Two-page research brief

MONETARY COMMUNICATION / RESEARCH BRIEF 1 / 2 Palić & Sikić | 17 September 2026 | Independent research; author review pending 2026-09-19.3 CROATIA / JANUARY 2021 - MAY 2026 HNB in a crowded inflation debate. Higher inflation accompanies a smaller institutional share in the primary models. Time controls qualify that finding. Petra Palić & Luka Sikić / Hrvatsko katoličko sveučilište -0.21 percentage points of weighted share per +1 percentage point of inflation Primary weekly model. The 95% interval is 0.11 to 0.31 percentage points lower. The full-sample models contain 268 fitted weeks or 60 months. This is a conditional association, not a causal effect. An inflation comparison: 2% to 10% Primary difference: -1.68 pp. Other variables held fixed; all models include media volume. Points show estimates; lines show 95% intervals. With year effects, the interval includes zero: the association is sensitive to time controls. What is being measured? Inflation relevance selects the items; prominence and potential reach determine their weights. The gap is the baseline share minus the current share. The baseline is a reference, not a target. Same numerator, smaller share. 4 / 100 = 4% | 4 / 200 = 2% Illustrative units: unchanged numerator, doubled denominator. The example does not establish an observed cause. -3 -2 -1 0 Weighted-share difference (percentage points) Primary model Period trends Year effects -1.68 (-2.46 to -0.90) -2.00 (-3.29 to -0.72) -1.08 (-2.63 to 0.47)

MONETARY COMMUNICATION / RESEARCH BRIEF 2 / 2 Palić & Sikić | 17 September 2026 | Independent research; author review pending 2026-09-19.3 THE FULL SAMPLE / PRIMARY CORPUS The whole timeline, in view. January 2021 - May 2026. Monthly weighted visibility and inflation. Separate scales; observed series, not fitted trends. Weighted visibility relates institutional prominence and potential reach to the inflation corpus. Potential reach is not observed readership or understanding. 01 / VISIBILITY TREND 02 / NETWORK POSITION 03 / EXPECTATIONS -0.076 pp per month April 2024 - May 2026 Controlled share trend: 95% interval -0.118 to -0.035 pp per month (25 months). Common sources also give a negative trend: -0.056 pp per month. This is a result for the stated period. Direction uncertain April 2024 - May 2026 The common-source centrality trend is -0.0051 index units per month. Its 95% interval (-0.0129 to 0.0027) includes zero. Lasting network decline is not established. Centrality is not trust. No stable predictive link January 2021 - May 2026 Article-share projections: no significant future horizons with linear weekly assignment; weeks 1-2 with step assignment; month 12 in monthly data, after correction. Controls also matter. Understanding and anchoring are not established by this pattern. PUBLICATION REPRESENTATION UNDERSTANDING For HNB: read publication activity, media representation and audience understanding together. This study measures representation; it complements communication records and audience research. Scope: selected media corpus, not a representative audience sample. Generic references can include the ECB. Earlier-paper numerical discrepancies remain unresolved. Sources: Sections 4.2-4.7; Appendices A12, A23, B2. Interactive evidence 0 6 12 18 1.61% Weighted institutional share (%) 01/2021 2022 2023 2024 2025 05/2026 0 4 8 12 4.9% Inflation in Croatia - HICP , year-on-year (%)