Using Excel to forecast US unemployment rate through second half of 2026

To forecast the US unemployment rate through the second half of 2026, I looked at the historical data from July 31, 2022, through June 30, 2026. I chose July 31, 2022 as the starting point because this is when the unemployment rate returned to the pre-pandemic level of 3.5% on February 29, 2020. Coincidentally, this also provided a data set of 48 consecutive months, for a sequence of exactly four 12-month cycles.

The Excel desktop software uses a triple exponential smoothing algorithm as its forecasting method. This is a moving average that applies exponentially decreasing weights to historical values as they move further into the past. This method applies additional weights based on the overall trend, plus the cycle of 12 months, that it automatically detects from the data.

Finally, the monthly values forecasted into the future are enclosed within upper and lower bounds based on a 95% confidence interval. This means that within two standard deviations of the historical data set, we can be confident that this prediction interval will capture the actual future values in 95 out of 100 forecasts repeated using the same algorithm.

The results of my analysis are displayed in the following table and chart.

For the first half of 2026, the US unemployment rate has ranged between 4.2% and 4.4%. The above forecast suggests that the middle of the interval will range between 4.3% and 4.4%. Additionally, we can be 95% confident that it will reach 4.7% at worst, or 4.1% at best.