KOFI - A Macromodel of the Swedish Financial Markets

The portfolio balance approach, on which KOFI is based, as well as its formal structure, has been described in an earlier Working Paper. The purpose of the present paper is to place the model in its institutional context and to illustrate its properties by simulations. Furthermore, previous work with Swedish and foreign financial models has shown that the statistical input has been a bottleneck hampering full exploitation of the portfolio balance approach. In order to widen the bottleneck as far as possible, extensive work has been undertaken in order to supply KOFI with relevant unpublished statistics. The number of sectors in KOFI is less than in the Swedish financial accounts, but larger than in similar foreign models, and thus requires a relatively large volume of statistics. That volume is manifold as data are collected on a yearly as well as a half-yearly and quarterly basis. To link up with KOSMOS - which is a semi-annual model - yearly statistics had to be transformed into half-yearly data. Another comlication has been frequent incompatibility between the data from Statistics Sweden and from other sources. The statistical work has focused on capital markets and foreign exchange markets i.e. on the areas which represent the core of a financial model. The selection of sectors and of the variables in the sector portfolios has been made with regarded to their importance in simulations of financial macroprocesses. In practice the handling of the model has required that - compared to the financial accounts - a number of items are excluded because they are relatively small, relatively constant or difficult to interpret. To the latter category belong the residuals ("other assets and liabilities") that appear on both the debit and the credit side of the balance sheets of the portfolios. The accounting framework of the model is shown in section 2. Section 3.1 contains an overview of the development in the Swedish financial markets. As a starting point is used a table of the changes of the portfolio of the total economy distributed between the various components: money, certificates, bonds and loans. This gives an idea about the structural changes in later years caused particularly by deregulation. Long run changes in supply are otherwise not so easily revealed in an analysis which centers on the demand behaviour of individual sectors. As further background material, the section contains tables of historical changes of demand and supply in the markets for bonds and certificates. These tables are derived from KOFI sector balances. Such tables are of vital importance for the interpretation of the simulations. Section 3.2-5 deals with various institutional aspects of KOFI`s sectors that are relevant for their modelling. Subsections 3.2 and 3.3 give brief accounts of the borrowing behaviour of mortgage institutions and of Central Government. Subsection 3.4 raises the question whether the model should distinguish or not between private insurance companies and the National Pension Fund. It appears there are good reasons for aggregating the two subsectors. The design of the "residual sector" of the model is described in subsection 3.5. The residual sector contains households, non-financial enterprises, local authorities, investment companies and finance companies. There are arguments both for and against having households and non-financial enterprises in a common sector. Some problems concerning the modelling of the foreign sector are treated in subsection 3.6. In section 4 certain formal aspects of KOFI are discussed as well as the econometric properties of the model. The latter concerns the determination of the long and the short term rates of interest, the exchange rate and the portfolio choice of the various actors. (These relationships are updated versions of those which have been more extensively described in the report referred to in footnote 1). Section 5 contains four simulations undertaken with KOFI. They concern the effects of respectively, a change of the strategy of government borrowing, an increase of the domestic short term rate, an increase of the foreign interest rate and, finally, an increase of the budget deficit. The simulations are desribed in the form of deviations from a given basic scenario and limited to developments within the financial markets. The developments are registered as changes of the balance sheets of the various sectors. The emphasis is mainly on the processes generated by the interaction between the sectors and not so much on the decision making by each actor. A typical outcome of simulation i the volume of open market operations required by the Central Bank to defend a targeted short-term rate of interest. Other outcomes concern e.g. the paths followed over a number of periods by lending and borrowing by various sectors. Section 6 gives some indications of the future development and use of KOFI.