I keep revising my manuscript titled ‘Climbing the right hill – an evolutionary approach to the European market of electricity’, in order to resubmit it to the journal Applied Energy. In my last update, titled ‘Still some juice in facts’, I used the technique of reverted reading to break the manuscript down into a chain of ideas. Now, I start reviewing the most recent literature associated with those ideas. I start with Rosales-Asensio et al. (2020)[1], i.e. with ‘Decision-making tools for sustainable planning and conceptual framework for the energy–water–food nexus’. The paper comes within a broader stream of literature, which I already mentioned in the first version of my manuscript, namely within the so-called MUSIASEM framework, where energy management in national economies is viewed as metabolic function, and socio-economic systems in general are equated to metabolic structures. Energy, water, food, and land are considered in this paper as sectors in the economic system, i.e. as chains of markets with economic goods being exchanged. We know that energy, water and food are interconnected, and all the three are connected to the way that our human social structures work. Yet, in the study of those connections we have been going into growing complexity of theoretical models, hardly workable at all when applied to actual policies. Rosales-Asensio et al. propose a method to simplify theoretical models in order to make them functional in decision-making. Water, land, and food can be included into economic planning as soon as we explicitly treat them as valuable assets. Here, the approach by Rosales-Asensio et al. goes interestingly against the current of something that can be labelled as ‘popular environmentalism’. Whilst the latter treats those natural (or semi-natural, in the case of food base) resources as invaluable and therefore impossible to put a price tag on, Rosales-Asensio et al. argue that it is much more workable, policy-wise to do exactly the opposite, i.e. to give explicit prices and book values to those resources. The connection between energy, water, food, and the economy is being done as transformation of matrices, thus as something akin a Markov chain of states.
The next article I pass in review is that by Al-Tamimi and Al-Ghamdi (2020), titled ‘Multiscale integrated analysis of societal and ecosystem metabolism of Qatar’ (Energy Reports, 6, 521-527, https://doi.org/10.1016/j.egyr.2019.09.019 ). This paper presents interesting findings, namely that energy consumption in Quatar, between 2006 and 2015, grew at a faster rate than GDP within the same period, and energy consumption per capita and energy intensity grew approximately at the same rate. That could suggest some kind of trade-off between productivity and energy intensity of an economy. Interestingly, the fall of productivity was accompanied by increased economic activity of the Quatar’s population, i.e. the growth of professionally active population, and thence of the labour market, was faster than the overall demographic growth.
In still another paper, titled ‘The energy metabolism of countries: Energy efficiency and use in the period that followed the global financial crisis’. Energy Policy, 139, 111304. https://doi.org/10.1016/j.enpol.2020.111304 (2020), professor Valeria Andreoni develops a line of research, where rapid economic change, even when it is a crisis-like change, contributes to reducing energy intensity of national economies. Still, some kind of blueprint for energy-efficient technological change needs to be in place, at the level of national policies. Energy-efficient technological change might be easier than we think, and yet, apparently, it needs some sort of accompanying economic change as its trigger. Energy efficiency seems to be correlated with competitive technological development in national economies. Financial constraints can hamper those positive changes. Cross-sectional (i.e. inter-country) gaps in energy efficiency are essentially bad for sustainable development. Public policies should aim at equalizing those gaps, by integrating the market of energy within EU.
Velasco-Fernández, R., Pérez-Sánchez, L., Chen, L., & Giampietro, M. (2020), in the article titled ‘A becoming China and the assisted maturity of the EU: Assessing the factors determining their energy metabolic patterns’. Energy Strategy Reviews, 32, 100562. https://doi.org/10.1016/j.esr.2020.100562 , bring empirical results somehow similar to mine, although with a different method. The number of hours worked per person per year is mentioned in this paper as an important variable of the MuSIASEM framework for China. There is, for example, a comparison of energy metabolized in the sector of paid work, as compared to the household sector. It is found that the aggregate amount of human work used in a given sector of the economy is closely correlated with the aggregate energy metabolized by that sector. The economic development of China, and its pattern of societal metabolism in using energy, displays increase in the level of capitalization of all sectors, while reducing the human activity (paid work) in all of them except in the services. In the same time, the amount of human work per unit of real output seems to be negatively correlated with the capital-intensity (or capital-endowment) of particular sectors in the economy. Energy efficiency seems to be driven by decreasing work-intensity and increasing capital-intensity.
I found another similarity to my own research, although under a different angle, in the article by Koponen, K., & Le Net, E. (2021): Towards robust renewable energy investment decisions at the territorial level. Applied Energy, 287, 116552. https://doi.org/10.1016/j.apenergy.2021.116552 . The authors build a simulative model in Excel, where they create m = 5000 alternative futures for a networked energy system aiming at optimizing 5 performance metrics, namely: the LCOE cost of electricity, the GHG metric (greenhouse gases emission) for the climate, the density of PM2.5 and PM10 particles in the ambient air as a metric of health, capacity of power generation as a technological benchmark, and the number of jobs as social outcome. That complex vector of outcomes has been simulated as dependent on a vector of uncertainty as regards costs, and more specifically: cost of CO2, cost of electricity, cost of natural gas, and the cost of biomass. The model was based on actual empirical data as for those variables, and the ‘alternative futures’ are, in other words, 5000 alternative states of the same system. Outcomes are gauged with the so-called regret analysis, where the relative performance in a specific outcome is measured as the residual difference between its local value, and, respectively, its general minimum or maximum, depending on whether the given metric is something we strive to maximize (e.g. capacity), or to minimize (e.g. GHG). The regret analysis is very similar to the estimation of residual local error.
That short review of literature has the merit of showing me that I am not completely off the picture with the method and he findings which I initially presented to the editor of Applied Energy in that manuscript: ‘Climbing the right hill – an evolutionary approach to the European market of electricity’. The idea of understanding the mechanism of change in social structures, including the market of energy, by studying many alternative versions of said structure, seems to be catching in literature. I am progressively wrapping my mind around the fact that in my manuscript, the method is more important than the findings. The real value for money of my article seems to reside in the extent to which I can demonstrate the reproducibility and robustness of that method.
Thus, probably for the umpteenth time, I am rephrasing the fundamentals of my approach, and I am trying to fit it into the structure which Applied Energy recommends for articles submitted to their attention. I should open up with an ‘Introduction’, where I sketch the purpose of the paper, as well as the main points of the theoretical background which my paper stems from, although without entering into detailed study thereof. Then, I should develop on ‘Material and Methods’, with the main focus on making my method as reproducible as possible, and now comes the time to develop on, respectively, ‘Theory’ and ‘Calculation’, thus elaborating on the theoretical foundations of my research as pitched against literature, and on the detailed computational procedures I used. I guess that I need to distinguish, at this specific point, between the literature pertinent to the substance of my research (Theory), and that oriented on the method of working with empirical data (Calculation).
Those four initial sections – Introduction, Material and Methods, Theory, Calculation – open the topic up and then comes the time to give it a closure, with, respectively: ‘Results’, ‘Discussion’, and, optionally, a separate ‘Conclusion’. Over the top of that logical flow, I need to decorate with sections pertinent to ‘Data availability’, ‘Glossary’, and ‘Appendices’. As I get further back from the core and substance of my manuscript, and deeper into peripheral information, I need to address three succinct ways of presenting my research: Highlights, Graphical Abstract, and a structured cover letter. Highlights are 5 – 6 bullet points, 1 – 2 lines each, sort of abstract translated into a corporate presentation on slides. The Graphical Abstract is a challenge – as I need to present complex ideas in a pictographic form – and it is an interesting challenge. The structured cover letter should address the following points:
>> what is the novelty of this work?
>> is the paper appealing to a popular or scientific audience?
>> why the author thinks the paper is important and why the journal should publish it?
>> has the article been checked by an expert native speaker?
>> is the author available as reviewer?
Now, I ask myself fundamental questions. Why should anyone bother about the substance and the method of the research I present in my article. I noticed, both in public policies and in business strategies, a tendency to formulate completely unrealistic plans, and then to complain about other people not being smart enough to carry those plans out and up to happy ending. It is very visible in everything related to environmental policies and environmentally friendly strategies in business. Environmental activism consumes itself, very largely, in bashing everyone around for not being diligent enough in saving the planet.
To me, it looks very similarly to what I did many times as a person: unrealistic plans, obvious failure which anyone sensible could have predicted, frustration, resentment, practical inefficiency. I did it many times, and, obviously, whole societies are perfectly able to do it collectively. Action is key to success. A good plan is the plan which utilizes and reinforces the skills and capacities I already have, makes those skills into recurrent patterns of action, something like one good thing done per day, whilst clearly defining the skills I need to learn in order to be even more complete and more efficient in what I do. A good public policy, just as a good business strategy, should work in the same way.
When we talk about energy efficiency, or about the transition towards renewable energies, what is our action? Like really, what is the most fundamental thing we do together? Do we purposefully increase energy efficiency, in the first place? Do we deliberately transition to renewables? Yes, and no. Yes, at the end of the day we get those outcomes, and no, what we do on a daily basis is something else. We work. We do business. We study in order to get a job, or to start a business. We live our lives, from day to day, and small outcomes of that daily activity pile up, producing big cumulative change.
Instead of discussing what we do completely wrong, and thus need to change, it is a good direction to discover what we do well, consistently and with visible learning. That line of action can be reinforced and amplified, with good results. The so-far review of literature suggests that research concerning energy and energy transition is progressively changing direction, from the tendency to growing complexity and depth in study, dominant until recently, towards a translation of those complex, in-depth findings into relatively simple decision-making tools for policies and business strategies.
Here comes my method. I think it is important to create an analytical background for policies and business strategies, where we take commonly available empirical data at the macro scale, and use this data to discover the essential, recurrently pursued collective outcomes of a society, in the context of specific social goals. My point and purpose is to nail down a reproducible, relatively simple method of discovering what whole societies are really after. Once again, I think about something simple, which anyone can perform on their computer, with access to Internet. Nothing of that fancy stuff of social engineering, with personal data collected from unaware folks on Facebook. I want the equivalent of a screwdriver in positive, acceptably fair social engineering.
How do I think I can make a social screwdriver? I start with defining a collective goal we think we should pursue. In the specific case of my research on energy it is the transition to renewable sources. I nail down my observation of achievement, regarding that goal, with a simple metric, such as e.g. the percentage of renewables in total energy consumed (https://data.worldbank.org/indicator/EG.FEC.RNEW.ZS ) or in total electricity produced (https://data.worldbank.org/indicator/EG.ELC.RNEW.ZS ). I place that metric in the context of other socio-economic variables, such as GDP per capita, average hours worked per person per year etc. At this point, I make an important assumption as regards the meaning of all the variables I use. I assume that if a lot of humans go to great lengths in measuring something and reporting those measurements, it must be important stuff. I know, sounds simplistic, yet it is fundamental. I assume that quantitative variables used in social sciences represent important aspects of social life, which we do our best to observe and understand. Importance translates as significant connection to the outcomes of our actions.
Quantitative variables which we use in social sciences represent collectively acknowledged outcomes of our collective action. They inform about something we consistently care about, as a society, and, at the same time, something we recurrently produce, as a society. An array of quantitative socio-economic variables represents an imperfect, and yet consistently construed representation of complex social reality.
We essentially care about change. Both individual human nervous systems, and whole cultures, are incredibly accommodative. When we stay in a really strange state long enough to develop adaptive habits, that strange state becomes normal. We pay attention to things that change, whence a further hypothesis of mine that quantitative socio-economic variables, even if arithmetically they are local stationary states, serve us to apprehend gradients of change, at the level of collective, communicable cognition.
If many different variables I study serve to represent, imperfectly but consistently, the process of change in social reality, they might zoom on the right thing with various degrees of accuracy. Some of them reflect better the kind of change that is really important for us, collectively, whilst some others are just sort of accurate in representing those collectively pursed outcomes. An important assumption pops its head from between the lines of my writing: the bridging between pursued outcomes and important change. We pay attention to change, and some types of change are more important to us than others. Those particularly important changes are, I think, the outcomes we are after. We pay the most attention, both individually and collectively, to phenomena which bring us payoffs, or, conversely, which seriously hamper such payoffs. This is, once again on my path of research, a salute to the Interface Theory of Perception (Hoffman et al. 2015[2]; Fields et al. 2018[3]).
Now, the question is: how to extract orientations, i.e. objectively pursued collective outcomes, from that array of apparently important, structured observations of what is happening to our society? One possible method consists in observing trends and variance over time, and this is what I had very largely done, up to a moment, and what I always do now, with a fresh dataset, as a way of data mining. In this approach, I generally assume that a combination of relatively strong variance with strong correlation to the remaining metrics, makes a particular variable likely to be the driving undertow of the whole social reality represented by the dataset at hand.
Still, there is another method, which I focus on in my research, and which consists in treating the empirical dataset as a complex and imperfect representation of the way that collectively intelligent social structures learn by experimenting with many alternative versions of themselves. That general hypothesis leads to building supervised, purposefully biased experiments with that data. Each experiment consists in running the dataset through a specifically skewed neural network – a perceptron – where one variable from the dataset is the output which the perceptron strives to optimize, and the remaining variables make the complex input instrumental to that end. Therefore, each such experiment simulates an artificial situation when one variable is the desired and collectively pursued outcome, with other variables representing gradients of change subservient to that chief value.
When I run such experiments with any dataset, I create as many transformed datasets as there are variables in the game. Both for the original dataset, and for the transformed ones, I can calculate the mean value of each variable, thus construing a vector of mean expected values, and, according to classical statistics, such a vector is representative for the expected state of the dataset in question. I end up with both the original dataset and the transformed ones being tied to the corresponding vectors of mean expected values. It is easy to estimate the Euclidean distance between those vectors, and thus to assess the relative mathematical resemblance between the underlying datasets. Here comes something I discovered more than assumed: those Euclidean distances are very disparate, and some of them are one or two orders of magnitude smaller than all the rest. In other words, some among all the supervised experiments done yield a simulated state of social reality much more similar to the original, empirical one than all the other experiments. This is the methodological discovery which underpins my whole research in this article, and which emerged as pure coincidence, when I was working on a revised version of another paper, titled ‘Energy efficiency as manifestation of collective intelligence in human societies’, which I published with the journal ‘Energy’(https://doi.org/10.1016/j.energy.2019.116500 ).
My guess from there was – and still is – that those supervised experiments have disparate capacity to represent the social reality I study with the given dataset. Experiments which yield mathematical transformations relatively the most similar to the original set of empirical numbers are probably the most representative. Once again, the mathematical structure of the perceptron used in all those experiments is rigorously the same, and what makes the difference is the focus on one particular variable as the output to optimize. In other words, some among the variables studied represent much more plausible collective outputs than others.
I feel a bit lost in my own thinking. Good. It means I have generated enough loose thoughts to put some order in them. It would be much worse if I didn’t have thoughts to put order in. Productive chaos is better than sterile emptiness. Anyway, the reproducible method I want to present and validate in my article ‘Climbing the right hill – an evolutionary approach to the European market of electricity’ aims at discovering the collectively pursued social outcomes, which, in turn, are assumed to be the key drivers of social change, and the path to that discovery leads through the hypothesis that such outcomes are equivalent to specific a gradient of change, which we collectively pay particular attention to in the complex social reality, imperfectly represented with an array of quantitative socio-economic variables. The methodological discovery which I bring forth in that reproducible method is that when any dataset of quantitative socio-economic variables is being transformed, with a perceptron, into as many single-variable-optimizing transformations as there are variables in the set, 1 ÷ 3 among those transformations are mathematically much more similar to the original set of observations that all the other thus transformed sets. Consequently, in this method, it is expected to find 1 ÷ 3 variables which represent – much more plausibly than others – the possible orientations, i.e. the collectively pursued outcomes of the society I study with the given empirical dataset.
Ouff! I have finally spat it out. It took some time. The idea needed to ripe, intellectually. As it is ripe, I can harvest.
[1] Rosales-Asensio, E., de la Puente-Gil, Á., García-Moya, F. J., Blanes-Peiró, J., & de Simón-Martín, M. (2020). Decision-making tools for sustainable planning and conceptual framework for the energy–water–food nexus. Energy Reports, 6, 4-15. https://doi.org/10.1016/j.egyr.2020.08.020
[2] Hoffman, D. D., Singh, M., & Prakash, C. (2015). The interface theory of perception. Psychonomic bulletin & review, 22(6), 1480-1506.
[3] Fields, C., Hoffman, D. D., Prakash, C., & Singh, M. (2018). Conscious agent networks: Formal analysis and application to cognition. Cognitive Systems Research, 47, 186-213. https://doi.org/10.1016/j.cogsys.2017.10.003