Stratégies financières

Je continue un peu dans la foulée d’analyse des rapports courants des sociétés de ma liste « technologies nouvelles en énergie ». Je fais de mon mieux pour développer sur les premières observations que j’ai déjà présentées dans « Mes lampes rouges » ainsi que dans « Different paths ». Comme je résume partiellement ce que j’ai lu, ma première conclusion est une confirmation de mes intuitions initiales. Ce que nous appelons l’industrie de l’hydrogène est en fait une combinaison des technologies de pointe (piles à combustible à la base d’hydrogène, par exemple) avec des technologies bien établies – quoi que sujettes à l’innovation incrémentale – comme l’électrolyse ou le stockage des gaz volatiles. Il semble y avoir d’importants effets d’échelle, probablement en raison de la complexité technologique. Les sociétés relativement plus grandes, comme Plug Power ou Fuel Cell Energy, capables d’acquérir d’autres sociétés et leurs technologies, semblent être mieux placées dans la course technologique que des indépendants qui développent des technologies propriétaires de façon indépendante. C’est un truc que j’ai déjà remarqué dans le photovoltaïque et dans l’industrie de véhicules électriques : oui, il y a des petits indépendants prometteurs mais la bonne vieille intégration industrielle, surtout en verticale, semble revenir comme stratégie de choix après de décennies de bannissement. 

J’ai remarqué aussi que la catégorie générique « technologies d’hydrogène » semble attirer du capital de façon un peu inconsidérée. Je veux dire qu’il semble suffisant de dire « Eh, les gars, on invente dans l’hydrogène » pour que les investisseurs se précipitent, peu importe si le modèle d’entreprise est viable et transparent, ou pas-tout-à-fait-vous-comprenez-c’est-confidentiel. Je vois dans l’industrie d’hydrogène le même phénomène que j’observais, il y a encore 4 ou 5 ans, dans le photovoltaïque ou bien chez Tesla : lorsqu’une technologie nouvelle commence à prendre son envol en termes de ventes, les organisations qui s’y greffent et développent sont un peu démesurées ainsi qu’exagérément dépensières et il faut du temps pour qu’elles se fassent vraiment rationnelles.    

Pour gagner un peu de distance vis-à-vis le business d’hydrogène, je commence à piocher dans les rapports courants d’autres sociétés sur ma liste. Tesla vient en tête. Pas sorcier, ça. C’est la plus grosse position dans mon portefeuille boursier. Je lis donc le rapport courant du 4 août 2022 qui rend compte de 13 propositions soumises à l’assemblée générale d’actionnaires de Tesla le même 4 août, ainsi que de l’opération de fractionnement d’actions prévue pour la seconde moitié d’août. Ce dernier truc, ça m’intéresse peut-être le plus. La version officielle qui, bien entendu, sera mise à l’épreuve par le marché boursier, est que le Conseil D’Administration souhaite rendre les actions de Tesla plus accessibles aux investisseurs et employés et procédera donc, le 17 août 2022, à un fractionnement d’actions en proportion trois-pour-une en forme d’une dividende-actions. Chaque actionnaire enregistré ce 17 août 2022 verra le nombre de ses actions multiplié par trois. Les nouvelles actions fractionnées entreront en circulation boursière normale le 24 août 2022.

Il est vrai que les actions de Tesla sont plutôt chères en ce moment : presque $900 la pièce et ceci après la forte dépréciation dans la première moitié de l’année. Formellement, le fractionnement en proportion trois-pour-une devrait diviser cette cotation par trois, seulement le marché, ça suit les règles d’économie, pas d’arithmétique pure. Je pense que par la fin de 2022 on aura trois fois plus d’actions de Tesla flottantes et cotées à plus qu’un troisième du prix d’aujourd’hui encore qu’entre temps, il y aura des turbulences, je vous le dis. J’attache donc ma ceinture de sécurité – en l’occurrence c’est une position en Apple Inc., bien plus stable et respectable que Tesla – et j’attends de voir la valse boursière autour de ces actions fractionnées.

A part cette histoire de fractionnement, les autres 12 propositions couvrent 4 qui ont été acceptées – dont une relative au fractionnement déjà signalé – ainsi que 8 propositions non-acceptées. Les 4 acceptées sont relatives à, respectivement :

>> la nomination de deux personnes au Conseil D’Administration

>> l’accès par procuration, proposition sans engagement présentée par actionnaires en minorité

>> la ratification du choix de PricewaterhouseCoopers LLP comme auditeur financier de Tesla pour l’année comptable 2022

>> l’accroissement du nombre d’actions ordinaires de Tesla par 4 000 000

Les 8 propositions rejetées se groupent en deux catégories distinctes d’une façon intéressante. Il y en a donc deux qui viennent des cadres gestionnaires de Tesla et qui postulaient de modifier l’acte d’incorporation de Tesla de façon à éliminer la règle de majorité qualifiée de 66 et 2/3% dans les votes, ainsi qu’à réduire à 2 ans le mandat des directeurs du Conseil d’Administration. Ces deux propositions-là ont perdu car elles n’avaient pas… de majorité qualifiée de 66 et 2/3%. Les 6 propositions restantes parmi les non-acceptées étaient toutes des propositions sans engagement de la part d’actionnaires minoritaires et toutes les 6 demandaient des rapports additionnels ou bien des changements afférents à, respectivement : la qualité de l’eau, travail forcé d’enfants, le lobbying, la liberté d’association, arbitrage dans les affaires d’emploi, la diversité au sein du Conseil d’Administration, les politiques internes contre le harassement et la discrimination.

Dans le vocabulaire politique de mon pays, la Pologne, nous avons l’expression « compter les sabres ». Elle désigne des votes qui sont perdus d’avance mais qui servent à compter la taille de la coalition possible que le proposant donné pourrait rallier pour quelque chose de plus sérieux. Je bien l’impression que quelqu’un chez Tesla commence à compter les sabres.

Je passe au rapport courant de Tesla du 20 juillet 2022 qui, en fait, annonce leur rapport financier du 2ème trimestre 2022. Ça a l’air bien. Le bénéfice net pour la première moitié de 2022 a triplé par rapport à la même période de 2021, le flux de trésorerie se fait plus robuste. Rien à dire.                

Je tourne vers un modèle d’entreprise beaucoup plus fluide, donc celui de Nuscale Power ( https://ir.nuscalepower.com/overview/default.aspx ). Lorsqu’on lit la présentation générale de ce business (https://ir.nuscalepower.com/overview/default.aspx ), tout colle à merveille : NuScale Power fournit des petits réacteurs nucléaires innovatifs, où un module peut fournir 77 mégawatts de puissance. Seulement, lorsque je commence à lire leur rapport annuel 2021, ça se corse, parce que le rapport est publié par l’entité nommée Spring Valley Acquisition Corporation, qui se présente comme une société coquille incorporée dans les îles Cayman, sous la forme légale de société exonérée. Le management déclarait, dans le rapport annuel 2021, que le but de Spring Valley Acquisition Corporation est de conduire une fusion ou bien une acquisition, un échange d’actions ou bien leur achat, une acquisition d’actifs, une réorganisation ou bien une autre forme de regroupement d’entreprises. Au mois de mars 2021 ; Spring Valley Acquisition Corporation est entrée en un accord tripartite, accompagné d’un plan de fusion, avec sa filiale en propriété exclusive, Spring Valley Merger Sub, Inc., incorporée dans l’état de Delaware, ainsi qu’avec Dream Holdings Inc., une autre société incorporée dans le Delaware, celle-ci sous la forme de société d’utilité publique. C’est une nouveauté dans la loi des sociétés dans le Delaware, introduite en 2013. Un article intéressant à ce sujet est accessible sur « Harward Law School Forum on Corporate Governance ».

Ainsi donc, en mars 2021, Spring Valley Acquisition Corporation, Spring Valley Merger Sub, Inc. et Dream Holdings Inc. avaient convenu de conduire un regroupement d’entreprises avec AeroFarms. Dream Holdings fusionne avec Spring Valley Merger Sub. En octobre 2021, l’accord en question a été résilié. En décembre 2021, Spring Valley Acquisition Corporation entre en un nouvel accord tripartite, encore une fois avec la participation de Spring Valley Merger Sub. Cette fois, Spring Valley Merger Sub est introduite comme une LLC (société à responsabilité limitée) incorporée dans l’état d’Oregon. La troisième partie de l’accord est NuScale Power LLC, aussi incorporée en Oregon. Poursuivant cet accord, Spring Valley Acquisition Corporation change de lieu d’incorporation des îles Cayman pour l’état de Delaware, pendant que Spring Valley Merger Sub LLC fusionne avec et en NuScale Power LLC. Après la fusion, Spring Valley change de nom et devient NuScale Power Corporation.

Comment a marché la combine ? Eh bien, voici une annonce courante de NuScale Power, datant d’hier (10 août), où NuScale donne un aperçu de leurs résultats pour le 2nd trimestre 2022. La perte d’exploitation pour cette première moitié de l’année 2022 était de 44,75 millions de dollars, un peu moins que dans la première moitié de 2021. Leurs actifs ont presque triplé en 12 mois, de $121,2 millions à $407,3 millions. Côté exploitation, une nouvelle entité opérationnelle est créé sous le nom de « VOYGR™ Services and Delivery (VSD) » avec la mission d’organiser les services, les fournitures et la gestion clients pour la technologie VOYGR™. Cette dernière est la technologie pour bâtir et exploiter des centrales nucléaires à puissance moyenne sur la base de « NuScale Power Module™ », soit avec 4 modules dedans et une puissance de 924 mégawatts de puissance électrique (VOYGR-4) soit avec 6 modules (VOYGR-6).

Cette comparaison rapide d’évènements relativement récents chez Tesla et NuScale Power me conduit à la conclusion que si je veux comprendre à fond un modèle d’entreprise, il faut que je m’intéresse plus (que je l’avais fait jusqu’à présent) à ce qui se passe dans les passifs du bilan. Je vois que des différentes phases d’avancement dans le développement d’une technologie s’accompagnent des stratégies financières très différentes et le succès technologique dépend largement du succès de ces mêmes stratégies.

Different paths

I keep digging in the business models of hydrogen-oriented companies, more specifically five of them: 

>> Fuel Cell Energy https://investor.fce.com/Investors/default.aspx

>> Plug Power https://www.ir.plugpower.com/overview/default.aspx

>> Green Hydrogen Systems https://investor.greenhydrogen.dk/

>> Nel Hydrogen https://nelhydrogen.com/investor-relations/

>> Next Hydrogen (previously BioHEP Technologies Ltd.) https://nexthydrogen.com/investor-relations/why-invest/

I am studying their current reports. This is the type of report which listed companies publish when something special happens, which goes beyond the normal course of everyday business, and can affect shareholders. I have already started with Fuel Cell Energy and their current report from July 12th, 2022 (https://d18rn0p25nwr6d.cloudfront.net/CIK-0000886128/b866ae77-6f4a-421e-bedd-906cb92850d7.pdf ), where they disclose a deal with a group of financial institutions: Jefferies LLC, B. Riley Securities, Inc., Barclays Capital Inc., BMO Capital Markets Corp., BofA Securities, Inc., Canaccord Genuity LLC, Citigroup Global

Markets Inc., J.P. Morgan Securities LLC and Loop Capital Markets LLC. Strange kind of deal, I should add. Those 10 financial firms are supposed to either buy or intermediate in selling to third parties parcels of 95 000 000 shares in the equity of Fuel Cell Energy. The tricky part is that the face value of those shares is supposed to be $0,0001 per share, just as it is the case with the ordinary 837 000 000 shares outstanding, whilst the market value of Fuel Cell Energy’s shares is currently above $4,00 per share, thus carrying an addition of thousands of percentage points of capital to pay.

It looks as if the part of equity in Fuel Cell Energy which is free floating in the stock market – quite a tiny part of their share capital – was becoming subject to quick financial gambling. I don’t like it. Whatever. Let’s go further, i.e. to the next current report of Fuel Cell Energy, that from July 7th, 2022 (https://d18rn0p25nwr6d.cloudfront.net/CIK-0000886128/77053fbf-f22a-4288-b702-6b82a039f588.pdf ). It brings updates on two projects:

>> The Toyota Project: a 2,3 megawatt trigeneration platform for Toyota at the Port of Long Beach, California.

>> The Groton Project: a 7.4 MW platform at the U.S. Navy Submarine Base in Groton, Connecticut.

Going further back in time, I browse through the current report from June 9th, 2022 (https://d18rn0p25nwr6d.cloudfront.net/CIK-0000886128/9f4b19f0-0a11-4d27-acd2-f0881fdefbc3.pdf ). It is the official version of a press release regarding financial and operational results of Fuel Cell Energy by the end of the 1st quarter 2022. As I am reading through it, I find data about other projects:

>> Joint Development Agreement with ExxonMobil, related to carbon capture and generation, which includes the 7,4 MW LIPA Yaphank fuel cell project

>>  a carbon capture project with Canadian National Resources Limited

>> a program with U.S. Department of Energy regarding solid oxide. I suppose that ‘solid oxide’ stands for solid oxide fuel cells, which use a solid, ceramic core of fuel, which is being oxidized and produces energy in the process.     

I pass to the current reports of Plug Power (https://www.ir.plugpower.com/financials/sec-filings/default.aspx ). Interesting things start when I go back to the current report from June 23rd, 2022 (https://d18rn0p25nwr6d.cloudfront.net/CIK-0001093691/36efa8c2-a675-451b-a41f-308221f5e612.pdf ). This is a summary presentation of something which looks like the company’s strategy. Apparently, Plug Power plans to have 13 plants with Green Hydrogen running in the United States by 2025, with a total expected yield of 500 tons per day. In a more immediate perspective, the company plans to locate 5 new plants in the U.S. over 2022 (total capacity of 70 tons per day) and 2023 (200 tons per day). Further, I read that what I thought was a hydrogen-focused company, has, in fact, a broader spectrum of operations: eFuel and methanol, ammonia, vehicle refueling, blending and heating, refining of natural oil, and the storage of renewable energy.  

As part of its strategy, Plug Power announces the acquisitions of companies supposed to bring additional technological competences: Frames Group (https://www.frames-group.com/ ) with power transmission systems and technology for building electrolyzers, ACT (Applied Cryo Technologies: https://www.appliedcryotech.com/ ) for cryogenics, and Joule (https://www.jouleprocess.com/about ) for the liquefaction of hydrogen. My immediate remark as regards those acquisitions, sort of intellectually straight-from-the-oven-still-warm-sorry-but-I-told-you-still-warm, is that Plug Power is acquiring a broad technological base rather than a specialized one. Officially, those acquisitions serve to enhance the Plug Power’s capacity as regards the deployment of hydrogen-focused technologies. Yet, as I am rummaging through the websites of those acquired companies, their technological competences go far beyond hydrogen.

Sort of contingent (adjacent?) to that current report is the piece of news, still on the Plug Power’s investors-relations site, from June 8th, 2022. It regards the deployment of a project in Europe, more specifically in the Port of Antwerp-Bruges (https://www.ir.plugpower.com/press-releases/news-details/2022/Plug-to-Build-Large-Scale-Green-Hydrogen-Generation-Plant-in-Europe-at-Port-of-Antwerp-Bruges/default.aspx ). This is supposed to be something labelled as a ‘Gigafactory’.

A little bit earlier this year, on my birthday, May 9th, Plug Power published a current report (https://d18rn0p25nwr6d.cloudfront.net/CIK-0001093691/203fd9c3-5302-4fa1-9edd-32fe4905689c.pdf ) coupled with a quarterly financial report (https://d18rn0p25nwr6d.cloudfront.net/CIK-0001093691/c7ad880f-71ff-4b58-8265-bd9791d98740.pdf ). Apparently, in the 1st quarter 2022, they had revenues 96% higher than 1Q 2021. Nice. There are interesting operational goals signaled in that current report. Plug Power plans to reduce services costs on a per unit basis by 30% in the 12 months following the report, thus until the end of the 1st quarter 2023. The exact quote is: ‘Plug remains focused on delivering on our previously announced target to reduce services costs on a per unit basis by 30% in the next 12 months, and 45% by the end of 2023. We are pleased to report that we have begun to see meaningful improvement in service margins on fuel cell systems and related infrastructure with a positive 30% increase in first quarter of 2022 versus the fourth quarter of 2021. The service margin improvement is a direct result of the enhanced technology GenDrive units that were delivered in 2021 which reduce service costs by 50%. The performance of these enhanced units demonstrates that the products are robust, and we expect these products will help support our long-term business needs. We believe service margins are tracking in the right direction with potential to break even by year end’.

When a business purposefully and effectively works on optimizing margins of profit, and the corresponding costs, it is a step forward in the lifecycle of the technologies used. This is a passage from the phase of early development towards late development, or, in other words, it is the phase when the company starts getting in control of small economic details in its technology.

I switch to the next company on my list, namely to Green Hydrogen Systems (Denmark, https://investor.greenhydrogen.dk/ ). They do not follow the SEC classification of reports, and, in order to get an update on their current developments, I go to their ‘Announcements & News’ section (https://investor.greenhydrogen.dk/announcements-and-news/default.aspx ).  On July 18th, 2022, Green Hydrogen Systems held an extraordinary General Meeting of shareholders. They amended their Articles of Association, as regards the Board of Directors, and the new version is: ‘The board of directors consists of no less than four and no more than nine members, all of whom must be elected by the general meeting. Members of the board of directors must resign at the next annual general meeting, but members of the board of directors may be eligible for re-election’. At the same extraordinary General Meeting, three new directors have been elected to the Board, on the top of the six already there.

To the extent that I know the Scandinavian ways of corporate governance, appointment of new directors to the Board usually comes with new business ties of the company. Those people are supposed to be something like intermediaries between the company and some external entities (research units? other companies? NGOs?). That change in the Board of Directors at Green Hydrogen Systems suggests something like the broadening of their network. That intuition is somehow confirmed by an earlier announcement, from June 13th (https://investor.greenhydrogen.dk/announcements-and-news/news-details/2022/072022-Green-Hydrogen-Systems-announces-changes-to-the-Board-of-Directors-and-provides-product-status-update/default.aspx ). The three new members of the Board come, respectively, from: Vestas Wind Systems, Siemens Energy, and Sonnedix (https://www.sonnedix.com/ ).

Still earlier this year, on April 12th, Green Hydrogen Systems announced ‘design complications in its HyProvide® A-Series platform’, and said complications are supposed to affect adversely the financial performance in 2022 (https://investor.greenhydrogen.dk/announcements-and-news/news-details/2022/Green-Hydrogen-Systems-announces-technical-design-complications-in-its-HyProvide-A-Series-platform/default.aspx ). When I think about it, design normally comes before its implementation, and therefore before any financial performance based thereon. When ‘design complications’ are serious enough for the company to disclose them and announce a possible negative impact on the financial side of the house, it means some serious mistakes years earlier, when that design was being conceptualized. I say ‘years’ because I notice the trademark symbol ‘®’ by the name of the technology. That means there had been time to: a) figure out the design b) register it as a trademark. That suggests at least 2 years, maybe more.

I quickly sum up my provisional conclusions from browsing current reports at Fuel Cell Energy, Plug Power, and Green Hydrogen Systems. I can see three different courses of events as regards the business models of those companies. At Fuel Cell Energy, broadly spoken marketing, including financial marketing, seems to be the name of the game. Both the technology and the equity of Fuel Cell Energy seems to be merchandise for trading. My educated guess is that the management of Fuel Cell Energy is trying to attract more financial investors to the game, and to close more technological deals, of the joint-venture type, at the operational level. It further suggests an attempt at broadening the business network of the company, whilst keeping the strategic ownership in the hands of the initial founders. As for Plug Power, the development I see is largely quantitative. They are broadening their technological base, including the acquisitions of strategically important assets, expanding their revenues, and ramping up their operational margins. This a textbook type of industrial development. Finally, at Green Hydrogen Systems, this still seems to be the phase of early development, with serious adjustments needed to both the technology owned and the team that runs it.

Those hydrogen-oriented companies seem to be following different paths and to be at different stages in the lifecycle of their technological base.

The real deal

I am blogging again, after months of break. My health required some attention, and my life priorities went a bit wobbly for some time, possibly because of the opioid pain killers which I took in hospital, after my surgery. Anyway, I am back in the game, writing freestyle.

Restarting after such a long break is a bit hard, and yet rewarding. I am removing rust from my thoughts, as if I were giving a new life to an old contrivance. I need to work up to cruise speed in my blogging. Currently, I am working on two subjects. One is my concept of Energy Ponds: a solution which combines ram pumps, hydropower, and the retention of water in wetlands. The other one pertains to business models in the broadly spoken industry of new sources of energy: electric vehicles (I am and remain a faithful investor in Tesla), technologies of energy storage, hydrogen and fuel cells based thereon, photovoltaic, wind and nuclear.

As I am thinking about it, the concept of Energy Ponds is already quite structured, and I am working on structuring it further by attracting the attention of people with knowledge and skills complementary to mine. On the other hand, the whole business models thing is foggy theoretically, and, at the same time, it is important to me at many levels, practical strategies of investment included. I know by experience that such topics – both vague and important – are the best for writing about on my blog.

Here comes the list of companies which I observe more or less regularly with respect to their business models:

>> Tesla https://ir.tesla.com/#quarterly-disclosure

>> Rivian https://rivian.com/investors

>> Lucid Group https://ir.lucidmotors.com/

>> Nuscale Power https://ir.nuscalepower.com/overview/default.aspx 

>> First Solar https://investor.firstsolar.com/home/default.aspx

>> SolarEdge https://investors.solaredge.com/

>> Fuel Cell Energy https://investor.fce.com/Investors/default.aspx

>> Plug Power https://www.ir.plugpower.com/overview/default.aspx

>> Green Hydrogen Systems https://investor.greenhydrogen.dk/

>> Nel Hydrogen https://nelhydrogen.com/investor-relations/

>> Next Hydrogen (précédemment BioHEP Technologies Ltd.) https://nexthydrogen.com/investor-relations/why-invest/

>> Energa https://ir.energa.pl/en

>> PGE https://www.gkpge.pl/en

>> Tauron https://raport.tauron.pl/en/tauron-in-2020/stock-exchange/investor-relations/

>> ZPUE  https://zpue.com/   

Two classifications come to my mind as I go through that list. Firstly, there are companies which I currently hold an investment position in: Tesla, Nuscale Power, Energa, PGE, Tauron et ZPUE. Then come those which I used to flirt with, namely Lucid Group, First Solar and SolarEdge. Finally, there are businesses which I just keep watching from a distance: Rivian, Fuel Cell Energy, Plug Power, Green Hydrogen Systems, Nel Hydrogen, and Next Hydrogen.

The other classification is based on the concept of owners’ earnings such as defined by Warren Buffett: net income plus amortization minus capital expenses. Tesla, PGE, Energa, ZPUE, Tauron, First Solar, SolarEdge – these guys generate a substantial stream of owners’ earnings. The others are cash-negative. As for the concept of owners’ earnings itself, you can consult both the investor-relations site of Berkshire Hathaway (https://www.berkshirehathaway.com/  ) or read a really good book by Robert G.Hagstrom « The Warren Buffett Way » (John Wiley & Sons, 2013, ISBN 1118793994, 9781118793992). I guess the intuition behind hinging my distinctions upon the cash-flow side of the house assumes that in the times of uncertainty, cash is king. Rapid technological change is full of uncertainty, especially when that change affects whole infrastructures, as it is the case with energy and propulsion. Besides, I definitely buy into Warren Buffett’s claim that cash-flow is symptomatic of the lifecycle in the given business.

The development of a business, especially on the base of innovative technologies, is cash-consuming. Cash, in business, is something we harvest rather than simply earn. Businesses which are truly able to harvest cash from their operations, have internal financing for moving to the next cycle of technological change. Those in need of cash from outside will need even more cash from outside in order to finance further innovation.

What’s so special about, cash in a business model? The most intuitive answer that comes to my mind is a motto heard from a banker, years ago: “In the times of crisis, cash is king”. Being a king means sovereignty in a territory, like “This place is mine, and, with all the due respect, pay respect or f**k off”. Having cash means having sovereignty of decision in business. Yet, nuance is welcome. Cash is cash. Once you have it, it does not matter that much where it came from, i.e. from operations or from external sources. When I have another look at businesses without positive owners’ earnings – Nuscale Power, Rivian, Fuel Cell Energy, Plug Power, Green Hydrogen Systems, Nel Hydrogen, and Next Hydrogen – I shift my focus from their cash-flow statements to their balance sheets and I can see insane amounts of cash on the assets’ side of the house. These companies, in their assets, have more cash than they have anything else. They look almost like banks, or investment funds.

Thus, my distinction between business models with positive owners’ earnings, on the one hand, and those without it, on the other hand, is a distinction along the axis of strategic specificity. When the sum total of net income and amortization, reduced by capital expenses, is positive and somehow in line with the market capitalization of the whole company, that company is launched on some clear tracks. The business is like a river: it is predictable and clearly traceable in its strategic decisions. On the other hand, a business with lots of cash in the balance sheet but little cash generated from operations is like lord Byron (George Gordon): those guys assume that the only two things worth doing are poetry and cavalry, only they haven’t decided yet the exact mix thereof.      

That path of thinking implies that a business model is more than a way of conducting operations; it is a vehicle for change through investment, thus for channeling capital with strategic decisions. Change which is about to come is somehow more interesting than change which is already there. Seen under this angle, businesses on my list convey different degrees of vagueness, and, therefore, different doses of intellectual provocation. I focus on the hydrogen ones, probably because in my country, Poland, we have that investment program implemented by the government: the hydrogen valleys.

As I have another look at the hydrogen-oriented companies on my list – Fuel Cell Energy, Plug Power, Green Hydrogen Systems, Nel Hydrogen, and Next Hydrogen – an interesting discrepancy emerges as regards the degree of technological advancement. Green Hydrogen Systems, Nel Hydrogen, and Next Hydrogen are essentially focused on making and supplying hydrogen. This is good old electrolysis, a technology with something like a century of industrial tradition, combined with the storage and transport of highly volatile gases. Only two, namely Fuel Cell Energy and Plug Power, are engaged into fuel cells based on hydrogen, and those fuel cells are, in my subjective view, the real deal as it comes to hydrogen-related innovation.

Ugly little cherubs

I am working on my long-term investment strategy, and I keep using the Warren Buffet’s tenets of investment (Hagstrom, Robert G.. The Warren Buffett Way (p. 98). Wiley. Kindle Edition.).

At the same time, one of my strategic goals is coming true, progressively: other people reach out to me and ask whether I would agree to advise them on their investment in the stock market. People see my results, sometimes I talk to them about my investment philosophy, and it seems to catch on.

This is both a blessing and a challenge. My dream, 2 years ago, when I was coming back to the business of regular investing in the stock market, was to create, with time, something like a small investment fund specialized in funding highly innovative, promising start-ups. It looks like that dream is progressively becoming reality. Reality requires realistic and intelligible strategies. I need to phrase out my own experience as regards investment in a manner, which is both understandable and convincing to other people.

As I am thinking about it, I want to articulate my strategy along three logical paths. Firstly, what is the logic in my current portfolio? Why am I holding the investment positions I am holding? Why in these proportions? How have I come to have that particular portfolio? If I can verbally explain the process of my so-far investment, I will know what kind of strategy I have been following up to now. This is the first step, and the next one is to formulate a strategy for the future. In one of my recent updates (Tesla first in line), I briefly introduced my portfolio, such as it was on December 2nd, 2021. Since then, I did some thinking, most of all in reference to the investment philosophy of Warren Buffett, and I made some moves. I came to the conclusion that my portfolio was astride a bit too many stocks, and the whole was somehow baroque. By ‘baroque’ I mean that type of structure, where we can have a horribly ugly little cherub, accompanied by just as ugly a little shepherd, but the whole looks nice due to the presence of a massive golden rim, woven around ugliness.

I made myself an idea of what are the ugly cherubs in my portfolio from December 2nd, and I kicked them out of the picture. In the list below, these entities are marked in slashed bold italic:

>> Tesla (https://ir.tesla.com/#tab-quarterly-disclosure),

>> Allegro.eu SA (https://about.allegro.eu/ir-home ),

>> Alten (https://www.alten.com/investors/ ),

>> Altimmune Inc (https://ir.altimmune.com/ ),

>> Apple Inc (https://investor.apple.com/investor-relations/default.aspx ),

>> CureVac NV (https://www.curevac.com/en/investor-relations/overview/ ),

>> Deepmatter Group PLC (https://www.deepmatter.io/investors/ ), 

>> FedEx Corp (https://investors.fedex.com/home/default.aspx ),

>> First Solar Inc (https://investor.firstsolar.com/home/default.aspx )

>> Inpost SA (https://www.inpost.eu/investors )

>> Intellia Therapeutics Inc (https://ir.intelliatx.com/ )

>> Lucid Group Inc (https://ir.lucidmotors.com/ )

>> Mercator Medical SA (https://en.mercatormedical.eu/investors/ )

>> Nucor Corp (https://www.nucor.com/investors/ )

>> Oncolytics Biotech Inc (https://ir.oncolyticsbiotech.com/ )

>> Solaredge Technologies Inc (https://investors.solaredge.com/ )

>> Soligenix Inc (https://ir.soligenix.com/ )

>> Vitalhub Corp (https://www.vitalhub.com/investors )

>> Whirlpool Corp (https://investors.whirlpoolcorp.com/home/default.aspx )

>> Biogened (https://biogened.com/ )

>> Biomaxima (https://www.biomaxima.com/325-investor-relations.html )

>> CyfrPolsat (https://grupapolsatplus.pl/en/investor-relations )

>> Emtasia (https://elemental-asia.biz/en/ )

>> Forposta (http://www.forposta.eu/relacje_inwestorskie/dzialalnosc_i_historia.html )

>> Gameops (http://www.gameops.pl/en/about-us/ )

>> HMInvest (https://grupainwest.pl/relacje )

>> Ifirma (https://www.ifirma.pl/dla-inwestorow )

>> Moderncom (http://moderncommercesa.com/wpmccom/en/dla-inwestorow/ )

>> PolimexMS (https://www.polimex-mostostal.pl/en/reports/raporty-okresowe )

>> Selvita (https://selvita.com/investors-media/ )

>> Swissmed (https://swissmed.com.pl/?menu_id=8 )  

Why did I put those specific investment positions into the bag labelled ‘ugly little cherubs in the picture’? Here comes a cognitive clash between the investment philosophy I used to have before I started studying in depth that of Warren Buffet and of Berkshire Hathaway. Before, I was using the purely probabilistic approach, according to which the stock market is so unpredictable that my likelihood of failure, on any individual investment, is greater than the likelihood of success, and, therefore, the more I spread my portfolio between different stocks, the less exposed I am to the risk of a complete fuck-up. As I studied the investment philosophy of Warren Buffet, I had great behavioural insights as regards my decisions. Diversifying one’s portfolio is cool, yet it can lead to careless individual choices. If my portfolio is really diversified, each individual position weighs so little that I am tempted to overlook its important features. At the end of the day, I might land with a bag full of potatoes instead of a chest full of gems.

I decided to kick out the superfluous. What did I put in this category? The superfluous investment positions which I kicked out shared some common characteristics, which I reconstructed from the history of the corresponding ‘buy’ orders. Firstly, these were comparatively small positions, hundreds of euros at best. This is one of the lessons by Warren Buffet. Small investments matter little, and they are probably going to stay this way. There is no point in collecting stocks which don’t matter to me. They give is a false sense of security, which is detrimental to the focus on capital gains.  

Secondly, I realized that I bought those ugly little cherubs by affinity to something else, not for their own sake. Two of them, FedEx and Allegro, are in the busines of express delivery. I made a ton of money of their stock, just as on the stock of Deutsche Post, during the trough of the pandemic, where retail distribution went mostly into the ‘online order >> express delivery’ pipeline. It was back then, and then I sold out, and then I thought ‘why not trying the same hack again?’. The ‘why not…?’ question was easy to answer, actually: because times change, and the commodity markets have adapted to the pandemic. FedEx and Allegro has returned to what it used to be: a solid business without much charm to me.  

Four others – Soligenix, Altimmune, CureVac and Oncolytics Biotech – are biotechnological companies. Once again: I made a ton of money in 2020 on biotech companies, because of the pandemic. Now, emotions in the market have settled, and biotech companies are back what they used to be, namely interesting investments endowed with high risk, high potential reward, and a bottomless capacity for burning cash. Those companies are what Tesla used to be a decade ago. I kept a serious position on a few other biotech businesses: Intellia Therapeutics, Biogened, Biomaxima, and Selvita. I want to keep a few of such undug gems in my portfolio, yet too much would be too much.

Thirdly, I had a loss on all of those ugly little cherubs I have just kicked out of my portfolio. Summing up, these were small positions, casually opened without much strategic thinking, and they were bringing me a loss. I could have waited to have a profit, but I preferred to sell them out and to concentrate my capital on the really promising stocks, which I nailed down using the method of intrinsic value. I realized that my portfolio was what it was, one week ago, before I started strategizing consciously, because I had hard times finding balance between two different motivations: running away from the danger of massive loss, on the one hand, and focusing on investments with a true potential for bringing long-term gains.

I focus more specifically on the concept of intrinsic value. Such as Warren Buffet used it, intrinsic value was based on what he called ‘owner’s earnings’ from a business. Owner’s earnings are spread over a window in time corresponding to the risk-free yield on sovereign bonds. The financial statement used for calculating intrinsic value is the cash-flow of the company in question, plus external data as regards average annual yield on sovereign bonds. The basic formula to calculate owner’s earnings goes like: net income after tax + amortization charges – capital expenditures). Once that nailed down, I divide those owner’s earnings by the interest rate on long-term sovereign bonds. For my positions in the US stock market, I use the long-term yield on the US federal bonds, i.e. 1,35% a year. As regards my portfolio in the Polish stock market, I use the yield 3,42% for Polish sovereign bonds on long-term.

I have calculated that intrinsic value for a few of my investments (I mean those I kept in my portfolio), on the basis of their financial results for 2020 and compared it to their market capitalisation. Then, additionally, I did the same calculation based on their published (yet unaudited) cash-flow for Q3 2021. Here are the results I had for Tesla. Net income 2020 $862,00 mln plus amortization charges 2020 $2 322,00 mln minus capital expenditures 2020 $3 132,00 mln equals owner’s earnings 2020 $52,00 mln. Divided by 1,35%, that gives an intrinsic value of $3 851,85 mln. Market capitalization on December 6th, 2021: $1 019 000,00 mln. The intrinsic value looks like several orders of magnitude smaller than market capitalisation. Looks risky.

Let’s see the Q3 2021 unaudited cash-flows. Here, I extrapolate the numbers for 9 months of 2021 over the whole year 2021: I multiply them by 4/3. Extrapolated net income for Q3 2021 $4 401,33 mln plus extrapolated amortization charges for Q3 2021 $2 750,67 minus extrapolated capital expenditures for Q3 2021 $7 936,00 equals extrapolated owner’s earnings amounting to $4 401,33 mln. Divided by 1,35%, it gives an extrapolated intrinsic value of $326 024,69 mln. It is much closer to market capitalization, yet much below it as for now. A lot of risk in that biggest investment position of mine. We live and we learn, as they say.

Another stock: Apple. With the economic size of a medium-sized country, Apple seems solid. Let’s walk it through the computational path of intrinsic value. There is an important methodological remark to formulate as for this cat. In the cash-flow statement of Apple for 2020-2021 (Apple Inc. ends its fiscal year by the end of September in the calendar year), under the category of ‘Investing activities’, most of the business pertains to buying and selling financial assets. It goes, ike:

Investing activities, in millions of USD:

>> Purchases of marketable securities (109 558)

>> Proceeds from maturities of marketable securities: 59 023

>> Proceeds from sales of marketable securities: 47 460

>> Payments for acquisition of property, plant and equipment (11 085)

>> Payments made in connection with business acquisitions, net (33)

>> Purchases of non-marketable securities (131)

>> Proceeds from non-marketable securities: 387

>> Bottom line: Cash generated by/(used in) investing activities (14 545)

Now, when I look at the thing through the lens of Warren Buffett’s investment tenets, anything that happens with and through financial securities, is retention of cash in the business. It just depends on what exact form we want to keep that cash under. Transactions grouped under the heading of ‘Purchases of marketable securities (109 558)’, for example, are not capital expenditures. They do not lead to exchanging cash money against productive technology. In all that list of investment activities, only two categories, namely: ‘Payments for acquisition of property, plant and equipment (11 085)’, and ‘Payments made in connection with business acquisitions, net (33)’ are capital expenditures sensu stricto. All the other categories, although placed in the account of investing activities, are labelled as such just because they pertain to transactions on assets. From the Warren Buffet’s point of view they all mean retained cash.

Therefore, when I calculate owner’s earnings for Apple, based on their latest annual cash-flow, I go like:

>> Net Income $94 680 mln + Depreciation and Amortization $11 284 mln + Purchases of marketable securities $109 558 mln + Proceeds from maturities of marketable securities $59 023 mln + Proceeds from sales of marketable securities $47 460 mln – Payments for acquisition of property, plant and equipment $11 085 mln – Payments made in connection with business acquisitions, net $33 mln + Purchases of non-marketable securities $131 mln + Proceeds from non-marketable securities $387 mln = Owner’s earnings $311 405 mln.

I divide that number by the 1,35% annual yield of the long-term Treasury bonds in the US, and I get an intrinsic value of $23 067 037 mln, against a market capitalisation floating around $2 600 000 mln, which gives a huge overhead in the former over the latter. Good investment.

I pass to another one of my investments, First Solar Inc. (https://investor.firstsolar.com/financials/sec-filings/default.aspx ). Same thing: investment activities consist most of all in moves pertinent to financial assets. It looks like:

>> Net income (loss) $398,35 mln

>> Depreciation, amortization and accretion $232,93 mln

>> Impairments and net losses on disposal of long-lived assets $35,81 mln

… and then come the Cash flows from investing activities:

 >> Purchases of property, plant and equipment ($416,64 mln)

>> Purchases of marketable securities and restricted marketable securities ($901,92 mln)

>> Proceeds from sales and maturities of marketable securities and restricted marketable securities $1 192,83 mln

>> Other investing activities ($5,5 mln)

… and therefore, from the perspective of owner’s earnings, the net cash used in investing activities is not, as stated officially, minus $131,23 mln. Net capital expenses, I mean net of transactions on financial assets, are: – $416,64 mln + $901,92 mln + $1 192,83 mln – $5,5 mln = $1 672,61 mln. Combined with the aforementioned net income, amortization and fiscally compensated impairments on long-lived assets, it makes owner’s earnings of $2 339,7 mln. And an intrinsic value of $173 311,11 mln, against some $10 450 000 mln in market capitalization. Once again, good and solid in terms of Warren Buffet’s margin of security.

I start using the method of intrinsic value for my investments, and it gives interesting results. It allows me to distinguish, with a precise gauge, between high-risk investments and the low-risk ones.

Je dois faire gaffe à la valeur intrinsèque

Me revoilà sur mon blog et je me concentre sur un truc : ma stratégie d’investissements boursiers. Je veux optimiser ma stratégie et à cette fin je me réfère à Warren Buffett et à sa philosophie d’investissement telle que vous et moi pouvons la trouver dans les rapports annuels de Berkshire Hathaway Inc. (https://www.berkshirehathaway.com/reports.html ). Je prends donc les principes de Warren Buffett et je les applique comparativement à deux compagnies : Tesla (https://ir.tesla.com/#tab-quarterly-disclosure ), la plus grande position dans mon portefeuille d’investissement, d’une part, et Selvita (https://selvita.com/investors-media/ ), une société polonaise de biotechnologie, sur les actions de laquelle je commence à développer un investissement sérieux.   

Avec Tesla, j’ai déjà entamé une analyse façon Warren Buffett (Tesla first in line) et maintenant je continue de manière comparative. C’est un truc qui marche : lorsque je veux comprendre quelque chose de complexe, je peux comparer cette chose complexe avec une autre chose complexe. Comparaison est une stratégie cognitive fondamentale. Elle me permet d’apercevoir les différences et les similarités entre des phénomènes complexes (tout est complexe, en fait) et de comprendre ainsi ce que la science cognitive désigne comme « saillance ».

Le genre de saillance sur laquelle je me concentre maintenant sont les traits distinctifs (et donc saillants et importants) de ces deux sociétés – Tesla (https://ir.tesla.com/#tab-quarterly-disclosure ) et Selvita (https://selvita.com/investors-media/ ) – en ce qui concerne les piliers conceptuels de la stratégie de Warren Buffett, qui sont :

>> l’entreprise en tant que telle : le modèle d’entreprise est-il simple et compréhensible ? l’entreprise a-t-elle une histoire cohérente d’exploitation ainsi que des perspectives favorables à long terme ?

>> la gestion : est-ce que la gestion de l’entreprise semble rationnelle ? Les gestionnaires semblent-ils agir dans le meilleur intérêt des actionnaires ? Les gestionnaires résistent-ils les

 modes et les pressions institutionnelles externes ?

>> la finance : quel est le retour sur capitaux propres ? quels sont les bénéfices agrégés pour les actionnaires ? Quelle est la marge de bénéfice dans les produits de l’entreprise ? Quelle est la concordance entre rétention de trésorerie d’une part et l’accroissement de valeur boursière ?   

>> le marché boursier : quelle est la valeur économique de la société en question ? comment cela concorde-t-il avec sa valeur boursière ?

Je me concentre sur la concordance entre la valeur économique de, respectivement, Tesla et Selvita, d’une part et leur valeur boursière d’autre part. Dans la stratégie modèle de Warren Buffett, la valeur économique d’une entreprise est égale au flux prévisible de trésorerie d’activités d’exploitation, escompté avec un taux de retour sur investissement sans risque. Pour donner un exemple pratiqe de cette méthode de base, je cite et traduis un passage du livre « The Warren Buffett Way » par Robert G. Hagstrom, plus précisément le fragment des pages 136 – 137 de l’édition Kindle, où la méthode de Buffett est démontrée dans son achat de Washington Post en 1973. Alors : « Nous commençons par calculer les revenus propriétaires pour l’année fiscale : bénéfice net de $13,3 millions plus dépréciation et amortissement de $3,7 millions moins les investissements capitalisables de $6,6 millions, ça donne un revenu propriétaire de $10,4 millions. Si nous divisons ce revenu par le taux de rendement des obligations souveraines long-terme de la Trésorerie Fédérale des États-Unis (6,81%), la valeur de Washington Post atteint $150 millions […]. Buffett dit qu’avec le temps, les investissements capitalisables d’un journal vont être égales au flux d’amortissement et de ce fait le bénéfice net devrait être une bonne estimation du revenu propriétaire. »

J’applique le même raisonnement à mes deux cas particuliers, donc Tesla et Selvita. Je vais chez https://ir.tesla.com/#tab-quarterly-disclosure et je sélectionne le rapport annuel pour 2020, soit https://www.sec.gov/Archives/edgar/data/1318605/000156459021004599/tsla-10k_20201231.htm. Je vais droit au rapport des flux de trésorerie https://www.sec.gov/Archives/edgar/data/1318605/000156459021004599/tsla-10k_20201231.htm#Consolidated_Statements_of_Cash_Flows .

Le bénéfice net de Tesla pour 2020 était de $862 millions, la charge d’amortissement montait à $2 322 millions, et le solde d’investissements capitalisables fût $3 132. J’obtiens un revenu propriétaire de $52 millions.  J’utilise deux taux de rendement comme référence : celui des obligations souveraines du Trésor Polonais, soit 3,242% (puisque j’investis à partir de Pologne), ainsi que celui des obligations souveraines long-terme de la Trésorerie Fédérale des États-Unis (1,35%), puisque mon résultat sur Tesla est déterminé par la valeur intrinsèque de Tesla telle qu’estimée par le marché financier pour Tesla se trouve aux États-Unis.

Après avoir divisé le revenu propriétaire de Tesla pour 2020 par ces deux taux alternatifs, j’obtiens une fourchette de valeur intrinsèque entre $1 603,95 milliards et $3 851,85 milliards. La capitalisation boursière de Tesla est couramment de $1 019 milliards, mais tout récemment, le 4 novembre, elle atteignait $1 248,43 milliards.

Je répète le même exercice – donc basé sur les résultats financiers pour l’année 2020 – pour Selvita. Je prends leur rapport financier annuel 2020 (https://selvita.com/wp-content/uploads/2021/03/Selvita-Group-Consolidated-Financial-Statements-2020.pdf ) et sur la page 8 je trouve le rapport des flux de trésorerie. Le bénéfice net était de PLN19 921 919, avec les charges d’amortissement de PLN13 525 722.

En ce qui concerne les d’investissements capitalisables, ça se corse. Je trouve un investissement en actifs matériels et immatériels de valeur totale de PLN15 003 636, ainsi que « l’acquisition d’autres actifs financiers » qui monte à PLN10 152 560. Selon la logique de Warren Buffett, l’investissement strictement dit, donc celui qui est déductible du revenu propriétaire, est celui en actifs productifs afférents à l’exploitation. L’acquisition d’actifs financiers est un placement, pas un investissement en actifs d’exploitation.

Je pense donc que je peux calculer le revenu propriétaire de Selvita de deux façons différentes. La première variante c’est « bénéfice net plus amortissement moins l’investissement en actifs matériels et immatériels » et dans la deuxième variante je considère l’acquisition d’actifs financiers comme un flux additionnel de trésorerie et je l’ajoute au solde du premier calcul. J’obtiens ainsi un revenu propriétaire façon Warren Buffett dans la fourchette entre PLN18 444 005 et PLN28 596 565. Je divise par le taux de rendement comme des obligations souveraines du Trésor Polonais (3,242%) et j’obtiens une fourchette correspondante de valeur intrinsèque entre PLN568 908 235 et PLN882 065 545. La dernière capitalisation boursière de Selvita est de PLN1 478 millions, avec un maximum sur les 12 mois derniers noté le 5 juillet 2021, égal à PLN2 894,7 millions.

Ma conclusion provisoire est que, sur la base des résultats financiers audités pour 2020, Tesla reste sous-valorisée par le marché boursier et ça donne des opportunités intéressantes. En revanche, Selvita semble être un peu gonflée en Bourse et je dois être sur mes gardes. Maintenant je passe à l’extension de l’exercice précèdent avec la méthode de MRQ ou « Most Recent Quarter », soit avec les résultats financiers non-audités de deux sociétés pour le troisième quart de 2021. Je fais un truc très primitif, qui est néanmoins utilisé fréquemment en analyse financière, donc j’extrapole les résultats de trois quarts de l’année fiscale en les multipliant par « 4/3 ». Oui, c’est simpliste et ça donne juste une estimation très provisoire de ce que les résultats annuels audités pour 2021 peuvent bien être. Néanmoins, cette méthode permet de simuler l’état d’esprit d’autres investisseurs qui – tout comme moi – utilisent la méthode de valeur intrinsèque façon Warren Buffett.

 Je commence par Tesla, encore une fois  (https://www.sec.gov/Archives/edgar/data/1318605/000095017021002253/tsla-20210930.htm#consolidated_statements_of_cash_flows . Bénéfice net $3 301 millions plus charge d’amortissement $2 063 moins investissements capitalisables de $5 952, ça donne… merde… – $588. Embarrassant, n’est-ce pas ? Revenu propriétaire négatif veut dire valeur intrinsèque négative. Esquive élégante : « Buffett dit qu’avec le temps, les investissements capitalisables d’un journal vont être égales au flux d’amortissement et de ce fait le bénéfice net devrait être une bonne estimation du revenu propriétaire. » Bon, Tesla, c’est presque comme un journal, quoi. Sauf que ça n’a rien à voir. Ce n’est même pas le même type fondamental de bien économique. Enfin, essayons avec l’équivalence « bénéfice net = revenu propriétaire ». J’extrapole le bénéfice net pour les 9 mois de 2021 sur les 12 mois de l’année fiscale et ça donne $4 401,33 millions. Je divise par le taux de rendement des obligations souveraines long-terme de la Trésorerie Fédérale des États-Unis (1,35%) et j’ai $326 024 milliards.

Je commence à comprendre la danse folle autour des actions de Tesla. Vous regardez le bénéfice net et ça a l’air de décoiffer (positivement). Vous jetez un coup d’œil sur les dépenses capitalisables d’investissement et vous commencer à vous poser des questions. Si le calcul très simple de Warren Buffett donne autant de doute, pas étonnant que plusieurs petits investisseurs se laissent prendre au jeu des grands fonds d’investissement futés.

Je passe à Selvita : https://selvita.com/wp-content/uploads/2021/11/Selvita-Group-Consolidated-Financial-Statements-Q3-2021.pdf . Bénéfice net sur les 9 mois de 2021 fait PLN8 844 005, les charges d’amortissement sur la même période montent à PLN17 764 894, acquisition d’actifs productifs est de PLN9 655 884 et celle d’actifs financiers c’est PLN3 172 566. Je somme des deux façons alternatives décrites plus tôt, j’extrapole sur les 12 mois, je divise par le taux d’intérêt sur les obligations de Trésor Polonais (3,242%)  et j’obtiens une valeur intrinsèque entre PLN660 936 257 et PLN784 623 041. C’est toujours très en-dessous de la capitalisation boursière de Selvita. Je dois faire gaffe.

Tesla first in line

Once again, a big gap in my blogging. What do you want – it happens when the academic year kicks in. As it kicks in, I need to divide my attention between scientific research and writing, on the one hand, and my teaching on the other hand.

I feel like taking a few steps back, namely back to the roots of my observation. I observe two essential types of phenomena, as a scientist: technological change, and, contiguously to that, the emergence of previously unexpected states of reality. Well, I guess we all observe the latter, we just sometimes don’t pay attention. I narrow it down a bit. When it comes to technological change, I am just bewildered with the amounts of cash that businesses have started holding, across the board, amidst an accelerating technological race. Twenty years ago, any teacher of economics would tell their students: ‘Guys, cash is the least productive asset of all. Keep just the sufficient cash to face the most immediate expenses. All the rest, invest it in something that makes sense’. Today, when I talk to my students, I tell them: ‘Guys, with the crazy speed of technological change we are observing, cash is king, like really. The greater reserves of cash you hold, the more flexible you stay in your strategy’.

Those abnormally big amounts of cash that businesses tend to hold, those last years, it has two dimensions in terms of research. On the one hand, it is economics and finance, and yet, on the other hand, it is management. For quite some time, digital transformation has been about the only thing worth writing about in management science, but that, namely the crazy accumulation of cash balances in corporate balance sheets, is definitely something worth writing about. Still, there is amazingly little published research on the general topic of cash flow and cash management in business, just as there is very little on financial liquidity in business. The latter topic is developed almost exclusively in the context of banks, mostly the central ones. Maybe it is all that craze about the abominable capitalism and the general claim that money is evil. I don’t know.

Anyway, it is interesting. Money, when handled at the microeconomic level, tells the hell of a story about our behaviour, our values, our mutual trust, and our emotions. Money held in corporate balance sheets tells the hell of a story about decision making. I explain. Please, consider the amount of money you carry around with you, like the contents of your wallet (credit cards included) plus whatever you have available instantly on your phone. Done? Visualised? Good. Now, ask yourself what percentage of all those immediately available monetary balances you use during your one average day. Done? Analysed? Good. In my case, it would be like 0,5%. Yes, 0,5%. I did that intellectual exercise with my students, many time. They usually hit no more than 10%, and they are gobsmacked. Their first reaction is WOKEish: ‘So I don’t really need all that money, right. Money is pointless, right?’. Not quite, my dear students. You need all that money; you just need it in a way which you don’t immediately notice.

There is a model in the theory of complex systems, called the ants’ colony (see for example: (Chaouch, Driss & Ghedira 2017[1]; Asghari & Azadi 2017[2]; Emdadi et al. 2019[3]; Gupta & Srivastava 2020[4]; Di Caprio et al. 2021[5]). Yes, Di Caprio. Not the Di Caprio you intuitively think about, though. Ants communicate with pheromones. They drop pheromones somewhere they sort of know (how?) it is going to be a signal for other ants. Each ant drops sort of a standard parcel of pheromones. Nothing to write home about, really, and yet enough to attract the attention of another ant which could drop its individual pheromonal parcel in the same location. With any luck, other ants will discover those chemical traces and validate them with their individual dumps of pheromones, and this is how the colony of ants maps its territories, mostly to find and exploit sources of food. This is interesting to find out that in order for all that chemical dance to work, there needs to be a minimum number of ants on the job. In there are not enough ants per square meter of territory, they just don’t find each other’s chemical imprints and have no chance to grab hold of the resources available. Yes, they all die prematurely. Money in human societies could be the equivalent of a pheromone. We need to spread it in order to carry out complex systemic changes. Interestingly, each of us, humans, is essentially blind to those complex changes: we just cannot wrap our mind around quickly around the technical details of something apparently as simple as the manufacturing chain of a gardening rake (do you know where exactly and in what specific amounts all the ingredients of steel come from? I don’t).  

All that talk about money made me think about my investments in the stock market. I feel like doing things the Warren Buffet’s way: going to the periodical financial reports of each company in my portfolio, and just passing in review what they do and what they are up to. By the way, talking about Warren Buffet’s way, I recommend my readers to go to the source: go to https://www.berkshirehathaway.com/ first, and then to  https://www.berkshirehathaway.com/2020ar/2020ar.pdf as well as to https://www.berkshirehathaway.com/qtrly/3rdqtr21.pdf . For now, I focus on studying my own portfolio according to the so called “12 immutable tenets by Warren Buffet”, such as I allow myself to quote them:

>> Business Tenets: Is the business simple and understandable? Does the business have a consistent operating history? Does the business have favourable long-term prospects?

>> Management Tenets: Is management rational? Is management candid with its shareholders? Does management resist the institutional imperative?

>> Financial Tenets Focus on return on equity, not earnings per share. Calculate “owner earnings.” Look for companies with high profit margins. For every dollar retained, make sure the company has created at least one dollar of market value.

>> Market Tenets: What is the value of the business? Can the business be purchased at a significant discount to its value?

(Hagstrom, Robert G.. The Warren Buffett Way (p. 98). Wiley. Kindle Edition.)

Anyway, here is my current portfolio:

>> Tesla (https://ir.tesla.com/#tab-quarterly-disclosure),

>> Allegro.eu SA (https://about.allegro.eu/ir-home ),

>> Alten (https://www.alten.com/investors/ ),

>> Altimmune Inc (https://ir.altimmune.com/ ),

>> Apple Inc (https://investor.apple.com/investor-relations/default.aspx ),

>> CureVac NV (https://www.curevac.com/en/investor-relations/overview/ ),

>> Deepmatter Group PLC (https://www.deepmatter.io/investors/ ), 

>> FedEx Corp (https://investors.fedex.com/home/default.aspx ),

>> First Solar Inc (https://investor.firstsolar.com/home/default.aspx )

>> Inpost SA (https://www.inpost.eu/investors )

>> Intellia Therapeutics Inc (https://ir.intelliatx.com/ )

>> Lucid Group Inc (https://ir.lucidmotors.com/ )

>> Mercator Medical SA (https://en.mercatormedical.eu/investors/ )

>> Nucor Corp (https://www.nucor.com/investors/ )

>> Oncolytics Biotech Inc (https://ir.oncolyticsbiotech.com/ )

>> Solaredge Technologies Inc (https://investors.solaredge.com/ )

>> Soligenix Inc (https://ir.soligenix.com/ )

>> Vitalhub Corp (https://www.vitalhub.com/investors )

>> Whirlpool Corp (https://investors.whirlpoolcorp.com/home/default.aspx )

>> Biogened (https://biogened.com/ )

>> Biomaxima (https://www.biomaxima.com/325-investor-relations.html )

>> CyfrPolsat (https://grupapolsatplus.pl/en/investor-relations )

>> Emtasia (https://elemental-asia.biz/en/ )

>> Forposta (http://www.forposta.eu/relacje_inwestorskie/dzialalnosc_i_historia.html )

>> Gameops (http://www.gameops.pl/en/about-us/ )

>> HMInvest (https://grupainwest.pl/relacje )

>> Ifirma (https://www.ifirma.pl/dla-inwestorow )

>> Moderncom (http://moderncommercesa.com/wpmccom/en/dla-inwestorow/ )

>> PolimexMS (https://www.polimex-mostostal.pl/en/reports/raporty-okresowe )

>> Selvita (https://selvita.com/investors-media/ )

>> Swissmed (https://swissmed.com.pl/?menu_id=8 )   

Studying that whole portfolio of mine through the lens of Warren Buffet’s tenets looks like a piece of work, really. Good. I like working. Besides, as I have been reading Warren Buffett’s annual reports at https://www.berkshirehathaway.com/ , I realized that I need a real strategy for investment. So far, I have developed a few efficient hacks, such as, for example, the habit of keeping my s**t together when other people panic or when they get euphoric. Still, hacks are not the same as strategy.

I feel like adding my own general principles to Warren Buffet’s tenets. Principle #1: whatever I think I do my essential strategy consists in running away from what I perceive as danger. Thus, what am I afraid of, in my investment? What subjective fears and objective risks factors shape my actions as investor? Once I understand that, I will know more about my own actions and decisions. Principle #2: the best strategy I can think of is a game with nature, where each move serves to learn something new about the rules of the game, and each move should be both decisive and leaving me with a margin of safety. What am I learning as I make my moves? What my typical moves actually are?

Let’s rock. Tesla (https://ir.tesla.com/#tab-quarterly-disclosure), comes first in line, as it is the biggest single asset in my portfolio. I start my digging with their quarterly financial report for Q3 2021 (https://www.sec.gov/Archives/edgar/data/1318605/000095017021002253/tsla-20210930.htm ), and I fish out their Consolidated Balance Sheets (in millions, except per share data, unaudited: https://www.sec.gov/Archives/edgar/data/1318605/000095017021002253/tsla-20210930.htm#consolidated_balance_sheets ).

Now, I assume that if I can understand why and how numbers change in the financial statements of a business, I can understand the business itself. The first change I can spot in that balance sheet is property, plant and equipment, net passing from $12 747 million to $17 298 million in 12 months. What exactly has happened? Here comes Note 7 – Property, Plant and Equipment, Net, in that quarterly report, and it starts with a specification of fixed assets comprised in that category. Good. What really increased in this category of assets is construction in progress, and here comes the descriptive explanation pertinent thereto: “Construction in progress is primarily comprised of construction of Gigafactory Berlin and Gigafactory Texas, expansion of Gigafactory Shanghai and equipment and tooling related to the manufacturing of our products. We are currently constructing Gigafactory Berlin under conditional permits in anticipation of being granted final permits. Completed assets are transferred to their respective asset classes, and depreciation begins when an asset is ready for its intended use. Interest on outstanding debt is capitalized during periods of significant capital asset construction and amortized over the useful lives of the related assets. During the three and nine months ended September 30, 2021, we capitalized $14 million and $52 million, respectively, of interest. During the three and nine months ended September 30, 2020, we capitalized $13 million and $33 million, respectively, of interest.

Depreciation expense during the three and nine months ended September 30, 2021 was $495 million and $1.38 billion, respectively. Depreciation expense during the three and nine months ended September 30, 2020 was $403 million and $1.13 billion, respectively. Gross property, plant and equipment under finance leases as of September 30, 2021 and December 31, 2020 was $2.60 billion and $2.28 billion, respectively, with accumulated depreciation of $1.11 billion and $816 million, respectively.

Panasonic has partnered with us on Gigafactory Nevada with investments in the production equipment that it uses to manufacture and supply us with battery cells. Under our arrangement with Panasonic, we plan to purchase the full output from their production equipment at negotiated prices. As the terms of the arrangement convey a finance lease under ASC 842, Leases, we account for their production equipment as leased assets when production commences. We account for each lease and any non-lease components associated with that lease as a single lease component for all asset classes, except production equipment classes embedded in supply agreements. This results in us recording the cost of their production equipment within Property, plant and equipment, net, on the consolidated balance sheets with a corresponding liability recorded to debt and finance leases. Depreciation on Panasonic production equipment is computed using the units-of-production method whereby capitalized costs are amortized over the total estimated productive life of the respective assets. As of September 30, 2021 and December 31, 2020, we had cumulatively capitalized costs of $1.89 billion and $1.77 billion, respectively, on the consolidated balance sheets in relation to the production equipment under our Panasonic arrangement.”

Good. I can try to wrap my mind around the contents of Note 7. Tesla is expanding its manufacturing base, including a Gigafactory in my beloved Europe. Expansion of the manufacturing capacity means significant, quantitative growth of the business. According to Warren Buffett’s philosophy: “The question of where to allocate earnings is linked to where that company is in its life cycle. As a company moves through its economic life cycle, its growth rates, sales, earnings, and cash flows change dramatically. In the development stage, a company loses money as it develops products and establishes markets. During the next stage, rapid growth, the company is profitable but growing so fast that it cannot support the growth; often it must not only retain all of its earnings but also borrow money or issue equity to finance growth” (Hagstrom, Robert G.. The Warren Buffett Way (p. 104). Wiley. Kindle Edition).  Tesla looks like they are in the phase of rapid growth. They have finally nailed down how to generate profits (yes, they have!), and they are expanding capacity-wise. They are likely to retain earnings and to be in need of cash, and that attracts my attention to another passage in Note 7: “Interest on outstanding debt is capitalized during periods of significant capital asset construction and amortized over the useful lives of the related assets”. If I understand correctly, the financial strategy consists in not servicing (i.e. not paying the interest due on) outstanding debt when that borrowed money is really being used to finance the construction of productive assets, and starting to service that debt only after the corresponding asset starts working and paying its bills. That means, in turn, that lenders are being patient and confident with Tesla. They assume their unconditional claims on Tesla’s future cash flows (this is one of the possible ways to define outstanding debt) are secure.   

Good. Now, I am having a look at Tesla’s Consolidated Statements of Operations (in millions, except per share data, unaudited: https://www.sec.gov/Archives/edgar/data/1318605/000095017021002253/tsla-20210930.htm#consolidated_statements_of_operations ). It is time to have a look at Warren Buffett’s Business Tenets as regards Tesla. Is the business simple and understandable? Yes, I think I can understand it. Does the business have a consistent operating history? No, operational results changed in 2020 and they keep changing. Tesla is passing from the stage of development (which took them a decade) to the stage of rapid growth. Does the business have favourable long-term prospects? Yes, they seem to have good prospects. The market of electric vehicles is booming (EV-Volumes[6]; IEA[7]).

Is Tesla’s management rational? Well, that’s another ball game. To develop in my next update.


[1] Chaouch, I., Driss, O. B., & Ghedira, K. (2017). A modified ant colony optimization algorithm for the distributed job shop scheduling problem. Procedia computer science, 112, 296-305. https://doi.org/10.1016/j.procs.2017.08.267

[2] Asghari, S., & Azadi, K. (2017). A reliable path between target users and clients in social networks using an inverted ant colony optimization algorithm. Karbala International Journal of Modern Science, 3(3), 143-152. http://dx.doi.org/10.1016/j.kijoms.2017.05.004

[3] Emdadi, A., Moughari, F. A., Meybodi, F. Y., & Eslahchi, C. (2019). A novel algorithm for parameter estimation of Hidden Markov Model inspired by Ant Colony Optimization. Heliyon, 5(3), e01299. https://doi.org/10.1016/j.heliyon.2019.e01299

[4] Gupta, A., & Srivastava, S. (2020). Comparative analysis of ant colony and particle swarm optimization algorithms for distance optimization. Procedia Computer Science, 173, 245-253. https://doi.org/10.1016/j.procs.2020.06.029

[5] Di Caprio, D., Ebrahimnejad, A., Alrezaamiri, H., & Santos-Arteaga, F. J. (2021). A novel ant colony algorithm for solving shortest path problems with fuzzy arc weights. Alexandria Engineering Journal. https://doi.org/10.1016/j.aej.2021.08.058

[6] https://www.ev-volumes.com/

[7] https://www.iea.org/reports/global-ev-outlook-2021/trends-and-developments-in-electric-vehicle-markets

The batteries we don’t need anymore

I continue on the thread I started to develop in my last update in French, titled ‘De quoi parler à la prochaine réunion de faculté’, i.e. I am using that blog, and the fact of writing, to put some order in the almost ritual mess that happens at the beginning of the academic year. New calls for tenders start in the ministerial grant programs, new syllabuses need to be prepared, new classes start. Ordinary stuff, mind you, this is just something about September, as if I were in Vivaldi’s ‘Four seasons’: the hot, tumultuous Summer slowly folds into the rich, textured, and yet implacably realistic Autumn.

My central idea is to use some of the science which I dove into during the summer holidays as an intellectual tool for putting order in that chaos. That almost new science of mine is mostly based on the theory of complex systems, and my basic claim is that technological change is an emergent phenomenon in complex social systems. We don’t know why exactly our technologies change the way they change. We can trace the current technologies back to their most immediate ancestors and sometimes we can predict their most immediate successors, but that’s about it. Futuristic visions of technologies that could be there in 50 years from now are already some kind of traditional entertainment. The concept of technological progress, when we try to find a developmental logic in the historically known technological change, is usually standing on wobbly legs, on the other hand. Yes, electricity allowed the emergence of medical technologies used in hospitals, and that saved a lot of human lives, but there is no way Thomas Edison could have known that. The most spectacular technological achievements of mankind, such as the Egyptian pyramids, the medieval cathedrals, the Dutch windmills from the 16th century, or the automobile, seen from the historical distance, look ambiguous. Yes, it all solved some problems, but it facilitated the emergence of new problems. The truly unequivocal benefit of those technological leaps, which could have been actually experienced by the people who made them, was to learn how to develop technologies.

The studies I did during the Summer holidays 2021 focused on four essential, mathematical models of emergent technological change: cellular automata, flock of birds AKA particle swarm, ants’ nest, and imperfect Markov chains. I start with passing in review the model of cellular automata. At any given moment, the social complexity can be divided into a finite number of social entities (agents). They can be individual humans, businesses, NGOs, governments, local markets etc. Each such entity has an immediate freedom of movement, i.e. a finite number of one-step moves. The concept is related to the theory of games and corresponds to what happens in real life. When we do something social, we seldom just rush forwards. Most frequently, we make one step, observe the outcomes, adjust, then we make the next step etc. When all social agents do it, the whole social complexity can be seen as a collection of cells, or pixels. Each such cell (pixel) is a local state of being in society. A social entity can move into that available state, or not, at their pleasure and leisure. All the one-step moves a social entity can make translate into a trajectory it can follow across the social space. Collective outcomes we strive for and achieve can be studied as temporary complex states of those entities following their respective trajectories. The epistemological trick here is that individual moves and their combinations can be known for sure only ex post. All we can do ex ante is to define the possible states, and just wait where does the reality go.

As we are talking about the possible states of social complexity, I found an interesting mathematical mindf**k at quite an unexpected source, namely in the book titled ‘Aware. The Science and Practice of Presence. The Groundbreaking Meditation Practice’ by Daniel J. Siegel [Penguin Random House LLC, 2018, Identifiers: LCCN 2018016987 (print), LCCN 2018027672 (ebook), ISBN 9780143111788, ISBN 9781101993040 (hardback)]. This is a mathematical way of thinking, apparently taken from quantum physics. Here is the essence of it. Everything that happens does so as 100% probability of the given thing happening. Each phenomenon which takes place is the actualization of the same phenomenon being just likely to happen.

Actualization of probability can be seen as collision of two vehicles in traffic. When the two vehicles are at a substantial distance from each other, the likelihood of them colliding is zero, for all practical purposes. As they converge towards each other, there comes a point when they become sort of provisionally entangled, e.g. they find themselves heading towards the same crossroads. The probability of collision increases slightly, and yet it is not even the probability of collision, it is just the probability that these two might find themselves in a vicinity conducive to a possible collision. Nothing to write home about, yet, like really. It can be seen as a plateau of probability slowly emerging out of the initial soup of all the things which can possibly happen.

As the two cars drive closer and closer to the crossroads in question, the panoply of possible states narrows down. There is a very clear chunk of reality which gains in likelihood, as if it was a mountain range pushing up from the provisional plateau. There comes a point where the two cars (and their drivers) just come on collision course, and there is no way around it, and this is a peak of 100% probability. Boom! Probability is being consumed.

What do those cars have in common with meditation and with the emergence of technological change? As regards meditation, thought can be viewed as a progressively emerging actualization of something that was just a weak probability, sort of a month ago it was just weakly probable that today I would think what I think, it became much more likely yesterday, as the thoughts from yesterday have an impact on the thoughts of today, and today it all comes to fruition, i.e. to the 100% probability. As regards emergent technological change, the way technology changes today can be viewed as actualization of something that was highly probable last year, just somehow probable 10 years ago, and had been just part of the amorphous soup of probability 30 years ago. Those trajectories followed by individual agents inside social complexity, as defined in the theory of cellular automata, are entangled together precisely according to that pattern of emergent probabilities. Two businesses coming up with two mutually independent, and yet similar technologies, are like two peak actualizations of 100% probability in a plateau of probable technological change, which, in turn, has been slowly emerging for some time.

Those other theories I use explain and allow to model mathematically that entanglement. The theory of particle swarm, pertinent to flocks of birds, assumes that autonomous social agents strive for a certain level of behavioural coupling. We expect some level of predictability from others, and we can cooperate with others when we are satisfactorily predictable in our actions. The strive for social coherence is, therefore, one mechanism of entanglement between individual trajectories of cellular automata. The theory of ants’ nest focuses on a specific category of communication systems in societies, working like pheromones. Ants organize by marking, reinforcing and following paths across their environment, and their pheromones serve as markers and reinforcement agents for those paths. In human societies, there are social pheromones. Money and financial markets make probably the most obvious example, but scientific publications are another one. The more scientific articles are being published on a given topic, the more likely are other articles being written on the same topic, until the whole thing reaches a point of saturation, when some ants (pardon me, scientists) start thinking about another path to mark with intellectual pheromones.

Cool. I have (OK, we have) complex social states, made of entangled probabilities that something specific happens, and they encompass technology. Those complex states change, i.e. one complex state morphs into another. Now, how the hell can I know, as a researcher, what is happening exactly? Such as the theory of complex systems suggests it, I can never know exactly, for one, and I need to observe, for two. As I don’t know exactly what is it exactly, that thing which I label ‘technological change’, it is problematic to set too many normative assumptions as for which specific path that technological change should take. I think this is the biggest point of contention as I apply my theory, such as I have just outlined it, to my main field of empirical research, namely energy economics, and technological change in the sector of energy. The more I do that research, the more convinced I am that the so-called ‘energy policies’, ‘climate policies’ etc. are politically driven bullshit based on wishful thinking, with not much of a chance to bring the positive change we expect. I have that deep feeling that setting a strategy for future innovations in our business/country/world is very much like that Polish expression ‘sharing the skin of a bear which is still running in the woods’. First, you need to kill the bear, only then you can bicker about who takes what part of the skin. In the case of innovation, long-term strategies in that domain consist in predicting what we will do when we have something we don’t even know yet what is it exactly.

I am trying to apply this general theory in the grant applications which I am in charge of preparing now, and in my teaching. We have that idea, at the faculty, to apply for funding to study the market of electric vehicles in Europe and in Poland. This is an interesting situation as regards business models. In the US, the market of electric cars is clearly divided among three categories of players. There is Tesla, which is a category and an industry in itself, with its peculiar strategy of extreme vertical integration. Then there are the big, classical car makers, such as Toyota, General Motors etc., with their business models based on rather a short vertical chain of value added inside the business, and a massive supply chain upstream of the house. Finally, there is a rising tide of small start-ups in the making of electric vehicles. I wonder what I could be in Europe. As our European market of electric vehicles is taking off, it is dominated by the incumbent big manufacturers, the old school ones, with Tesla building a factory in Germany, and progressively building a beachhead in the market. There is some timid movement towards small start-up businesses in the field, but it is really timid. In my home country, Poland, the most significant attempt at starting up an electric vehicle made in Poland is a big consortium of state-controlled companies, running under the name of ‘Electromobility Poland’.  

I have that intuition, which I provisionally express as a working hypothesis, namely that business models are an emergent property of technologies which they use. As regards the market of electric vehicles, it means that Tesla’s business model is not an accidental explosion of Elon Musk’s genius mind: it is an emergent characteristic of the technologies involved.

Good. I have some theory taking shape, nice and easy. I let it ripen a bit, and I start sniffing around for facts. What is a business model, in my mind? It is the way of operating the chain of value added, and getting paid for it, in the first place. Then, it is the way of using capital. I noticed that highly innovative environments force businesses to build up and keep large amounts of cash money, arguably to manage the diverse uncertainties emerging as technologies around morph like hell. In some cases, e.g. in biotech, the right business model for rapid innovation is a money-sucker, with apparently endless pay-ins of additional equity by the shareholders, and yet with a big value in terms of technological novelty created. I can associate that phenomenon of vacuum cleaning equity with the case of Tesla, who just recently started being profitable, and had gone through something like a decade in permanent operational loss. That is all pertinent to fixed costs, thus to the cash we need to build up and keep in place the organizational structure required for managing the value chain the way we want to manage it.

I am translating those loose remarks of mine into observable phenomena. Everything I have just mentioned is to be found in the annual financial reports. This is my first source of information. When I want to study business models in the market of electric vehicles, I need to look into financial and corporate reports of businesses active in the market. I need to look into the financial reports of Mercedes Benz, BMW, Renault, PSA, Volkswagen, Fiat, Volvo, and Opel – thus the European automotive makers – and see how it is going, and whether whatever is going on can be correlated with changes in the European market of electric vehicles. Then, it is useful to look into the financial reports of global players present in the European market, e.g. Tesla, Toyota, Honda and whatnot, just to see what changes in them as the European market of electric vehicles is changing.

If my intuition is correct, i.e. if business models are truly an emergent property of technologies used, the fact of engaging into the business of electric vehicles should be correlated with some sort of recurrent pattern in those companies.         

Good. This is about the big boys in the playground. Now, I turn toward the small ones, the start-up businesses. As I already said, it is not like we have a crowd of them in the European industry of electric vehicles. The intuitive axis of research which comes to my mind is to look at start-ups active in the U.S., study their business models, and see if there is any chance of something similar emerging in Europe. Somehow tangentially to that, I think it would be interesting to check whether the plan of Polish government regarding ‘Electromobility Poland’, that is the plan to develop it with public and semi-public money, and then sell it to private investors, has any grounds and under what conditions it can be a workable plan.

Good. I have rummaged a bit in my own mind, time to do the same to other people. I mean, I am passing to reviewing the literature. I type ‘electric vehicles Europe business model’ at the https://www.sciencedirect.com/ platform, and I look at what’s popping up. Here comes the paper by Pardo-Bosch, F., Pujadas, P., Morton, C., & Cervera, C. (2021). Sustainable deployment of an electric vehicle public charging infrastructure network from a city business model perspective. Sustainable Cities and Society, 71, 102957., https://doi.org/10.1016/j.scs.2021.102957 . The abstract says: ‘The unprecedented growth of global cities together with increased population mobility and a heightened concern regarding climate change and energy independence have increased interest in electric vehicles (EVs) as one means to address these challenges. The development of a public charging infrastructure network is a key element for promoting EVs, and with them reducing greenhouse gas emissions attributable to the operation of conventional cars and improving the local environment through reductions in air pollution. This paper discusses the effectiveness, efficiency, and feasibility of city strategic plans for establishing a public charging infrastructure network to encourage the uptake and use of EVs. A holistic analysis based on the Value Creation Ecosystem (VCE) and the City Model Canvas (CMC) is used to visualise how such plans may offer public value with a long-term and sustainable approach. The charging infrastructure network implementation strategy of two major European cities, Nantes (France) and Hamburg (Germany), are analysed and the results indicate the need to involve a wide range of public and private stakeholders in the metropolitan areas. Additionally, relevant, and fundamental patterns and recommendations are provided, which may help other public managers effectively implement this service and scale-up its use and business model.

Well, I see there is a lot of work to do, as I read that abstract. I rarely find a paper where I have so much to argue with, just after having read the abstract. First of all, ‘the unprecedented growth of global cities’ thing. Actually, if you care to have a look at the World Bank data on urban land (https://data.worldbank.org/indicator/AG.LND.TOTL.UR.K2 ), as well as that on urban population (https://data.worldbank.org/indicator/SP.URB.TOTL.IN.ZS ), you will see that urbanization is an ambiguous phenomenon, strongly region-specific. The central thing is that cities become increasingly distinct from the countryside, as types of human settlements. The connection between electric vehicles and cities is partly clear, but just partly. Cities are the most obvious place to start with EVs, because of the relatively short distance to travel between charging points. Still, moving EVs outside the cities, and making them functional in rural areas, is the next big challenge.

Then comes the ‘The development of a public charging infrastructure network is a key element for promoting EVs’ part. As I studied the thing in Europe, the network of charging stations, as compared to the fleet of EVs in the streets is so dense that we have like 12 vehicles per charging station on average, across the European Union. There is no way a private investor can have it for their money, when financing a private charging station, with that average density. We face a paradox: there are so many publicly funded charging stations, in relation to the car fleet out there, that private investment gets discouraged. I agree that it could be an acceptable transitory state in the market, although it begs the question whether private charging stations are a viable business in Europe. Tesla has based a large part of its business model in the US precisely on the development of their own charging stations. Is it a viable solution in Europe?

Here comes another general remark, contingent to my hypothesis of business models being emergent on the basis of technologies. Automotive technologies in general, thus the technology of a vehicle moving by itself, regardless the method of propulsion (i.e. internal combustion vs electric) is a combination of two component technologies. Said method of propulsion is one of them, and the other one is the technology of distributing the power source across space. Electric vehicles can be viewed as cousins to tramways and electric trains, with just more pronounced a taste for independence: instead of drinking electricity from a permanent wiring, EVs carry their electricity around with them, in batteries.

As we talk about batteries, here comes another paper in my cursory rummaging across other people’s science: Albertsen, L., Richter, J. L., Peck, P., Dalhammar, C., & Plepys, A. (2021). Circular business models for electric vehicle lithium-ion batteries: An analysis of current practices of vehicle manufacturers and policies in the EU. Resources, Conservation and Recycling, 172, 105658., https://doi.org/10.1016/j.resconrec.2021.105658 . Yes, indeed, the advent of electric vehicles creates a problem to solve, namely what to do with all those batteries. I mean two categories of batteries. Those which we need, and hope to acquire easily when the time comes for changing them in our vehicles, in the first place, and those we don’t need anymore and expect someone to take care of them swiftly and elegantly.       

I have proven myself wrong

I keep working on a proof-of-concept paper for the idea I baptized ‘Energy Ponds’. You can consult two previous updates, namely ‘We keep going until we observe’ and ‘Ça semble expérimenter toujours’ to keep track of the intellectual drift I am taking. This time, I am focusing on the end of the technological pipeline, namely on the battery-powered charging station for electric cars. First, I want to make myself an idea of the market for charging.

I take the case of France. In December 2020, they had a total of 119 737 electric vehicles officially registered (matriculated), which made + 135% as compared to December 2019[1]. That number pertains only to 100% electrical ones, with plug-in hybrids left aside for the moment. When plug-in hybrids enter the game, France had, in December 2020, 470 295 vehicles that need or might need the services of charging stations. According to the same source, there were 28 928 charging stations in France at the time, which makes 13 EVs per charging station. That coefficient is presented for 4 other European countries: Norway (23 EVs per charging station), UK (12), Germany (9), and Netherlands (4).

I look up into other sources. According to Reuters[2], there was 250 000 charging stations in Europe by September 2020, as compared to 34 000 in 2014. That means an average increase by 36 000 a year. I find a different estimation with Statista[3]: 2010 – 3 201; 2011 – 7 018; 2012 – 17 498; 2013 – 28 824; 2014 – 40 910; 2015 – 67 064; 2016 – 98 669; 2017 – 136 059; 2018 – 153 841; 2019 – 211 438; 2020 – 285 796.

On the other hand, the European Alternative Fuels Observatory supplies their own data at https://www.eafo.eu/electric-vehicle-charging-infrastructure, as regards European Union.

Number of EVs per charging station (source: European Alternative Fuels Observatory):

EVs per charging station
201014
20116
20123
20134
20145
20155
20165
20175
20186
20197
20209

The same EAFO site gives their own estimation as regards the number of charging stations in Europe:

Number of charging stations in Europe (source: European Alternative Fuels Observatory):

High-power recharging points (more than 22 kW) in EUNormal charging stations in EUTotal charging stations
201225710 25010 507
201375117 09317 844
20141 47424 91726 391
20153 39644 78648 182
20165 19070 01275 202
20178 72397 287106 010
201811 138107 446118 584
201915 136148 880164 016
202024 987199 250224 237

Two conclusions jump to the eye. Firstly, there is just a very approximate count of charging stations. Numbers differ substantially from source to source. I can just guess that one of the reasons for that discrepancy is the distinction between officially issued permits to build charging points, on the one hand, and the actually active charging points, on the other hand. In Europe, building charging points for electric vehicles has become sort of a virtue, which governments at all levels like signaling. I guess there is some boasting and chest-puffing in the numbers those individual countries report.  

Secondly, high-power stations, charging with direct current, with a power of at least 22 kWh,  gain in importance. In 2012, that category made 2,45% of the total charging network in Europe, and in 2020 that share climbed to 11,14%. This is an important piece of information as regards the proof-of-concept which I am building up for my idea of Energy Ponds. The charging station I placed at the end of the pipeline in the concept of Energy Ponds, and which is supposed to earn a living for all the technologies and installations upstream of it, is supposed to be powered from a power storage facility. That means direct current, and most likely, high power.   

On the whole, the www.eafo.eu site seems somehow more credible that Statista, with all the due respect for the latter, and thus I am reporting some data they present on the fleet of EVs in Europe. Here it comes, in a few consecutive tables below:

Passenger EVs in Europe (source: European Alternative Fuels Observatory):

BEV (pure electric)PHEV (plug-in-hybrid)Total
20084 1554 155
20094 8414 841
20105 7855 785
201113 39516313 558
201225 8913 71229 603
201345 66232 47478 136
201475 47956 745132 224
2015119 618125 770245 388
2016165 137189 153354 290
2017245 347254 473499 820
2018376 398349 616726 014
2019615 878479 7061 095 584
20201 125 485967 7212 093 206

Light Commercial EVs in Europe (source: European Alternative Fuels Observatory):

BEV (pure electric)PHEV (plug-in-hybrid)Total
2008253253
2009254254
2010309309
20117 6697 669
20129 5279 527
201313 66913 669
201410 04910 049
201528 61028 610
201640 926140 927
201752 026152 027
201876 286176 287
201997 36311797 480
2020120 7111 054121 765

Bus EVs in Europe (source: European Alternative Fuels Observatory):

BEV (pure electric)PHEV (plug-in-hybrid)Total
20082727
20091212
2010123123
2011128128
2012286286
2013376376
201438940429
2015420145565
2016686304990
20178884451 333
20181 6084862 094
20193 6365254 161
20205 3115505 861

Truck EVs in Europe (source: European Alternative Fuels Observatory):

BEV (pure electric)PHEV (plug-in-hybrid)Total
200855
200955
201066
201177
201288
20134747
20145858
20157171
201611339152
2017544094
201822240262
201959538633
2020983291 012

Structure of EV fleet in Europe as regards the types of vehicles (source: European Alternative Fuels Observatory):

Passenger EVLight commercial EVBus EVTruck EV
200893,58%5,70%0,61%0,11%
200994,70%4,97%0,23%0,10%
201092,96%4,97%1,98%0,10%
201163,47%35,90%0,60%0,03%
201275,09%24,17%0,73%0,02%
201384,72%14,82%0,41%0,05%
201492,62%7,04%0,30%0,04%
201589,35%10,42%0,21%0,03%
201689,39%10,33%0,25%0,04%
201790,34%9,40%0,24%0,02%
201890,23%9,48%0,26%0,03%
201991,46%8,14%0,35%0,05%
202094,21%5,48%0,26%0,05%

Summing it up a bit. The market of Electric Vehicles in Europe seems being durably dominated by passenger cars. There is some fleet in other categories of vehicles, and there is even some increase, but, for the moment, in all looks more like an experiment. Well, maybe electric buses turn up sort of more systemically.

The proportion between the fleet of electric vehicles and the infrastructure of charging stations still seems to be in the phase of adjustment in the latter to the abundance of the former. Generally, the number of charging stations seems to be growing slower than the fleet of EVs. Thus, for my own concept, I assume that the coefficient of 9 EVs per charging station, on average, will stand still or will slightly increase. For the moment, I take 9. I assume that my charging stations will have like 9 habitual customers, plus a fringe of incidental ones.

From there, I think in the following terms. The number of times the average customer charges their car depends on the distance they cover. Apparently, there is like a 100 km  50 kWh equivalence. I did not find detailed statistics as regards distances covered by electric vehicles as such, however I came by some Eurostat data on distances covered by all passenger vehicles taken together: https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Passenger_mobility_statistics#Distance_covered . There is a lot of discrepancy between the 11 European countries studied for that metric, but the average is 12,49 km per day. My average 9 customers would do, in total, an average of 410,27 of 50 kWh charging purchases per year. I checked the prices of fast charging with direct current: 2,3 PLN per 1 kWh in Poland[4],  €0,22 per 1 kWh in France[5], $0,13 per 1 kWh in US[6], 0,25 pence per 1 kWh in UK[7]. Once converted to US$, it gives $0,59 in Poland, $0,26 in France, $0,35 in UK, and, of course, $0,13 in US. Even at the highest price, namely that in Poland, those 410,27 charging stops give barely more than $12 000 a year.

If I want to have a station able to charge 2 EVs at the same time, fast charging, and counting 350 kW per charging pile (McKinsey 2018[8]), I need 700 kW it total. Investment in batteries is like $600 ÷ $800 per 1 kW (Cole & Frazier 2019[9]; Cole, Frazier, Augustine 2021[10]), thus 700 * ($600 ÷ $800) = $420 000 ÷ $560 000. There is no way that investment pays back with $12 000 a year in revenue, and I haven’t even started talking about paying off on investment in all the remaining infrastructure of Energy Ponds: ram pumps, elevated tanks, semi-artificial wetlands, and hydroelectric turbines.

Now, I revert my thinking. Investment in the range of $420 000 ÷ $560 000, in the charging station and its batteries, gives a middle-of-the-interval value of $490 000. I found a paper by Zhang et al. (2018[11]) who claim that a charging station has chances to pay off, as a business, when it sells some 5 000 000 kWh a year. When I put it back-to-back with the [50 kWh / 100 km] coefficient, it gives 10 000 000 km. Divided by the average annual distance covered by European drivers, thus by 4 558,55 km, it gives 2 193,68 customers per year, or some 6 charging stops per day. That seems hardly feasible with 9 customers. I assume that one customer would charge their electric vehicle no more than twice a week, and 6 chargings a day make 6*7 = 42 chargings, and therefore 21 customers.

I need to stop and think. Essentially, I have proven myself wrong. I had been assuming that putting a charging station for electric vehicles at the end of the internal value chain in the overall infrastructure of Energy Ponds will solve the problem of making money on selling electricity. Turns out it makes even more problems. I need time to wrap my mind around it.


[1] http://www.avere-france.org/Uploads/Documents/161011498173a9d7b7d55aef7bdda9008a7e50cb38-barometre-des-immatriculations-decembre-2020(9).pdf

[2] https://www.reuters.com/article/us-eu-autos-electric-charging-idUSKBN2C023C

[3] https://www.statista.com/statistics/955443/number-of-electric-vehicle-charging-stations-in-europe/

[4] https://elo.city/news/ile-kosztuje-ladowanie-samochodu-elektrycznego

[5] https://particulier.edf.fr/fr/accueil/guide-energie/electricite/cout-recharge-voiture-electrique.html

[6] https://afdc.energy.gov/fuels/electricity_charging_home.html

[7] https://pod-point.com/guides/driver/cost-of-charging-electric-car

[8] McKinsey Center for Future Mobility, How Battery Storage Can Help Charge the Electric-Vehicle Market?, February 2018,

[9] Cole, Wesley, and A. Will Frazier. 2019. Cost Projections for Utility-Scale Battery Storage.

Golden, CO: National Renewable Energy Laboratory. NREL/TP-6A20-73222. https://www.nrel.gov/docs/fy19osti/73222.pdf

[10] Cole, Wesley, A. Will Frazier, and Chad Augustine. 2021. Cost Projections for UtilityScale Battery Storage: 2021 Update. Golden, CO: National Renewable Energy

Laboratory. NREL/TP-6A20-79236. https://www.nrel.gov/docs/fy21osti/79236.pdf.

[11] Zhang, J., Liu, C., Yuan, R., Li, T., Li, K., Li, B., … & Jiang, Z. (2019). Design scheme for fast charging station for electric vehicles with distributed photovoltaic power generation. Global Energy Interconnection, 2(2), 150-159. https://doi.org/10.1016/j.gloei.2019.07.003

Ça semble expérimenter toujours

Je continue avec l’idée que j’avais baptisée « Projet Aqueduc ». Je suis en train de préparer un article sur ce sujet, du type « démonstration de faisabilité ». Je le prépare en anglais et je me suis dit que c’est une bonne idée de reformuler en français ce que j’ai écrit jusqu’à maintenant, l’histoire de changer l’angle intellectuel, me dégourdir un peu et prendre de la distance.

Une démonstration de faisabilité suit une logique similaire à tout autre article scientifique, sauf qu’au lieu d’explorer et vérifier une hypothèse théorique du type « les choses marchent de façon ABCD, sous conditions RTYU », j’explore et vérifie l’hypothèse qu’un concept pratique, comme celui du « Projet Aqueduc », a des fondements scientifiques suffisamment solides pour que ça vaille la peine de travailler dessus et de le tester en vie réelle. Les fondements scientifiques viennent en deux couches, en quelque sorte. La couche de base consiste à passer en revue la littérature du sujet pour voir si quelqu’un a déjà décrit des solutions similaires et là, le truc c’est explorer des différentes perspectives de similarité. Similaire ne veut pas dire identique, n’est-ce pas ? Cette revue de littérature doit apporter une structure logique – un modèle – applicable à la recherche empirique, avec des variables et des paramètres constants. C’est alors que vient la couche supérieure de démonstration de faisabilité, qui consiste à conduire de la recherche empirique proprement dite avec ce modèle.    

Moi, pour le moment, j’en suis à la couche de base. Je passe donc en revue la littérature pertinente aux solutions hydrologiques et hydroélectriques, tout en formant, progressivement, un modèle numérique du « Projet Aqueduc ». Dans cette mise à jour, je commence par une brève récapitulation du concept et j’enchaîne avec ce que j’ai réussi à trouver dans la littérature. Le concept de base du « Projet Aqueduc » consiste donc à placer dans le cours d’une rivière des pompes qui travaillent selon le principe du bélier hydraulique et qui donc utilisent l’énergie cinétique de l’eau pour pomper une partie de cette eau en dehors du lit de la rivière, vers des structures marécageuses qui ont pour fonction de retenir l’eau dans l’écosystème local. Le bélier hydraulique à la capacité de pomper à la verticale aussi bien qu’à l’horizontale et donc avant d’être retenue dans les marécages, l’eau passe par une structure similaire à un aqueduc élevé (d’où le nom du concept en français), avec des réservoirs d’égalisation de flux, et ensuite elle descend vers les marécages à travers des turbines hydroélectriques. Ces dernières produisent de l’énergie qui est ensuite emmagasinée dans une installation de stockage et de là, elle est vendue pour assurer la survie financière à la structure entière. On peut ajouter des installations éoliennes et/ou photovoltaïques pour optimiser la production de l’énergie sur le terrain occupé par la structure entière.  Vous pouvez trouver une description plus élaborée du concept dans ma mise à jour intitulée « Le Catch 22 dans ce jardin d’Eden ». La faisabilité dont je veux faire une démonstration c’est la capacité de cette structure à se financer entièrement sur la base des ventes d’électricité, comme un business régulier, donc de se développer et durer sans subventions publiques. La solution pratique que je prends en compte très sérieusement en termes de créneau de vente d’électricité est une station de chargement des véhicules électriques.   

L’approche de base que j’utilise dans la démonstration de faisabilité – donc mon modèle de base – consiste à représenter le concept en question comme une chaîne des technologies :

>> TCES – stockage d’énergie

>> TCCS – station de chargement des véhicules électriques

>> TCRP – pompage en bélier hydraulique

>> TCEW – réservoirs élevés d’égalisation

>> TCCW – acheminement et siphonage d’eau

>> TCWS – l’équipement artificiel des structures marécageuses

>> TCHE – les turbines hydroélectriques

>> TCSW – installations éoliennes et photovoltaïques     

Mon intuition de départ, que j’ai l’intention de vérifier dans ma recherche à travers la littérature, est que certaines de ces technologies sont plutôt prévisibles et bien calibrées, pendant qu’il y en a d’autres qui sont plus floues et sujettes au changement, donc moins prévisibles. Les technologies prévisibles sont une sorte d’ancrage pour the concept entier et celles plus floues sont l’objet d’expérimentation.

Je commence la revue de littérature par le contexte environnemental, donc avec l’hydrologie. Les variations au niveau de la nappe phréatiques, qui est un terme scientifique pour les eaux souterraines, semblent être le facteur numéro 1 des anomalies au niveau de rétention d’eau dans les réservoirs artificiels (Neves, Nunes, & Monteiro 2020[1]). D’autre part, même sans modélisation hydrologique détaillée, il y a des preuves empiriques substantielles que la taille des réservoirs naturels et artificiels dans les plaines fluviales, ainsi que la densité de placement de ces réservoirs et ma manière de les exploiter ont une influence majeure sur l’accès pratique à l’eau dans les écosystèmes locaux. Il semble que la taille et la densité des espaces boisés intervient comme un facteur d’égalisation dans l’influence environnementale des réservoirs (Chisola, Van der Laan, & Bristow 2020[2]). Par comparaison aux autres types de technologie, l’hydrologie semble être un peu en arrière en termes de rythme d’innovation et il semble aussi que des méthodes de gestion d’innovation appliquées ailleurs avec succès peuvent marcher pour l’hydrologie, par exemple des réseaux d’innovation ou des incubateurs des technologies (Wehn & Montalvo 2018[3]; Mvulirwenande & Wehn 2020[4]). L’hydrologie rurale et agriculturale semble être plus innovatrice que l’hydrologie urbaine, par ailleurs (Wong, Rogers & Brown 2020[5]).

Ce que je trouve assez surprenant est le manque apparent de consensus scientifique à propos de la quantité d’eau dont les sociétés humaines ont besoin. Toute évaluation à ce sujet commence avec « beaucoup et certainement trop » et à partir de là, le beaucoup et le trop deviennent plutôt flous. J’ai trouvé un seul calcul, pour le moment, chez Hogeboom (2020[6]), qui maintient que la personne moyenne dans les pays développés consomme 3800 litres d’eau par jour au total, mais c’est une estimation très holistique qui inclue la consommation indirecte à travers les biens et les services ainsi que le transport. Ce qui est consommé directement via le robinet et la chasse d’eau dans les toilettes, ça reste un mystère pour la science, apparemment, à moins que la science ne considère ce sujet comment trop terre-à-terre pour s’en occuper sérieusement.     

Il y a un créneau de recherche intéressant, que certains de ses représentants appellent « la socio-hydrologie », qui étudie les comportements collectifs vis-à-vis de l’eau et des systèmes hydrologiques et qui est basée sur l’observation empirique que lesdits comportements collectifs s’adaptent, d’une façon profonde et pernicieuse à la fois, aux conditions hydrologiques que la société en question vit avec (Kumar et al. 2020[7]). Il semble que nous nous adaptons collectivement à la consommation accrue de l’eau par une productivité croissante dans l’exploitation de nos ressources hydrologiques et le revenu moyen par tête d’habitant semble être positivement corrélé avec cette productivité (Bagstad et al. 2020[8]). Il paraît donc que l’accumulation et superposition de nombreuses technologies, caractéristique aux pays développés, contribue à utiliser l’eau de façon de plus en plus productive. Dans ce contexte, il y a une recherche intéressante conduite par Mohamed et al. (2020[9]) qui avance la thèse qu’un environnement aride est non seulement un état hydrologique mais aussi une façon de gérer les ressources hydrologiques, sur ma base des données qui sont toujours incomplètes par rapport à une situation qui change rapidement.

Il y a une question qui vient plus ou moins naturellement : dans la foulée de l’adaptation socio-hydrologique quelqu’un a-t-il présenté un concept similaire à ce que moi je présente comme « Projet Aqueduc » ? Eh bien, je n’ai rien trouvé d’identique, néanmoins il y a des idées intéressement proches. Dans l’hydrologie descriptive il y a ce concept de pseudo-réservoir, qui veut dire une structure comme les marécages ou des nappes phréatiques peu profondes qui ne retiennent pas l’eau de façons statique, comme un lac artificiel, mais qui ralentissent la circulation de l’eau dans le bassin fluvial d’une rivière suffisamment pour modifier les conditions hydrologiques dans l’écosystème (Harvey et al. 2009[10]; Phiri et al. 2021[11]). D’autre part, il y a une équipe des chercheurs australiens qui ont inventé une structure qu’ils appellent par l’acronyme STORES et dont le nom complet est « short-term off-river energy storage » (Lu et al. 2021[12]; Stocks et al. 2021[13]). STORES est une structure semi-artificielle d’accumulation par pompage, où on bâtit un réservoir artificiel au sommet d’un monticule naturel placé à une certaine distance de la rivière la plus proche et ce réservoir reçoit l’eau pompée artificiellement de la rivière. Ces chercheurs australiens avancent et donnent des preuves scientifiques pour appuyer la thèse qu’avec un peu d’astuce on peut faire fonctionner ce réservoir naturel en boucle fermée avec la rivière qui l’alimente et donc de créer un système de rétention d’eau. STORES semble être relativement le plus près de mon concept de « Projet Aqueduc » et ce qui est épatant est que moi, j’avais inventé mon idée pour l’environnement des plaines alluviales de l’Europe tandis que STORES avait été mis au point pour l’environnement aride et quasi-désertique d’Australie. Enfin, il y a l’idée des soi-disant « jardins de pluie » qui sont une technologie de rétention d’eau de pluie dans l’environnement urbain, dans des structures horticulturales, souvent placées sur les toits d’immeubles (Bortolini & Zanin 2019[14], par exemple).

Je peux conclure provisoirement que tout ce qui touche à l’hydrologie strictement dite dans le cadre du « Projet Aqueduc » est sujet aux changements plutôt imprévisible. Ce que j’ai pu déduire de la littérature ressemble à un potage bouillant sous couvercle. Il y a du potentiel pour changement technologique, il y a de la pression environnementale et sociale, mais il n’y pas encore de mécanismes institutionnels récurrents pour connecter l’un à l’autre. Les technologies TCEW (réservoirs élevés d’égalisation), TCCW (acheminement et siphonage d’eau), et TCWS (l’équipement artificiel des structures marécageuses) démontrant donc un avenir flou, je passe à la technologie TCRP de pompage en bélier hydraulique. J’ai trouvé deux articles chinois, qui se suivent chronologiquement et qui semblent par ailleurs avoir été écrits par la même équipe de chercheurs : Guo et al. (2018[15]), and Li et al. (2021[16]). Ils montrent la technologie du bélier hydraulique sous un angle intéressant. D’une part, les Chinois semblent avoir donné du vrai élan à l’innovation dans ce domaine spécifique, tout au moins beaucoup plus d’élan que j’ai pu observer en Europe. D’autre part, les estimations de la hauteur effective à laquelle l’eau peut être pompée avec les béliers hydrauliques dernier cri sont respectivement de 50 mètres dans l’article de 2018 et 30 mètres dans celui de 2021. Vu que les deux articles semblent être le fruit du même projet, il y a eu comme une fascination suivie par une correction vers le bas. Quoi qu’il en soit, même l’estimation plus conservative de 30 mètres c’est nettement mieux que les 20 mètres que j’assumais jusqu’à maintenant.

Cette élévation relative possible à atteindre avec la technologie du bélier hydraulique est importante pour la technologie suivante de ma chaîne, donc celle des petites turbines hydroélectriques, la TCHE. L’élévation relative de l’eau et le flux par seconde sont les deux paramètres clés qui déterminent la puissance électrique produite (Cai, Ye & Gholinia 2020[17]) et il se trouve que dans le « Projet Aqueduc », avec l’élévation et le flux largement contrôlés à travers la technologie du bélier hydraulique, les turbines deviennent un peu moins dépendantes sur les conditions naturelles.

J’ai trouvé une revue merveilleusement encyclopédique des paramètres pertinents aux petites turbines hydroélectriques chez Hatata, El-Saadawi, & Saad (2019[18]). La puissance électrique se calcule donc comme : Puissance = densité de l’eau (1000 kg/m3) * constante d’accélération gravitationnelle (9,8 m/s2) * élévation nette (mètres) * Q (flux par seconde m3/s).

L’investissement initial en de telles installations se calcule par unité de puissance, donc sur la base de 1 kilowatt et se divise en 6 catégories : la construction de la prise d’eau, la centrale électrique strictement dite, les turbines, le générateur, l’équipement auxiliaire, le transformateur et enfin le poste extérieur. Je me dis par ailleurs que – vu la structure du « Projet Aqueduc » – l’investissement en la construction de prise d’eau est en quelque sorte équivalent au système des béliers hydrauliques et réservoirs élevés. En tout cas :

>> la construction de la prise d’eau, par 1 kW de puissance  ($) 186,216 * Puissance-0,2368 * Élévation -0,597

>> la centrale électrique strictement dite, par 1 kW de puissance  ($) 1389,16 * Puissance-0,2351 * Élévation-0,0585

>> les turbines, par 1 kW de puissance  ($)

@ la turbine Kaplan: 39398 * Puissance-0,58338 * Élévation-0,113901

@ la turbine Frances: 30462 * Puissance-0,560135 * Élévation-0,127243

@ la turbine à impulsions radiales: 10486,65 * Puissance-0,3644725 * Élévation-0,281735

@ la turbine Pelton: 2 * la turbine à impulsions radiales

>> le générateur, par 1 kW de puissance  ($) 1179,86 * Puissance-0,1855 * Élévation-0,2083

>> l’équipement auxiliaire, par 1 kW de puissance  ($) 612,87 * Puissance-0,1892 * Élévation-0,2118

>> le transformateur et le poste extérieur, par 1 kW de puissance 

($) 281 * Puissance0,1803 * Élévation-0,2075

Une fois la puissance électrique calculée avec le paramètre d’élévation relative assurée par les béliers hydrauliques, je peux calculer l’investissement initial en hydro-génération comme la somme des positions mentionnées ci-dessus. Hatata, El-Saadawi, & Saad (2019 op. cit.) recommandent aussi de multiplier une telle somme par le facteur de 1,13 (c’est donc un facteur du type « on ne sait jamais ») et d’assumer que les frais courants d’exploitation annuelle vont se situer entre 1% et 6% de l’investissement initial.

Syahputra & Soesanti (2021[19]) étudient le cas de la rivière Progo, dotée d’un flux tout à fait modeste de 6,696 mètres cubes par seconde et située dans Kulon Progo Regency (une region spéciale au sein de Yogyakarta, Indonesia). Le système des petites turbines hydroélectriques y fournit l’électricité aux 962 ménages locaux, et crée un surplus de 4 263 951 kWh par an d’énergie à revendre aux consommateurs externes. Dans un autre article, Sterl et al. (2020[20]) étudient le cas de Suriname et avancent une thèse intéressante, notamment que le développement d’installations basées sur les énergies renouvelables crée un phénomène d’appétit d’énergie qui croît à mesure de manger et qu’un tel développement en une source d’énergie – le vent, par exemple – stimule l’investissement en installations basées sur d’autres sources, donc l’hydraulique et le photovoltaïque.  

Ces études relativement récentes corroborent celles d’il y a quelques années, comme celle de Vilanova & Balestieri (2014[21]) ou bien celle de Vieira et al. (2015[22]), avec une conclusion générale que les petites turbines hydroélectriques ont atteint un degré de sophistication technologique suffisante pour dégager une quantité d’énergie économiquement profitable. Par ailleurs, il semble qu’il y a beaucoup à gagner dans ce domaine à travers l’optimisation de la distribution de puissance entre les turbines différentes. De retour aux publications les plus récentes, j’ai trouvé des études de faisabilité tout à fait robustes pour les petites turbines hydroélectriques, qui indiquent que – pourvu qu’on soit prêt à accepter un retour d’environ 10 à 11 ans sur l’investissement initial – le petit hydro peut être exploité profitablement même avec une élévation relative en dessous de 20 mètres (Arthur et al. 2020[23] ; Ali et al. 2021[24]).

C’est ainsi que j’arrive donc à la portion finale dans la chaîne technologique du « Projet Aqueduc », donc au stockage d’énergie (TCES) ainsi que TCCS ou la station de chargement des véhicules électriques. La puissance à installer dans une station de chargement semble se situer entre 700 et 1000 kilowatts (Zhang et al. 2018[25]; McKinsey 2018[26]). En dessous de 700 kilowatt la station peut devenir si difficile à accéder pour le consommateur moyen, due aux files d’attente, qu’elle peut perdre la confiance des clients locaux. En revanche, tout ce qui va au-dessus de 1000 kilowatts est vraiment utile seulement aux heures de pointe dans des environnements urbains denses. Il y a des études de concept pour les stations de chargement où l’unité de stockage d’énergie est alimentée à partir des sources renouvelables (Al Wahedi & Bicer 2020[27]). Zhang et al. (2019[28]) présentent un concept d’entreprise tout fait pour une station de chargement située dans le milieu urbain. Apparemment, le seuil de profitabilité se situe aux environs de 5 100 000 kilowatt heures vendues par an.  

En termes de technologie de stockage strictement dite, les batteries Li-ion semblent être la solution de base pour maintenant, quoi qu’une combinaison avec les piles à combustible ou bien avec l’hydrogène semble prometteuse (Al Wahedi & Bicer 2020 op. cit. ; Sharma, Panvar & Tripati 2020[29]). En général, pour le moment, les batteries Li-Ion montrent le rythme d’innovation relativement le plus soutenu (Tomaszewska et al. 2019[30] ; de Simone & Piegari 2019[31]; Koohi-Fayegh & Rosen 2020[32]). Un article récent par Elmeligy et al. (2021[33]) présente un concept intéressant d’unité mobile de stockage qui pourrait se déplacer entre plusieurs stations de chargement. Quant à l’investissement initial requis pour une station de chargement, ça semble expérimenter toujours mais la marge de manœuvre se rétrécit pour tomber quelque part entre $600 ÷ $800 par 1 kW de puissance (Cole & Frazier 2019[34]; Cole, Frazier, Augustine 2021[35]).


[1] Neves, M. C., Nunes, L. M., & Monteiro, J. P. (2020). Evaluation of GRACE data for water resource management in Iberia: a case study of groundwater storage monitoring in the Algarve region. Journal of Hydrology: Regional Studies, 32, 100734. https://doi.org/10.1016/j.ejrh.2020.100734

[2] Chisola, M. N., Van der Laan, M., & Bristow, K. L. (2020). A landscape hydrology approach to inform sustainable water resource management under a changing environment. A case study for the Kaleya River Catchment, Zambia. Journal of Hydrology: Regional Studies, 32, 100762. https://doi.org/10.1016/j.ejrh.2020.100762

[3] Wehn, U., & Montalvo, C. (2018). Exploring the dynamics of water innovation: Foundations for water innovation studies. Journal of Cleaner Production, 171, S1-S19. https://doi.org/10.1016/j.jclepro.2017.10.118

[4] Mvulirwenande, S., & Wehn, U. (2020). Fostering water innovation in Africa through virtual incubation: Insights from the Dutch VIA Water programme. Environmental Science & Policy, 114, 119-127. https://doi.org/10.1016/j.envsci.2020.07.025

[5] Wong, T. H., Rogers, B. C., & Brown, R. R. (2020). Transforming cities through water-sensitive principles and practices. One Earth, 3(4), 436-447. https://doi.org/10.1016/j.oneear.2020.09.012

[6] Hogeboom, R. J. (2020). The Water Footprint Concept and Water’s Grand Environmental Challenges. One earth, 2(3), 218-222. https://doi.org/10.1016/j.oneear.2020.02.010

[7] Kumar, P., Avtar, R., Dasgupta, R., Johnson, B. A., Mukherjee, A., Ahsan, M. N., … & Mishra, B. K. (2020). Socio-hydrology: A key approach for adaptation to water scarcity and achieving human well-being in large riverine islands. Progress in Disaster Science, 8, 100134. https://doi.org/10.1016/j.pdisas.2020.100134

[8] Bagstad, K. J., Ancona, Z. H., Hass, J., Glynn, P. D., Wentland, S., Vardon, M., & Fay, J. (2020). Integrating physical and economic data into experimental water accounts for the United States: Lessons and opportunities. Ecosystem Services, 45, 101182. https://doi.org/10.1016/j.ecoser.2020.101182

[9] Mohamed, M. M., El-Shorbagy, W., Kizhisseri, M. I., Chowdhury, R., & McDonald, A. (2020). Evaluation of policy scenarios for water resources planning and management in an arid region. Journal of Hydrology: Regional Studies, 32, 100758. https://doi.org/10.1016/j.ejrh.2020.100758

[10] Harvey, J.W., Schaffranek, R.W., Noe, G.B., Larsen, L.G., Nowacki, D.J., O’Connor, B.L., 2009. Hydroecological factors governing surface water flow on a low-gradient floodplain. Water Resour. Res. 45, W03421, https://doi.org/10.1029/2008WR007129.

[11] Phiri, W. K., Vanzo, D., Banda, K., Nyirenda, E., & Nyambe, I. A. (2021). A pseudo-reservoir concept in SWAT model for the simulation of an alluvial floodplain in a complex tropical river system. Journal of Hydrology: Regional Studies, 33, 100770. https://doi.org/10.1016/j.ejrh.2020.100770.

[12] Lu, B., Blakers, A., Stocks, M., & Do, T. N. (2021). Low-cost, low-emission 100% renewable electricity in Southeast Asia supported by pumped hydro storage. Energy, 121387. https://doi.org/10.1016/j.energy.2021.121387

[13] Stocks, M., Stocks, R., Lu, B., Cheng, C., & Blakers, A. (2021). Global atlas of closed-loop pumped hydro energy storage. Joule, 5(1), 270-284. https://doi.org/10.1016/j.joule.2020.11.015

[14] Bortolini, L., & Zanin, G. (2019). Reprint of: Hydrological behaviour of rain gardens and plant suitability: A study in the Veneto plain (north-eastern Italy) conditions. Urban forestry & urban greening, 37, 74-86. https://doi.org/10.1016/j.ufug.2018.07.003

[15] Guo, X., Li, J., Yang, K., Fu, H., Wang, T., Guo, Y., … & Huang, W. (2018). Optimal design and performance analysis of hydraulic ram pump system. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Puissance and Energy, 232(7), 841-855. https://doi.org/10.1177%2F0957650918756761

[16] Li, J., Yang, K., Guo, X., Huang, W., Wang, T., Guo, Y., & Fu, H. (2021). Structural design and parameter optimization on a waste valve for hydraulic ram pumps. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Puissance and Energy, 235(4), 747–765. https://doi.org/10.1177/0957650920967489

[17] Cai, X., Ye, F., & Gholinia, F. (2020). Application of artificial neural network and Soil and Water Assessment Tools in evaluating Puissance generation of small hydroPuissance stations. Energy Reports, 6, 2106-2118. https://doi.org/10.1016/j.egyr.2020.08.010.

[18] Hatata, A. Y., El-Saadawi, M. M., & Saad, S. (2019). A feasibility study of small hydro Puissance for selected locations in Egypt. Energy Strategy Reviews, 24, 300-313. https://doi.org/10.1016/j.esr.2019.04.013

[19] Syahputra, R., & Soesanti, I. (2021). Renewable energy systems based on micro-hydro and solar photovoltaic for rural areas: A case study in Yogyakarta, Indonesia. Energy Reports, 7, 472-490. https://doi.org/10.1016/j.egyr.2021.01.015

[20] Sterl, S., Donk, P., Willems, P., & Thiery, W. (2020). Turbines of the Caribbean: Decarbonising Suriname’s electricity mix through hydro-supported integration of wind Puissance. Renewable and Sustainable Energy Reviews, 134, 110352. https://doi.org/10.1016/j.rser.2020.110352

[21] Vilanova, M. R. N., & Balestieri, J. A. P. (2014). HydroPuissance recovery in water supply systems: Models and case study. Energy conversion and management, 84, 414-426. https://doi.org/10.1016/j.enconman.2014.04.057

[22] Vieira, D. A. G., Guedes, L. S. M., Lisboa, A. C., & Saldanha, R. R. (2015). Formulations for hydroelectric energy production with optimality conditions. Energy Conversion and Management, 89, 781-788. https://doi.org/10.1016/j.enconman.2014.10.048

[23] Arthur, E., Anyemedu, F. O. K., Gyamfi, C., Asantewaa-Tannor, P., Adjei, K. A., Anornu, G. K., & Odai, S. N. (2020). Potential for small hydroPuissance development in the Lower Pra River Basin, Ghana. Journal of Hydrology: Regional Studies, 32, 100757. https://doi.org/10.1016/j.ejrh.2020.100757

[24] Ali, M., Wazir, R., Imran, K., Ullah, K., Janjua, A. K., Ulasyar, A., … & Guerrero, J. M. (2021). Techno-economic assessment and sustainability impact of hybrid energy systems in Gilgit-Baltistan, Pakistan. Energy Reports, 7, 2546-2562. https://doi.org/10.1016/j.egyr.2021.04.036

[25] Zhang, Y., He, Y., Wang, X., Wang, Y., Fang, C., Xue, H., & Fang, C. (2018). Modeling of fast charging station equipped with energy storage. Global Energy Interconnection, 1(2), 145-152. DOI:10.14171/j.2096-5117.gei.2018.02.006

[26] McKinsey Center for Future Mobility, How Battery Storage Can Help Charge the Electric-Vehicle Market?, February 2018,

[27] Al Wahedi, A., & Bicer, Y. (2020). Development of an off-grid electrical vehicle charging station hybridized with renewables including battery cooling system and multiple energy storage units. Energy Reports, 6, 2006-2021. https://doi.org/10.1016/j.egyr.2020.07.022

[28] Zhang, J., Liu, C., Yuan, R., Li, T., Li, K., Li, B., … & Jiang, Z. (2019). Design scheme for fast charging station for electric vehicles with distributed photovoltaic power generation. Global Energy Interconnection, 2(2), 150-159. https://doi.org/10.1016/j.gloei.2019.07.003

[29] Sharma, S., Panwar, A. K., & Tripathi, M. M. (2020). Storage technologies for electric vehicles. Journal of traffic and transportation engineering (english edition), 7(3), 340-361. https://doi.org/10.1016/j.jtte.2020.04.004

[30] Tomaszewska, A., Chu, Z., Feng, X., O’Kane, S., Liu, X., Chen, J., … & Wu, B. (2019). Lithium-ion battery fast charging: A review. ETransportation, 1, 100011. https://doi.org/10.1016/j.etran.2019.100011

[31] De Simone, D., & Piegari, L. (2019). Integration of stationary batteries for fast charge EV charging stations. Energies, 12(24), 4638. https://doi.org/10.3390/en12244638

[32] Koohi-Fayegh, S., & Rosen, M. A. (2020). A review of energy storage types, applications and recent developments. Journal of Energy Storage, 27, 101047. https://doi.org/10.1016/j.est.2019.101047

[33] Elmeligy, M. M., Shaaban, M. F., Azab, A., Azzouz, M. A., & Mokhtar, M. (2021). A Mobile Energy Storage Unit Serving Multiple EV Charging Stations. Energies, 14(10), 2969. https://doi.org/10.3390/en14102969

[34] Cole, Wesley, and A. Will Frazier. 2019. Cost Projections for Utility-Scale Battery Storage.

Golden, CO: National Renewable Energy Laboratory. NREL/TP-6A20-73222. https://www.nrel.gov/docs/fy19osti/73222.pdf

[35] Cole, Wesley, A. Will Frazier, and Chad Augustine. 2021. Cost Projections for UtilityScale Battery Storage: 2021 Update. Golden, CO: National Renewable Energy

Laboratory. NREL/TP-6A20-79236. https://www.nrel.gov/docs/fy21osti/79236.pdf.

We keep going until we observe

I keep working on a proof-of-concept paper for my idea of ‘Energy Ponds’. In my last two updates, namely in ‘Seasonal lakes’, and in ‘Le Catch 22 dans ce jardin d’Eden’, I sort of refreshed my ideas and set the canvas for painting. Now, I start sketching. What exact concept do I want to prove, and what kind of evidence can possibly confirm (or discard) that concept? The idea I am working on has a few different layers. The most general vision is that of purposefully storing water in spongy structures akin to swamps or wetlands. These can bear various degree of artificial construction, and can stretch from natural wetlands, through semi-artificial ones, all the way to urban technologies such as rain gardens and sponge cities. The most general proof corresponding to that vision is a review of publicly available research – peer-reviewed papers, preprints, databases etc. – on that general topic.

Against that general landscape, I sketch two more specific concepts: the idea of using ram pumps as a technology of forced water retention, and the possibility of locating those wetland structures in the broadly spoken Northern Europe, thus my home region. Correspondingly, I need to provide two streams of scientific proof: a review of literature on the technology of ram pumping, on the one hand, and on the actual natural conditions, as well as land management policies in Europe, on the other hand.  I need to consider the environmental impact of creating new wetland-like structures in Northern Europe, as well as the socio-economic impact, and legal feasibility of conducting such projects.

Next, I sort of build upwards. I hypothesise a complex technology, where ram-pumped water from the river goes into a sort of light elevated tanks, and from there, using the principle of Roman siphon, cascades down into wetlands, and through a series of small hydro-electric turbines. Turbines generate electricity, which is being stored and then sold outside.

At that point, I have a technology of water retention coupled with a technology of energy generation and storage. I further advance a second hypothesis that such a complex technology will be economically sustainable based on the corresponding sales of electricity. In other words, I want to figure out a configuration of that technology, which will be suitable for communities which either don’t care at all, or simply cannot afford to care about the positive environmental impact of the solution proposed.

Proof of concept for those two hypotheses is going to be complex. First, I need to pass in review the available technologies for energy storage, energy generation, as well as for the construction of elevated tanks and Roman siphons. I need to take into account various technological mixes, including the incorporation of wind turbines and photovoltaic installation into the whole thing, in order to optimize the output of energy. I will try to look for documented examples of small hydro-generation coupled with wind and solar. Then, I have to rack the literature as regards mathematical models for the optimization of such power systems and put them against my own idea of reverse engineering back from the storage technology. I take the technology of energy storage which seems the most suitable for the local market of energy, and for the hypothetical charging from hydro-wind-solar mixed generation. I build a control scenario where that storage facility just buys energy at wholesale prices from the power grid and then resells it. Next, I configure the hydro-wind-solar generation so as to make it economically competitive against the supply of energy from the power grid.

Now, I sketch. I keep in mind the levels of conceptualization outlined above, and I quickly move through published science along that logical path, quickly picking a few articles for each topic. I am going to put those nonchalantly collected pieces of science back-to-back and see how and whether at all it all makes sense together. I start with Bortolini & Zanin (2019[1]), who study the impact of rain gardens on water management in cities of the Veneto region in Italy. Rain gardens are vegetal structures, set up in the urban environment, with the specific purpose to retain rainwater.  Bortolini & Zanin (2019 op. cit.) use a simplified water balance, where the rain garden absorbs and retains a volume ‘I’ of water (‘I’ stands for infiltration), which is the difference between precipitations on the one hand, and the sum total of overflowing runoff from the rain garden plus evapotranspiration of water, on the other hand. Soil and plants in the rain garden have a given top capacity to retain water. Green plants typically hold 80 – 95% of their mass in water, whilst trees hold about 50%. Soil is considered wet when it contains about 25% of water. The rain garden absorbs water from precipitations at a rate determined by hydraulic conductivity, which means the relative ease of a fluid (usually water) to move through pore spaces or fractures, and which depends on the intrinsic permeability of the material, the degree of saturation, and on the density and viscosity of the fluid.

As I look at it, I can see that the actual capacity of water retention in a rain garden can hardly be determined a priori, unless we have really a lot of empirical data from the given location. For a new location of a new rain garden, it is safe to assume that we need an experimental phase when we empirically assess the retentive capacity of the rain garden with different configurations of soil and vegetation used. That leads me to generalizing that any porous structure we use for retaining rainwater, would it be something like wetlands, or something like a rain garden in urban environment, has a natural constraint of hydraulic conductivity, and that constraint determines the percentage of precipitations, and the metric volume thereof, which the given structure can retain.

Bortolini & Zanin (2019 op. cit.) bring forth empirical results which suggest that properly designed rain gardens located on rooftops in a city can absorb from 87% to 93% of the total input of water they receive. Cool. I move on and towards the issue of water management in Europe, with a working paper by Fribourg-Blanc, B. (2018[2]), and the most important takeaway from that paper is that we have something called European Platform for Natural Water Retention Measures AKA http://nwrm.eu , and that thing have both good properties and bad properties. The good thing about http://nwrm.eu is that it contains loads of data and publications about projects in Natural Water Retention in Europe. The bad thing is that http://nwrm.eu is not a secure website. Another paper, by Tóth et al. (2017[3]) tells me that another analytical tool exists, namely the European Soil Hydraulic Database (EU‐ SoilHydroGrids ver1.0).

So far, so good. I already know there is data and science for evaluating, with acceptable precision, the optimal structure and the capacity for water retention in porous structures such as rain gardens or wetlands, in the European context. I move to the technology of ram pumps. I grab two papers: Guo et al. (2018[4]), and Li et al. (2021[5]). They show me two important things. Firstly, China seems to be burning the rubber in the field of ram pumping technology. Secondly, the greatest uncertainty as for that technology seems to be the actual height those ram pumps can elevate water at, or, when coupled with hydropower, the hydraulic head which ram pumps can create. Guo et al. (2018 op. cit.) claim that 50 meters of elevation is the maximum which is both feasible and efficient. Li et al. (2021 op. cit.) are sort of vertically more conservative and claim that the whole thing should be kept below 30 meters of elevation. Both are better than 20 meters, which is what I thought was the best one can expect. Greater elevation of water means greater hydraulic head, and more hydropower to be generated. It pays off to review literature.

Lots of uncertainty as for the actual capacity and efficiency of ram pumping means quick technological change in that domain. This is economically interesting. It means that investing in projects which involve ram pumping means investment in quickly changing a technology. That means both high hopes for an even better technology in immediate future, and high needs for cash in the balance sheet of the entities involved.

I move to the end-of-the-pipeline technology in my concept, namely to energy storage. I study a paper by Koohi-Fayegh & Rosen (2020[6]), which suggests two things. Firstly, for a standalone installation in renewable energy, whatever combination of small hydropower, photovoltaic and small wind turbines we think of, lithium-ion batteries are always a good idea for power storage, Secondly, when we work with hydrogeneration, thus when we have any hydraulic head to make electricity with, pumped storage comes sort of natural. That leads me to an idea which looks even crazier than what I have imagined so far: what if we create an elevated garden with strong capacity for water retention. Ram pumps take water from the river and pump it up onto elevated platforms with rain gardens on it. Those platforms can be optimized as for their absorption of sunlight and thus as regards their interaction with whatever is underneath them.  

I move to small hydro, and I find two papers, namely Couto & Olden (2018[7]), and Lange et al. (2018[8]), which are both interestingly critical as regards small hydropower installations. Lange et al. (2018 op. cit.) claim that the overall environmental impact of small hydro should be closely monitored. Couto & Olden (2018 op. cit.) go further and claim there is a ‘craze’ about small hydro, and that craze has already lead to overinvestment in the corresponding installations, which can be damaging both environmentally and economically (overinvestment means financial collapse of many projects). Those critical views in mind, I turn to another paper, by Zhou et al. (2019[9]), who approach the issue as a case for optimization, within a broader framework called ‘Water-Food-Energy’ Nexus, WFE for closer friends. This paper, just as a few others it cites (Ming et al. 2018[10]; Uen et al. 2018[11]), advocates for using artificial intelligence in order to optimize for WFE.

Zhou et al. (2019 op.cit.) set three hydrological scenarios for empirical research and simulation. The baseline scenario corresponds to an average hydrological year, with average water levels and average precipitations. Next to it are: a dry year and a wet year. The authors assume that the cost of installation in small hydropower is $600 per kW on average.  They simulate the use of two technologies for hydro-electric turbines: Pelton and Vortex. Pelton turbines are optimized paddled wheels, essentially, whilst the Vortex technology consists in creating, precisely, a vortex of water, and that vortex moves a rotor placed in the middle of it.

Zhou et al. (2019 op.cit.) create a multi-objective function to optimize, with the following desired outcomes:

>> Objective 1: maximize the reliability of water supply by minimizing the probability of real water shortage occurring.

>> Objective 2: maximize water storage given the capacity of the reservoir. Note: reservoir is understood hydrologically, as any structure, natural or artificial, able to retain water.

>> Objective 3: maximize the average annual output of small hydro-electric turbines

Those objectives are being achieved under the corresponding sets of constraints. For water supply those constraints all turn around water balance, whilst for energy output it is more about the engineering properties of the technologies taken into account. The three objectives are hierarchized. First, Zhou et al. (2019 op.cit.) perform an optimization regarding Objectives 1 and 2, thus in order to find the optimal hydrological characteristics to meet, and then, on the basis of these, they optimize the technology to put in place, as regards power output.

The general tool for optimization used by Zhou et al. (2019 op.cit.) is a genetic algorithm called NSGA-II, AKA Non-dominated Sorting Genetic Algorithm. Apparently, NSGA-II has a long and successful history of good track in engineering, including water management and energy (see e.g. Chang et al. 2016[12]; Jain & Sachdeva 2017[13];  Assaf & Shabani 2018[14]). I want to stop for a while here and have a good look at this specific algorithm. The logic of NSGA-II starts with creating an initial population of cases/situations/configurations etc. Each case is a combination of observations as regards the objectives to meet, and the actual values observed in constraining variables, e.g. precipitations for water balance or hydraulic head for the output of hydropower. In the conventional lingo of this algorithm, those cases are called chromosomes. Yes, I know, a hydro-electric turbine placed in the context of water management hardly looks like a chromosome, but it is a genetic algorithm, and it just sounds fancy to use that biologically marked vocabulary.

As for me, I like staying close to real life, and therefore I call those cases solutions rather than chromosomes. Anyway, the underlying math is the same. Once I have that initial population of real-life solutions, I calculate two parameters for each of them: their rank as regards the objectives to maximize, and their so-called ‘crowded distance’. Ranking is done with the procedure of fast non-dominated sorting. It is a comparison in pairs, where the solution A dominates another solution B, if and only if there is no objective of A worse than that objective of B and there is at least one objective of A better than that objective of B. The solution which scores the most wins in such peer-to-peer comparisons is at the top of the ranking, the one with the second score of wins is the second etc. Crowding distance is essentially the same as what I call coefficient of coherence in my own research: Euclidean distance (or other mathematical distance) is calculated for each pair of solutions. As a result, each solution is associated with k Euclidean distances to the k remaining solutions, which can be reduced to an average distance, i.e. the crowded distance.

In the next step, an off-spring population is produced from that original population of solutions. It is created by taking relatively the fittest solutions from the initial population, recombining their characteristics in a 50/50 proportion, and adding them some capacity for endogenous mutation. Two out of these three genetic functions are de facto controlled. We choose relatively the fittest by establishing some kind of threshold for fitness, as regards the objectives pursued. It can be a required minimum, a quantile (e.g. the third quartile), or an average. In the first case, we arbitrarily impose a scale of fitness on our population, whilst in the latter two the hierarchy of fitness is generated endogenously from the population of solutions observed. Fitness can have shades and grades, by weighing the score in non-dominated sorting, thus the number of wins over other solutions, on the one hand, and the crowded distance on the other hand. In other words, we can go for solutions which have a lot of similar ones in the population (i.e. which have a low average crowded distance), or, conversely, we can privilege lone wolves, with a high average Euclidean distance from anything else on the plate.  

The capacity for endogenous mutation means that we can allow variance in all or in just the selected variables which make each solution. The number of degrees of freedom we allow in each variable dictates the number of mutations that can be created. Once again, discreet power is given to the analyst: we can choose the genetic traits which can mutate and we can determine their freedom to mutate. In an engineering problem, technological and environmental constraints should normally put a cap on the capacity for mutation. Still, we can think about an algorithm which definitely kicks the lid off the barrel of reality, and which generates mutations in the wildest registers of variables considered. It is a way to simulate a process when the presence of strong outliers has a strong impact on the whole population.

The same discreet cap on the freedom to evolve is to be found when we repeat the process. The offspring generation of solutions goes essentially through the same process as the initial one, to produce further offspring: ranking by non-dominated sorting and crowded distance, selection of the fittest, recombination, and endogenous mutation. At the starting point of this process, we can be two alternative versions of the Mother Nature. We can be a mean Mother Nature, and we shave off from the offspring population all those baby-solutions which do not meet the initial constraints, e.g. zero supply of water in this specific case. On the other hand, we can be even meaner a Mother Nature and allow those strange, dysfunctional mutants to keep going and see what happens to the whole species after a few rounds of genetic reproduction.

With each generation, we compute an average crowded distance between all the solutions created, i.e. we check how diverse is the species in this generation. As long as diversity grows or remains constant, we assume that the divergence between the solutions generated grows or stays the same. Similarly, we can compute an even more general crowded distance between each pair of generations, and therefore to assess how far has the current generation gone from the parent one. We keep going until we observe that the intra-generational crowded distance and the inter-generational one start narrowing down asymptotically to zero. In other words, we consider resuming evolution when solutions in the game become highly similar to each other and when genetic change stops bringing significant functional change.

Cool. When I want to optimize my concept of Energy Ponds, I need to add the objective of constrained return on investment, based on the sales of electricity. In comparison to Zhou et al. (2019 op.cit.), I need to add a third level of selection. I start with selecting environmentally the solutions which make sense in terms of water management. In the next step, I produce a range of solutions which assure the greatest output of power, in a possible mix with solar and wind. Then I take those and filter them through the NSGA-II procedure as regards their capacity to sustain themselves financially. Mind you, I can shake it off a bit by fusing together those levels of selection. I can simulate extreme cases, when, for example, good economic sustainability becomes an environmental problem. Still, it would be rather theoretical. In Europe, non-compliance with environmental requirements makes a project a non-starter per se: you just can get the necessary permits if your hydropower project messes with hydrological constraints legally imposed on the given location.     

Cool. It all starts making sense. There is apparently a lot of stir in the technology of making semi-artificial structures for retaining water, such as rain gardens and wetlands. That means a lot of experimentation, and that experimentation can be guided and optimized by testing the fitness of alternative solutions for meeting objectives of water management, power output and economic sustainability. I have some starting data, to produce the initial generation of solutions, and then try to optimize them with an algorithm such as NSGA-II.


[1] Bortolini, L., & Zanin, G. (2019). Reprint of: Hydrological behaviour of rain gardens and plant suitability: A study in the Veneto plain (north-eastern Italy) conditions. Urban forestry & urban greening, 37, 74-86. https://doi.org/10.1016/j.ufug.2018.07.003

[2] Fribourg-Blanc, B. (2018, April). Natural Water Retention Measures (NWRM), a tool to manage hydrological issues in Europe?. In EGU General Assembly Conference Abstracts (p. 19043). https://ui.adsabs.harvard.edu/abs/2018EGUGA..2019043F/abstract

[3] Tóth, B., Weynants, M., Pásztor, L., & Hengl, T. (2017). 3D soil hydraulic database of Europe at 250 m resolution. Hydrological Processes, 31(14), 2662-2666. https://onlinelibrary.wiley.com/doi/pdf/10.1002/hyp.11203

[4] Guo, X., Li, J., Yang, K., Fu, H., Wang, T., Guo, Y., … & Huang, W. (2018). Optimal design and performance analysis of hydraulic ram pump system. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy, 232(7), 841-855. https://doi.org/10.1177%2F0957650918756761

[5] Li, J., Yang, K., Guo, X., Huang, W., Wang, T., Guo, Y., & Fu, H. (2021). Structural design and parameter optimization on a waste valve for hydraulic ram pumps. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy, 235(4), 747–765. https://doi.org/10.1177/0957650920967489

[6] Koohi-Fayegh, S., & Rosen, M. A. (2020). A review of energy storage types, applications and recent developments. Journal of Energy Storage, 27, 101047. https://doi.org/10.1016/j.est.2019.101047

[7] Couto, T. B., & Olden, J. D. (2018). Global proliferation of small hydropower plants–science and policy. Frontiers in Ecology and the Environment, 16(2), 91-100. https://doi.org/10.1002/fee.1746

[8] Lange, K., Meier, P., Trautwein, C., Schmid, M., Robinson, C. T., Weber, C., & Brodersen, J. (2018). Basin‐scale effects of small hydropower on biodiversity dynamics. Frontiers in Ecology and the Environment, 16(7), 397-404.  https://doi.org/10.1002/fee.1823

[9] Zhou, Y., Chang, L. C., Uen, T. S., Guo, S., Xu, C. Y., & Chang, F. J. (2019). Prospect for small-hydropower installation settled upon optimal water allocation: An action to stimulate synergies of water-food-energy nexus. Applied Energy, 238, 668-682. https://doi.org/10.1016/j.apenergy.2019.01.069

[10] Ming, B., Liu, P., Cheng, L., Zhou, Y., & Wang, X. (2018). Optimal daily generation scheduling of large hydro–photovoltaic hybrid power plants. Energy Conversion and Management, 171, 528-540. https://doi.org/10.1016/j.enconman.2018.06.001

[11] Uen, T. S., Chang, F. J., Zhou, Y., & Tsai, W. P. (2018). Exploring synergistic benefits of Water-Food-Energy Nexus through multi-objective reservoir optimization schemes. Science of the Total Environment, 633, 341-351. https://doi.org/10.1016/j.scitotenv.2018.03.172

[12] Chang, F. J., Wang, Y. C., & Tsai, W. P. (2016). Modelling intelligent water resources allocation for multi-users. Water resources management, 30(4), 1395-1413. https://doi.org/10.1007/s11269-016-1229-6

[13] Jain, V., & Sachdeva, G. (2017). Energy, exergy, economic (3E) analyses and multi-objective optimization of vapor absorption heat transformer using NSGA-II technique. Energy Conversion and Management, 148, 1096-1113. https://doi.org/10.1016/j.enconman.2017.06.055

[14] Assaf, J., & Shabani, B. (2018). Multi-objective sizing optimisation of a solar-thermal system integrated with a solar-hydrogen combined heat and power system, using genetic algorithm. Energy Conversion and Management, 164, 518-532. https://doi.org/10.1016/j.enconman.2018.03.026