©Risk Concerm. All rights reserved.

### In our previous report, we used the World Bank's 'GDP by unit of energy used' indicator (Link) to determine the greenest, most energy-efficient economies, and environmentally worst-performing economies. An important question that still needs further analysis: Does it pay to go green or be highly energy efficient? Does it benefit the Investors?

As the topmost energy-efficient economies are highly efficient, i.e., requiring less energy input, a fundamental requirement in productivity, to produce output; it wouldn't be illogical to assume that firms in these economies should have a higher level of profitability, compared to firms in economies that aren't highly energy efficient. Investing in these economies, thus, should yield higher returns compared to others, theoretically.

For example, if a cluster of firms in a green economy can produce productivity worth $100, by using $10 of energy, these firms should, in theory, be more profitable than a cluster of firms in a less energy-efficient economy that produces $100 worth of productivity, by using $50 of energy. Also, anecdotal evidence overwhelmingly supports a connection between going green and profitability.

### However, facts can only be uncovered through analysis of data, not by anecdotal evidence, which is overwhelming.

### Is this the case? Is investment in energy-efficient and green economies more profitable?

To scrutinize this assumption, we have tested the top economies from our report "What are the greenest economies" (see below). The benchmark used in the analysis is the S&P 500, which, of course, is US-based; however, the firms included in the S&P 500 also have significant international exposure. The GDP by unit of energy used for the US is 8.6, close to the overall global figure of 8.3 (GDP by unit of energy by World Bank).

The GDP by unit of energy for the top ten most energy-efficient economies in our list is 22, 156% higher than that of the US.

### Below an analysis of Hong Kong's Hang Seng index vs. S&P 500, Irelands ISEQ vs. S&P 500, and Switzerland's SMI vs. S&P 500 is presented:

hypothesis testing is employed to analyze the difference between the returns of the S&P 500 and the Indices tested in this analysis. The pooled variance test has been conducted to test the returns; The F test has been conducted to test the difference in the risk profiles.

Pooled variance test is conducted as:

H0: µ1 -µ2 = 0, vs H1: µ1 -µ2 ≠ 0

The mean returns of Index 1 minus Index 2 (comparison) are analyzed for equivalence.

F test has been conducted as:

H0: σ1 = σ2, vs H1: σ1 ≠ σ2

The variance of the returns of Index 1 is compared to the variance of the returns of Index 2 (comparison) to assess equivalence.

### Hang Seng vs. S&P 500 (374 months, or more than three decades of data analyzed)

Pooled variance test:

Two sample t-test (pooled variance), using T distribution (DF=748.0000) (two-tailed) (validation)

1. H0 hypothesis Since p-value > α, H0 is accepted. The average of Hang Seng's population is considered to be equal to the average. of the S&P 500's population. In other words, the difference between the average of the Hang Seng and S&P 500 populations is not big enough to be statistically significant. 2. P-value p-value equals 0.774412, ( p(x≤T) = 0.612794 ). This means that if we would reject H0, the chance of type I error (rejecting a correct H0) would be too high: 0.7744 (77.44%). The larger the p-value the more it supports H0. 3. The statistics The test statistic T equals 0.286712, is in the 95% critical value accepted range: [-1.9631 : 1.9631]. x1-x2=0.0012, is in the 95% accepted range: [-0.008200 : 0.001288]. The statistic S' equals 0.00420 4. Effect size The observed standardized effect size is small (0.021). That indicates that the magnitude of the difference between the average and average is small.

The null is not rejected, therefore.

### What does this mean, simplistically? Hang Seng is not more profitable than S&P 500?

There is no evidence with any statistical significance backing the assumption that Hang Seng's returns are different from the returns of the S&P 500. The returns are similar, and there is statistical evidence supporting this conclusion.

### What about the risk? Is the risk of Hang Seng and S&P 500 the same?

F test has been conducted for definitive answers:

F test for variances, using F distribution (dfnum=374,dfdenom=374) (two-tailed) (validation)

1. H0 hypothesis Since p-value < α, H0 is rejected. The sample standard deviation (S) of Hang Seng's population is considered to be not equal to the sample standard deviation (S) of S&P 500's population. In other words, the difference between the sample standard deviation (S) of the Hang Seng and S&P 500 populations is big enough to be statistically significant. 2. P-value The p-value equals 0, ( p(x≤F) = 1 ). It means that the chance of type I error (rejecting a correct H0) is small: 0 (0%). The smaller the p-value the more it supports H1. 3. The statistics The test statistic F equals 2.7081, which is not in the 98% region of acceptance: [0.7858 : 1.2726]. S1/S2=1.65, is not in the 98% region of acceptance: [0.8865 : 1.1281]. The 98% confidence interval of σ12/σ22 is: [2.1281 , 3.4462].

F test for variances, using F distribution (dfnum=374,dfdenom=374) (right-tailed) (validation)

1. H0 hypothesis Since p-value < α, H0 is rejected. The sample standard deviation (S) of Hang Seng's population is considered to be greater than the sample standard deviation (S) of S&P 500's population. 2. P-value The p-value equals 0, ( p(x≤F) = 1 ). It means that the chance of type I error (rejecting a correct H0) is small: 0 (0%). The smaller the p-value the more it supports H1. 3. The statistics The test statistic F equals 2.7081, which is not in the 98% region of acceptance: [-∞ : 1.2371]. S1/S2=1.65, is not in the 98% region of acceptance: [-∞ : 1.1122]. The 98% confidence interval of σ12/σ22 is: [2.1892 , Infinity].

### What does this mean, simplistically? Does Hang Seng have the same risk as S&P 500?

There is evidence with statistical significance backing the assumption that Hang Seng's risk is different from the risk of the S&P 500. Hang Seng is riskier than S&P 500, and the risk profiles are not similar; there is statistical evidence backing this conclusion.

### Ireland Stock Market (ISEQ) vs. S&P 500 (210 months of data analyzed for ISEQ, 374 months of Data analyzed for S&P 500)

Pooled variance test:

Two sample t-test (pooled variance), using T distribution (DF=583.0000) (two-tailed) (validation)

1. H0 hypothesis Since p-value > α, H0 is accepted. The average of ISEQ's population is considered to be equal to the average. of the S&P 500's population. In other words, the difference between the average of the ISEQ and S&P 500 populations is not big enough to be statistically significant. 2. P-value p-value equals 0.453096, ( p(x≤T) = 0.226548 ). This means that if we would reject H0, the chance of type I error (rejecting a correct H0) would be too high: 0.4531 (45.31%). The larger the p-value the more it supports H0. 3. The statistics The test statistic T equals -0.750767, is in the 95% critical value accepted range: [-1.9640 : 1.9640]. x1-x2=-0.0031, is in the 95% accepted range: [-0.008100 : 0.001654]. The statistic S' equals 0.00410 4. Effect size The observed standardized effect size is small (0.065). That indicates that the magnitude of the difference between the average and average is small.

The null is not rejected, therefore.

### What does this mean, simplistically? ISEQ yields more than S&P 500?

There is no evidence with any statistical significance backing the assumption that the returns of ISEQ are different from the returns of the S&P 500. The returns are similar, and there is statistical evidence supporting this conclusion.

### What about the risk? Is the risk of ISEQ and S&P 500 the same?

F test has been conducted for definitive answers:

F test for variances, using F distribution (dfnum=209,dfdenom=374) (two-tailed) (validation)

1. H0 hypothesis Since p-value < α, H0 is rejected. The sample standard deviation (S) of ISEQ's population is considered to be not equal to the sample standard deviation (S) of S&P 500's population. In other words, the difference between the sample standard deviation (S) of the ISEQ and S&P 500 populations is big enough to be statistically significant. 2. P-value The p-value equals 0.000002943, ( p(x≤F) = 1 ). It means that the chance of type I error (rejecting a correct H0) is small: 0.000002943 (0.00029%). The smaller the p-value the more it supports H1. 3. The statistics The test statistic F equals 1.7482, which is not in the 98% region of acceptance: [0.7482 : 1.3227]. S1/S2=1.32, is not in the 98% region of acceptance: [0.865 : 1.1501]. The 98% confidence interval of σ12/σ22 is: [1.3217 , 2.3367].

F test for variances, using F distribution (dfnum=209,dfdenom=374) (right-tailed) (validation)

1. H0 hypothesis Since p-value < α, H0 is rejected. The sample standard deviation (S) of ISEQ's population is considered to be greater than the sample standard deviation (S) of S&P 500's population. 2. P-value The p-value equals 0.000001472, ( p(x≤F) = 1 ). It means that the chance of type I error (rejecting a correct H0) is small: 0.000001472 (0.00015%). The smaller the p-value the more it supports H1. 3. The statistics The test statistic F equals 1.7482, which is not in the 98% region of acceptance: [-∞ : 1.2802]. S1/S2=1.32, is not in the 98% region of acceptance: [-∞ : 1.1315]. The 98% confidence interval of σ12/σ22 is: [1.3655 , Infinity].

### What does this mean, simplistically? Is ISEQ riskier than S&P 500?

The is evidence with statistical significance backing the assumption that the risk of ISEQ is different from the risk of the S&P 500. ISEQ is a riskier investment compared S&P 500, and the risk profiles are not similar; there is statistical evidence backing this conclusion.

### Swiss Market Index (SMI) vs. S&P 500 (374 months, or more than three decades of data analyzed)

Pooled variance test:

Two sample t-test (pooled variance), using T distribution (DF=748.0000) (two-tailed) (validation)

1. H0 hypothesis Since p-value > α, H0 is accepted. The average of SMI's population is considered to be equal to the average. of the S&P 500's population. In other words, the difference between the average of the SMI and S&P 500 populations is not big enough to be statistically significant. 2. P-value p-value equals 0.646256, ( p(x≤T) = 0.323128 ). This means that if we would reject H0, the chance of type I error (rejecting a correct H0) would be too high: 0.6463 (64.63%). The larger the p-value the more it supports H0. 3. The statistics The test statistic T equals -0.459155, is in the 95% critical value accepted range: [-1.9631 : 1.9631]. x1-x2=-0.0014, is in the 95% accepted range: [-0.006200 : 0.001288]. The statistic S' equals 0.00315 4. Effect size The observed standardized effect size is small (0.034). That indicates that the magnitude of the difference between the average and average is small.

The null is not rejected, therefore.

### What does this mean, simplistically? SMI yields a higher return than S&P 500?

There is no evidence with any statistical significance backing the assumption that the returns of SMI are different from the returns of the S&P 500. The returns are similar, and there is no statistical evidence to support that the returns of SMI and S&P 500 are different.

### What about the risk? Is the risk of SMI and S&P 500 the same?

F test has been conducted for definitive answers:

F test for variances, using F distribution (dfnum=374,dfdenom=374) (two-tailed) (validation)

1. H0 hypothesis Since p-value > α, H0 is accepted. The sample standard deviation (S) of SMI's population is considered to be equal to the sample standard deviation (S) of S&P 500's population. In other words, the difference between the sample standard deviation (S) of the SMI and S&P 500 populations is not big enough to be statistically significant. 2. P-value The p-value equals 0.4103, ( p(x≤F) = 0.7948 ). It means that the chance of type I error, rejecting a correct H0, is too high: 0.4103 (41.03%). The larger the p-value the more it supports H0. 3. The statistics The test statistic F equals 1.089, which is in the 98% region of acceptance: [0.7858 : 1.2726]. S1/S2=1.04, is in the 98% region of acceptance: [0.8865 : 1.1281]. The 98% confidence interval of σ12/σ22 is: [0.8557 , 1.3858].

The null is not rejected, therefore.

### What does this mean? SMI is riskier than S&P 500?

There is no evidence with any statistical significance backing the assumption that the variance of SMI is different from the variance of S&P 500. The two have similar variance, and there is no statistical evidence to support that the returns of SMI and S&P 500 have a different level of risk.

### What are the fundamental takeaways from this analysis? Being energy efficient and green does not mean higher profits?

We can conclude with significant confidence by analyzing over 100 years of data altogether that being energy efficient and green does not translate to higher profits for investors. Overall market indices of top green and energy-efficient economies reveal that the returns of markets in these economies are equivalent to the returns of S&P 500.

However, the risk profiles are a different story; the risk of 2 out of the three most energy-efficient economies is greater than that of the S&P 500. Similar returns but higher risk . . . presently, therefore, going green and being highly energy efficient is not correlated with higher profitability for investors, unfortunately.

Another important point here is the growth rate in the energy efficiency of the topmost energy-efficient economies. The growth rate of energy efficiency of the top energy-efficient economies has been significantly higher than the global average; nonetheless, the returns have not differentiated from the benchmark of the S&P 500.

Higher risk (variance) but similar returns cannot be considered the 'cost of going green,' with an added supposition that this cost, in the long run, would materialize in higher profits.

This is because if the risk was attributed to the cost of going green, this amount should have circled back into the economies and the industries in the green economies, resulting in higher profitability of the firms in these economies and, overall, higher returns in the indices in the green economies; this, however, is not observed.

The cost of going green, very likely, seems external, i.e., this cost appears to be an expense that firms in the green economies incur to international firms, and thus, markets in the green economies have not yielded a higher return.

It may be argued that the returns of the firms in green economies are not higher because labor in energy-efficient economies commands higher wages. For example, going back to the analogy used in the first part of this report: "if a cluster of firms in a green economy can produce productivity worth $100, by using $10 of energy, these firms should, in theory, be more profitable than a cluster of firms in a less energy-efficient economy that produces $100 worth of productivity, by using $50 of energy."

This condition, it may be argued, is explicable as labor in the green economies perhaps earns more, which offsets the advantages of being green, in terms of profitability. The assumption here being that individuals are better off in green economies.

This assumption is not backed by data, however. Highly energy efficient economies have not facilitated their labor sector with higher wages. Furthermore, it may be argued that the wages are not higher compared to, say the US, but this is an average figure, and the workers earn a higher median salary in energy-efficient economies due to lower inequality. This assumption was tested, and it is false; there is no statistical evidence backing the claim that inequality, as measured by the GINI index, is lower in energy-efficient economies:

Applying a prism of risk parity theory, we can state that in the long-run, the returns of the indices of highly energy-efficient economies should: 1. Increase at a slightly higher rate than the S&P 500, or 2. Their risk (variance) should reduce in magnitude.

In conclusion, investing in international indices based on their energy efficiency or green credentials has not resulted in higher returns; thus, recommending these indices is not an option currently.

It is worth mentioning here that as elaborated in our report titled "What are the greenest economies" (see below), "several market-moving financial institutions, such as Norway's sovereign wealth fund, have committed to investments that strictly adhere to ESG best practices." As more money chases green opportunities, we might see an appreciation in asset prices in green economies.

Presently, significant growth, as per the economic fundamentals of growth in aggregate labor hours, labor productivity, and total factor productivity (TFP) (neoclassical model), does not present a strong case for allocating capital to the most energy-efficient countries. For the mid-term horizon, 3-5 years, the S&P 500 or the Wilshire 5000, comparatively, appear more accessible and more manageable.

### Growth Data:

The growth rate (geometric mean) in Ireland's GDP per unit of energy used (2005-2015) stood at 4.5%;

The growth rate (geometric mean) in Hong Kong's GDP per unit of energy (2005-2015) stood at 2.7%;

The growth rate (geometric mean) in Switzerland's GDP per unit of energy (2005-2015) stood at 2.5%;

The growth rate (geometric mean) in United states' GDP per unit of energy (2005-2015) stood at 2.2%;

The growth rate (geometric mean) in Global GDP per unit of energy (2004-2014) stood at 1.5%.

Analysis Data sheet:

Keywords: Green economies, green investment, esg investment, clean green investments, clean green etfs, clean green investments, hang seng vs s&P 500, ISEQ vs s&P 500, SMI vs S&P 500, investing in green economies, are energy-effiicient companies more profitable, which economies are greem, are green investments good, investing in green economies, green investments more profitable, gdp by unit of energy, investing in hong kong better than us, investing in ireland better than us, investing in switzerland better than us.