We reported previously that the Loss-making Power Generation Sector in Singapore requires Bold Steps to improve its gross profit margin (i.e. Spark Spread) and profitability. Investments in new power plants during 2011-2013, partly incentivised by the LNG Vesting Scheme, resulted in a structural over-capacity of efficient CCGT generation.

This over-capacity combined with Singapore’s Market Model which determines the prices for electricity leaves little room for the power generating companies (the Gencos) to recover their fixed costs. It is a feature of these markets that they clear at the marginal cost of production. And the marginal cost of production of a CCGT power plant does not recover its fixed costs.

The Vesting Contracts introduced by the Energy Market Authority (EMA) to curb the exercise of market power by the Gencos reveal the extent of the issue. Numbers derived from the Vesting Contracts show that the gross profit margin (VC Adjusted Spark Spread) of CCGT power plants is not sufficient to cover costs like fixed and other non-fuel related costs, financing costs and a return on equity1See Loss-making Power Generation Sector in Singapore requires Bold Steps..

As Vesting Contract prices are set at the start of each calendar quarter, the fuel costs projected in these prices are not necessarily representative for the actual fuel cost. Fuel costs changes are passed on to the Vesting Contract prices, but with a delay of a quarter. This clarifies why in some periods the Adjusted Spark Spread as calculated from the Vesting Contracts is negative. It suggests that during these periods the Gencos would run at a gross margin loss. This is however not the case. The Vesting Contracts projected cost of fuel (natural gas) has been larger than the actual cost of fuel during these periods as a result of the delay.

This is illustrated by figure 1 in which the Adjusted Spark Spread is calculated using our estimates of the realised costs of Natural Gas in Singapore.

Figure 1: Adjusted Spark Spread vs Non-Fuel Costs

The periods with a negative VC Adjusted Spark Spread (based on the the Vesting Contract implied cost of Natural Gas) are now showing positive when based on the realised cost. This is most profound during Q3-Q4 2008 when commodity prices drop rapidly following the global financial crisis. And likewise during the decline during the period Q3-Q4 2014 and Q3-Q4 2015.

Figure 2: Dated Brent Oil Price 

It follows that the realised Adjusted Spark Spreads are larger during periods with falling fuel costs than the VC Adjusted Spark Spreads as implied by the Vesting Contract prices. Nevertheless, the conclusion we made before2See Loss-making Power Generation Sector in Singapore requires Bold Steps. is the same. The current electricity prices are not sufficient to cover all costs (with one exception in Jul 2015)3In July 2015, some coincidental events resulted in an unexplained price spike. Although investigated by the EMA and Market Surveillance & Compliance Panel, no prove of market rigging or irregularity was found. and take away the ability of the power industry to be profitable.

, , , , ,
Martin J. van der Lugt

Martin works as an independent consultant for the energy industry. He gained wide ranged experience in the gas and power markets in Northwest Europe working for European and American energy companies and as an independent consultant. He has a particular interest in the developments of the Southeast Asian markets and is open to discus opportunities.


  1. Martin, interesting article highlighting the key risks the CCGT industry faces today. It’s a matter of time before some of the newer gencos which don’t benefit fully from the vesting contracts start getting into solvency issues. The USEP prices today are still way below the sustainable long-run cost of running CCGTs. Seems like the situation for Gencos are pretty bleak today. While some of these Gencos scramble to find alternative off-take than the Vesting Contracts for their electricity via the contestable market, the EMA giving out more than 25 retail licenses in such a short has eaten up all the margins which has existed for a long while.

    I can’t seem to understand why the EMA has decided to take such drastic moves in reducing the vesting quantities and introducing an unhindered wave of independent power retailers during this period of low commodity prices (and historically low spark-spreads) to destroy the electricity sector of Singapore.

    1. Hi Andrew,

      The newer Gencos have mainly contracted LNG supplies as fuel and are therefore covered by the LNG Vesting Scheme. It are the incumbent Gencos that are now almost fully exposed to the low USEP as a result of the low Vesting Contract levels and reduced turnover as the new plants take production away from the incumbent Gencos.

      When the EMA determines the Vesting Contract levels, it only considers the potential for market power of Gencos in influencing the determination of the USEP in Singapore’s Market Model. With the addition of the additional capacity, this market power is significantly reduced, if not non-existent. As such the EMA is correct to reduce the Vesting Contract Levels.

  2. Firstly, a very insightful article into the thinning of Genco margins and its reasons; I learnt a lot about the power generation sector that I didn’t previously know about.

    Just a question though: while I have no doubt that the adjusted spark spread would approximately reflect the margins of the Gencos, I would like to ask about whether you believe the impact of the thin spread on their bottomline is mitigated by whatever hedging strategies the Gencos may have engaged in, be it in Brent or HSFO futures?

    It’s just inexplicable to me that with profits highly dependent on their COGS, the Gencos would not engage in trading activities that substantially limits their exposure!

    1. Hi Samuel,

      The squeeze of the spark spread of CCGT power plants is a result of the over-investment in new CCGT capacity during the 2012-2014 period following the introduction of the LNG Vesting Scheme in 2010.

      The LNG Vesting Scheme provides a seamless spark spread hedge for this capacity. It is therefore understandable that new Gencos (Pacific Light and Power, Sembcorp and Tuaspring) made these investments as they are not affected by the squeeze of the spark spread (which was a result of these investments).

      The incumbent Gencos (Senoko Power, Tuas Power and YTL PowerSeraya) take the bulk of the loss in profitability on their existing plants as far as not covered by the Vesting Contracts. In 2013, the Vesting Contracts Level for these plants was still 55%, but it dropped to 19% today. On top of that, the new plants take production away from the incumbent Gencos.

      A hedge of a spark spread would be a combination of a long-term sale of power and an equally long-term purchase of fuel. This is normally done through a Tolling Agreement1. Effectively, the Gencos would have to sell the capacity against a fee.

      This is highly unlikely (if not impossible) in a market with an over-capacity as large as in Singapore. Even if a party has an interest in doing so, it would certainly not be willing to pay for this capacity to the extent it covers all non-fuel costs to restore the profitability of the plant. The Market Model provides much better prices and the over-capacity keeps the risk for price hikes very limited.

      Note that the Adjusted Spark Spread already includes the effects the hedging as a result of the Vesting Contracts and the LNG Vesting Scheme.

      1 Tolling agreement: A processing agreement for the conversion of an input product for a fee. In the energy sector, tolling agreements are contracts where one party – the toller – provides a company with one form of fuel to be converted into another form of fuel on their behalf. In the electric power market, tolling agreements are typically between a power buyer and a power generator, under which the buyer supplies the fuel and receives an amount of power generated based on an assumed heat rate at an agreed cost.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.