The Conference Board’s Saskatchewan Survey

Asking business whether to get rid of government regulations is much like asking a barber whether to get a haircut. Nevertheless, the Conference Board’s main (and only) evidence in projecting TILMA’s benefits for BC and Saskatchewan were surveys of business organizations and government departments. In our paper, Marc and I noted some serious problems with the BC survey as well as the arbitrariness of the economic estimates inferred from it. Dr. Helliwell made the latter point with respect to the Saskatchewan assessment.

However, Dr. Howe, the other professor who reviewed the Saskatchewan assessment, defends the survey methodology. In his view, “There are arbitrary elements to the quantitative analysis by the Conference Board just as there are arbitrary elements to any quantitative analysis” (p. 3 of cross-review). While he may be correct that “the standard approaches to measuring barriers to trade do not work” (p. 7 of review), at least these standard approaches are purportedly based on explicit, if arbitrary, models of the economy. The Conference Board does not even pretend to measure inter-provincial trade barriers in estimating TILMA’s supposed benefits. Howe’s contention seems to be that, since some arbitrariness is unavoidable, any amount of arbitrariness is permissible.

He concludes, “Professor Helliwell believes that he understands the size of trade barriers better than the businesspeople who deal with these barriers on a day-to-day basis. Readers should be offended.” However, on the previous page, Howe himself questions the business responses: “you must adjust for the self-serving biases of those who answer. Some Saskatchewan businesses that have successfully sought protection from outside competition will not be pleased with the lowering of the trade barriers, giving them an incentive to under report the gains from Saskatchewan signing TILMA. Consequently, we should expect that the gains from TILMA would be greater than suggested by the Conference Board’s survey” (pp. 2-3 of cross-review). But Helliwell asks, “would not the gainers be as likely to over-estimate the gains, for analogous reasons?” (p. 2 of cross-review).

Whatever the theoretical merits of these arguments, there are serious practical problems with the Saskatchewan survey. The Conference Board reports that “The survey was sent to a total of 118 persons: 17 representing the public sector and 111 from the private sector.” Unfortunately, 17 plus 111 does not equal 118. Fortunately, Appendix A indicates that there were, in fact, 17 public-sector entities and 101 private-sector entities (pp. 47-49).

The Conference Board goes on to explain, “we received a total of 34 responses, 9 from the public sector and 23 from the private sector.” (p. 5). Unfortunately, 9 plus 23 does not equal 34. It subsequently reports receiving 31 complete responses: 9 from the public sector and 22 from the private sector (p. 31). Perhaps it also received 3 incomplete responses from the private sector, for a grand total of 34 responses.

Despite this ambiguity, a couple of points are fairly clear. First, more than three-quarters of private-sector entities did not respond to the survey, which suggests that alleged inter-provincial barriers are not an important issue for Saskatchewan business. More generally, this low response rate casts doubt on the survey’s validity.

Second, most of the private-sector organizations surveyed are industry associations or chambers of commerce, which are ideologically committed to deregulation via TILMA and other means. Howe might be correct that some individual businesses benefit from provincial “protection.” However, since so few actual businesses were surveyed, it is extremely unlikely that this factor biased the results against TILMA. The pro-TILMA bias suggested by Helliwell is far more likely.

I do not know how much the Government of Saskatchewan paid the Conference Board for this assessment, but it should probably demand a refund.

Leave a Reply

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