Prediction-error expansion (PEE) method in reversible watermarking is an important technique which can embed large payload into digital images with low distortion. This technique makes use of prediction error to embed data into an image. In this technique, the correlations among prediction-errors are not considered and not utilized in current PEE based techniques. But, the prediction-errors are modified individually in data embedding. In this paper we propose a new idea in reversible data hiding framework which is using prediction-error expansion (PEE) with Pair in image authentication. This algorithm is named as pairwise prediction-error expansion. We propose that every two adjacent prediction-errors compositely and make pair. All pair consider to generate a sequence which consisting of prediction-error. Pixel selection is used for sorting in decreasing order of prediction pair. To make an entry of prediction error pairs based on magnitude. Based on the sequence, we making 2D prediction-error histogram. A more efficient embedding strategy can be designed where expanding and shifting in histogram bins. To achieve an improved performance we use pairwise prediction error expansion and, a reduced size of location map. This may help us to improve capacity into the image with less distortion.