You are reviewing the monthly loss-reserve forecast for a mid-sized specialty insurer. Over the last six months, actual incurred losses have been running above forecast, and your CFO wants to know whether the forecasting process is drifting or simply noisy. You have the forecast and actuals for the same line of business, and you need to explain the size and direction of the error before next month's close.
| Month | Forecast Losses | Actual Losses |
|---|---|---|
| Jan | $4,800,000 | $5,100,000 |
| Feb | $4,900,000 | $5,050,000 |
| Mar | $5,000,000 | $5,250,000 |
| Apr | $5,100,000 | $5,400,000 |
| May | $5,200,000 | $5,550,000 |
| Jun | $5,300,000 | $5,700,000 |
How would you measure forecast accuracy here, diagnose whether the model is systematically underforecasting, and what would you recommend changing before the next forecasting cycle?