Shostak’s Town and My Statistical Settlers

Peter Shostak is a Canadian artist renowned for depicting the lives of people in rural Alberta, especially Ukrainian-Canadians. I happened to see one of his silkscreen prints, entitled What is the Name of this Town? at the Ukrainian cultural centre in Toronto in 1980. Shostak tells me this was a demo of 45 copies he did when he was beginning his career, and first exhibited at the Pysanka Ukrainian festival in Vegreville, Alberta in 1979. I had the good luck of buying one that hadn’t been sold in Vegreville.

The Print Itself

The print is only 6.5 inches by 9 inches but tells quite a story. In the bottom third, there are green and yellow fields of grain. The town consists of several small homes, two large red grain elevators, and a white onion-domed Orthodox Church. The sky dominates the rest of the print, with pillowy clouds and lines of water vapour rising. The point of view is approaching the town but still several miles away, either on foot or driving. The town, therefore, exists as a thin margin between the deep fields and the vast sky. I can imagine a traveler asking the town’s name as they see the town in the distance. After forty years of farm consolidation and rural depopulation, the question might seem sad or ironic if the town is now abandoned and forgotten.

In part, I was drawn to this print because it is so well-executed, with strong and contrasting colours, the image of the town as a margin, and its poignant question. I was also drawn to it because at the time I was close to completing a paper about western Canadian settlement that was taking almost a decade to write.

My Statistical Settlers

In 1972, I began the doctoral program in Economics at Harvard. One of the required courses was economic history, taught by Prof. Alexander Gerschenkron who was, to quote the title of an excellent biography by Nicholas Dawidoff, “an exceptional character.” Gerschenkron was an expert in European economic history who combined pioneering statistical research with a deep understanding of political economy. (I posted a blog about his reinterpretation of Hans Christian Andersen’s story The Most Incredible Thing.)

Gerschenkron required an essay, and the word in the common room was to use lots of statistics and to write about a part of the world he didn’t know as well as Europe. I decided to write about western Canadian settlement. This was influenced by the American economic historian Robert Fogel’s iconoclastic 1964 book Railroads and American Economic Growth, which argued that railroads didn’t greatly hasten settlement of the American west. Fogel came to this conclusion by comparing the economic impact of the actual railroad network in 1890 with that of a hypothetical network of waterways, canals, and wagon trails.

Canada attempted to hasten settlement by making free land available to homesteaders, advertising widely in Europe, and building a dense railway network. I wrote a paper for Gerschenkron about the impact of these policies, which he liked. That led me to take Gerschenkron’s economic history seminar in 1973, which required another paper. That year I was also taking econometrics from Zvi Grilliches, an imaginative scholar who developed new statistical techniques to study technological change. I decided to expand the previous year’s research about western Canadian settlement to write a paper I could mutatis mutandis use for both courses. (If you understand the Latin phrase you will realize I wasn’t cheating.) I expanded a time series of total homesteads started between 1882 and 1914 (32 observations) to homesteads started in 14 administrative regions over that period (32 x 14 = 448). This was not easy, because the boundaries of the administrative regions kept changing. The benefit was that the fourteen-fold larger data set provided much more scope for statistical analysis.

Both term papers were well-received. As I was beginning my academic career, I presented a paper based on this research at a Canadian economic history conference. My new colleagues were very impressed by my sophisticated statistical work. I could have published a paper at that point, but I still wasn’t completely satisfied so I decided to spend more time refining the independent variables in the database, namely those influencing the level of homesteading. This paper became a lower priority than my doctoral dissertation on airport planning, papers based on the dissertation, and a new project on the bilingual air traffic control project. But I kept chugging away on the homesteading paper, and completed it in mid-1981, just in time to have “Western Canadian homesteading in time and space” accepted by the Canadian Journal of Economics when I was being considered for tenure.

A Bittersweet Story

In writing this post, I took a serious critical look at the article 40 years after its publication. (If you are interested, here is the complete article.


To my credit, I built an extensive data base to document western Canadian homesteading as well as factors influencing the level of settlement, including Canadian government expenditure on settlement (including advertising); global wheat prices; the cost of transporting wheat from the farm gate to European markets on a year-by-year and region-by-region basis; and regional weather data, such as summer rainfall and frost-free days (pp. 20-21).

(With apologies to some readers, this paragraph is a bit technical.) I also displayed the econometrics I learned from Zvi Grilliches. Grilliches always talked about “having a feel for the data,” which meant trial and error. It specifically involved comparing the residuals between actual and estimated values of the dependent variable (the number of homesteads) and seeing if there were patterns indicating the residuals weren’t really random. If so, there were various econometric techniques that could be used to fix these problems (see p. 23).

Now I will be my own harshest critic.

The word “indigenous” doesn’t appear once in my paper. Like other colleagues who wrote about western Canadian settlement, we completely ignored First Nations. Implicitly, we saw them as an impediment to settlement; when they were rounded up and put on reserves, the west was now “open for settlement.” We didn’t realize it then, or perhaps it didn’t matter to us, but we were writing the colonizers’ economic history.

I also asked the so what question about my paper. Unlike Robert Fogel’s study of railroads, my results were not counter-intuitive. I showed that homesteaders were economically rational, but would we have expected otherwise? I showed that government policy (advertising and building railroads) had some impact on the rate of settlement, and my econometric model predicted that if more money had been spent on advertising and more branch lines built, settlement would have been a few years faster. My conclusions were similar to those of other colleagues who used much smaller databases. I expended a lot more effort to get similar results.

I looked the paper up on Google scholar, and it has had a paltry 17 citations over the 40 years it has been published. Many of my other papers, particularly on public sector innovation, have done much better. At the end of the appendix about the data, I wrote “This data set is available from the author on request.” No one ever asked.

The paper might have served as a calling card, introducing me to the small community of Canadian economic historians and the considerably larger community of econometricians using the same approach (panel data, which is discussed in the obituary for Zvi Grilliches linked above). But I chose not to follow up with either or both groups because I was more interested in other topics, so the paper represents a deliberate missed opportunity.

I think back on my research for this paper as a folly of youth – a willingness to attack an enormous task with boundless energy and an assumption that time is also unlimited. In subsequent years, when doctoral students described projects that reminded me of this one, I cautioned them, and sometimes they listened.

Despite this self-criticism, I do want to end with some sweetness. I had persevered and brought a project that took almost a decade to fruition. I had done the best possible work I could do, analyzing a large data set using what in 1980 was state-of-the-art econometrics and computer technology. I’m proud of both.

Finally, the research project led to my purchasing Peter Shostak’s excellent print. Shostak vividly reminds me of the physical beauty of the western prairie, and he also reminds me that the statistical settlers I wrote about were real people, with a deep connection to the soil and to their communities.

Leave a Reply

Your email address will not be published. Required fields are marked *