A Comparative Analysis of GDP Growth: Manmohan Singh vs. Narendra Modi — Part 2 Relative to world.
In my previous analysis, I concentrated solely on India’s GDP growth during the tenures of Narendra Modi and Manmohan Singh. I illustrated that under Manmohan Singh’s UPA government, GDP growth slightly surpassed the anticipated levels (derived from 1000 simulations), whereas GDP growth during Narendra Modi’s NDA government fell short of expectations. However, a notable limitation of the study was its failure to evaluate India’s GDP growth in the context of global economic performance. This analysis aims to address this limitation.
In the realm of economic analysis, evaluating the performance of different governments is a complex task that goes beyond simply comparing GDP numbers. To provide a more comprehensive assessment of India’s economic performance during the tenures of Prime Ministers Manmohan Singh and Narendra Modi, we turn to relative GDP growth. This approach allows us to gauge India’s economic performance in comparison to other major economies, using GDP growth data and inflation data from the World Bank for India, Russia, China, Brazil, and the global economy.
GDP growth and Inflation Variation
The two graphs above clearly illustrate that India’s relatively restrained GDP growth during the UPA period, notably in its initial phase from 2004 to 2009, was accompanied by a substantial surge in consumer price inflation. Throughout much of the UPA’s term, inflation in India surpassed not only the global average but also exceeded the rates observed in Russia, China, and Brazil.
The Significance of Relative GDP Growth
When assessing the economic performance of governments, it’s essential to consider how a country’s GDP growth compares to the rest of the world. Comparing India’s GDP growth to that of other countries provides valuable insights into whether the country is outperforming or underperforming relative to its global peers.
To conduct this analysis, we calculated the relative GDP growth, which is the difference between India’s GDP growth and the GDP growth of each of the selected countries. This relative measure allows us to see how India’s economic growth stacks up against other major economies during the tenures of Manmohan Singh and Narendra Modi.
The Superiority of the Modi Government
Unlike the previous analysis, my initial examination of the data revealed that, on average, the Modi government outperformed the Singh government in terms of relative GDP growth. This suggests that India experienced stronger economic growth during Modi’s tenure when compared to other major economies, which was not as pronounced during Singh’s time in office.
Sampling Variability in GDP Growth
Nevertheless, it’s vital to recognize that drawing definitive conclusions from our analysis should be approached with caution. GDP growth rates can exhibit year-to-year fluctuations due to a multitude of factors, including economic policies, global economic conditions, and unforeseeable events. Additionally, the presence of sampling variability in GDP data further compounds this issue. This inherent variability stems from the fact that GDP data is frequently derived from a subset of economic entities (e.g., businesses, households), rather than encompassing the entire population. Consequently, this inherent variability within GDP growth data introduces an element of uncertainty into our findings.
To address this, we performed hypothesis testing on the mean of relative GDP growth for both regimes. Hypothesis testing helps us determine whether the observed differences are statistically significant or if they could have occurred due to random variation.
Hypothesis Testing: Bootstrap vs. 2-Sample t-Test
We utilized two statistical methods for hypothesis testing: the bootstrap hypothesis test and the 2-sample t-test. In both cases, the null hypothesis was that there is no significant difference in relative GDP growth between the Manmohan Singh and Narendra Modi governments. The alternate hypothesis was that relative GDP growth during Modi’s tenure was better.
Bootstrap Hypothesis Test:
The bootstrap test involves simulating multiple random samples from the available data and comparing the means of relative GDP growth for both regimes.
The bootstrap test for inflation-adjusted relative GDP growth data showed for all cases, the bootstrap test failed to reject the null hypothesis, implying that the observed differences in relative GDP growth might not be statistically significant in those instances.
Two sample t-test
The 2-sample t-test compares the means of relative GDP growth for the two regimes, assuming that the data follows a normal distribution.
Similar to the bootstrap test, when we compare the results to the t-test, we arrive at similar results where we fail to reject the null hypothesis.
Note: Even though the p-value for India-China inflation adjusted relative GDP growth between UPA and NDA falls below 0.05 (our significance level), it isn’t able to reject null hypothesis because of Bonferroni correction being applied.
Bonferroni Correction
Since we conducted multiple pairwise tests (one for each country), we employed the Bonferroni correction to mitigate the risk of making Type I errors (false positives). This correction adjusts the significance level to account for the multiple comparisons. This correction brings down the effective p-value threshold from 0.05 to 0.0125 (0.05/number of pairs =4).
Conclusion
In summary, while a surface-level comparison of GDP growth might seem to favor the Modi government, a more in-depth analysis using hypothesis testing on inflation-adjusted relative GDP growth reveals a more nuanced perspective. Both the bootstrap test and t-test failed to detect statistically significant differences in inflation-adjusted relative GDP growth across all countries, suggesting that the observed variations in these instances could potentially be attributed to random fluctuations.
The analysis suggests that India’s economic growth relative to China during Narendra Modi’s tenure, while notably better than during Manmohan Singh’s period as Prime Minister, did not reach a level of statistical significance. This pattern holds true for other cases as well.
Update: I had earlier made the error of subtracting inflation from real gdp growth data (which already accounted for inflation). Hence, this earlier led to a some fale statistically significant conclusions.
You can find the analysis notebook here