In his textbook "Elements of Forecasting", Francis Diebold presents on page 130 the following series describing it as the Canadian Employment Index, quarterly, seasonally adjusted.

enter image description here

The data for this series can be obtained from the author's site: https://www.sas.upenn.edu/~fdiebold/Textbooks.html under Forecasting-DataAndCode/R_Others/R_Misc/R_Employment

However, It is not clear how this series is indexed. The series doesn't take the value 100 in any quarter. Also it is now clear what exactly is plotted? Is it Employment level over time? If so, in FRED this data looks very different (see below) and indexing it shouldn't change much the way it looks. Is it Employment Rate?

enter image description here

The data source is not listed. I reached out to the author to ask these details about this data series but am still waiting for a response. Anybody know exactly what this series plots, where it comes from, and how I can reproduce it using publicly available data?


1 Answer 1


Why do you think it should be the OECD data you found on FRED?

100 is a base year, not quarter (seems to be 1991 here, assuming it starts in 1962.1 because it's the average quarterly value in that year that is set to equal 100). The source is more likely a national statistics provider and not the OECD (Statistics Canada here). I haven't found the original series (didn't spend much time but here is one for regions with an index). Newer values are most certainly re-based, so you need to search old datasets. The index is not a rate (rate means it's percent).

Ultimately, it's a intro to statistical forecasting book, even if the numbers were completely made up it wouldn't change anything. It's a bunch of numbers to conduct elementary statistics with.

  • $\begingroup$ Thank you for this answer. 1. Your point that the data in indexed to a year as opposed to a quarter was very good. It is indexed to year 1990 (with series starting in Q1 1961). 2. So the Employment Index just plots the number of people employed over time, indexed to a year? Did I get that right? 3. Makes sense as you say that newer datasets would be reindexed. I guess one part I got wrong is that indexing should not change much the shape of a dataset. Indexing can change the shape dramatically depending on which year we're indexing to. $\endgroup$ Commented Dec 27, 2021 at 9:00
  • $\begingroup$ economics.stackexchange.com/a/15956/37817 has some explanation. It's common to use base years (and it's usually never a shorter period but always averages thereof, like quarterly here). Also, definitions of (un)employment definitions vary a lot over time and between countries. www150.statcan.gc.ca/n1/pub/75-005-m/75-005-m2016001-eng.htm shows revisions in the Canadian survey. $\endgroup$
    – AKdemy
    Commented Dec 27, 2021 at 9:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.