Central banks objectives

In order for Central banks to quickly react to unpredicted developments and steer monetary policy to achieve its objectives, it is mandatory to have comprehensive systems for acquiring, sharing, and analysing economic and financial data. (Granger, 1992)

The question arises whether this analysis has to either largely depend on judgment, or whether mathematic, economical models can be used to improve this process? In specific, as policies have to be adjusted in relation to the future changes of the economy, can models be used to forecast its direction? This report aims to discuss the various types of models which central banks may use, and the advantages and disadvantages of each type. (Granger, 1992)


The primary task of most central banks is to operate monetary policy to secure price stability. In general, it can be said that central banks aim to be prepared for future economic changes. (Kennedy, 1993)

However, changes in monetary policy take time to influence the economy, in particular changes in inflation. A solution to that problem could be an analysis of these econometric data in a clear and logic framework. In order to do so, there are various models available which can be used to explain and analyse the economy. (Granger, 1992)

In the 1960s & 1970s computers have first been used to run mathematic models of the economy. However, as computer power was limited, only a small volume of data could be used. Nowadays, data volume and computer power have massively increased which enabled Central banks to compute models with higher precision and accuracy. (Kennedy, 1993)

Yet, central banks reduced the importance of these models. This was mainly due to insufficient efficiency due to the high number of economists needed to work on a model. Therefore, there has been a move to smaller & non-quantitative models.


In general, market economies aren't steered by state orders but by the decisions of millions of consumers, workers, manager, entrepreneurs and of course investors.

Hence, this results a gigantic task to predict everyone's behaviour, thoughts and expectations. The next sub-chapter will elucidate this issue.


If for example, the government raised the sales tax on electrical goods, the consumer would not know whether this tax increase would be passed on fully to consumers, or whether retailers or manufacturers might absorb part of the tax increase by reducing their profit margins. If however, the amount of price increase could be estimated, then the consumers could be expected to buy fewer electrical goods.

Probably, the consumers would rather spend more in the purchase of cheaper products than expensive, electrical goods.

However, If the tax increase on electrical goods led consumers to fear, would there be a further increase in the tax on electrical goods? Wouldn't it be more likely for the consumer to buy more electrical goods before prices for electronic goods went up again? Or perhaps, the tax rise will be merely temporary, in which case consumers could be expected to defer purchases of electrical goods and temporarily increase their savings.

The issue in this example clearly demonstrates the variety of different outcomes of a monetary policy decision and shows, that a precise prediction on consumer buying behaviour is very difficult and that public's expectations can play a very important role in determining the impact of a policy measure.


The effects of an interest rate rise will depend in large parts on what happens to real interest rates. However, these expectations are themselves not directly observable or measurable. Problems in measuring and predicting expectations make the modelling of monetary and financial behaviour particularly difficult. (Whitley, 1994)

A further consideration is that, even where the formulation of expectations appear to be well-modelled, people may react in a different way. For example, the public might interpret an unexpected interest rate rise as an indication that the central bank is more serious about cutting inflation than they had previously thought. If so, the public might anticipate further rises in interest rates and at the same time revise downwards their inflationary expectations.

The interest rate rise would then have a greater effect in reducing inflation than would otherwise have been the case. In such ways, models which have performed well in the past can be invalidated by policy changes.

Also, wrong interpretation of unexpected interest rise might cause greater effect in reducing inflation. (Whitley, 1994)


In general, the main task of a central bank model on macroeconomic stabilization is the design of a consistent set of policies intended to move an economy toward internal and external balance. They give central importance to fiscal policies to control the fiscal deficit. A main monetary policy tool ? the level of interest rates- enters only through its effect on money velocity and money demand. It effects on aggregate demand are ignored. Thus, a major ingredient in monetary policy is not used. When there is a need to reduce an excessive rate of inflation to ?normal? levels, more direct action may be needed. The Central Bank has three possible routes:

  1. The CB can use a ?nominal foreign exchange anchor?. Monetary policy aims at pegging the exchange rate to a country with low inflation to bring inflation to this level. This option however has made countries more vulnerable to balance of payments crises. (Granger, 1992)
  2. The CB can use a ?money growth anchor?: monetary policy aims at controlling the rate of growth of money supply. This policy however has been almost abandoned now by CBs, since the relation between money supply and inflation is unstable. (Granger, 1992)
  3. And finally, the CB can use ?inflation targeting? as a nominal anchor to bring inflation down to desired levels. Inflation targeting has worked well in many countries for the following reasons:
    • It is consistent with modern view of the power and limits of monetary policy.
    • It addresses directly the need to determine a long-run inflation objective.
    • It provides more transparency and accountability to monetary policy.
    • It provides a stable policy guidance (forces the CB to take into account long-term consequences of short-term actions).
    • It shields CB from political pressures from government and private sector to relax monetary policy for short-term gains.
    • It provides a ?framework of constrained discretion?


Under the IT system, the central bank manages monetary policy instruments with the direct goal of containing inflation over the medium run. In this setup, inflation becomes the overriding goal of monetary policy.

All the other indicators (output gap, money stock growth, the exchange rate, etc) become auxiliary variables; the central bank will take them into account only if this information helps it to improve its inflation forecast.

IT involves the following three steps:

  1. The Central bank sets a ?target? rate of inflation for the country for the ?medium? term: experience with monetary policy management in developed countries has shown that the impact of monetary policy changes on inflation works its effects with a significant lag (at least nine months, and up to two years for a full impact).
  2. The Central bank makes forecasts of the likely rate of inflation based on current conditions. This step requires the developing of sound inflation forecasting methods. This inflation forecast may indicate that inflation may be likely to be above the ?target?.
  3. The Central bank uses monetary policies (particularly the level of interest rates) to close the gap between the likely inflation rate and the ?target? rate. This requires the central bank to have a good model of the ?transmission mechanisms? from changes in monetary policy (such as interest rates) into aggregated demand and then into inflation rates.


Most Central Banks have two types of models:

  1. Models to Forecast Inflation based on current conditions without any monetary policy action.
  2. Models to Target Inflation based on the adjustment over time of interest rates and their transmission mechanisms to affect demand and inflation.

For these models, CBs follow a well-defined sequence of steps for preparation and presentation to policy-makers and public:

  1. Gathering information
  2. Input data into the database
  3. Maintaining a system to monitor and record news that could affect inflation outlook
  4. Presenting interest rate scenarios to Monetary Policy Committee
  5. Discussing technical proposals of the staff at policy level
  6. Interact with staff on impact of alternative scenarios for future policy interest rates
  7. Communicating inflation forecast to the market
  8. Follow up


From this graphic model you can see that there are two main variances for forecasting future economic movements.

Short-term forecasting includes:

  1. Consumer Surveys and Expert Judgments
  2. Statistical (time series) Econometric Techniques

Whereas, medium-term forecasting explains the structural models based on economic parameters.:

Empirical studies show that no one single model predict inflation consistently better than others. For this reason, many researchers and central banks use a combination of various models to forecast inflation.



In general, the usage of models in a central bank imply some sort of assistance for the bank in its decision making process. However, the usage of these models can only be exercised under certain restrictions leaving the central bank in a spot having to compromise and weight the different options, advantages and disadvantages of each model available.

There are three main areas of macroeconomic analysis, which are generally supported through the application of models:

  1. Understanding and quantifying how the economy and monetary policy work
  2. Monitoring where the economy stands and making short-term projections
  3. Making loner-term projections to determine what monetary policy to follow

In order to gather the information required normally econometric testing of hypotheses is used in order to highlight the different aspects of economic behaviour as well as to outline the correlations between different variables.

As an example, the central bank may wish to study and measure the effects, which interest rate changes have had in the past on personal savings and on business investment, and the impact of changes in the exchange rate on exports and imports. (Lionel Price, 1996)

In order to keep such research manageable, normally central banks generate series of separate projects, either equation by equation, or mostly in small groups of equations on related topics.

The outcomes of these research projects are then put together into a model and used to understand and quantify how the economy works. In addition it serves the purpose of monitoring the economy.

However, the problem is that the defined equations never fit the data perfectly and that each equation will therefore include an error term. Through observation, the error can then be understood better and adjustments can be made in order to improve the model which then additionally improve the monitoring of the economy.

Since the central bank wants a more complete understanding to make its projections the outcome of the research projects has to be brought together into a more comprehensive model. The difficulties embedded in this process are clear and simple. Complex interactions might be misinterpreted, a lot of the linkages can be missed and therefore the conclusions drawn might be wrong.


Structural models are relationship-based models, which are mainly based on economic behaviour. These models only take equations into account, which are plausible, even though variables might have shown some arithmetic relationship in the past.

This type of modelling is concentrated on the demand side of the economy, building up GDP as the sum of its main expenditure components:

  • Private consumption
  • Business investment
  • Public spending
  • Balance of trade

Normally these models do not include any details on production or employment in different economic sectors, the reason being that it is not reasonable to expect the central bank to have any sort of influence on this development in the near future.

So the focus is also set towards matters, on which the central bank can or is expected to have influence. Also these models are rather large as they are. Further disaggregation could lead to models becoming too complicated or to costly to operate.

The trend has been to build rather small models, which are based on micro-economic principles rather than to estimate coefficients from past data.

In chapter 4, the authors will be taking a closer look of how the process of modelling is conducted in the Bank of England. Yet, it is interesting to note, that in comparison, today the Bank of England uses a structural model of some 30 variables, in the 70s however they were using a model with over 600 variables.


These models are also called reduced-form models, since they do not attempt to explain the underlying economic behaviour.

Structural models include several equations with plausible relationships between the variables. The reduced-form model will only include a single equation showing the relationship between the two variables (showing evidence of association, not causation). These equations are referred to as black boxes. Symbolically one variable is put into the box the other variable comes out of the box without any explanation in regards to the relationship between both. The outcome then is analysed through statistical testing with past data.


A VAR model includes a number of atheoretic relationships. Therefore, VAR models do not use economic theory in order to define inter-relationships of different variables. The causal economic transmission mechanism between the variables is not explained in a VAR model. It may only be used to create short-term estimations and to outline statistical linkages between two variables. In other words, it can show a past trend, which outlines, that the values of two or more variables are linked between each other. In a VAR model, each variable depends on its own past values and past values of all the other variables in the system.

For example, if a VAR model contains just three variables: inflation, output, and money. Then inflation is explained by past inflation, output and money; output is explained by past output, money and inflation; and money is explained by past money, inflation and output.

Some of the main disadvantages of a VAR models lie within the requirement of large amounts of data and the estimates obtained are often puzzling and unexpected. They can easily be invalidated by changes in the structure of the economy. Moreover, their lack of economic rationale means they cannot normally be used for exploring the effects of policy changes.



The model used by the Bank of England until the 1990s includes 350 variables. One hundred were exogenous (set outside of the model) and the remaining 250 were determined within the model either by behavioural relationships or by statistical identities. Starting in the 1970s, the model contained 600 variables, but the tendency in the Bank of England was to use smaller models, as they are clearer and easier to understand and operate.

A team of four economists, who maintain and use the relatively small structural model, produce an inflation report every 3 months since 1993. The most critical input in the model does not come from the variables, but from the judgement of the forecasters. The outcome of the forecast can be greatly influenced by the modifications made in the equations of the model. However, every modification made, because of the personal judgment of a forecaster must be justified and explained to the users of the forecast. First, the forecasters will discuss the projection with the Chief Economist and other members of the research and analysis divisions. After the discussion, they send the modified forecast to the Governor and other directors of the Bank. If they believe that changes should be made, they will be quite small. Finally, the Inflation Report is passed on to the senior management. The final draft should be clear and it shouldn't include unnecessary information. The report can still be modified between the quarterly projections, but only if there is new data, which will influence the forecast. Such new data might include a new government budget or a sudden exchange rate change. Monetary policy in the UK is conceived in the medium-term, looking forward up to five years. However, such projections are not reliable, nor useful.


In comparison to the Bank of England, the Bank of Canada makes longer-term economic projections. Twice a year they make a medium-term projection, which looks at a 7 years horizon. However, they also have quarterly updates twice a year, mid-quarter assessments 4 times a year and monitoring each Friday, which is with two-quarter horizon (Morgan & Butter, 2000).

The Bank of Canada uses models because they are mathematical representations of the economy and they are designed to be simplifications of the complex reality. By using a model, a policy maker can see the impact of a particular economic development on the economy. Also, economic models help settle debates, which cannot be settled by theory alone (Bank of Canada, 2010).

There are three types of models used by the Bank of Canada. The first types are single-equation or indicator models. They are used for short-run predictions of inflation, output growth, and the exchange rate. A well-known single-equation model is used to forecast the Canada/ US real exchange rate, which is important factor for the Canadian economy.

The second type of models is the small multi-equation or reduced-form models. There are several models like this used by the Bank of Canada. One of the models is called NAOMI or the North American Open Economy Macroeconomic Integrated model. It consists of six behavioural equations, which determine output growth, core and GDP inflation, the real exchange rate, and short and long term interest rates. The expectations from this model are based on recent behaviour of the factors. It is therefore useful for short-term forecasting.

Another reduced for model they use is called USM or the United States Model. They use it to forecast some of the key US variables, because they affect Canada as well. It has only 3 equations.

The last type of models is the medium sized, dynamic general equilibrium models (DGEMs). The economic projection is made by one core model that reflects the mainstream view of the key macroeconomic linkages in the economy. The model is called ToTEM or Terms of Trade Economic Model. In 2005, it replaced the model used before ? Quarterly Projection Model (QPM). ToTEM is an open-economy, dynamic general equilibrium model that contains producers of four distinct finished products: consumer goods and services, investment goods, government goods, and export goods. ToTEM also contains a commodity-producing sector. The behaviour of almost all key variables in ToTEM is traceable to a set of fundamental assumptions about the underlying structure of the Canadian economy. Compared to the previous model, ToTEM is easier to use, modify and to be learnt by staff.

The first stage of making a projection in the Bank of Canada is gathering data and making projections for the rest of the world using models such as the USM. The second stage is inputting the data in the ToTEM model. The third stage is monitoring the model and the economy and on the next stage the staff of the bank can input their own judgement in the model and they modify factors, which are not captured by the ToTEM. The outcome is the model, including monitoring and judgement. This is visualised in the following graph.

Please be aware that the free essay that you were just reading was not written by us. This essay, and all of the others available to view on the website, were provided to us by students in exchange for services that we offer. This relationship helps our students to get an even better deal while also contributing to the biggest free essay resource in the UK!