Nuclear vs renewable energy and the critical importance of independent research

This is an adapted version of a blog by Prof Benjamin K. Sovacool and Prof Andy Stirling, to accompany the publication of their paper “Differences in carbon emissions reduction between countries pursuing renewable electricity versus nuclear power” in Nature Energy. A University of Sussex press release also summarises the paper’s findings and policy recommendations.

The role of nuclear power in a low-carbon future has been subject to a long and contentious debate. Is a nuclear or a renewables pathway the best way forward, or do we need a “do everything” approach where every deployable technology is rolled out to decarbonise our electricity supply as soon as possible?

Many influential climate scientists and international organisations argue that a global shift towards nuclear power offers the best pathway to tackling the climate emergency and meeting the world’s increasing demands for electricity.

Others argue that renewable sources of energy are the best pathway towards a low-carbon electricity system and assert that they are cleaner, safer and more economically sustainable than nuclear.

In an attempt to negotiate these contending positions, a frequent mantra is that energy strategies should “do everything” in order to address the climate emergency. But – as a number of commentators have noted (for example, here and here) – this would actually be a highly irrational course of action.

Where “doing everything” involves making investments that are slower or less cost effective, which divert resources away from preferable options, or which in some other way impede them, the result would be potentially disastrous for carbon emissions mitigation.

Amidst many uncertainties, the real questions we should be addressing are about which investments offer the most cost-effective and beneficial ways forward.

Our new paper, Differences in carbon emissions reduction between countries pursuing renewable electricity versus nuclear power, seeks to contribute towards this debate.

Nuclear vs renewable energy – what this paper tells us

Our paper focuses specifically on situations in which real-world constraints mean strategic choices must be made on resource allocation between nuclear or renewables-based electricity.

Our research explores this dilemma retrospectively, examining past patterns in the attachments (i.e. investments) of different countries to nuclear or renewable strategies. Our paper addresses three hypotheses:

  1. A “nuclear climate mitigation” hypothesis: that countries with a greater attachment to nuclear power will tend to have lower overall carbon emissions.
  2.  A “renewables climate mitigation” hypothesis: that countries with a greater attachment to renewables will tend to have lower overall carbon emissions.
  3. A “crowding out” hypothesis: that countries with a greater attachment to nuclear will tend to have a lesser attachment to renewables, and vice versa

Across the study countries as a whole we found that the “nuclear climate mitigation” hypothesis is not sustained by the evidence at an appropriate level of statistical significance. The renewable climate mitigation hypothesis is confirmed with substantial significance. And the crowding out hypothesis is also significantly sustained.

Put plainly – if countries want to lower emissions as substantially, rapidly and cost-effectively as possible, they should prioritise support for renewables rather than nuclear power. Pursuit of nuclear strategies risks taking up resources that could be used more effectively and suppressing the uptake of renewable energy.

A windmill next to a body of water

Description automatically generated

What causes these patterns?

What might explain these patterns? Technologically, nuclear systems have been prone to greater construction cost overruns, delays, and longer lead times than similarly sized renewable energy projects. Thus, per dollar invested, the modularity of renewables projects offers quicker emissions reductions than large-scale, delay-prone, nuclear projects.

Furthermore, renewables tend to display higher rates of positive learning where increased deployment results in lower costs and improved performance, especially for wind farms and solar energy parks. This contrasts with the experience of nuclear power in France which has been prone to negative learning,” rising costs or reduced performance with the next generation of technology.

In terms of policy, the incidents at Three Mile Island (1979), Chernobyl (1986), and Fukushima (2011), all resulted in significant tightening of regulatory requirements for nuclear reactors.

Finally, wider social factors may also work against nuclear energy, and for renewable energy, facilitating faster acceptance, permitting and deployment.

Of course, these are just informed speculations, beyond the scope of the paper itself. Other commentators will favor contrasting interpretations.

But here, perhaps the most important issue – especially given the prominence of the topic and the scale of what is at stake – is that this kind of analysis has been so remarkably neglected over recent years.

Given how highly charged and hotly contested the associated policy controversy is, it is rather strange that there is not a large body of work on these questions. Either way, the many open questions and issues of detail acknowledged in the paper show that much work remains to be done.

The critical importance of independent research – our view

We have presented the findings of our research. Now we must acknowledge the uncertainties and errors, divergent interpretations and clashing interests that make it difficult to achieve the comprehensive prioritising analysis called for at the beginning of this blog – while making a case for the vital importance of scientific scrutiny.

In an ideal world of “evidence-based policy”, energy and climate policy would only go ahead after comprehensive research into every relevant positive or negative aspect of all possible energy resources.

The resulting self-evident “facts” would be examined by objective analysts and any uncertainties eliminated, until a point where a single unambiguous ‘truth’ is determined – with grateful policy makers adopting the identified energy pathway or portfolio.

Unfortunately, we do not live in an ideal world.

Across various energy debates – and not restricted to any political constituency – crucial roles are often played by deliberate mis-representation of information, manipulation of discourse, co-option of leading opposing voices, direct subversion of opponents and stifling of meaningful public debate. 

Under conditions like this, the line between advocacy and scholarship (porous at the best of times) can become especially loose when analysts become passionate about their topic. The reasons for such passion can be as trivial as disciplinary identities or sectoral interests, or as deep as wider political ideologies. On all sides “theorising” can be reduced to a search for validation, and “investigation” to the selective collection of data. 

Energy debates suffer gravely from these syndromes. “Energy evangelists” on all sides are convinced they have found “the solution” to societies’ energy problems—whether this be solar energy, hydrogen fuel cells or nuclear reactors. The intensity of this advocacy (and the scale of the interests often behind it) can lead to everyone else’s solutions being treated as sacrilegious.

So, exchanges of ideas can become hostile battlefields where proponents are unable to reconcile their underlying differences.

There seems to be an especially pernicious asymmetry in this field. Those whom comparative analysis leads to be generally critical of nuclear power are labelled “anti-nuclear”, whilst no such generally-established terminology exists to the same degree for those who are (entirely legitimately – if debatably) critical of renewable energy.

The situation is aggravated by so much research in this field being (unlike our own) funded (directly or indirectly) by organisations with prior entrenched interests on one side or another.

Despite this, we have often found valued opportunities to bridge the divide with those who hold “opposing” views, but with similar open mindedness and good faith.

It is in this spirit that our analysis is offered. We are open about its background and limitations. We acknowledge that our evidence does not compel only one supposedly definitive interpretation. We are clear about the conditions attached to our own interpretations. By publishing our full dataset and the detailed procedures undertaken in our regression analyses, we offer a basis for others to contest our findings.

The “truth” of our study is in this sense not something arrived at by particular analysts claiming individually-transcendent authority, but by contrastingly-oriented analysts contending with each other in an open and pluralistic way, such as to arrive collectively at more robust understandings. This is the organised skepticism of independent science.

If our analysis stimulates reactions in the same vein, then the cause of scientific scrutiny is reinforced. If, on the other hand, it leads to less qualified assertions and ad hominen labelling, then the chance of bridging the polarised divides is sadly diminished. We hope it will do the former.

Follow Sussex Energy Group Facebooktwitterlinkedin
Tagged with: , ,
Posted in All Posts
13 comments on “Nuclear vs renewable energy and the critical importance of independent research
  1. Colin Megson says:

    “…nuclear systems have been prone to greater construction cost overruns, delays, and longer lead times than similarly sized renewable energy projects…”

    GE Hitachi’s BWRX-300 Small Modular Reactor (SMR) changes all of that. It has a 2-year build programme – no different to Wind And Solar Plants (WASPs). GE Hitachi openly proclaim on their BWRX-300 webpage, a capital cost of US$2,250/kW for the NOAK. The FOAK will be operational in 2027 and should be available for manufacture under licence in the UK by 2030, for a capital investment of £587 million.

    For 3,200 MW, the capital investment required would be £5.6 billion, that’s 72% below Sizewell C’s 2 x 1600 MW EPRs, at £20 billion. And it’s easy to see why, if you have a bit of engineering nous.

    These pseudo-green fund managers will be clawing at one another’s throats to get their pots out of WASPs and into BWRX-300 NPPs. When the BWRX-300 is up and running, it will be the beginning of the end of the insane decades of our energy-inept politicians being suckered into gambling on get-rich-quick WASP technologies by obfuscating lobbyists, like renewableUK, not-of-the-real-world academics and other slick individuals.

    Government ‘money’ is not necessary. Commercial investors can earn many times more for every £1.00 invested in advanced Nuclear Power Plants (NPPs) than £1.00 invested in WASPs:
    Onshore Wind – 7.2X more. Search for: bwrx-300 blogspot. Scroll to Sunday 24 May 2020.

    Offshore Wind – 12X more. Search for: bwrx-300 blogspot. Scroll to Sunday 10 May 2020.

    Solar PV – 15.5X more. Search for: bwrx-300 blogspot. Scroll to Sunday 17 May 2020.

    • Andy Stirling says:

      Thank you Colin for this detailed account of your views on one particular example among what are many different prospective candidates for an entirely new (as yet largely untested) generation of nuclear power reactors. There is much wider discussion of the history, pros and cons of these kinds of claims. But in short, our article does not set out to address this issue. We are simply looking at empirical patterns over the past 25 years.

  2. This post talks about the importance of independent review, which I think is laudable. I think sharing one’s data and information about what one has done in their analyses should be a basic requirement of all publication. That said, is it possible for people to obtain a copy of the article itself in a similar fashion?

    I know paywalls are a part of life for academia, but they can be a painful hurdle for many people who might be interested in learning about new work.

    • Andy Stirling says:

      Thank you Brandon for making this very relevant point, with which I entirely agree. We do publish all our data in its entirety through Nature Energy. If you email me or any other among the authors, we would be happy to send you the paper and additional material.

    • Hi Brandon, happy to share, here is an open access link to ReadCube (free to all, but it cannot be downloaded or printed): Any other questions, feel free to ask.

  3. Nick Greenacre says:

    “Nuclear vs renewable energy”

    Is that a reasonable premise?
    Is it even in question – after all, no one plans exclusive nuclear.
    We are currently pursuing a mix of renewables WITH nuclear power.

    It is Storage that large scale RE would need were we to remove Firm nuclear from the mix so surely the more pertinent [and perhaps less ideological?] area for study would be: ‘RE firmed with Storage -vs- Nuclear’
    And of course that we will need to retain a grid wide ‘Black Start’ capability to recover from grid failures.

    So what would storage and intermittent support cost without nuclear?
    There’s alarmingly little on this.
    There’s not even a commercial flat grid storage.

    UKERC estimate that a 50% penetration of RE would add up to £50/MWh to wholesale costs.
    They didn’t model beyond that 50%

    If we don’t ask the correct questions now, then we could be reconciling ourselves to a future of RE supported by unabated gas.


    • Andy Stirling says:

      Thank you Nick for this interesting and important comment. As someone who has worked intensively on issues around energy diversity (references in the article we are discussing), I entirely sympathise with your point that it is not helpful unduly to polarise the comparison of nuclear and renewable -based strategies, nor to exclude the many other zero carbon energy service alternatives. This said, it does remain relevant against this background to interrogate carefully (as we do in this paper) particular angles on the relative merits of what must be acknowledged to be these two highest profile broad clusters of electricity supply options. Sadly, it’s not possible to do everything in one paper.

      The questions you raise about costs of addressing intermittency and inter-seasonality are also important. But please don’t be “alarmed”, because it was not the purpose of our article to undertake this kind of comprehensive appraisal. In order to be rigorous, we are simply looking at extant empirical patterns and how they relate to three hypotheses.

      But I would mention by reference to a variety of other recent major reviews of this issue (like the recent UK BEIS estimates of levelised electricity generating costs), that it is now becoming quite difficult to avoid accepting that (even taking account of high estimates for these grid integration costs), that very large tranches of renewable resources are typically available at more significantly favourable costs than nuclear power. So the reasonable burden of proof now goes in the opposite direction to the way you suggest.

  4. Michael Hoffmann says:

    After reviewing your article’s regression analyses, referenced data, and supplemental information, I was very surprised to see that your article made no reference to change in electricity production from fossil fuels (namely coal, natural gas, and fuel oil) between 1990 and 2014.

    (Q1) Did you consider how the consumption of fossil fuels in electricity production might have affected the CO2 emissions per capita figure and by extension, affected the association between increase in nuclear power or renewable energy production and CO2 emissions per capita?

    The main reason I ask this question is because between the period of 1990 and 2014 many countries greatly expanded their electricity production capacity mostly through fossil fuels, especially the most rapidly developing countries such as the BRICS countries — all of which also happen to be nuclear power countries. In the case of BRICS the expansion of nuclear power-based electricity would appear to also increase emissions per capita because of the relatively greater magnitude expansion in fossil fuels-based electricity.

    With this in mind, (Q2) did you consider performing a regression on only those countries that are both nuclear and renewable power producing countries? I imagine this would give a more accurate depiction of the association between CO2 emissions per capita and nuclear power or renewable energy because it eliminates the averaging out effect of including the ~100 renewable-only (i.e. not nuclear) countries.

    • Goetz Walter says:

      Dear Mr. Hoffmann,

      thank you for your post. In short my answer:

      Regarding your first question, we understand your point, and we think it is incorporated in our study design. The changes you describe are usually strongly associated with economic growth. And the first variable we put into our model is “GDP per capita”. Hierarchical regression analysis works based on semi-partial correlation. Meaning the effect of independent variables is measured while keeping constant the effects of variables put earlier into the model. Meaning: All the effects of electricity production we found were calculated under the condition that GDP per capita is constant, because GDP per capita was for this very reason put into the model first.

      Regarding your second question, this is a good idea. However, the sample you describe actually is our nuclear country sample (n=30), since all the countries in this sample also have some renewable electricity production.

      If you have any further questions, or if there is some further clarification needed, do not hesitate to write us once more.

      Best Regards,

      Goetz Walter

  5. Stephen de Souza says:

    Thanks for sharing an interesting and thought-provoking article. As you say, correlation is not causation and, in such a polarised debate, we will all rush to conclusions. For me the takeaways are:
    1. That energy policy does need to make choices, because resources (in the broadest sense) are finite. And that probably doesn’t just apply to nuclear vs. renewables.
    2. It is, as you say, rather strange that this hasn’t been looked at before.
    3. I wonder if its fair to characterise nuclear as a solution looking for a problem. As a physicist, the energy yield promised by nuclear fission is very attractive. Maybe the reason nuclear doesn’t perform well on a carbon reduction metric is because that’s not what it was pursued for. Maybe, rather than jump on the tCO2/£ bandwagon, proponents of different nuclear technologies should be clearer about the pros and cons of their technology.

  6. Many thanks Stephen, for this very acute observations. I would agree strongly with all the enumerated points and queries you raise.

    Some possible light is cast on the third issue you raise by some parallel research that has been underway at SPRU for the last few years. In case this is of interest, anyone can find links to some associated research outputs here:

  7. Eero Hirvijoki says:

    Dear Authors

    In your study that appeared in Nature Energy, I think you have overlooked three major factors:

    (1) You have used the total fossil-fuel-emissions-per-capita to assess the mitigation potential of nuclear and renewables, although these energy sources effectively only supply to the power sector.

    (2) You list possible benefits of distributed energy systems but fail to recognize that during 1990-2014, over 90% of the renewable-energy production was from hydropower, which is clearly a non-flexible source with respect to being small in scale and distributable.

    (3) According to your supplementary material, majority of the coutries with a high (>50%) renewable share of the electricity supply were poor developing countries. These countries have overall small emissions not because of high percentage of renewables but because they are poor and have low overall energy-use-per-capita (transportation, agriculture, industry etc).

    Because of these reasons, it is difficult to accept the metric you have chosen and,consequently, the conclusion regarding the hypotheses you come to on the basis of your statistical analysis.

    Eero Hirvijoki

Leave a comment

Follow Sussex Energy Group on Twitter


The views and opinions expressed here are solely those of the individual authors and do not represent Sussex Energy Group.

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 87 other subscribers.


Subscribe to Sussex Energy Group's quarterly newsletter