South America · Country reference

Earthquake Risk in Peru

Peru sits above one of the most heavily loaded segments of the Nazca subduction zone. The stretch of megathrust facing Lima has not fully ruptured since 1746, when a magnitude 8.8 earthquake and a 24-metre tsunami destroyed the capital and erased Callao, and the published recurrence for that rupture is about 305 years. Lima and Callao now concentrate close to half of Peru's GDP and 11 million people, while roughly 3% of the country's homes carry earthquake cover. Few markets anywhere combine a sharper tail with a thinner cushion.

Modelled and reviewed by Dynamis, earthquake engineering consultantsLast reviewed 13 July 2026

Primary source
Lima megathrust segment
Highly coupled; no full rupture since 1746
Reference scenario
M8.8, 110,000+ deaths
INDECI official scenario for Lima and Callao
Exposure concentration
~46% of GDP
Lima and Callao (INEI departmental accounts)
Homes with earthquake cover
~3%
APESEG, 2025; about 7% within Lima
Sovereign protection
US$60M paid, 2019
Pacific Alliance cat bond; matured 2021, not renewed
Design code
E.030
In force since 1977; last updated May 2026

What drives the risk

Offshore Peru, the Nazca Plate dives beneath South America along the Peru-Chile Trench, producing some of the largest earthquakes ever recorded on the continent. The margin behaves as a chain of segments: each stores strain for decades to centuries and ruptures in great earthquakes, sometimes with tsunamis. Inland, the subducted plate keeps generating deep events beneath the Andes and the Amazon, and crustal faults thread the highlands near cities like Cusco. Where a segment has stayed quiet longest is where the ledger is largest, and the quietest segment of all faces Lima.

Risk in Peru is driven by a segmented subduction margin, and the segments are not equal. The central segment, bounded roughly by the Mendana Fracture Zone and the Nazca Ridge, is the one facing Lima: geodetic studies show it highly coupled, and the magnitude 8.0 earthquakes of 1940, 1966, 1974 and 2007 released only a fraction of the strain accumulated since 1746. The south last broke fully in 1868, in an earthquake felt across the Pacific. The deep events beneath the Amazon, like the magnitude 8.0 of 2019, shake enormous areas but tend to spare structures, a contrast that matters for pricing.

The second driver is concentration. Lima and Callao hold close to half the country's output, three quarters of its insured homes, and the port through which most of its trade moves. A portfolio can be diversified across all of Peru on paper and still be, in seismic terms, a single position on the central segment.

The fault map of Peru

The fault sources and subduction geometry of the Xpectral Hazard Model, the same geometry that drives our platform, plus every M6+ earthquake since 1960 from the USGS live catalogue. Click any event for magnitude, date and depth.

Loading map…
active faultsubduction trench< 70 km70–300 km> 300 kmM ≥ 6 since 1960 · USGS, live

City exposure, from our model

Tiers summarise the Xpectral Peru model at the design return period, combined with each city's dominant source. They are a portfolio screening view: for underwriting we resolve this per site, because within one city the answer moves sharply with ground conditions and construction vintage.

CityExposure tierDominant sourceUnderwriting note
Lima & CallaoSevereCentral megathrust segment, coupled since 1746Close to half the country's GDP in one rupture footprint, plus tsunami exposure at the port. The national accumulation ceiling is set here.
ArequipaHighSouthern subduction segment (2001 magnitude 8.4)The 2001 event damaged the sillar masonry of the historic centre and sent a tsunami into Camana. Heritage stock and modern stock behave very differently here.
Tacna & MoqueguaHighFar-southern segment, last full rupture 1868The 1868 earthquake is the regional benchmark; the segment's quiet since then is the reason, not reassurance.
CuscoModerateShallow crustal faults of the high AndesLocal faults near the city have produced damaging events; adobe and heritage construction concentrate the vulnerability.
TrujilloModerateNorthern subduction segmentLower coupling than the centre, but the 1970 Ancash earthquake showed what the northern coast can take from one event.
IquitosLowerDeep events under the Amazon basinThe 2019 magnitude 8.0 passed 110 km beneath the region with two fatalities: depth, not luck. Shaking hazard exists; loss potential is structurally lower.

Tiers are a qualitative summary of the model output. Asset-level figures (AAL, EP curves, PML) come from the platform, resolved at the site.

Reference scenario

The Lima segment, loaded since 1746

On 28 October 1746, the megathrust off Lima ruptured in an earthquake estimated near magnitude 8.8. The capital was left with 25 houses standing; the tsunami that followed ran up 24 metres at Callao and killed nearly all of its five to six thousand inhabitants. Geodesy says that segment is again fully locked, that the magnitude 8.0 events of 1940, 1966, 1974 and 2007 barely dented the accumulated strain, and that a 1746-type rupture recurs roughly every 305 years, plus or minus 40. The arithmetic is uncomfortable and it is official.

Peru's civil defence agency INDECI plans against a magnitude 8.8 rupture in front of Lima and Callao: 110,313 deaths, more than two million injured, 353,497 dwellings destroyed and eight million people affected. Published economic estimates run from US$17.7 billion (IDB, magnitude 8.5 with tsunami) to more than US$35 billion (Lloyd's City Risk Index), with industry voices citing US$100 billion for a shallow rupture directly under the capital. This is the stress case that should size a Peruvian earthquake accumulation, and it is precisely the footprint our model resolves.

Loading map…

Metropolitan Lima and Callao. The coupled central segment of the megathrust lies some 150 km offshore (the amber interface on the country map); the pins mark the districts whose exposure the INDECI scenario stresses, from the port of Callao to the soft-ground districts of the east and south.

Major historical earthquakes

The events that shaped how Peru is underwritten today.

YearMagnitudeEventImpact
2019M8.0Loreto (Lagunas) earthquake
Amazon basin, 110+ km deep
Two deaths from a magnitude 8.0: depth defanged it. It still triggered a US$60 million payout on Peru's catastrophe bond, the country's only documented earthquake insurance recovery at sovereign scale.
2007M8.0Pisco earthquake
Ica, south of Lima
595 killed, 148 of them in one church collapse; around 58,000 homes destroyed. Economic losses above US$1 billion against roughly US$120 million paid by insurers: a 10% insured share.
2001M8.4Arequipa (Camana) earthquake
Southern Peru
The largest earthquake worldwide in 36 years: 74 killed plus 64 missing in the Camana tsunami, 17,584 homes destroyed, and the cathedral of Arequipa's historic centre badly damaged.
1996M7.7Nazca earthquake
Ica / Nazca
Around 4,000 houses collapsed for 14 deaths: an afternoon event in low-rise towns, and a lesson in how timing and building stock shape the loss.
1970M7.9Ancash earthquake
Chimbote / Huaraz
About 70,000 dead, the deadliest earthquake in the history of the Americas. The shaking released an avalanche from Huascaran that buried the town of Yungay, and the catastrophe created Peru's civil defence system.
1868M8.8Arica earthquake and tsunami
Then-southern Peru
Estimated near magnitude 9: around 25,000 deaths, 12 to 16 metre waves at Arica and a tsunami recorded across the Pacific. The far-southern segment has not fully ruptured since.
1746M8.8Lima-Callao earthquake and tsunami
Lima / Callao
Lima flattened, Callao erased by a 24-metre tsunami with barely 200 survivors. The rupture the Peruvian market still measures itself against.

The earthquake insurance market

Peru's insurance penetration was 2.05% of GDP at mid-2025, with property and casualty lines below one point: roughly half the Latin American average. Earthquake cover is written as part of property policies and is not required by any law; in practice banks demand fire and earthquake cover on mortgaged homes, which is why cover exists at all. APESEG, the insurers' association, counts around 3% of Peruvian homes with earthquake protection, about 7% within Lima, and reports that only 12 in 100 households keep the policy once the mortgage is paid. Among the country's three million small businesses, documented estimates of property cover run between 5% and 20% depending on scope.

The loss history makes the gap concrete. The 2007 Pisco earthquake caused economic losses above US$1 billion; insurers paid out around US$120 million, a tenth. The regulator has been active on the underwriting side for two decades: since 2005 the SBS requires insurers to quantify catastrophe exposure with an approved earthquake model and to hold a dedicated catastrophe reserve, and in 2026 it opened a parametric insurance regulation to consultation. GEM's public risk profile for Peru estimates an average annual loss of US$273 million and a probable maximum loss around US$4.2 billion for the building stock, numbers that dwarf the premium base collecting against them.

At sovereign level, Peru placed a US$200 million tranche in the 2018 Pacific Alliance catastrophe bond, the World Bank's largest sovereign risk transaction at the time. The deep 2019 Loreto earthquake triggered a 30% payout, US$60 million, within four weeks. The bond matured in 2021 and no renewal is documented; Peru's disaster risk financing now leans on contingent credit lines above US$1.2 billion. For international (re)insurers the reading is simple: the tail is among the sharpest in the hemisphere, the insured base is thin, and both the regulator and the sovereign have shown appetite for structured solutions.

Building stock and the seismic code

Peru has had a national seismic design norm since 1977, revised in 1997, 2003, 2016 and 2018, and most recently updated in May 2026 with modern ground classification and bidirectional analysis requirements. E.030 is issued by the housing ministry with SENCICO as technical secretariat, and the record of engineered buildings in recent earthquakes is respectable. The problem is what the code never touched: a large share of Lima's homes are self-built, and the INDECI scenario's 353,497 destroyed dwellings are concentrated precisely there.

For underwriting, vintage and formality are the two questions that matter most in Peru. The microzonation studies of CISMID map how the capital's soils amplify shaking district by district, and in 2025 the regulator and CISMID published a structural typology manual specifically to improve residential earthquake underwriting. The information to price Lima street by street exists and is public; portfolios rated at province level are leaving that information on the table.

What it means for your book

For a (re)insurer, Peru is a tail-and-concentration market. Premium volume is modest and loss frequency is manageable, but the capital charge is set by one scenario: the central segment rupturing in front of Lima. Nearly half the country's GDP, three quarters of its insured homes and its main port sit inside that footprint. Accumulation control against the segment, not the country, is the discipline that matters.

This is where the three Xpectral layers enter the book directly. Sismicus resolves the ground-shaking hazard at the site, validated against E.030, not at the province. Fragility turns that shaking into a defensible loss for each asset, from sillar heritage in Arequipa to self-built masonry in Lima's hills. Risco converts the portfolio into the average annual loss, exceedance-probability curve and probable maximum loss that feed pricing, accumulation limits and Solvency II capital.

Frequently asked

Why is Peru so earthquake-prone?

The Nazca Plate subducts beneath South America along the Peru-Chile Trench, directly offshore Peru's coast. The margin ruptures in great earthquakes segment by segment, has produced events near magnitude 9 (1746, 1868), and keeps generating deep earthquakes beneath the Andes and the Amazon. The deadliest earthquake in the history of the Americas, Ancash 1970 with about 70,000 dead, was Peruvian.

What is the official earthquake scenario for Lima?

INDECI, Peru's civil defence agency, plans against a magnitude 8.8 rupture on the megathrust segment facing Lima and Callao: 110,313 deaths, more than two million injured, 353,497 dwellings destroyed and around eight million people affected. Published economic estimates range from US$17.7 billion (IDB) to more than US$35 billion (Lloyd's City Risk Index).

When did Lima last have a great earthquake?

The segment facing Lima last ruptured fully in 1746, in an event estimated near magnitude 8.8 whose tsunami destroyed Callao. The magnitude 8.0 earthquakes of 1940, 1966, 1974 and 2007 released only a fraction of the strain accumulated since; published recurrence for a 1746-type rupture is about 305 years, plus or minus 40.

Is earthquake insurance mandatory in Peru?

No law requires it. Banks require fire and earthquake cover on mortgaged homes for the life of the loan, which is the main reason cover exists; APESEG reports only 12 in 100 households keep the policy after the mortgage ends. Overall, around 3% of Peruvian homes carry earthquake cover.

How large is Peru's earthquake protection gap?

In the 2007 Pisco earthquake, insurers paid roughly US$120 million against economic losses above US$1 billion, about a tenth. With around 3% of homes insured nationwide and GEM estimating an average annual loss of US$273 million for the building stock, the vast majority of Peru's earthquake risk is uninsured.

Does Peru have a catastrophe bond?

It did. Peru placed a US$200 million tranche in the 2018 Pacific Alliance catastrophe bond arranged by the World Bank. The 2019 magnitude 8.0 Loreto earthquake triggered a 30% payout of US$60 million. The bond matured in February 2021 and no renewal is documented; sovereign protection currently relies on contingent credit lines.

Modelled and validated

The Xpectral Peru model is built and validated by Dynamis, engineering consultants in earthquake engineering and structural dynamics with 100+ projects in 20+ countries across five continents, and benchmarked against the national code and the GEM global reference before any loss figure reaches a portfolio.