Changing states of the economy

Author: Bill Eadie

The analysis described in this article arose from a need to describe concisely the detailed behaviour of the components of the economy, to provide explanation of the development of all kinds of port traffic, in particular following major shocks to the economy. The application of sophisticated statistical techniques to understanding how port traffic behaves is unique. The common approach applied by economists to such forecasts does not take account of the complex inter-relationships within the economy or the structural changes to the economy which arise during a major shock. The net result is that conventional economic forecasts assume that the shapes of activity return to pre-shock patterns of development. Through the application of statistical techniques it has been possible to show that this is not the case.

The conclusions are that the economy can be described as being in a relatively steady state with sector contributions changing smoothly relative to each other, while GDP in total follows a well-defined steady trend of growth. A major shock disrupts these smooth patterns. After such a major fluctuation, a similar state is eventually attained, with sectors again developing smoothly relative to each other. But the characteristics of these relationships are quite different from those holding before the shock. GDP also returns to a well-defined growth pattern, parallel to but displaced relative to the continuation its trend before the shock.

Placard at public sector cuts demonstration.

One union estimated up to half a million

people traveled from across the country to

demonstrate as the Conservative–Liberal Democrat

coalition implements wide-ranging spending cuts.

Image by Mark Ramsay/Wikimedia.

This combined situation, described by the level and trend path of the economy and the linear relationships between sectors, specifies a state of the economic process, which is maintained until the next major shock. The impact of such major fluctuations is to change the state of the economy to another, different state, with changed differentials between sectors, and the economy in total at a different level.

There is an obvious similarity to ecosystems subject to sudden shocks. An extreme example is a meteorite collision in the Jurassic period, after which a new quasi-stable system emerges minus most dinosaurs, with mammals evolving to fill spaces faster than birds. Another example is the behaviour of a complex atom with several electrons orbiting around a nucleus. The state of such an atom is defined by the levels of the orbits of the circulating electrons, and the direction and momentum of the atom as a whole. Each state of such an atom is maintained until a collision with another atom or particle occurs, at which point the orbits of the electrons are changed, and the trajectory of the atom as a whole is displaced to a different smooth path, followed until the next collision.

Recognition of these changing states of the economy is essential in constructing an explanation of port traffic in relation to changes in economic activity. Each type of port traffic relates to a specific combination of sector activity at points in time. Because the relationships of the sectors to each other, and to GDP in total, are different after a major shock, any model description of port traffic development must take account of such changes. In particular, any connection with growth in GDP in total will change, reflecting the discrete realignment in the composition of the economy that occurs after such shocks. To enable efficient estimation of the linkage between traffic, GDP and individual sectors, the fluctuations of each sector around its trend with GDP have to be understood. These residual variations are highly correlated and maintain their characteristics before and after major shocks. Identification of their eigenvalues enables most of the variation in economic sectors to be described by a small number of variables.

This process results in a concise and parsimonious description of variations in overall economic activity that can be used in simple but precise models of the development of port traffic components.

Analysis

Sector composition of Gross Domestic Product (GDP)

There are 21 sections in the Standard Industrial Classification (SIC 2007), generally labelled A to U. For the purposes of UK port traffic modelling we compress this classification to the following 13 categories:

For geographically wider-ranging studies, such as trade route analyses, the usual data source for comparable national data is the UN Macroeconomic Database, which uses a 7 category classification as follows:

In the following charts and analyses, these series are expressed in the form of indices relative to a selected base year. Figure 1 shows the UK sector variables from 1970 to 2008, with the total GDP series illustrating the average pattern of growth:

Figure 1: UK Sector Indices, 1970 - 2008

The complex patterns of development and change in the composition of the economy may be clarified by looking at the change of each sector relative to the total GDP; that is by assessing the changes in each sector’s share in the total economy. This view is shown in Figure 2:

Figure 2: Sector variation relative to GDP

It is evident that each sector follows a trend relative to GDP followed by relatively short periods of irregularity before returning to another steady trend. There are relatively smaller variations about these trends, of similar magnitude throughout the whole period. The interesting characteristics of these patterns are:

- The timing of the irregularities is similar for all sectors across the economy;
- The changes in the trend relationships with GDP after each irregularity.

The development of each sector relative to GDP is different after each period of irregularity, while still following a distinct trend relative to GDP. The overall pattern of these trends after the irregularity is quite different from the pattern before the irregularity. The behaviour of the economy may be described as being in a different (multi-dimensional) state after each irregularity, specified by the set of parameters describing these trends.

As illustrated in Figure 2 the changes of state of the economy correspond with the occurrence of major recessionary shocks to the economy.

GDP development

The changes of state of the economy occur at the time of major fluctuations in the economy. The effects of these shocks on the overall development of GDP are shown in Figure 3, plotting the GDP index on a semilog scale.

Figure 3: UK GDP, 1960 – 2010

It is evident that the irregularities of sector behaviour occur at times of major shocks to the trend of overall economic development. Such major shocks were described by Christopher Dow in his book ‘Major Recessions’ in which he identified important characteristics of these events that differentiated them from ‘normal’ recessions. His preferred description was ‘major fluctuations’. In particular he identified that such events destroyed productive capacity in its widest sense, including not only physical capacity in buildings and machinery, but also knowledge, technical skills, organisational and other intellectual capacity relevant to the production of goods and services. When demand recovers from such major fluctuations, it does so in a system with a reduced potential capacity, and which has to be restructured in order to enable the trend rate of growth to be reinstated.

From Figure 3, this pattern is clear for the UK. The major fluctuations in 1975, 1980, 1991 and 2008 set back the level of GDP, with future growth following a path parallel to the trend before the shock, but at a lower absolute level. As illustrated above in Figure 2, these major shocks also shift the pattern of sectoral development into a different state.

More detailed analysis shows that the pattern of GDP growth between major fluctuations follows a consistent pattern throughout the whole period. These patterns can be represented by a simple growth function such as the modified exponential, after adjustment for the displacements by the major shocks. For the UK this analysis is consistent with a slowly decelerating growth rate.

General applicability

This analysis has been carried out for many countries around the world at different stages of development, and subjected to major fluctuations of various kinds. Similar structural changes occur in all economies when struck by such shocks. When developing models to describe specific port traffic in relation to the relevant economic activities, it is essential to take account of these changes of state of the economy, and also to take their future likelihood into account in any projection of future traffic. It is obvious that when analysing changes in trade in specific commodities, each of which may be expected to relate to particular sectors of the economy, their relationships with GDP development will be different in each inter-shock period, because the links between sectors and GDP change.

In such traffic models, it is also important to include the sectoral variations around their trend relationships with GDP. These variations are highly correlated, so that most of the variance can be described by one or two principal components. These components commonly have periodic and lagged behaviour. If such variations are not included, any estimates of relationships with GDP are likely to be biased.