(cont.)Can statistics tell the truth? --some tips and techniques for interpreting statistics--

    Then, what should we do for interpreting statistics? There are three major ways. Each way is introduced one by one below.

1) period-on-period (year, month, quarter) analysis:
By comparing the same period value (but different year), it gets possible to eliminate seasonally trends to some extent.

fig.3 : the developments of nominal GDP growth rate (nominal, original series) in Japan


2) moving average analysis:
    Moving average (MA) is used to reduce volatility. It’s good for grasping the basic trend of certain data, but may overlook shocks. In the case of figure 4, the decrease caused by so-called “Lehman crisis” appears one year after it really occurred.

fig.4 : Development of nominal GDP (4 quarter MA)