Exit demo
An interactive primer for competition economists

GUPPI.

The Gross Upward Pricing Pressure Index — the intuition, the algebra, the sensitivities, and the numbers regulators have actually used.

GUPPI1 = D12diversion to partner × m2partner's margin × p2p1relative price

The value of sales diverted to the merger partner, expressed as a share of the revenue lost on product 1.

§ 01 — THE INTUITION

A merger internalizes lost sales

Pre-merger, firm 1's price satisfies its own first-order condition. Customers who walk away when it raises price are simply gone — where they go is irrelevant to firm 1's profit. Post-merger, some of those customers walk into the merger partner's shop. Their business is no longer lost; it is recaptured at the partner's margin.

That recapture acts like a new opportunity cost of selling product 1 — every unit sold at the old price forgoes profitable diversion to product 2. GUPPI measures the size of that opportunity cost, scaled by product 1's price. It is gross: no efficiencies, no rival responses, no pass-through — just the raw incentive.

Interactive · click a customer to reroute them

Where do the marginal customers go?

Firm 1 nudges its price up and these {{ nDots }} customers walk. Click each one to send them to the merger partner, an outside rival, or out of the market. Only the first group matters for GUPPI.

to partner (product 2) to outside rivals exit the market
Diversion ratio to partner
{{ dotDPct }}%
{{ dotPartner }} of {{ nDots }} switchers land on product 2
Implied GUPPI at m₂ = 30%, p₂ = p₁
{{ dotGPct }}%
Rivals and exits dilute diversion — only recaptured customers create upward pressure.
Fig. 1 — Diversion is the share of firm 1's marginal customers recaptured by the partner, not the share who leave.
§ 02 — THE CALCULATOR

Three numbers, one index

Everything GUPPI needs is measurable: a diversion ratio (from surveys, switching data, or share proportionality), the partner's price–cost margin, and relative prices. Drag the sliders.

{{ dPct }}%
Share of firm 1's lost sales captured by product 2
{{ mPct }}%
(p₂ − c₂) / p₂, at pre-merger prices
{{ rpStr }}
Converts partner-units of margin into product-1 revenue terms
GUPPI₁
{{ gStr }}%
0 5% 10% 30%+

Of every £1.00 of product-1 sales given up, £{{ divRev }} reappears as product-2 revenue, earning £{{ divProfit }} of margin — a {{ gStr }}% tax-like credit on cutting product-1 output.

Fig. 2 — GUPPI₁ = D₁₂ · m₂ · (p₂/p₁). Reference ticks at the 5% and 10% levels often discussed as screening bands.
§ 03 — THE DERIVATION

From the first-order condition

Six steps, no shortcuts. Differentiated Bertrand, single-product firms 1 and 2, constant marginal costs.

Derivation · step {{ step }} of 6
1.
Pre-merger, firm 1 maximizes π₁ = (p₁ − c₁)·q₁(p₁, p₂). Its first-order condition prices in only its own margin on lost sales:
q₁ + (p₁ − c₁) · ∂q₁/∂p₁ = 0
2.
Post-merger, the firm maximizes joint profit π₁ + π₂. The FOC for p₁ gains a new term — product 2's margin on the sales p₁ pushes its way:
q₁ + (p₁ − c₁) · ∂q₁/∂p₁ + (p₂ − c₂) · ∂q₂/∂p₁ = 0
3.
Define the diversion ratio — of the sales product 1 loses, the fraction product 2 gains:
D₁₂ = (∂q₂/∂p₁) / (−∂q₁/∂p₁)
4.
Substitute. The new term equals D₁₂(p₂ − c₂) per unit of lost sales — the merger acts exactly like an increase in product 1's marginal cost (or a specific tax on it) of that amount:
c₁  ⟶  c₁ + D₁₂ · (p₂ − c₂)
5.
Scale the opportunity cost by product 1's price to make it unit-free, and rewrite the partner's absolute margin as m₂ · p₂:
D₁₂ · (p₂ − c₂) / p₁  =  D₁₂ · m₂ · (p₂/p₁)
6.
That expression is the GUPPI. It is the whole first-order incentive — gross of efficiencies, holding p₂ and rivals fixed, and silent on how much of the pressure passes through to price:
GUPPI₁ = D₁₂ · m₂ · (p₂/p₁)
Fig. 3 — Step through the algebra; step 4 delivers the "tax analogy" — the merger raises product 1's effective marginal cost by the value of diverted sales.
§ 04 — SENSITIVITY

The margin × diversion surface

GUPPI is bilinear in its two contested inputs, so screening thresholds are hyperbolas: a high-margin industry hits any given level at low diversion. Hover to read the surface (p₂ = p₁).

MARGIN m₂ → 80% 0
D₁₂ {{ hoverD }}%  m₂ {{ hoverM }}%
GUPPI {{ hoverG }}%
0DIVERSION RATIO D₁₂ →75%
Fig. 4 — Iso-GUPPI contours. The dashed curve marks 2.75%, the decision-rule threshold the CMA applied to supermarkets in Sainsbury's/Asda (2019); earlier UK cases intervened around 5–10%.
§ 05 — BEYOND GROSS

What the G leaves out

Three standard refinements. The first two update live — the sliders below are the same D₁₂ and m₂ as the calculator, plus each card's own controls.

{{ dPct }}%
{{ mPct }}%
GUPPI₁
{{ gStr }}%
EFFICIENCIES · CMCR

The compensating cost reduction

The marginal-cost saving that exactly offsets the upward pressure (symmetric Bertrand, Werden 1996):

CMCR = m·D / [(1−m)(1−D)]
At your settings: {{ cmcrStr }}% of marginal cost — note how it outruns GUPPI as D and m grow.
{{ ePct }}%
Net pressure: {{ gStr }}% − {{ ePct }}% = {{ netStr }}% — {{ netVerdict }}
PASS-THROUGH

Pressure ≠ price rise

GUPPI is a cost shock in disguise, so the first-order price effect is GUPPI × pass-through. Linear demand gives ½; log-concavity less, iso-elastic more.

Δp₁/p₁ ≈ ρ × GUPPI₁,  ρ ≈ ½ (linear)
{{ ptRate }}%
Your GUPPI of {{ gStr }}% implies ≈ {{ ptStr }}% on p₁, before rival feedback. {{ ptNote }}
VERTICAL · vGUPPI

The vertical cousin

Moresi & Salop (2013) rebuild the logic for vertical mergers: vGUPPIu scores the upstream unit's incentive to raise input prices to downstream rivals; vGUPPIr the resulting rival cost pass-through; vGUPPId the downstream unit's own softened pricing. Same engine — diversion × margin — routed through the input market.

§ 06 — IN THE CASE FILES

Numbers that decided mergers

CMA · 2019 · PROHIBITED

Sainsbury's / Asda

The CMA used GUPPI as a local decision rule across hundreds of overlap areas, with SLC thresholds of 2.75% for supermarkets and 3.25% for convenience stores — after crediting efficiencies. Nationally weighted GUPPIs: 2.5% (Sainsbury's) and 3.3% (Asda).

2.75%
SUPERMARKET
THRESHOLD
3.25%
CONVENIENCE
THRESHOLD
CMA · 2016–17 · CLEARED WITH REMEDIES

Ladbrokes/Coral · Tesco/Booker

Earlier UK retail cases drew intervention lines at markedly higher GUPPI levels — roughly 5–10% — making the Sainsbury's/Asda thresholds a step-change in stringency.

DOJ · 2011 · BLOCKED IN COURT

H&R Block / TaxACT

The first US litigated merger in which a court engaged seriously with upward-pricing-pressure evidence, a year after diversion-and-margin logic entered the 2010 Horizontal Merger Guidelines as "the value of diverted sales."