Lena Malesevic

1
Investigating non-linearities in the
inflation-growth trade-off in transition countries

• Lena Malesevic
• Teaching assistant at the Faculty of Economics
  Split, Croatia
• PhD student at Staffordshire University, UK

Dubrovnik, 27.6.2007
2
Theory

• Akerlof et al. (2000)
• Palley (2003)
• Impact on growth

Empirical work

• Khan and Senhadji (2000) - the threshold is
  estimated to be at 1-3 percent for industrial
  countries and 7-11 percent for developing
  countries.

3
The model(s)
Our sample consists of 8 transition countries
(Bulgaria, Croatia, Czech Republic, Hungary,
Poland, Romania, Slovakia and Slovenia) and 13
years (1991-2003), i.e. 104 observations per each
variable.
(1)
(2)
(3)
(4)
4
(No Transcript)
5
Dependent variable GDP per capita growth Static panel (1) Static panel (2) Dynamic panel (1) Dynamic panel (2)
GDP growth (lag 1)     -0.052 0.126
GDP growth (lag 2)     -0.242 -0.281
GDP growth (lag 3) -0.130  
GDP gap (USA) 0.0364 -0.326 -0.814 -0.486
Investment/GDP 0.358 0.299 0.317 0.373
Govt. expenditure/GDP -0.157 -0.124 -0.036 -0.034
Population growth -0.680 -0.812 -1.225 -1.819
Life expectancy -1.618 -1.997 -0.624 -1.101
Secondary school enrolment 0.0593 0.0338 0.070 0.0298
(Exports imports)/GDP -0.0167 -0.006 0.054 0.029
Transition index -0.0147 0.411 3.064 3.703
War dummy -0.206 -0.240 0.314 0.126
Inflation -0.015   -0.065  
Ln (inflation)   -1.456   -0.836
Squared inflation 5.31e-06   -0.052  
, and denote 1, 5 and 10 percent level of
significance, respectively. Robust standard
errors were used and year dummies included.
6
(3)
The coefficient (?2) on the dummy variable,
without year dummies, static panel
D1 8.176 D6 2.369 D11 1.938 D16 -1.500 D21 -3.672
D2 6.114 D7 3.047 D12 1.537 D17 -1.500 D22 -3.672
D3 6.464 D8 2.612 D13 0.513 D18 -1.500 D23 -3.689
D4 4.396 D9 1.858 D14 -0.244 D19 -2.594 D24 -3.538
D5 3.336 D10 1.688 D15 -0.706 D20 -2.920 D25 -3.538
, and denote 1, 5 and 10 percent level of
significance, respectively. Robust standard
errors were used and year dummies included.
The coefficient (?2) on the dummy variable,
without year dummies, dynamic panel
D1 2.647 D6 0.675 D11 1.463 D16 -0.175 D21 -1.515
D2 1.542 D7 1.058 D12 1.816 D17 -0.175 D22 -1.515
D3 2.137 D8 1.478 D13 1.803 D18 -0.175 D23 -0.933
D4 1.669 D9 1.473 D14 0.609 D19 -1.526 D24 0.457
D5 0.323 D10 1.295 D15 0.769 D20 -1.879 D25 0.457
7
R2 for different values of threshold inflation
(?), without year dummies, static panel

? R2 ? R2 ? R2 ? R2
1 0.5685 6 0.6055 11 0.6289 16 0.6405
2 0.5740 7 0.6113 12 0.6333 17 0.6410
3 0.5857 8 0.6162 13 0.6364 18 0.6413
4 0.5942 9 0.6206 14 0.6385 19 0.6411
5 0.6006 10 0.6240 15 0.6397 20 0.6404
(4)
8
Wald (chi2) for different values of threshold
inflation (?), without year dummies, dynamic
panel

? chi2 ? chi2 ? chi2 ? chi2
1 2824.77 6 1223.84 11 1311.25 16 242.99
2 2601.39 7 1112.86 12 1375.98 17 235.43
3 1965.14 8 1186.02 13 251.68 18 229.67
4 1434.92 9 1188.26 14 250.53 19 226.56
5 1317.81 10 1275.43 15 249.82 20 221.54
9
Conclusion

• Static vs. dynamic model
• Non-linearities
• Placement of the kink?