From Platform Power to Policy Drift: Techno-Oligarchy, Electoral Polarization, and Authoritarian Agenda-Setting Across Election Cycles

This study examines techno-oligarchic drift, a process by which electoral debates and policy agendas increasingly favor platform-based data and algorithmic oversight, co-evolving with partisan polarization. Hence, drawing on polarization, agenda-setting and party organization scholarship, we explore: (RQ1) whether techno-oligarchy salience has risen over the past two national elections; (RQ2) has this discourse becomes more polarized across parties, topics, and audience communities; and (RQ3) do post-election policy agendas shift toward control-oriented governance rationalized by platform/data imperatives. To answer these questions we compiled a rich, multi-source corpus that combines party manifestos, speeches/rallies, official social posts, YouTube transcripts, mainstream/digital news, fact-checks, and post-election artifacts (bills, executive orders, regulatory advisories, budget speeches). Our measurement relies on a vectionary pipeline that integrates dictionary lexicons, embedding models and expert validation, implemented in a fully cross-lingual framework. Seed lexicons for platform power, data extractivism, content moderation, surveillance/rights, competition/interoperability, and electoral microtargeting are semantically expanded within language (fastText/word2vec and transformer neighbors), aligned across languages, and iteratively pruned via coder adjudication and inter-rater checks. We estimate a Salience Index (length-normalized, source-weighted), stance/moral-frame classifiers (promote/oppose/neutral; rights-security trade-offs), and polarization metrics: topic segregation (Louvain modularity on party/media × topic bipartite graphs), valence asymmetry, and audience echo modularity (retweet/quote/reply networks). A Governance Drift Index links campaign frames to enacted/attempted policies using semantic similarity weighted by regulatory specificity markers. Identification combines pre/post event studies around campaign milestones, and measurement audits (split-sample, leave-one-source-out); as-if exogenous platform outages/policy windows serve as instruments in sensitivity analyses. Preliminary evidence indicates rising techno-oligarchy salience, increasing topic/community segregation, and measurable campaign→policy alignment post-election. Substantively, we recast polarization as a governance technology that gates which techno-governance proposals become legitimate, and we formalize platform power as narrative infrastructure with agenda-setting capital. Methodologically, we contribute a replicable cross-lingual computational pipeline that improves construct validity over static dictionaries and enables comparative diagnostics beyond the Global North.