Week 22 crop size estimates

In this weeks estimates you will note that the corrected model is now above the uncorrected model. Now I've mentioned that from the begining of the season that conditions decline more often then they improve. The model is calibrated to the 25th week in my growing year (for the most part) and so they should be very close or exact for most states in that week. However, it appears that instead of a lowering of estimates to correct for condition changes, for some states over the next week or two there is a positive correction. This would suggest that conditions improve between now and week 25 on average. Perhaps this indicates a certain pesimism at this time of the year or simply an statistical discrepency, but it indicates that the estimates for this week would on average underestimate final production.

Week 22 spreadsheet (3mb)

Week 21

Greetings from Los Banos Philippines. I have moved across the great firewall of China and can now post my blog directly, although the time difference means a slight delay. Being here has kept me from making an first hand observations about the crop in the US and I'm separated from the buzz, so I was quite surprised in the improvement in crop conditions this week but I should say here that I don't take any one week's movement very seriously as other factors have been know to push the condition ratings for a short amount of time.

In corn and soybeans it was a nearly universal held or improved yields. Illinois did show a very slight decline in the uncorrected model, but the western half of the corn belt, NE, KS and SD showed above average gains. The models are clearly converging as we approach harvest. Cotton, as is common, is a mixed bag where the distributional effects can be substantial as abandonment moves around. The corrected model held flat, as the needed correction declined, the uncorrected model which rigidly tracks week to week condition movements decline, but only modestly for production. I'm still quite a bit higher but I've been getting emails which suggest I'm probably on the correct side of USDA, but so far nobody will anonymously post supporting comments so you have to take my word for it.

I added the corrected model after some discussion with others about how conditions tended to evolve over the season. This year may be one of those years where it wasn't strictly needed. The upward climb of the corrected model isn't due to improving conditions, it is because there is less time for 'something to go wrong'. Should I ever be made of time, I'll go back and historically track the two models in previous years to see how they behave under different situations. In theory, the corrected model should offer the best year to year estimate, however, it could simply be the influence of the 'bad' years such as 1988 or 1993.

Week 21 Spreadsheet (3mb)

Week 20

Hello from Beijing China. The commentary will be short this week as the Chinese are blocking my blog and I've had to enlist some help to log-in and get the results up. I sent out an email to several folks last week after the USDA report which showed the models prediction relative to the USDA. The mode results are in close agreement for corn. For soybeans my yield and production numbers are a bit higher than USDA and my cotton results are higher than USDA and grew this week. You can examine the state by state results, while the totals may agree for corn we get there with a handful of noticeable differences by state (Check out IL and OH for example).


Week 20 Spreadsheet

Week 20 production and yield graphs

Week 19 production and yield results

The uncorrected model stands at 159.3 bushels an acre while the corrected comes in one bushel lower at 158.3. To reiterate the difference between the two models for anybody joining just, the uncorrected model is the estimate if we finish the year with current conditions while the corrected model takes into account the average decline we see in conditions throughout the season. That is to say conditions normally decline as we go through the year. However, although perhaps obvious, it is worth noting that in the really exceptional yields, conditions are flat and don't show this decline. These two models will converge as we approach mid-September and the adjustment shrinks quickly from this point. In the corrected model I'm predicting record yields in some minor states but also a record yield in Iowa of 182.9 bushels an acre. I think this is very plausible given growing conditions in the state, but I'm also predicting a record yield in Ohio at 172.7 bushels an acre, this would be a full 8% above the previous record and seems to be a stretch despite the reported crop conditions, if someone wants to give some anonymous feedback it would be most appreciated. I'm also carrying near records in Wisconsin and Nebraska. For production, I take the USDA planted acreage as given and I'm running just above a 92% harvesting rate nation wide. This gives me production numbers, but corrected and uncorrected that are at the high end or even above the range I've seen floated by industry. The USDA can of course resurvey area and I'll take those numbers into consideration, so while I'm comfortable with the yield, the production number can be adjusted by any change you feel that may be forthcoming from the USDA.

For soybeans the uncorrected yield of 42.7 bushels an acre and the corrected yield of 42.0 bushels an acre would put me in the thick of industry estimates as would my production numbers of just over 3.2 billion bushels. I

For cotton the only information I've seen to date has been comments by O.A. Cleveland and Carl Anderson which suggested I may be a bit high on yield and production but I've seen little else on the crop size estimates since then and I don't seem to be on the list for circulation of other folks models like I am for corn and soybeans, so if someone out there wants to share their results you can send me an email at meyerse@missouri.edu . I'd be happy to see what other's thoughts are on the size of this years cotton crop.

Week 19 Spreadsheet (3mb)

2007 and 2008 corn production results

This week's results are just below in the following post, but let us look at a bit of historical performance in this post. I've pulled these results out of the models for the 2007 and 2008 crop year. The lines represent my model results while the red dots represent the USDA reports in the approximate week they were released. Again, I do not calibrate in any way to the USDA reports and planted area (not harvested area) is taken as given. The models are re-estimated at the end of each year to make use of the additional year of historical data. The first year where condition data in five categories was available nation-wide was 1986 so each additional observation remains quite valuable. In 2007 you can see that the the crop size grew through September in the USDA's estimates and the model tracked this but was at a lower level and kept growing through October, where conditions usually end as harvest is well underway. The November and January USDA numbers showed a decline. The model estimates were competitive through October, showing a good estimate. I've examined the model and will share some additional look at model errors, but the conclusion is the model is very competitive against USDA estimates up to the October estimate, at which point no new information from the model is available but the USDA continues to collect information and the November USDA estimate usually shows only modest revisions in January. We can see that in 2008 the corrected model showed a good estimate of the final production estimate while the uncorrected model tracked USDA estimates quite well. I wouldn't draw a broad conclusion from this as I expect that the USDA estimates would be by their nature a corrected model but perhaps somebody with more insight into their procedure can comment. you can also see the nature of convergence of the two models where estimates in September, where the equations are estimated are identical.

Again the model predicts the final production, does not seek to estimate the USDA estimate nor does the USDA estimate enter into the prediction of the final yield. That is to say, I don't take the USDA estimate in August and try to use this information to predict final yields. I may put up some historical error analysis between the model and USDA performance if there is interest. Also if there are other graphs such as this that you would like to see, leave a comment and I'll see what I can do to put up the desired graphics.

Week 18 Production and Yield results

Modest changes from last week but as the season continues, the correction shrinks and the corrected model estimate rises. For soybeans, my current estimate of less than 43 bu/ac is at the lower end of the estimates I've been seeing. However, the existing record of 43 bu/ac occurred in 2005 under what would probably be considered excellent growing conditions. We have had a few years to improve yields since then. I'm running an Iowa soybean estimate of 52.3 bu/ac in the uncorrected model as well as 47.6 bu/ac in Illinois and 40.3 bu/ac in Missouri. Those are all respectable yields. My Indiana yield, at 44.4 bu/ac would show some weakness in the state. For Corn, there was some modest improvement in conditions. The uncorrected model gives a yield of 159.3 bu/ac while the corrected model result is 157.8 bu/ac. Once again, the uncorrected model gives the result if we finish the year with current conditions, while the corrected model takes account of the 'average' change in conditions. Given the adequate moisture this year across the corn belt, I think it is harder to make the case for a 'normal' decline. You can download the detailed spreadsheet at the link below. New Mexico has not reported conditions yet this week so results reflect carrying forward last week's conditions for cotton.

Week 18 Spreadsheet (3mb)