Fair value is defined as the price that would be received to sell an asset or paid to transfer a liability in an orderly transaction between market participants at the measurement date. The definition implies objectivity: a market price, determined by willing parties, in an arm’s-length transaction.
In practice, fair value measurement spans a spectrum from directly observable (a listed equity security with an active market price) to entirely model-dependent (an illiquid investment in an unlisted entity with no comparable transactions and no observable inputs).
Ind AS 113 organises this spectrum into three levels:
Level 1: Quoted prices in active markets. Observable. Objective. Verifiable.
Level 2: Observable inputs other than Level 1 prices. Interest rates, credit spreads, comparable transaction prices. Partially observable. Partially judgment-dependent.
Level 3: Unobservable inputs. Management’s own assumptions about the inputs market participants would use. Not observable. Entirely judgment-dependent.
The further an asset moves from Level 1 toward Level 3, the more the “fair value” becomes management’s model of what a fair value would be if a market existed. And the more the reported number depends on assumptions that management controls.
Where Fair Value Becomes Judgment
Unlisted investments
A company holds a 15% stake in an unlisted entity acquired for Rs 20 crore three years ago. There is no active market. There have been no recent transactions in the entity’s shares. The entity’s financial performance has been mixed.
The fair value is determined by a valuation model: DCF, comparable companies, comparable transactions or a combination. Each approach requires assumptions about future cash flows, growth rates, margins, discount rates, and the selection and adjustment of comparable companies.
The resulting fair value may range from Rs 12 crore to Rs 30 crore, depending on the methodology and assumptions. Management selects one approach, one set of assumptions and one number. The financial statements report the number with the precision of a fact. It is, in reality, management’s judgment.
Complex derivatives
A derivative instrument without an active market price (a bespoke interest rate swap, a commodity option with non-standard terms, a structured product) requires a valuation model. The model requires inputs: forward curves, volatility surfaces, correlation assumptions and credit adjustments. Each input involves judgment, particularly for long-dated instruments where the observable data is limited.
Investment property
Investment property measured at fair value requires a valuation that may be based on comparable transactions (if available), income capitalisation (rental income divided by a capitalisation rate), or DCF (projected rental income discounted to present value). Each approach involves assumptions about future rentals, vacancy rates, capitalisation rates and discount rates.
Biological assets
Agricultural companies measuring biological assets (livestock, standing timber, unharvested crops) at fair value face measurement challenges because active markets for growing biological assets rarely exist. The fair value is typically estimated based on the expected selling price at harvest less the costs to complete growing and to sell.
The Governance Challenge
The governance challenge with Level 3 fair values is that the number in the financial statements, presented alongside Level 1 and Level 2 measurements with the same degree of apparent certainty, is qualitatively different. It is not a market price. It is a model output. And the model’s output depends entirely on assumptions that management selects.
In Northrop Management Private Limited’s forensic and financial reporting work, fair value analysis focuses on Level 3 measurements, which are the most judgment-dependent and the most consequential to reported results.
The methodology: identify every Level 3 fair value measurement on the balance sheet. For each, examine the valuation model, the assumptions, the sensitivity to changes in key inputs and the consistency of the methodology over time. Test whether the assumptions are supported by external evidence, whether the methodology has changed (and if so, what the impact was) and whether the fair value movement is consistent with the underlying economics of the asset.
Ashish Chaudhary, frames the reporting reality directly: “The further an asset moves from observable market evidence, the more accounting becomes an exercise in judgment. A Level 1 fair value is a fact. A Level 3 fair value is a model. The financial statements present both with the same apparent precision. The board should know which is which, because the risk profile of a balance sheet depends on how much of its value is observable and how much is modelled.”
Questions for the Boardroom
- What percentage of our total assets measured at fair value are Level 3 (model-dependent, unobservable inputs)?
- For each significant Level 3 measurement, what model is used, what assumptions drive it and how sensitive is the valuation to changes in key assumptions?
- Has the valuation methodology or any significant assumption changed in the last three years, and what was the impact?
- If we changed the most sensitive assumption in each Level 3 model by 10% in the unfavourable direction, what would the aggregate impact on reported net worth be?
- Are our Level 3 fair value measurements independently verified by a party other than the management team that benefits from the valuation?
Closing Implication
Fair value is a powerful measurement concept. At Level 1, it is the most objective measurement available: a market price, determined by willing participants, verifiable in real time. At Level 3, it is the most subjective: a model output, determined by management’s assumptions, verifiable only by examining the model itself.
The board that knows the distribution of its balance sheet across the fair value hierarchy knows how much of its reported value is fact and how much is model. The board that does not knows only the total, which may conceal a concentration of model-dependent values that are one assumption change away from a material write-down.