Quantifying Renewable Energy Viability Via Leading Composite Index
As we hurtle into the future, the recurring emphasis on renewable energy sources grows stronger and stronger, while reality fails to reflect this urgency in its implementation.
Repeatedly, the issue in both the public eye and from the policymaker’s mouth is perceived cost; the incentive is just not there for firms to implement renewable energy on a large scale, at least in the manner that was forecast at the beginning of the millennium.
This highlights the problem being addressed– the viability of renewable energy sources, and how a relationship between these sources and the economy is reflected in boom-and-bust cycles. The issue at hand is that renewable energy has had a difficult time usurping the status quo of fossil fuels, especially on a nation-wide scale.
In its Energy Sector Rating, Schwab analysts bring this issue to light by stating “...as the energy transformation advances, many companies in the [renewables] sector haven't ramped up capital expenditures and production along with rising oil prices, as they have done historically.” If we are able to define some relationship between the economy and these energy sources, it may provide insight on when the best time for investment and implementation may be, or could also detail the symbiotic relationship between renewable energy spending and the health of the economy.
There has been a balancing act in the energy industry, with renewables gaining ground at a pace slower than many expected around 20 years ago. To have an index that assesses different components to derive a quantitative figure that charts renewable viability would prove to be invaluable in this fight. The International Renewable Energy Agency summarizes the problem succinctly– “Renewable energy has entered a virtuous cycle of falling costs, increasing deployment and accelerated technological progress. [Photo-voltaic] module prices have fallen 90% since the end of 2009, while wind turbine prices have fallen by 55-60% since 2010.
The public debate around renewable energy, however, continues to suffer from the dated perception that renewable energy is not competitive, forming a significant and unnecessary barrier to its deployment.” The trade-off between renewables (defined here as hydro, solar, wind, and nuclear) and non-renewables (defined as petroleum derivatives, natural gas, and coal) undergirds a lot of dissonance around the discussion of energy; as each commodity experiences it’s own market fluctuations, it makes the decision to jump from carbon-based to renewable is a difficult one.
The Total Renewable Viability Index (TRVI) helps gauge the current state of the renewable energy industry, specifically focusing on nuclear, hydro-electric, solar, and wind-generated power. This index consists of a basket of relevant ETFs and stocks who’s financial performance is largely dictated by investment in and utilization of clean energy sources. The TRVI aims to quantify the strength of the renewable energy industry, and in-turn provide some insight into the current state of the economy. In Business Cycle Co-Movements Between Renewables Consumption And Industrial Production, a relationship between economic development and renewable energy utilization was identified, with the author stating how “renewable(s) consumption has positive significant impacts on industrial production, hence, on economic growth, in both lower and higher frequencies in the US.” Another academic paper, The Repercussions Of Business Cycles On Renewable & Non-Renewable Energy Consumption Structure, came to the following conclusion– “When business cycles change (expansion or recession), fluctuation in the economy affects the production/consumption pattern, which ultimately influences the overall demand for renewables and non-renewables.
Besides, both sources have a different effect on the environment… accordingly, the findings showed that the economy reduces renewable and non-renewable energy demand during a recession.” The thesis underlying the TRVI is as follows: during periods of economic expansion (as identified by NBER chronology), renewables investment and correlated industry value should increase; in times of economic recession or contraction, that spending and investment would contract as well.
The goal here is to create an index that details the trade off of a basket of renewables versus a basket of non-renewables on a cost basis, and to overlay this index with business cycle chronology (using NBER methodology) in order to build insights on when the best possible time would be to transition from one type of energy source to another. The TRVI is a leading index; the aim was to create something that reflects economic peaks and troughs before they’re properly defined, and financial data relating to the energy industry can be a good gauge of future economic performance.
The Energy Information Administration emphasizes this link between electricity consumption and economic performance– in their 2014 publication Electricity Use as an Indicator of U.S. Economic Activity, the relationship is summarized as follows: “We argue for the resurrection of an old idea: electricity use as an indicator of U.S. economic activity. Our analysis relies on associations–the 40-year correlation between growth rates in real GDP and electricity use can be as high as 89% –and intuition. Electricity use and economic conditions should move together. The vast majority of goods and services are still produced using electricity; services may require less electricity, but they still require some.”
Full disclosure– while nuclear energy is not technically a renewable energy source, it is widely considered “green” energy and is currently the most viable replacement for fossil fuels at scale. According to the World Nuclear Association, nuclear plants produce no greenhouse gas emissions during operation, and over the course of its life-cycle, nuclear produces about the same amount of carbon dioxide-equivalent emissions per unit of electricity as wind, and one-third of the emissions per unit of electricity when compared with solar.
To go more in depth on the exact data sources, the aim was to split the data into two categories: commodity-specific and general renewable. For commodity specific, I chose the largest market-cap ETFs for each sub commodity (sans hydroelectric). This includes the following: Invesco Solar ETF, URA Uranium ETF, and First Trust Global Wind Energy ETF. Since a lot of hydroelectric power is publicly funded or comes from legacy systems, getting specific data can be challenging (as no ETFs or ETNs exist currently). To substitute, I will be looking at Duke Energy (who operates 31 hydroelectric plants in the Carolinas) and Edison (who’s southern California subsidiary is a major hydroelectric player– Southern California Edison's peak system demand for hydropower is more than 22,000 megawatts, which comes from 36 powerhouses and 79 generating units). Conversely, general data captures sources that are more broad-spectrum, namely the iShares Global Clean Energy ETF and First Trust NASDAQ Clean Edge Green Energy Index Fund.
In constructing this index, a number of transformations occurred on each component. After experimenting with different data modifications, the final index consisted of components being weighted to reflect both their importance in the world of renewable energy and on the state of the economy as a whole (using the leading economic indicator and S&P 500 index to create a standard “image” of the real economy).
Commodity-specific ETFs (Invescon Solar ETF, First Trust Wind ETF, and Uranium ETF) were each weighted at 0.0833, since their scope was limited to that of a single renewable component. Edison and Duke, the two energy firms included in this index, were both weighted equally at 0.1667 (as a note, neither firm is included in any of the subsequent ETFs), and lastly the iShares Global Clean Energy ETF and First Trust NASDAQ Clean Edge Green Energy Index Fund were both weighted the heaviest at 0.208. Both of the latter ETFs were emphasized due to their scope and inclusion of a basket of renewable-based commodities and firms (instead of focusing on a single commodity). The weighing in this sense is similar to how standardization is done in creating the Leading Economic Indicator, which reflects how strongly a particular component is represented in the overall index.
Another data transformation was the creation of a seasonal average to backfill the uranium ETF component (URA). URA was first traded in December of 2010, however, the data for all other components goes back into 2008 or earlier. In order to include the 2008 recession in the data, a monthly average of the adjusted closing price was taken for the first three years of the uranium ETF. This was then used to backfill the missing data from July of 2008 through December of 2010–not so much to “create” new data, but to smooth the index trend and to create a less-jarring inclusion of URA starting in late 2010.
Each component’s data is made up of their adjusted closing price from July of 2008 through April of 2022 (sans URA). The adjusted closing price was utilized over standard closing price because it’s adjusted for splits and dividend and/or capital gain distributions. This took a lot of legwork out of normalizing stock data. For each component, a monthly percent-change figure was calculated using the symmetric percent change formula, or 200*(current-previous)/(current+previous).
These figures were then standardized (weighted) with each subsequent quotient as mentioned earlier. From there, the sum of all standardized component figures was symmetrically adjusted using the formula 100*(200+current)/(200-current). Note how the calculations are similar to what’s utilized in calculating Lead Economic Indicator– this was done intentionally to create a fair comparison across the two indices. No further adjustments were made to the data as a whole. A time series of the now-complete Total Renewable Viability Index gives the following:
A quick-glance provides the following insights: as the 2008 recession continued, the TRVI precipitously declined, and bottoms out as NBER chronology defined the recession as 5 “over”. The TRVI experiences some fluctuation throughout 2009 - 2012, bottoming out at it’s all-time low of 39.52 in November of 2012. From this point through late 2019, the TRVI experiences a fluctuating-yet-onward expansion, with a pre-pandemic high of 74.12 in January of 2020.
However, beginning in February of 2020 (and in-line with the NBER-defined peak), the TRVI experiences a recession (assumedly) due to some relationship with the COVID-19 pandemic and subsequent mini-recession. From that point onward, the TRVI skyrockets to a new all-time high, coinciding with the economy-wide expansion during this time. McKinsey assessed this period as follows– “the trends driving the economy today shows a pattern of great acceleration: many of those present before the pandemic have experienced dramatic surges.” Again, from Schwab’s energy sector rating, they pinpoint that “the energy sector has outperformed the overall market since the COVID-19-crisis-related market lows in March 2020.” While on a simple visual analysis, the TRVI aligns with overall economic trends. Below, the TRVI is represented with the LEI and NBER-defined recessions:
Measuring both indices on the same axis, it’s obvious that the TRVI experiences more emphasized expansions and contractions than the LEI. With that, there's something important to notice: the TRVI hits it’s 2008-recession trough two months before the LEI does (and subsequently ends the recession).
This pattern is also reflected in the COVID-19 recession, with the TRVI recovering back to it’s pre-pandemic levels one month before the LEI hits its trough (and again subsequently ends the pandemic). Although the time series may be brief in scope, 6 these findings show that the TRVI is in fact a leading economic indicator, as it twice predicts future changes in the economic cycle. Below, the TRVI and LEI can be assessed with figures from the S&P 500 in the same time period:
Here, the close alignment between the TRVI and S&P is obvious– with the S&P almost perfectly reflecting the same degree of change during the 2020 recession as the TRVI, while also reflecting the soaring heights the index came to in the recession’s recovery. Notice the intricacies of each of the three indices, especially their relationship during business cycle troughs.
With the understanding that this was more of a niche index, I was surprised to see a number of scholarly sources cite the relationship between energy source and business cycle chronology, as well as other relevant intricacies. From Erciyes University in Turkey, Dr. Faik Bilgili wrote Business Cycle Co-Movements Between Renewables Consumption And Industrial Production to assess the relationship between renewable energy consumption and productivity in the United States, with Dr. Bilgili yielding that “renewables consumption has positive significant impacts on industrial production, hence, on economic growth, in both lower and higher frequencies in the US.” Another interesting source came from a cohort out of Panzhihua University with the publication of The Repercussions Of Business Cycles On Renewable & Non-Renewable Energy Consumption Structure. This paper centers around “the choice” a nation makes between renewable and non-renewable energies during different business cycle phases, and how this can affect economic growth and energy demand respectively.
It was refreshing to see a lot of the business cycle logic we’ve covered in this course be reflected in a publication that balances the discussion on energy sources and their different economic effects; here, Panzhihua University supplied the following outline that reflects the relationship between energy sources, investment, and R&D, all as determining factors in business cycle trajectory. The paper came to a conclusion that reflects the thesis of the TRVI– “When business cycles change (expansion or recession), fluctuation in the economy affects the production/consumption pattern… the economy reduces renewable and non-renewable energy demand during a recession.” These findings are clearly reflected in the above graphs, showing a precipitous decline in TRVI during times of economic contraction, and massive gains during periods of economic expansion. My personal theory is that spending on renewable energy is seen as expendable by policymakers due to it’s redundancy with legacy fossil fuel systems. Due to it’s perceived expendability, during times of economic hardship, spending is abandoned at a higher rate than what’s seen in other sectors of the economy; during times of economic expansion however, opportunists are more willing to partake in this budding industry and the heights in which the TRVI hits are again exacerbated more than other economic sectors.
Two other sources I found meditated on the same idea, being An Investigation Of Renewable And Non-Renewable Energy Consumption And Economic Growth Nexus Using Industrial And Residential Energy Consumption, which was born out of a cross-collaboration from CUNY Brooklyn and the RMIT University, and From Volatility Spillover To Risk Spread: An Empirical Study Focuses On Renewable Energy Markets from the Yunnan University of Finance and Economics. The latter source focused on the economic risk associated with disparate energy markets, or specifically, the “volatility spillover in the main renewable energy markets from 2010 to 2019 and effective management to reduce financial system risks.” This gives some insight into a time period that the TRVI seems the least reliable, the period of economic expansion between the 2008 and 2020 recessions. If there is indeed volatility spillover amongst the renewable energy industry, this could explain the rickety-upward trend that was seen in the TRVI during the 2010s, as this was a time period where solar panels greatly fluctuated in cost (the cost of solar modules fell from $2 per watt to $0.20/W during the 2010s), and nuclear energy faced a number of regulator setbacks (from the beginning of the TRVI time series through the end of the 2010s, zero contracts were signed for new nuclear power plants in the United States).
Lastly, to round out the peer-reviewed sources, I came across Valuation For Renewable Energy: A Comparative Review, which reviews the valuation/evaluation of renewable energy resources and summarizes relevant valuation methods, as well as discerns the four main streams of valuation in renewable energy. While not all of these sources pinpoint the same thesis, I believe that an index with such a wide breadth requires diverse sources of reliable input in order to be truly well-rounded. There are inevitable limitations of this model, due in part to the nature of indices as a whole as well as the components that make it up– as mentioned earlier, there was no clear data source for energy sources such as hydroelectric or biomass, whose absence detracts from the “total renewable” aspect of this index. Also, there was some dissonance in the available time periods for each component, mainly with the uranium ETF; while the monthly averages assisted in smoothing the trend, it does not reflect reality in the way other components did during the months of July of 2008 through November of 2010. Although it is a leading index, the Total Renewable Viability Index is not able to predict the future, nor future events that could impact the energy market. A big part of this is that the energy sector is largely a victim of externality, and events like war, natural disaster, and geopolitical tension can wreak havoc on energy markets worldwide, throwing a wrench in any empirical prediction.
With that said, there are a number of industry-specific factors that can make one energy source less viable than others, and this too greatly affects the outcome of this index since it is dependent on how the energy source is categorized. Say, for instance, that in the next two months, hydrogen power is shown to be greatly detrimental to aquatic life and is not as effective at electricity generation as previously thought. For starters, it’s back to the drawing board with our nation’s dams, but also, this would greatly hinder the progress made by renewable energy. If hydrogen’s initial claims were incorrect, then who can we trust? The public would demand the whole renewable system to be audited before any other major project, hydro or not. Now, lets say it’s found that natural gas extraction is massively more harmful than initially claimed, and is responsible for a number of cancer clusters around the US (similar to later findings on fracking and shale extraction). Undoubtedly there would be calls for carbon-based sources to go, as we’ve seen, but in the short-term, natural gas would simply be substituted for another legacy source. These two cause-and-effect scenarios would have drastic effects on the index as a whole, and for different reasons— it’s continually an uphill battle for green energy sources to usurp the legacy carbon-based sources.
To improve the accuracy of the Total Renewable Viability Index, one aim is to make it more authoritative on renewable energy as a whole. This would mean including more all-renewable ETFs (such as the NASDAQ Clean Energy ETF [ICLN]), but also have each sub-commodity be represented by it’s own set of data in a manner that’s translatable across all sub-commodities; in this case, that would mean having separate ETFs (or equivalents) for things like hydroelectric power, biomass, geothermal energy, or fuel cells. Also, a big issue with the TRVI as it’s presented is the limited scope of it’s data– it was challenging to find anything prior to 2008, but in order to truly backtest the accuracy of the index, there would need to be data going back through previous economic contractions beyond 2008 and 2020.
To summarize, in creating this index there were key learning moments in both cleaning the data as a whole and in the selection of it’s components. Technically, this is the TRVI version 2, as the previous version utilized components that consisted of commodity-specific production data, installation metrics, and a random assortment of renewable energy-related data sources (such as National rooftop Potential, total hydroelectric load, etc). The goal was also to include the inverse of some sort of fossil fuel data, to create an index that measured the industrial competitiveness of renewable energy sources. Obviously this has a loose link to business cycles (if at all), and the random assortment of components created an index that really had no strong basis in reality. With some input, I pivoted away from this trajectory and focused solely on stock data, mainly stock composites or ETFs. These new components gave more clear insight into the relationship between business cycles and renewable energy sources, their financial performance, and overall viability. Pivoting away from my initial approach was a big learning moment in being able to create this linkage.
