Oil prices drive market volatility

Forecasting the price of oil is by no means an easy task. West Texas Intermediate (WTI), the benchmark for New York Mercantile Exchange (NYMEX) crude oil, had traded, for the most part, below US $30/bbl prior to 2002. Beginning in 2002, a steady rise in oil prices ensued, supported by record oil demand growth as the global economy expanded at a record pace.

A culmination of factors led to prices skyrocketing to an all-time high of roughly $145/bbl in 2008. Less than six months later, oil prices had fallen to around the historical $30/bbl range. At the time this article went to press, WTI closed at a two-year high of $87.06.

What causes such volatility, and how can future swings in oil prices be evaluated?

Record oil demand growth followed by years of underinvestment in the expansion of oil supply culminated at a time when industry costs to develop either an upstream or downstream asset had more than doubled. At the same time, political instability in some major oil-producing countries led to incremental supply disruptions. OPEC had to increase production to meet demand, drawing down spare crude capacity. As these factors pushed up the fundamental cost of oil, the relaxed monetary policy of the US Federal Reserve drove investors to seek alternative investments as the dollar weakened.

Speculative investors began to pump large amounts of liquidity into the oil market, and negative real interest rates – low cost of borrowing – provided the perfect environment for investors to raise their risk tolerance and jump into new asset classes such as commodities and oil. Commodity price inflation ensued across precious, hard, and soft commodities. The result was a massive price shock, all-time high oil prices that burst with the near collapse of global financial markets, and concomitant global economic recession by 2008.

Not only did the absolute price of oil change over the past decade, but so did volatility between any particular day or month. Between 1990 and 2000, the month-to-month price of WTI crude typically fluctuated around a band of +/- $3. This trend broke in the early part of the decade, with the price of WTI crude swinging within a +/- $10 band on a monthly basis. When presented as a percentage, the volatility in the price of WTI in the same period had only slightly increased by an average of about two percentage points from the previous decade (1990s average: +/- 5%). Despite the minimal increase in price volatility in percentage terms, the frequency at which price fluctuations have occurred has increased significantly.

Forecasting oil price
Uncertainty surrounding the price of oil, and the volatility for which it trades, led Hart Energy Publishing to develop both a short-term and long-term WTI crude price forecast. The longest-term price reflects the cost of producing the marginal barrel of oil; the economic price of oil underpins the price without any market friction – it is useful for pure analytical purposes and, more specifically, sets the price floor for oil. Hart references short-term pricing as the seasonal price, which denotes the average price over a quarterly basis.

In addition to the long-term fundamental influences of supply and demand, the seasonal price is heavily influenced by variables such as:
• Market-based indices (S&P 500);
• Purchasing power stability (price of gold);
• Unit of exchange/currency (USD/EUR); and
• Speculative pressure (traders’ net positions).

These factors tend to influence the daily price of oil and have created a new paradigm of oil prices with the role of non-fundamental influences beyond that from conventional demand, stock, and supply amplifying the price of oil and impacting short-term prices.

It is clear that prices no longer solely depend on the fundamentals – that oil no longer works solely on conventional demand and supply balances as was the case before 2002. Nor is it based only on the new paradigm that currently seems to be driving another upswing in prices.

Short-term economic outlooks and future long-term oil market fundamentals appeared to be destabilizing the pre-crisis dynamics of the price of oil. For example, traditionally oil prices had an inverse correlation to the stock market, but lately they appear to be going against the conventional wisdom that a stronger stock performance usually leads to a fall in prices.

Realizing this change in price structure, Hart has pioneered a proprietary price forecasting model that takes into account the new financial and monetary parameters as well as the traditional price drivers such as demand, stock, and supply. Quarterly datasets have been used in fitting the price forecasting model through a variety of econometric techniques such as regression, auto-regressive integrated moving average, and value at risk. Heavier weightings are put onto the recent 2000-2006 relationships among the parameters and with the price of WTI. The period between 2006 and 2009 was used to test the validity of the fitted model, which is adjusted seasonally to account for the evolving relationship between and influences on oil price from the individual factors under changing macro-circumstances.

The short-term quarterly forecast (for the next three years) and the long-term nominal (annual outlook to 2030) were launched in November 2010. Early in 2011, Hart will include differentials (heavy versus light crude), and the forecasts will be used with AspenTech’s Process Industry Management System model to forecast refined petroleum products (gasoline, diesel, and jet fuel).