During the cement-making process, calcium carbonate is reduced at 1,600 degrees centigrade in a kiln, producing an intermediate product, called clinker, which is then pulverised into grey cement powder. The process is highly energy intensive, both in terms of fuel for the kilns and the electricity used to operate the grinding mills.

"The percentage of overall expenditure that goes on energy is 25%, which is the general benchmark for the industry," says Andrew Poulter, chief financial officer at Adelaide Brighton Cement, an Australian manufacturer of cement, lime and prepackaged dry blended products.

Adelaide Brighton runs over 20 mills in South Australia and Western Australia. Such mills can vary greatly in size, capacity and efficiency, but are often long-lived installations – one company mill is over 70 years old. Smaller mills are less efficient on a tonnage-produced basis, as they consume more energy pro-rata. Mills account for a large part of the power used, while the company's nine kilns run on natural gas or a mixture of gas and coal.

"Coal and petroleum coke are the fuels used most by the cement industry worldwide," says Poulter. "In Australia we're a little different as we use natural gas because of its abundance and low cost – but things are changing on both points."

The company takes out three-year contracts through a buyers' pool for its gas supplies, thereby avoiding cost spikes. However, the company operates on the spot market for electricity in South Australia. At times of high demand, particularly during hot summers, the spot price can vary significantly. There is a pricing cap of $10,000 per megawatt hour (MWh), though the price normally runs between$45 and $50 per MWh. Adelaide Brighton takes an ingenious approach to operating in such a volatile market. Price forecasts are relayed to mills at five-minute intervals conjoined into half-hour periods. When prices are predicted to peak, mills can be turned off. "Effectively the way we hedge our power pricing is by managing our demand, because the cement mills can be turned off at virtually the flick of a switch. We have our own in-built hedge through the way we run our plants." Price is reviewed on a half-hourly but also look-ahead basis. The plant operator monitors the price throughout the day and turns equipment off according to matrix of prices versus stock levels. As a back up to human monitoring, the price system is also linked to the plants' process control systems, which give off an audible alarm when prices reach certain levels. Product engineers and production managers also monitor forward prices for the coming day to highlight any imminent price spikes. "We even have an accountant who sits in the control room as well, who acts as a back up," says Michael Williams, operations manager at Adelaide Brighton. "He'll ring round the other plants' control rooms to make sure that they're on the ball if there a price spike forecast." If a significant and prolonged peak in power prices is detected, the company will also turn its core cement kilns down. "If you take those down for any period of time you have to go through a proper ramp down then ramp up," says Poulter. "It's quite expensive in terms of efficiency losses but we have done it in the past to avoid peak pricing. If you turn your mills and kiln off you're not producing anything." If kilns are turned off, the ramp up process can take from 24 to 36 hours. "If prices are thought to be high for four to six hours we'll look at turning the kilns off," says Williams. "That happens quite rarely, but turning the mills off happens very regularly." Adelaide Brighton runs its plants at capacity in terms of clinker production. If kilns are turned off, the production of that tonnage of cement is lost forever. "Turning off the mills is not quite as punitive on our bottom line," says Poulter. "We often have to turn them off anyway because we have a full inventory." The volatile conditions of Australia's National Electricity Market (NEM) forced Adelaide Brighton to review their energy risk management policy. It was Williams who wanted to understand the merits of hedging, and to take an alternate view of operating on the spot market. "It was about two years ago, and we actually believe, on balance, that we've saved a figure in the millions of dollars over what we would have paid if we'd been in the buyers' pool," says Poulter. There are downsides to the policy, however. In 2004 there was large interruption in supply after a fire at an electricity substation. "Very quickly you have hundreds of thousands of dollars in additional power cost," says Poulter. "If you get any serious force majeure and prices stay high for a long time, we have to continue to operate, to keep the cement market supplied." Retail suppliers build the possibility of such events into their price model. "In effect you pay for the option of managing your own risk versus paying a huge premium for someone else to take on that risk. Our analysis showed it was much better for us to manage our own risk," says Poulter. However, inefficiency in the NEM can sometimes benefit the end-user. On two occasions during summer months the company was actually paid to use electricity. "In other words the different power stations and companies were all out there bidding, and rather than reducing their baseload they wanted to make sure they sold their product; we actually got credit for a period – several half hours in one day," says Poulter. However such circumstances are rare. "What is the norm is the really extreme peaks in pricing where you get that peak of$10,000 per MWh," says Williams. "It just beggars belief how much the market can be distorted."

The NEM does offer a reserve trading mechanism, which allows companies to be automatically taken off the market in times of acute demand. Demand is projected over a six to eight week period. However, as the NEM bills clients for committing to the scheme, and Adelaide Brighton chose not to be a part of it. Poulter is, however, satisfied with the company's current system. "In my experience it's quite an unusual way to operate," he says. "Hedging like this costs you over time, but it takes out the peaks in pricing."

#### 7 days in 60 seconds

###### VM change, Libor fallback and DTCC blockchain

The week on Risk.net, November 10-16, 2017