Solar grew by 50%: compared to what exactly?

In line with previous post (on the claim that solar is to surpass nuclear by the end of the year), there is this Bloomberg article titled Solar Grew Faster Than All Other Forms of Power for the First Time. A new IEA report claimed that:

solar powered by photovoltaics, or PVs, grew by 50 percent, with almost half of new plants built in China

There was no link to the report, there was only a link to a Bloomberg page with the company profile of IEA. Searching the IEA website revealed that the article “
Solar PV grew faster than any other fuel in 2016, opening a new era for solar power” was rather similar to the Bloomberg article and reported the same claims. It described this report: Market Report Series: Renewables 2017 – Analysis and Forecasts to 2022.

Unfortunately, the report is paywalled and I will need to reach for my wallet first to get access to it. Which is not something I wanted to do. Luckily, there is the BP data that I used in the last few posts. It also has the photovoltaics installed capacity data. The second part of the claim was easily found. The difference of installed photovoltaic 2015 versus 2016 for China is 34,530 MW. Worldwide difference is 75,093 MW. Meaning roughly 46%. So “almost half of new plants built in China” seems to be confirmed by the BP data.

However, I had some problems confirming the first part of the claim (that “PVs grew by 50%”). According to BP, the worldwide increase of photovoltaics capacity is 33.2%:

That is considerably less than the 50% claimed by the IEA.

That is quite puzzling. I don’t think it is a rounding issue. It is very unlikely that 33.2 (or 0.332) is rounded to 50 (or 0.5). So I assumed that the IEA data is rather different from the BP data and since I no access to that data, I initially forgot about it.

That changed when I became curious of how different those two datasets were and what were those differences. I remembered that there was also a bar graph in the Bloomberg article that depicted the new installed capacity in 2016 of five countries (China, US, Japan, India and UK):

Comparison with the BP data showed that the two datasets were much closer to each other than I expected:

The bar of China reaches just below 35 GW. The BP table mentions 34.54 GW
The US bar flirts with the 15 GW line. The BP table mentions 14.73 GW
I guess that Japan is around 8 GW in the graph. The BP table mentions 8.599 GW
India is just below the 5 GW line. The BP table mentions 3.97 GW
UK is just below half the 5 GW line, so somewhere around 2 GW. The BP table mentions 2,039 GW.

Although it was only a very rough check, the figures in both datasets seems to match rather well. At least for the countries with the largest increases of photovoltaic capacity (in absolute numbers).

This made me even more curious, so I decided to search for the “IEA Photovoltaic Power Systems Programme” (one of the sources used in the BP data) and found an overload of information about solar energy. Very interesting was the Snapshot of global photovoltaic markets 2016. On page 4, there was the statement that capacity of PV grew by 75 GW worldwide.

Which is exactly what the BP data showed. See above.

On page 10, I found the increases of the top 10 of countries with the biggest increase in PV in 2016. Indeed, those numbers were all in the same ballpark as the BP numbers. Confirming that the two datasets are bit far off each other.

Then I found the PV capacity worldwide on page 7:

HOW MUCH PV CAPACITY IS PRODUCING ELECTRICITY IN THE WORLD TODAY?
The total installed capacity at the end of 2016 globally amounted to at least 303 GW.

Later on that page, the number 303.1 GW is given (the BP number is 301.473 GW).

If the global capacity was 303 GW and the increase 75 GW, then the capacity in 2015 should be somewhere around 228 GW. I quickly found the Snapshot of global photovoltaic markets 2015 and found this on page 7:

HOW MUCH PV CAPACITY IS PRODUCING ELECTRICITY IN THE WORLD TODAY?
The total installed capacity at the end of 2015 globally amounted to at least 227,1 GW.

If that is really true, then it is not even possible to have a 50% increase with their own numbers!!!

According to IEA’s own number that is a (303.1 – 227.1)/227.1 x 100 = 33.5% increase, similar to what BP calculated (33.2%). At this point I was at a loss how IEA could claim that “PV grew by 50 percent”.

Then I read page 7 of the Snapshot of global photovoltaic markets 2016 report again and slowly bells began to ring (my emphasis):

In 2016, the PV market broke again several records again and continued its global expansion, with a 50% growth bringing the market to at least 75 GW. After a limited development in 2014,
and a 25% growth in 2015, the market continued its growth, with many regions of the world contributing to PV development.

Aha, I think that their definition of “PV grew 50%” is different than mine. To test my hypothesis, I searched for the global installed capacity for 2014, which I found in the Snapshot of global photovoltaic markets 2014 report on page 6:

HOW MUCH PV CAPACITY IS PRODUCING ELECTRICITY IN THE WORLD TODAY?
The total installed capacity at the end of 2014 globally amounted to at least 177 GW.

I think I now know how IEA came to their claim:

In 2015 there was a growth of 227.1 – 177 = 50,1 GW
In 2016, there was a growth of 303.1 – 227.1 = 76 GW
Therefor the increase in 2016 (76 GW) was roughly 50% bigger than the increase in 2015 (50.1 GW), coming to the conclusion that “PV grew by 50 percent”.

But it was not PV that grew by 50% compared to the previous year (it obviously grew by 33%), it was the increase of PV that grew by 50% as compared to the increase of the previous year. 76 GW was added in 2016 compared to 50 GW the year before.

I wonder whether IEA reinterpretated their definition of PV growth on purpose or whether they were confused about it themselves. Whatever the case, it is rather clever communication. By doing so, the IEA framed a 33% growth as a 50% growth and therefor giving the impression that PV is in fact doing much better than is warranted by the numbers.