Our reinsurance segment reported gross premiums written of $172 million and $1.8 billion in the quarter and full year 2012, respectively, compared to $145 million and $2.0 billion in the corresponding periods of 2011. The full-year reduction was largely due to the repositioning of our catastrophe reinsurance portfolio.
Underwriting losses in our reinsurance segment were $28 million and $7 million in the fourth quarters of 2012 and 2011, respectively and reflected combined ratios of 106.0% and 101.5%. The current accident year loss ratios of 84.4% and 85.3% in the fourth quarters of 2012 and 2011, respectively, were both significantly impacted by natural-catastrophe losses. The current quarter's ratio includes aggregate pre-tax net losses (net of reinstatement premiums) of $153 million, or 33.6 points, related to Storm Sandy and an aggregate $15 million, or 3.3 point, reduction in our estimate for events of the first nine months (primarily Hurricane Isaac and second quarter U.S. weather events). Comparatively, the ratio for the fourth quarter of 2011 included aggregate pre-tax net losses (net of reinstatement premiums) of $111 million, or 24.0 points, related to the Thai floods and an aggregate increase in our estimate for catastrophe and weather-related events of the first nine months. Exclusive of these amounts, the primary driver of the reduction in the fourth quarter current accident year loss ratio was reduced exposure and loss experience related to aggregate property reinsurance of regional insurance companies in the U.S.
Net favorable prior period reserve development was $24 million, or 5.2 points, this quarter compared with $49 million, or 10.3 points, in the fourth quarter of 2011. The increase in fourth quarter general and administrative expenses was the result of higher performance-related compensation costs.
For the full year, our reinsurance segment reported underwriting income of $198 million, compared to an underwriting loss of $362 million for 2011. The significant decrease in the level of natural catastrophe activity was the primary driver of this variance.