Working Paper Series: Special Edition of 2016 to 2018 Interns

3.2 The Error Correction Model (ECM)

An augmented growth model represented in a Cobb-Douglas production function with constant returns to scale as in Equation (10) underpins the ECM: " = " ∙ c ∙ e ∙ (%$c$e) " " " where Y is output, A the level of technology or Total Factor Productivity, K physical capital, H human capital (proxied by average years of schooling) and L is labour. Loening (2005) sought to address the possibility of multicollinearity between capital and labour by standardising output and capital stock by labour units. This also imposes the restriction that the scale elasticity of the production factors is equal to unity. The same is done in this study with the following logarithmic expression: (10) average human capital. An ECM is used to mitigate against the issues that arise after the time-series is transformed to stationarity by first differencing. It combines long-run information with short-run adjustment mechanisms to account for unit roots common in macroeconomic time-series taking the form: ∆ " = + % ∙ Δ log " + A ∙ Δ log "$% + D ∙ log "$% + M ∙ log "$% + 9 ∙ log ℎ "$% + ∑ H H ∙ H," + " in equation 11 . As with Loening (2005), it is considered a measure of the speed of adjustment in which the system moves towards its equilibrium on the average. The constant term represents a freely moving technology parameter. Dummy variables representing shocks to GDP were included to chronicle the growth experience of St Vincent and the Grenadines. The 1972 dummy captures the effects of a tightly contested election cycle, while the 1975 dummy accounts for the shift in the timing of the government’s financial year and spill-off effects from the 1973-74 energy crisis. (12) Estimates of the parameter γ D are used to calculate the elasticities of and (11) " = log " + ∙ log " + ∙ log ℎ " + " = q are output and capital in intensive terms, and where = q and = q stands for

212