Digital creative clusters in Slovakia need a local critical mass—about 1,000 inhabitants/km²—to become self-sustaining, while arts require less density; research hubs boost local creative activity but siphon talent from neighbouring districts, and manufacturing legacies block digital diversification, so scattershot, place-neutral policies are unlikely to seed enduring multi-segment clusters.

Main Finding

Creative agglomeration in Slovakia is conditional on reaching segment-specific density thresholds and is shaped by asymmetric, cross-district spillovers and industrial legacy. Science acts as a local hub that simultaneously supports within-district digital and artistic activity and extracts creative capacity from neighbouring districts. Digital creative activity requires a higher local density (≈1,000 inhabitants/km²) to become self-sustaining than the arts, and manufacturing path-dependency independently suppresses digital creative specialization. Consequently, place-neutral dispersal policies are unlikely to create self-sustaining multi-segment clusters; strategic reinforcement of districts approaching critical density is needed.

Key Points

• Segmentation matters: digital technology, science, and arts follow different agglomeration dynamics and thresholds.
• Science segment has a dual spatial role:
  • Positive local (within-district) Jacobian effect: it supports co-located digital and artistic activity.
  • Negative regional (between-district) backwash: it tends to draw creative capacity away from neighbouring districts.
• Density thresholds are non-linear and segment-specific:
  • The digital segment is more density-constrained than the arts.
  • A critical mass of about 1,000 inhabitants/km² was identified as a structural precondition for multi-segment cluster formation; districts below this show no evidence of self-sustaining creative agglomeration despite policy inputs.
• Industrial legacy matters separately from density:
  • High manufacturing specialization (path-dependency) acts as an independent suppressor of digital creative specialization.
  • Thus, density and industrial legacy are distinct structural barriers to creative diversification.
• Policy implication: broad, place-neutral creative policies are unlikely to work; instead, concentrate resources on districts near critical density and address manufacturing lock-in.

Data & Methods

• Spatial scale and context:
  • District-level analysis for Slovakia, a semi-peripheral, monocentric economy (capital-region dominance).
  • Creative economy disaggregated into three functionally distinct segments: digital technology, science (research/knowledge production), and arts (cultural/creative practice).
• Key variables:
  • Segment-specific location quotients as measures of local specialization.
  • Population density (inhabitants/km²) as the main density variable.
  • Manufacturing specialization as the indicator of industrial legacy/path-dependency.
• Analytical approach:
  • Spatial Durbin model to estimate both local (within-district) and spatial spillover (between-district) effects.
  • Random forest simulation to explore nonlinearity and interactions between predictors.
  • Piecewise (segmented) regression and Davies tests to formally detect and verify density breakpoints (i.e., critical mass thresholds).
• Identification strategy:
  • Combined spatial econometric and machine-learning diagnostics to establish non-linear density dependence, cross-district spillovers, and independent effects of industrial legacy.
• Interpretation caveats:
  • Results are conditioned on the Slovakian context (semi-peripheral, monocentric); thresholds and magnitudes may differ in polycentric or core economies.
  • The summary does not report time span or dynamic analysis specifics (as not provided in the brief).

Implications for AI Economics

• Threshold-dependent clustering for AI: Like digital creative industries, AI and data-science ecosystems may require local density above a segment-specific critical mass to become self-sustaining; policies should identify and prioritize places near that threshold rather than scatter investment evenly.
• Role of science/research nodes: Research-intensive nodes (universities, labs) can catalyze local AI activity but may also centralize talent and create regional extraction effects; regional coordination is needed to manage positive local spillovers while mitigating negative backwash.
• Industrial legacy as an independent barrier: Regions with strong manufacturing path-dependency may be structurally less able to develop AI specialization; conversion strategies (reskilling, complementary infrastructure, incentives) are necessary in addition to density-focused policies.
• Modeling recommendations for AI economics:
  • Incorporate spatial econometrics (e.g., spatial Durbin models) to capture local and spillover effects.
  • Use machine-learning (random forests) alongside econometric models to detect nonlinearity and interactions.
  • Test formally for breakpoints (piecewise regression and Davies tests) rather than assuming linear relationships between density and specialization.
• Policy design:
  • Targeted, place-sensitive interventions: invest heavily in districts that are close to the identified density threshold to tip them into self-sustaining AI clusters.
  • Strengthen science/knowledge institutions in non-extractive ways (partnerships, satellite labs) to spread benefits regionally.
  • Combine density-focused clustering with active strategies to overcome manufacturing lock-in where relevant (retooling firms, workforce development, upgrading infrastructure).
• Transferability note: The ≈1,000 inhabitants/km² threshold is empirically derived for Slovakia’s digital segment; practitioners should empirically establish local thresholds before applying the same numeric cutoff in other national or regional contexts.