Skip to main content

Table 5 Investment equations, III.

From: Asymmetric real-exchange-rate effects on capital accumulation: evidence from non-linear ARDL models for Mexico

 

(5.1)

(5.2)

(5.3)

(5.4)

Estimated coefficients

Depreciation component, RERD

0.306 (0.00)

0.477 (0.00)

1.083 (0.00)

0.628 (0.00)

Appreciation component, RERA

0.465 (0.00)

0.829 (0.00)

0.480 (0.12)

1.043 (0.00)

Industrial production index

2.950 (0.00)

1.893 (0.04)

  

Government investment

− 0.277 (0.00)

− 0.289 (0.00)

− 0.838 (0.00)

− 0.412 (0.00)

Nominal interest rate, NIR

− 0.628 (0.00)

− 0.875 (0.00)

− 1.055 (0.07)

− 1.161 (0.00)

Annual inflation rate, INF

0.368 (0.00)

0.482 (0.00)

0.476 (0.14)

0.624 (0.00)

Manufacturing exports

 

0.589 (0.03)

 

0.997 (0.00)

Gross domestic product

  

1.252 (0.00)

0.826 (0.00)

Error-correction coefficient, ρ

− 0.739 (0.00)

− 0.475 (0.00)

− 0.175 (0.00)

− 0.292 (0.00)

Intercept shifts

1995–2016, post-tequila crisis

− 3.48 (0.00)

− 5.99 (0.00)

− 6.01 (0.00)

− 7.32 (0.00)

2013–2016, global slowdown

− 1.25 (0.00)

− 1.63 (0.00)

− 5.09 (0.00)

− 2.62 (0.02)

Diagnostics

Jarque–Bera

0.880

0.216

0.978

0.892

Breusch‒Godfrey, n = 4

0.384

0.718

0.851

0.896

ARCH

0.695

0.112

0.544

0.465

Ramsey’s RESET

0.106

0.702

0.656

0.944

Adjusted R-squared

0.870

0.889

0.899

0.908

Wald and bounds tests

Wald RERD = RERA

0.086

0.052

0.001

0.109

Wald NIR = − INF

0.003

0.002

0.074

0.005

Bounds t statistic

− 8.69***

− 3.75

− 3.46

− 4.34*

Bounds F statistic

19.09***

5.76***

8.88***

6.52***

  1. Dependent variable: private fixed investment
  2. Long-run coefficients from non-linear EC ARDL models, 1988Q1‒2016Q3, 115 quarterly observations
  3. For illustrative purposes, the p values for the di coefficients from Eq. (3a) in the text are shown in parenthesis, next to the estimated coefficients
  4. For an explanation of the diagnostic, Wald, and bounds tests, see Tables 1 and 2
  5. The intercept was omitted from columns (5.3) and (5.4) due to lack of significance; column (5.1) includes quarterly outlier dummies for 1997Q4 and 2009Q3 (not shown)
  6. The estimated intercept shifts are reported as % of the investment levels observed in 1994 and 2012
  7. All the variables are measured in natural logs times 100, except NIR and INF, which are in %
  8. These estimations use the investment (and corresponding GDP) series shown as series B in Fig. 2