Emin Köksal
Research

AI-native economic research

Using AI as the method to ask economic and competition-law questions too large to ask by hand

Say a careful reader needs half an hour for one decision of the Turkish Competition Board. The Board has published 9,995 of them, so reading the lot once would take some 5,000 hours, two and a half working years, before anyone drew a single conclusion. That arithmetic is why the record had never been studied as a whole, and why I rebuilt part of my research practice around AI. The two projects below are the result. The publications underneath show where the questions come from.

Method

AI-powered analysis of legal corpora: regulatory text turned into economic evidence

The method combines three steps:

  1. 01

    Large-scale AI analysis of legal texts

    Treat a national authority’s full decision record as a single queryable dataset, then use AI to extract concepts, parties, holdings, and reasoning patterns at a granularity manual review cannot reach.

  2. 02

    Citation network analysis

    Map how authority, precedent, and reasoning move through a body of decisions over time, turning a flat case list into an intellectual genealogy.

  3. 03

    Economic interpretation

    Use economics (industrial organization, competition theory, regulatory economics) to make sense of what the first two steps surface, and to find the next question.

The combination is what makes the work AI-native rather than AI-flavored: each step depends on the others, and none works as well alone.

Flagship projects

Two projects, one program

Both papers study the Turkish Competition Board. It is the authority I have worked on for twenty years, and it publishes its decisions with their reasoning, which is what makes a full corpus possible. The method should travel to any authority that does the same.

Published· SSRN, 2025

A Quarter-Century Analysis of the Turkish Competition Board’s Decisions

Co-authors · with C. Peker and M. Üyer

We took all 3,369 infringement decisions the Turkish Competition Board published over roughly twenty-five years and analyzed them as one dataset, which nobody had done before. Combining NLP with economic analysis, the paper traces how the Board’s reasoning and rulings moved across industries and decision types. Patterns that would have taken a team months to find by hand showed up in days.

The dataset itself is part of the contribution. Building it opened the door to a research program, including the citation network study below.

Why it matters

  • It shows that a national agency’s full decision record can be treated as a research dataset.
  • It maps how the agency’s practice has moved across industries and over time.
  • And it makes a larger claim. As we put it in the paper: “this research demonstrates the potential of large language models to open up new avenues of inquiry in text-intensive disciplines such as law and economics.”
Published· SSRN, 2026

The Intellectual DNA of the Turkish Competition Board

Co-authors · with C. Peker and M. Üyer

The follow-up to the 2025 paper. An AI model read all 9,995 published decisions of the Turkish Competition Board, roughly thirty years of case law, and extracted the 18,513 citations connecting them. Network analysis did the rest. The paper draws the Board’s “intellectual DNA” as a graph: which decisions carry the case law, how its concepts descend from one another, and where the doctrine changed course.

Methodologically, the paper combines two techniques that competition-law research rarely uses together: AI-assisted extraction of structured information from regulatory text at scale, and network analysis of the resulting citation web. Nothing like it exists for the Turkish setting, and there are few examples anywhere.

What we found

  • The Board’s most influential decision attracts just 10 direct citations. Authority in this case law is spread across many rulings, and no small set of landmarks dominates.
  • Reliance on EU case law follows a U-curve: early dependence, then autonomization, then re-engagement.
  • Doctrine-shifting decisions become the canon within three years. How fast a decision gets cited predicts its staying power.
Selected publications

Where the questions come from

The AI-native work above grows out of twenty years of research on competition policy, network industries, and digital regulation. A curated subset is below; the full list lives on the CV page.

Books & edited volumes

  • 2024

    Eroğlu, M., Finger, M., & Köksal, E. (Eds.). (2024). The Economics and Regulation of Digitalisation: The Case of Türkiye. Routledge.

  • 2024

    Köksal, E., & Bakış, O. (2024). Digitalization of Society: Türkiye Digital Society Index. In The Economics and Regulation of Digitalisation (pp. 136–154). Routledge.

  • 2024

    Finger, M., Köksal, E., & Eroğlu, M. (2024). Setting the Scene. In The Economics and Regulation of Digitalisation (pp. 1–20). Routledge.

Selected peer-reviewed journal articles

  • 2022

    Köksal, E. & Ardıyok, Ş. (2022). Turkish Competition Authority’s First Hub-and-Spoke Cartel Decision. Journal of European Competition Law & Practice, 13(8), 566–570.

  • 2019

    Aydınonat, N. E. & Köksal, E. (2019). Explanatory Value in Context: The Curious Case of Hotelling’s Location Model. The European Journal of the History of Economic Thought, 26(5), 879–910.

  • 2018

    Köksal, E. & Ardıyok, Ş. (2018). Regulatory and Market Disharmony in the Turkish Electricity Industry. Utilities Policy, 55, 90–98.

  • 2018

    Köksal, E. & Ardıyok, Ş. (2018). Diverging Approaches in Europe for the Most Favoured-Customer Clauses: How Turkish Competition Authority’s Decision for the Online Food Ordering Market Contributed. Journal of European Competition Law & Practice, 9(2), 119–123.

  • 2016

    Gürakar, E. & Köksal, E. (2016). Institutional Evolution and Economic Development in Iran and Turkey. Middle East Development Journal, 8(1), 32–64.

  • 2015

    Köksal, E. & Ardıyok, Ş. (2015). Reviewing Regulatory Policy for Broadband in Turkey. Competition and Regulation in Network Industries, 16(4), 354–377.

  • 2011

    Köksal, E. (2011). Network Neutrality and Quality of Service: A Two-Sided Market Analysis. International Journal of Management and Network Economics, 2(1), 39–57.

Selected book chapters

  • 2024

    Köksal, E. & Ak, A. (2024). Neo-Brandeisian Traces in the New E-Commerce Act. On İki Levha. (in Turkish)

  • 2023

    Köksal, E., İkiler, B., & Canbeyli, A. (2023). Law and Economics of the 2021 Retail Decision. On İki Levha. (in Turkish)

  • 2021

    Köksal, E. (2021). Regulation of Fiber and the Internet. In Finger & Eroğlu (Eds.), Regulation of Turkish Network Industries (pp. 383–401). Springer.

Selected reports & working papers

  • 2026

    Köksal, E., Peker, C. & Üyer, M. (2026). The Intellectual DNA of the Turkish Competition Board: Mapping Three Decades of Case Law Through Citation Networks. SSRN.

  • 2025

    Köksal, E., Peker, C. & Üyer, M. (2025). A Quarter-Century Analysis of the Turkish Competition Board’s Decisions: An AI-Supported Examination. SSRN.

  • 2023

    Köksal, E. & Bakış, O. (2023). Türkiye Digital Society Index. Betam Research Note, 23/269.

  • 2022

    Bakış, O., Tetikol-Dalgıç, D. E., Deniz, P., Finger, M., Gümüş, İ., & Köksal, E. (2022). Assessment of a Carbon Tax as a Tool to Decarbonize Turkey’s Energy Supply 2050. IC4R Report Series No. 1.

Full publication list: /cv.

Research interests

Active interests

  • Generative AI and economics: research methods, agentic workflows, AI-assisted analysis of legal and regulatory corpora.

  • Industrial economics & competition policy: competition cases, theory of harm, market definition, cartels, vertical restraints.

  • Platform economics & digital regulation: multi-sided markets, platform governance, digital market regulation.

  • Digitalization: measurement, policy, structural change.

  • Climate economics: carbon pricing, decarbonization, regulation.

Collaborating on AI-native economic or legal research?

If you work on AI methods for legal or regulatory text, on citation networks, on competition-policy research, or on research infrastructure for economists and lawyers, send me a note. There are not many of us yet.