August 12, 2013

Doug Hadden### FreeBalance Clients on the Road to Success, Part 8

*Carlos Lipari, FreeBalance Advisory Services*

This post covers the interpretation of the data analysis on FreeBalance customers and improved government effectiveness.

In order to analyze the statistical association between governance indicators on economic growth we considered several variations of a multiple regression model. 8 scenarios were considered. A benchmark scenario (1) and 7 additional specifications. In (1) we include every variable (all 6 governance indicators and 6 control variables). In (2), (3), (4), (5), (6), (7) and (8) certain control variables are removed. This is done in order to increase robustness of results.

a) Our dependent variable (Y) is the 2010 economic growth rate and is presented as a level variable.

b) Governance Indicators (all 6) are presented as log variables. Therefore we have a level-log functional form. Δy=(β1/100)% Δx

c) Control Variables are presented as level variables. In this case we have a level-level functional form where Δy= β1 Δx

Among the dependent variables analyzed in our *baseline scenario*, only effectiveness, HDI and Investment are statistically significant at standard levels (1%, 5% or 10%). In *alternative scenarios* Voice & Accountability also becomes significant but only at 10% level (it will not be considered).

**Government Effectiveness**: the results state that a 1% increase in government effectiveness is, everything else constant, on average, associated with a 0.14 percentage point increase in the economic growth rate.

**Human Development Index (HDI)**: results suggest that a 1 percentage point in the HDI index is, on average and everything else constant, associated with a 0.088 percentage point decrease in the economic growth rate. Therefore, a country that has a 0.9 HD, which is a very high HDI, is expected to, on average, have 1.76 percentage points less of economic growth rate than a country with a relatively high HDI of just 0.7. In simple terms, less developed countries are, ceteris paribus, expected to grow faster than more developed countries. Barro (1991) found evidence that it is, on average, easier for countries with lower income per capita to grow at faster rates than higher income per capita countries. This, together with a high correlation of income per capita with life expectancy and levels of schooling might explain much of this tendency of less developed countries growing faster than highly developed.

**Investment**: one percentage point of GDP increase in Investment is, on average and everything else constant, associated to an increase of 0.105 percentage point of the economic growth rate. Therefore, if we had two set of similar countries, one investing 30% of its GDP and another investing just 20%, if the first group registered an average growth of 4% a year, it is reasonable to expect the second group to grow about 3% a year. Without surprise, this result highlights the importance of investment on economic growth.

### Cross-Sectional Analysis Results and its limitations

It appears that only governance effectiveness, HDI and Investment are statistically significant at standard levels (1%, 5% or 10%). In alternative scenarios Voice & Accountability level becomes also significant at the 10% level. Even though only one governance indicator appears as statistically significant, its economic significance appears to be very large: a 1% increase in governance effectiveness being associated with a 0.14 percentage point increase of the economic growth rate.

A more complete study would be a Panel Data research (combining cross section and time series analysis) to include data of previous years. A panel data specification this would also allow us to better deal with omitted variable bias, since first differencing the model would eliminate individual specific non time varying unobservable, which may be correlated with the error term.

The fact that certain governance indicators are not statistically relevant in this research does not necessarily mean they do not impact economic growth. Not only we are focusing in just one year, the quality of the information currently provided by such governance indicators might also need some improvement.

Contents

- Introduction to the Study
- Government Revenue
- Government Expenditures
- Government Gross Debt
- Country Investment
- Real Investment Growth Rates
- Government Effectiveness
- Interpretation
- Conclusions

#### Doug Hadden

#### Latest posts by Doug Hadden (see all)

- What is the ‘Smart’ in Smart Government? - October 25, 2016
- The (IT) Project was a Success, but the Patient Died [Part 2] - September 21, 2016
- The (IT) Project was a Success, but the Patient Died [Part 1] - September 20, 2016
- Have we over-complicated the ‘smart’ in smart government? - September 8, 2016