Natural Gas (NG) demand is widely modeled using Heating Degree Days (HDD). At Planette we can forecast monthly average HDD 1-to-3 months in advance. The benefit of this data is obvious. We can also forecast the volatility of temperature. The benefit of this data, however, is not obvious. In this post, we venture to show how it can generate alpha.
The spot and future prices of any commodity are related to each other by this equation:
Where, Fr , P0 are the spot and future price of commodities, respectively, and r , c, y are the borrowing rate, storage cost and convenience yield, respectively. A similar relationship exists between the prices of different contracts as well. Figure 1 depicts the spot price of natural gas as well as the generic 1st, 2nd, 3rd and 4th month contracts:
You can see how the prices zig and zag together emphasizing the direct connection between them. This means that lower supply or higher demand at any point in time pushes up not only the price of spot and all Futures contracts simultaneously. Planette’s HDD forecasts can be used to take positions in any contract or all contracts (swaps).
It’s interesting to note that while the prices all zig and zag together, the spreads between them increase and decrease over time. These fluctuations are the result of changing borrowing rates, storage costs and convenience yields. Borrowing rates can be modeled as a spread over the treasury forward rates, while the storage costs are easily predictable since storage follows a sinusoidal pattern and storage quantity data is published by EIA weekly. The convenience yield, however, isn’t predictable, but Planette’s volatility forecasts can help model it.
Traders can thus monetize our forecasts by taking the right positions - long or short - on calendar spread contracts.
The following figures show two measures of volatility in monthly temperature data for Chicago. Figure 2 shows the average of daily minimums for the month of January in the last 20 years. You can see how there is no discernible pattern in the data. Convenience yield is just the benefit of having natural gas available to meet unforeseen increase in demand. The year-to-year variation in daily minimums is a good proxy for this.
Another proxy for unexpected demand is the standard deviation of daily average temperatures, as seen in Figure 3. Technically, it is the volatility of the volatility! If this standard deviation was more or less the same for every year, one could gauge convenience yield reasonably well. But it clearly isn’t the same every year.
It’s possible to forecast both these factors of volatility. In fact, we can forecast the daily distribution of temperatures as well as Low temperature days (LTD). LTD is the number of days in a month with minimum temperature below a specific temperature value. For example, LTD (32F) is the number of days where the daily temperature was less than 32F.
You can transform your natural gas trading strategy with Planette's Sura™ forecasting system. We deliver high-precision forecasts, 2 weeks to 1 year ahead, in formats that integrate seamlessly with your trading workflow.
Sign up for a 90-day Free Trial of Sura™ to incorporate our temperature forecasting data into your own volatility forecasts so you can capitalize on convenience yield trends before the market prices them in.
