We have seen that productivity is a ratio. The numerator is some measure of output, the denominator is some measure of input. There are physical productivity measures. We can calculate the tonnage of wheat grown per hectare; we can count the number of cars produced per worker; we could calculate the number of examination passes per 100 students. Leave aside whether this would capture or explain everything we are interested in. These physical measures of productivity can only take us so far. Economists are more interested in inputs as a cost and outputs as value. This potentially gives us the three different types of productivity measure, as set out in Figure 2.1.

If we take land productivity, we can still have a single-input measure that divides the monetary value of output by a physical input or a value input. In agriculture the value of agricultural output per unit of land input is an example. We can also look at labour productivity in terms of the value of output per person (effectively an output per head calculation), per worker or, best of all if the data are available, per hour worked. These measures are some of the most basic and widely used measures of productivity. When it comes to the productivity of capital things are different. We cannot add up a plough, a hammer and a computer. We can only measure the value of output against the value of the capital used. To measure the value of output against all inputs too, we need prices. We cannot add hectares of land to numbers of workers and machines as inputs. And we cannot add the weight of wheat to the number of cars or the meals cooked as outputs. We have no choice but to add up the cost of the inputs and the value of outputs.

Table 2.1 sets out the basic productivity measures that economists use following this logic (OECD 2001).

Notice that there is an additional element shown in Table 2.1. This is the idea of an unknown component, or “knowledge” component, embodied in the changing technology of doing things.