Full text: Extension of the empirical stock-flow consistent (SFC) model for Austria (163)

IHS—Miess, Schmelzer/SFC Model Austria—27
3 The Model
The model described in the following follows the basic logic of national accounting as put forth
in section 2.1.1.
Firstly, non-financial transaction flows (current accounts) are set forth, including flows deter-
mined by behavioural equations for agents (e.g. consumption, investment), exogenous variables
(e.g. exports, imports), and flows implied by stocks of the previous period (e.g. interest or
divident payments). Outcome is primarily observed as change in GDP and NLNB.
Secondly, financial transactions (accumulation accounts) are depicted, including behavioural
equations (portfolio choice), and variables exogenous in the model (revaluation of assets, amount
of balance sheet extension). Outcome is observed by the actual holdings of different assets by
agents, and their net worth at the end of the period, thus obtaining the closing balance sheet for
this period (the opening balance sheet for next period).
The method to calibrate the model is as follows: firstly, we reformulate the equations taking
the variables as fixed (taken from past NASA data 1995 - 2014) and the parameters as the
unknowns. We then calculate parameter values and obtain trends for the parameters for these
time series for the past, which are shown for the most important parameters below. We then
use past trends to forecast the development of the parameters into the future. In most cases,
we have used the simplest possible forecasting method for the parameter - taking the last value
available in the data and keeping it fixed. This was due to 1. Time restrictions in the model
construction stage due to the large amount of data work we had to manage, but also 2. since
we do not want to distort the dynamics of the model too much by strong assumptions on the
trends of the parameters. Since the trends in most parameters are very stable, it is possible to
get a look at the dynamics of the behavioural assumptions in the model itself, not influenced too
strongly by the trends in parameters.
After constructing the business as usual scenario, we obtain the effects of the policy mea-
sure by comparing the scenario simulation with the business as usual scenario. Currently, the
forecasting horizon of the model runs until 2025.
Notation: below, parameters are denoted by lower case Greek letters, variables by capitalised
Latin letters. Index t signifies time, index s economic sectors (institutional units), direct means
the direction of payment: received (RECV) or paid. The index finpos relates to the financial
positions of a sector, i.e. whether the financial instrument is held as an asset (ass) or as a liability
(liab). The subscript fa relates to the different classes of financial assets in the model.
3.1 Non-financial Transaction Flows
3.1.1 Behavioural Equations and Parameters
The core behavioural equations that decisively regulate the model behaviour are partly con-
structed in reference to the literature, mostly based on Godley and Lavoie (2007), or specified
according to empirical evidence put forth in Schmelzer (2015). Most importantly, for our empir-
ical SFC model, we obtain the parameter values directly from national accounting data. Since
this is a very preliminary version of the model, the extrapolation procedures for parameter trends
are work in progress, and are certainly open to discussion.
Household Consumption Ct is taken as a fixed fraction ?1,t of disposable household income
INCt as determined in equation 16 plus a fixed fraction ?2,t of household’s last period’s hold-
ings of deposits DEPt (their primary means of payment, and their storage of liquid means for

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.