In finance, be cautious of anyone who uses historical correlation to back up their argument. In shipping, just flat out run from them. Shipping’s notorious Baltic Dry Index, which is an index of spot rates for shipping dry bulk commodities such as coal and iron ore around the world, achieved deathdefying heights and then, well, death-causing lows, in the course of 2008, falling 90% from its peak, and attracted a lot of attention in the process both on the way up and down. The BDI meme is still alive, especially given a recent rally, and we have quite a few people claiming it as a quality indicator, or even the best indicator (sheesh) for the direction of stock markets or the world economy. Unfortunately, a lot of smart people misunderstand what the BDI represents.
Not a Reliable Leading Indicator – The BDI
Recently, the French economist and commodities guru Philippe Chalmin touted the index at a conference in Cannes for example. Closer home to the web, Bespoke Investment Group raised the question of whether or not there is a relationship between the stock market and the BDI here. And there are others all over the web. If anyone can send me a major house’s piece on BDI/market correlation I’d be elated.
The Baltic Dry Index is currently riding an eleven day winning streak during which the index has gained 43%. Year to date, the index is now up 228%. Given that it is a measure of shipping rates, the increase in the Baltic Dry Index is regarded by many as an important indicator of an improving global economy. How this translates to the stock market, however, is unclear.
Over the long term (since 1985), the Baltic Dry Index and the S&P 500 have had a positive correlation of 0.5 (1 = perfect correlation, -1 = perfect inverse correlation). Like everything else recently, though, that relationship has been turned completely upside down. As shown in the chart below, the S&P 500 and the Baltic Dry Index have been moving in opposite directions for most of 2009. As one has risen, the other has declined, and when one falls, the other seems to rise. Looking at the correlation between the two shows that year to date, they have had a negative correlation of -0.4, which implies a significant inverse relationship between the two.
Bespoke sort of leaves things hanging, but seems to imply that we shouldn’t care about the BDI when wondering where the stock market will go. I hope so at least.
Because there is no way the BDI is a reliable leading indicator of much, if anything, let alone the stock market. If one wants to track commodities demand as a signal of the economy, just look at the actual nominal commodities demand. But before we go here, let’s confront correlation studies such as the above and as frequently exist elsewhere.
I have actually done a rather in-depth BDI correlation study in the past during my tenure as shipping analyst at Citi. One important thing to note about correlation is that you need to look at correlations for many different time periods because if you just calculate a single correlation then it will be highly dependent on your start and end dates. For example, in the Bespoke correlation above, they use a period where the US markets went up over 20+ years and when at the same time the BDI also went up. Surprise, surprise, the correlation is positive over this long period. But if you dig into the details, then one can find many examples where correlation was very negative, and to be fair they show this with their 2009 data. In addition to this, as I have found by doing correlation studies in the past on the BDI vs. shipping stocks in particular, the outcome for any sort of trading strategy based on correlation can differ greatly from your long term correlation. Correlations can reverse for long enough that you are wiped out if you follow some sort of period correlation trading strategy.
Now to make things more clear, I think its even best to step back from numbers or historical correlations and think about some very important differences between the nature of the BDI and the nature of a stock price for a company or market. The BDI measures the rate to ship something around the world, ie. the COST to ship something around the world. On the other hand, the stock of a company represents ownership of a productive asset, a company. (well, at least usually they are productive!) A company is seeking ways to be more efficient, grow the size of its business and generally increase its value.
What this means is that while shipping rates should track the cost to ship something around the world, which actually in the long term should be declining on a real basis due to human progress (and has been), a company (when managed properly), or collection of stocks, ie. a stock market, should actually be increasing in value over time on a real basis due to human progress. If we want to think about the value of all the companies in the world versus the cost to ship a ton of ore around the world, in the long term the value of global business will be increasing while the costs to float ore will be falling. For shipping costs to rise in the long term, humans would have to somehow become less and less efficient at transportation over time. Unlikely.
Why do shipping rates seem to jump all over the place? Due to near term supply of ships versus demand for commodities. Its just a matter of bottleneck problems. If rates go up it can come from either of two things, not enough ships at the time or too much commodities demand at the time. In a situation where ship owners match demand, which over the long run they will, then rates won’t sky rocket and will just track their costs plus some margin for their effort.
Now in regards to the BDI as an economic indicator, and there are tons of similar views out there, I have just pasted an example below, where economist Susan Lee says the following.
I suggest you watch an index that will tell you when the world economies are starting to perk up and when trade conditions are really starting to ease. It’s called the Baltic Dry Index. Essentially the Baltic Dry tracks the average daily price for shipping dry bulk like coal, iron ore, wheat and soybeans. There are three things that make it such a good leading indicator. One, the index looks at raw materials, so it captures activity at the very beginning of the production process. Two, it looks at ocean shipping, so it reveals what’s happening to international trade — the critical driver of global growth. And, three, the shipping business depends heavily on credit, so the Baltic Dry indicates whether credit is tight or loose. Back in 2005, when the world’s economies were just fine and credit was abundant, the Baltic Dry looked like a powerhouse. But it peaked in May of 2008. And it’s been heading almost straight down ever since — losing about 90 per cent of its value.
I don’t mean to pick on the quote above alone, a lot of people are making the same argument, the quote above was just the most convenient at hand.
But essentially one problem with using the BDI for economic forecasting is that the BDI could feasibly go up in an environment where commodities demand was shrinking, if the supply of ships was shrinking even faster. These would be negative economic factors. This is because the BDI’s value is not solely driven from the demand side. To me, it makes far more sense to just look at nominal demand for commodities rather than the BDI since the BDI has the complicating factor of vessel supply growth one needs to consider. The other thing is that the BDI is a measure of spot rates for dry bulk commodities consumers who, generally, are in the near term forced to pay whatever it takes to get their raw materials shipped (A steel plant needs to keep operating despite some higher ore transportation cost). On the flipside, vessel owners are in a similar boat (no pun intended), and in the near term are generally forced to take whatever rate they can get to fill their ships. (A ship sitting around is just a cost, ie. fixed costs are high, thus using a ship at a loss is usually better than not using it at all)
Because of these inelastic characteristics of supply and demand, and since the BDI is a measure of spot rates, the BDI is thus absurdly volatile. I can explain why via the following simplified example, which I used to use frequently at Citi.
Imagine you have 10 loads of iron ore and 9 ships, and that every load of iron ore must be sent no matter what while every ship must be filled no matter what. Imagine the bidding war between those 10 iron ore consumers fighting over just 9 ships. Shipping cost would skyrocket since they all need to ship regardless of cost. Now imagine if a week later two more ships enter the market. Now imagine the bidding process. Suddenly the tables have completely changed. You have 11 ships, that all need to be filled no matter what, and only 10 loads of ore. Shipping rates would plunge, despite a period of just a week passing by. This is, in a simplified nutshell why the BDI is so volatile.
Now, add to this the fact that predicting ship supply and commodities demand has a pretty high margin of error, at the same time remembering how sensitive the BDI is to small mismatches due to the inelastic nature of its underlying supply and demand, and you quickly realise that predicting the BDI is a fool’s game and also that it is not a reliable forward indicator given that it is a spot rate index in a market where both sides are basically forced to close a deal due to high fixed costs. The BDI is measure of supply/demand mismatch at the moment, and can change drastically on a dime. Its little else beyond this. It hit its peak not when the global economy was in its healthiest state, but in early 2008 when things were already starting to come apart, but Chinese commodities demand growth still had some steam and just kept outstripping stagnant vessel supply growth. For a moment. And then it all collapsed. And BDI correlators got annihilated in popular stocks such as DryShips (DRYS). Thus, let’s hope that we put to rest any talk of the BDI as a reliable leading indicator, even if in six months someone datamines some new, latest correlation.
This post originally appeared at Reserach Reloaded.