Turkey Earthquakes 2023: USD 104,000M in Losses, Only 5% Insured — the Protection Gap of the Decade
An unprecedented seismic doublet (M7.8 + M7.7 in 9 hours) exposed the largest protection gap of the 21st century — and the critical limitations of risk models that fail to capture directivity and correlated loading.
April 10, 2026
10 min readDynamis Associates × Appgile
On February 6, 2023, at 04:17 local time, a magnitude Mw 7.8 earthquake struck southeastern Turkey near Kahramanmaras, along the East Anatolian Fault. Just 9 hours later, at 13:24, a second Mw 7.7 earthquake hit the same region — an unprecedented modern seismic doublet that turned a disaster into a catastrophe.
With more than 59,000 fatalities (the deadliest earthquake globally since Haiti 2010), over 300,000 buildings damaged or destroyed, and economic losses exceeding USD 104,000 million in Turkey alone, the event exposed an uncomfortable truth for the insurance industry: only 5.3% of losses were covered. The largest protection gap of the decade revealed deep deficiencies in both risk modeling and catastrophe insurance penetration.
First earthquake Mw 7.8. Bilateral rupture of ~300 km along the East Anatolian Fault (EAF). Focal depth: 17.9 km. Left-lateral strike-slip mechanism. Directivity effects toward the northeast amplified ground motions in Hatay and Kahramanmaras.
04:17 – 05:00
Massive building collapse. Thousands of 5-10 story reinforced concrete structures suffered “pancake collapse” — stacked floor collapse. Victims were trapped under debris in the early morning hours.
13:24 local time — 9 hours later
Second earthquake Mw 7.7 on the Cardak Fault, 95 km north of the first. Structures already damaged by the first event collapsed. This second impact is what turned a severe disaster into a historic catastrophe.
Week 1 – 4
Damage assessment. Over 300,000 buildings damaged or destroyed across 11 Turkish provinces. Displaced: 3 million people. Sub-zero temperatures hampered rescue operations.
Q1 – Q2 2023
Claims crisis and criminal charges. TCIP/DASK (Turkish catastrophe pool) processed 870,000+ policies. Arrest warrants were issued for over 600 contractors and developers for construction deficiencies.
Why losses exceeded estimates
Insurance portfolios in Turkey in 2023 faced three fundamental deficiencies in seismic risk modeling that this doublet exposed in devastating fashion:
Issue 1
Erroneous construction classification
Buildings classified as “moderate-code” (Turkish 2007 code) were in reality pre-code quality. The Turkish seismic code is modern and adequate, but construction execution was catastrophically deficient: low-strength concrete, insufficient stirrups, smooth reinforcement bars.
Issue 2
Near-fault directivity not captured
Directivity effects generated velocity pulses that exceeded the design PGA by 2-3x. Recorded motions of up to 0.50g and velocities of up to 80 cm/s in the near field — something standard ergodic GMPEs did not predict.
Issue 3
Temporal correlation of the doublet
Two M7.7+ earthquakes within 9 hours. Loss models assume independent events, but the second earthquake impacted already weakened structures. The cumulative damage is not the sum of two events — it is exponentially worse.
Recorded PGA vs. design PGA by zone
Maximum observed acceleration at selected stations vs. design level from the 2007 Turkish seismic code
Sources: AFAD (Disaster and Emergency Management Presidency of Turkey), USGS ShakeMap, Melgar et al. (2023). Selected accelerograph stations.
What a modern PSHA would have revealed
A full probabilistic analysis — integrating 3D geometry of the East Anatolian Fault, near-fault directivity models, regionally adjusted GMPEs, and fragility based on actual construction quality — would have produced risk indicators drastically different from those insurers were working with in 2023.
EP Curves: Traditional model vs. full PSHA
Exceedance Probability for a C1-M building (concrete, 7 stories, actual pre-code quality) in Hatay. Replacement value: USD 2M
Illustrative values based on actual parameters from the 2023 earthquake and HAZUS-MH fragility curves for C1-M pre-code typology (actual construction quality).
Key indicators: Traditional model vs. full PSHA
For a C1-M building (reinforced concrete, 7 stories, actual pre-code quality despite nominally moderate code) located in Hatay province, with a replacement value of USD 2,000,000:
Indicator
Traditional Model (2023)
Full PSHA
Difference
PGA 475yr
0.25g
0.50g
+100%
PML 475yr
USD 640,000
USD 1,400,000
+119%
PML 2475yr
USD 960,000
USD 1,840,000
+92%
AAL
USD 8,400
USD 21,600
+157%
AAL Ratio
0.42%
1.08%
+157%
Loss Ratio 475yr
32%
70%
+119%
Direct implication for insurers
Under the traditional model, the technical premium for this asset was calculated at approximately USD 8,400/year. A full PSHA — incorporating near-fault directivity and actual construction quality — would have indicated a technical premium of USD 21,600/year — a 61% premium shortfall. Across a portfolio of thousands of buildings along the EAF, this shortfall translates to hundreds of millions in insufficient reserves.
Expected damage distribution
HAZUS-MH fragility curves applied to the actual PGA recorded in Hatay province (~0.50g) show that a C1-M building with pre-code construction quality had a 70% probability of sustaining extensive damage or collapse — consistent with destruction rates observed in the region, where over 60% of mid-rise concrete buildings were rendered uninhabitable or destroyed.
Damage probability distribution | PGA = 0.50g
C1-M building type (concrete, mid-rise, actual pre-code quality), Hatay
3% — No damage
8% — Slight damage (cosmetic cracks)
19% — Moderate damage (repairable)
32% — Extensive damage (irreparable)
38% — Complete collapse
Insurance market impact
Consequences for the Turkish and international insurance sector (2023-2024)
The largest protection gap of the decade and its ramifications
The 2023 seismic doublet generated a cascade effect across the global insurance market:
The protection gap of 94.7% — USD 98,500M uninsured — became the largest catastrophe coverage deficit of the decade. The Turkish state absorbed the majority as fiscal expenditure.
TCIP/DASK (Turkish Catastrophe Insurance Pool) paid a record USD 3,200M in claims. The pool covered residential only, not commercial or industrial. Its risk model severely underestimated directivity-driven losses.
Global reinsurers (Swiss Re, Munich Re, Lloyd's) reported an additional USD 2,300M. Catastrophe reinsurance premiums for Turkey rose 45% at the July 2023 renewal.
Earthquake insurance penetration in the affected zone was just 5.3% for residential and virtually nil for commercial — revealing an enormous untapped protection market.
“The Turkey earthquake demonstrated that the seismic protection gap is not a problem of the future — it is a crisis of the present. Models that assumed construction quality based on the prevailing code, without verifying actual execution, underestimated losses by a factor of 2 to 3.”
— Swiss Re Institute, Natural Catastrophes Review 2023
How Xpectral addresses these deficiencies
Our seismic intelligence platform is specifically designed to solve the three problems exposed by the 2023 doublet:
Complete global catalog
USGS + ISC + GEM since 1900
We automatically integrate the USGS FDSN catalog, the ISC Bulletin (historical events), and the GEM active faults database (GAF-DB). The East Anatolian Fault is fully modeled with 3D geometry and calibrated slip rates.
Site-specific Vs30
1 km² resolution
Vs30 data from USGS at 1 km² resolution. We identify soft alluvial soils in Hatay and Iskenderun (Vs30 180-270 m/s) that significantly amplified seismic waves vs. models using generic regional values.
Calibrated HAZUS fragility
36 typologies × 4 levels
Classification based on actual construction quality, not the nominally prevailing code. A building constructed under the 2007 Turkish code but with low-strength concrete is classified as pre-code — reflecting its true vulnerability.
Insurance metrics
PML, AAL, EP, SCR
Direct output in the metrics the industry needs: PML at multiple return periods, AAL, EP curves (OEP/AEP), damage distribution, and inputs for Solvency II SCR calculation.
The lesson: the protection gap closes with better modeling
The 2023 Kahramanmaras doublet was not an unexpected event. The East Anatolian Fault is one of the most studied seismic structures in the world. Turkey has suffered destructive earthquakes in Erzincan (1939, M7.8), Izmit (1999, M7.6), Van (2011, M7.1), and now Kahramanmaras (2023, M7.8+M7.7). The hazard was well documented.
The problem was not seismological science. The problem was threefold: (1) risk models that assumed construction quality based on the prevailing code without verifying actual execution, (2) GMPEs that did not capture near-fault directivity effects, and (3) catastrophically low insurance penetration that left 95% of property owners without coverage.
For insurers, the lesson is clear: the Turkish protection gap was not inevitable. With risk models that reflect the true vulnerability of structures and risk-adequate pricing, the earthquake insurance market in Turkey and similar regions represents an opportunity of USD 5,000-10,000 million annually in potential premiums — currently uncaptured.
By the numbers
Had portfolios exposed in the EAF region used a full PSHA model with construction classification based on actual quality and near-fault directivity, technical premiums would have been 100-157% higher for near-field assets — yet still insurable. The cost of PSHA analysis per asset is a fraction of the 0.01% of uninsured losses that materialized.
Disclaimer: Numerical values presented in comparisons are illustrative, calculated using actual event parameters but for a hypothetical asset. Each building requires an individual analysis with its specific coordinates, typology, and site conditions.
Frequently asked questions
Key data for insurance professionals and catastrophe risk modelers
Market What were the total losses from the 2023 Turkey earthquake?
Total economic losses from the 2023 Turkey double earthquake (M7.8 + M7.7) were estimated at USD 104,000 million according to the World Bank. Insured losses amounted to just USD 5,500 million, revealing a protection gap of 94.7% — one of the largest protection gaps in recent seismic catastrophe history.
Technical Why did the 2023 Turkey earthquake cause so many deaths despite having a modern seismic code?
Turkey adopted a modern seismic code in 2007, but construction execution was catastrophically deficient. An estimated 70% of collapsed buildings had severe structural deficiencies: use of sea sand in concrete, insufficient stirrups, undersized columns, and floors added without engineering calculations. The code was adequate; the construction was not.
Market What is the seismic protection gap and why was it so high in Turkey?
The protection gap is the difference between total economic losses and losses covered by insurance. In the 2023 Turkey earthquake, of USD 104,000M in losses only USD 5,500M were insured (5.3%). Contributing factors: TCIP/DASK only covered residential (not commercial), low penetration outside urban areas, and insured sums below actual replacement value.
Market What is TCIP/DASK and how much did it pay out in the 2023 earthquake?
TCIP (Turkish Catastrophe Insurance Pool), known locally as DASK, is the mandatory catastrophe insurance pool for residential properties in Turkey. In the 2023 earthquake it paid a record USD 3,200 million in claims, depleting a significant portion of its reserves and requiring activation of its reinsurance contracts and cat bonds.
Technical What PGA was recorded in the 2023 Turkey earthquake and how did it compare to the design code?
PGA values of up to 0.72g in Iskenderun and 0.58g in Nurdagi were recorded, exceeding the Turkish code design PGA of 0.4g for that zone by 2-3 times. The near-fault directivity effect, not captured by traditional ergodic models, generated velocity pulses that exceeded design spectra at critical structural periods (0.5-2.0 seconds).
Technical What impact did the double earthquake (M7.8 + M7.7) have on structures?
The second M7.7 earthquake, occurring just 9 hours after the M7.8, impacted structures already damaged by the first event, causing collapses in buildings that would have survived a single earthquake. This temporal correlation is not captured in standard risk models that treat each event as independent. An estimated 15-20% of collapses in the second event were due to this cumulative effect.
Cat Modeling What is the PML for a mid-rise concrete building in the Turkey earthquake zone?
For a C1-M building (concrete, mid-rise, pre-code construction quality despite nominally moderate code) in Hatay province with a replacement value of USD 2M, the PML at 475 years with full PSHA is USD 1,096,000 (loss ratio 54.8%). Traditional models estimated USD 500,000, a shortfall of 119%.
Pricing How did the 2023 Turkey earthquake affect reinsurance premiums?
Catastrophe reinsurance premiums for seismic risk in Turkey and the EMEA region increased by 45% at the 2023-2024 renewals. Several reinsurers reduced their risk appetite for the region, and Turkish cat bonds saw significantly wider spreads. The retrocession market also hardened due to the magnitude of losses.
Cat Modeling What is near-fault directivity and why does it matter for insurance?
Near-fault directivity is a phenomenon where fault rupture generates amplified velocity pulses in the direction of propagation. In Turkey 2023, this produced PGA values 2-3 times above the code in zones near the fault. Ergodic GMPE models average out this effect, systematically underestimating near-fault risk — critical for portfolios with concentrated exposure near active faults.
Market What market opportunity does the seismic protection gap in Turkey represent?
The 95% protection gap in Turkey represents an estimated market opportunity of USD 5,000-10,000 million annually in uncaptured premiums. To access this market, insurers need PSHA models that capture actual construction quality (not nominal), directivity effects, and temporal correlation between events — capabilities that traditional catastrophe models do not offer.