Energy Production Committee

Understanding the Petroleum Industry

Price Overview | Demand | Supply | Trade and Imports | Refining | Stocks
What's Hot: Demand | Supply | Trade and Imports | Refining | Stocks

Stocks

• Stocks are critical to an understanding of oil market behavior because, as in any commodity market, stocks and price movements are inextricably intertwined. Oil delivered into a market can come from "fresh" sources -- for crude oil, these would be production or imports, and for product, refinery output or imports -- or from the market buffer, inventories. At any given time, stocks supply only a tiny amount of the oil delivered to the market. However, they provide information about whether a particular market segment: a region, a product, a company, etc., has too much, too little, or just the right amount of oil. They are a constant indicator of the relative availability of the next delivery. Thus, when stocks are low in a particular market, prices are sensitive on the upside -- likely to show upward movement and thus encourage extra supply or reduce demand (whether through fuel switching or other means)-- and when stocks are high, prices are sensitive on the low side -- likely to fall.
• Because of this linkage, the market focuses very closely on both reported and projected stock levels. The Energy Information Administration and American Petroleum Institute data on U.S. stocks, the world’s only non-proprietary weekly data on inventory, and the International Energy Agency's Monthly Oil Market Report, the most widely available perspective on world stocks, are particularly keenly read.
• Stocks are not all created equal with respect to their roles in moving prices. Typically, only 10-15% of the world’s stocks are usable by the industry when and how it chooses (see below). It is movements in this small subset of the world’s oil stocks -- "discretionary" stocks -- that link to prices. Because of this, an inventory increase or decrease that appears small relative to the total stock level may result in a large price change, because the stock change represents an important difference in the usable level of inventories.
• The vast majority of the world’s stocks are necessary -- non-discretionary. They exist because the huge, global, oil supply system cannot physically operate without them or because governments have mandated them. Non-discretionary stocks include
  • Oil in transit from supplier to purchaser, but outside the routine surveys of stock volumes. This could be oil in a tanker on its way from a producing country to a refiner, or even gasoline in a tank truck traveling from a supply terminal to a gasoline station.
  • Oil that cannot be used without also impairing delivery reliability. Pipeline "fill," the amount of oil that must stay in the pipeline (analogous to plumbing pipes) is such an example. Pipeline fill is enormous in the United States; the Alaskan pipeline alone, for instance, holds about 9 million barrels of crude oil, always. Another is "tank bottoms," the amount of oil below the outtake valve at a terminal tank. These volumes are a function of the system's physical characteristics.
  • Oil whose storage governments have mandated to allay concerns about supply security. These "strategic" or "compulsory" stocks may be government-owned, as in the United States, or a combination of government-owned and required industry-owned stocks as in Japan and Europe.
  • Oil held to smooth flows that are neither homogeneous nor constant nor totally predictable. For example, continuous flow from a well or pipeline has to be meshed with the discrete volumes required by tankers; crudes and products have to be blended to meet the specific needs of refiners and consumers; deliveries can be disrupted or demand increased by bad weather, etc.

• By the end of 1997, stocks held worldwide by governments and by significant industry players, such as producers, refiners and terminallers, totaled around 6 billion barrels -- the equivalent of 90 days of consumption – and were worth around $100 billion dollars.

• In addition, small industry players, like gasoline retailers, and oil consumers, like petrochemical plants, manufacturers, homeowners that depend on oil for heating, and almost every vehicle owner, held another 1-2 billion barrels of stocks. Data collection challenges mean that these stocks, known as secondary and tertiary stocks (see What's Hot), are almost always excluded from estimates of stock levels, as they are here.
• In the decade after the 1986 oil price decline, world oil stocks barely changed from one year’s end to the next. Even including the rise in response to Iraq’s invasion of Kuwait, the range from high to low was only 5%. This stability is deceptive, because it applies to the total, but not to the constituent parts.
• After 1986, the volume of usable stocks declined sharply, especially relative to demand, which had been growing (see Demand). In 1987, usable stocks were equal to 17 days consumption. By 1996, they had dropped to less than 500 million barrels worldwide, half of their 1987 level, and were sufficient to supply the world for just 1 week. Therefore, prices rose sharply in the wake of the market's supply and demand surprises of late 1995 and early 1996, reaching their highest levels since early 1991 and Desert Storm.
• Subsequently, the world changed dramatically: Iraq returned to the international market; the Asian crisis undermined demand; and OPEC production, sanctioned by higher quotas, soared. These trends allowed global usable stocks, and prices, to return to more normal levels in 1997.
• Stocks, however, continued to rise well into 1998, both in terms of volumes and the days cover provided by usable stocks, setting the stage for this year's unusually low prices. Adherence to the two rounds of production cutbacks agreed to in March and July (see World Supply) is critical to any reversal of the build in usable stocks that occurred during the first half of 1998, and therefore to price recovery.
• The seasonality of world demand – an average 3.5 million B/D swing between the highest demand quarter, the fourth, and the lowest, the second -- is much more pronounced than that of world supply. Consequently, world oil stocks are seasonal, with a seasonality that mimics demand’s but is more muted. Thus, stocks are generally drawn down most rapidly in the middle of the winter, and built most rapidly in the spring. It is these stock movements that create the tendency for world prices to be strongest in the fall and weakest in the spring.
• The Organization for Economic Cooperation and Development (OECD) has the world’s most reliable stock data. All OECD member countries have monthly reporting systems for oil stocks, but reporting is patchy everywhere else. Thus, world stocks are derived primarily by subtracting world supply from world demand, and stocks outside the OECD are estimated by subtracting OECD stocks from world stocks. Non-OECD stocks can therefore be distorted if there are problems in any of the data on world supply or demand.

• Although world stocks were relatively static over the decade after the third oil price shock, OECD stocks climbed steadily, at a rate of 40 million barrels a year. This growth was all in government mandated stocks: stocks required by governments because of national security concerns. Some governments moved to shoulder more of the compulsory stock burden they had previously left to the industry, while others, new to the OECD, started to build aggressively to meet the organization’s higher requirements.
• However, this growth masks a sea change in views about how high compulsory stocks should be. Several governments have started backing away from their traditional posture of setting compulsory stock targets well in excess of those required by international agreements. This reflects a reassessment of the costs and benefits of holding more compulsory stocks than necessary, in the light of lower prices, strong non-OPEC supply growth, an ever weaker cartel, and a more flexible global market.
• Another factor, of course, has been the political preoccupation with balanced budgets. The U.S. has been in the vanguard on this. Since 1996, it has conducted several sales of crude oil from the Strategic Petroleum Reserve (SPR), reducing volumes by almost 30 million barrels from their 591 million barrel peak.
• There is an equally sharp difference in trends when OECD stocks are divided by region instead of by type. While the U.S. share of the growth in inventories between 1985 and 1995 got progressively smaller, its share of both the 1995/96 stock decline and the 1997/98 stock build was disproportionately large. The next section explains why.

• U.S. stocks (excluding the SPR) peaked in 1981. By 1996, stocks were not just substantially lower than in 1981, but were suddenly reaching or breaking through lows not seen since the 1970’s. On a volume basis, this decline followed an extended period of relatively flat stocks. On a days supply basis -- the number of days of demand stocks can meet, calculated as stocks divided by demand -- it marked an acceleration of what had been a pronounced and relatively steady reduction in both crude and total product inventories.

• Stocks cost money to hold, whatever the reason for holding them. At a minimum, this is the cost of the working capital tied up in owning the oil. A function of interest rates and the oil price, this cost has ranged from 8-20 cents per barrel per month in recent years. If a company has to hire storage, that adds between 15-25 cents for crude and 30-40 cents for product to this cost. Thus, every 1 million barrels of oil a company holds in inventory increases its overhead by $1.0-$7.0 million a year.
• Because of these costs, as improving financial performance has become increasingly important, so too has keeping stocks at their most cost-effective level. Companies’ stock holding policies are therefore now continuously updated in line with structural changes in the marketplace. In the U.S. in particular, structural changes have been plentiful. Of these, declining production and industry restructuring are the main ones responsible for the post-1981 downward trend in stocks.

• The extended decline in crude oil production (see Supply) reduced the volumes of oil in the parts of the distribution system that specifically relate to domestic crude, such as waiting at the lease to be collected by tank trucks, or in the case of Alaskan North Slope, the only waterborne U.S. crude, tied up in long voyages to the East or Gulf Coast.
• The wave of oil industry restructuring, via mergers, acquisitions, closures and joint ventures (see Refining), scaled down the U.S. supply and distribution infrastructure, i.e. cut the numbers of refineries and terminals, and, in some deals involving overseas producers, also severely restricted the number of grades being run. It therefore significantly reduced both the numbers of storage tanks and the total volume of stocks being held in them.

• The sharp increase in crude imports might reasonably have been expected to lead to rising stocks. However, there was an important structural shift: the switch to short haul crude imports (see Trade and Imports).

• With the switch to short-haul crudes, refiners can operate with lower stock levels. Firstly, stocks of a particular grade are equal on average to at least half the typical volume delivered, and short haul crudes are delivered in smaller ships than long haul crudes. Secondly, inventory theory shows that as delivery times shrink, so too should the stock cushions used to protect the supply system from breaking down, because resupply is nearer at hand.
• There is less price risk to the buyer on short haul crudes. To minimize discounts to its customers yet maintain its share of the highly competitive U.S. market, Saudi Arabia transformed itself into a pseudo short-haul supplier by setting up a supply base in the Caribbean. Just like companies that switched to real short haul crudes, Saudi Arabia’s customers cut their stocks.

• The first short-term factor that caused stocks in 1996 to drop sharply below the established, declining trend was Mexico’s dramatic, late 1995, 40 million barrel production loss because of Hurricane Roxanne. Thus, the U.S., the destination of over ¾ of Mexico’s exports, started 1996 with unusually depressed levels of stocks, particularly of crude oil on the Gulf Coast.
• A return to normal stock levels was prevented initially by a cold snap in January and February and then by the high prices, particularly for readily available or "prompt" crude, that characterized 1996. These high prices signaled that the world oil market was very tight. Since the U.S. is structurally more responsive than other countries to such prices, not only was the U.S. unable to spread the impact of the Mexican shortfall gradually among other nations but it also had its stocks drained as the market tried to handle the rest of the global shortfall.
• The U.S. is more responsive because its strategic stock obligations are met predominately via the SPR, and have no bearing on how the companies set their stock levels. In many other industrialized countries, the strategic stocks the industry has to hold are frequently so large that they, and not the infrastructure's features, define the minimum level for company stocks. In such countries, the industry cannot reduce stocks, regardless of what the price signals indicate.
• In addition, thanks to the possibility of reducing the risk of inventory-holding by hedging with NYMEX futures contracts, the U.S. is the safest place to build discretionary stocks when the market encourages storage with higher expected prices in the future (see What's Hot), as it was prior to 1996. Consequently, when the 1996 market switched to a sharp and enduring expectation of lower prices in the future, thus discouraging storage, there were more discretionary stocks to draw down in the U.S. than elsewhere.
• Underscoring the sensitivity of discretionary stocks in the U.S. to swings in expected future prices, stocks in the U.S. have been trending up since 1997. By mid-1998, they had climbed to levels not seen on a sustained basis since the first half of the 1980’s. The 1996 worry that the system would run out of oil was replaced in 1998 by the worry that it would run out of tanks. Neither happened. Providing prices are not artificially constrained and panic is avoided, neither should happen.
• Although historically, stock swings used to be more concentrated in crude, product stocks have been the more responsive to recent price gyrations. They dropped much more dramatically than crude stocks in 1996, incidentally also confounding the widespread view that the introduction of specialty grades (see What's Hot) had raised the stock lows for gasoline and distillate. Both set at least 15-year lows that year. The situation has been inverted in 1998, with product stocks rising much more sharply than crude stocks. This product stock responsiveness is good news for the consumer, because it tends to mean lower product prices, but bad news for the refiner, whose margins get squeezed even harder.
• The responsiveness reflects the greater integration of futures transactions into the refining and marketing businesses. Refiners, who always want to keep their runs and their utilization rates as high as possible to spread their costs, hedge any product stockbuilds on the NYMEX when markets show higher future prices. Thus, when markets are in surplus, as they were in 1994/95 and have been again in 1997/98, U.S. refiners aggressively convert surplus crude into surplus product, knowing their hedging leaves them with little risk. When the markets reflect an expectation of falling prices, as they did in 1996, then the consequent drawdown is predominately of products, because it was products that were built previously.
• There are significant regional differences in stock levels. Taking 1997 as an example, stocks ranged from over 400 million barrels, or 40% of the total, in the Gulf Coast (PAD 3) down to under 30 million, or 3% of the total, in the Rockies (PAD 4).

• This variability reflects structural differences between the regions. Demand is obviously a key one of these differences, as the smaller regional variability in days cover than in absolute volumes confirms. However, since days cover is not constant, there must be other important differences too. There are two in particular, both of which relate mainly to infrastructure.

• Crude stocks provide less than half the days cover on the East Coast (PAD 1) that they do anywhere else because it is the only region with no crude oil trunklines and its crude oil production is insignificant.
• The Gulf Coast has 2/3 higher product days cover than elsewhere because it is the swing refining region for the U.S. (see Refining), and also its main gas processing and petrochemical region.

• Reflecting the region-specific importance of heating oil and LPG as winter heating fuels, there are just three stock series that are highly seasonal: distillate fuel oil on the East Coast and natural gas liquids on the Gulf Coast and in the Midwest. There is also a substantially more muted, and inverse, seasonality in gasoline stocks in all regions, and in asphalt stocks in the Midwest.
• How much oil is built seasonally, and when the build starts and ends, is affected by the behavior of the forward price curve, especially in the Northeast, which is both highly dependent on the seasonal build of distillate fuel oil and the site of "delivery" of the NYMEX futures contract. When there was a strong inducement to store, based on expectation of higher prices, as there was through the second half of both 1993 and 1994, distillate stocks in PAD 1 set 8-year peaks, despite the steady downward trend in demand. But in 1995, when the market saw the opposite signal -- lower prices in the future -- at the start of the winter, and in 1996, when the switch came during the summer, the stockbuilds were cut off prematurely. The winter 93/94 peak for East Coast distillate stocks was 25 million barrels higher and two months earlier than the 1996 peak.
• In 1998, with the return of a weak market today and an expectation of a stronger market tomorrow, distillate stocks on the East Coast soared to a 16- year high that was 60 % higher than their 1996 peak. Despite this leap, the question about how high these stocks can go remains unanswered.

For more info, go to What's Hot in Stocks

• Secondary and Tertiary Stocks
• Contango, Backwardation and Stocks
• Specialty Products and Market Fragmentation, from the Inventory Side

To see graphs, go to Stocks Graphs

< back to top >