In the Bank of Lebanon data for October 2025, exposures above £50 billion account for 87.9% of the portfolio studied and 89 % of the debt at risk. This weight explains why an improvement on small amounts is not enough to change the overall indicator. It also shows why the decline in a ratio does not always prove an improvement in repayments.
The case of at-risk claims in Lebanon is not an average. BDL’s macroeconomic review of December 2025, released in April 2026, compares several tranches between February and October. It allows simultaneous observation of portfolio composition and changes in each group. These dated data do not constitute a current assessment of each bank.
Risk claims in Lebanon: respect the perimeter of the figure
The BDL includes loans in risk categories 4, 5 and 6. It shall report their amount to the exposures within its scope, specifying that the denominator excludes depreciated claims designated as « credit-impaired loans ». According to this definition, the ratio reached 86.3% in October 2025, as in February.
This rate is therefore not a proportion of all assets in the banking system. It does not mean that 86.3% of depositors would have lost their savings, nor that the same share of the amounts concerned would be definitively uncollectible. It describes a specific classification and scope. A comparison with another country would first require checking that the definitions are identical.
The classification of a claim gives information on its risk. The final loss also depends on future repayments, guarantees, costs and recovery times. Without a file review, the aggregate ratio does not make it possible to estimate what will actually be recovered. Nor is it possible to assign individual responsibility.
Why a drop can come from the denominator
In the less than £10 billion range, the ratio rose from 78.4% to 69.9%. The decrease is 8.5 percentage points. However, the amount of risk claims decreases by only around 3%, while the total exposures in this tranche increase by 9%. The improvement in the ratio is therefore largely due to the increase in the denominator.
A purely pedagogical example helps to understand this mechanism. If a portfolio contains 80 risk claims out of 100 exposures, its ratio is 80 %. If claims at risk remain at 80 but exposures increase to 110, the ratio drops to 72.7%. No risk claims were repaid in this example. Numbers illustrate a calculation, not a real bank.
The £10 to £20 billion range is even clearer. Its ratio decreased from 83.4% to 80%, while the amount of risk claims increased by 1%. Total exposures increased faster by 6%. The decrease in the percentage therefore coexists with an increase in the amount classified at risk.
| In billions of pounds | Ratio February 2025 | Ratio October 2025 |
|---|---|---|
| Less than 10 | 78,4 % | 69.9% |
| 10-20 | 83.4% | 80.0% |
| From 20 to 30 | 86.7% | 85.8% |
| 40-50 | 86.8% | 88,0 % |
| More than 50 | 86.8% | 87.4% |
These tranches are expressed in billions of pounds, not dollars. The table presents a selection of documented categories; it does not constitute a list of banks or a ranking of enterprises. The figures do not give the number of borrowers corresponding to each amount.
The determining weight of the largest exposures
In the tranche over £50 billion, the amount of risk claims decreased by 3.1 %. However, exposures decreased further by 3.7%. The ratio thus rises from 86.8% to 87.4%. This is the opposite of that of the small tranches: a decrease in the amount at risk can accompany a deterioration of the percentage.
This group dominates the consolidated result because it represents nearly nine tenths of the exposures studied. A sector average is weighted by the amounts. It does not give the same weight to each tranche or borrower. The evolution of a very important set can thus neutralise progress elsewhere.
The BDL presents a simulation to measure this sensitivity: reducing risk claims from the largest tranche by 20%, keeping total exposures constant, would increase the overall ratio from 86.3% to about 70.9%. This calculation is neither an announced objective nor a repayment forecast. It isolates the effect of the dominant group under an explicit hypothesis.
Average does not replace file review
The maintenance of the overall ratio of 86.3% between February and October therefore masks different movements. Some amounts decrease, others increase; portfolio sizes also change. To monitor the quality of the credits, the rate, its numerator and denominator must be read together.
In the separate issue of banking restructuring, the IMF communiqué of 13 February 2026 stressed the need for an independent, transparent and effective process. Aggregated statistics cannot replace this assessment. Two institutions exposed to different portfolios do not necessarily have the same recovery capacity, even when the national average appears stable.
The next useful element will therefore be sufficient detailed information on the movements of the portfolio: refunds, new classifications, exits and guarantees. The 2025 data already establish a precise point: the weight of the largest exposures largely controls the overall indicator, while the percentage changes must be separated from the changes in the amounts at risk.
Comments