Revisiting growth empirics based on IV panel quantile regression

Abstract

We analyse the well-known issue of economic growth convergence using quantile regression. Most previous studies have used a least squares (LS) method or variation, which focuses on the issue only at the mean of the growth rate. Therefore, such results cannot provide a satisfactory answer to what can happen if the growth rate is far from the conditional mean level. For example, we consider the following question: do we still have economic growth convergence or is the convergence speed changed in a low growth period such as the ‘Great Recession,’ that started in 2008? We propose using instrumental variable panel quantile regression to answer this question. Our empirical findings demonstrate that economic growth convergence occurs at all quantiles over the entire conditional distribution, but that the convergence speed does depend on quantiles; the convergence speed is much higher when the GDP growth rate is at either high or low quantiles.

Keywords:

• quantile regression
• panel data
• endogeneity
• growth convergence

JEL Classification:

• O4
• C2

Notes

¹ See Barro and Sala-i-Martin (2003) for more details.

² We have also tried to estimate the augmented Solow model. Estimation results show that the estimated coefficient for human capital is either insignificant or wrongly signed, which is consistent with Islam (1995). Therefore, we will employ the original Solow model only in our empirical study.

³ The ‘Great Recession’ termed in Stock and Watson (2012) is different from the previous postwar recessions because it (i) is persistent, (ii) has a lower mean growth rate and (iii) shows much slower recovery.

⁴ Following the convention in the literature, we use the lagged values of explanatory variables as IVs in our empirical study; specifically, we use the first lagged values as our instruments.

⁵ Mankiw et al. (1992) excluded oil producers from their data-set because the major portion of their recorded GDP is from the extraction of oil resources, and as a result, the standard growth theory fails to explain their GDP growth.

⁶ When we compute $\lambda$ from ${\beta }^{\ast }$ using the equation $\lambda =-ln\left({\beta }^{\ast }\right)/\tau$, we set $\tau =5$ given that 5-year time intervals are used in the estimation.

⁷ Confidence intervals are constructed from the relevant percentiles of 1000 cross-sectional bootstrapping replications, following Kato et al. (2010).

⁸ It is important to note that the type of convergence under consideration is ‘conditional convergence’ because we control for all the included covariates, which can be important determinants of the steady-state level for an economy.

Additional information

Funding

Tae-Hwan Kim is grateful for the financial support provided by the DAELIM SUAM FOUNDATION[grant number DAE-LIM SUAM 2013-04].