טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentJulia Polak
SubjectAsymmetric Cointegration: Applications to Macro Variables
in Israel and the OECD
DepartmentDepartment of Industrial Engineering and Management
Supervisor Mr. Lieberman Offer
Full Thesis text - in Hebrew Full thesis text - Hebrew Version


Abstract

This research examined the relationship between government expenditure (GE), its components[1] and gross domestic product (GDP), by means of asymmetric cointegration. The empirical examination included Israel and 10 members of the OECD[2]. Each country was examined separately. The basic assumption of this research is that governments behave differently in different market circumstances.


Cointegration assumes long-run relationship between two or more time series. When one or more of the series deviate from the long run equilibrium, a short-run correction is applied to restore the long-run balance.


The assumption of asymmetric cointegration was tested using a Lagrange Multiplier (LM) test under the null model. The asymptotic distribution of the test was calculated by a  bootstrap procedure.


Current research lead us to a number of conclusions. Governments tend to keep the GE (or its component) to GDP ratio constant over years. When the ratio changes, policy-makers adjust the government expenditure to achieve the desirable ratio (long-term balance). The reaction to economy changes is different when the economy is in "routine" state from the reaction when the economy is in "crisis". Additionally, examination of the relationship between GE and GDP, as well as of the relationship between Government Final Consumption Expenditure and GDP, in terms of asymmetric cointegration, allowsus  to identify for most of the countries in this study a "punishment effect", which means that the economy reacts negatively to government expenditure growth and that all countries except Italy and Spain make a substantial effort to limit the GE to GDP ratio.


We also identified asymmetric relationship between Government Investments and GDP, between Interest Payments and GDP and between Central Government Debt and GDP. We found no empirical evidence for asymmetric cointegration between Public Spending on Education and GDP and between Current Transfers and GDP, though there is still evidence of classical cointegration.


To further investigate the relationship between GE and GDP, we recommend to develop models with time varying  thresholds. Also, we recommend to extend the research to other countries and additional relationships, e.g. the relationship between GDP and investment in infrastructure, investment in health, volume of trade with other countries, etc. Examination of relationships between three or more series is also recommended.



[1] Current Transfers, Investments, Interest Payments, Central Government Debt, Final Consumption Expenditure and Public Spending on Education.

[2] USA, United Kingdom, Australia, Canada, France, Germany, Italy, Japan, South Korea and Spain.