Coffee Traders' Forum - A Discussion about Coffee Futures Trading

Coffee Traders Discussion Forum

Coffee Traders' Forum

Rumor / Reports / Stat Forecast vs Reality

In the 90s I had the opportunity to work for the FNC, being part of a small group (we were only 3 people in addition to an external consultant), which had the objective of developing a scientific methodology (statistics) to forecast the size of the harvest in Colombia.

Based on that first-hand experience, I can assure you that the scientific basis with which FNC works today is ideal (Scientifically speaking). It is a sampling process where areas that must be visited and measured (green bean count) are randomly selected. This type of construction produce a forecast (6-7 month prior to the actual crop) with an error of +- 5%.

Even tough on paper the forecast is robust, with a tolerable margin of error, in practice things get complicated. Bear in mind that Colombia is a mountainous country and coffee is planted in distant areas, of difficult access, steep slopes, making not reasonable to expect that data collected works be extent of human errors.

Unfortunately, there is no way to calculate the size of this human error source, because it is basically unpredictable event (unknown). This means that the error that theoretically is estimated at 5%, in reality is greater in an unknown quantity.

In short words, in practice it is impossible to develop a statistical forecast with a totally known margin of error.

Fortunately, in practice we also found that the margin of error did not go off in an uncontrolled manner and the forecast was accurate especially in determining a trend of change% with respect to the previous crop. That is, the sampling design served to obtain valuable information.

If it is so difficult to obtain a precise measurement of a future crop based on statistical calculations, one should consider what can be the margin of error of other types of non scientifically estimates.

After having worked on this defying theoretical job, I worked in other positions more related with commercialization, being in charge of coffee purchasing operations in several Latin American countries and based on this complementary experience I can tell you that, I am 100% sure the only reliable (non subject to errors or surprises) indicator of the size of a crop, is the actual coffee flow that arrives day after day at the door of your warehouse.

For me, there is a huge distance between a rumor, a report and even a technical measurement of a crop and the reality. This is key for a physical buyer.

But talking about futures prices, it is well known that perception is reality, if market participants believe in a rumor, report or statistically constructed forecast, the price will react accordly. Example the 60 millions bags expected from Brazil 18/19, at this stage of the crop nobody can know for sure if it will be a fact until the flow show up, but 60m is reality these days in the futures arena.

Messages In This Thread

Rumor / Reports / Stat Forecast vs Reality
Re: Rumor / Reports / Stat Forecast vs Reality
Re: Rumor / Reports / Stat Forecast vs Reality
Re: Rumor / Reports / Stat Forecast vs Reality