An Intertemporal Benchmark Model for Turkey’s Current Account

actual consumption smoothing current account, if the national cash flow inclusive of interest
payments, (y_t - i_t - g_t ) and consumption c_t are I(1) processes and both are stationary in first
differences and a cointegrating vector exists, then the consumption tilting parameter θ can be
computed as the cointegrating vector between the two series (in levels). In the following section
we test for stationarity and cointegration and provide results for θ.

IV.   Data and Results

Data

For the IBM we require the real world rate of interest and national cash flow and
consumption data for Turkey. The world rate of interest (r_t ) is computed by using 10-year US
treasury nominal bond yields and converting them to real rates by inflation rates from US CPI.
Data for the world rate of interest are available from the Bureau of Labor Statistics website.

National cash flow is made up of three components (y_t - i_t - g_t ) . The first component
y_t includes GDP (q_t ) and net investment income (r b_t ) . We do not need to compute net
investment income as it is one of the components of the current account for which data is
available directly. The other two components of national cash flow are investment (i_t ) and
government expenditure (g_t ) . In addition, we require consumption (c_t ) . Data for all these series
are available from the Central Bank of Turkey website. All data is in constant Turkish liras.

We use quarterly data from 1992 to 2004. Our choice for the period is restricted to data
availability. While annual data was available for a longer period, we wanted to focus on the
current state of the Turkish economy. Therefore, we opted for quarterly data for a shorter period.
All data for the econometric analysis is seasonally adjusted.

Tests of IBM for Turkey

12

<<   10   11   12   13   14   15   16   >>

More intriguing information