Loading...
Loading...
Borys Ulanenko
CEO of ArmsLength AI
Get the latest transfer pricing insights, AI benchmarking tips, and industry updates delivered straight to your inbox.
Loss-making comparables require investigation, not automatic exclusion. The OECD states that loss-making comparables "should not however be rejected on the sole basis that they suffer losses" (). Instead, analyze the cause of the losses and whether those causes are relevant to your tested party's situation.
Direct answer: Yes—loss-making companies can be valid comparables. The key is whether the losses reflect normal business conditions and comparable risks (potentially includable), or non-comparable distress / extraordinary events / different risk profiles (exclude). Document the cause and tie your conclusion to comparability factors and .
Decision Framework:
| Loss Situation | Action | Rationale |
|---|---|---|
| Industry-wide downturn affecting tested party | Include | Loss reflects normal market conditions |
| Start-up phase (if tested party also in start-up) | Include | Comparable business stage |
| One-year anomaly with otherwise healthy performance | Include (use multi-year data) | Smooths out volatility |
| Persistent losses (3+ consecutive years) | Generally exclude | Common audit threshold; often indicates failing business or different risk profile |
| Losses from restructuring or extraordinary events | Exclude | Not reflective of normal business conditions |
| Tested party profitable, comparable persistently loss-making | Exclude | Different economic circumstances |
For related guidance, see our Benchmarking Study Guide, IQR vs Full Range Guide, and Multi-Year Averaging Methods.
When selecting comparables for a TNMM or CPM analysis, loss-making companies create a fundamental question: do their losses reflect normal business outcomes that could apply to the tested party, or do they indicate something fundamentally different?
Including loss-makers:
Excluding loss-makers:
: "Generally speaking, a loss-making uncontrolled transaction should trigger further investigation in order to establish whether or not it can be a comparable"—not automatic rejection.
Including inappropriate loss-makers can:
Excluding appropriate loss-makers can:
Not all losses are equal. Understanding the type of loss helps determine whether to include or exclude a comparable.
| Loss Type | Description | Comparability Impact |
|---|---|---|
| Cyclical Losses | Losses during economic or industry downturns; profits return in up cycles | May include if tested party faces same cycle; use multi-year data to capture full cycle |
| Start-up Losses | Initial years of heavy investment, low revenue, market-building pricing | Include only if tested party is also in start-up phase; losses should be temporary (2–3 years is a common practitioner expectation, not a regulatory cap) |
| Restructuring Losses | One-time costs from plant closures, severance, write-downs | Generally exclude unless tested party underwent similar restructuring; consider adjusting financials if data available |
| Market Exit Losses | Losses from winding down operations, liquidating inventory | Exclude—not a going-concern benchmark; distress pricing doesn't reflect arm's length outcomes |
| Persistent Losses | Losses in 3+ consecutive years with no return to profit | Strong case for exclusion—indicates failing business or non-arm's length conditions |
| One-Time Anomaly | Single bad year from isolated event; profitable in other years | May include using multi-year averages; investigate the cause of the bad year |
When you encounter a loss-making comparable:
1. Industry-Wide Downturn
If the losses occurred during a period of broad economic stress affecting the entire industry, and your tested party was also impacted, including loss-makers provides a realistic arm's length range.
Example: Many companies had losses in 2020 due to COVID-19. If your tested party also had a loss that year, excluding all loss-making comparables would artificially raise the benchmark and make the tested party's result look non-arm's length.
2. Loss-Making Tested Party (Context, Not a Standalone Justification)
A tested party showing losses can make loss-making comparables more relevant, but it is not a sufficient reason to include them. Otherwise the analysis becomes circular: "we included loss-makers because we needed loss-makers." What matters is whether independent companies with a similar functional and risk profile would plausibly accept similar losses under similar market conditions.
If the tested party is a routine / limited-risk entity, persistent losses are often inconsistent with the delineated profile. In that scenario, including persistent loss-makers may weaken the case by implying the tested party bears risks it claims it does not. If the tested party bears market / capacity / inventory / entrepreneurial risk, and the market environment plausibly produces losses for independent firms, then retaining loss-making comparables that reflect the same conditions can strengthen the analysis.
Bottom line: treat "tested party is loss-making" as a prompt to explain the economic story, not a free pass for selecting loss-makers.
3. Single-Year Anomaly
If a comparable has one off-year of loss but is profitable in other years, and its multi-year average is positive, you may include it using multi-year analysis. support using multiple-year data to understand the conditions affecting results, including business and product life cycles.
4. Comparable Economic Circumstances
If the loss arose from the same market factors affecting the tested party—and you can document this parallel—inclusion strengthens the argument that the tested party's strategy and resulting losses are commercially rational.
Documentation is critical: When including a loss-maker, explicitly state: "Company X had a loss in 2020 due to the pandemic, which also affected the tested party. Including Company X reflects market conditions."
1. Losses Not Reflective of Normal Conditions
Exclude when losses result from extraordinary events unrelated to routine business—major strikes, natural disasters, fraud, or one-off legal settlements.
2. Different Risk Profile
Exclude when the comparable undertook risks the tested party doesn't bear. For example, a comparable that lost money on speculative R&D ventures isn't comparable to a stable distribution subsidiary.
3. Persistent Losses (3+ Years)
If a company has losses in all years of your analysis period, it suggests a failing business or non-arm's length support. An independent business wouldn't tolerate indefinite losses without restructuring or exit ().
Practical threshold: Three consecutive loss-making years is a common practitioner heuristic that raises serious concern. In Germany, the GAufzV documentation trigger formally kicks in after more than three consecutive fiscal years (4+), but three years is already a widely used red flag in practice. This is not a universal rule, but a reasonable audit-defense threshold.
4. Tested Party Profitable, Comparable Not
If the tested party was profitable throughout the period while the comparable consistently lost money, the comparable's situation is clearly different—exclude on grounds of different economic circumstances.
5. Restructuring or Exit Scenarios
Companies undergoing major restructuring, winding down, or exiting markets have distorted financials that don't reflect ongoing business operations.
| Years of Losses | Typical Treatment |
|---|---|
| 1 of 5 | Often acceptable if explained |
| 2 of 5 | Borderline—investigate cause carefully |
| 3 of 5 | Commonly excluded in practice; note that the German GAufzV documentation trigger is 4+ consecutive years |
| 4+ of 5 | Generally excluded absent strong justification |
Note: These thresholds reflect common practice and German guidance. The OECD does not prescribe specific cutoffs—investigation is always required.
The OECD's core position on loss-making comparables:
"Generally speaking, a loss-making uncontrolled transaction should trigger further investigation to establish whether or not it can be a comparable." (¶3.65)
"Loss-making comparables that satisfy the comparability analysis should not however be rejected on the sole basis that they suffer losses." (¶3.65)
The OECD also emphasizes () that independent enterprises don't indefinitely tolerate losses—a key principle when evaluating persistent loss situations.
US Treasury Regulations §1.482-1 don't explicitly forbid loss-making comparables but emphasize reliability. Where multiple comparables remain and statistical methods are appropriate to improve reliability, the regulations permit use of the interquartile range (25th–75th percentile) as the arm's length range (§1.482-1(e)(2)(iii)). This can reduce the influence of extreme results—including large losses—depending on the dataset, though that is an effect rather than the stated purpose of IQR.
German authorities have specific expectations, drawing from two sources:
Administrative Principles (2024):
GAufzV Documentation Regulation:
This combination creates a practical expectation that persistent losses will trigger scrutiny and require robust justification.
Outside the OECD’s general framework, some tax authorities scrutinize losses more aggressively for limited-risk (“routine”) entities, where sustained losses may be viewed as inconsistent with the delineated risk profile. If you’re in a high-scrutiny jurisdiction, keep the analysis principle-based: document the cause of losses, confirm the tested party’s risk profile, and show how your comparable set reflects the same economic circumstances.
In SNF Australia (FCAFC 74, 2011), the Commissioner attributed the taxpayer's trading losses to non-arm's-length related-party pricing, while the taxpayer argued the losses were due to market competition and business factors. The Full Federal Court accepted SNF's evidence and held that the transfer prices were arm's length. This case reinforces that the arm's length question turns on evidence and comparability, not losses alone—investigate why before concluding anything about pricing.
The Italian Supreme Court held that loss-making comparables cannot be excluded solely because they are loss-making—exclusion requires a comparability-based reason (e.g., losses driven by exceptional or non-comparable circumstances). The case involved a call center service provider whose markup had been recalculated by tax authorities after excluding entities with negative results. This aligns with and supports retaining loss-makers when the cause of losses is explainable and comparable. See our full Convergys case analysis for details.
The Czech Supreme Administrative Court upheld the exclusion of loss-making comparables for a contract manufacturer with a limited functional and risk profile, reinforcing stricter profitability expectations for routine entities. The court rejected multi-year averaging to justify persistent losses, affirming that routine entities should achieve at least minimal operating profitability. Together with Convergys, this illustrates that courts go both ways depending on the tested party's risk profile—routine entities face higher scrutiny.
START: Comparable shows loss in one or more years
Q0: Is the tested party's delineated risk profile consistent with losses?
Q1: Are losses explained by industry-wide factors?
Q2: Is this a one-year anomaly or persistent (3+ years)?
Q3: What's the business reason for losses?
Q4: Sample size impact?
Document in your accept/reject matrix:
| Element | Example Language |
|---|---|
| Financial results | "Company X: Operating margin -5% (2021), -8% (2022), -10% (2023)" |
| Cause identified | "Per annual report, losses attributed to failed product launch and subsequent restructuring" |
| Tested party comparison | "Tested party was profitable in all three years with no comparable restructuring" |
| Decision | "Excluded—losses not reflective of tested party's conditions per " |
| Element | Example Language |
|---|---|
| Financial results | "Company Y: Operating margin +2% (2019), -5% (2020), +3% (2021)" |
| Cause identified | "2020 loss attributed to COVID-19 impact on retail sector" |
| Tested party comparison | "Tested party also experienced losses in 2020 due to pandemic; recovered in 2021" |
| Decision | "Included—loss reflects industry conditions affecting tested party; 3-year average margin is positive" |
| Company | Years of Losses | Loss Cause | Decision | Rationale |
|---|---|---|---|---|
| Company A | 2021-2023 (3 years) | Persistent; no turnaround | Excluded | Persistent losses not reflective of normal conditions; tested party profitable |
| Company B | 2020 only | COVID-19 downturn | Included | Industry-wide downturn affecting tested party; multi-year average positive |
| Company C | 2022-2023 (2 years) | Market exit mode | Excluded | Not going-concern comparable |
When the tested party itself has losses, special considerations apply.
A consistently loss-making affiliate while the broader MNE group is profitable is a major audit trigger. Tax authorities will ask: is this entity adequately compensated by the group?
If the tested party is routine (limited-risk distributor, contract manufacturer, service provider):
If the tested party is non-routine (bears market / entrepreneurial risk):
| Scenario | Acceptable? | Documentation Required |
|---|---|---|
| Start-up phase (typically 2–3 years in practice) | Yes | Business plan showing path to profitability |
| Economic downturn | Yes | Industry data showing independent firms also had losses |
| One-off adverse event | Possibly | Evidence that event was extraordinary and temporary |
| 4+ consecutive years | Very difficult | Robust justification; expect scrutiny |
Prepare a "Loss Justification Pack":
If losses appear driven by specific measurable differences (e.g., working capital, capacity utilization, accounting classification), consider whether a comparability adjustment is feasible and document the decision either way. See our Working Capital Adjustments Guide and Comparability Adjustments Guide.
Key question: Would an independent party in the tested party's role ever agree to this outcome? If the honest answer is "rarely, if ever," you may need a transfer pricing adjustment or restructuring.
Audit-defense framing: When the tested party is loss-making, the goal is not to "find loss-makers," but to construct an objectively selected comparable set that reflects the same economic circumstances. Loss-making comparables are retained only where the losses are explainable and consistent with the tested party's functions and risks, or where market evidence shows that independent firms in the same position also incurred losses during the relevant period. Otherwise, persistent loss-makers are excluded as failing businesses or as having a different risk profile.
Guides:
Glossary:
This article is based on guidance from the OECD Transfer Pricing Guidelines (2022):
→ Search the full OECD Guidelines
No. OECD Guidelines state that loss-making comparables "should not however be rejected on the sole basis that they suffer losses" (). Each loss-making comparable should be investigated, not reflexively eliminated. If the loss arose from normal business risks or market conditions relevant to the tested transaction, you may retain it. If it arose from factors that don't affect the tested party, exclude it. The decision should be case-by-case, not a blanket rule.
Three or more consecutive loss-making years is a common practitioner heuristic that raises serious concern. In Germany, the formal GAufzV documentation trigger is more than three consecutive fiscal years (i.e., 4+), though three years is already treated as a red flag in practice. The OECD does not set a specific numeric cutoff—it requires further investigation, not automatic exclusion at any number. One or two loss years may be acceptable if explained; three or more usually indicates exclusion unless the tested party has a similar multi-year loss pattern.
Yes—this is exactly when you might need to. If your tested party has losses, including comparables that also suffered losses (from the same economic conditions) can support the argument that the tested loss is arm's length. However, ensure the comparables' losses stem from similar causes as the tested party's. Document the parallels carefully.
Options include: (1) broaden your search criteria geographically or by industry to find additional profitable comparables, (2) use multi-year data to retain comparables with only one loss year, or (3) keep borderline loss-makers with heavy documentation. As a practical workflow consideration (not a regulatory minimum), if excluding all loss-makers leaves fewer than ~5 comparables, explain why those remaining are the only reliable ones and consider supplementing with alternative approaches. The OECD does not prescribe a minimum number of comparables.
Don't just state "it was loss-making." Provide comparability-based reasons: "Company X had three consecutive loss years indicating circumstances not comparable to the tested party (which was profitable). Per its annual report, losses resulted from a major restructuring—a situation the tested party did not experience." Tie the exclusion to recognized comparability factors (functions, risks, conditions) and reference .
Yes. Loss-making comparables typically produce results at the lower end of the distribution and may fall outside the IQR (below the 25th percentile) depending on the dataset. However, IQR is designed to address reliability when material differences can't be adjusted—not specifically to exclude losses. If many comparables have losses (e.g., during a downturn), the median and even quartile boundaries may shift into negative territory. This could be appropriate if the industry truly experienced widespread losses—but it makes the analysis more difficult to defend.
A failing business has persistent losses with no viable path to profitability—it's essentially exiting the market in slow motion. A loss-making comparable may have a temporary loss due to market conditions, start-up phase, or one-off events, but is otherwise viable. The distinction matters: failing businesses aren't reliable benchmarks for ongoing operations; companies with temporary losses may be.
During widespread crises, many companies have losses. The OECD's COVID-19 guidance (2020) stated that loss-making comparables should not be rejected solely because of losses in affected periods—provided comparables assume similar levels of risk and were similarly impacted by the crisis. Include loss-makers from crisis periods if your tested party was also impacted and faced comparable risks; document that the period was extraordinary and that both tested party and comparables were affected. Consider testing 2020 separately rather than averaging it with normal years.