58 Generating Economic Cycles 



" In order to determine the fluctuations we require 

 to know not only the actual price, but whether that 

 price is above or below the average for that time. 

 It becomes necessary then to average the prices in 

 some way so as to obtain a standard for each year, and 

 we can then determine whether the price for any 

 particular year is high or low according as it is above 

 or below that standard. I have found it sufficient 

 to average for ten years at a time, that is, I have taken 

 as the standard for each year the average of the 

 ten years of which that year is the fifth. If a curve 

 be drawn whose ordinates represent these standard 

 prices, it will be seen at once that all the larger 

 irregularities are nearly smoothed down. . . . 



'' It would now be possible to represent the rises and 

 falls in price by comparing the price for each year 

 with the standard for that year. But there are so 

 many irregularities of short duration, say two or 

 three years, that it is more convenient to take, in- 

 stead of the price for each year, the average for a 

 short period, and for this purpose I adopt four 

 years. The price for any one year then to be com- 

 pared with the standard, is the average for the four 

 years of which that year is the second. 



''Were there only very small variations in the 

 standard, it would be sufficient to take the differ- 

 ence between the ten-yearly and the four-yearly 

 averages. But the standard varies very considerably 

 . . . The higher the standard, the greater are the 

 differences between it and the four-yearly average. 

 To obtain results which may be compared with each 

 other at different times, this effect of change of 



