by R.W.M.Johnson Ministry of Agriculture Wellington
This paper describes methodologies and results achieved with local and regional models for agricultural impact analysis both within the agricultural sector and on other sectors in New Zealand. While the methodologies were developed for drought impacts, the applications to other change situations are straightforward and within the capability of the software. Attention is directed to the incorporation of existing modelling approaches and the need for some consistency in definition of variables and aggregation procedures.
The work described here arose out of the need to assess the economy-wide impacts of a severe drought which affected the eastern parts of New Zealand in the winter and spring of 1988.
In order to trigger Government assistance for the affected farmers (5500 in the South Island and 5500 in the North Island) officials were required to prepare an assessment of the economic damage to the economy of a continuation of the drought for a certain period and the mitigation of that damage if assistance proposals were implemented.
In New Zealand adverse climatic event assistance is available in the event of drought, flood, wind storm, snow storm and cyclone damage. Floods and droughts have to be demonstrated to have a frequency of occurrence exceeding 1: 50 and 1: 20 respectively while other climatic disasters may be assessed on their merits. When this criteria is met, the Minister declares that an adverse event has occurred and then planning for restoration may start. Measures at this stage would include assistance for immediate physical restoration aimed at securing property, standard welfare entitlements if needed, labour assistance schemes and income tax relief through an income equalisation scheme (as in Australia) .
In the case of major cyclones and droughts these measures are inadequate to deal with the situation and specially designed schemes have to be drawn up and approved through the parliamentary process.
It is in the latter stage of the assessment that local and regional models are useful and can fulfil a worthwhile function in providing accurate and representative data on the agicultural and non-agricultural impacts of a particular event.
The 1988 drought provides an example of the process involved in such an analysis. The threat of a drought was recognised during the winter months of 1988 and the failure of spring rain was confirmed by the middle of October of that year. Ministry analysts immediately started on an assessment of the economic effects of two scenarios; rain by the end of December and rain by the end of May of the following year.
Data at the local level was derived from the Ministry's continuous survey of the financial state of the main farming systems in the area concerned. This data base is updated every six months and in the case of the South Island drought area has six different farm type models represented. Each of these models is a representative farm budget base adjusted by intentions data collected from a panel of farmers.
At the reqional level the impact of the drought on agriculture was assessed from the weighted aggregate results of the individual farm type models. The impact on the reqional economy was assessed by applying existing multiplier coefficients from external sources and formulating these in terms of output, national income and employment impacts. In this particular case the national agricultural and non-agricultural impacts were not assessed. At a later stage the same procedure was followed for the regions affected by drought in the North Island where the initial impact was slightly delayed.
As a result of the emergency procedures described above, the internal process for assessment was re-specified and improved in subsequent months to make it more representative and to add in the link to the national economy. A national workshop was held to coordinate various local and regional modelling research endeavours and further research objectives were identified for future investment.
As in happens a climatic event of the size of Cyclone Bola or the 1988 drought has not occurred in the period up to the present hence the policy advice functions have not been mobilised in recent years. However, further research work has been conducted on regional and national input-output analysis and on aggregation principles and these procedures are readily available in suitable software at the present time. This is discussed in more detail in the rest of this article.
The particular local model used for the initial drought impact work was a farm financial or farm business model. At the workshop held in November 1989 other local models were reported on and assessed (Johnson, 1989). In particular it became clear that there were a number of physical models in use in the scientific community which could be incorporated in a system of models which could have appropriate aggregation properties and which could provide the necessary policy advice outputs when needed (See contributions by Jarvis, Bywater, McCall, Jamieson and Thiele in Johnson 1989) .
It was therefore decided to call for an integrated systems approach to the representation and aggregation problems that incorporated all available models and data. Chart 1 shows the flow chart of the resulting agricultural production-income model that was developed for the workshop (Johnson 1989 p 11). It was envisaged that a land use data base would be required to store the census information on farms and to provide the correct aggregation or raising factors (sampling fractions). Farm physical models could incorporate highly detailed technical information and relate for example dry matter production to animal production which could in turn be assessed in whole farm economic entities.
Farm business models could be drawn from the Ministry's data base on farm budgets (MAE Monitoring Report 1989) or could be drawn from other data bases with economic information on farms (New Zealand Meat ands Wool Boards' Economic Service 1989; New Zealand Dairy Board 1989) .
National farm income estimates could be derived from suitably aggregated samples of farm business models, but to obtain both reqional and national impacts of disaster events it was necessary to have a linkage to regional and national input-output tables. The regional tables deyeloped are described by Butcher (1985) and by Neild in Johnson (1989); and adjusted national tables and coefficients are described by Narayan and SriRamaratnam (1992).
The input-output tables published in New Zealand are only prepared at the national level hence to obtain the necessary data base to carry out regional analysis a special project was set up to dis-aggregate the published tables into suitable regions (Neild 1989). Once this base is established, however, software for manipulating the table is readily available and changes in projections can be incorporated or even new rows and columns can be introduced to assess the non-agricultural impacts (West 1988, Neild and Forbes 1990).
The risk component of the research programme was assessed at two levels. First, further research was clearly required into the trigger mechanism for droughts and floods. There are considerable measurement problems with climatic data (Baird 1989), and clearer criteria are really required for major assistance programmes that might result. Internal research is continuing in this area.
Second, it was felt that the behaviour of farmers under risk should be more clearly understood. A series of contractual projects were commissioned for this purpose and the contributants were brought together for another workshop in June 1990 (Johnson 1990).
The results of this workshop showed that farm level risk behaviour was not well understood and that wider education was needed at the farmer and the professional level. The Ministry subsequently launched an extension programme for farmers nationwide with suitable reading and display materials (MAF 1990) . The proceedings of the workshop were disseminated widely among professional groups and special articles written where appropriate (Johnson 1992).
Baird D (1989) 'The Analys is of Return Periods for Extreme Climatic Events' in Johnson R.W.M Integrated Systems Analysis and Climate Impacts, Ministry of Agriculture, Wellington.
Butcher G V (1985) 'Regional Income, Output and Employment Multipliers: Their Uses and Estimates of Them' in R Forbes Cost Benefit Handbook, Economics Dision, Ministry of Agriculture, Wellington.
Johnson R W M (1989) Integrated Systems Analysis and Climate Impacts, Proceedings of a Workshop on November 30, Ministry of Agriculture, Wellington.
Johnson R W M (1990) Management of Risk in New Zealand Farming, Proceedings of a Workshop on May 9, Mini stry of Agriculture, Wellington.
Johnson R W M (1992) View' Review (forthcoming) .
'Risk and the Farm Firm: A Corporate Finance of Marketing and Agricultural Economics,
Ministry of Agriculture (1989) National Farm Monitoring Report, Wellington.
Ministry of Agriculture (1990) Drought Proofing Your Farm, MAFTech, Christchurch.
Narayan P and SriRamaratnam S (1992) 'Pastoral Sector Impacts on the New Zealand Economy' MAF Policy Technical Paper 9211, Wellington.
Neild J (1989) 'The Role of Input-Output Analysis' in R W M Johnson Integrated Systems Analysis and Climate Change, Wellington.
Neild J and Forbes R (1990) Employment Impacts of the New Zealand Sub-Tidal Clam Fishery, A Report to Regional Fisheries Manager, Nelson.
New Zealand Dairy Board (1989) Survey of Dairy Farm Incomes, Internal paper, Wellington.
New Zealand Meat and Wool Boards' Economic Service (1989) The New Zealand Sheep and Beef Farm Survey, Wellington.
West G R (1988) GRIMP Input-Output Analysis Computer Programme, Users Reference Manual, Version 6.01, University of Queensland.