Methodological Choices in ICT

Some methods are used to estimate the contribution of the new economy to hourly productivity growth. The accounting methodology used here is based on a standard transfer function.

This is the method whose purpose is to take into account the improvement of the quality of the factors of production and to highlight the substitutions between the different categories of inputs used.

Although this method is commonly used for this type of exercise, it is used to calculate hourly productivity and compile the series needed to identify its sources. It is not exempt from assumptions and difficulties. The main ones are related to the calculation of the share of ICT investments in the added value and the volume/price distribution.

Self-employment accounting, which is necessary to ensure consistency with value-added series across the economy, poses particular problems. Often not available. Data on hours worked and compensation are subject to estimates.

Average working time of employees applies to all jobs, employees or self-employed. Gross operating surplus is calculated by default as value added less wages.

It is therefore adjusted for the mixed income of the self-employed attributable to wages. For France, INSEE separates the share of value-added EBITDA from mixed income.

Therefore, we start from the EBITDA series to determine the fee. Measuring prices is particularly important for such applications. Indeed, it is very important to anticipate developments in volume. It is derived from data on values and prices. Growth alone is taken into account in the accounting breakdown.

In terms of value, it is necessary to better distinguish between price effects and volume effects in ICT investments. National accountants in the United States and Europe use alternative methods of matching methods or the cost approach of factors: hedonic regression methods.

Rather than simply measuring the evolution of the unit price of products, these methods take into account the evolution of their performance.

In the case of products such as computers, the exponential evolution of their performance is considered. Whether it's computing capacity, storage capacity, or functionality, nuance matters.

Hedonic methods that take these changes into account provide many conveniences. It makes it possible to determine the extraordinary decrease in the prices of computer equipment and the increase in the volumes produced. (It is no longer measured in mere units of goods, but in performance-equivalent units.)

In France, INSEE produced such series only for computers from 1990 onwards. The extremely rapid development of performance is not taken into account for other ICT equipment.

To correct this bias, the ICT equipment price series are aligned in the estimates presented below using the suggested methods.

The ratio of the price of non-ICT investments to the price of ICT investments is looked at. It involves re-estimating the evolution of ICT investment prices, assuming that this ratio is equal to that observed in the United States.

The contribution of France's ICT investments to hourly productivity growth is 0.09 points higher than would be achieved. In this case, this fix is not trivial.

The residuals are calculated as: TFP growth, partly cyclical phenomena, partly structural measurement errors, excluded variables and assumptions underlying the neoclassical production function.

SECTOR ANALYSIS

ICTs can be a source of productivity gains when produced or used. In most of the empirical studies conducted on the American economy, there have been significant productivity gains in the ICT producing sector. It highlighted an increase in ICT capital accumulation in all sectors.

To embody this observation and assess its scope, in our next article we will present an industry-level analysis of the impact of ICT on hourly productivity growth in France, the United Kingdom and the United States.

Yaşam Ayavefe