Single Firm-Wide Discount Rate

The single firm-wide discount rate is the empirical pattern, due to Kruger, Landier, and Thesmar ( $2015$ ), that corporate managers apply one discount rate to value projects across all of the firm’s divisions, rather than computing project-specific or division-specific discount rates as the rational benchmark requires. The bounded-rationality interpretation is that constructing accurate division-specific discount rates is computationally costly and informationally demanding, and managers economize by applying the firm’s overall weighted-average cost of capital (WACC) as a uniform benchmark.

The rational benchmark. A division of the firm has its own systematic risk — its own beta with the market, its own exposure to industry shocks. The division’s correct discount rate is determined by its own beta, not by the firm’s average beta across all divisions. The CAPM gives:
$r_{division} = r_{f} + β_{division} \cdot (r_{m} - r_{f}) .$
The high-beta divisions of a multi-segment firm should be discounted at higher rates than the low-beta divisions. Using a single firm-wide discount rate over-discounts low-beta divisions (rejecting good projects) and under-discounts high-beta divisions (accepting bad projects).

The Kruger-Landier-Thesmar prediction. If managers apply the firm-wide WACC uniformly:

High-beta divisions receive too low a discount rate — many marginal projects look NPV-positive that should be NPV-negative. Result: over-investment in high-beta divisions.

Low-beta divisions receive too high a discount rate — many marginal projects look NPV-negative that should be NPV-positive. Result: under-investment in low-beta divisions.

Empirical confirmation. Kruger-Landier-Thesmar test the prediction using a sample of multi-segment U.S. firms. They classify each segment by its industry beta (based on stand-alone industry comparables) and test whether segment investment behaves according to the rational benchmark or according to the single-discount-rate bias. The data confirm the bias: segments with high industry betas receive higher investment-to-assets ratios than the rational CAPM benchmark predicts, and segments with low industry betas receive lower investment.

Why managers do this. The bounded-rationality interpretation rests on cognitive cost. Computing a project-specific or division-specific discount rate requires (i) estimating the project’s beta, (ii) updating the cost-of-capital model with the project’s risk profile, and (iii) explaining the choice to the board and to investors. The firm-wide WACC is a single, well-known number that is easy to apply and easy to defend. Managers economize by using it everywhere, accepting the over- and under-investment costs as the price of cognitive simplicity.

Intuition

The textbook NPV rule says project value depends on project-specific cash flows discounted at a project-specific risk-adjusted rate. In practice, virtually no firm computes division-specific or project-specific discount rates routinely; instead, almost all firms use a single firm-wide weighted-average cost of capital (WACC) as the discount rate for all projects across all divisions. This is not because finance professors have failed to teach the right rule — corporate-finance textbooks emphasize project-specific discount rates extensively — but because applying the right rule is operationally difficult and the simplification is convenient.

The cost of the simplification is not symmetric. High-beta divisions get over-invested because their too-low discount rate makes marginal projects look NPV-positive; low-beta divisions get under-invested because their too-high discount rate makes marginal projects look NPV-negative. The firm’s overall investment policy is therefore distorted toward higher-risk, higher-return projects relative to the value-maximizing benchmark — a pattern visible in the cross-sectional data on multi-segment firms.

The Kruger-Landier-Thesmar pattern is a clean example of bounded rationality in corporate decision-making. The decision-makers are sophisticated, well-trained MBAs and senior executives. The failure is not lack of knowledge of the right rule but a cognitive economy that trades off computational complexity against decision quality. The empirical magnitude — the over-investment in high-beta segments is economically significant — suggests that the cognitive savings come at meaningful welfare cost.

Properties

(Rational Benchmark) Discount rate equals project-specific cost of capital.

The CAPM-based rational benchmark requires that each project (or each division as an aggregate of projects) be discounted at its own risk-adjusted rate $r_{f} + β_{project} (r_{m} - r_{f})$ . A division producing automobiles in a multi-segment industrial firm should be discounted at the auto-industry beta; a division producing pharmaceuticals should be discounted at the pharma-industry beta. Using a single firm-wide WACC is correct only if all divisions have identical betas — a condition that essentially never holds for multi-segment firms.

(Empirical Pattern) Firms use one WACC for all divisions.

Survey evidence (Graham-Harvey $2001$ and Jagannathan-Matsa-Meier-Tarhan $2016$ ) confirms the practice: the vast majority of large U.S. firms report using a single firm-wide WACC as the discount rate for capital-budgeting decisions across all divisions. The minority that report using division-specific discount rates typically use only two or three different rates rather than truly project-specific rates. The practice is widespread enough that it dominates corporate capital-budgeting in any large multi-segment firm.

(Distortion Direction) High-beta divisions over-invest; low-beta divisions under-invest.

The Kruger-Landier-Thesmar prediction follows mechanically from the rational-benchmark vs. firm-wide-WACC comparison. If a division’s true beta exceeds the firm-wide average, its true cost of capital exceeds the firm-wide WACC, so applying the WACC under-discounts the division and over-states its NPV. The opposite holds for divisions below the average. The firm therefore systematically over-invests in its high-beta divisions and under-invests in its low-beta divisions.

(Empirical Test) Kruger-Landier-Thesmar (2015) confirm the cross-sectional pattern.

Kruger, Landier, and Thesmar ( $2015$ ) test the prediction using a sample of multi-segment U.S. firms from $1990$ onward. They classify each segment by its industry beta (using stand-alone industry comparables that operate only in that segment) and regress segment investment-to-assets on the segment’s beta minus the firm’s overall WACC. The coefficient is statistically significant and economically meaningful: segments with above-average betas have $5$ – $10$ percentage points higher investment-to-assets ratios than the rational benchmark predicts, and segments with below-average betas have correspondingly lower investment.

(Robustness) The pattern survives controls for division-specific opportunities.

A natural alternative explanation: high-beta divisions might genuinely have better investment opportunities than low-beta divisions, justifying their higher investment levels even under correct discount rates. Kruger-Landier-Thesmar control for division-specific Tobin’s $Q$ and other opportunity proxies and find that the bias survives. Even after the rational-benchmark variation in opportunities is netted out, high-beta divisions invest more than they should and low-beta divisions invest less.

(Bounded-Rationality Interpretation) Cognitive cost of project-specific rates is high.

The interpretation rests on the cognitive cost of computing accurate project-specific discount rates. The CAPM beta of a stand-alone division is not directly observable; it must be estimated from comparable-firm betas, with judgment calls about which comparables to use, how to adjust for leverage differences, and how to incorporate the firm’s own debt-equity mix into the WACC. Each step introduces both computational complexity and discretion, the latter of which can be politically inconvenient (division managers may dispute the rate assigned to them). Using the firm-wide WACC avoids all of these complications at the cost of accepting the over- and under-investment distortion.

(Aggregate Implications) Investment misallocation across the economy.

If the single-firm-wide-discount-rate bias holds across the population of multi-segment firms, the aggregate implication is meaningful misallocation of capital: too much investment in industries with high systematic risk, too little in industries with low systematic risk. The pattern would show up as over-investment in cyclical industries (autos, basic materials) and under-investment in defensive industries (utilities, consumer staples) relative to the value-maximizing allocation. Whether the aggregate pattern is large enough to matter for macro-level capital allocation is an open empirical question.

Examples

A conglomerate with two divisions

A conglomerate operates two divisions: an industrial-chemicals division (industry beta $\approx 1.4$ ) and a regulated-utility division (industry beta $\approx 0.6$ ). The firm’s overall WACC, weighted by division size, is $10%$ . The rational division-specific costs of capital would be approximately $12%$ for chemicals and $7%$ for utilities (using risk-free rate $4%$ and market risk premium $5%$ ).

The chemicals division evaluates a marginal project with expected return $11%$ . Under the rational benchmark, $11% < 12%$ — the project is NPV-negative and should be rejected. Under the firm-wide WACC of $10%$ , the project appears NPV-positive ( $11% > 10%$ ) and is accepted. Over-investment.

The utility division evaluates a marginal project with expected return $8%$ . Under the rational benchmark, $8% > 7%$ — the project is NPV-positive and should be accepted. Under the firm-wide WACC of $10%$ , the project appears NPV-negative ( $8% < 10%$ ) and is rejected. Under-investment.

The same uniform WACC pushes investment toward the higher-risk division and away from the lower-risk division — the Kruger-Landier-Thesmar pattern.

Why firms persist with the practice despite knowing the right rule

Senior executives at any large firm have either MBAs or extensive corporate-finance training; they know about CAPM, project-specific discount rates, and the welfare cost of using a single WACC. They persist with the single-WACC practice for several reasons. First, computing accurate division-specific betas is operationally difficult and produces estimates with wide standard errors. Second, division managers have political incentives to dispute any rate assigned to them; using the firm-wide WACC sidesteps the conflict. Third, the firm’s external communications (to investors, board, analysts) often presume a single discount rate, making the single-WACC practice administratively simpler. The cognitive economy is real, but it comes at the cost of the misallocation that Kruger-Landier-Thesmar document.

A simple test you could run on a public conglomerate

Take a publicly-traded multi-segment firm. Compute each segment’s industry-beta using stand-alone industry comparables. Compute the firm’s overall WACC. Compare each segment’s investment-to-assets ratio to the segment’s industry-average investment-to-assets ratio (which should reflect how high-beta-segment firms invest when they are not weighted down by a low-beta-segment-pulled WACC). The Kruger-Landier-Thesmar prediction is that high-beta segments inside the conglomerate over-invest relative to their stand-alone industry peers, and low-beta segments under-invest. The pattern can be tested with publicly available data and is the kind of empirical exercise that course problem sets often include.

tags	domain/finance, type/concept, level/intermediate, status/complete
aliases	WACC Misuse, Kruger-Landier-Thesmar, Single-Discount-Rate Bias

Project Alexandria

Explorer

Single Firm-Wide Discount Rate