Nuclear Was One Quarter of Japan’s Power Supply, Not “One Third”

*** UPDATE October 1, 2013 ***
A well recognized energy analyst based in Japan and I had a fairly in depth discussion in the comments of this post. Here is a quick summary of my take away on this issue. Please see the comments section for the details:

1. Nuclear was in fact only about a quarter (25.4%) of Japan’s total electricity generated for the five years up to the Fukushima catastrophe. Nuclear was not “about one third,” nor was it “about 30%.”

2.  One reason that the share of nuclear power has been inflated is due to calculations which fudge the math to show an exaggerated share for nuclear. Some people are dividing the IEA’s Production from nuclear number into the Final consumption number instead of the Total production figure without subtracting any of the transmission/distribution and other losses. This calculation alone inflates the nuclear percentage by about 12% for 2010 alone.

3. Another reason is that some analysts are neglecting to count power generated by non-utilities such as TEPCO and the other regional monopolies. There are various reasons why specialized analysts may want to do this based on their definition of the electricity market, however, this discounts much of the electric power actually generated in Japan and further exaggerates the importance of nuclear power.

Unless a journalist, blogger, or others in the media are describing specific and clearly defined subsets of the Japanese electricity situation, the correct approximation to use is “about a quarter.”

*** End of UPDATE ***

Journalists are consistently misreporting the amount of electricity that nuclear power provided the Japanese market prior to the Fukushima nuclear radiation catastrophe in 2011. Many reports falsely state that nuclear energy provided “one third” or “30%” of Japan’s electricity before Fukushima’s reactors devastated the nation. This is simply not true. In fact, nuclear was only providing about a quarter of Japan’s electrical power in the years prior to the catastrophe.

Here are some examples of inaccurate reporting:

“used to provide almost a third of the nation’s electricity”
Reuters story by Linda Sieg

“plans to raise nuclear capacity from one-third to over half of total demand”
AP story by Elaine Kurtenbach (published on Huffington Post)

“no firm date for bringing back an energy source that had covered about a third of the country’s electricity needs”
Reuters story by Tetsushi Kajimoto

“nuclear reactors provided close to a third of the electricity to keep the $5-trillion economy going before the Fukushima disaster”
Reuters story by Osamu Tsukimori

“until three years ago provided 30% of the electricity to power the world’s third largest economy”
The Guardian story by Phillip Inman & Terry Macalister

There are numerous similarly incorrect statements in other reports.

In 2010, the year prior to the horrific catastrophe, Japan only generated 25.8% of its electric power with nuclear according to the International Energy Agency’s (IEA) statistics. Here are actual production numbers from the IEA for the five years leading up to the Fukushima radiation disaster:

Japan's Nuclear Production

Obviously, nuclear never got even close to a third of Japan’s electricity production, nor was it even thirty percent.

Why then do journalists consistently misreport this information? Well, the nuclear industry can be to blame at least in part. A friend on twitter pointed out that the World Nuclear Association’s (WNA) profile on Japan makes claims that could confuse journalists. The WNA is, however, a public relations arm of the nuclear industry and can be expected to exaggerate nuclear’s importance – it is in their vested interest.

My twitter friend also pointed out that Japan’s Federation of Electric Power Companies (FEPC) also makes a similar “about 30%” claim. The FEPC is also highly vested in promoting nuclear power and can be expected to similarly “spin” atomic energy in a positive light.

Further, another friend on twitter who is a prominent and respected journalist pointed out to me that the confusion over the numbers could possibly be since the nuclear portion of total demand, rather than total supply, might be seen as higher than 25%. This would be since actual demand or consumption of electricity is generally smaller than supply due to various losses such as transmission loss. This explanation does not make sense to me for several reasons.

One glaring reason is that neither Japan nor the IEA publishes consumption of electricity by source, only as a part of total production. It is not clear which source, nuclear or other, is being consumed. So it is very difficult, if not impossible for a journalist to report that “nuclear is X% of consumption.”

Another reason is that nuclear suffers more from transmission loss than other forms of electric power generation. Due to the very real risks that nuclear poses, atomic energy plants are necessarily located far from population centers where the bulk of electricity is consumed. As a portion of electricity consumed, nuclear would suffer more from losses than other forms of power and therefore it seems unlikely that the Japanese power industry would want to show that off.

My own best guess at this point is that the industry spinmeisters have been exaggerating the importance of nuclear power for a long time, and unfortunately, some journalists are not taking the time to check their sources or the math.

In any event, nuclear is now zero percent of Japan’s electric supply. Hopefully it will stay that way for a very long time.


18 thoughts on “Nuclear Was One Quarter of Japan’s Power Supply, Not “One Third”

  1. The WNA wording is “have provided some 30% of the country’s electricity”. Maybe the trick is to go back to the mid-2000s somewhere and get a number just above 27.5 and round that up to 30%. Then the media will helpfully turn the 30% into “one third”.

    The other dodge you see a lot in energy numbers – also used when boosting renewables – is to use a number for “installed capacity” with the maximum you could put out if you turned everything on at once. Japanese nuclear power stations always seemed to spend a lot of their time shut down even before Fukushima, either for regular maintenance or because of safety / seismic / compliance issues. So 30% might plausibly have been the number for installed capacity.

  2. I’ve long noticed this trend of the nuclear industry to be a little sneaky. Indeed, I was just over trawling through stats in the IAEA website and they report the EJth thermal output ) of nuclear reactors which, given a typical thermal efficiency of 33-40%, is about 3-2.5 times higher than the EJe (e as in electrical output) figure, which by the way is given in the IEA stats under the section for electricity. They then compare this EJth figure to the output figures for renewables, the EJ of electricity for renewables that is!

    This would be plain sloppy as it was, only the sources of renewable energy that output predominately in the form of heat (such as biomass or solar thermal) seem to be ignored!

    And the nuclear industry wonders why so few people trust them!

    • LOL! I agree wholeheartedly. As Mark Twain once said there are three kinds of lies, “lies, damned lies, and statistics.” I think the nuclear industry aggressively pursues all three types.

  3. Kevin, hi… I’m on the team that produces No.1 Shimbun, the magazine of the Foreign Correspondents’ Club of Japan. Could we condense your blog post into a sidebar for our next issue? It would be useful to alert our members to your astute observation. Please contact us. John R. Harris

  4. Mr. Meyerson, thanks for your entertaining blog postings. I
    always enjoy reading your theories. I hesitate to offer my
    thoughts, but it might be of interest to some readers why some
    journalists slowly began to use qualifiers like “around,” “about”,
    “roughly”, etc. First, it would be appropriate to look at the data
    in a time-series. No one should cherry-pick one year or only a few
    years for political purposes. I use the 有価証券報告書 (annual reports of
    the power companies) for individual markets and 電気事業便覧 (Handbook of
    Electric Power Industry) for a macro-view of Japan’s power market
    dating back to the 1960s. The OECD/IEA would be fine (because
    ultimately they’re simply getting it from the same source I am),
    but I wasn’t able to replicate your table. The numerator is easy
    enough to establish. That’s is amount of generated electricity on a
    megawatt-hour (MWh) basis by the nine electric power companies
    (sometimes analysts include the extra generated nuclear power from
    the Japan Atomic Power Corporation). The denominator is the tricky
    part. Does one use the total generated electric power of the nine
    companies? Does one include the net electricity supplied to the
    consumer after losses, transfers, and purchases? Or does one
    include the total electricity of the entire market, including
    off-grid generation by IPPs? There is no correct answer. Nuclear
    power as a percentage of total electricity generated first hit the
    30% mark in 1986. It rose to 40% by 1998, and declined thereafter
    for several reasons including (but not limited to) the decline in
    the numerator as a result of the Chuetsu Off-shore Earthquake
    forcing TEPCO’s Kashiwazaki-Kariwa NPS to go offline for a few
    years. Nevertheless, by 2010 (one year before Fukushima) nuclear
    power recorded an annual 33.2% share of total electricity generated
    by the nine power companies. Obviously, if you included Okinawa in
    the denominator, which has no nuclear power generation, the share
    would fall slightly to 33% in 2010. If the denominator were to
    include total electricity supplied (at user end) — which is the
    net power supplied to the end-user after losses, purchases from
    other utilities, etc., you’ll find a similar pattern. 1987 was the
    first year that nuclear hit the 30% mark as a share. It rose
    steadily thereafter. By 2010, nuclear was 30.1% of total
    electricity supplied (at user end). I could continue going through
    all the various permutations of conceivable shares in a
    time-series, but the shares will be “around”, “about”, “roughly”
    the same in each case. I suspect that many journalists use the
    qualified 30% figure simply because it’s both convenient and avoids
    confusing the general reader who probably knows very little about
    wholesale, retail, net generation, load factors, etc. Keeping it
    simple and within a word limit seems to be one of the mantras of
    some newspaper articles. I’m not convinced that there is anything
    sinister involved by “pro-nuclear” lobbyists. In fact, I suspect
    that your numbers (which I haven’t been able to verify yet) are
    likely to be total electric power including off-grid from IPPs. If
    that were the case, the share would fall somewhat but not
    dramatically. In any case, I have to get back to work. Best wishes,

  5. Regarding the International Energy Agency (IEA), I haven’t been able to find the table that Mr. Meyerson posted above.

    However, the IEA publishes various booklets every year. One interesting publication was “Energy Policies of IEA Countries: Japan 2008 Review” (OECD/IEA, 2008). On page 26, the IEA records the actual supply-demand in Table 3 for 1990 and 2005. They were 27% and 31%, respectively. As I suspected, the IEA was basing the table’s shares on generated nuclear power data by the general electric utilities, or 一般電気事業者 in Japanese, which shouldn’t be surprising.

    There is no “correct” number when it comes to 50 years of nuclear power generation in Japan, but it ultimately depends on what data you are selecting, as well as when and why.

    • Paul,

      Thanks for your comments.

      I’m surprised that you are unable to duplicate the results in the table. The numbers come from the IEA’s “Japan: Electricity and Heat for 20xx” which is linked in the article. I produced the table with these numbers using a spreadsheet application.

      Here is the shortened link to the 2010 data for your convenience:

      The numerator is “Production from: – nuclear” which shows the nation of Japan’s total GWh for nuclear power produced for the given year.

      The denominator is the “Total production” number for each year which shows the total amount of electricity the nation of Japan produced. This number is also in GWh.

      I agree numbers should not be cherry picked, and that is why I included 5 years of data leading up to the Fukushima nuclear catastrophe in the table. Five years of data is significant enough for the purposes of finding out how much Japan depended on nuclear power prior to the radiation catastrophe.

      I see how some people would want to only insert consumption numbers as the denominator. From the point of view of the nuclear industry and its supporters, this inflates the importance of nuclear power in popular media. I disagree that this inflation is insignificant. The numbers have been inflated by up to 130% – a significant number. This can lead to an exaggerated perception of the importance of nuclear power which is not based in fact.

      Regarding whether FEPC, WNA or other nuclear lobbyist numbers are “sinister” (a term that you used, not me), industry lobbies are clearly vested interests and to believe that they will not try to “spin” information in favorable light is naive, IMHO.

      Regarding whether individual power companies or individual sectors in Japan generated more or less than 25% of the electricity from nuclear: of course certain subsets of Japan get more from nuclear – Kansai Electric is often mentioned in the media as being particularly dependent on nuclear. This is irrelevant in my opinion as none of the news media’s comments were discussing specific subsets of Japan – the entire nation is being misrepresented via the exaggerations.

      By the way, I am curious as to your thoughts on how much T&D losses can be attributed to various sources of electric power? A former high voltage power engineer friend of mine told me that nuclear suffers from significantly more losses than other sources due to the necessarily long distances that nuclear is transmitted. I have not been able to find data that addresses this question specifically.

      Thanks again for your thoughts.


      P.S. I also recalculated the numbers again to include the entire ten year period from 2001 through 2010. Not surprisingly, the average percentage of nuclear power is only slightly higher – 26.2% for the decade – or still about a quarter, not “about a third” as some less accurate sources suggest.

  6. Kevin,

    Thank you for the reply. I never doubted that you found some data somewhere to support your argument. You tend to be punctilious on this score. What I suspected, and it turned out to be true after I read the footnotes to their website (thanks for the direct link), is that the IEA is simply including the IPPs in the denominator. That means all of the off-grid electric power produced by what the Japanese call 自家発電者, or self-generators, is included in the total denominator.

    Is that the only way to calculate the percentage, which incidentally comes to about 29% of the total (or 288230/999728)*100 in that particular year? Not really. As I said before, there is no “correct” percentage. Technically, the self-generators were never part of the market because they were not true buyers and sellers. The self-generation from IPPS were — as the name suggests — used for their own industrial purposes because they wanted to get off the grid starting in the 1970s-80s when electricity prices were intolerably high. If you’re interested in examining the data in great detail (and I encourage everyone to do so), you should really read 電気事業便覧. You’ll notice that the ex IPP market data has fluctuated in the 30% to 40% range for the past twenty years. Most respected energy analysts, financial analysts, and economists usually use the ex IPP data because that’s what actually is traded. Consumers don’t technically “buy” self-generation from IPPs (although since phase 1 of liberalization in 1995 some IPPs were able to sell their limited surplus to the general electric utilities to help reduce incumbent supplier prices).

    Again, perhaps I wasn’t clear the first time so I apologize. The numbers I cited above that fall within the 30% to 40% range in a time-series were the AGGREGATE of the general electric utilities. I am puzzled why you won’t acknowledge the point and instead keep insisting that there is only one way to look at an issue.

    As for the charge that I am somehow “naive”, what can I say? I’ve been a professional energy analyst for fifteen years. I am neither “pro” nor “anti” anything; I am concerned with what is (empirical) and not what should be (normative). I really have a difficult time understanding why some people tend to think in black-and-white normative terms, but I suppose there is a certain attraction to thinking that everyone must somehow have a political agenda, everyone must somehow be powerful (but only when the preferred policy agenda of the author is not adopted), and everyone is somehow trying to “spin” something. Having spent years in libraries analyzing data, transcripts, interviews, etc I can say with some confidence that the only “naive” position is the assumption that everything is somehow black and white.

    Personally, I’ve never found it too useful to think in normative black and white terms (i.e., “I’m pro renewable energy”, “I’m pro nuclear”) because in my experience it’s difficult to prove objectively the type of causality you assume must be happening, but also because it’s difficult to predict the future when someone is so clearly biased. Then again, you might be right. It could very well be that “pro-nuclear” sources are attempting to “spin” data in the media somehow. Unfortunately, the problem with your argument — aside from the obvious that you are simply assuming this to be true — is that it leaves you vulnerable to the same criticism. Are you attempting to “spin” carefully selected data to support an anti-nuclear argument rather than acknowledging that there are several possible denominators and none of them are the “correct” one?

    I’ll let readers decide.

    Best wishes,

    • Paul,

      First off, please allow me to apologize. I do not think you are in any way naive, nor was it my intention to insinuate so. I was speaking to the situation you described, not to what I think of you personally.

      Regarding the charge of me being ‘punctilious’, after living in a country for so long where people are upset at the trains being late more than a minute or two, I think my punctiliousness is within reason, although my wife (who happens to be Japanese) believes quite differently…

      As to the math (which may or may not be black or white IMO), I’m still not convinced that the percentage of nuclear should be calculated as you are suggesting. Going back to the IEA’s numbers for 2010, you are stating that it is OK to divide the nuclear production total by the final consumption total to get a percentage. Using your argument, this calculation should apply to all sources of electric power production, not just nuclear.

      This does not make mathematical sense, IMHO. By using your suggested calculation for all sources of production, electric power production in Japan adds up to 112%, a clearly incorrect result. The math requires the percentages to add up to one hundred percent – Japan is not really making more electricity than it is making, so to speak.

      Here is the IEA’s 2010 data for Japan that we have been referring to along with the calculations that you and I are suggesting:

      *the above link is a graphic file showing the calculations. I tried to insert it into my comment but it did not display properly. I obviously need to learn how to use better…

      The column titled “% of Final Consumption” is based on your suggested way of counting. Does it make sense to you for the numbers to add up to significantly more than 100%?

      Regarding the issue of IPPs, if the analysis is targeted only at the power market made up by the *EPCOs, then removing the IPPs from the equation makes sense to me. However, these are tumultuous times in Japan’s energy market. Many, many numbers of consumers are faced with new choices on how to get electricity. These choices did not exist in most people’s minds (other than a limited number of large power users) prior to the Fukushima nuclear catastrophe. My personal opinion is that leaving them out of the equation when describing the Japanese power market leaves the observer without an accurate picture. A more holistic picture of Japan’s power market prior to, and after the catastrophe is critical.

      FWIW, my thoughts and opinions are greatly shaped by my long experience in the internet business. Prior to deregulation in the mid 1990s many industry experts claimed that the ensuing internet revolution in Japan just would not happen. In fact it did happen despite these claims, and it happened far more quickly than most thought possible at the time. Personally, I think the revolutionary energy transformation that is currently occurring in Japan, and other places such as Germany is very similar to the situation in the telecom market prior to 1995. Change is already happening quickly and it will continue to accelerate in my opinion.

      Looking forward to your thoughts!

      Best regards,


      • Kevin,

        Thanks for the reply. I really need to get back to work, so let me offer a few more clarifications and I’ll let you have the last word.

        Regarding the difference between initial generation and end-user consumption numbers not adding up, that is understandable. It will be difficult to explain in a blog posting without Excel to illustrate what is happening. Suffice it to say that generated electricity is lost along the transmission grid until it reaches the end-user. This is not a generation source issue; it is a transmission grid issue. Historically, the loss ratio for the EPCOs (送配電損失率) in the post-war period fluctuated between 5% and 12% depending on the state of the grid and the distance from the source to the end-user. The total loss rate (総合損失率), if one were to include the power being transmitted from wholesale electric utilities (卸電気事業者), municipal electric utilities (公営電気事業者), joint-ventures (共同火力発電事業者), and special electric utilities (特定電気事業者) sometimes raised the loss ratio further.

        In my fifteen years as an analyst, I have never seen data to support your assertion that “nuclear suffers more from transmission loss than other forms of electric power generation.” From the way you worded your sentence, it seems a little more like speculation than fact to me.

        Regarding the IPPs, people sometimes make the mistake of confusing the off-grid IPPs (自家発電者) with the new entrants, or Special Power Producer and Suppliers (特定規模電気事業者), in the liberalized market and think this is all part of the “potential”. But the former are not the latter.

        Let me put this in a slightly different way, and show why most people end up saying “roughly” 30% of nuclear power.

        Energy analysts distinguish between 電気事業用 and 自家用. 電気事業用 consists of the generation from the 10 EPCOs (一般電気事業者), the wholesale suppliers (i.e., J-Power, JAPC, and others), the Special Electric Utilities (特定電気事業者), and the new entrants into the liberalized market. In total, they generated about 918,200 GWh in 2010. These are net buyers and sellers. Full stop. 自家用 are the independent power producers (IPPs). They generate an additional 239 GWh. Roughly. Most of it is conventional thermal power because it was the cheapest and quickest they could build. IPPs were not allowed in the post-liberalized market to build nuclear reactors, even if they wanted to build them (and some actually did). In any case, this added power brings us to around 1,157 GWh for the country as a whole. The total (合計) is before we even start talking about transmission and distribution losses on the 電気事業用 side, which will bring the end-user units down somewhat.

        The IPPs are not part of the “market”. They never were. They only recently can sell a small portion of their surplus power to consumers, and this is assuming that the IPPs are connected to the grid (most aren’t). If you were interested in market potential for consumers in post-Fukushima Japan, as you say, then it makes very little sense to include them because they are never going to surrender their cheaper electric power for internal operations (e.g., steel, ceramics, beer, non-ferrous metals) to the general consumer. That is *not* why they were created. I suppose if these industries went bankrupt, one could sell their plants into the market. That would mean a reduction of 自家用 and an increase in 電気事業用. Anything is theoretically possible, but also highly unlikely.

        So, to suggest in your blog posting that the FEPC, WNA, and others are also presenting an inaccurate picture of the market in order to push up the share of nuclear in the denominator by not including the IPPs is not a very compelling argument, in my opinion. It misunderstands the distinction between 電気事業用 and 自家用. And aside from not having evidence that these actors were deliberately trying to report “inaccurate” information, it would also be presenting a somewhat misleading picture of market “potential” to the general reader that is divorced from the IPP reality.

        Best regards,

      • Hi Paul,

        Thanks again for the detailed response.

        Regarding transmission loss, I was able to confirm via google that a number of authorities do in fact acknowledge that nuclear power has significantly more transmission loss than other forms of power generation. Nuclear plants are farther away from consumers and therefore have more loss.

        For example, the IPCC (Intergovernmental Panel on Climate Change) attributes nearly double the transmission/distribution loss rate to nuclear power (7%) versus fossil and biofuel power generation (4%). The IPCC numbers are based on research done in Japan at CRIEPI (Central Research Institute of Electric Power Industry). CRIEPI is funded mainly by Japan’s regional electric power monopolies. CRIEPI is quite vested in the nuclear industry.

        Incidentally, the CRIEPI acronym is quite humorous on several levels…

        Nuclear transmission loss in Japan is probably even worse according to some. This is because much of Japan’s world leading pumped hydro generation capacity was built in tandem with nuclear plants and the T&D losses for these hydro plants are arguably part of nuclear’s own loss rates.

        Without subtracting these significant losses attributable to nuclear from the production figures, using the “about 30%” or “close to a third” generalizations are simply not accurate and require the reader to have specialized information for proper understanding.

        On your point of whether independent power producers were/weren’t part of the market, it seems a bit unnecessarily esoteric to me. I can certainly understand why you as an analyst want to formally define the market for academic purposes. From a different point of view, electricity consumers have a choice of whether to buy power from the grid or make it and consume it themselves. They are still part of the broader market in Japan.

        In any event, I think you and most readers will agree that mass media news reports should not report inaccurately calculated information, or at least should not do so if they are not going to clarify the inaccuracies in the articles. Lengthy explanations are not realistic. It makes much more sense to publish accurate information that does not require lengthy explanation and esoteric information.

        Best regards,


  7. I’m glad to see that someone else is trying to uncover these distortions about energy statistics in Japan. I wrote about it on my blog a few months ago. I tried to point out that energy imports are not the main issue in the trade deficit, and the increase in fossil fuel consumption used for electricity is much smaller when you consider the overall consumption for transport, heating and other non-electric uses.

    • Thanks for your feedback! I agree. The supposedly massive increase in fossil fuels is also highly discounted when one takes into account the truly massive efficiency increases that Japan has realized since the Fukushima nuclear catastrophe. Much of the nuclear power generation lost has already been made up by efficiency gains.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s