Worthwhile Canadian Initiative: Lumbering Along in the Finding Data Process

Well, all I was trying to do was introduce a set of lecture slides on the nineteenth century timber trade with a simple overview of the Canadian logging industry's employment in the twentieth century. Well, three hours later it has proven to be a more frustrating exercise than I would have expected but here is what I have.

Historical Statistics of Canada has two series on employment in the lumber industry. Series L156 covers the period 1957 to 1975 while Series L160 covers the period 1917 to 1959. The two series do not coincide exactly for the 1957 to 59 period (eg. in 1957 L160 says there are 50,664 employees in Canada's lumber industry while L156 says there are 46,044) but their overlap is reasonable close. According to the notes accompanying the tables, the two tables provide for a three year over-lap for the 1957 to 1959 period to bridge the changes brought about by the 1961 revision of the S.I.C.

Of course,it would be nice to have a twentieth century that does not end in 1975 so I went off to the Statistics Canada site to see f I could pin down some logging employment numbers. There are a wide range of series and numbers which in the end reminded my of a line from one of books I used to read my children at bedtime..."Number, numbers infinitesimal, I like numbers, especially decimals..." Something I suppose only an economist could love.

I did find a monthly SEPH employment index from 1961 to 1983 (V74974) that goes from 110.2 in 1961-01 to 67.7 by 1972-01 and then 47.7 by 1983-01. I found monthly employment in forestry, logging and support services for 2001 to 2012 (v98436752) where employment starts at 70,896 in January 2001 and finishes in September 2012 at 37,978. Then there is another monthly series (v1596767) on forestry, logging and support employment that runs from 79,047 in January 1991 to 80,150 by December 2000. There is also a another monthly seasonally adjusted forestry employment series (v79311) that runs from 1961-01 to 1972-09 with employment going from 70,436 to 57,834. There is another seasonally un-adjusted monthly series (v248841) for employment in logging and forestry running from 1983-01 to 2000-12 that goes from 60,416 to 46,970 as well as an annual series for 1983 to 2000 (v249124) and runs from 65,744 to 67,973 employees. I'm pretty sure there may be others.

Well, aside from the employment index which I suppose I could use to try and interpolate something, I was still unable to easily find enough data for a complete series leaving a gap from 1976 to 1982. As well, the series that cover 1983 to 2015 show a fair amount of variation though two of them (V98436752 and V249124) seem to mesh rather nicely.

What is a simple country economist to do? Well, I decided to plot a number of them together and see if that might provide a reasonable picture as to the evolution of lumber industry employment. The accompanying figure puts six series together and if you decide to ignore v1596767 (and maybe V79311), then the remaining series paint an overview of growth up until the 1980s, a leveling off with some variation during the 1990s and then a precipitous decline after 2000. I suppose the gap between 1975 and 1983 can be left up to the imagination but there was the 1981-82 recession during this period though the average of the monthly employment index series suggests a period of falling employment. However, the employment index really complicates things because it suggests the whole 1961 to 1983 period saw declining lumber industry employment whereas the Historical Statistics of Canada series suggests growing employment.

Anyway, I'm not a labour economist or an expert on labour data so perhaps I am missing something here. Actually, I suspect I am missing quite a lot in terms of changes in definitions, coverage, survey methods,etc...Still, it has been frustrating. I think I'll skip the stuff on pulp mills.