Abstract

The article focuses on interpretation of seasonally adjusted data. It has been held that, upon seeing an actual rise of 40 percent in monthly data, and finding that the seasonal indices showed a normal rise of 20 percent, most people would conclude that the seasonally deflated data would show a residual of 20 percent. The article examines what actually happens, and the resulting interpretations, by using a simplified numerical example. First, the way seasonal adjustment usually occurs with the help of a table indicating monthly cement production, seasonal index and seasonally adjusted production. The proportionate variation from one month to the next in seasonally adjusted data is equal to the proportionate variation of the actual value for the second month from that value which would have occurred for the second month if only the normal seasonal change had taken place from the first month to the second. The author holds that the correct interpretation of seasonally adjusted data is more complicated than commonly supposed.