Presentation is loading. Please wait.

Presentation is loading. Please wait.

Helga Kristjánsdóttir, 15. maí 2007

Similar presentations


Presentation on theme: "Helga Kristjánsdóttir, 15. maí 2007"— Presentation transcript:

1 Helga Kristjánsdóttir, 15. maí 2007
Estimating the Impact of Time-Invariant Variables on FDI with Fixed Effects (Mat á áhrifum tíma óháðra breytna á beina erlenda fjárfestingu þegar tekið er tillit til fastra áhrifa) Helga Kristjánsdóttir, 15. maí 2007 Faculty of Economics and Business Administration University of Iceland

2 Faculty of Economics and Business Administration
Inngangur Rannsóknin mælir föst áhrif panel gagna með vektor sundurtekningar metli (XTFEVD). XT fixed time FE fixed effects VD vector decomposition Notað við mat þriggja gagnasetta beinnar erlendrar fjárfestingar að teknu tilliti til fastra áhrifa UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

3 Faculty of Economics and Business Administration
Inngangur Niðurstöður leiða m.a. í ljós að sé ekki tekið tillit til fastra áhrifa þá sé hætt við hneigðu mati breytna sérstaklega þeirra sem standa fyrir viðskiptakostnað og menningu Að teknu tilliti til fastra áhrifa bendir mat margra tíma óháðra breytna til mikilvægis lóðréttrar fjárfestingar UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

4 Faculty of Economics and Business Administration
Inngangur Að teknu tilliti til fastra áhrifa Fæst jafnframt að sé hneigðu mati eytt (eliminating biases) hverfur matsmunur milli gagnasetta að mestu Þetta bendir til þess að Ólíkar fræðiniðurstöður sem rekja má til mismunar gagnasetta má e.t.v. skýra með XTFEVD matsaðferðinni UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

5 Faculty of Economics and Business Administration
Introduction This paper applies the panel fixed effects with vector decomposition estimator. Applied to three FDI datasets to estimate the impact of time-invariant variables on FDI while including fixed effects. UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

6 Faculty of Economics and Business Administration
Introduction We find that the omission of fixed effects significantly biases several of these variables, especially those proxying for trade costs and culture. After including fixed effects, we find that many time-invariant variables indicate the importance of vertical FDI. UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

7 Faculty of Economics and Business Administration
Introduction We also find that by eliminating these biases, the differences across datasets largely disappear. Thus, controversies in the literature that are driven by differences in data sets may be resolved by using this estimation technique. UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

8 Faculty of Economics and Business Administration
Probability Density Function for estimators The point estimator is said to be an unbiased estimator of the parameter if UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

9 Faculty of Economics and Business Administration
Literature Plümper and Troeger (forthcoming) have developed a three-stage estimation procedure that efficiently estimates the impact of time-invariant variables while also controlling for fixed effects. The primary goal of this paper is to apply this panel fixed effects with vector decomposition (XTFEVD) method to three commonly used FDI data sets. fevd stands for ”fixed effects vector decomposition” UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

10 Faculty of Economics and Business Administration
Literature Hofstede (2001) 5 cultural distances 1. Power Distance (PD) 2. Individualism (Independence) 3. Masculinity (Masculinity) 4. Uncertainty tolerance (Uncertainty) 5. Long-term orientation (Time) UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

11 Faculty of Economics and Business Administration
Model Setup The basic equation estimated can be described as: UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

12 Faculty of Economics and Business Administration
Data Three most commonly used FDI sets of data: OECD data World Bank data WDI (World Development Indicators) BEA Bureau of Economic Anlysis bea.doc.gov data UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

13 Faculty of Economics and Business Administration
Estimation Results Estimation Results (World Bank Data) Omission of fixed effects leads to biased estimates for trade cost and culture etc. Inbound Outbound OLS FE XTFEVD GDP 0.977*** 1.371*** 1.357*** 2.254*** 2.448*** 2.467*** Pop -0.193* -3.698*** -3.703*** -0.933*** -5.419*** -5.412*** TC *** *** *** *** -2.508 -1.909 School -0.768*** -0.478 -0.495 -0.215 1.614*** 1.613** PD 0.240* 3.723*** 0.405*** 4.449*** UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

14 Estimation Results, cont.
Inbound Outbound OLS FE XTFEVD Independence 0.819*** -0.733*** 0.251 -1.870*** Masculinity 0.042 3.334*** -0.174* 4.304*** Uncertainty -0.109 -3.875*** -0.759*** -5.012*** Time -0.058 1.155*** 0.341** 2.189*** Area 0.431*** 1.161*** 0.143*** 0.952*** Island 0.002 -3.777*** 0.408*** -4.234*** Landlock -0.046 -3.092*** 0.059 -4.633*** UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

15 Faculty of Economics and Business Administration
Estimation Results We cannot apply the conventional fixed effects estimator to the time invariant variables - and the OLS estimator provides us with biased results for these (unobserved characteristics usually dealt with by fixed effects) When OLS then of unobserved country heterogeneity UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

16 Faculty of Economics and Business Administration
Estimation Results If time-invariant variable then estimating fixed difference leads to multicollinearity (cross-sectional difference) Time invariant variable include Hofstede index, common language, landlocked, distance XTFEVD provides unbiased estimates, get consistent results between datasets OECD, World Bank and US data UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

17 Faculty of Economics and Business Administration
Conclusions The goal of this paper has been to employ the panel fixed effects with vector decomposition method developed by Plümper and Troeger (forthcoming) to examine the extent to which the omission of fixed effects gives misleading coefficients in FDI regressions. We find that the problems created by omission of fixed effects are quite significant and that they vary across the three most commonly used FDI datasets. As a result, when including fixed effects, we find much more consistency in the estimates of time-invariant variables across data sets. In general, these new estimates find additional evidence of the importance of trade costs. UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

18 Faculty of Economics and Business Administration
Conclusions Markusen (2002) finds horizontal FDI to be the highest when trade costs are moderate to high This implyes that lower trade cost will decrease horizontal FDI. When corrected for with conventional fixed effects and the XTFEVD procedure, trade costs are estimated to be lower. This is suggestive of vertical FDI. Nevertheless, the distance between capitals indicates horizontal FDI. UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

19 Faculty of Economics and Business Administration
Conclusions Since XTFEVD yields far more comparable results across our datasets, using fixed effects may be very useful in removing the problems this variation causes when industry level data are not available. In addition to our main results, we find that the cultural variables constructed by Hofstede (2001) are significant determinants of FDI. UNIVERSITY OF ICELAND Faculty of Economics and Business Administration

20 Faculty of Economics and Business Administration
UNIVERSITY OF ICELAND Faculty of Economics and Business Administration


Download ppt "Helga Kristjánsdóttir, 15. maí 2007"

Similar presentations


Ads by Google