SYDNEY (Reuters) - Last week the International Energy Agency published another World Energy Outlook (WEO), as it has every year since 1998. Coronavirus meant this year the report wasn’t launched in a London room packed with the world’s top energy journalists. But it still generated dozens of news stories. This year, two headlines stood out. IEA Executive Director Fatih Birol unveiled the beginnings of a scenario outlining how the energy sector might develop if the world decided to get serious about limiting global warming to 1.5 degrees Celsius above pre-industrial times. And he declared solar power to be the “king of electricity”, acknowledging that it is growing capacity faster than any other electricity source, and can cost less to install than existing coal-fired power facilities do to run. Long-term IEA-watchers no doubt greeted all this with a bittersweet smile. The WEO is much more than just an overview. The report’s numbers and tables really matter. Those in the central scenario, which map out future energy pathways to 2040 in five-year increments, are plugged into estimates, models, and reports by governments, companies and investors around the world as a baseline, business-as-usual scenario. The IEA describes the WEO as the “gold standard in long-term energy analysis”. It’s hard to agree. The IEA may participate in many climate-related initiatives and be well-regarded for many of its other data-gathering activities. But a common complaint is that it has been consistently pessimistic on clean energy’s growth potential, particularly the scope for costs to fall as technological learning curves accelerate and display unwarranted support for the status quo. Last year dozens of investor groups, scientists, business leaders and climate policy experts wrote to Birol himself, asking for reforms. IEA assumptions on technology and economic optimisation are buried deep within its model, so it can be tricky to prove the point decisively. But compare how the IEA’s past projections have performed against reality, and the pattern becomes obvious. For ages, the WEO has severely underestimated solar growth. Year after year, the actual deployment of solar photovoltaic panels outstrips the IEA’s projection in its central scenario. This has caused some embarrassing volte-faces. The IEA cut its view of the current cost of solar photovoltaic panels in the United States almost in half in just a year - from an inexplicably high $95 per megawatt hour in the 2019 WEO, to a more realistic $50. Analysing the annual reports back to 2000 and comparing them with actual expected energy use in 2020, assuming no Covid-19, shows how systematic the problem is. The WEOs have in almost every year underestimated the growth of wind power – only 2016 veered towards the too optimistic. The WEO only began to acknowledge offshore wind power as a major power source in its own right in 2019. Fossil fuel forecasts, meanwhile, have been too optimistic. Despite the U.S. shale boom and the explosion in liquefied natural gas facilities, the IEA’s “central” scenarios mostly overestimated how much gas would be produced this year. Coal production in 2020 was also more accurately predicted by the “climate-constrained scenarios” in much of the past decade than by the “central” scenarios. Meanwhile, the WEO has made extremely optimistic assumptions about the potential for carbon capture and storage, a technology fundamental to continuing use of coal and natural gas in power generation in a carbon-constrained world. The IEA first featured carbon capture and storage (CCS) in its carbon-constrained scenario, then called the Alternative Policies Scenario, in 2007. Just three years later, the WEO’s 2010 carbon-constrained scenario, then known as 450S, projected CCS-enabled coal plant capacity would total 400 GW by 2030. By 2019, it was only expecting 215 GW by 2050 – half the amount, and 20 years later. The IEA’s 2020 improvements are welcome, but they remain far from perfect. The new net-zero forecast is really a mini-scenario, or “case”, because it only shows 10 years of a 1.5 degrees Celsius pathway up to 2030. This is a long way from adopting such an outlook as a central scenario. Punches are still being pulled. The WEO is also a bit of a black box. The IEA publishes limited detail about its methodologies, including its self-developed “value-added levelised cost of energy”. VALCOE is supposed to be a handy way to compare the cost of solar and wind when allowances are made for the fact that it’s not sunny or windy all the time. Compared to the most established standard of “levelised cost of energy”, or LCOE, the IEA’s VALCOE makes solar and wind energy more expensive, while gas and coal become cheaper. But it’s not possible to understand the moving parts. The IEA has a difficult balancing act to strike, and its 2020 forecasts are a step in the right direction. But it has much further to go to shake off the energy community’s scepticism.