What Really Drives GDP Growth (And Why Stimulus Doesn’t Guarantee It)
By: Jivraj Karwa
“Stimulus” and “steroids” are loaded words, but they have more in common than it seems at first blush.
A catch-all word for a lot of different, more nuanced interventions? Check.
Performance enhanced when using? Check.
Potential side effects? Double check.
But perhaps the biggest similarity is that both words assume growth is a guaranteed result. But what you’re using matters.
In this case, not all stimulus is created equal.
Generally, modern economics looks at two broad policy levers for driving economic growth – fiscal stimulus (otherwise known as government spending), and monetary stimulus (central bank asset growth). There are myriad details and variations to each – how the money is spent, what central banks should buy, etc. – but the core components remain.
But which of those two is the better driver or indicator of growth?
This question is not explored in a vacuum; existing literature already supports some general conclusions to guide our focus. For instance, a 2017 article by the Federal Reserve Bank of St. Louis posits that government spending is not an effective way to boost GDP growth rate.[a] Many government programs fail to generate adequate rate of returns; though that’s not the only marker of the success or failure of government programs, policy makers should understand that funding programs with moneys more productively utilized by the private sector will result in less economic growth.[b] Policymakers should determine whether spending for a given program yields enough benefits to offset the corresponding loss of money to the private sector. Additionally, the implications of central bank balance sheet development cannot be neglected. That is, policy makers should continuously investigate the impact of changes in central bank balance sheets on financial markets and the real economy.[c]
So what is the first impression from the “real economy?” In the years leading up to the 2008 financial crisis, both fiscal and monetary stimulus were relatively modest. To create an accurate picture of growth patterns in a global economy, we needed a reliable cross-section of the economy. With five larger economic bodies – US, Euro Region, England, China, Japan – and a number of smaller, developing economies as our dataset, we dug in.
Here’s what we found.
From 2000-2019, government spending as a share of GDP for developed economies grew overall by 2% (Figure 1). Although the nineteen-year growth was only two percentage points, there was a significant peak in government spending due to the 2008 financial crisis, then a steady decline to pre-crisis levels, with an upswing once again underway at the time of this writing. Developing economies in the same period saw government spending growth of 11% from 22% to 33% (Figure 2). Developing economies also saw a sharp increase due to the financial crisis, but unlike the developed economies, there was no subsequent drop-off but a continual steady increase after the financial crisis.
This increased drastically during the financial crisis when governments and central banks increase spend and balance sheets during the crisis period. Figure 3 also depicts an overall growth of >250% over a six year period, with a CAGR of 22% from 2006-2011. For the same economies during the total nineteen year period (2000-2019), aggregate nominal GDP has grown from $30T to $68T at a CAGR of 4%. 
Fig. 1 – Share of Govt. Spend in GDP (Developed economies)
Fig. 3 – Total Central Bank Assets -Global ($Tn)
Fig. 2 – Share of Govt. Spend in GDP (Developing economies)
Fig. 4 – Nominal GDP – Global ($Tn)
The relationship between the share of government spend in GDP, growth in the assets of the central bank of the country, and growth in non-Government GDP with future growth in non-Government GDP is unclear. But these figures combined depict two key points – growth in government spend as a share of GDP and growth in central bank assets. For the latter, one could reasonably assume an even higher number after 2020’s data is finalized.
*Economies considered: US, England, Euro Region, Japan, China, India, Indonesia, Canada, Turkey, Mexico, Brazil, South Korea
Our objective in this analysis is to understand the impact of fiscal and monetary activities on future growth in non-Government GDP. Given the general trend among economies to favor monetary stimulus over fiscal stimulus, our cross-section of developed and developing economies lays the bedrock for statistical analysis. From there, we might see if we can infer which of our two main economic policy levers has a better track record of driving growth.
Through our analysis, we also hoped to gain answers to some key questions.
1. How well – if at all – does the non-government GDP growth rate of developing economies reflect the effect of fiscal activities vs monetary activities?
2. What is the main factor of influence for future growth of non-Government GDP?
3. What is the impact of non-Government GDP growth rate in developing vs developed economies?
4. In developed economies, which variable has the strongest impact on the future non-Government GDP growth rate?
RESEARCH AND ANALYSIS
Of several independent variables considered (Table 1), growth in non-Government GDP shows a strong correlation – both positive and negative – with three variables from prior year: share of Government spend in GDP, growth in central bank assets, and nominal GDP growth rate. The correlation did not hold going back two years, however, making prior year the favorable time frame for comparison (Figure 5).
Fig. 5 – Correlations of Independent Variables with Non-Government GDP Growth Rate
Running CARTree regression on the data, non-Government GDP growth rate changes with independent variables as shown in table 2. Finally, Figure 6 depicts the CARTree analysis in more detail, confirming our hypothesis from our correlation analysis. Central bank asset growth rate from prior year for developed economies are negatively correlated to this year’s nominal GDP growth rate. The initial node segregates according to prior year non-government GDP growth rate, with developed economies less than 8.3% and developing economies greater (operating on a general assumption that developing economies tend to have higher GDP growth rates). The approximate percentage split of developed to developing economies is 62% to 38%, respectively.
All this supports an intriguing observation: the maturity of developed economies reduces the perceived impact from either kind of stimulus.
Figure 6 – CARTree Regression Decison Tree for Developed and Developing Economies
Thanks to this analysis, we can confidently make a few assertions regarding our earlier guiding questions. First, non-Government GDP growth rate of developing economies does not reflect the effect of fiscal activities strongly, while there is a moderate impact of monetary activities. Second, future growth of non-Government GDP mainly depends on the growth stage at which the economy lies. If the current non-Government GDP growth rate is high, the subsequent year growth is also likely to be higher. Going from developing to developed economies, the impact of non-Government GDP growth rate lessens.
Crucially, we noted the future non-Government GDP of the developed economies declines with an increase in the country’s central bank assets and share of Government spending in the GDP. Of variables assessed, growth in a country’s central bank assets has the strongest impact on the future non-Government GDP growth rate, followed by share of government spending in the GDP. Meanwhile, as the economies move from developing to developed, share of Government spending in GDP increases.
With this analysis as a baseline, future work could be done to project growth in an economy based on its developmental stage and the specific activities of governments and central banks, as well as their particular economic stage. There are some additional provocative implications here concerning the ability of governments in developing economies to influence GDP – small details with a big impact on the global economic stage.
Data tells a story. Curious what insights data analytics can provide for your organization?
1. IMF Cross Country Macroeconomic Statistics
2. IMF Cross Country Macroeconomic Statistics
3. [US] Federal Reserve Economic Data (FRED: https://fred.stlouisfed.org/series/WALCL)
[EURO] Federal Reserve Economic Data (https://fred.stlouisfed.org/series/ECBASSETSW)
[JAPAN] Federal Reserve Economic Data (https://fred.stlouisfed.org/series/JPNASSETS)
[ENGLAND] Bank of England annual reports (2000 – 2019)
[CHINA] People’s Bank of China annual reports (2000 – 2019)
[INDIA] Reserve Bank of India annual reports (2000 – 2019)
[INDONESIA] Bank Indonesia annual reports (2003 – 2019)
[BRAZIL] Central Bank of Brazil (Banco Central do Brasil) annual reports (2000 – 2019)
[MEXICO] Banco de México annual reports (2000 – 2019)
[CANADA] Central Bank of Canada annual reports (2000 – 2019)
[SOUTH KOREA] Bank of Korea annual report (2000 – 2019)
[TURKEY] Central Bank of the Republic of Turkey annual reports (2000 – 2019)
4. World Development Indicators – Worldbank (GDP (constant LCU))
Will COVID Finally Give Big Banks Their...
What India's Infant Mortality Rate Tells Us...
The Seesawing Journey of Fintechs During...