Accounting Scripts and the Politics of Compliance: Understanding Accountants’ Roles in AML through Sentiment and Script Theory

How accountants shaped anti‑money laundering in Canada’s non‑banking sectors

Accountants are frequently discussed as procedural technicians – record‑keepers and number crunchers – but the Cullen Commission proceedings and follow‑up interviews show they are also cognitive and institutional actors whose professional scripts materially shape AML outcomes. Analysis of sworn accountant testimony at the Cullen Commission, combined with interviews of accountants, notaries, regulators and legal actors and machine/deep learning sentiment models, reveals that accountants enact distinct evaluative scripts – rooted in training, ethics codes and organizational practice – that both enable and constrain AML performance across real estate, luxury vehicle sales and gaming.

Accountants as scripted gatekeepers

Script theory helps explain why accountants do what they do when confronted with suspicious activity. Preexisting schemata – professional education, codes of conduct, prior audit experience, and established client relationships – provide baseline expectations about responsibility and acceptable practice. Many accountants’ testimonies express positive sentiment grounded in this foundation: confidence that codes of professional conduct, suspicious transaction reporting (STR) channels and established compliance procedures exist and serve a legitimate role. That positive script supports routine compliance behavior, collaboration with FINTRAC and a belief that the profession has a central role in maintaining financial transparency.

But scripts also incorporate practical boundaries. Assimilation processes – where accountants integrate new regulatory developments, typologies of laundering and operational realities – produce mixed sentiments. Neutral statements in testimony frequently reflect technical descriptions of procedures, limited mandates of provincial bodies, or discussions of monitoring tools and registries. Those neutral scripts underscore pragmatic orientations: accountants aiming to be precise, to avoid overreach, and to preserve client confidentiality where law permits, while still seeking to fulfill statutory obligations.

Accommodation occurs when scripts must shift because new evidence or institutional pressures make existing responses inadequate. Negative sentiments in testimonies and interviews – frustration about fragmented guidance, limited statutory authority, weak enforcement, low rates of industry inspections, and jurisdictional ambiguities – reveal accommodation in progress. Accountants report rational calculations about enforcement odds and detection limits, and express concern that compliance duties may be undermined by unclear definitions (for example, which practitioners are formally “accountants” under PCMLTFA ) or by the practical reality that certain sectors (notably real estate) place primary screening duties on other professions (lawyers, notaries).

Bastian Schwind-Wagner _Follow

"Accountants play a decisive role in AML outcomes because their professional scripts – formed by training, codes, and everyday practice – shape how suspicious transactions are recognized and acted upon. Strengthening clarity in statutory roles, targeted training, and visible enforcement can help align those scripts with public policy goals and reduce regulatory gaps.

However, structural constraints such as jurisdictional ambiguity, uneven inspection regimes, and reliance on client disclosure limit accountants’ effectiveness in high‑risk sectors. Policymakers should address these systemic issues alongside improved professional guidance to make AML regimes more resilient and actionable."

Sentiment patterns and what they mean

Applying machine and deep learning sentiment analysis to the accounting testimonies provided two complementary benefits.

1. Models (CNN and RNN+LSTM performed strongly) recovered systematic evaluative patterns across the small but institutionally important corpus of Cullen Commission statements.
2. When triangulated with semi‑structured interviews, these patterns translated into meaningful script categories rather than isolated affective judgments.

Positive sentiment clustered around adherence to regulatory frameworks and proactive compliance: references to professional conduct rules, CPA involvement with federal AML regimes, STR filing as part of professional duty, and the view that accountants provide valuable data for law enforcement. Neutral sentiment mapped to technical, factual explanations of procedures, limitations of mandate (for example, provincial professional bodies focusing on member resources rather than direct regulation of money laundering), and descriptions of monitoring tools (e.g., registries, analytic approaches in casinos). Negative sentiment concentrated on structural and political constraints: role ambiguity between professions, limited enforcement capacity, under‑reporting concerns, and legislative wording that may leave some facilitators outside statutory obligations.

These results suggest that sentiment is less a measure of individual attitudes than a linguistic trace of how professional scripts are enacted under scrutiny. When accountants present affirmative scripts, they emphasize rule compliance and cooperation. When they present neutral scripts, they foreground procedure and technical clarity. When they present negative scripts, they surface institutional gaps – points where scripts collide with the limits of authority, resources, or law.

Practical barriers that emerged from testimony and interviews

The combined qualitative and computational analysis highlights several recurring barriers that reduce AML effectiveness in non‑banking financial institutions:

• Role and jurisdictional ambiguity. In real estate transactions, for example, lawyers and notaries handle trust accounting and many cash components, so accountants’ direct line of sight into the most vulnerable transaction elements is often limited. Legislative definitions that narrowly define who must report can leave gaps.
• Fragmented guidance and uneven training. Multiple witnesses noted that AML guidance and practice vary by firm size, professional designation, and jurisdiction. Smaller firms may receive little enforcement scrutiny; compliance audits are infrequent; and AML remains a relatively new reporting regime for many professionals.
• Low detection/low enforcement calculus. Several accountants described AML as an “odds game”: authorities cannot audit everyone, and many suspicious transaction reports do not culminate in remediation or visible enforcement outcomes. That reality reduces perceived deterrence and can dampen proactive reporting incentives.
• Conflicts of interest and reliance on client disclosure. Where STR decisions involve client‑sourced narratives, accountants may default to cautious, objective descriptions to avoid over‑accusation – sometimes at the cost of more assertive reporting. The realistic risk that complicit practitioners would not self‑report further complicates reliance on voluntary disclosure.

Implications for policy and practice

Understanding accountants as actors who follow and adapt scripts reframes AML policy responses.

Three directions follow from the findings:

• Clarify statutory roles and reduce jurisdictional gaps. Lawmakers should revisit definitions in AML statutes to ensure key facilitators in high‑risk sectors are consistently covered, and coordinate responsibilities in transactions that involve multiple professionals (e.g., real estate closings).
• Strengthen guidance, training and sectoral outreach. Regulators and professional bodies should expand targeted AML training, practical typology guidance for sector‑specific risks, and accessible tools for identifying red flags – especially for small and medium practices that may lack dedicated compliance teams.
• Increase assurance and enforcement visibility. More frequent, risk‑based compliance reviews or inspections, coupled with public reporting on enforcement activity, would raise the expected cost of non‑compliance and help close the perception gap that “reporting leads nowhere”. Visible enforcement outcomes reinforce professional incentives to follow through on STRs.

Limitations and next steps

The study’s jurisdictional focus on British Columbia and Canada constrains direct generalization; other jurisdictions with different regulatory architectures or professional scopes may display different script dynamics. Sentiment models can misclassify nuance, irony or highly technical phrasing, so computational outputs were triangulated with interview evidence to avoid overinterpretation. Future comparative work across jurisdictions and larger corpora of sworn testimony would help test the generality of the script‑based account and refine automated methods for capturing institutionalized professional discourse.

Concluding observation

Accountants are not merely passive enforcers of AML checklists; they are scripted professionals whose sense‑making, professional boundaries and rational calculations help determine how AML rules are applied in practice. Improving AML performance in non‑banking financial institutions therefore requires policy fixes and enforcement resources, but also a deeper engagement with the professional scripts that govern how accountants interpret obligation, risk and the limits of their authority. Addressing legislative ambiguity, strengthening targeted training and making enforcement more visible will help align scripts with the public policy goal of preventing illicit finance from corrupting legitimate markets.

Dive deeper

• Research ¦ Mark E. Lokanan, Accounting scripts and the politics of compliance: Understanding accountants’ roles in anti-money laundering through sentiment and script theory, Journal of Economic Criminology, Volume 11, 2026, 100207, ISSN 2949-7914, https://doi.org/10.1016/j.jeconc.2026.100207. ¦ Link ¦ licensed under the following terms, with no changes made: CC BY 4.0