5. The macroeconomic forecast

The forecast is produced four times a year based on analytical work lasting about a month, during which data inputs from the domestic economy and abroad are processed. Results and projections obtained using a wide range of analytical and forecasting tools are discussed and modified on an ongoing basis by the Monetary Department’s economists based on their expert judgement.

The first step in the process is the preparation of the outlook for developments abroad and the impact of domestic fiscal policy (the macroeconomic impact of relevant items of the public budget system). These factors are largely independent of the Czech economy and therefore enter the subsequent stages of the forecast as assumptions on which the resulting forecast is conditional. The main focus of the foreign outlook is on the “effective euro area” – the six euro area countries with which the Czech economy has the strongest trade links (Germany, Slovakia, France, Italy, Spain and Austria). The fiscal outlook is captured in the forecast through general government consumption and the “fiscal impulse”. The latter quantifies the impact of fiscal measures on the various components of GDP.

A near-term forecast (NTF) is then produced for the domestic economy. The NTF is very detailed and more accurate for the next few quarters than the outlook from the core projection model. It is thus used to set the initial conditions for the medium-term forecast in the core projection model as well as possible. The more distant quarters of the NTF serve as a complementary view of possible future economic developments from the perspective of the experts that produce the NTF. At this stage, the NTF team also prepares an outlook for domestic administered prices and expected tax changes for the entire forecast horizon.

These first outputs (the external environment and fiscal policy outlooks and parts of the NTF), supplemented by feedback from the Bank Board, are then integrated as assumptions into the core projection model, which is used to construct the medium-term forecast. The latter stretches approximately two years into the future and thus covers the entire monetary policy horizon. The core projection model forms the unifying framework for the overall forecasting process, in which all the partial analytical and data inputs are collected.

The final step of the forecasting process is the integration of the medium-term forecast in areas not directly included in the core projection model. Based on the variables predicted by the core model, forecasts are calculated in the following satellite areas: the balance of payments, the complete fiscal forecast, the quantity of money in the economy, bank loans and other price categories outside the consumer price index. The analytical tools and models used thus produce mutually consistent forecasts complementing the core model’s forecast in greater detail and in a broader context.

Figure – Production of the CNB forecast

The CNB uses a “core projection model” as its main tool for producing macroeconomic forecasts. Since 2008, it has been a DSGE macroeconomic model. This model is (D)ynamic, which means it describes the paths of macroeconomic variables over time, with the past state of the economy influencing its future. The model is also (S)tochastic, which means it uses the concept of random shocks deflecting the economy from equilibrium. The interpretation of these shocks helps to explain the observed past evolution of the economy and also allows for modelling of expected events over the forecast horizon. After the shocks have passed, the model economy tends to return to (G)eneral (E)quilibrium, with economic growth and the exchange rate at their equilibrium levels and inflation in line with the central bank’s inflation target. The model equations describe the economic mechanisms that determine how the economy responds to shocks and how quickly it returns to equilibrium.

The CNB’s core projection model captures the characteristics of a small open economy. This is consistent with the Czech economy’s strong export orientation and its relatively small weight in the global and European economies. Foreign economic activity is therefore crucial for the domestic economy, whereas the foreign economy is independent of the Czech one. Projections for the external environment are therefore included in the model forecast as exogenous assumptions and are thus taken as given from the perspective of the domestic economy.

In the model, the domestic economy is simplistically divided into firms, households, general government and the central bank. Firms produce goods and services for the domestic and foreign markets. Their inputs are labour, capital, technology and imported intermediate goods. Firms aim to maximise profits. Households seek to maximise their current and expected future utility from consumption and leisure subject to their budget constraints. Their main income comes from wages and salaries. Transfers from general government are another source of income. Some households are also owners of capital and thus recipients of firms’ profits. The government collects taxes, redistributes transfers and consumes a portion of the goods produced. It can also issue bonds and thus create public debt.

The model central bank targets inflation. Its behaviour is described by a reaction function taking into account the deviation of future inflation from the inflation target (the Taylor rule). This means that the central bank sets short-term interest rates according to the expected path of inflation. The path of the exchange rate depends on the interest rate differential between the Czech economy and the euro area (the uncovered interest rate parity condition) and takes the export performance of the economy into account. The model captures the main channels of monetary policy transmission in a simplified form – from the setting of monetary policy rates through to their impact on inflation and other economic variables.

Expectations about future developments play a key role in the decision-making of households and firms in the model. These expectations are formed within the model based on all the information available. However, actual future developments can also be affected by shocks, which are very difficult to predict. The role of the central bank in the model is not only to anticipate expected economic developments, but also to respond to these unexpected shocks on an ongoing basis, thereby stabilising (anchoring) inflation expectations and steering inflation towards the 2% target.

Figure – Scheme of the links in the CNB model

The CNB constantly updates its model framework. The original near-term forecasting (NTF) methods were supplemented in 2002 by a core quarterly projection model (QPM). This was replaced in 2008 by a dynamic stochastic structural general equilibrium (DSGE) model called g3. This model combines real business cycle theory with nominal rigidities according to New Keynesian theory.

In 2019, on the basis of experience and current trends in macroeconomic modelling, the CNB introduced the g3+ model, which contains a more elaborate structure of foreign economies and enhanced modelling of the domestic household sector. These changes allowed for more credible consideration of the specifics of the Czech economy, improved forecasting properties and a more detailed description of economic developments.

In 2024, g3+ was further upgraded to better reflect the extraordinary economic developments seen in previous years and the forecasting experience gained from its regular use in forecasting exercises. Further adjustments will soon be made in response to the external review of the CNB’s monetary policy analytical and modelling framework that took place in summer 2024.

Figure – Development of projection models at the CNB

The economic forecast used as the basis for the CNB Bank Board’s monetary policy decision-making is produced by a team of experts. The baseline scenario of the forecast represents the most likely expected future path of the Czech economy as seen by the Monetary Department’s economists. However, the forecast, together with any scenarios, is discussed with the Bank Board during its preparation. For simplicity and clarity for the general public, the forecast is communicated as the “CNB forecast”.

Consultations with the Bank Board during the forecast production process take place at two preparatory meetings, where the assumptions, the position of the domestic economy in the business cycle and the current underlying themes and outlines of the baseline scenario of the forecast are discussed. During these meetings, the Bank Board discusses the risks and uncertainties of the emerging forecast and instructs the Monetary Department to develop additional scenarios or simulations that offer alternative views on possible future developments. The baseline and other scenarios are subsequently presented at the monetary policy meeting and published in the Monetary Policy Report.

At the monetary policy meeting, the Bank Board may express reservations about the baseline forecast by referring to upside or downside risks relative to the baseline. If the Bank Board has requested the preparation of alternative scenarios, individual board members may, depending on their perception of the situation, express a preference for one of these alternatives, including one involving a different monetary policy response.

The CNB uses fan charts as a communication tool to indicate that the forecast is always inherently uncertain and cannot be expected to be perfectly fulfilled. In this type of chart, the expected path of an economic variable is not shown by a single line but is surrounded by a fan of lines (or bands). The fans illustrate the uncertainty of fulfilment of the forecast, i.e. the extent to which the future path of the variable may deviate from the baseline scenario together with the relevant probability. The CNB publishes fan charts for five key macroeconomic variables – headline and monetary policy-relevant inflation, GDP growth, interest rates and the exchange rate of the koruna against the euro. The current forecast, including fan charts, can be found here.

The wider the fans, the greater the uncertainty about the future path of the variable. The width of the fans is based on the historical deviations of the forecasts from the subsequently published outcomes, in other words the forecast errors or the degree of (non)fulfilment of past forecasts. In general, if past forecasts have been fulfilled well, it can be expected that future ones will also be fulfilled well, and the fan expressing the forecast uncertainty will be narrower.

To calculate the width of the fans, the CNB uses the root mean square error (RMSE). The widths of the fan charts for the CNB forecast are normally updated once a year in January. The calculation incorporates the last 20 forecasts for which the outcomes are already known over the entire forecast horizon shown in the fan charts, as well as those forecasts for which the forecast horizon has only partially elapsed (those prepared in the last less than two years). However, the errors in older forecasts have less weight in the calculation than those in more recent ones.

In addition to the calculated RMSEs of past forecasts, the expert opinion of the Monetary Department’s economists is used in the construction of the fans. This occurs in situations where there are good reasons to believe that past forecast errors are not a good approximation of future ones. A case in point is the “exchange rate commitment” of 2013–2017, when, by definition, the forecasts for interest rates (which were technically at the zero lower bound for a long time) and the exchange rate (which was close to the intervention level) were being fulfilled exactly. Therefore, this period was not included in the calculation of the RMSEs of these variables in subsequent years. The fan charts used by the CNB thus cannot be interpreted as capturing past forecast errors purely mechanically. Nor should they be understood as an estimate of the degree of economic uncertainty surrounding one particular forecast.

The fan charts are constructed symmetrically. However, the Bank Board (and the Monetary Department in its analyses) may assess the risks and uncertainties of the forecast as being skewed in one direction or the other. The risk of such bias may be perceived differently by different board members. In addition, some risks and uncertainties are very difficult to quantify. The CNB’s approach to calculating the fan charts therefore does not aim to capture the perception of risks in the given situation.