Economic Impact of the 2001 Foot and Mouth Disease Outbreak in Scotland
Annex B: The modelling framework
The modelling results given in this report are generated using a specially constructed variant of AMOS, a computable general equilibrium (CGE) modelling framework that is parameterised on 1999 data for Scotland. 12 A very brief description is presented in this section: a full listing of the AMOS model is provided in Harrigan et al (1991).
The current version of AMOS has 3 transactor groups: households; corporations; and government. 13 There are 25 commodities and activities and two exogenous external transactor groups, namely the rest of the UK (RUK) and the rest of the world (ROW). There are four major components of final demand: consumption; exports; government expenditure; and investment. Of these, consumption is a linear homogeneous function of real disposable income. Exports (and imports) are generally determined via an Armington link (Armington, 1969) and are therefore relative-price sensitive with trade substitution elasticities of 2.0 (Gibson, 1990). Real government expenditure in Scotland is taken to be exogenous. Investment is initially set equal to depreciation although, as explained later in this annex, in subsequent periods investment demand in each sector is generated through a capital stock adjustment process, where net investment is a proportion of the difference between actual and desired capital stock.
Production is determined through cost minimisation with multi-level production functions. These are generally of a CES form but with Leontief and Cobb-Douglas available as special cases. For simplicity, all domestic intermediate transactions are assumed to be of the Leontief form in this paper. Otherwise we assume CES technology (notably for the production of value-added from capital and labour services).
In CGE models, market processes are specifically modelled. A wide variety of forms of product or factor market are possible. In this report, commodity markets are taken to be competitive (though potentially subject to government intervention). By this we mean that product prices are determined by the interaction of supply and demand, although production and consumption can be taxed or subsidised, and existing taxes and subsidies are automatically incorporated in the cost structure of individual industries. We do not explicitly model financial flows, our assumption being that Scotland is a price-taker in competitive UK financial markets and, under the small open economy assumption, the Bank of England's Monetary Policy Committee's interest-rate-setting decisions are taken to be exogenous to Scotland.
In all of the simulations in this report we impose a single Scottish labour market characterised by perfect sectoral mobility. The Scottish wage is determined through a bargaining function in which the regional real consumption wage is directly related to workers' bargaining power, and therefore inversely to the regional unemployment rate (Minford et al, 1994). This hypothesis has received considerable support in the recent past from a number of authors. Here we take the bargaining function from the regional econometric work reported by Layard et al (1991):
| (B.1) |
where w s and u s are the natural logarithms of the Scottish real consumption wage and the unemployment rate respectively, t is the time subscript and a is a calibrated parameter. 14 Empirical support for this "wage curve" specification is now widespread, even in a regional context (Blanchflower and Oswald, 1994).
A key implicit assumption made concerning the operation of the labour market is that a change in the demand for labour will be translated into a change in the number of workers employed, rather than an adjustment in the actual or effective average number of hours worked. Given that there are hiring and firing costs, it might be that model therefore overestimates the employment change generated by what are perceived to be temporary economic shocks. That is to say, firms might chose to maintain employment levels in the face of temporary downturn in activity in order to maintain worker goodwill, firm-specific skills, etc.. The model also assumes that labour is mobile between sectors so that where labour is released from one sector, downward pressure is put on the wage rate and some of that labour is re-employed in other sectors.
We report the results for period by period simulations, where each period is interpreted as a year. In each period, all commodity markets are assumed to clear and the wage is determined by the bargaining function, with a fixed population. In any one period, the capital stock is fixed, not just in aggregate but to individual sectors. Between periods, we have population updating, via migration, and capital updating, through investment.
The population adjustment is driven by a relationship whereby Scottish net migration is positively related to the real wage differential and negatively to the unemployment rate differential with the rest of the UK (RUK). This variant of the Harris and Todaro (1970) model is commonly employed in studies of US migration (e.g. Greenwood et al, 1991; Treyz et al, 1993). It is parameterised here from the econometrically estimated model reported in Layard et al (1991):
| (B.2) |
where m is the net in-migration rate (as a proportion of the indigenous population); w r and u r are the natural logarithms of the RUK real consumption wage and unemployment rate and b is a calibrated parameter. 15
In some of the agriculture simulations, the particular disturbances are modelled as efficiency changes. In AMOS these operate in the production of value-added. Three options are available: Hicks-, Harrod- or Solow-neutral technical change. Hicks-neutral technical change affects the efficiency of both capital and labour equally. Harrod-neutral technical change only changes the efficiency of the labour input. Solow-neutral technical change only changes the efficiency of the capital input.
The parameter values used in the model are either based on econometric work or are our standard "best guess" estimates. The key differences in the parameters used in these simulations come from the database, and these are explained in more detail in the next chapter.
One important point concerning the model is that it is not a forecasting model, but rather a model for scenario experiments. The model assumes that the region is in equilibrium in the base period. It then identifies comparative static impacts of changes in exogenous variables or parameters. This means that with unchanged parameters and exogenous variables, the model replicates the base values continuously. The model is also an economic, rather than financial model. It is tracking the real impacts on economic activity.
When the model reacts to a change in demand, it is most appropriate to think about it as an Input-Output (I-O) model except with supply constraints, competitiveness effects and production and consumption flexibility. The model has the same multi-sectoral focus as I-O but attempts to more realistically replicate market mechanisms. Therefore unlike the I-O case where an increase in exogenous demand benefits activity in all sectors, here wage and price increases will reduce the competitiveness of the economy as a whole and adversely effect some sectors. Therefore, especially over relatively short time periods as discussed here, there would be crowding out in some sectors. For supply-side changes the inter-relationships are more complex and the analysis more involved.