Update: Feb 10th
At the special workshop on Feb 12th at 5pm the head of finance for the BGSD will share a power point presentation. For full disclosure here is a link to that presentation without any comments or questions: https://www.boarddocs.com/wa/bgps/Board.nsf/files/AVPVK67842BC/$file/BGSD%20Presentation%202-12-18.pdf
We added one more question and response below from a Feb 12th email exchange
Update Feb 9th
After reading the report from Ed Hovee I sent him a list of questions that he kindly responded to. Here is a link to the questions and his responses: BGSD Student Enrollment 10 Yr Forecast Memo from Ed Hovee Feb 7 2018 – Richard Rylander Response
After reading his responses there was one more question that occured to me that I sent him. Following is the follow up question I sent and his response. Nothing was changed. This is a copy/paste. I’ll leave it to you, the reader, to interpret the questions and responses and make your own judgements. Please note that the original report is further down and should be read to get the background. Again, to be clear, the document is from Ed Hovee’s group; the highlights are mine and the comments/questions in the document are mine.
The term “modeling” is used. In my view there is no forumula. He uses a mix of factors and subjectively weighs them to come to a conclusion. His conclusions may be accurate…or not. The only way to know is to wait 5 – 10 years and then look to see if he was right or not.
After reading your responses one other topic comes to mind that I would appreciate your thoughts on. When I look at any projection/model I attempt to determine if it is predictive and accurate. To do that I go back in time and then apply it. If I go back say 3 years (which has been a strong period of residential growth) what do we see? Your Baseline suggests ~1500 more students over the next 10 years with the Alternative suggesting ~2700. Given the strength of the last 3 years is there any reason to believe your model shouldn’t be predictive?
Using the student enrollment data reported by the BGSD to OSPI in May of each year we see:
2016/17 = 13,498
2015/16 = 13,477
2014/15 = 13,564
2013/14 = 13,263
From 2013/14 to 2014/15 the change = +301
From 2014/15 to 2015/16 the change = -87
From 2015/16 to 2016/17 the change = +21
Net change in those 3 years = +235
The baseline model suggests +150/year (on average) = +450. The Alternative model would have predicted +270/year or +810 for the most recent 3 years.
The real vs. projected = 52% (using the Baseline model)
Given that the past 3 years have seen the most activity in a decade and the actual student numbers are 1/2 the predicted how would you assure the reader (and public) that the baseline (or alternative) can be predictive of the future? (Factoring in that we are at year 10 of a “recovery” which, historically, would lead to an economic decline in the coming years) Shouldn’t your model be predictive if it is used historically? Is there something about the coming 10 years that makes it more likely to see double the student growth of the past 3 years?
In the attached document is the residential (SF + MF) issued permits for unincorporated Clark County along with the numbers from the City of BG. (I requested these and have the emails with the data on file) As you will see 2017 represents a decline in permits which would suggest a slowing rather than continuance or growth in student population. Does this data, in conjunction with your projections, support the model or raise questions? Given your expertise your views are important. If I am misinterpreting data or arriving at erroneous conclusions please help me understand.
Response from Mr. Hovee:
Your suggested approach of testing the reliability of a forecast model from historical data can be a worthwhile exercise. However, in this case it appears you are using data and an approach that is not fully consistent with the modeling methodology utilized with this forecast process:
- It is less useful to draw from the all-years approach to student generation (of 0.45 students per household) as the most recent 6-year experience indicates a lower generation rate (0.38 students per household) based on actual matching of assessor’s data for new construction with students now living in these recently constructed homes and apartments.
- Also noted is that the forecast is not driven by new development alone. Other factors of importance are birth rates and age-cohort specific grade-to-grade enrollment changes. For example, a large cohort of primary students will offset what was previously a declining number of middle schoolers. This is what can be expected in the next three years. Middle school enrollment declined in the last three years but is forecast to increase in the next three years due to larger incoming primary classes. However, the smaller middle school classes of the last three years will likely now likely cause high school enrollments to decline somewhat in the three years ahead.
The base case forecast scenario involves relatively modest enrollment growth across the full K-12 system – increasing by only 155 students in the next three years to 2020 (or just a bit over 50 students per year). It’s not until about 2021 that base case enrollments are expected to move up more sharply (once currently relatively small middle school classes complete their K-12 education up through the district’s high school grade levels).
Eric Hovee – Principal
Feb 12th: Do you have a history that shows the accuracy?
Over the last three forecast cycles for Vancouver, forecasts have averaged to a 1.4% deviation from actuals. The most problematic forecast was in 2012, underestimating the strength of the long-awaited rebound from the recession as it finally took hold in Clark County.
In the Fall of 2017 the BG School Board signed a consultant contract with Hovee Consulting in the amount of $15,000. Hovee was tasked with preparing a report projecting the 2017-2027 student enrollment for the BGSD. Hovee has done this for other SD’s (most recently Evergreen). We filed a Public Information Request when we learned of the report and just received it on January 31st. In this blog article we will (1) provide the report; (2) analyze the report and (3) list questions that still need to be answered. We suggest that you read it and come to your own conclusions.
This report will affect your vote as it relates to the claims of “growth” being a driver for supporting the bond.
Note: On February 12, 2018, at 5pm, there will be a “work session” at the Lewisville Campus (north of the Post Office). The public will have an opportunity to ask questions. If at all possible and if this report is of interest to you please attend. If you can’t attend and you have questions or comments we would be happy to be your conduit. If you will email us with the subject line being: Feb 12th Hovee Work Session, we’ll deliver your comments in writing. Our email address is: swweducation at gmail.com
Now…on to the report
Report: BGSD Enrollment Forecast Report (1-30-18)_Rylander_Notes You will see what looks like little chat bubbles in the document. You can “mouse-over” or click them to read the comments. This does require Adobe Acrobat to read. The highlighted areas are ours for emphasis. If you would like us to email you a copy of the report please drop us an email and provide your email address.
Analysis and Comments:
- The report covers the recent history of enrollment going back to 2007
- They discuss their methodology for forecasting
- They provide two (2) scenarios: A baseline and an alternative. They say that baseline is conservative and that alternative is more aggressive.
- They indicate a small annual growth rate of slightly over 1% for the baseline case
- They note that any growth will be smaller in the first 5 years of the report (2017-2022)
- They indicate that a drop in elementary and middle school enrollment may occur in the next 2-3 years
- Prairie HS may see a slight increase but BG HS will see a drop in enrollment
- This report tells us that there is NO need for a new campus on 152nd for the next 10 years
- Boundary line shifts can take care of much of the southern end growth
- Birthrate is a major driver and it will continue to drop with a rebound in about 10 years
There are some holes in the data and projections. On January 31st we sent the following list of questions/issues that need answers:
- What about GMA Growth Controls which inhibit subdividing land?
- This report assumes that the past 10 years of economic recovery will continue. The likelihood of that is virtually non-existent. A normal recovery is 8 years and we are now at 10. Their projections appear to contemplate this will continue for another 10 years for a total 20 year recovery which is not realistic. Thus, what data did them contemplate relative to future grown prospects? There is no evidence that they spoke with any economists nor consider factors suggesting we are nearing another peak (given property values now at or above the last peak; a stock market hyper valued and interest rates climbing to name but a few).
- There is no discussion about affordability. What impact will higher prices (including assessed values) have on their forecasts?
- This report doesn’t address the “empty nest” and those who do not move/sell once their children graduate. Given that ~62% of the households have no children do they see that changing and if so how and will that impact their projections?
- There is no “down” projection. This is important because if we go through another 2007/2008 recession will the numbers mimic the last one?
- What impact will rising interest rates have on construction? (The Fed is projecting 2 or 3 rate increases in 2018)
- What impact has (and will) home schooling and private (including religious) schools have on the projections. What has been their past impact and what does it look like for the future?
What say you? What questions or comments? Let us know.
Again, Feb 12th at 5pm at Lewisville if you have questions or email them to us and we will submit them. You can have your name on your comments or they can be anonymous.