Joint Economic Forecast 2022 - 2026

Tasks

The research project "Gemeinschaftsdiagnose" (GD) analyzes and forecasts the economic situation in Germany. The diagnoses are prepared twice a year, in spring and autumn. The GD's forecasts provide orientation for the Federal Government's projections. The GD is a joint research project of several economic research institutes. Through their cooperation, the analysis and the forecast in the dialogue are substantiated in the best possible way.

Currently, the GD is being jointly prepared by the following contractors:

• ifo Institute - Leibniz Institute for Economic Research at the University of Munich e.V. in cooperation with:  Austrian Institute of Economic Research (WIFO)
• RWI - Leibniz Institute for Economic Research e.V., Essen, in cooperation with: Institute for Advanced Studies Vienna
• Leibniz Institute for Economic Research Halle (IWH) and
• Kiel Institute for the World Economy (IfW Kiel).

The reports examine and present the most important national and international economic conditions and developments. On this basis, the analysis of the economic situation in Germany, the forecast for the short term and the forecast for the medium term, including an estimate of potential output, are carried out. The estimation of potential output is carried out in accordance with the procedure applied by Germany under the European budgetary surveillance and agreed with the European Commission. The report may contain an analysis of economic policy and economic policy recommendations. One focus topic with reference to current economic developments is dealt with in depth. The forecasts are prepared within the framework of the national accounts on the basis of the quarterly results of the Federal Statistical Office. All important aggregates of the national accounts are forecast. 

Methods

The ifo Institute relies on forecasting that is anchored in survey-based business cycle research and makes use of current empirical methods and theoretical models. In the area of forecasting methodology, special attention is paid to the problems of large data volumes, different frequencies and various publication delays, both in active research and in forecasting practice. 

One research focus is on methods for information condensation and selection. The focus is on the question of whether microdata from business surveys are suitable for forecasting macroeconomic time series. Among other things, different aggregations of the microdata (by participants, industries or regions) are examined for their influence on the forecasting quality. One example is boosting techniques, which can be used to identify core information (e.g. companies with high forecasting power). In addition, methods such as Bayesian model averaging or models with mixed frequencies, which are already in constant use, are examined for their suitability and further developed. Another research focus is the analysis of new and alternative data sources for forecasting macroeconomic indicators in the context of the ifo-wide Big Data strategy. Here, the focus is primarily on high-frequency indicators such as electricity consumption, truck toll mileage, mobility indicators, reservation data in restaurants, transaction data from financial service providers or text analyses of newspaper articles.

In forecasting practice, those methods are used and adapted to the German data situation whose suitability has previously been verified according to scientific criteria. For the forecast of gross domestic product in the very short term, i.e. the estimate for the current and the following quarter, the ifo Institute emphasizes the importance of survey-based indicators. The extensive results of the ifo Business Survey play a central role in forecasting the German economy. These include on the one hand the time series available at monthly frequency over a period of several decades, such as the ifo Business Climate Index, the ifo Employment Barometer or the ifo Price Expectations, and, on the other hand, the results of special questions introduced depending on the situation, on the basis of which, for example, monthly estimates of the current extent of short-time work in Germany have been compiled and published since May 2020. In addition, other leading indicators from official statistics and recently published high-frequency indicators are also taken into account. In order to condense the wealth of available information and indicators into a forecast, the ifo Institute follows two approaches. First, many forecasts are produced using many indicators, which are then condensed by averaging ("pooling of forecasts"). This approach has been used successfully at the ifo Institute since 2009. On the other hand, the information is already condensed before the forecast is prepared and a single forecast is produced ("Pooling of Information"). Since 2020, for example, the ifo Institute has been using ifoCAST to produce a continuous and automated forecast of German GDP for the current and subsequent quarters on the basis of a dynamic factor model, which is updated every two weeks and made available to the public. 

However, a wide range of alternative models is also used for forecasting horizons beyond this in order to be able to identify model imperfections in good time with the help of cross-checks. Different types of models from time series analysis, econometrics and applied macroeconomics are used. On the one hand, vector autoregressive models with mixed frequencies are used for both the use-side and the emergence-side forecasts. On the other hand, the ifo Institute uses a DSGE (Dynamic Stochastic General Equilibrium) model developed in-house and estimated for the German economy. A detailed trend analysis of the gross domestic product and its aggregates on the consumption and output side plays a key role in forecasting the further course of the business cycle. This becomes all the more important the more underlying trend growth rates are likely to change in the forecast period, e.g. as a result of demographic change. In addition to the EU reference method for determining potential output, alternative time series models are also used for this purpose, which, among other things, eliminate the cyclical components of the time series with the help of the ifo capacity utilization. 

Great importance is attached to a scientifically sound theoretical, institutional and empirical analysis. This makes it possible to identify cyclical and structural development trends at an early stage and to estimate the effects of policy measures or changes in the institutional framework. The research focuses on the impact of financial market frictions and uncertainty as well as the effects of monetary and fiscal policy measures on real and financial variables. Many issues arise from the experience and observations in connection with the world financial and euro crisis as well as the Corona crisis. Methodologically, model-theoretical approaches, such as DSGE models, empirical time series methods, such as vector autoregressive models, and microeconometric estimation methods are used. The empirical work often makes use of survey data from the ifo Institute. The micro data of the ifo Business Survey are used, for example, to identify financial constraints and firm-level uncertainty, to analyze the resilience of German SMEs to the crisis, and to study the response of firms to the Corona crisis in terms of their investment, pricing and hiring behavior. 

The variety of forecasting models and analysis methods is combined with the available expert knowledge in an iterative analytical procedure and condensed into a forecast that is consistent with the framework of national accounting. In addition, the forecasts from the expenditure side, for which a smaller number of indicators are traditionally available, are checked for plausibility against the forecasts from the production side. The overall uncertainty of economic forecasts is openly communicated. For example, the ifo Institute publishes an interval for its own forecast of German GDP. In addition, important forecast risks are described. Since an economic forecast is always based on certain assumptions, e.g. regarding exchange rate developments or economic policy, scenarios are also discussed. These may concern the various policy options and alternative reactions of the financial markets. The assumptions of the baseline forecast and possible alternative scenarios are disclosed. Such scenario calculations were particularly important in the initial phase of the corona crisis, when uncertainty about the further course of the infectious event and about the reaction of private economic actors and economic policy was especially high.

Of particular importance for the ifo Institute's forecasting philosophy is also the communication of the forecasting methodology, the analytical background information and the relevance of the ifo survey results for business cycle analysis. Members of the forecasting team regularly present the most important results of their own studies to the public in articles in the ifo Schnelldienst and in technical boxes in the institute's own economic forecasts.

Data and Other Sources

German Federal Statistical Office, Deutsche Bundesbank, German Federal Employment Agency, European Central Bank, Eurostat, IWF, OECD and own calculations.