On productivity

1. Introduction

Last year, the World Bank released its Information and Communications for Development Report. The study made clear that access to high-speed internet and mobile phone services is paramount for generating economic growth and job creation, in particular in developing countries.

Interestingly, the report highlights the role of e-government by suggesting that it is an ideal instrument to increase ICT penetration in any country – “…e-government could be a key driver for the penetration of ICT in almost any country regardless of its ICT infrastructure and ICT penetration" R. Sudan, Lead ICT Policy Specialist. The report also adds that the world is moving fast towards knowledge-based activities.

Provided that ICT are the foundation of knowledge-based economies the World Bank suggests that countries must improve and increase its use and application.

It is generally accepted that ICT improve standards of living and incentive economic growth. However, when it comes to measure the impact of such technologies and their contribution to productivity levels in the economy, figures tend to reveal otherwise.

This article tries to identify difficulties and challenges faced by governments in measuring the performance of public services and their contribution to the economy.

  

2. Productivity – Impact in the Economy and Government Policy

Productivity is a fundamental measure for business leaders, investors and policy-makers. In its most basic concept it can be defined as the relationship between output and one or more inputs used in the production process of a good or service (National Academy of Sciences, Washington D.C., 1979).

Its uses include analysing sources of economic growth and measuring and forecasting such economic magnitudes as potential GDP and labour requirements.

A slowdown in productivity invariably means a slower rate of growth in income per capita and consequently, in the standard of living. 

Alan Greenspan, addressing the productivity slowdown of the decade of 1990 in the US, mentioned that it was virtually unimaginable half a century ago “the extent to which concepts and ideas would substitute for physical resources and human brawn in the production of goods and services.” He also remarked that while the economic output is only modestly higher than 50 years before, value added has risen well over threefold.

Greenspan’s comments surged in a period when productivity rates were lower than expected and ICT was a key factor to attract investment. Moreover, statistics were showing the services sector causing the slowdown, establishing grounds for the views put forward by Solow (1987) and Baumol (1967).

Greenspan also touched a topic under debate for a long time – the “productivity paradox” as defined by Solow (1987). However, recent research such as the work carried out by Brynjolfsson (1994) have challenged both Solow and Baumol’s theories - “it appears that the shortfall of IT productivity is as much due to deficiencies in our measurement and methodological tool kit as to mismanagement by developers and users of IT”

In fact, Triplett and Bosworth (2002) analysed productivity growth in the US over the 1995-2000 period, and found that ICT equipment contributed to between 30% and 37% of growth in labour productivity of business services, wholesale trade and transportation.

 3. Measuring the Performance of Service Industries

There are at least two reasons why we need to get a clearer picture of the performance of service industries.

Firstly, services contribute towards an increasing share of economic activity - in 2005 the world average output share of services was almost 70% (Image 1). Secondly, services have become so integrated in manufacturing and industrial activity that even when a service is not traded, potential higher productivity of service providers can play a fundamental role in increasing a country’s market share in industrial products (Van Ark, Monnikhof and Mulder, 1999). It shall be noted that I refer to potential in the same way that David (1990) refers to the introduction of skylights in factories, turning them into lighter and cleaner workshops. In other words, there are a number of qualitative indirect benefits neglected by the current methods of measuring productivity that “come part as a package containing other gains” that “take the form of more readily quantifiable resource savings” (David, 1990).

Image 1 – Manufacturing vs Services. Source: World Resources Institute

 

Measuring productivity of services is neither an accurate process nor even an established one. Commonly, economists apply the same formulae used for measuring productivity in manufacturing industries.

However, there are substantial differences between these two dimensions of productivity. Services, as opposed to goods, are intangible and difficult to quantify. Consequently, it is often virtually impossible to identify what the output of a service is (Andersen and Corley, 2006).

Measuring productivity of producing goods disregards their actual consumption. Take as an example a car manufacturer that produces more vehicles than the rest of the automotive industry with the same input. As per current methods, this firm will be highly productive in spite of the fact that its vehicles might have no buoyant market.

On the other hand, the measurement of productivity of services introduces a new dimension – the consumer. Thus, outputs of service industries are analysed on a ‘transaction perspective’, i.e. services are only considered produced if they are also consumed (Andersen and Corley, 2006).

This raises a number of issues since a service typically involves a mixture of object entities, types of transformations and transformations spheres. Hence, services are generally heterogeneous (Andersen and Corley, 2006).

In all sciences, and indeed all aspects of life, categories consist of homogeneous individual components. Current measures of productivity recognise only services that can be quantifiable (Andersen and Corley, 2006) and grouped into categories. However, services are highly heterogenic making actual productivity measures inadequate and inaccurate, in particular when we consider bundled services. 

4. Productivity in the Public Sector

Similarly to other sectors, there is little evidence about how the widespread adoption of ICT in government bodies contribute towards an increase of productivity. Given the nature of their task, public services are bound to be more problematic than most businesses.

Garicano and Heaton (2007) examined how ICT contribute to organisational change and improved productivity in police organisations using a data set of police departments covering 1987-2003, in the US.

One of their major conclusions was that the adoption of ICT has no net relationship with an increase in productivity - measured in ‘clearance rates’, i.e. the number of crimes ‘cleared’ divided by the total number of crimes recorded - and is actually associated with an increase in offence rates.

Shocking as it may seem, one possible explanation is that offences might be higher in places with higher ICT adoption purely because officers may be more willing to file incident reports on a computer rather than by hand. Another explanation could be that it is now easier to report crimes than ever before.

Other topic that stood out is the assumption that greater use of ICT leads to a larger share of non-production personnel with little impact on productivity.

These findings reveal the static nature of the current methods for measuring productivity missing out on a variable that has been subject to increasing discussion – quality.

4.1  Quality

In the services sector quality is intrinsically associated to price. However, quality change is not reflected in price indexes, hence the adoption of hedonic pricing in some countries.

In the case of public services, the output is generally a non-market one - it is free or at prices not economically significant – thus, adding complexity to identifying the nature of outputs and measuring ‘transactions’.

There is an increasing emphasis on delivering quality services both in the private and in the public sector. “As a result there are greater demands on, and expectations of, measures of government output” (National Statistician, 2005)

In the UK, the Atkison Review (2005) which measures government output and productivity for the national accounts has corroborated previous research by arguing that “the impact of technological change may appear with a delay” and that “the benefits of ICT may appear in quality change that is not recorded.”

Analysing the impact of quality in the provision of a service is a difficult task and the outcome will always be subject to relativity, regardless of the method.

The Eurostat Handbook suggests that “part of the quality change can be captured by differentiating as many qualities of a product as possible. These different qualities are then in fact treated as different products” – an approximate concept to the one of hedonic pricing. However, it is also recognised that the quality adjustment cannot be exclusively limited to differentiation.

In the case of public transportation, Hill (1977) suggests that “the amount of services produced must be based primarily on the number of passengers transported and the distance over which they are transported” but that the quality of the service can be considered to account for comfort, speed, punctuality and safety.

If we make use of the Eurostat Handbook view, and calculate a quality adjustment factor for the example above, the result would be an aggregated indicator of quality which would be a combination of the different dimensions of quality, being comfort, speed, punctuality and safety - perhaps a way forward.

4.2  The Transaction Multiplier Effect

Another factor to account for is the difficulty in separating the different functions that contribute to an output. Going back to the police case, governments are faced with the challenge of quantifying the contribution of police, courts and prisons to the final output. These functions are in most cases interdependent and the effectiveness of each depends on the effectiveness of the others.

This is a very similar situation to the role of different departments within a private firm or even to the transaction multiplier effect in e-commerce. (Image 2)

Image 2

The Transaction Multiplier Effect of a Crime

Simple Illustration of the Transaction Multiplier Effect of a Purchase on Amazon

  

5. Conclusions

Statistics show that in spite of evidence of increased utilization of ICT in services sector, productivity growth has not shown signs of acceleration.

Nonetheless, there are several service industries revealing high productivity growth rates. One possible reason is the presence of returns to scale which might result from network effects in the production and use of ICT (Wölfl, A., 2003). This is particularly visible in “social productivity” where the benefits of using a certain technology increase more than proportionately to the number of users.

Firm share of value added in the service sector within the OECD have increased for market-related services such as finance, insurance and business services (Wölfl, A., 2003). At the same time, OECD studies have shown that small firm size is often found in service industries where productivity growth is negative or low, which is the case of the financial services.

Whatever the reason for low productivity growth of services, there is significant evidence that the problem lies to a great extent in the way it is measured. Research also suggests that there is a potential under-estimation of the contribution of services to the production of intermediate goods - as opposed to services provided to the final consumer.

David (1990) and Wolf (2003) also suggest that we did not allow enough time yet to see technological advancements having a real impact on productivity.

Garicano and Heaton (2007) on their research on police organisation, also suggest that there is a learning process associated with ICT. Hence the adoption of ICT can only become synonymous with productivity growth if police officers have the necessary skills.

Another problem that subsists in the knowledge intensive services is that paper-based procedures are retained alongside the new ones. The so-called Big 4 firms (KPMG, PwC, Ernst & Young and Deloitte) illustrate this by having an established procedure of keeping paper records alongside electronic ones.

There are also problems in categorisation and accounting. Accounting does not record novel technologies and therefore quality changes associated to the introduction of such technologies are not possibly to measure (David 1990).

David (1990) made a parallel between the dynamo and ICT and suggested that the widespread adoption of ICT is a key complementary element in foregoing innovation.

This reveals an urgent need for reviewing the way we measure productivity in the service industries in which quality is a central concern and should be accounted for. There are many dimensions of quality but there is way forward. If quality adjustments cannot be comprehensive, they should be representative of a relevant range of dimensions. That could not make more sense in a measure that assumes involvement of the consumer.