15.10. *Converting units of quantity*. There are two means of converting reported units of quantity to standard HS units of quantity, namely, (a) mathematical conversion of the reported units to the standard units and (b) conversion from one unit to another unit using, for example, the specific gravity of the commodity or commodities involved.

15.11. *Mathematical conversion*. Annex XV.A below gives examples of (multiplication) factors with which specific non-standard units can be converted to standard HS units of quantity. The table contains mostly units of quantity of the systems of measurement of the United States of America and the United Kingdom of Great Britain and Northern Ireland. Those factors are applied by the United Nations Statistics Division to convert units of quantity to WCO standard unit of quantity for certain HS headings. It is good practice to establish a comprehensive list of conversion factors and to publish and circulate this list among all agencies involved in the collection of trade statistics. There are other country-specific units of measurement, many of which apply to a single commodity. Commodity boards and other organizations publish conversion factors for some of those.^{[10]} Many of those commodity-specific sources are gathered together in other reference sources.^{[11]} Still other references deal with smaller groups of commodities.^{[12]}

15.12. *Specific gravity*. The use of specific gravity to convert, for instance, litres of a certain commodity into kilograms constitutes a much more complicated and less accurate approach, since it is based on empirical rather than mathematical principles. HS headings often contain a multitude of products which can all differ in, say, weight per volume or weight per unit. Even such seemingly homogeneous commodities as crude oil and milk will have different weight-per-volume indices, depending on country of origin and, for example, on the sweetness (for crude oil) or the concentration of fat or time of collection (for milk) (for examples of the various conversion factors, see annex XV.B below).

15.13. *Use of specific versus broad-based conversion factors.* The best conversion of volume into weight or of pieces into weight is carried out at the national or even the subnational level. For instance, to convert to cubic metres (m^{3}) the volume of sawnwood originally reported by Canada and the United States in 1,000 board feet, the Food and Agriculture Organization of the United Nations applies the country-specific conversion factor of 2.36 m^{3} per 1,000 board feet, as sawing conventions in those countries generally result in a volume that is smaller than the nominal volume.^{[13]}* *In contrast, broad-based conversions at the national or international level are inaccurate by definition and can serve the purpose only of making quantity (especially weight) estimates for general trade or transport analyses. The use of broad-based conversion factors by FAO is reflected in the examples below:

(a) When countries record coconuts in number instead of weight, amounts are converted to weight on the assumption that, on an average basis, 1,000 nuts = 1 metric ton, unless official conversion factors are available;

(b) Refined sugar is converted to raw sugar equivalent using the factor 1.087 for all countries;

(c) Quantities of wine, vermouth and similar beverages are expressed in weight; for countries recording their statistics in volume, it is assumed that 1,000 liters = 1 metric ton.

[10] See, for example, Quarterly Bulletin of Cocoa Statistics.

[11] See, for example, The Economist Desk Companion: How to Measure, Convert, Calculate and Define Practically Anything (New York, John Wiley and Sons, 1998).

[12] See, for example, “Weights, measures, and conversion factors for agricultural commodities and their products”, Agricultural Handbook No. 697 (Washington, D.C., United States Department of Agriculture, June 1992).

[13] It is estimated that for the United States, taking coniferous and non-coniferous data together, the actual average volume of rough green sawnwood would be 3 per cent less than the nominal volume, while the weighted average for surfaced dry coniferous and rough dry non-coniferous sawnwood would be 27 per cent less than the nominal volume”. Further details are available from the FAO website (www.fao.org).