Benzene forecasting

Автор Jeff Eickholt, Commercial Analyst, and Simon Palmer, Vice President, Global Aromatics

When I started working in the US Gulf coast aromatics markets in the early nineties, I was given the simple guidance that US benzene contract prices averaged a buck twenty five a gallon (~$374/t), plus or minus twenty five cents (~$75/t).

Really, I thought there really must be more to it than just this simple plus or minus rule, but that was actually a pretty good ready-reckoner which lasted me well, and for quite some time.

You’d load up when benzene neared a buck a gallon and start to short sell it when it kissed a buck fifty. Short-term price trends were readily identifiable and much easier to jump on and jump off than they are now. This was indeed a simpler time, when the US market possessed a ready-made balancing mechanism in the form of large toluene dealkylation units (THDA or HDA) which could make on-purpose benzene at the rate of several millions tons a year from toluene and T/X mix.

If the benzene price rallied to the upside, these units had the ability to quickly ramp up on redirected and/or purchased feed and capture a few months of good margin before bringing the price back into line. When benzene prices moved to the downside, these plants could be slowed or idled, with the feed being sold to the large merchant toluene market or for blending into gasoline.

The company I worked for at the time had one of the largest THDA units in Europe and we played the game to good effect. For the market-at-large, this balancing mechanism allowed market forces to function with a relatively limited margin for risk, both up and down.

As the years have passed however, this rather effective balancing process has broken down and a brave new world has emerged, where simple rules of engagement have been replaced by more complexity and most definitely greater risk. The returns to the dealkylation of toluene started to falter in the mid-1990’s as new fangled STDP (selective toluene disproportionation) units started making higher value paraxylene from toluene.

These units were integrated with refineries and consumed large volumes of toluene which had previously been supplied to the merchant market. They avoided lower purity toluene and mixed TX feeds, but were able to pay higher feed values than the existing dealkylation plants, and indeed conventional toluene disproportionation (TDP) plants.

As if to add even more salt to the wound, they also produced co-product benzene, much like a TDP, and were intended to run ratably, not to be campaigned as many of the THDA units were by this time. These new plants increased the underlying supply of benzene and started to make the run-ups in price less frequent and of shorter duration.

The benzene they produced was also marketed more as a co-product and in some cases, more aggressively than traditional chemical companies had historically done. All this also happened during a time when refineries on the Gulf coast were long octane and were incentivized to maximize aromatics recovery and export large volumes of mixed xylenes and paraxylene to the rapidly expanding polyester industry growing east of Suez.

Over the course of time, and through periods of much attrition, refinery-based THDA’s started to be shut down on a more permanent basis, and only those integrated to naphtha-based ethylene crackers appeared to have a future in the business. These mostly processed lower grade toluene and fractionated TX streams with much lower alternative values in, and access to, the gasoline pool.

By the time that the industry experienced its first real demand-pull cycle in 2004-2005, most of the swing THDA capability had been shut down, or at least idled, and the market showed some of its new colors, as a dynamic marketplace without a ready balancing mechanism. More and more benzene was being produced as a co-product, with less and less on-purpose. As can be seen from the chart below, one dollar benzene disappeared from the market and four dollar benzene appeared on the scene.

The historic king of the aromatics business was not going to be usurped without a fight!

US benzene monthly contract price history

Since the mid-2000’s therefore, the US benzene market has been functioning largely without a ready balancing mechanism, save for the importation of large volumes from increasingly far away suppliers. Balancing using imports is inefficient and by definition the market rarely achieves the balance it pursues. Molecular tourism has become a phenomenon, with ships passing in the night with the same products onboard, but going in completely opposite directions. So, in this context of continuous imbalance, how does one forecast future benzene prices and how reliable can these forecasts really be?

Forecasting prices for readily traded commodity aromatics has always been a process of defining a likely range for prices, rather than striving for absolute precision. Due to the range of sources, from refineries, petrochemical plants and steel manufacturing complexes, and their nature as both components in gasoline and petrochemical feedstocks, aromatics tend to have rather jekyll and hyde behaviors, switching personalities with some regularity.

If you can predict these changes and switches in direction with any degree of reliability then you have mostly cracked the code to effective forecasting. Benzene is perhaps the most challenging of all the aromatics as it has significant tranches of production which are by-product in nature, large supplies which are as a subordinate co-product, and very little which can be said to be on-purpose.

Paraxylene is one of the few on-purpose products in the aromatics portfolio, and mixed xylenes and toluene are inherently hitched to the PX wagon. They also have a real alternative value in the US gasoline pool which is not subject to legislated limits in the way that benzene is. Prices therefore can move between a floor based on their alternative value in gasoline, and a ceiling based on their value for conversion to paraxylene, to feed the polyester value chain.

Benzene on the other hand, only has an alternative value in gasoline below the federally mandated 0.62pc maximum content by volume. Above this limit, benzene carries a penalty in the form of the cost to process it such that it can be blended (as hydrogenated C6A), the cost to adapt operations not to make it, or the cost to purchase benzene credits. This greatly complicates the floor pricing mechanism, and really pushes us towards using historical price patterns as a helpful predictor of the future.

Aromatic hydrocarbons for the most part all derive from crude oil. The value of crude oil impacts directly, or indirectly, the cost of all hydrocarbons within a refinery and steam cracking olefins complex. The crude oil price is the rising tide which lifts all boats and then the bath water rushing out when the plug is pulled. Even by-product petrochemical commodities, such as benzene, can not fight crude oil prices for any sustained period of time.

It values the benzene which occurs naturally in crude, and the naphtha which feeds the reformers which generate the bulk of the benzene which arises within refineries. These same reformers make the toluene which feeds the STDP units which make the bulk of the co-product benzene. Crude costs therefore fundamentally help drive benzene recovery and conversion cost, and have to be a component in any floor price mechanism.

How do we look at the relationship between the price of crude and prices for benzene? We could look at the absolute spread between the two, or indeed the ratio. As crude is an appreciating commodity in nominal dollar terms, the absolute spread relationship carries some of that appreciation with it. A spread relationship however is a simple concept, which people can easily wrap their minds around.

This would be good if the correlation was high but unfortunately it is not. There is so much short-term noise in the data that it undermines the validity of having an easy to understand link to base the forecast on. The spread chart below provides an illustration of the degree to which short-term noise has impacted twenty or so years of (monthly average) data.

US benzene price spread to WTI crude futures

Now, the alternative is to use the crude to benzene ratio, which in full disclosure, is what we use at Argus to help set our view for floor prices. The ratio accounts for the nominal appreciation in crude prices and irons out some of the volatility in the absolute data. It also provides a more consistent result over time and reflects some of the fundamental changes which are impacting benzene prices.

As can be seen from the chart below, it provides a more “manageable” data set and helps us determine a ready rule to live by. The rule I have used through most of my more recent time in the business as my “Keep It Simple Simon” rule is 1.9 times WTI plus or minus 10pc. This sets the floor around 1.7 times crude as WTI, as illustrated in the ratio chart below. Now Argus uses a slightly lower floor in its current multi-facetted forecasting regime but conceptually the approaches are the same.

Now, this obviously puts a significant onus on the validity of the crude oil price forecast which underpins many of the fuels and chemical price forecasts which Argus generates through the course of business. The process could be said to be more of a “margin and spread” outlook rather than an absolute price forecast, even though price is really the end game here, so if you start with a wayward crude price outlook, then you are certainly going to end up with a directionally challenged product price forecast. All this being said, this is how we elect to generate our base price forecast floor, alongside which we then align our incremental production cost outlook.

US benzene contract as a ratio to WTI crude futures

Now the relationship to crude oil helps determine the cost drivers for underlying benzene supply in the US but we also need to look at incremental supply cost. For this we focus on the conversion cost for the STDP units, which make up the majority of the co-product supply. By using our forward view for toluene feed cost, which has a much more sturdy gasoline-related floor price mechanism, and our forecast for on-purpose paraxylene prices, we can derive our benzene cost to reach breakeven for this significant tranche of supply.

STDP/TPX plants account for around 12.5 pc of the 11 mn t of US benzene capacity currently available for service. On the surface it would not appear to be enough to swing the market but when one bears in mind that most benzene recovery units run at let’s say sub-60pc, then it is clear that they can make a structurally short market become rapidly shorter.

This then is why we include their economics as an incremental mechanism to help us set floor prices for benzene. It also allows us to better reflect the impact of alternative values for aromatics in the US gasoline pool. From the chart below of US benzene STDP breakeven values, it appears that there is some kind of functional floor value, well up until the most recent variance from market norms!

US benzene contract price spread over stop breakeven

So in combining our floor ratio to the crude price and the breakeven value for benzene from STDP, we can derive our underlying benzene “floor value”. We generally take the lower of these two calculated values as the basis for the floor for each forecast period, then we move on to derive a forecast benzene ceiling price. The ceiling price we use is intended to reflect a combination of several ceiling price mechanisms that set the high side of the forecast price range.

Firstly for the purposes of identifying a price ceiling, and likely most importantly, is the arbitrage from benzene regional markets east of Suez. The structural length in country markets such as Korea, India and Saudi Arabia allow for a global balancing mechanism of sorts, if the US market moves strongly and independently to the upside. Given a few months lead time, US imports can move from sub-100,000t per month to 300,000t if the right arbitrage incentives exist.

This magnitude of swing in supply can be sufficient to act as a progressive cap on the market if sustained. We use a judgment of product availability and current freight rates to derive the ceiling value for imports forward two months and beyond. This calculation would be quite straightforward were it not for increasing competition for internationally traded barrels from other markets in western Europe and of course, China.

This competition is taken into account when deriving a ceiling value and a premium over forecast prices in the source regions is set, before freight and other arbitrage costs are added. Another ceiling mechanism is benzene affordability for the major consuming derivatives. These mechanisms can become complex, but revolve around identifying whether any of the main derivative chains might face affordability challenges from the new forecast track for benzene prices.

We use a ceiling mechanism for US General Purpose Polystyrene as the volume risk is significant enough to the benzene market at large to make a difference, but for the other derivatives we use more subjective judgements based on some more generic affordability formulas. We then roll these arbitrage and affordability measures into a ball to derive the general track of ceiling prices over the same forecast period, and as such we now have a likely track of the min-max trading range for prices.

Once we have established the forecast range for the period, we then work up where in that range we expect prices to likely land. This is applied in our “market adjustment” for each time period, which is an adder to the floor price. This adjustment is based upon where we judge a combination of other factors such as short-term supply and demand, seasonal influences, turnarounds and outages, logistics constraints, tax and duty changes and so forth to drive the market upwards or downwards within the range. We primarily use internal Argus data resources to support this more subjective element to the process.

Lastly, we specifically model the polyester business cycle as an overlay to the forecasting process. The polyester business is notoriously cyclical in nature as investments are now of such a scale that one world-scale paraxylene train or complex can be the equivalent of many major individual country’s annual consumption.

Investments also tend to be like London buses, you wait for an hour and then three show up at once. This on-off investment character prompts operating rates to cycle wildly, and in general when the polyester business is in a downswing then benzene prices will tend to rise above their underlying trendline, and then suffer downward pressure when the polyester business cycle is in an upswing.

This may sound counterintuitive but a period of robust demand and growth for polyester means more feedstock paraxylene is required and greater volumes of co-product benzene will be produced. This good fortune for polyester means generally lower profitability for benzene recovery, and vice versa.

At the end of this rather involved process, we emerge with a benzene price forecast, which is worked up to look something like the price forecast chart below. Arbitrage windows are created to allow for necessary product movements between the major regions and the required regional production incentives such as not to redirect benzene away from the chemical markets. As a result, a global price picture is generated, allowing our clients to look forward from wherever their particular vantage point may be.

Global benzene prices

Argus publishes this 24 month price outlook for clients towards the end of each calendar month as part of the Benzene Outlook report . The need to take a stab at forecasting monthly prices out this far means that the process has to be rather involved. I’m not entirely sure how confident anyone can be that the price forecast today for May 2024 will land anywhere close to that level as there is so much that will inevitably change in the interim. It is however a task that calls for rigor and process, and as such the challenge should mean that our forecasts of prices at little closer to hand are as meaningful as possible.

If you want to learn more about what Argus does in the Aromatics consulting space, please do get in touch with either of us, Jeff Eickholt or Simon Palmer.

This article was built with proprietary data, analysis and insight sourced from:
Argus Benzene and Derivatives
Argus Benzene Outlook 
Argus Toluene, Xylenes and Isomers/PET
Argus Toluene and Xylenes Outlook

Похожие записи

30 мая 2022

Argus European monthly methanol CP eases to €515/t

The Argus European monthly methanol contract price (CP) for June was assessed at €515/t on 30 May, down by €5/t from the May price.


Asia-Pacific English Europe North America Нефтехимия Global

24 мая 2022

Chemical Conversations: Chlor-Alkali Outlook

Lauren Williamson, VP for Commercial Products, talks to leading chlor-alkali experts George Eisenhauer in Houston, Bernard Law in Kuala Lumpur, and analyst Dhanish Kalayarasu in London, about the launch of Argus Chlor-Alkali Outlook service


English Global Методология Обзоры рынка Нефтехимия

16 мая 2022

The explosion in octane values

Traditionally newcomers to the aromatics business in the US have been taught that the measure of “blend value” alternatives for aromatic products like toluene and mixed xylenes was based on the spread between premium and regular gasoline, or the “regrade spread”.


Нефтехимия Нефтепродукты North America English