Chapter 2 discussed the different views about modeling speed in the literature. Vessel speed op- timization plays an important role in earlier maritime economics models, while later structural models assume a constant speed given charter party clauses stating a fixed speed, weather con- ditions or the small margin that it was believed a ship should be operated under. In the matching
model, the extent to which speed can be optimized depends on whether the ship is employed or not. It was therefore necessary to distinguish between a “matched” speed and an “unmatched” speed. The conclusions drawn from the literature, data and interviews with industry is that speed optimization is important but the extent to which it is being optimized in the matched state depends on the oil trader. Rather there appears to be some stickiness in changing the speed in contracts, especially for oil majors which have speed clauses. In recent years, the simultane- ous downturn of freight rates and higher bunker fuel prices relative to the bunker price priced into the benchmark flat rate has meant that it would be in shipowners best interest to have more flexibility over the contract speed. An optimal speed and constant speed were simulated for the model’s matched state. The optimal speed in the matched state was specified as a function of the trader’s oil revenue, the shipment cost, and the ship’s option value. The sensitivity of speed to this parameter was a function of the specification of the equation, and therefore the speeds from the model reflect this. The optimal speed is sensitive to the assumption that the oil trader has already purchased the oil cargo and therefore has to store the cargo until the ship arrives, thus imposing a cost for the days it has to wait until the ship arrives. In some cases however, oil traders make a simultaneous decision to purchase oil provided there is a ship available to hire. This would have the effect of lowering the ship’s speed. For the ship, it was important to include not only the impact of speed on the current voyage in terms of cost, but also the opportunity cost of time. This opportunity cost was quantified by including the ship’s option value. The optimal speed equation for the matched state contains some opposing forces; a slower speed lowers the trader’s oil revenue and the potential rental revenue to the ship while at the same time lowering the shipment costs. The impact of the ship’s option value depends on its sign; a negative ship option value would mean that it is better to go slower since there are negative profits to be made from the discharge area, while a positive option value would be negatively affected by a slower speed. Ultimately, the results showed that the optimal speed depended on the relative magnitude and the effect of speed on these parameters.
In the unmatched state, ships are unemployed and have to decide what speed to travel in their repositioning voyage to a waiting area. In this case, the optimal speed depends on the shipowner’s decision. This decision is a tradeoff between the repositioning costs (a function of distance, the fuel price, and the fuel efficiency) and the ship’s option value to be in the waiting area. When this option value is negative, it makes sense to go as slow as possible, whereas a positive value will offset the benefits of minimizing costs to relocate.
Anecdotal evidence points to some ships operating at slow-steaming speeds in 2011; ac- cording to Lloyd’s List (2011), “Maersk’s fleet of 11 VLCCs are traveling at speeds as slow as
8.1. Introduction 187 8.5 knots when transiting without cargo to the next load port.” In practice, ships have a time window to arrive at a load area once they match, which is not included in the model. It would not be difficult however to extend the model to include an average time window. In addition, a charter party will provide a speed range rather than one speed (common for oil majors) and can contain an “utmost dispatch clause” requiring the ship to sail at full speed. According to Lloyd’s List (Lloyd’s List, 2011), one major London broker of VLCCs said, “Certainly owners try to have low and high speeds in the charterers’ options of 14.5-15.5 knots and I know a lot of owners say ‘we’re not going to do that’ [when higher speeds are asked for] and say “we’re not going to give you the upside on speed because we don’t get compensated for it.”
Aside from just anecdotal evidence, average speed for the 2011 option value scenario was validated using AIS data as described in Chapter 5. The AIS sample data (Smith et. al., 2013) shows that VLCC ships sailed at an average of 13.24 (9.38-15.55) knots in laden and slightly higher in ballast at 13.52 (9.31-15.92 range) knots in 2011. This contrasts with evidence from Maersk stating super-slow speeds of 8.5 knots in ballast. Both speeds are higher than what the model predicts should be the optimal speed. There are a few plausible reasons why the observed speeds are closer to 13 knots in ballast and not 8. It is well-known that the tanker and bulk shipping companies have expressed concerns about super-slow steaming due to the belief that operating a ship well below its as-designed speed (around 15 knots) might damage a ship’s engine. According to Maersk Tankers, “What we have found out is that during times of difficulties owners have gone down to the most economical speed, which is about 13 knots,” said Maersk Tankers head of crude Claus Gronborg (Lloyd’s List, 2011). “If you go below that speed there are some precautions your crew need to take onboard the vessel but in contrast to common beliefs, no engine modifications as such have to be made,” which Mr. Gronborg said were technical lessons learned from super-slowing steaming within its Maersk-Line container ship fleet. The practice was introduced 18 months prior to 2011 and the company “now decides on a case-to-case basis at what speed each VLCC will travel in ballast to potential crude loading ports in search of employment.” Another explanation for the difference in economical sailing speeds among shipowners is that ships have different engine efficiencies which deliver different savings from slow steaming. This “case-by-case” speed determination by some of the leading companies is however consistent with the model’s forward-looking optimal speed scenario. From a modeling perspective, it is important to understand the implications of both the constant and optimal speed case. The fact that the model predicts slower speeds suggests that traders have the bargaining power and this is justified by the fact that the aggregate market is in their favor. In other words, ships have to consider that a higher a speed is better than no fixture.
8.1.4 Impact of supply side (fuel price increase, physical ship characteristics)