Same reaction here--that sure looks like an anomaly or just plain Excel contamination. Then I put on my thinking cap and asked: what if I modeled this as a naive time series? Does assuming that the current month will be identical to the immediately past month tell me anything? Well, I won't hijack your thread, but I did find an interesting result concerning autocorrelation of residuals.
Same reaction here--that sure looks like an anomaly or just plain Excel contamination. Then I put on my thinking cap and asked: what if I modeled this as a naive time series? Does assuming that the current month will be identical to the immediately past month tell me anything? Well, I won't hijack your thread, but I did find an interesting result concerning autocorrelation of residuals.