ANNUAL REPORT 2009COMPARISON OF CANADIAN PRICES TO FOREIGN PRICESTables 13 and 14 provide detailed statistics comparing the foreign prices of patented drug products to their Canadian prices. Each table provides four sets of average price ratios. These are differentiated according to (1) the averaging formula applied, and (2) the method by which foreign prices were converted to their Canadian dollar equivalents. The tables also give the numbers of drug products (DINs) and the volume of sales encompassed by each price ratio reported.^{18}
The PMPRB has traditionally reported average foreigntoCanadian price ratios constructed as salesweighted geometric means of individual ratios. Such results are included in Tables 13 and 14 (under Geometric Mean). The tables also provide results obtained using a salesweighted arithmetic average (under Arithmetic Mean).^{19} These statistics provide answers to questions of the type:
“How much more/less would Canadians have paid for the patented drug products they purchased in 2009 had they paid Country X prices rather than Canadian prices for these products?” ^{20}
For example, Table 13 states that the 2009 average FrenchtoCanadian price ratio obtained using the arithmetic mean was 0.90. This means Canadians would have paid 10% less for the patented drug products they purchased in 2009 had they bought these products at French prices.
For many years, the PMPRB has reported average foreigntoCanadian price ratios with foreign prices converted to their Canadian dollar equivalents by means of market exchange rates. (More exactly, the 36month moving averages of market rates the PMPRB normally uses in applying its Guidelines.)
Table 13 also reports foreigntoCanadian price ratios with currency conversion at purchasing power parity (PPP). The PPP between any two countries measures their relative costofliving expressed in their own currencies. In practice, costofliving is determined by pricing out a standard set or “basket” of goods and services at prices prevailing in each country. Because PPPs are designed to represent relative costofliving, they offer a simple way to account for differences in national price levels when comparing individual prices, incomes and other monetary values across countries. When applied to the calculation of average foreigntoCanadian price ratios they produce statistics answering questions of the form:
“How much more/less consumption of other goodsandservices would Canadians have sacrificed for the patented drug products they purchased in 2009 had they lived in Country X?”
Questions of this type cannot be answered by simply comparing drug prices. Rather, one must first calculate what each price represents in terms of goodsandservices foregone. PPPs are designed for such purposes.
TABLE 13 Average ForeigntoCanadian Price Ratios, Bilateral Comparisons, 2009
At Market Exchange Rates

Canada 
France 
Italy 
Germany 
Sweden 
Switzerland 
United Kingdom 
United States 
Geometric Mean 
1.00 
0.84 
0.80 
1.08 
0.93 
0.98 
0.90 
1.71 
Arithmetic Mean 
1.00 
0.90 
0.86 
1.15 
0.99 
1.03 
0.96 
1.85 
Number of DINs 
1,180 
762 
765 
868 
851 
817 
855 
992 
Net Revenues ($Millions) 
13,334.0 
11,606.8 
10,884.6 
11,811.4 
11,915.2 
11,827.4 
11,728.6 
12,661.1 
At PurchasingPowerParities

Canada 
France 
Italy 
Germany 
Sweden 
Switzerland 
United Kingdom 
United States 
Geometric Mean 
1.00 
0.72 
0.73 
0.98 
0.76 
0.74 
0.81 
1.85 
Arithmetic Mean 
1.00 
0.76 
0.79 
1.05 
0.80 
0.78 
0.87 
2.00 
Number of DINs 
1,180 
762 
765 
868 
851 
817 
855 
992 
Net Revenues ($Millions) 
13,334.0 
11,606.8 
10,884.6 
11,811.4 
11,915.2 
11,827.4 
11,728.6 
12,661.1 
BILATERAL COMPARISONS
Table 13 provides bilateral comparisons of prices in each of the PMPRB´s seven comparator countries to corresponding Canadian prices. Focusing on the results with currency conversion at market exchange rates, it appears that, as in previous years, Canadian prices were roughly in the middle of the pack on average. Prices in Italy and France were, on average, appreciably lower than Canadian prices. As in previous years, U.S. prices were substantially higher than prices in Canada or any other comparator country.
Average price ratios obtained with currency conversion at PPPs, provided at the bottom of Table 13, tell a somewhat more dramatic story. Once international differences in costofliving are accounted for, it appears that Canadians incurred a substantially greater consumption cost for the patented drug products they purchased in 2009 than did residents of every comparator country other than Germany and the United States.
Figure 11 puts these results in historical perspective. In 1998, Canadian prices were, on average, higher than prices in France and Italy but below prices in the five other comparator countries. This pattern was largely unchanged as of 2003. In 2009, Canadian prices were, on average, decidedly above prices in Italy and France, much below prices in the United States, but within a margin of plus/minus 10% of prices in Germany, Sweden, Switzerland and the United Kingdom.
AVERAGE FOREIGNTOCANADIAN BILATERAL PRICE RATIOS: ANALYSIS OF CHANGES
By and large, the international price comparisons reported above are very similar to those reported in last year´s Annual Report. The largest change involves the average U.S.toCanadian price ratios obtained at market exchange rates, which have risen considerably (from 1.63 to 1.71 in the case of the geometric mean, and from 1.76 to 1.85 in the case of the arithmetic mean). In light of the method used to derive these ratios, there are five factors that might account for this change:
(1) a change in currency conversion factors that acts to raise the Canadiandollar equivalents of U.S. prices
(2) rising U.S. prices
(3) declining Canadian prices
(4) a change in the set of encompassed drug products that on balance favours products with higher U.S.toCanadian price ratios
(5) a shift in salesweights that on balance favours drug products with higher U.S.toCanadian price ratios
Further analysis reveals the rise in average U.S.toCanadian price ratios was entirely the result of rising U.S. prices. Changes in salesweights acted to somewhat moderate the impact of U.S. price increases. Changes in other factors had little impact on the average price ratios.
MULTILATERAL PRICE COMPARISONS
Table 14 provides average foreigntoCanadian price ratios using several multilateral measures of foreign prices. The median international price (MIP) is the median of prices observed among the seven comparator countries. Other multilateral price ratios compare the minimum, maximum and simple mean of foreign prices to their Canadian counterparts.
TABLE 14 Average ForeigntoCanadian Price Ratios, Multilateral Comparisons, 2009
At Market Exchange Rates

Median 
Minimum 
Maximum 
Mean 
Geometric Mean 
0.98 
0.73 
1.85 
1.11 
Arithmetic Mean 
1.04 
0.79 
1.98 
1.16 
Number of DINs 
1,112 
1,112 
1,112 
1,112 
Net Revenues ($Millions) 
13,044.6 
13,044.6 
13,044.6 
13,044.6 
At Market Exchange Rates

Median 
Minimum 
Maximum 
Mean 
At PurchasingPowerParities
Geometric Mean 
0.84 
0.64 
1.92 
1.02 
Arithmetic Mean 
0.90 
0.71 
2.06 
1.07 
Number of DINs 
1,112 
1,112 
1,112 
1,112 
Net Revenues ($Millions) 
13,044.6 
13,044.6 
13,044.6 
13,044.6 
Source: PMPRB
Focusing again on results at market exchange rates, the average MIPtoCanadian price ratio stood at 0.98 in 2009 applying the geometric mean, and 1.04 at the arithmetic mean. (The corresponding values for 2008 were 0.97 and 1.02.)
Figure 12 puts this result in historical perspective. MIPs were on average 19% less than corresponding Canadian prices in 1987. By 1998, MIPs were on average 14% higher than Canadian prices. The average MIPtoCanadian price ratio has remained within 3% of parity since 2006.
Results obtained with other multilateral measures are much as one would expect. Interestingly, it appears that mean foreign prices typically produce higher foreignto Canadian price ratios than do MIPs. This is readily explained by the influence of U.S. prices, which are typically much higher than prices elsewhere. U.S. prices nearly always figure importantly in the calculation of the mean foreign price but almost never emerge as median international prices.
As with the bilateral comparisons, differences between results obtained at market exchange rates and at PPPs are striking. These affirm that while Canada may be a “medium price” country in purely monetary terms, Canadians actually sacrifice appreciably more consumption to acquire patented drug products than do residents of most comparator countries. With currency conversion at PPPs, the average MIPtoCanadian price ratio (calculated as a geometric mean) was 0.84 in 2009, substantially less than the value of 0.98 obtained at market exchange rates.
Figure 13 offers more detail on the productlevel MIPtoCanadian ratios underlying the averages reported in Table 14. This figure distributes the 2009 sales of each patented drug product according to the value of its MIPtoCanadian price ratio (more exactly, according to the range into which the ratio fell).^{21} These results show substantial dispersion in productlevel price ratios: while patented drug products with MIPtoCanadian price ratios between 0.90 and 1.10 accounted for 35.8% of sales, those with ratios less than 0.90 accounted for 34.1% of sales, and products with ratios exceeding 1.10 accounted for 30.1%.
^{18} The number of drug products and sales encompassed vary among comparator countries because it is not always possible to find a matching foreign price for every patented drug product sold in Canada. It is worth noting in this regard that all of the average price ratios reported in Tables 13 and 14 cover at least 81% of 2009 Canadian sales. The reported U.S.toCanada price ratios cover about 95% of 2009 sales.
^{19} Let RG represent the average price ratio obtained using the geometric method, RA the average price ratio obtained using the arithmetic method. Let p(i) represent the Canadian price of drug i, pf(i) its foreign price (converted to Canadian dollars) and w(i) its share of Canadian sales. Then RG = Ð [pf(i)/p(i)]w(i) (where Ð signifies multiplication over all patented drug products), while RA = Ó w(i)[pf(i)/p(i)] (where Ó signifies summation over all patented drug products).
It is readily demonstrated that RG can never exceed RA. It is also possible to show that the difference between RA and RG will increase with the extent of variation among individual price ratios, and that RG will equal RA only in the special case where all productlevel price ratios have the same value.
^{20} The difference between these two statistics, however, is that while the geometric mean provides an approximate answer, the arithmetic mean provides an exact answer. Consequently, as of 2010, the PMPRB will be using only the arithmetic mean.
^{21} To produce the results in this figure, foreign prices were converted to their Canadiandollar equivalents using market exchange rates.
