2025: A Year of Automotive Fraud
CEO’s Report

Automotive finance in Canada underwent a quiet but consequential shift in 2025. Fraud, long treated as a manageable operational risk, became a material factor shaping credit quality, portfolio performance, and institutional confidence.

Data from Equifax Canada show a sharp rise in automotive fraud over the past year, driven primarily by application misrepresentation and identity misuse. Much of this activity involved real borrowers using real identities, with income, employment, or financial obligations overstated or selectively omitted. These cases often performed initially, masking risk until economic pressure revealed underlying fragility. Identity-based fraud, including synthetic identities, also continued to evolve, surfacing later and across multiple lenders.

Three dynamics explain the acceleration.

• First, sustained economic pressure reduced household financial resilience. Higher interest rates, inflation, and housing costs widened the gap between credit eligibility and lived reality, increasing the incidence and plausibility of misrepresentation.
• Second, digital origination advanced faster than risk integration. Systems optimized for speed and convenience outpaced identity verification, behavioural analytics, and cross-platform visibility, creating exploitable gaps.
• Third, fragmentation persisted across the ecosystem. Lenders, dealers, and investigators each held partial information, but rarely in a form that could be integrated in time to prevent loss.

The impact of fraud extends beyond charge-offs. Misstated borrower quality distorts underwriting, weakens pricing discipline, and undermines residual value assumptions. In many cases, fraud is absorbed into credit losses rather than identified explicitly, obscuring its true scale and delaying corrective action.

The industry response, while uneven, has strengthened over the year. Lenders moved closer to the point of sale, working more closely with dealers to surface behavioural risk earlier. Digital identity verification, real-time analytics, and targeted dealer training improved detection. Leading institutions elevated fraud prevention from a back-office function to an enterprise risk discipline, with clearer ownership, escalation, and board-level visibility.

The Canadian Lenders Association has supported this shift by working with industry participants to explore compliant, near real-time approaches to sharing bad actor intelligence, particularly at the dealership level. These efforts reflect a growing recognition that repeat and organized fraud cannot be addressed effectively in isolation.

Looking ahead, the risk environment will become more complex. Generative artificial intelligence will lower the cost of document fabrication, synthetic identities will mature further, and cross-border schemes will increase. At the same time, defensive capabilities are improving. Identity orchestration, real-time verification, and collaborative intelligence models are moving from concept to practice.

This report argues that fraud prevention is no longer a narrow control function. It is a strategic capability that affects growth, pricing, inclusion, and trust. Institutions that integrate fraud risk coherently across systems, governance, and leadership will be better positioned to navigate the next phase of automotive finance in Canada.

10 Key Points this Report will examine

1. Fraud moved from the margins to the centre of auto finance risk in 2025.
   What was once treated as a manageable operational issue became a material driver of credit quality, portfolio performance, and institutional confidence.
2. The sharp rise in fraud was driven primarily by application misrepresentation, not identity theft alone.
   Much of the activity involved real borrowers using real identities, overstating income, employment stability, or financial obligations, often performing initially before risk surfaced.
3. Economic pressure materially increased both the incidence and plausibility of fraud.
   Higher interest rates, inflation, and housing costs widened the gap between credit eligibility and lived reality, making misrepresentation more common and harder to distinguish from legitimate volatility.
4. Digital origination outpaced risk integration.
   Systems optimized for speed and convenience evolved faster than identity verification, behavioural analytics, and cross-platform visibility, creating exploitable gaps between systems and stakeholders.
5. Fraud thrived in fragmentation across the ecosystem.
   Lenders saw credit data, dealers saw behaviour, investigators saw outcomes, but these perspectives were rarely integrated in time to prevent loss, even at the leadership level.
6. Fraud distorted decision-making long before losses materialized.
   Misstated borrower quality weakened underwriting discipline, mispriced credit, and undermined residual value assumptions, with many cases absorbed into credit losses rather than identified explicitly as fraud.
7. The industry response strengthened meaningfully over the year.
   Lenders moved closer to the point of sale, increased collaboration with dealers, adopted real-time identity and behavioural tools, and invested in targeted dealer training and governance.
8. Leading institutions elevated fraud prevention to an enterprise risk discipline.
   Clear ownership, defined escalation, integrated data governance, and board-level visibility distinguished more resilient portfolios from those where losses persisted.
9. Collective intelligence is becoming essential to addressing repeat and organized fraud.
   The CLA’s work to explore compliant, near real-time sharing of bad actor intelligence reflects recognition that many fraud patterns only emerge when viewed across institutions.
10. Fraud prevention is now a strategic capability, not a compliance function.
    As generative AI, synthetic identities, and cross-border schemes intensify risk, institutions that prioritize coherence across systems, governance, and leadership will be best positioned to protect growth, inclusion, and trust.

The End of Innocence

For much of its recent history, automotive finance occupied a relatively stable corner of consumer credit. Loss curves were predictable. Fraud existed, but it rarely threatened portfolio integrity. Dealers acted as effective gatekeepers. Physical presence introduced friction, and friction discouraged abuse.

Digitization altered that balance.

Remote applications, automated approvals, and integrated dealer platforms accelerated transactions but diluted scrutiny. Identity, once verified face to face, became a data point. Income, once corroborated, became asserted. Convenience increased. Certainty declined.

Fraud did not exploit technology itself so much as the assumptions embedded within it. The assumption that borrowers would be broadly truthful. That dealers would reliably catch anomalies. That post funding controls would compensate for weaker front end verification.

In 2025, those assumptions were tested. Many did not hold.

Equifax Canada’s findings offer the clearest snapshot of the shift. Automotive fraud rose sharply year over year, increasing by more than 50 percent. Crucially, the majority of cases involved application fraud rather than identity theft alone.

This distinction matters.

First party fraud is more difficult to detect precisely because it involves real people using real identities. The deception is incremental rather than absolute. Income is inflated. Employment is overstated or unstable. Liabilities are selectively omitted. These borrowers often perform initially, giving portfolios the appearance of health until economic pressure exposes the underlying fragility.

Identity based fraud remains a material risk, particularly synthetic identities that blend legitimate and fabricated data. These profiles are patient. They mature across systems, pass traditional checks, and fail later, often across multiple lenders at once.

Geography plays a role. Ontario stands out, reflecting both its market scale and economic strain. But the problem is not regional. It is systemic.

Reported fraud also understates reality. Many cases surface months after origination. Others are absorbed into credit losses rather than classified as fraud. The economic cost is therefore higher, and the signal noisier, than headline figures suggest.

The consequences extend well beyond losses.

Fraud distorts decision making long before defaults materialize. Misstated borrower quality leads to mispriced credit. Approval rates appear strong while underlying risk deteriorates. Residual value assumptions weaken when vehicles disappear, are damaged, or are encumbered by competing liens.

Electric vehicles add further complexity. Higher price points, evolving resale dynamics, and unfamiliar repair economics create new avenues for misrepresentation and exploitation.

Fraud, in short, is not merely a loss event. It is a strategic pollutant. It clouds data, erodes confidence, and quietly undermines the foundations on which automotive finance decisions are made.

Three Reason that Fraud Accelerated in 2025

1. Economic stress

Elevated interest rates, inflation, and housing costs compressed household budgets. For some borrowers, misrepresentation became a means of accessing transportation deemed essential. For organized actors, economic stress increased the plausibility of false narratives.

The past year placed Canadian households under sustained financial pressure. Elevated interest rates raised borrowing costs across the board. Inflation eroded real incomes. Housing costs absorbed a growing share of monthly cash flow, particularly in major urban centres. For many households, resilience thinned to the margins.

Transportation, meanwhile, remained non-negotiable. Work, caregiving, and basic participation in the economy still depend on mobility. Public alternatives are uneven. For millions of Canadians, access to a vehicle is not discretionary. It is infrastructural.

This tension reshaped borrower behaviour. In such conditions, misrepresentation often presents itself not as fraud but as necessity. Income is rounded up. Employment stability is implied rather than proven. Liabilities are postponed into silence. The intent may not be criminal in the traditional sense, but the outcome is identical. Risk is mispriced. Loss is deferred.

Economic stress does more than motivate individual misrepresentation. It supplies credibility to deception.

For organized fraud actors, a strained economy expands the universe of plausible stories. Interrupted employment, gig income, multiple jobs, recent relocations, and non-linear credit histories become common enough to evade scrutiny. What once looked suspicious now appears ordinary.

This is the crucial shift. Economic volatility blurs the boundary between anomaly and norm. Underwriters are trained to spot deviations from expected patterns. When patterns themselves fragment, detection becomes harder.

High interest rates further complicate the picture. As payments rise, the tolerance for error shrinks. Loans that might have survived mild misrepresentation in a low-rate environment now fail quickly. Fraud surfaces faster, but not before capital is deployed.

Importantly, economic stress does not create fraudsters. It lowers friction for fraud.

It compresses time horizons. It reduces patience for documentation. It encourages speed over scrutiny on both sides of the transaction. In such an environment, weak controls are not merely exploited. They are rationalised away.

For lenders and dealers, this dynamic presents a delicate challenge. Tightening controls indiscriminately risks excluding legitimate borrowers already under strain. Loosening them invites exploitation. The solution is not blunt restriction, but precision.

Economic stress demands better risk segmentation, not broader exclusion. It requires deeper verification for higher-risk profiles and faster pathways for well-substantiated ones. It calls for intelligence, not austerity.

The lesson from 2025 is sobering. Fraud does not rise simply because people become dishonest. It rises because systems fail to adapt when economic reality shifts.

Those institutions that understand this will design controls that are empathetic but firm, flexible but disciplined. Those that do not will continue to confuse social stress with credit risk, and pay for it twice.

2. Digital acceleration outpaced risk integration.

Origination systems evolved faster than identity verification, behavioural analytics, and cross-platform visibility. Fraudsters exploited gaps between systems and stakeholders.

Digital transformation delivered exactly what it promised to automotive finance: speed, scale, and convenience. Credit decisions that once took days now take minutes. Dealer desks became digital terminals. Identity became data. Friction disappeared.

So did many of the safeguards.

Origination systems evolved rapidly, optimised for throughput and customer experience. Risk integration did not keep pace. Identity verification, behavioural analytics, device intelligence, and cross-platform visibility were bolted on unevenly, if at all. The result was an ecosystem that moved quickly but saw narrowly.

Fraudsters recognised the asymmetry immediately. They did not need to defeat systems. They only needed to traverse them.

Applications flowed cleanly through individual platforms while inconsistencies accumulated across them. A borrower’s story made sense in isolation but collapsed when viewed holistically. Income claims conflicted with employment patterns. Identity signals diverged subtly across channels. Device behaviour hinted at orchestration rather than coincidence. Yet no single system held the full picture.

The problem was not technology failure. It was architectural sequencing. Speed was prioritised before signal integration. Controls were layered after scale had already been achieved.

Cross-platform visibility remains the critical weakness. Dealer systems, lender origination engines, identity vendors, and bureau data often operate in semi-detached environments. Data moves, but insight does not. Alerts trigger locally rather than systemically. Patterns emerge late, after losses harden.

This fragmentation creates a perverse advantage for fraud. Criminal actors adapt faster than institutions because they face fewer constraints. They test workflows repeatedly, probing for latency, thresholds, and blind spots. Once identified, those gaps are exploited at scale.

Digital acceleration did not cause fraud. But it changed the economics of it.

In a slower, paper-based world, fraud was expensive and risky. In a fast, digitised one, it is cheap, repeatable, and hard to attribute.

Closing this gap will require more than incremental improvements. It will require re-thinking origination architecture itself. Identity, behaviour, device, and credit signals must be orchestrated, not merely appended. Risk must travel at the same speed as approval.

This is no longer a question of tooling. It is a question of design.

Institutions that integrate risk natively into digital workflows will slow fraud without slowing growth. Those that do not will continue to trade speed today for losses tomorrow.

3. Fragmentation persisted

Lenders see credit data. Dealers see behaviour. Investigators see outcomes. Rarely were these views fully integrated. Even one leader is locked out of what other leaders see. 

Fraud thrives in fragmentation. 2025 provided ample opportunity.

In Canada’s automotive finance ecosystem, fragmentation remains the quiet enabler of loss. Lenders see credit files, payment histories, and bureau signals. Dealers see behaviour: urgency, inconsistencies, body language, unusual documentation, pressure tactics. Investigators, repossession agents, and insurers see outcomes: missing vehicles, fabricated repairs, lien stacking, repeat patterns across jurisdictions.

Each perspective is valid. None is sufficient.

Rarely are these views fully integrated. Even more rarely are they shared in real time. What one party flags as suspicious may never reach another until losses are crystallized. In many cases, the warning signs were visible early, just not to the same institution.

More troubling still is the asymmetry at the leadership level. Even within the same transaction chain, one executive may be locked out of what another executive sees. Data silos persist not only between organizations, but within them. Risk teams, credit teams, dealer relations, and fraud units often operate on parallel tracks, governed by different incentives and reporting lines.

Fraud exploits this architecture relentlessly. It moves laterally across institutions while information remains vertical. It thrives in the gaps between responsibility and authority.

The implication is uncomfortable but unavoidable. Fraud is not merely a detection problem. It is a coordination failure.

Until credit data, behavioural intelligence, and investigative outcomes are meaningfully connected, fraud prevention will remain reactive. The industry will continue to learn lessons after the fact, rather than preventing loss before it occurs.

The next phase of automotive finance risk management will not be defined by better tools alone, but by better integration: shared signals, aligned incentives, and leadership visibility across the full lifecycle of a transaction.

Our Industry’s Response

The response was not immediate, but it has been substantive.

As losses accumulated and patterns became harder to dismiss, the industry began to recalibrate. The first shift was practical. Lenders moved closer to the point of sale, recognising that fraud prevention could not remain a back-office function. Lenders attempted to collaborate with dealers. Fraud awareness swam upstream and was embedded earlier in the transaction lifecycle, where behavioural signals are most visible and intervention is cheapest.

Digital identity verification tools gained wider adoption, particularly those capable of operating in real time and across channels. Behavioural analytics began to complement static checks, assessing how applications were completed rather than merely what they contained. Velocity controls, device intelligence, and anomaly detection reduced some of the most obvious abuse. Detection improved not because fraud diminished, but because institutions learned where and how to look.

Dealer training emerged as one of the most effective and least glamorous controls. Staff were trained to recognise inconsistencies, unusual urgency, narrative gaps, and documentation anomalies. These signals rarely prove fraud in isolation, but they often indicate elevated risk. Hiring practices tightened, particularly in finance and insurance functions. Internal audits became more targeted, focusing on process deviations rather than checklist compliance.

The deeper shift, however, was institutional.

Lenders began to reposition themselves not merely as capital providers but as risk partners. Fraud prevention moved upstream. Dealers were no longer expected to absorb operational risk alone, and lenders no longer assumed that post-funding controls would compensate for weak front-end verification. Responsibility became shared rather than deferred.

Institutions that fared best shared common traits. Fraud risk had a clear owner. Escalation protocols were defined and exercised, not improvised. Data governance improved, allowing signals from origination, servicing, and investigation to inform one another. Most importantly, leadership had visibility. Fraud was discussed alongside credit risk, not beneath it.

Metrics were tracked consistently. Trends were analysed rather than explained away. Lessons from individual incidents were institutionalised and fed back into underwriting logic, dealer engagement models, and control design. Controls were stress-tested against emerging scenarios, not merely calibrated to past losses.

A particularly important development has been the Canadian Lenders Association’s work with industry participants to explore mechanisms for sharing bad actor intelligence in a compliant and near real-time fashion. This initiative reflects a growing recognition that many fraud patterns, especially at the dealership level, only become visible when viewed across institutions.

Bad actors rarely confine themselves to a single lender or financing partner. They test processes repeatedly, moving laterally across dealer networks and credit providers, exploiting the fact that each institution typically sees only a fragment of the activity. Individually, these incidents may appear ambiguous or isolated. Collectively, they form a clear and actionable pattern.

Historically, coordination has been constrained by privacy obligations, legal risk, and competitive sensitivities. The CLA’s work focuses on overcoming these barriers through compliant frameworks that allow risk signals to be shared without compromising consumer rights or commercial integrity. The objective is not the creation of blunt blacklists, but shared awareness. Early warning signals, repeat behavioural markers, and anomalous dealership-level activity can be surfaced before losses crystallise.

If implemented effectively, such approaches have the potential to capture a significant portion of repeat and organised bad actor activity, particularly where dealership processes are being systematically exploited. This marks a meaningful shift from isolated defence to collective resilience.

The contrast across the industry remains instructive.

Where fraud controls were treated as compliance obligations, losses persisted. Effort was expended, but insight remained limited. Where fraud prevention was embedded into enterprise risk frameworks and supported by shared intelligence, resilience improved. Detection became earlier. Losses became smaller. Confidence returned.

For the CLA community, the implication is clear. Fraud strategy cannot be delegated downward indefinitely. It must be owned at the highest level, where trade-offs between speed, growth, inclusion, and risk are made deliberately.

If leadership does not own fraud risk, circumstances will. And circumstances, unlike institutions, rarely act in the industry’s long-term interest.

The Next Wave of Risk

Generative artificial intelligence is already lowering the cost and skill required to fabricate convincing financial artefacts. Pay stubs, employment letters, bank statements, and identification documents can now be produced in minutes, tailored to specific underwriting thresholds and adjusted iteratively until they pass automated checks. What once required specialist forgery now demands little more than prompt engineering.

This shift matters because it erodes a foundational assumption of modern credit systems: that documentation, while imperfect, is directionally reliable. In an AI-enabled environment, documents are no longer evidence. They are inputs to be tested.

Synthetic identities will evolve in parallel. Early versions relied on obvious inconsistencies and thin credit files. The next generation will be patient. They will season. They will accumulate data across platforms, build transactional histories, and behave conservatively until leverage is maximized. These identities will not look fraudulent. They will look ordinary.

Cross-border schemes will add another layer of complexity. Fraud no longer respects jurisdictional boundaries. Data, devices, and vehicles move fluidly across borders, while regulatory and enforcement frameworks remain largely national. A vehicle financed in one province may be sold, exported, or re-registered elsewhere before risk systems react.

This convergence of AI, synthetic identity, and cross-border mobility will stretch traditional controls to their limits.

Yet this is not a story of inevitable defeat.

Defensive capabilities are advancing as well. Identity orchestration, once discussed abstractly, is becoming operational. Rather than relying on single-point verification, leading institutions are combining identity signals across lifecycle stages, devices, behaviour, and third-party data. Verification is no longer a moment. It is a process.

Real-time analytics are replacing static rules. Behavioural models now assess how an application is completed, not just what it contains. Velocity checks, device fingerprinting, and network analysis are beginning to expose coordination that would otherwise appear random.

Shared intelligence, long constrained by legal and operational silos, is inching forward. Lenders, dealers, insurers, and investigators are experimenting with controlled data sharing, anonymised pattern recognition, and collaborative alerts. The goal is not universal transparency, but actionable awareness.

Still, the temptation remains to treat fraud as a tooling problem. Add another vendor. Layer another dashboard. Deploy another score.

This approach will fail.

The institutions that succeed will not be those with the most tools, but those with the most coherence. Coherence of data. Coherence of governance. Coherence of accountability.

Fraud systems must speak to credit systems. Dealer intelligence must inform underwriting. Investigative outcomes must flow back into origination logic. Leadership must see across the lifecycle, not just within functions.

In the next phase of automotive finance, resilience will be defined less by detection rates than by integration. The winning institutions will design systems that assume adversarial behaviour and respond holistically, not episodically.

Fraud is evolving into a strategic discipline. Those who recognise this early will not just protect portfolios. They will redefine how trust is engineered in a digital economy.

See you at the Summit on Feb 10th

Automotive Finance Canada 2026 arrives at a critical inflection point. https://financeevents.ca/events/2026-automotive-finance-canada/

If 2025 revealed anything with clarity, it is that automotive fraud is no longer an episodic risk to be managed at the margins. It is a structural feature of modern auto finance, shaped by economic pressure, digital acceleration, and institutional fragmentation. It cannot be wished away, delegated downward, or solved with tools alone.

The year marked the end of plausible deniability. Fraud is no longer something that happens occasionally to well-run portfolios. It is something that happens systematically when speed outpaces integration, when data remains siloed, and when governance lags complexity.

Fraud prevention is no longer a defensive exercise. It is a strategic capability. Properly designed, it improves pricing, protects inclusion, strengthens trust, and sharpens competitive advantage. Poorly designed, it becomes a tax on growth and a drag on confidence.

For the CLA community, the implication is clear. The question is no longer whether fraud risk can be reduced. It is whether the industry is prepared to treat fraud as a shared, strategic responsibility rather than a downstream inconvenience.

Automotive finance has always been about more than vehicles. It is about mobility, access, and trust at scale. In a digital economy, trust must be engineered deliberately. That work is underway, but it is unfinished.

What happens next will determine whether 2025 is remembered as a breaking point or a turning point.

See you at the summit.