3.1.1A Supply and Demand Analysis of IT and Productivity

At the core of the infamous productivity paradox is the assumption that computers and information technologies in general must have visible effects on productivity. Berndt and Malone (1995) described this common belief:

‘From the simple observation that computers can do certain kinds of things much faster and less expensively than individual people can, it is a natural leap to assume that replacing selected employees of a business with computers will greatly increase the speed and reduce the costs of certain business activities.’

Thus, in the early 1980s people were buying computers expecting important productivity gains. Berndt and Malone believed these gains were visible until the late 1980s when the productivity paradox was revealed. How is it possible for information technology equipment to affect productivity?

Using a neoclassical supply and demand analysis approach, Denison (1985, 1989) pioneered an original framework for assessing the role of information technology in economic growth. Sichel (1997) developed this analytical framework, which helps understand why productivity growth has remained sluggish while information technologies were booming. Note that in Sichel’s analysis, “computer hardware” only is considered when measuring IT capital, whereas this study considers the broader category information processing equipment. Hence, I will alternatively use the word “computers” for “information technology.” Sichel also distinguished two sectors in the economy: the computer-producing sector and the computer-using sector. He argued that the large increase in computer spending over the last twenty years was mainly driven by a considerable price decline that resulted from important productivity gains in the computer-producing sector. These gains were the results of dramatic technological progress, mainly in the manufacture of computer components. The supply-driven price drop caused real investment in computers to increase every year, as illustrated in the simple supply and demand framework in Figure 3.1.

In a neoclassical world, economic agents always make optimal investment decisions and all types of capital earn the same marginal return. After comparing returns on investment and costs of capital, firms stop buying computers at point A where the economy is in equilibrium. Now suppose technological progress in the computer producing sector shifts the supply curve from S1 to S2. The price of computers drops from P1 to P2. Then, the cost of computer capital is less than its return and firms invest. Hence, the quantity of computers increases from KIT1 to KIT2. As a result, the equilibrium point shifts from A to B. The question is how and by how much this increase in computers will affect output growth?

Figure 3.1Supply and Demand Framework for Computers

In the neoclassical framework, suppose computers earn a competitive return r COMP. Let ΔKIT represent the change in the capital stock of computers from one year to the next (KIT 2 – KIT 1). The neoclassical boost to the level of output from computers only is the product of the change in the stock of computers and the competitive return to computers. If Y 1 and Y 2 denote output levels at time t-1 and t, respectively, then

Y 2 – Y 1 = r COMP ΔKIT(3.1)

If output Y and capital input K are measured in real quantities and depreciation affects the return to computers then

Y 2 – Y 1 = [(r COMP + d) (P KIT / P Y)] ΔKIT(3.2)

Where d is the rate of depreciation, P KIT and P Y express the respective price index of IT capital and output. This is an expression of the increase in real output due to an increase in IT capital input. Sichel then divided both sides by Y 1 and multiplied the right-hand side by KIT 1 / KIT 1, which led to

(Y 2 – Y 1 ) /Y 1= [(r COMP + d) (P KIT / P Y) (KIT 1/ Y 1)] (ΔKIT / KIT 1 )

⇔ gr (Y 1) = [(r COMP + d) (P KIT KIT ) / (P Y Y)] gr(KIT ) (3.3)

where gr(.) represents the growth rate of the variable in parenthesis and P KIT KIT is the nominal stock of IT capital, which earns a return of r COMP + d. The product of these terms yields the total income flow generated by IT capital. Dividing this term by total income gives the share of income generated by IT capital, s IT such as

gr (Y 1) = sIT gr (KIT)(3.4)

where s IT is (in nominal terms)

s IT = [ ( r COMP + d ) KIT] / Y(3.5)

A decline in computer price leads to higher computer investment, which induces growth in output as seen in equation 3.4. Another way to measure the impact of the growth in IT capital on output growth is discussed next.