Douglas County School District (DCSD) plans to propose a property tax increase that will impact Douglas County residents. At a high level, this proposal has two parts:
Voters will be asked to approve or reject each part of this proposal in the November 2023 elections.
I created this presentation to help myself properly understand the MLO portion of this proposal, and specifically the teacher salary problem this proposal aims to address, since the majority of the funding would be spent on this. If you're a fellow Douglas County resident, I hope it helps you too.
I've tried to stick to facts, but where I've stated something that is a matter of opinion, I've indicated it with a yellow box like this.
If you'd like to discuss or debate anything in this presentation, reach me via email at .
DCSD's proposal aims to raise an additional $66.0M annually via the MLO increase. Most residents will want to know the bottom line: how much more will residents have to pay out of pocket in property taxes for this? Very roughly speaking, the share of the $66.0M a property owner will be responsible for paying is proportional to the share of the county's total property value that they own. However, several factors make reailty more complicated than that, including:
That said, according to slide 24 of this presentation from DCSD, it is estimated that an additional 5.912 mills will need to be levied assuming an average increase of 35% in total property valuations. This means a residential property valued at $1M after the current assessment increases will owe an additional $399.95.
Enter a property value (after assessment increases) in the calculator below to see an estimate of how your annual tax burden will increase if this initiative passes. You can find the 2023 "Actual Value" for your property using the Douglas County Assessor's search tool here. NOTE: This information will not be stored or seen by anyone other than you.
| Your 2023 actual property value: | ||
| x | Assessment rate: | 6.765% |
| x | Additional mills needed: | 5.912 |
| ÷ | 1000 | |
| = | Annual property tax increase: |
DCSD has published figures stating tax rates will increase $20 per $100,000 in home value (see here). DCSD has published figures stating tax burdens will increase $200 for a $1M home (see here). How do these statements reconcile with the $399.95 figure above?
There is very little information available to the public to understand how these figures were determined. Reading between the lines, it appears that the logic used to figure the rate increase of $20 per $100,000 of home value is valid but there is some mistake somewhere and the resulting figure is not quite accurate. More concerning, however, is that it appears the logic used to figure the burden increase of $200 for a $1M home is not valid and the resulting figure might be significantly inaccurate; there's a big difference between owing $200 and owing $399.95 so at least one of these is wildly inaccurate. NOTE: Even though the $399.95 figure is presented with greater precision (i.e. more decimal places) than the $200, that alone does not imply that it's more accurate.
Despite the lack of public information, it's important to try to understand these figures because, at the end of the day, this is what will be coming out of taxpayers' pockets. In order to understand the rate increase, there are first three important things to note:
Now the increase to the effective tax rate resulting from 5A (and 35% property value increases) can be figured as follows:
It's also worth noting the change in effective tax rate per $100,000 of actual property value if 5A fails: (6.704 - 9.050) ÷ 1000 x 6.765% x 100,000 = -$16. It bears repeating that this doesn't imply that your tax burden will decrease, it's just the effective tax rate which is decreasing to counterbalance the effective of the taxable property values increasing. All else being equal, your tax burden for MLOs will neither increase nor decrease if 5A fails.
This $24 figure matches what DCSD has published on slide 24 here where it pertains to 5A alone, independent of 5B. We can try to apply the same logic as above to understand the impact of 5B:
On slide 25 of that same presentation, DCSD publishes a different figure, -$4. Again, since DCSD's methodology is not completely transparent, it's not clear why the logic above works to explain DCSD's MLO figures on slide 24 but the same logic fails to be consistent with the bond figures on slide 25. This is where it appears there was some mistake somewhere and DCSD's figure of -$4 is not quite accurate. The combined impact of 5A and 5B on the effective tax rate is an increase of $24 - $6.40 = $17.60 per $100,000 of home value, somewhat different than DCSD's published figure of $24 - $4 = $20. It's not a trivial discrepancy, but not an enormous one either.
Turning to the tax burden impact, the discrepancy is enormous. DCSD states the tax burden increase for an owner of a $1M property would be $200. It seems this was determined by simply multiplying their effective tax rate increase of $20 per $100,000 by the home value of $1M, which is indeed $200. However, this is not valid. In other words:
(new tax rate - old tax rate) x (new property value) ≠ tax burden change
The correct way to compute this is:
(new tax rate x new property value) - (old tax rate x old property value) = tax burden change
Since the discrepancy between $399.95 and $200 is so large, it's worth sanity-checking the $399.95 figure. This was calculated using the districts published figures about how the current mills levied for the existing MLO, and the expected changes to the mills depending on whether 5A passes or fails. But there's an alternative "outside-in" approach to calculating the tax burden that doesn't look at mills. Consider the following assumptions:
Then a property valued at $1M after the assessment increases would be valued at $1M ÷ 1.35 = $740,740.74 prior to the assessment increases, and the tax burden increase for this property would be is:
| Amount being raised: | $66,000,000.00 | |
| x | Your property value prior to assessment increases: | $740,740.74 |
| x | Assessment rate: | 6.765% |
| ÷ | Total assessed value prior to assessment increases: | $8,144,600,544.00 |
| = | Annual property tax increase: | $406.08 |
In words, this calculation determines the share of Douglas County's total property value that a $1M home represents, and assumes that's the share of the $66M that the property owner will owe, to determine the owner's tax burden increase for 5A. This figure of $406.08 is extremely close to the figure of $399.95 computed above.
I believe that the increased tax burden for the MLO increase alone (5A) for a $1M (after assessment increases) is closer to $400 rather than DCSD's published figure of $200. I have no reason to believe this discrepancy is due to any deceitful intentions on behalf of the school district. DCSD staff have been very willing to engage with me to get to the bottom of this. It could be an honest mistake; going from believing the rate will change $20 per $100,000 of home value to concluding that the tax burden will change by $200 for a $1M home is an easy mistake to make. I think it is reckless for voters to dismiss 5A on the speculation that there's some deceit in how DCSD is presenting 5A. I think it is possible to independently analyze the nature of the teacher pay problem in the district, and have done so in the analysis that follows. I think it's easy to independently verify whether money is being spent on teacher pay using the same Colorado Department of Education financial transparency data that is used to analyze the state of the present problem, and trust that this money won't be misspent even if one doesn't completely trust everything DCSD tells the public. And it's incumbent on voters to do their own cost-benefit analysis of 5A whether you believe DCSD's cost figures, the figures presented in this analysis, or something else, and not just dismiss the whole issue because you don't trust DCSD's figures.
The $66.0M is expected to be used as follows:
| Compensation | Security | Total | |
|---|---|---|---|
| District (Non-Charter) Allocation (76%) | $45.6M | $4.6M | $50.2M |
| Charter Allocation (24%) | $14.4M | $1.4M | $15.8M |
| Total | $60.0M | $6.0M | $66.0M |
Non-charter licensed teacher salaries will be increased by 9.2% on average under this proposal. In the 2021-2022 school year, DCSD spent $597,832,880 on salaries and benefits; $60.0M represents about 10% of that figure. Is the additional amount this MLO proposal seeks to raise for compensation equal to the additional amount it plans to spend on compensation? It's hard to verify precisely, but it definitely seems at least roughly true.
Douglas County residents should want a community where top teachers, police, firefighers, and other key contributors to the community can actually afford to live and thrive in the community. To compete for top teachers, their pay needs to be competitive relative to the other large, neighboring school districts with similar costs of living. The district's ability to retain teachers is at risk due to insufficiently competitive pay.
Generally speaking, average salaries for licensed classroom teachers in DCSD have always been low relative to peer districts. However, salary growth has been noticeably stagnant from 2008 to present compared to growth in peer districts, with Denver County and Jefferson County surpassing DCSD over that time period. Looking at DCSD alone, this wage stagnation seems untenable giving that the cost of living has not stagnated over that time period. This problem is even more acute when viewed through a competitive lens comparing DCSD pay to peer districts.
Lower average pay is not necessarily an indication of uncompetitive pay. Hypothetically, it could be that DCSD offers equal pay for equal experience, and simply employs a greater proportion of less experienced teachers. Assuming the overall quality of education is comparable to peer districts, this would be great news. It would imply DCSD is able to do more with less. However, Colorado school districts publish pay schedules, so it's possible to compare salaries across districts at the same experience level. For instance, for licensed classroom teachers with a BA degree, no additional credits towards an MA degree, and no relevant work experience, DCSD pays significantly less than peer districts:
You can find published salary schedules for each district below:
It stands to reason that uncompetitive pay would lead to attrition, meaning DCSD would struggle to retain teachers, especially top teachers. If a teacher departs, where might they go?
Retirement (#1) is natural, and is likely driven more by the age of a teacher than their pay. Some amount of teachers leaving the profession (#2) altogether is also natural. Demand for teachers is generally proportional to the number of school-aged children, and their numbers are decreasing as part of a broader demographic trend affecting all highly advanced countries. The US is no exception. And at a more local level, Douglas, Jefferson, and Boulder counties have all had a decrease in the child population from 2010 to 2019. Those three counties plus Denver and Arapahoe counties are projected to see a decrease in the child population from 2019 to 2030.
One plausible dynamic is that uncompetitive pay leads to an increased turnover rate amongst teachers in DCSD, especially amongst top teachers. These departing teachers would most likely go to competing nearby districts (#3) rather than leaving state (#4), since that's an extremely disruptive life change, or switching to private schools (#5), since public schools are much larger employers than private schools. Those districts can't just hire increasingly more teachers and keep pay high, especially given demographic shifts in the child population, without also losing some teachers. Those districts would have lower, relatively more stable turnover rates (slowly shrinking to slowly growing), perhaps especially amongst lower-performing teachers.
It's hard to find data on turnover rates by performance level, but there is data on teacher turnover rates in general:
DCSD is at near-all-time highs for its annual teacher turnover rate, the rate at which the turnover rate is growing, and for how high it sits up amongst the range of its peer districts' annual teacher turnover rates. This data seems very consistent with the scenario of high teacher turnover in DCSD among top performers, balanced with moderate teacher turnover in peer districts.
Is DCSD currently spending existing funds elsewhere that would be better off redirected towards salaries?
One way to look at this is to consider overall expenditures per pupil, and per-pupil expenditures within certain categories and subcategories of spending, to see whether DCSD is a high-spending outlier relative to peer districts, and relative to all Colorado school districts. By and large, the data suggest that, no, DCSD is not excessively spending existing funds elsewhere.
First, total expenditures per pupil:
This data goes back to the 2006-2007 school year. DCSD actually spent more per pupil than the other five highlighted peer districts that year. The latest data from 2020-2021 has DCSD spending less than the other five. These figures are in current dollars, meaning nominal dollar amounts not adjusted for inflation. DCSD per-pupil spending has essentially remained flat despite inflation, while all other districts' spending have grown. So, overall, DCSD is very frugal relative to its peers.
Looking beyond just peer districts, the thin gray lines in the chart above represent most other school districts in Colorado†. DCSD per-pupil spending is low-to-average relative to most districts. This reinforces that DCSD is relatively very frugal.
†Graphs exclude all BOCES and the following districts which had inconsistent data: School District 27J, Westminster School District, Custer County School District C-1, Elizabeth School District, District 49, Fremont RE-2, Pueblo County 70, Upper Rio Grande C-7, Revere School District, Weld RE-3J, and Weld RE-8.
Despite DCSD spending less in total, it's possible that the specific ways money is being spent is simply sub-optimal, and there could be money being spent in some categories which should be redirected to instructor compensation. It's worth drilling down into several of the largest subcategories of spending. First, before looking for categories where spending might be too high, it makes sense to look at instructional salaries and benefits and confirm this spending is consistent with what the previous section showed, namely that teacher salaries are insufficiently competitive:
Since 2008-2009, DCSD has pretty much always spent the least per-pupil on instructor salaries amongst its peers (occasionally being nearly tied with Jefferson County for spending the least). It's worth noting that, aside from the nebulous "Other" expenditure category, "Instructional Salary" is the largest category of spending.
When it comes to benefits, DCSD is more middle-of-the-road, both amongst peers and amongst all Colorado districts as a whole. Denver appears to spend much less on benefits, but when figuring total compensation (salary + benefits), Denver spent $4,959.91 + $927.58 = $5,888.49 per-pupil on instructor compensation in 2020-2021 whereas Douglas County only spent $4,340.98 + $1,451.16 = $5,792.14. The only peer district relative to which Douglas County spends more is Jefferson County, and the margin is small.
Next, administrative spending. Bloated administrative spending is a perennial concern in education (and in many industries across the US).
DCSD is middle-of-the-road here. Not very frugal, but by no means excessive.
Note, this chart looks at School Administration expenses, which is the total of school-specific administrative expenses. There is another category of administrative expenses, referred to as General Administration, which accounts for central or district-wide administrative expenses. General Administration is relatively small category of spending, however, especially for DCSD and its very large peer districts that benefit from economies of scale when it comes to central administrative spending.
While "administrative bloat" is sometimes hotly debated as there are often questions or allegations of "wasteful bureaucracy", there is another category of spending that is just as large though less often discussed. Here is expenditure data for Operations & Maintenance:
Here, it's evident that DCSD has consistently spent less than its peers, and is in fact one of the most spendthrift districts in all of Colorado.
Lastly, the aforementioned nebulous "Other" categories. Unfortunately, due to lack of standardized accounting between districts, there can be huge variations from district to district and year to year in terms of what gets lumped into "Other". And these "Other" categories are in fact very large spending categories, so while being extremely difficult to scrutinize, they can't simply be ignored. To complicate matters, there are different kinds of "Other". At a high level, spending is broken down into Instructional, Support (for Instructional Activities), Community Services, and Other. Within the Support category, there is an "Other Support" subcategory. Both the top-level "Other" category and the "Other Support" subcategory are sizable, so we'll look at both:
Since 2014-2015, DCSD has been a low spender on "Other" relative to peers districts. "Other Support" is generally not as big as "Other" as a spending category, but in this category DCSD is a relatively high spender, and has been so consistently. When you add both of these "Others" together, only Cherry Creek spends less per pupil than DCSD amongst peer districts.
When analyzing this data, there are 3 important caveats to call out:
With the caveats in mind, I believe DCSD should consider spending less on School Administration if they haven't considered this already. For example, decreasing spending 20% here from $961.39 per pupil to $769.12 would put DCSD's spending on par with Littleton's spending of $797.42. And this could be used to contribute significantly towards closing the gap in Instrucitonal Salary expenditures: DCSD's spending could increase from $4,340.98 per pupil to $4,533.26 compared to Littleton's $4,984.57, reducing the gap 30% from $643.59 to $451.31. However, it would definitely not completely close the gap, and DCSD's overall spending patterns are otherwise relatively very frugal, so significant additional revenues will still need to be raised to achieve competitive teacher pay.
Douglas County resident property taxes are about to go up significantly with the increases in assessed values. Will that already pay for this? If not, does the additional revenue really need to come from Douglas County resident property taxes?
School funding in Colorado is extremely complicated. The "Total Program" defined in the Public School Finance Act (1994) represents the majority of school districts' funding. The Total Program amount that each district gets is determined by the state, and reflects what the state believes the district needs based on:
The state applies a formula to determine the Total Program for each district, and then if the sum of all districts' Total Program excessively exceeds the previous year's total, all districts' Total Program amounts are reduced by a common factor to effect Budget Stabilization. The post-stabilization Total Program is what each district receives, and is determiend by the state, but the funding for this comes from both state and local taxes.
Funding for this starts with a fixed-percentage carve-out from local property taxes and vehicle registration fees. Any shortfall towards a district's Total Program from these local funds is paid for by the state. When property values increase, the local share of Total Program funding increases, and the state share decreases, but the Total Program amount stays the same.
Districts which need additional funding beyond the state's Total Program calculation can raise additional funding through local property taxes. There are legal limitations to the purposes for which funds can be raised, and the amount by which property tax revenue can be increased for each of those purposes. One of the most important mechanisms is the mill levy override; school districts can seek to raise an additional 25% of the Total Program value through local property tax increases, and can use this for operating expenses including salaries (but not capital expenses such as building construction).
There are other potential sources for revenue that could be used for compensation increases, namely:
But these sources represent a small fraction of DCSD revenue compared to property taxes. For example, in 2020-2021 these represented somewhere between $41M and $126M, whereas property taxes represented $324M. Furthermore, residents have the power to directly vote on the MLO increase, there's much less ability to directly influence how much more funding can come from those other sources and ensure those funds can be used for compensation increases.
The charts in this presentation were created with and served from this spreadsheet. All data used in that spreadsheet were obtained from the Colorado Department of Education (CDE) website. The data on average teacher salaries was manually copied from CDE into the spreadsheet; the data on turnover rates and expenditures was downloaded in XLS format, processed and output into CSV format, and then uploaded into the spreadsheet. The un-processed XLS files downloaded from CDE and the source code for processing the CSV files are shared publicly:
The processing code was bootstrapped using ChatGPT, see chat here.
The source code for this presentation can be found here.
My wife and I moved from San Francisco to Highlands Ranch in 2021. We moved here for the beautiful nature, low crime, friendly people, great schools, political diversity, and housing affordability relative to San Francisco.
My parents immigrated from India to Canada. I was born in Toronto in 1984, and grew up in the suburbs of Toronto. I moved to the US in 2007 to pursue a PhD in mathematics at UC Berkeley. In 2012 I left academia to pursue a career in tech, working for 4 years as a software engineer, 2 years as a product manager, and 5 years as a director of product management, leading teams of product managers. My professional area of focus has been in developing and bringing to market large scale, cloud-based application and data management platforms targeting the market of IT departments of large enterprise companies. I have specialized in working for mid-stage startups growing rapidly through to IPO and beyond.
I do not work for, nor represent in any way, the Douglas County School District, its elected officials, or its staff. I do not have kids enrolled in DCSD, though I plan to in the future. I am not an expert in DCSD operations, Colorado school financing laws, or the laws and regulations related to taxation in the state of Colorado or in Douglas County.
I think schools are important, teachers are important, financial responsibility is important, and an engaged citizenry who understands the details and holds governments accountable is important.
I have written about the recent changes to DCSD's equity policy: Douglas County School District Makes Its Equity Policy More Pro-Human. I have also interviewed some of the school board candidates running in the November 2023 election: Interviews with DCSD BoE 2023 Candidates.