Skip to main content

The Official Journal of the Pan-Pacific Association of Input-Output Studies (PAPAIOS)

Table 1 Panel unit root test

From: Re-estimation of the savings retention coefficient in OECD countries: a new measure of home country bias

 

Case 1

Case 2

Case 3

Case 4

Periods

1960–1974

1960–1974

1960–1969

1970–1979

1980–1989

1990–1999

2000–2008

2009–2014

2000–2014

CIPS of \(\widetilde{I}_{it}\)

− 1.5946*

− 2.1248*

− 5.8300***

− 1.7953**

− 2.1470*

− 2.4335**

− 1.6222**

− 5.7017***

− 1.6654**

ADF Type

None

Constant

Const. + Tre

None

Constant

Constant

None

Const. + Tre

None

IPS

− 1.5648*

− 2.1770**

− 2.2987**

− 1.6769**

− 2.7004***

− 1.8643**

− 1.9130**

− 3.6768***

− 2.5682***

CIPS of \(\widetilde{S}_{it}\)

− 1.5312*

− 2.1149*

− 3.1990***

− 5.6749***

− 2.2202*

− 3.6409***

− 2.8321*

− 6.1013***

− 1.3800

ADF Type

None

Constant

Constant

 Const. + Tre

Constant

Const. + Tre

Const. + Tre

Const. + Tre

 

IPS

− 2.2421**

− 2.8813***

− 2.8096***

− 2.8813***

− 2.1426**

− 2.3230**

− 0.0147

− 4.5989***

− 1.9219**

  1. *, **, and *** indicate significance at the 10%, 5%, and 1% levels, respectively. The CPIS tests which assume serial correlated errors are based on the t-statistics of the OLS estimate of \({\rho }_{i}\) in the following regression for the series: \({\left\{{y}_{it}\right\}}_{i=1,\cdots ,N}^{t=0,\cdots ,T}\),
  2. \(\Delta {y}_{it}={\alpha }_{i}+{\gamma }_{i}t+{\rho }_{i}{y}_{it-1}+{c}_{i}{\overline{y} }_{t-1}+\sum_{j=0}^{p}{d}_{ij}\Delta {\overline{y} }_{t-j}+\sum_{j=1}^{p}{\beta }_{ij}\Delta {y}_{it-j}+{\mu }_{it},\)
  3. where \({\overline{y} }_{t-1}=\left(\frac{1}{N}\right)\sum_{i=1}^{N}{y}_{it-1}\,and\,\Delta {\overline{y} }_{t}=\left(\frac{1}{N}\right)\sum_{i=1}^{N } \Delta {y}_{it}\).
  4. Three ADF types were tested, such as Type 1 (Const. + Tre.): \({\alpha }_{i}\ne 0\) and \({\gamma }_{i}\ne 0\), Type 2 (Constant): \({\gamma }_{i}=0\) and Type 3 (None): \({\alpha }_{i}\ne 0\) and\({\gamma }_{i}=0\). Repeated CPIS tests were applied from Type 1 to Type 3 until the null hypothesis was rejected at the 10%, 5%, and 1% significance levels. The test results of the CPIS tests are reported in Table 1, where the results of the IPS tests (panel unit root test with cross-sectional independence) were also reported for comparison with the CIPS test. The lag length p was automatically chosen based on AIC.