| Author | Message | ||
     Octofoil |
 | ||
| Octofoil |
The post above contains a chart with an index of median resale prices of homes in Maplewood fromm 1988 through 2000. The source for the data is the Certified Valuations spreadsheet dated 1/19/01 and distributed by Jerry Ryan. The data is suggestive of a couple of things: 1) The appreciation in overall resale prices in Maplewood appears to have been largely a one-year phenomenon-so far. That is, while some modest gains occured in the '97 - '99 period, the majority of the appreciation occured in 2000. 2) The data begs the question as to whether or not such outsized one-year gains as seen in 2000 will be sustained. Given that such gains are rarely repeatable,likely scenarios include a) a modest retrenchment of values, then holding steady at the higher levels; or b) advancing or declining at rates more consistent with and reflective of underlying economic conditions (which includes the possiblility of a sharp fall off in values as the slowing economy begins to impact the area's economy in a more significant way). 3) Some of the "we've been overpaying for all these years" language from some taxpayers appears to be overstating the case. There is no denying the inequity produced by the size of relative gains in 2000, but to continue to describe the situation as though it had been extant since the mid- to late-1980's doesn't square with this data. I hope to fill in the data for 1980 - 1988 to better see the complete data set from last reval to this one. Any errors in the extraction and preparation of this data are mine alone. Any errors in the source data belong to those who prepared the Excel workbook. | ||
| Townie |
Octofoil, Thanks for doing this. Three questions: 1) This shows median sales for ALL of Maplewood, right? Is it possible to see just the Jefferson alone? 2) If all areas were rising and falling like this, everybody's taxes would have been "fair," in the least inflammatory sense of that word, right? (Presuming the 1981 reval didn't unfairly overassess and underassess some.) 3) Does your work thus far tell you anything about when between '91 and today would have been the best time to do the reval? Thanks, kathleen | ||
| Octofoil |
1) Right. The index is for sales for all of Maplewood. An ongoing project is to break out by neighborhoods and regions (east of Valley, west of Valley, for example). And, as I mentioned, to fill in the "front end data" from 1980 onward. 2) Remember, as noted above, this is aggregate data so it doesnt reflect compositional (neighborhood) changes. It doesn't show which areas increased the most or the least. Hence, the need to break out by neighborhoods. 3) No, unfortunately, it doesn't, in a specific sense, show when would have been the best time to do the reval. The reason being that it doesn't show the tools that the assessor's use (e.g., the so-called "coefficient of variation"), to gauge when market prices are out-of-line with assessed values. However, as a general observation, there were several years of 5% or greater increases. After a couple of years of 5% plus increases, my assumption is that the coefficient of variation would have dropped below the threshold level. But my assumption could be wrong. A lot depends on the specifics of how the assessors measure things. For example, the assessors seem to employ arithmetic averages for many measurements and ratios, which is totally inappropriate due to the skewed nature of the data, IMHO. The data as reflected in the chart has limited usefullness, IMHO, but in the interest of adding some perspective to the timing of the increases in values, there it is. | ||
| Melidere |
you've presented absolutely no data that bears on the question of 'overpaying' and underpaying'. I think you know fully well that the revaluation has nothing to do with the absolute value of homes. It is about the 'relative' value of homes, and this chart doesn't speak to that. I can come up with an infinite set of sale prices that will produce that chart, all of which tell a completely different story. | ||
| Dytunck |
Octo, Thanks for the effort. This was a valiant attempt, but I'm sorry to say there are too many flaws presented. 1) The term is "Coefficient of Deviation" 2) The Maplewood Home Prices index numbers are way way off. I can only assume 134.3 is $134,300. That's nowhere close. 3) There were several mistakes in the spreadsheet made public by Jerry. Did you account for these? 4) The spreadsheet data drops off as you go back in time, because it shows only the most recent sale. If a house sold in 1991, and again in 1997, only the 1997 price will be given. 5) There is no valid data for 1989-1991. 6) Your data won't stop Nakaille from being sarcastic. | ||
| Nakaille |
Dytunck: I don't think that was Octo's intention, do you? We all have different ways of expressing anger. That's just one of mine. Better than shooting up schools, don't you think? Bacata | ||
| Gerardryan |
Octo: which column in the sheet did you graph? Or what formula did you graph? Index of what? thanks PS Dytunck's right, the sheet only has most-recent sale | ||
| Octofoil |
Melidere: 1) No, you can't come up with an "infinite" set of data that will produce this chart. This chart is the linked price relatives of the median sales prices of the set of sales data extracted from the CV data base as furnished by Jerry Ryan. 2) You are correct, the chart shows nothing about the relative value of homes. As I said in my post, this is the aggregate data townwide; neighborhood/region comparisons will be forthcoming (when myself or someone else finds the time or initiative to do it). Dytunck: 1) Yep, the term is "coefficient of deviation". Apologies for the typo. 2) These are index numbers, not prices. 134.3 does not represent 134,300. These are linked price relatives that indicate percentage changes, not absolute price levels. 3) No, I did not account for any errors in the spreadsheet because I don't know what they are/were. If you or anyone would advise me of them, I will correct the data. 4) If you or anyone else can point or supply me with more complete data, I'll use it. Until then, this data does represent a statistically valid sample of sales (except for any errors that might be present in the source data). 5) The "cert-011901-3.xls" spreadsheet contains data for 1989-1991. Why is not valid? Is there a note or post or any documentation to that effect anywhere? Point me to it and I'll make whatever adjustments appear appropriate. 6) Stop Nakaille/Bacata from being sarcastic? Not in a million years! Nor would I ever try! While I readily admit that I find some aspects of Bacata's posts annoying, even grating sometimes, I also admit that they have a redeeming essential correctness about them. I may not like her style, which is sometimes bullying, arrogant and threatening, in my humble opinion, but most of the time, the underlying point is quite valid. Look, I prepared this data and posted for one reason: to add some dimension to the questions concerning the timing of the reval. Looking at the data, one can get a hint (not definitive, but a suggestion) as to why the assessor was encouraged to skip a reval for several years when, in hindsight, it probably shoud not have been skipped. I've been very critical of the TC in their handling of this whole thing, feeling among other things, that the elapsed time between revals was inexcusable and such contributed hugely to the magnitude of the tax increases some/I now face. Though I still disagree with not doing it sooner, I now get a better sense of why they didn't do a reval sooner. The data suggests that perhaps the sales-data-based case for doing a reval in the intervening years was borderline and wasn't as solid as I had thought. Legal considerations/requirements are another story. In closing, it appears that some don't understand index numbers. Perhaps I assumed too much. In this instance, I've used the simplest possible approach to constructing the index due to the paucity of the data. I said it earlier and I'll repeat it now: it isn't definitive, it is only suggestive. The data isn't exhaustive or complete enough to do otherwise. | ||
| Lseltzer |
I still don't get it. Index of what? It's not like "Home Price Index" is a standardized term. Please give us your formula. | ||
| Octofoil |
Jerry Ryan and Lseltzer: Technically, the term is "link relatives". If p1, p2, p3,..., represent prices during successive intervals of time 1, 2, 3,..., then p1|2, p2|3, p3|4,..., represent the price relatives of each time interval relative to the preceding time interval. I first extracted the sales prices from the "cert-011901-3.xls" xl workbook, then determined the median price for each year. The index of link relatives was constructed from the median prices. According to Dytunck, "cert-011901-3.xls" contains some errors. Can you point them out to me? | ||
| Gerardryan |
Can you email me the resultant sheet that you made so that I can understand the formula? Send it to revalinfo07040@aol.com. Time to dig out the old stats textbook; I don't remember "link relatives". Can you explain it? What does it demonstrate? In concept I mean. I don't know what errors Dyntunck might be referring to. It is the case that only the most recent sale is included. It's also true that some sales that are not arms-length transactions are included in that mix as well. | ||
| Melidere |
Octo, a median simply means that half the sales were above and half were below. the median could rise by precisely the same amount if all the houses were rising by the same percentage OR if some of the houses were rising rapidly and others were falling rapidly. i've looked at this data and i've studied it. i think you have too, and i think you know that you cannot make statement number 3 from this data. you left out a really important piece of the data, to boot, which is the original assessments. I think that if you added them you would see that the movement from 1981 to 1991 is quite significant. there was a roaring bull market in this town in 86 and 87. | ||
| Aruba18 |
Octofoil- The other major reason for not doing a reval earlier is strictly political--it is not a popular thing to do! | ||
| Dytunck |
Octo, Sorry if I misunderstood your chart. I think LSeltzer is right when he says further explanation of your methodology is needed. (Now you must feel like Certified Valuations!) Be honest, without explanation, who would know what 134.3 in 2000 would mean? 34% above the prior year? I'm still confused. As for the data errors, I found a couple glitches when I was performing a similar study to yours. Maybe your file is not corrupted. I found suspect data in the following addresses: 1609-19 SPRINGFIELD AVE. 494 VALLEY STREET 767 VALLEY STREET 769 VALLEY STREET 2203-2211 MILLBURN AVE R 2185-2209 MILLBURN AVENUE 2211 MILLBURN AVE 2213 MILLBURN AVE These are commercial properties. I don't know if you filtered out the class 1s, 4s, 6s and 15s. Your chart does say "Home" price index, so you probably did, but still, there was some erroneous data affecting the 2000 numbers for me. Also, the version I downloaded had no 1989 sales data whatsoever. And one sale from Jan 90, one from Feb, one from March. Surely not enough statistical data. There was a major drop-off in data the further back in time you went. The data gets more misleading the further back in time you go also, because as I said, only the last sale is listed. The best data is the 2000 data. 1999 could have properties that sold again in 2000, so those won't be listed. Likewise 1998. So, the further back you go, the more compromised the data, IMHO. I am interested to know more. Thanks for your analysis. I check back for your response. Dytunck | ||
| Octofoil |
Melidere, Well, at least you got the definition of "median" right. As to statement number 3, why not? The statement clearly refers, and necessarily so, only to the data presented, nothing more, nothing less. I would like to be able to answer definitively the extrapolation that you're attempting to make, but don't have the data available as of this moment. If I can locate a data source for 1980 forward, I'll add it and recalc and then we'll see what conclusions can be drawn about the entire 1981 - 2000 period. Any ideas as to a source? I was here in '86 and '87 (been here since early 1980) and recall a period of rising prices (though I wouldn't have referred to it as a roaring bull market, but thats just our differing impressions). My impression also is of some softness and retrenchment in the years following, particularly during the economic slowdown that followed. | ||
| Melidere |
Octo, Overpaying and underpaying have to do with the relative movements of the values of property in different areas of town. How could we come to any sort of conculusion about that using townwide data that obscures those relative movements? just because the townwide median was relatively stable doesn't mean that the median in hilton wasn't taking a nosedive while the median on wyoming was rising. or perhaps it would be ok with you if your doctor ordered a drug for you because the studies showed that the median for all the people in the test study was right in the middle, half the people showed great improvement and half showed serious side effects. Would it change your mind to know that the half showing serious side effects were male and the half showing great improvement were female? one of the better companies that specialize in residential real estate price indices uses 4 'tiers' for measuring movements in price, recognizing that those tiers move very differently. They use Tier 1: from 0 to 75% of the median Tier 2: from 75% to 100% of the median Tier 3: from 100% to 125% of the median Tier 4: from 125% of the median and up. Home Price Indices They developed these tiers over many years of studying the movements in the prices of homes including something like 95% of the homes in the united states over a significant period of time. They use them because the data is close to meaningless without them. | ||
| Octofoil |
Dytunck, As to indexes, in this day and time, with the equity markets the center of attention, I rather thought that folks would be more comfortable with indexes, i.e., the S&P, the NASDAQ. Granted, those indexes are constructed differently (some market cap weighted, some not), but the essentials are quite similar; the underlying principles the same. Yep, I filtered out all but residential properties. In the version that I downloaded, there are a few records that contain data from prior to 88 - 89, but I chose to use remain with only those years in which there were at least 100 obervations, thereby giving some confidence of a decent sample size. As a check, the sample data for 1998-2000 agreed with that provided in a separate file by Jerry Ryan ("sales-011901.xls") As to misleading the further back in time...not really. A transaction is still a transaction. And as long as the sample size is sufficient, you can still draw inferences. For example, if one is constructing an index of equity trades, one might use closing prices. But what about all those other trades during the day? We are drawing samples from a population, not recording every trade in the population. While it is always desirable to work with more transactions rather than less, i.e., including subsequent sales of same properties-the larger the sample size the more reliable the inference, the sample size in all years shown is sufficient for indicative purposes. As I said, the data can only be considered as suggestive, not definitive. But even suggestive is, to me, informative. Everyone else will have to decide for themselves. Moreover, this is the same data set (excluding distressed sales, etc.) upon which the CV/assessor made their initial set of valuation decisions. Thanks very much for taking the time to look at the data and respond. | ||
| Octofoil |
Melidere, See my post at 5:01 pm on Mar 11, above. The breaking out by neighborhood in order to identify the relative changes is next on the list. | ||
| Octofoil |
Melidere, Mortgagerisk.com is but one of the data mining firms that uses tiered indexes for specialized purposes, such as targeted mailing lists for home equity loans, etc. Got any fliers in your mailbox lately? This is one of the companies that provides mailing lists. Notice the tiers are keyed to percent of median value? Thats one way in which they target home equity loan fliers. They also provide indexes with all observations. | ||
| Melidere |
i've used their database for many years. The home price tiers are used because the tiers behave very differently in up/down markets. If you take a look at the next offer for a home equity line that you receive (because yes, they do use this database for that) you might take a look at how incredibly accurate that amount is. That's because a combination of zip code level data plus tiering provides an very accurate assessment of the equity in your home. Without the tiering the data is nearly useless. the basis for most of their observations is using what they call 'pairs' which are two datapoints for each house in the database. That gives them accurate data about the relative movements in prices without the complicating factor of adding in apples and oranges. They pull out outliers, like homes with significant improvements/additions. I don't mean to be overly critical, and perhaps, if you look at the data using tiers you will see why i'm so disappointed with that chart. I understand this is a volunteer effort and am not unappreciative. If you teir the database using the 81 assessement and then follow it through using the sales data from 88 on, you will paint a very different picture, i'm certain, based only what i've seen and observed. | ||
| Octofoil |
Melidere, If you've used their database for years, how about getting a look at their index for 07040? Theirs is complete for the period that we would like to examine and is based on a much larger sample. Should prove very instructive. | ||
| Melidere |
i wish i could sighs the cost of their data is way over my head. | ||
| Octofoil |
Melidere, I hear that. Any other thoughts on sources of data? | ||
| Octofoil |
Gerardryan et al, Sorry, forgot to respond to your request about sending spreadsheet...will do so this evening (its on the box at home, dont have it on the work box...even though the work box is a laptop and I'm never without it, I minimize the personal stuff on it). | ||
| Nakaille |
Octofoil: thanks for your comments about the substance of my writing versus the style of it (at least on these reval threads.) If you read other threads you probably know that I am generally much more civilized (ie. less aggravated/sarcastic and more inclined to be helpful) than here. That is my general demeanor in life and it actually makes me laugh to read myself (or my writing) described as arrogant, bullying, etc. If my old shrink could see me now she'd probably applaud (and then want to talk about it, of course!) Growing up in a violent family made any overt expression of anger dangerous. Sarcasm (for me) is simply a defense, although I do understand that some people find particularly offensive. I find some others' assumptions about money and privilege particularly offensive, too. I do appreciate the fact that there are people like yourself willing to look beyond the emotional tone, though. Again, thanks. Bacata |
