Scottish Government

Economic Impact of the 2001 Foot and Mouth Disease Outbreak in Scotland

3. THE DIRECT SHOCKS ASSOCIATED WITH FOOT AND MOUTH DISEASE

In this section we identify the direct (exogenous) shocks to the Scottish economy that occurred as a result of the Foot and Mouth Disease (FMD) outbreak. These shocks are then entered into the model and the resulting impacts on the Scottish economy as a whole are reported in Section 4. The shocks occur in the Agriculture and Meat Processing sectors and to tourism demand. A qualitative summary of the nature of the exogenous changes that we are using to model the impact of FMD in Scotland is given in Table 3.1 below.

Table 3.1 identifies a number of general points. First, the impacts on "Agriculture" are varied: they include demand- and supply-side disturbances. On the other hand, the direct impacts generated via tourism are all of a demand-side nature. Second, the impacts of the agricultural direct effects fall almost exclusively on agricultural sectors, with "Meat Processing" also hit by the export ban. However, the direct impacts from changes in tourist behaviour are much more evenly spread across sectors.

Table 3.1: Summary of exogenous shocks to the model

Most of the direct effects of the FMD outbreak are thought to last for only one period of the model. This corresponds to the year 2001. These effects are therefore modelled as a temporary, one-period shock. The only action that has a different time dimension is the direct impact of the loss of the breeding livestock. First of all, we here assume that the associated production and demand effects begin in year two, that is, 2002. Second, these direct effects continue to year four, given the time-scale of restocking.

3.1 The Shocks to the Agriculture and Meat Processing sectors

For agriculture there are four broad separate exogenous shocks associated with the outbreak of Foot and Mouth Disease (FMD). These are:

• The export ban

• The production implications of the cull and compensation payments made for trading animals

• The movement restrictions

• The production implications of the cull and compensation payments made for breeding animals

We will deal with these in turn.

3.1.1 Export ban

In identifying the impact of the export ban imposed as a result of FMD we assume that the impact is realised through its effect on the exports of Meat Processing. In effect, the export of live animals is small and was already affected by BSE. However, the importance of agricultural products as intermediate inputs into "Meat Processing", a relationship which is captured in the model, means that there are powerful indirect effects on the demand for agricultural commodities. The strength of these indirect links is indicated in Tables 3.2 and 3.3.

Table 3.2 shows the proportion of output of each farm type sector that is sold as an intermediate input to the "Meat Processing" sector, so for example, 20.71% of the "Specialist Sheep" sector output is purchased by the "Meat Processing" sector. "Meat Processing" buys between 10% and 25% of the output of the seven disaggregated farming sectors. Table 3.3 gives the cost breakdown of the "Meat Processing" sector. This table reveals that the "Meat Processing" sector buys 52.11% of its total inputs from the seven farm-type agricultural sectors. We therefore expect strong backward linkages from "Meat Processing" to these farming sectors.

Table 3.2: Linkages between "Meat Processing" and other Agriculture sectors: percentage of farm-type output going to "Meat Processing" and exports

Source: Scottish Executive (2002b), Scottish Council for Development of Industry Database (2003)

Table 3.3: The share of total input costs of the "Meat Processing" sector taken up by each farm-type sector

Source: Scottish Executive (2002b), Scottish Council for Development of Industry Database (2003), Scottish Executive (1999)

To establish an estimate for the shock to ROW "Meat Processing" exports generated by the FMD export ban, we used the Scottish Council for Development of Industry Database (2003) figures for exports of Livestock and Products from the Primary Sector Exports (Table 9) from 2000 to 2001. These data indicate a 55% reduction in exports of these items, which we assume is all directed through the "Meat Processing" sector. The regional impact of this direct shock is assumed to be in proportion to the distribution of employment within this sector across the urban, uninfected rural and infected rural regions.

There is clearly uncertainty about the absolute size of the shock that accompanies the export ban. Further, the export ban is the agricultural shock that generates the largest impact on the Scottish economy. We have therefore subjected this simulation to sensitivity analysis. In section 4, we report results from high and low estimates corresponding to ROW export adjustments to the "Meat Processing" sector of 65% and 45% respectively.

3.1.2 Production implications of the cull and compensation payments made for trading stock animals

To estimate the effect of the cull and compensation payments made for trading stock animals, we begin with the total compensation payments of 171million paid to farmers. These payments were made either because the farm had confirmed cases of FMD, or because it fell within the contiguous cull boundaries (with animals destroyed due to the Livestock Welfare Disposal Scheme). The total payments were then split out across farm sector types, based on the different numbers of each type of animal that were destroyed in each farm type. The compensation payments per animal, obtained from data compiled for McDonald and Roberts (2002), are shown below.

Table 3.4: Compensation payments per animal

Source: Scottish Executive (2002a) (Tables C34 and C35), Scottish Executive (2001) (Table 5), cull statistics data from McDonald and Roberts (2002).

Based on data from the 1999 agricultural census on animal numbers across the sectorally disaggregated farm types and regions, these compensation payments were allocated across the sub-regions of Scotland - urban, uninfected rural and infected rural - for each of the seven disaggregated farm types. These total compensation payments to each farm type were then divided into compensation for breeding animals and compensation for trading animals. This division was based on the aggregate value of each type of animal in each farm-type sector. For instance, if 50% of the value of animals on a typical "Specialist Sheep" farm are breeding stock, then 50% of the compensation payments to that farm type are given to breeding stock, the remainder going to trading stock. In this section we discuss the direct impact of the cull and compensation payments for trading animals. In the next section we discuss the implications of the cull and compensation for breeding animals.

The figures for compensation for trading animals broken down by farm-type and region are given in Table 3.5 below. Total payments made under these two schemes for trading animals sum to just over 72 million in 1999 prices ⁵. The largest payment by farm-type goes to "Dairy" farms, which receive over a third of all payments. The "Specialist Beef" and "Specialist Cattle and Sheep" farm types also get a significant share of this aid: "Specialist Beef" over a quarter and "Specialist Cattle and Sheep" over 15% of the total payments. As would be expected, over 99% of all the compensation payments are made to farms in the infected rural region.

Table 3.5: Compensation payments for trading animals broken down by region and farm-type, 1999 prices ()

Source: Regional Breakdown from Scottish Executive (2002a), cull information from McDonald and Roberts (2002).

We treat the impact of the cull of trading animals by imposing a corresponding reduction in the supply of animals to the market. This reduction in supply has a positive impact on prices. It implies that without compensation, farms suffering the cull would experience a reduction in income, but other farms would gain. The impact of compensation is modelled in a very straightforward way, as an exogenous demand injection. Essentially we treat the government as acting as the consumer for culled trading animals instead of private Meat Processing firms and final demand categories. This therefore accounts for the production cost of the traded animals which did not reach market. Both the supply restriction and demand shocks apply for only one period: that is, only operates in the year 2001. The impact of the cull and compensation for breeding animals is treated in a rather different way.

3.1.3: The production implications of the cull and compensation paid for breeding animals

In the previous section we outlined the way in which we model the impact of the cull and compensation payments that affected trading animal stock. In this section we explain how we model the effect of the cull of breeding animals. Essentially we treat such animals as elements of the capital stock of the relevant sectors. The effect of the cull on breeding animals is to reduce the productive capacity, and therefore subsequently output, in the infected farms. The first issue is therefore to identify the impact of the cull on capital stock in the different agricultural farm-type sectors. A second issue concerns the speed and extent of restocking. In so far as restocking replaces culled animals, the capital stock and productive capacity of affected sectors will be restored. However, there is also a demand impact of restocking, in that the demand for the output of some Scottish agricultural sectors will be increased as a result of the demand for replacement breeding animals.

The figure for the rate of capital stock decline was obtained by taking the value of destroyed breeding stock (defined as capital assets on the farm's balance sheet) as a proportion of the total assets held by farms. Total assets by farm type came from the Farm Accounts Survey and included all assets held by farms. The breeding stock were animals that, due to the length of the production cycle, would have been kept on the farm for more than one year, but were destroyed in the response to FMD.

Table 3.6 gives the total value of compensation payments made for breeding livestock, broken down by type of farm and region. For each farm-type sector we know the distribution of payments between trading and breeding animals. Similarly for each farm sector we know the distribution of total compensation payments between regions. The regionally disaggregated figures presented in Table 3.6 are calculated on the basis that within any one farm-type, the distribution of compensation payments between trading and breeding animals does not vary between regions. The final row of Table 3.6 expresses the total reduction in the value of breeding animals as a share of the total capital stock in that farm-type sector. This is the reduction in capital stock in each farm-type sector that is used to shock the model to replicate the capacity-reducing effects of the cull of breeding animals.

Table 3.6: The absolute values and percentage change in total capital stock as a result of the cull of breeding animals, 1999 prices ()

Source: Calculated from Scottish Executive (2001) (Tables 1 and 5), Scottish Executive (1999) and statistics from McDonald and Roberts (2002).

In the absence of more concrete figures we have taken expert advice about the extent of restocking ⁶. We assume that in the period subsequent to the onset of FMD, that is year 2002, the decline in capital stock will be as in Table 3.6. That is to say, in 2001 there is a cull in the number of breeding animals which is carried forward to 2002 when restocking begins. However, this restocking is not assumed to affect production in the rural infected region until year 3 (2003). In this third year, we assume there is a 75% recovery in breeding stock levels. In the fourth and fifth periods, we assume further 10% and 7.5% recoveries, leaving a residual of 7.5%. It is thought that this residual will never be replaced. These estimates determine the extent and timing of the restocking supply- and demand-side impacts.

The restocking is associated with a demand injection in the form of an increase in exogenous investment demand. We assume that all restocking is domestic and comes from uninfected areas, though this assumption is probably not fully accurate in that a small proportion is thought to have been sourced from England, Wales and Holland. The time path of reinvestment in breeding animals determines the timing of these demand shocks.

These exogenous changes are modelled as follows. For the supply side effects, in year 2 we introduce a capital augmenting reduction in productive efficiency in each farm-type sector ⁷. The size of the reduction is equal to the figure given in the total column in Table 3.6. These adjustments act to reduce the capital efficiency in each sector and therefore are equivalent to corresponding reductions in the capital stock. In subsequent time periods, the efficiency loss is reduced as just outlined.

For the demand impacts, we assume that all farms are restocked from the uninfected regions of Scotland. This generates a demand for the output of the livestock sectors in Scotland whose timing is the same as the changes in the capital stock.

3.1.4 Estimating the costs of the movement ban - by farm type

Four types of cost are identified as being associated with the movement ban. These are:

• Retention Costs, which include the additional feed costs during the longer than normal period of retention.

• The Price Penalty from retaining fattened animals beyond their optimum sale date.

• The Capacity Costs, which reflect the additional feed costs resulting from shifting stock densities / patterns away from the optimum.

• The Production Inefficiencies, which reflect the lower fattening capability which arises during the longer period of retention of store animals.

Costings are estimated as gross before making any allowances for adjustments to business practices (to respond to the 20-day standstill) or before making any allowance for any offsetting gains. For example, there are cases where additional costs to one farm (i.e. in producing store animals) may be offset by gains to other (fatteners) farms.

Estimates of the size of these impacts are given by Animal Movement Standstill: Economic Assessment of the Impact Upon the Livestock Industry in England and Wales (DEFRA, 2001, Table 4). For Scotland, they are calculated by proportioning (by farm types) the English and Welsh costs to Scottish Farms, using the total number of holdings for Scotland and England (obtained from the regional trends database on http://www.statistics.gov.uk/CCInsclasp?ID =5019); and data on agricultural holdings by farm type, size and county (from http://www.defra.gov.uk/esp/work_htm/publications/cf/fiuk/current/chapter4/c4t5.xls).

Table 3.7: Direct costs of the livestock movement restrictions, broken down by farm-type and regions (, millions)

Source: DEFRA (2001) (Table 4), Foot and Mouth Website. Regional trends database on http://www.statistics.gov.uk/CCInsclasp?ID=5019 and http://www.defra.gov.uk/esp/work_htm/publications/cf/fiuk/current/chapter4/c4t5.xls

The farm-type and regional disaggregation of the costs associated with the movement ban are given in Table 3.7. These show that the costs are concentrated almost wholly in three farm-type sectors. These are "Specialist Beef", "Specialist Cattle and Sheep" and "Mixed Farming". Note also that almost 60% of the costs associated with the movement ban are borne by farms in the Uninfected Rural region. This reflects the relative geographic size of this region.

The direct impact of the movement ban is simulated as a reduction in the efficiency of production in the individual farm-types. The shock enters as an improvement in technical progress that applies equally between all elements of value added, and takes a value equal to the estimated cost to that farm-type sector divided by the total value added of that sector. ⁸ Therefore, for example, if the cost of the movement ban to Specialist Beef is estimated to be 1 million and the total value added in Specialist Beef is 20 million, the efficiency shock that would be administered would take the value 1/20, or 5%. The shock is run for one period only. This implies the ban operates only for 2001.

3.2 Tourism related shocks

The aggregate impact of the adjustment to tourist demand in Scotland generated by FMD is separated into five distinct elements. These are:

• Rest of the world (ROW) Tourism

• Rest of the UK (RUK) Tourism

• Scottish Tourism in Scotland

• Daytrips

• Displacement to Scottish Household Expenditure

The structure of tourism demand within Scotland is shown in Table 3.8 below. The regional and industrial disaggregation methodology is given in greater detail in Annex C, and copies of the disaggregated final demand columns are given in Annex D.

Table 3.8: Structure of Scottish tourism, 1999

Note: this table simply reproduces Table 2.1

Annual Tourism from the Rest of the World (ROW) was estimated to be 859.55 million, which was substantially less than the 2,538.75 million attributable to Rest of the UK (RUK) Tourism. The two types of tourism normally included within Household Consumption in the Input-Output Tables are Scottish Tourism in Scotland and Daytrips. These were identified as 1,368.63 million and 4,726 million respectively.

ROW and RUK tourism covers spending by non-Scottish tourists in Scotland. Scottish Tourism in Scotland is the expenditures of Scottish nationals on tourism in Scotland. Daytrip expenditures are also made by Scots, but whereas in domestic tourism an overnight stay must be involved, with daytrips an overnight stay does not occur.

The final element here is the displacement of household expenditure. There are two similar, but distinct, possible mechanisms here. The first is the substitution of some types of Scottish holiday or daytrip destinations for others. The second is the displacement of previous tourist and daytrip expenditure by Scots in Scotland by other forms of consumer expenditure. In this report we simply identify displacement expenditure: we have not attempted to calculate any location substitution.

In calculating the impacts of each of these tourist-related changes in demand, two aspects are important. These are:

• The absolute size of the demand shock.

• The distribution of the expenditure across different commodities.

In the case of tourism demand there is general uncertainty over the direct affect of FMD on tourist expenditures. We therefore provide a central estimate and a high-low band of alternative estimates of these direct impacts.

3.2.1 Absolute size of the shock

After looking closely at the trends identified in the available tourism data for Scotland and the judgements made in UK studies, a separate estimate is given for the impact of FMD on each tourism type, with upper and lower ranges. These figures are shown below in Table 3.9.

Table 3.9: Estimated exogenous shocks to tourism expenditures caused by foot and mouth disease, 2001 (, millions, 1999 prices)

Source: ROW figure from International Passenger Survey (2001), RUK and Scottish Tourism in Scotland figure from VisitScotland (2001), Daytrip data are based on estimates in MORI Leisure Survey Data (2002), Displaced Consumption data from VisitScotland (2001).

To make the ROW Tourism central estimate, we first identify the absolute fall in ROW tourist expenditure in Scotland between 2000 and 2001 from the International Passenger Survey (IPS) (2001). We then assume that 50% of the real (1999 prices) fall in ROW tourism expenditure in Scotland between 2000 and 2001 was due to FMD. For sensitivity analysis, we give a low figure where 40% of the fall is due to FMD and a high estimate where 60% of the fall is due to FMD.

For RUK and Scottish Tourism in Scotland we used the VisitScotland data on expenditure by these different types of tourists. Again the real changes from 2000 to 2001 (in 1999 prices) were calculated. However the central estimate assumed that 60% of this fall was due to FMD, with a low and high range of 50% and 70%. The larger percentage of the actual decline ascribed to FMD reflects the fact that the terrorist attacks of September 11th reduced American foreign tourism, therefore some of the reduction in ROW tourism should be attributed to this cause. However, this should not affect UK or Scottish Tourism in Scotland.

There is a lack of annual data on the size of expenditure for Daytrips. We therefore had to estimate the shock using the best of the available information. We used two methods. First, evidence from MORI Leisure Survey Data (2002) suggested that daytrips expenditure fell by 7% in Scotland due to FMD. Second, we assumed that the decline in daytrip expenditure due to FMD, as compared to the initial daytrip total expenditure, was equal to the corresponding impact on Scottish tourist expenditure in Scotland. Scottish tourist expenditure in Scotland was estimated to have experienced a 10.6% decline as a result of FMD. These two estimates were used to provide high and low bounds, with a central estimate of the decline in daytrips expenditure being taken as 9%.

Displaced Consumption was calculated using data from VisitScotland. The basic rationale was that consumption not spent by households on domestic or day tourism would be spent elsewhere in the economy in the pattern of typical, non-tourism, domestic household expenditure. While, in principle, this should be a positive shock to household consumption equal to the reduced value of Scottish tourist expenditure in Scotland and Scottish daytrips expenditure, indications are that there was some substitution of increased tourism expenditure overseas.

Data are given in the IPS (2001) for visitors expenditure abroad by Scottish residents from 2000 to 2002, quarter by quarter. For both the second and third quarter of 2001 we calculated a counterfactual total expenditure based on the average expenditure in the corresponding quarter in 2000 and 2002. The actual expenditure was compared to this counterfactual. In the second quarter of 2001, the actual was greater than the estimated counterfactual by 42 million. For the third quarter, the difference is not so great, with an expenditure of 7.5 million greater than the counterfactual. We therefore estimated a combined total of 49.5million that was not spent by households in Scotland, but was a leakage to overseas destinations.

The total positive shock to households is therefore the reduced daytrips expenditure plus the reduced Scottish tourist expenditure in Scotland, minus the expenditure lost to overseas (49.5million), giving a central estimate of 518.36million. The high and low estimates of this shock were taken from summing the high and low domestic and daytrips shocks and subtracting the lost 49.5million in each case.

3.2.2 Regional breakdown

Information on the geographical pattern of tourist expenditures within Scotland came from VisitScotland (2001, 2002). These give expenditure for both ROW and RUK tourists broken down by the 11 area tourist boards for both 2000 and 2001. For modelling purposes we require the proportionate changes in demand for each of the 25 industrial sectors for the three sub-regions of Scotland. The geographical regions are shown diagrammatically in Annex A and specified in Annex B.

The changes in expenditure by different types of tourist, in each of the three areas, are obtained in two steps. First, the VisitScotland figures for the change in expenditure between 2000 and 2001 were calculated in real terms (1999 prices) using the GDP deflator from the UK Treasury. The share of the reduction between the three geographical areas is then applied to the estimated IPS (2001) figure for the total Scottish reduction in ROW tourist expenditure. The result is shown in Table 3.10.

Table 3.10: Tourism expenditure changes due to FMD by type and sub-region of Scotland (, millions)

Source: International Passenger Survey, 2001 and VisitScotland (2000/2001)

As is seen above, the impact of the decline in ROW tourism was especially felt in the urban region, although a significant portion was also experienced in the uninfected rural area. This area includes the Highlands of Scotland, an especially popular destination for overseas visitors to Scotland, third in expenditure terms for 2000 behind only Edinburgh and Glasgow. Indeed, overseas expenditure in 2000 in the Aberdeen and Grampian and Highland Area Tourist Boards (ATBs) was greater than Glasgow receipts from overseas tourism in the same year.

To identify the particular demand impacts generated by the change in ROW tourist expenditure, the figures for each region gave the aggregate change. This was allocated across commodities using the disaggregation identified in Annex D. Note expenditure patterns for each type of tourist vary across the sub-regions. These expenditure changes were then expressed as a percentage of the total figure for exports to the Rest of the World from each of the sectors, giving us percentage shocks to apply to the model to simulate the effects of the decline in overseas visitors to Scotland.

The same general procedure was then used for the other types of tourist expenditure (RUK Tourism, Scottish Tourists in Scotland and Daytrips). The regional breakdowns from VisitScotland, and the change in receipts in 1999 prices, were scaled to the central estimates we have calculated for tourism.

A major difference was that the VisitScotland figures do not separately identify RUK and Scottish Tourism in Scotland figures: rather these data are aggregated under the heading "UK tourism". This required the assumption that RUK and Scottish Tourism in Scotland have the same geographical dispersion. Given the lack of any other reliable data, this seems a reasonable assumption. We have made a similar assumption for Daytrips. That is to say, the regional distribution of changes in UK tourism in Scotland is taken to be the same as the regional distribution of changes in daytrips. This seems a little more questionable, but appears unavoidable, given the lack of alternative data. These results are given in Table 3.10.

The fall in tourist expenditure from RUK, Scottish Tourism in Scotland and Daytrips is much more heavily concentrated in the uninfected rural region: over 70% of the reduction is identified as being in this region. However, the geographic distribution of the increased household expenditure, is assumed to match the regional distribution of GDP, so that over 60% is in the urban region.

Again the estimates achieved in this way for RUK, Scottish Tourism in Scotland and Daytrip expenditure by region were used as control totals and proportioned out across the 25 sectors using the corresponding expenditure patterns which are given in Annex D.