Search

In this paper, a set of neural network (NN) models is developed to compute short-term forecasts of regional employment patterns in Germany. Neural networks are modern statistical tools based on learning algorithms that are able to process large amounts of data. Neural networks are enjoying increasing interest in several fields because of their effectiveness in handling complex data sets when the functional relationship between dependent and independent variables is not specified explicitly. The present paper compares two NN methodologies. First, it uses NNs to forecast regional employment in both the former West and East Germany. Each model implemented computes single estimates of employment growth rates for each German district, with a 2-year forecasting range. Next, additional forecasts are computed, by combining the NN methodology with shift-share analysis (SSA). Since SSA aims to identify variations observed among the labour districts, its results are used as further explanatory variables in the NN models. The data set used in our experiments consists of a panel of 439 German (NUTS 3) districts. Because of differences in the size and time horizons of the data, the forecasts for West and East Germany are computed separately. The out-of-sample forecasting ability of the models is evaluated by means of several appropriate statistical indicators.

By using a shift-share regression approach the contribution of services to the development of employment in eastern Germany is analysed. The results obtained with highly differentiated data from the employment statistics show that services contributed more to a favourable path of development than other industries. This is due in part to general world-wide processes of structural change and the special situation in eastern Germany. Many subsidies were transferred to the East, which stabilises the special segment of the economy related to local demand. This segment is made up mostly of services in eastern Germany. Processes of industrial change can be explained using structural change approaches. It is shown that processes of de-concentration play a role in explaining regional disparities, since inverse localisation and positive urbanisation effects are visible. For the empirical analyses an augmented approach is applied which uses a generalisation of an econometric analogue from the common shift-share meth...

The paper provides an outline of a method useful for forecasting problems. The approach is based on a combination of top-down and bottom-up techniques. It is applied to project employment in all 327 (western) German districts for a time span of two years. The most important step in the preparation of the forecast uses the ENTROP method, which is an entropy optimizing procedure, a generalization of common RAS techniques, newly developed for the estimation of matrices from heterogeneous information. In a defined sense the estimated matrix is the most probable one. The method chosen is very flexible and uses any available information extensively. Therefore, the estimates are reliable as is shown in an ex-post forecast. There is a double purpose for the forecast of employment. First, it helps to gain insights in the causal processes generating regional developments and spatial disparities on labour markets. Second, it is useful for regional labour market policies, e.g. the budgetary planning of the Federal Em...