![]() |
|
|
|
Agricultural land forecast prepared by Technical Forecasts Limited The agricultural land forecastThis is the sixth time we have published an agricultural land forecast in the Property Market Report and it has been prepared again by Technical Forecasts. There are two forecasts, one for equipped land and one for unequipped land and they cover all land types in the UK as described in the Property Market Report. The data set for the unequipped land has very short history as valuations on this basis were only prepared from 2002. For this reason the model has very little information on historic patterns of unequipped land values and the resulting forecasts are not as well founded as those where more information is available. In particular, any above trend change in a period comprised within a short data series will be magnified compared to a longer run of data. Once again the forecast model has predicted that the rate of increase in the value of unequipped land will outstrip equipped land. The forecasts for equipped land show a slow rate of increase in values during 2008 into early 2009, followed by a steady rate of increase to the end of the forecast period. The rate of increase is higher than the forecast of 6 months ago. The forecasts for unequipped land continue to show consistent rises in value, but the suspicion the short data run may be disproportionately influencing the results remains. However, the sharp increase in the price of bare land at the end of 2007 and into 2008 is in accordance with the market direction predicted by the earlier forecasts. As with all other forecasts, but even more so in this particular case, the figures should not be taken in isolation but considered together with all other market information that is taken into account in any decision making process. The agricultural market is complex, as it is driven by farming and non farming factors, with a significant current element of purchasers being driven by non-farming objectives, in addition to the now consistent influence of overseas buyers. The historic data will of course not include the effect on farmer confidence of the rise in cereal prices seen in the last 12 months, nor the forthcoming changes in the rate of Capital Gains Tax.
Background information The forecasts have been provided by Technical Forecasts Limited, an established company that is at the forefront in the development of forecasting techniques for the property sector. The techniques adopted for the original series of forecasts published in the Property Market Report have been undergoing continual development. These new systems are based on Bayesian principles and techniques for finding the best average forecasts, given current market forces. Data source The agricultural data is based on opinion of value of four types of farm, arable, dairy, mixed and hill at typical locations throughout the UK. A more detailed description of the typical farm is provided in the methodology of the report. Methodology The forecast uses Technical Forecasts' latest methods for forecasting property time series. Radial basis functions have been adopted (a type of neural network) together with information from economic, financial and other property series. The programs select the best economic, financial and property series that give the maximum amount of information to the series. Bayesian statistical methods are then used to give the most probable future course of the Land series, given the past data. The forecasts are made on the basis of relationships that exist in the past history of the data series along with associated economic and financial data. The methods used Construction Orders, Bank of England base rates, Government data series and also Halifax and Nationwide data series as well as other public and private sources of technical data; over 100 other sources of data were scanned and the relevant patterns extracted in order to inform the forecasting process. Data such as property income received was additionally found helpful in informing the agricultural forecasts. Advanced methods of error reduction in the past time series have also been developed . The training of the models is controlled in such a way as to make sure that only the relevant data patterns are used in the creation of the models, eliminating, or at least minimising, any errors that may exist in the data. In this manner only the most relevant past patterns i.e. those that determine the future, are taken into account. Application It is important to remember that future values do not necessarily depend on past performance and mathematical techniques cannot take into account sudden social and political changes or other shocks to the market. All forecasts should be treated with caution and should be used in conjunction with professional opinion but they do provide an analysis of relationships within existing data series and present a degree of objectivity over their subjective equivalents. Technical information about this forecast and on additional property forecasts is available from: Technical Forecasts Limited PO Box 828 Horsham RH12 9EZ tel 0845 2 300 456 Email: mandy.bradley@tfl.biz www.propertyforecasts.co.uk
Agricultural land forecast ( equipped )
Agricultural land forecast ( unequipped )
|