ORDEEN · Choice Modelling · 2 MIN READ

What is the Krinsky-Robb Method for Willingness-to-Pay Intervals?

The Krinsky-Robb method constructs honest confidence intervals for willingness-to-pay by simulating from the joint distribution of coefficients.

Book a 20-min demo What is the Krinsky-Robb Method for Willingness-to-Pay Intervals?

The Krinsky-Robb method is a simulation-based procedure for constructing confidence intervals around willingness-to-pay (WTP) estimates in choice modelling. It is the honest alternative to the delta-method approximation — and the only reliable choice when the price coefficient is small, uncertain, or near zero.

Why WTP intervals are hard

WTP is a ratio of two estimated utilities:

$$\mathrm{WTP} = \frac{\beta_{\text{attribute}}}{|\beta_{\text{price}}|}$$

The price coefficient sits in the denominator. When $\beta_{\text{price}}$ is well-identified, the ratio is well-behaved. When $\beta_{\text{price}}$ is small or straddles zero, the ratio explodes — a 95% CI from the standard delta method can span negative infinity to positive infinity without anyone noticing. Reporting a point WTP as if it were reliable, when the distribution is pathological, is a classical analytical failure.

How Krinsky-Robb works

Instead of linearising the ratio, Krinsky-Robb takes draws from the full joint distribution of the estimated coefficients and recomputes the ratio at each draw:

  1. Draw $K$ (typically 1,000–10,000) samples $(\tilde{\beta}_{\text{attribute}}^{(k)}, \tilde{\beta}_{\text{price}}^{(k)})$ from the estimated multivariate normal distribution of the coefficients, respecting their covariance.

  2. Compute $\mathrm{WTP}^{(k)} = \tilde{\beta}_{\text{attribute}}^{(k)} / |\tilde{\beta}_{\text{price}}^{(k)}|$ at each draw.

  3. Report the empirical quantiles (2.5% and 97.5% for a 95% CI) of the resulting WTP distribution.

In a Bayesian setting, the equivalent uses posterior draws directly — no reliance on asymptotic normality, no reliance on linearisation.

When Krinsky-Robb fails loudly (and should)

The method surfaces the pathology instead of hiding it. If the price coefficient’s posterior straddles zero, the WTP distribution has infinite variance and the empirical CI will be absurdly wide — correctly signalling that the data cannot support a confident WTP. The right response is to refuse the WTP estimate rather than report an incompetent point value.

Why Ordeen uses it

Every whole-sample WTP is computed at each posterior draw (the Bayesian form of Krinsky-Robb) and reported with a 95% credible interval alongside the point estimate. When the price effect changes sign across draws, the WTP is flagged as unreliable and refused outright — a wide-and-useless number is worse than silence. Hole (2007) compares this credible-interval approach with the delta method and other alternatives; the alternative — a plain delta-method CI shipped without a sanity check — is the quickest way to an indefensible pricing recommendation.

Related terms

Willingness to pay · Hierarchical Bayes · Full methodology

The Krinsky-Robb method is a simulation-based procedure for constructing confidence intervals around willingness-to-pay (WTP) estimates in choice modelling. It is the honest alternative to the delta-method approximation — and the only reliable choice when the price coefficient is small, uncertain, or near zero.

Why WTP intervals are hard

WTP is a ratio of two estimated utilities:

$$\mathrm{WTP} = \frac{\beta_{\text{attribute}}}{|\beta_{\text{price}}|}$$

The price coefficient sits in the denominator. When $\beta_{\text{price}}$ is well-identified, the ratio is well-behaved. When $\beta_{\text{price}}$ is small or straddles zero, the ratio explodes — a 95% CI from the standard delta method can span negative infinity to positive infinity without anyone noticing. Reporting a point WTP as if it were reliable, when the distribution is pathological, is a classical analytical failure.

How Krinsky-Robb works

Instead of linearising the ratio, Krinsky-Robb takes draws from the full joint distribution of the estimated coefficients and recomputes the ratio at each draw:

  1. Draw $K$ (typically 1,000–10,000) samples $(\tilde{\beta}_{\text{attribute}}^{(k)}, \tilde{\beta}_{\text{price}}^{(k)})$ from the estimated multivariate normal distribution of the coefficients, respecting their covariance.

  2. Compute $\mathrm{WTP}^{(k)} = \tilde{\beta}_{\text{attribute}}^{(k)} / |\tilde{\beta}_{\text{price}}^{(k)}|$ at each draw.

  3. Report the empirical quantiles (2.5% and 97.5% for a 95% CI) of the resulting WTP distribution.

In a Bayesian setting, the equivalent uses posterior draws directly — no reliance on asymptotic normality, no reliance on linearisation.

When Krinsky-Robb fails loudly (and should)

The method surfaces the pathology instead of hiding it. If the price coefficient’s posterior straddles zero, the WTP distribution has infinite variance and the empirical CI will be absurdly wide — correctly signalling that the data cannot support a confident WTP. The right response is to refuse the WTP estimate rather than report an incompetent point value.

Why Ordeen uses it

Every whole-sample WTP is computed at each posterior draw (the Bayesian form of Krinsky-Robb) and reported with a 95% credible interval alongside the point estimate. When the price effect changes sign across draws, the WTP is flagged as unreliable and refused outright — a wide-and-useless number is worse than silence. Hole (2007) compares this credible-interval approach with the delta method and other alternatives; the alternative — a plain delta-method CI shipped without a sanity check — is the quickest way to an indefensible pricing recommendation.

Related terms

Willingness to pay · Hierarchical Bayes · Full methodology

Related guides

Contract Analysis

Jurisdiction-Grounded AI Contract Review & Risk

Why generic chatbots hallucinate legal contract risk, and how Ordeen grounds contract review, portfolio risk scoring and renewal drafting in governing law.

Read the guide
Forecasting

Intermittent Demand Forecasting: Croston, TSB & MinT

How to forecast lumpy, sparse SKU demand and reconcile hierarchies: Croston's method, SBA, TSB, MinT reconciliation and out-of-sample AI signal validation.

Read the guide
Price Monitoring

Retail Checkout Price Monitoring vs Sticker Scrapers

Why basic retail price scrapers miss real checkout undercutting: how delivery thresholds, variant SKU matching and screenshot evidence protect brand RRP.

Read the guide