Models for short-term and long-term forecasts in conjunction with assessments

Forecasts for the real economy and inflation and analyses of the effects of monetary policy on financial markets, the real economy and inflation are central to the monetary policy decision-making materials.^[13] Nyman and Söderström (2016) discuss what the Riksbank's inflation-targeting regime means for the forecasting work at the Riksbank, and how the forecasts can therefore differ from those of other forecasters. The work on forecasts is mainly carried out in the Monetary Policy Department. The methods for making forecasts are continuously evolving, and the Riksbank has a set of models that are used to make as good an analysis as possible of the current economic situation, economic developments in the short to medium term and the effects of monetary policy measures. These models are based on economic theory and empirical studies of various correlations in the economy. In addition to a number of different time series models, the Riksbank also uses general equilibrium models in its forecasting work.^[14] The Riksbank has, for instance, developed the model MAJA, which belongs to the group DSGE models of the Neo-Keynesian type, see Corbo and Strid (2020), and Ringqvist, Stockhammar and Strid (2020). They can help explain economic events and assume that market mechanisms and economic policies will bring the economy back to normal (long-term equilibrium) after a disturbance. General equilibrium models are an important tool for studying alternative economic scenarios and how monetary policy responds over time to disturbances.

To form an idea of where the economy is to start with, different types of indicators are used in addition to outcome data, for example, for production, the state of the labour market, pricing plans and confidence among households and companies. When making short-term forecasts (one to two quarters ahead), greater emphasis is placed on time series models and indicator models that have good short-term forecasting characteristics. Long-term forecasts are made more widely with the aid of macroeconomic models and assumptions on long-term equilibrium, which together with the monetary policy assumptions provide a coherent assessment of the driving forces of economic development. In addition, smaller models are used that focus solely on selected economic relationships.

Since models are simplifications of reality, model forecasts are a good starting point that helps structure the discussions and focus on the most central issues for monetary policy. To varying degrees, models assume that history repeats itself or that economic relationships are stable over time, and therefore they also need to be supplemented by analyses and judgements by experts who have an insight into mechanisms and economic developments that the models do not always capture. The overall information from the models is finally weighed together with other information to produce forecasts for economic activity and inflation.^[15] An analysis by Lindé and Reslow (2017) points to the models not explaining the Riksbank’s forecasts to the extent sometimes claimed by external critics and reviewers. The experts’ assessments become particularly important when unusual events and/or structural changes that alter the function of the economy take place. The coronavirus crisis has been such a period in which analytical methods based on historical relationships, observed outside times of crisis, have had difficulty in providing relevant support in the decision-making. So-called real-time indicators, which show progress without a time lag, have therefore been an important tool. Examples include indicators of mobility in society measured from mobile phone data, the booking situation in hotels and restaurants, and data for card transactions.^[16] See Ewertzh, Falk, Hesselman, Hull, Löf, Stigland and Tibblin (2020).

Process for producing decision-making data

At the beginning of the monetary policy process, APP employees discuss with the Executive Board the issues on which the preparation will focus. The purpose of this first meeting is for the department to receive guidance from the Executive Board: Is the proposal for prioritisation presented by the department appropriate and does it respond to the needs of the Executive Board? This may involve in-depth analysis work, which also leads to publication in the Monetary Policy Report or in the Riksbank's other publications. The main focus of the discussions is on the work ahead of the next monetary policy decision, but longer-term planning is also discussed. In this context, the main uncertainties and risks surrounding the economy and monetary policy are also discussed. APP provides suggestions for alternative scenarios that can help illustrate how the forecasts would be different if specific events occur in the economy.^[17] For example, the effects on Swedish inflation and monetary policy that changes in the price of oil, productivity, the rate of wage increases or international demand may result in. The scenarios are sometimes published in the Monetary Policy Report. There is also sometimes another introductory meeting with the Executive Board, where APP briefly presents the new information that has been received since the previous monetary policy meeting.

For about three weeks, APP works internally on producing a main scenario for economic activity and inflation in Sweden. In the first phase of this process, it is a question of clarifying the conditions for the forecast, that is, what happens independent of how one chooses to formulate monetary policy in Sweden at the next monetary policy meeting. This includes events abroad, the current financial conditions and the current economic situation in the Swedish economy, the 'current situation’ (nowcasting). APP's discussions on the outlook for international economic events and inflation have then focused particularly on Europe and the United States, which are Sweden's main trading partners. The discussions about financial markets relate, among other things, to the exchange rate and interest rates, to assess the transmission of monetary policy. At a meeting on the current assessment of the Swedish economy, new information since the previous monetary policy meeting is discussed in depth, and how this relates to the Riksbank's forecasts for the various components of GDP, the labour market, wages and inflation. Since new information is published continuously, APP continuously updates the current assessment and forecasts for the Swedish and international economy until the date of the monetary policy decision.

In the next phase, APP uses macroeconomic models to produce a proposal for the main scenario for the entire forecast period. APP first forms an idea of overall macroeconomic developments by focusing on some key variables such as inflation, GDP, unemployment, the repo rate and the exchange rate. Based on the forecast for the macroeconomic outlook, more detailed forecasts are then formulated, for example for the labour market and inflation.^[18] Documents that clarify and document how the final assessment forecast relates to different model results are shared with the Executive Board in connection with the meeting of the monetary policy group.

At the same time as the proposal for the main scenario is being prepared, APP also works on producing, with the aid of material from AFS and AFM departments, a proposal for monetary policy. The proposed monetary policy is not a recommendation from the department, but the aim is that the proposal that is produced and presented to the Executive Board should reflect how the Board has normally chosen to act on the basis of economic and inflation prospects. When there are unusual disturbances to the economy and many new complementary tools are used alongside the policy rate, this work becomes more complex. Cooperation with other policy departments at the Riksbank is therefore of great importance. In this context, monetary policy alternatives are also produced to illustrate the effects on economic activity and inflation prospects.