M. Pericàs Riera, “Quantum r-security-guidelines/guidelines-for-cryptography technologies: Ranking germany’s ambition and international reputation,” German Council on [27] Cyber Security Agency of Singapore, “QuantumForeign Relations (DGAP), DGAP Policy Brief 10, safe handbook and critical information infrastructure 2025. [Online]. Available: https://dgap.org/en/resear milestones,” 2025. [Online]. Available: https: ch/publications/quantum-technologies-ranking-ger //www.csa.gov.sg/resources/publications/quantu manys-ambition-and-international-reputation m-safe-handbook-and-quantum-readiness-index/ [18] B. Mazumdar, “Quantum veil protocol (QVP) global [28] Government of India, “National quantum mission system 2026: Sovereign digital security index post-quantum cryptography roadmap,” 2025. [On(QSSI),” 2026, self-published preprint. line]. Available: https://dst.gov.in/sites/default/files /Report_TaskForce_PQMigration_4Feb26%20(v1). [19] Y. Baseri, V. Chouhan, A. Ghorbani, and A. Chow, pdf “Evaluation framework for quantum security risk assessment: A comprehensive strategy for quantum- [29] State Cryptography Administration of China, safe transition,” Computers & Security, vol. 150, p. “Commercial cryptography standards sm2, sm3 104272, 2024. and sm4,” 2021. [Online]. Available: https: //www.oscca.gov.cn/sca/xxgk/bzgf.shtml [20] L. Floridi, “The fight for digital sovereignty: What it is, and why it matters, especially for the EU,” Phi- [30] Institute of Commercial Cryptography Standards, losophy & Technology, vol. 33, no. 3, pp. 369–378, “Global call for post-quantum cryptographic algo2020. rithms,” 2025. [21] Federal Office for Information Security (BSI), “Post- [31] National Intelligence Service (Republic of Korea), quantum cryptography: Migration recommendations “Winners of the korean post-quantum cryptography and hybrid mandates,” Joint statement, 2025. (kpqc) competition,” 2025. [Online]. Available: [Online]. Available: https://www.bsi.bund.de/Shared https://www.kpqc.or.kr Docs/Downloads/EN/BSI/Crypto/Migration_to_Po [32] International Telecommunication Union, “Global st_Quantum_Cryptography.pdf cybersecurity index 2024,” 2024. [Online]. Available: [22] Agence Nationale de la Sécurité des Systèmes https://www.itu.int/hub/publication/d-hdb-gci-01-2 d’Information, “Anssi views on the post-quantum 024/ transition and 2027 certification milestones,” 2025. [33] e-Governance Academy, “National cyber security [Online]. Available: https://cyber.gouv.fr/en/publica index,” 2026. [Online]. Available: https://ncsi.ega.ee/ tions/anssi-views-post-quantum-cryptography-tra [34] Saudi National Cybersecurity Authority, “National nsition cryptographic standards (ncs),” 2022. [Online]. [23] European Commission, “Quantum europe strategy,” Available: https://nca.gov.sa/en/regulatory-docum 2025. [Online]. Available: https://digital-strategy.ec ents/frameworks-and-standard-list/ncs/ .europa.eu/en/library/quantum-europe-strategy [35] Communications Authority of Kenya and ICT [24] UK National Cyber Security Centre, “Timelines Authority, “Kenya national public key infrastructure,” for migration to post-quantum cryptography,” 2025. 2023. [Online]. Available: https://www.ca.go.ke/nat [Online]. Available: https://www.ncsc.gov.uk/guid ional-public-key-infrastrure ance/pqc-migration-timelines [36] G. Banegas, P. S. L. M. Barreto, B. O. Boidje, P.-L. [25] Canadian Centre for Cyber Security, “Guidance Cárdenas-Bonett, P.-L. Cayrel, G. N. Dione, K. Gaj, on becoming cryptographically agile and the C. T. Gueye et al., “Dags: Key encapsulation from Page 24 of 25

Research Article

M. A. Bouke

dyadic GS codes,” in Selected Areas in Cryptography, 2018. [37] Université Cheikh Anta Diop, “A home-authored code-based key encapsulation submission,” 2018, nIST PQC Round 1 submission. [Online]. Available: https://csrc.nist.gov/projects/post-quantum-cryptog raphy/post-quantum-cryptography-standardization /round-1-submissions [38] Technology Innovation Institute, “Abu dhabi’s TII unveils first post-quantum cryptography library in the UAE,” C and FPGA implementation integrated into secure communication products, 2022. [Online]. Available: https://www.tii.ae/news/abu-dhabis-tii-u nveils-first-post-quantum-cryptography-library-u ae [39] United Nations, “General assembly endorses the equal earth projection to show countries in their true relative size,” UN News, 2026. [Online]. Available: https://news.un.org/en/story/2026/09/1168284

Page 25 of 25