So, this post may be coming a bit late to be of much help, but maybe it will be useful to someone. This stems from some casual research I was doing. I've been hoping for a contract to forecast Napa County Cabernet Sauvignon. It's such a huge question going forward, as to what extent Napa Cab will follow a steady, upward trajectory and to what extent it is still vulnerable to cyclical effects and macroeconomic headwinds.

To really look at an issue like that would take, in my estimation, 60 hours. It's a tough nut to crack and no contract has materialized. I can't afford to do that just for R&D. But I have done some looking for clues. In doing that research, I forecasted some 2018 Lake County prices. In this post, I'm sharing my forecast for the USDA-NASS Table 6 price for "All Wine Varieties."

Note that this is a purely quantitative forecast. There is no consideration of any actual information from growers. It is based on Crush Report data, macroeconomic data, wine consumption data and Acreage Report data, plus my own, proprietary forecasts of such data. But I don't consider, for instance, if I see a client's prices rising.

I know the following charts are a bit much for just a single price, but bear with me. First off, the charts include the component models. That way I can publicly track the accuracy of each one. They also include an average and a trimmed average. The trimmed average discards the highest and lowest value and this is my "official" forecast.

The bottom chart is the forecast in 2017 dollars. The top chart is in 2018 dollars. I assumed inflation of 2.5%. Which price should you use? I will grade myself based on 2017 dollars, adjusted for inflation. But there are reasons to use 2018 dollars or an average of the two.

I also give the probability distribution, which is identical for the average and the trimmed average. For instance, there is a 2.5% chance that prices will be below $1,233.59, in 2018 dollars and 20% chance that they will be above $2,042.52. For longer-term forecasts, or at least more current ones, these probabilities can be used for a wide range of strategic and financial analysis. They're included here, though, at least to some extent for assessing the model's accuracy when we get the actual prices from the Crush Report.

One last note: Because this is not prepared for actual use by a client, I did not delve into it very deeply. I suspect that had I jumped in deeper, the probability distribution would look different. For one, as you can see the univariate model has a very wide probability distribution, which is the one I used. In actuality, I could have accurately tightened it, based on the stronger models. In fact, it would be pretty reasonable to use the probability distribution for the Demand Model.

Second, I likely would have skewed the probability distribution. That is, they are rarely symmetrical, as indicated in this case. I'm not sure, but I would guess that further research would indicate a higher likelihood that the Crush Report indicates higher prices than lower prices.

Anyways, take a look at the charts below or just make note of the most important number there: I think the price will be $1,793.98.

So, what are each of these models?

The Univariate Model is a simple extrapolation from the price history. It includes no causal variables.

The Macro Models all use various macroeconomic indicators as causal variables. They involve an added bit of uncertainty, since they rely on my own forecasts of these variables. For instance, Macro Model 1b assumes GDP per capita in 2018 of $56,535. If this number is actually higher, we would expect the grape price to be higher and vice-versa. These might be the most interesting models to me, since I am not an economist, but the economy does, of course, influence prices.

The Supply Model incorporates as causal variables, measures of supply, such as bearing acreage of various regions.

The Demand Model incorporates domestic wine consumption measures as causal variables.

By the way, I have now set up a way to subscribe to this blog. At the bottom of the main blog page, you can enter your e-mail address and get automatic updates. If you want the most reliable quantitative analysis about the business of growing grapes, go ahead and subscribe!