## External Debt And Gdp Relationship Economics Essay

## Abstraction

External debt and GDP relationship is largely investigated through empirical observation for different states in the literature. Most of these surveies employ co-integration and Granger causality trials and presume a inactive nexus. This survey differentiates from bing documents by using Kalman Filter algorithm in order to analyse clip changing empirical nexus between external debt and GDP.

In this paper, we investigated external debt and GDP relationship for 1989Q4-2011Q4 periods. First, we analyzed stationarity features of series using assorted unit root trials and all series found in I ( 1 ) . Then, we investigated co-integration relationship between GDP and external debt using Bound trial attack proposed by Peseran at Al. 2001 and found long run co-integration relationship between external debt and GDP. Finally, we analyzed external debt snap of income dynamically by using Kalman Filter theoretical accounts.

The empirical consequences suggest that external debt snap of income is diminishing in 1994 and 2001 crises period and so increasing after last one-fourth of 2002 which coincides with the political stableness term of Turkey. These findings support success of efficient debt direction and financial policy after crises period.

Keywords: External Debt, Bound Test, Kalman Filter, Turkey

JEL Codes: H63 ; C22,

## I.INTRODUCTION

The chief aim of the states is to increase the gross domestic merchandise every bit good as to better the societal and economic development of society. This aim could be realized with sustainable economic growing which becomes a really of import challenge for developing states confronting the budget shortage caused by the bit by bit spread outing current history shortage and the high degree of debt service, particularly external debt service.

The domestic resources of developing states remain deficient to keep a sustainable economic growing. This insufficiency of domestic resources is particularly due to inter-countries differences and unjust distribution of resources among states. Actual salvaging spread in developing states raises the demand of external financess for funding of investings. This spread is besides feasible for the Turkish economic system. Particularly after 1950, due to the deficiency of qualified labour force and engineering, external debt has increased to keep sustainable economic growing.

External debt was foremost defined in 1984 by the working group of external debt statistics of four international establishments including IMF, World Bank, Bank of International Settlements ( BIS ) and OECD as follows ; external debt, at any given clip, is the outstanding sum of those existent current, and non contingent, liabilities that require payment ( s ) chief and/or involvement by the debitor at some point ( s ) in the hereafter and that are owed to nonresidents by occupants of an economic system ( Klein, 1994 ) .

Between 1994-2001 period Turkey faced three economic crises and two of them had reflected Turkey ‘s ain economic instabilities straight. After 2001 crises period, Turkish economic system became more stable because of the political and structural transmutation. The external debt stock of Turkey was $ 113.6 billion in 2001 and reached $ 208.4 billion by 2006 and $ 306.7 billion in 2011.

Most of the surveies in the literature look into the relationship between external debt and economic growing using co-integration and Granger causality trials and presume a inactive nexus. This survey differentiates itself from bing plants by using the Kalman filter algorithm in order to account for clip changing empirical nexus between external debt and GDP.

The paper is organized as follows. The following subdivision briefly introduces literature reappraisal. Section 3 nowadayss informations and methodological analysis. Section 4 shows empirical consequences and subdivision 5 includes the decision.

## II. LITERATURE REVIEW

Many empirical surveies particularly focus on to a great extent indebted hapless states and analyze external debt and economic growing relationship via different econometric techniques including co-integration, causality analysis and GMM and panel informations analysis in the literature.

Chowdhury ( 1994 ) investigated public and private external debt and GNP relationship for developing and debitor Asiatic and Pacific states between 1970 and 1988 by using Granger causality trial. He found public and private sector external debts have a little and negative consequence on GNP.

Clements et Al. ( 2003 ) analyzed how external debt affects economic growing for low income states. They investigated external debt and economic growing relationship for 55 low income states for 1970-1999 periods using panel informations analysis and GMM method. They found that a lessening in the external debt stock has a direct positive consequence on GDP per capita growing in low income states, which means that there is negative relationship between external debt and economic growing.

Scharek ( 2004 ) investigated the relationship between external debt and economic growing utilizing informations set of 59 developing and 24 developed states for 1970-2002 period using panel informations analysis and generalised method of minutes ( GMM ) theoretical account. He found negative and important relationship between external debt and economic growing in developing states. Furthermore, he analyzed effects of public and private sector external debt on economic growing individually and found negative relationship from public external debt to economic growing but non frailty versa.

Cordella et Al. ( 2005 ) investigated how the relationship between external debt and economic growing alterations harmonizing to debt degree in developing states using panel informations and General Methods of Moments ( GMM ) theoretical account for 79 developing states covering the period 1970-2002. They found negative relationship between external debt and economic growing in the instances of no inordinate debt load.

Wijeweera et Al. ( 2005 ) investigated external debt and GDP and economic growing relationship for Sri Lanka for the 1952-2002 period using co-integration analysis and mistake rectification theoretical account. They found negative but statistically undistinguished relationship between external debt service and Gross Domestic Product ( GDP ) in the long term and statistically undistinguished relationship in the short-run.

Butts ( 2009 ) investigated causality relationship between external debt and economic growing for 27 Latin American and Caribbean states covering the period 1970-2003. He found unidirectional causality from economic growing to external debt for 13 Latin American and Caribbean states in both long term and short term.

Ogunmuyiwa ( 2011 ) investigated external debt and economic growing relationship for Niger utilizing clip series informations for the period 1970 – 2007 within the model of an mistake rectification theoretical account. Harmonizing to the empirical consequences, he found weak and statistically undistinguished relationship between external debt and economic growing and he could non happen any causality relationship between economic growing and external debt.

There are besides surveies which investigated the external debt and economic growing relationship for Turkey. These surveies are as Karagol ( 2002 ) , Javed and Azahinoz ( 2005 ) , A°pek and YaAYar ( 2008 ) and Uysal ( 2009 ) .

Karagol ( 2002 ) examined the relationship between external debt load and economic growing for Turkey utilizing five variables VAR theoretical account using co-integration and Granger causality analysis. He found external debt load negatively affects economic growing in the long tally. Furthermore, he found unidirectional causality from external debt to GDP for Turkey.

Javed and Azahinoz ( 2005 ) investigated effects of external debt on investing, export and economic growing for Turkey covering the 1983-2002 period and stated that external debt has a negative consequence on investing and positive consequence on exports, nevertheless no consequence on economic growing.

A°pek and YaAYar ( 2008 ) analyzed the relationship between economic growing and external debt within the model of the co-integration and causality analysis with informations from 1989 to 2007. They concluded that there is a co-integration between external debt and economic growing and they found bidirectional causality between external debt and economic growing in both long term and short term.

Uysal et Al. ( 2009 ) investigated external debt and economic growing relationship for Turkey covering the 1965-2007 period by using VAR theoretical account and found negative relationship between external debt and economic growing.

Most of these surveies in the literature dressed ore on external debt and economic growing relationship. In this survey, we investigated external debt and GDP relationship different from literature.

## III. DATA AND METHODOLOGY

In this survey, we investigate the relationship between external debt and Gross Domestic Product ( GDP ) . We use quarterly existent GDP and existent external debt series covering the period 1989Q4-2011Q4.

The existent GDP series in 1998 changeless thousand Turkish Liras were obtained from Turkish Statistical Institute. Real GDP series between the 1989Q4-1997Q4 period were obtained by using 1987 based changeless GDP growing rates. External debt series were obtained from Undersecretariat of Treasury converted to Turkish Liras utilizing average the exchange rate and deflated by using GDP deflator. Both existent GDP and existent external debt series were seasonally adjusted utilizing the Tramo-Seats methodological analysis and measured in natural logarithms similar to the empirical literature and natural logarithms of GDP and external debt were denoted as LY and LDB severally.

In the empirical analysis, we foremost analyzed stationarity belongingss of the series by using largely used unit root trials in empirical surveies including ADF, PP and Ng-Perron trials in empirical analysis.

After finding stationarity belongingss for series, we investigated the being of the long term co-integration relationship between existent GDP and existent external debt series using Bound Test developed by Pesaran et Al. ( 2001 ) . Bound trial attack has some advantages over the conventional co-integration theoretical accounts. First, The Bounds proving attack can be employed irrespective of whether the regressors are strictly I ( 0 ) or I ( 1 ) ( Peseran et al. , 2001 ) . Second, the bound trial co-integration attack has superior belongingss in little sample sizes than other co-integration attacks ( Narayan and Narayan, 2004 ; Mangir and ErtuAYrul, 2012 ) .

Finally, we investigated dynamic relationship between external debt and GDP by following a dynamic attack using Kalman filter to picture the clip changing interaction between external debt and economic growing.

## IV. Result

## IV.1. Unit Root Trials

First, we investigate stationarity features of series. In this regard, we employed ADF ( 1979 ) , PP ( 1988 ) and Ng-Peron ( 2001 ) trials. The consequences of unit root trials are presented in Table 1.

Table 1: Unit of measurement Root Test Results

ADF Test Results

LY

-2.645

I”LY

-9.188*

LDB

-1.907

I”LDB

-8.290*

ADF critical values for LY and LDB % 1=-4.066 and % 5=-3.462

ADF critical values for I”LY and I”LDB % 1=-3.507 % 5=-2.895

PP Test Results

LY

-2.883

I”LY

-9.196*

LDB

-2.094

I”LDB

-8.289*

PP critical values for LY and LDB % 1=-4.066 and % 5=-3.462

PP critical values for I”LY and I”LDB % 1=-3.507 % 5=-2.895

Ng-Perron Test Results

Mutual savings bank

MPT

LY

-12.550

-2.492

0.199

7.333

LDB

-6.625

-1.809

0.273

13.762

I”LY

-32.902

-4.053

0.123

0.753

I”LDB

-55.287

-5.256

0.095

0.448

Ng-Peron critical values for LY and LDB series ; MZa, MZt, MSB, MPT severally ;

% 1 significance degree -23.80, -3.42, 0.14 and 4.03

% 5 significance degree for -17.30, -2.91, 0.17 and 5.48.

Ng-Peron critical values for I”LY and I”LDB series ; MZa, MZt, MSB, MPT severally ;

% 1 significance degree -13.80, -2.58, 0.17 and 1.78

% 5 significance degree for -8.10, -1.98, 0.23 and 3.17

* denote % 1 significance degree

** denote % 5 significance degree

Harmonizing to Table 1,

For ADF and PP trials, the void hypothesis suggests that the series include unit root. The deliberate T statistics for all variables are less than the critical values in their degree signifiers for both ADF and PP trials. Therefore, the void hypothesis can non be rejected, proposing that all variables are nonstationary in their degree signifiers. The consequences of the first differenced variables show that the deliberate T statistics for all variables are greater than the critical values at % 1 degrees, proposing that all variables are integrated of order I ( 1 ) harmonizing both ADF and PP trials.

For Ng-Peron trial, harmonizing to, tests the void hypothesis indicates that the series have unit root and harmonizing to MSB and MPT tests the void hypothesis indicates that the series are stationary. For, trials, the deliberate T statistics LY and LDB are less and for MSB and MPT tests the deliberate T statistics for LY and LDB are greater than the critical values proposing that LY and LDB are non-stationary in their degree signifiers. For the first difference of series, harmonizing to, trials, the deliberate T statistics for LY and LDB are greater and for MSB and MPT tests the deliberate T statistics for LY and LDB are less than the critical values at 1 % degrees, proposing that LY and LDB become stationary after differencing so that LY and LDB series are I ( 1 ) harmonizing to Ng-Peron trials.

In drumhead, unit root trials shows that LY and LDB series are stationary after differencing so that LY and LDB series are I ( 1 ) .

IV.2. Bound Test Co-Integration Approach

After look intoing stationarity of series, we analyze co-integration relationship between existent GDP and existent external debt by using Bounds Test attack developed by Pesaran et Al. ( 2001 ) . For the Bound trial analysis, we foremost formed the Unrestricted Error Correction theoretical account ( UECM ) . UECM specification for our survey is shown in equation 1.

( 1 )

Where, LY is log of existent GDP, LDB log of existent external debt. In UECM theoretical account in equation 1, “ m ” represents figure of slowdowns and “ T ” represents tendency variables.

For proving co-integration relationship, the statistic underlying the process is the Wald or F-statistic in a generalised Dickey-Fuller type arrested development, which is used to prove the significance of lagged degrees of the variables under consideration in a conditional UECM ( Narayan and Narayan, 2004 ) . Null hypothesis for F trial is established as for our survey and calculated F statistics is compared with table underside and upper critical degrees in Pesaran et Al. ( 2001 ) . If the computed F-statistic falls outside the critical bounds, a conclusive deAcision can be made sing co-integration without cognizing the order of inteAgration of the regressors. For case, if the empirical analysis shows that the estimated F statistics is higher than the upper edge of the critical values, so the void hypothesis of no co-integration is rejected. If the estimated F statistics is lower than the bottom edge of critical values, there is no co-integration relationship between the series ( Narayan and Narayan, 2004 ) . If the deliberate F statistics is between the underside and upper critical values, no exact sentiment can be made ( Karagol, Erbaykal and ErtuAYrul, 2007 ) .

Maximal slowdown figure for UECM theoretical account is taken 8 and harmonizing to Schwarz standards and slowdown figure is found 1[ 3 ]. After finding lag figure of UECM theoretical account, we investigate co-integration relationship by comparing the computed F-statistic from UECM theoretical account with table underside and upper critical degrees in Pesaran et Al. ( 2001 ) . Table 2 shows the bound trial consequences.

## Table 2. Bound Test Results

## K

## F statistics

## Critical Value at % 5 Significance Level

Bottom Bound

Upper Bound

1

8.33

6.56

7.30

K is figure of independent variable figure in equation 1. Critical values are taken from Table C1.v at Pesaran et. Al. ( 2001:300 )

Harmonizing to Table 2, F statistics is higher than the upper edge of the critical values ; therefore the void hypothesis of no co-integration is rejected. As a consequence, we found a important long run co-integration relationship between existent GDP and existent external debt harmonizing to Bound trial analysis.

## IV.3. Dynamic Approach

In order to look into dynamic relationship between existent GDP and existent external debt, we employ dynamic attack on a classical mention of Harvey ( 1990 ) that introduces Kalman filter attack. The Kalman filter attack is based on a signifier of province infinite representation. A additive province infinite of the kineticss of an equation can be represented as

( 2 )

( 3 )

where in our instance is a vector of unseen province variables, where, and are adaptable vectors and matrices, and whereandare vectors of mean nothing, Gaussian perturbations. As stated in equation ( 3 ) , unseen province vector is assumed to alter over clip as a first-order vector auto-regression. The Kalman filter recursively estimates the parametric quantities by updating the appraisal with every extra observation ( ErtuAYrul, 2012 ) .

The Kalman filter specification used in our survey is presented in equations 4 and 5 below.

( 4 )

( 5 )

coefficient in equation ( 4 ) indicates external debt snap of income. External debt snap of income shows per centum alteration in income in response to a one per centum alteration in external debt. If external debt snap of income additions, same proportion external debt alteration will do more income alteration.

The clip changing parametric quantity estimations for external debt snap of income using Kalman Filter attack in 1994Q1-2011Q3 period are shown in Figure 1. The parametric quantity estimations for all variables of Kalman Filter attack are statistically important.

## Figure 1: Parameter Estimates for External Debt Elasticity of Income

Our consequences show that external debt snap of income lessenings in crises period of Turkey including 1994 and 2001 and so begins to increase after last one-fourth of 2002 which coincide with the political stableness term of Turkey.

## V. CONCLUSION

The domestic resources of developing states remain deficient to keep a sustainable economic growing. This insufficiency of domestic resources is particularly due to inter-countries differences and unjust distribution of resources among states. Actual salvaging spread in developing states raises the demand of external financess for funding of investings.

External debt and economic growing and GDP relationship is largely investigated through empirical observation for different states by using assorted econometric techniques in the literature. Most of these surveies employ co-integration and Granger causality trials and presume a inactive nexus. This survey differentiates from bing documents by using Kalman Filter algorithm in order to analyse clip changing empirical nexus between external debt and GDP.

In this survey, we try to look into dynamic relationship between external debt and GDP for Turkey. The survey uses quarterly existent Gross Domestic Product ( LY ) and existent external debt ( LDB ) series covering the period 1989Q4-2011Q4.

In empirical analysis foremost, we analyzed stationarity belongingss of the series by using ADF, PP and Ng-Perron trials and all unit root trials shows that LY and LDB series are stationary after differencing so that LY and LDB series are I ( 1 ) .

Then, we investigated the being of the long term co-integration relationship between existent GDP and existent external debt series using Bound Test developed by Pesaran et Al. ( 2001 ) and found a important long run co-integration relationship between existent GDP and existent external debt harmonizing to Bound trial analysis.

Finally, we investigated dynamic relationship between external debt and GDP by following a dynamic attack using Kalman filter. Our consequences show that external debt snap of income lessenings in crises period of Turkey including 1994 and 2001 and so begins to increase after last one-fourth of 2002 which coincide with the political stableness term of Turkey. These findings are compatible with prudential financial and pecuniary policy stance of Turkey after 2001 crises period and support success of efficient debt direction and financial policy after crises period.