David Rathgeber's

Charts & Statistics

Introduction

This is an in-depth analyses of real estate statistics, how they are developed, what they mean, and how they should, and should not, be used. Too often we see and hear real estate information that sounds important, but is patently useless. It is hoped that buyers and sellers of resale homes can use this information to see through the fog and to confirm their ability to discern what constitutes actionable information. You will not see this anywhere else!

Also see What are they talking about?

This page contains the following sections:

Comparing data

First consider the source. It is safer to assume that data from one reporting source cannot be compared to that from another. Next, be sure of what area is being reported. There are significant regional variations even within our own metropolitan area. In short: Compare apples to apples.

Our real estate market is seasonal. As data is generally reported in monthly increments, 2 comparisons make sense:

The first comparison tends to remove seasonality and provides a long-term view. The second comparison tends to include seasonality and provides a short-term view. However, they are both comparisons of 2 bits of data. If the individual monthly data points (e.g. averages) contain a lot of noise or bounce, the resulting comparison is subject to the same inaccuracy.

Always check to see whether the data is recent, or several months old. Some data reported in the media is 6 months old and is therefore useless. From what perspective is the data analyzed? For longer-term data; how were the start and end dates selected? For example, we get very different views if we calculate home price appreciation to date starting in 1990 versus starting in 1998!

Why was the data analyzed? The methodology will influence its usefulness. Was it prepared to help:

Massaging the data is a bit more complex than it seems on the surface. How it is done can matter and big mistakes are made. If you are looking for actionable information, you need to be sure.

Try to determine what is included in the data. Is the data for resale homes? Is it diluted by inclusion of:

Is it limited by inclusion of only:

Does the data include short-sales, and what is the effect? Is it seasonally adjusted? Is it inflation adjusted? All of the above are complicating factors in the assembling of data that we hear frequently. But no one takes the time to explain them to us, and they can make a significant difference, especially in the approximation of average home prices where the number of data points controls the accuracy of the result. As a buyer or seller of a resale home, this is very important to you and, sadly, these are distinctions often lost on the folks collecting, massaging, and reporting the data.

And finally, we are a regional market; there is no national real estate market. While national averages are important for macro-economists they are merely useless bits of information for individual homeowners.

Also see the comments on Statistical Significance.

[Click here to return to top]

Demand

Index Demand alone is of almost no importance to home buyers and home sellers. Without supply, it is only one side of the story. It generally need not be reported as it would only serve to dilute any main message.

However, note that peak sales occur in April, not the summer "when the kids are out of school" as you might have been told. The graph above is based on contracts entered, not sales closed. Buyers, sellers, and agents need to know about contract activity. They could care less when, or even if, the sales go to settlement. Further, those interested in demand should realize that it not only fluctuates seasonally, but also varies with macro-economic trends on a 15 to 20 year cycle, in recent decades.

Note to Realtors: Demand is of critical importance to you. There can be significant, periodic, and sometimes random fluctuations in one's own business volume. So be sure to bury some acorns in the boom times to provide for the times when the seasonal, macro-economic, and personal cycles hit a concurrent low. Remember that the alternative is getting a real job. Ugh!
[Click here to return to top]

Supply

Index Supply alone is of almost no importance to home buyers and home sellers. Without demand, it is only one side of the story. It generally need not be reported as it would only serve to dilute any main message, but do notice how little supply varies from May to October. This is discussed later under Market Seasonality.

However, the percentage of foreclosures, short-sales, and vacant homes compared to all homes on the market can sometimes be of importance. They should be reported when they have the potential to affect buyers, sellers, or average home prices. Lower percentages of distressed properties are better for sellers. Higher percentages are better for buyers, because distressed properties imply market softness and can apply downward pressure on prices.
[Click here to return to top]

Market Strength

Index The single most important real estate statistic is the Months Supply of homes on the market shown above. A number greater than 5.0 indicates that buyers have the upper hand. A number lower than 3.0 favors sellers, and less than 1.5 indicates a hot market.

To obtain this figure, the number of homes on the market is divided by the number of contracts entered (sometimes called pending sales) for the previous month. The focus is on resale homes: Rental properties and brand new homes are excluded from the calculation.

Home sellers can use the information as a measure of:

Home buyers can use the information as a measure of:

Months Supply is the most important statistic because it incorporates both supply, demand, and the effects of home prices, interest rates, and local economic conditions into a single real-time indicator. A monthly reading can be immediately available with essentially no delay or distortion by using the number of contracts entered, that is pending sales, not closed sales. Nearly all pending sales proceed to closing. More importantly, pending sales represent actual contracts; whether they close or not is irrelevant.

One way to view the Months Supply of Homes: With a 4 Months Supply, if buyers kept buying at the current rate and no new homes entered the market, after 4 months there would be no homes left on the market. Of course this never happens, but merely serves to illustrate the concept.

There is a distinct seasonality (see the section below) to the real estate market, which is reflected in the Months Supply. This means that certain times of the year are somewhat more advantageous for home sellers (March?) and other times are more advantageous for home buyers (December?). You cannot beat timeliness coupled with accuracy!

Note: The inverse of Months Supply is sometimes reported as the Absorption Rate. For example, a 4 Months Supply is the same as a 25% Absorption Rate.
[Click here to return to top]

Market Seasonality

(The Market Index - Months Supply of Homes)

graph
The graph above incorporates supply and demand into a single number and shows the seasonality of the real estate market. This is the classic case of how the market should behave. It shows sellers having their best market in March, not June and July as you often hear. It shows buyers doing best in December. Note the September "slump."

The graph below shows the seasonality of our real estate market for several recent years. Sellers still have their best market in March, not June and July as you often hear. But buyers now do best in September and then months supply decreases for the rest of the year.
graph
What has changed? Here are some clues:

  1. The seasonal pattern and shape of the number of homes sold (demand) chart has not changed with peak sales occurring in April most years.
  2. The number of homes on the market (supply) is about two-thirds of the previous historical value.
  3. Both supply and demand are decreasing in the final months of the year as always, but the supply decrease has become the controlling factor.
  4. The seasonal pattern and shape of the supply chart has flattened. There is now very little change in supply from May to October.
  5. Average number of days on the market is much lower than before, so a home spends a month or two less on the market. This tends to flatten the increasing supply trend earlier in the year, and steepen the decreasing trend later in the year.

The supply chart flattening effect due to homes selling faster (#5 above) seems to be the best explanation.
[Click here to return to top]

Short-Sales: A disturbing influence

Short-sales have had a disturbing influence on market data in recent years. While foreclosures and vacant homes generally are sold and closed in relatively normal time frames, short-sales can take much longer. Further, a significant proportion of them never get to closing. Therefore, many buyers and agents will not even view them. So including them in homes on the market increases Months Supply figures un-naturally. In other words, they look like inventory but really are not. Excluding them altogether from the calculation would be equally misleading because they are of interest to a limited number of buyers.

On the other hand, many short-sale properties result in more than one contract, as buyers get discouraged and bail out, making way for second, third, and more successive contracts on the same property. While this tends to offset the effect of inflated inventory described above, there is no telling to what extent this actually happens.

Fortunately, lenders (with government urging) have been improving their ability to deal with short sales. Further, the number of short-sales has declined significantly since 2008. At this point, the disturbing influence on our data is present, but not great.
[Click here to return to top]

Market Prices

graph
The graph above displays the Average Home Price trend for our area from 1991 to date. Note the unusually slow market of the early 1990's when home appreciation was nil for 6 or 7 years. Then the market heated up dramatically until the average price peaked at $568,074 in June 2006 then started to decline, hitting a low of $359,660 in December of 2008.

Our ears always perk up when we hear home price data, but the information is generally recreational rather than actionable. The average home price can rise or fall in two ways:

When we try to distinguish between these two effects by looking at the medians instead of average prices, we find that the medians and averages generally rise and fall together. So this seems to be of no help. Nevertheless, if a trend is indicated, and it has resulted from general appreciation or depreciation, it can be valuable in the following ways:

For home sellers:

For home buyers:

Note that a price calculation can be done for homes currently on the market or for only those homes that have recently sold. The on the market calculation yields average prices much greater than the average for recently sold homes. But the numbers generally quoted in the media are for sold homes, so buyers and sellers who are contemplating action need to remember that this is old news. How old is it? It is 2 to 4 months old: Two unavoidable months old due to data collection limitations (the lag between contract and closing) and another month or two for data assembly, review, comment, and publication. It is critical to understand this, because taking action on 2 to 4 month old data can be a disaster. To compound the problem, the media generally reports the data as "last month's numbers" since an explanation such as that above is much too time consuming.

The concept of statistical significance is germane to this discussion as many of the average prices reports are worse than meaningless; they are misleading. Be sure to see the segment entitled Statistical Significance.
[Click here to return to top]

Statistical Significance

Statistical Significance is a mathematical concept, too infrequently applied to real estate data. It states that in order for a conclusion to be meaningful, it must be based on a sufficient number of observations or individual data points. A complete treatment of the subject is far beyond the scope of this author, but the concept is important: Depending on the set of data being analyzed, there is a certain minimum number of individual data points required in order for the conclusion to be reliable, valid, or useful. Inaccuracy due to an insufficient number of data points is sometimes called "statistical bounce." Such inaccuracy exists in most published reports of average home prices, often to an alarming extent.

Of course, there is a reliable statistic called average home price which changes over time in some orderly fashion (i.e. without excessive bounce). The intent of an average price calculation is to measure the change in price for all homes in a specific area by calculating an average of those homes that have sold recently. In other words we are trying to measure an entire population with data from a small portion of that population.

An empirical analysis was performed on a 3 month sample of Northern Virginia sold homes. There were nearly 10,000 data points (sold homes) in the sample; enough to provide a reliable average. When random samples roughly representing one month (3,300 solds) were analyzed, the average price was found to be within 4% of the correct value about half the time. When random samples of 20 solds were analyzed, the average price was found to be within 10.6% of the correct value about half the time.

To summarize, an average of 10,000 solds results in a reasonable confidence level (i.e. accuracy) for an average price calculation in order to provide valuable, actionable information. As few as 3,300 solds (roughly one month of Northern Virginia solds) can be useful if viewed with skepticism: When the average Northern Virginia home price usually changes less than 1% in any month, a monthly number that can be 4% off is worthy of skepticism! So when you see published data that bounces around from one report to the next, you will understand why.

For example, an average price can be calculated for any zip code you wish, but to get to 10,000 solds would take years and years of data, so any accurate reading would be based on data which is, on the average, many years old; again, meaningless. This is why average home prices in any zip code for any month, or even a year, are patently meaningless: The data sample size is not large enough over a short enough time. Nevertheless, we often hear average price calculations which are based on less than 20 data points! Many media folks have no concept of statistical significance and will report anything that sounds like news, useful or not. The fact that someone found some numbers and calculated an average might be good enough for them, but not for savvy folks like us.
[Click here to return to top]

Days on the Market

graph
Days on the Market is an important statistic for both buyers and sellers but the interpretation is often tricky: The calculation can be done for all homes on the market, or for only those homes that have sold recently. To further compound the problem, some MLS systems compute Days on the Market using only the current listing of a home, some account for previous un-sold listing entries of the same home, some systems provide both figures, and some provide no information at all. Our regional MLS system, MRIS, provides both figures. Stellar!

To add still another level of problem-compounding-complexity: The very useful Days on the Market at the current price is often difficult or impossible to determine. And finally, at the risk of TMI, no one ever tells you which number they are reporting!

Using Days on the Market for home sellers:

When the market speaks, sellers need to listen. But remember that any sold data lags by 1 or 2 months, plus any time needed for data assembly and publication. So use Days on the Market as a very rough guide rather than a precise indicator.

Using Days on the Market for home buyers:

Delays generally favor a seller, so buyers should never delay a decision.

But it is suggested that even when suitable data are available, Days on the Market should be used with caution because the calculations reported and the resulting expectations can vary widely from actual results for any individual case. Yes, in this one instance, your home sale (or purchase) is unique.
graph
The graph above illustrates the general relationship between price and time on the market. The shape of the curve, the intercept, et cetera will change with market conditions and with the magnitude and frequency of any required price adjustments. "Overpriced" is defined as the relationship between a home's initial asking price and it's actual market value. Note well: When a home enters the market, no one knows the real market value. (No, not even Zillow!)

Whenever considering Days on the Market, be sure to check whether average prices are rising or falling annually as well as seasonally, and be sure to check the all-important Months Supply.
[Click here to return to top]

Showing Traffic

The very best market indicator for home sellers is the number of showings your home receives every week. Reasonable expectations vary greatly by the health of the market, seasonality, price range, and proximity to Washington D.C. which is the center of our region's market. But if no one came to see your home in the last 2 weeks you very likely are in trouble: The answer is most likely on the sheet that comes out of the printer when an agent searches for properties to show. Often, but not always, it is your home's asking price. This is not rocket science, just attention to details, which includes knowing which details need attention. There is no magic involved in selling a home.
[Click here to return to top]

Selling-Price to Asking-Price Ratio

graph
The selling-price to asking-price ratio, expressed as a percentage, is one of the most important statistics available. For a buyer, it provides an indication of how much can be negotiated off the average asking price so that expectations are kept within a reasonable range, and so that time is not wasted viewing properties that are priced out of reach. For a seller, the ratio tells how much "fat" is required in the asking-price. Sellers and buyers alike are surprised to find the ratio is so high, almost always greater than 95%. In fact, in a hot market, the ratio averages over 100%!

To set the expectations of buyers making an offer when there are no other buyers, it will be helpful to calculate an average of only the transactions for which the ratio is less than 100%. A similar calculation of only the transactions for which the ratio is greater than 100% is useful information for buyers in multiple-offer situations.

To calculate the selling-price to asking-price ratio, select recently sold homes from the MLS database. Divide the total of all the selling prices (reduced by any seller credit) by the total of all the asking prices to find the selling-price to asking-price ratio. It is important to include at least 200 individual sales to ensure an accurate result. The selling-price to asking-price ratio represents critical information about our market; do not rely on a guess or the 95% figure noted above. Obtain an exact, up to date reading, tailored to your needs.

One might also use the selling-price to asking-price ratio as a measure of an agent's negotiating performance: Compare an individual's record with the local average. Wow!
[Click here to return to top]

Tax Assessments and other schemes

A tax assessment is the value a local government assigns to a home for real estate taxation purposes. It has only a very general relationship to that home's market value: It is so general as to be useless. But some will tell you there is a direct relationship. They have not done an analysis, are relying on a very few bits of data, or are repeating "what everybody knows." A careful analysis of the statistics proves them wrong, both over a wide area as well as in an area as limited as a development or subdivision.

One who attempts to predict a market value or contract price from a tax assessment needs to guess a percentage factor to multiply by the tax assessment to obtain the predicted contract price. That would mean correctly guessing a number (the percentage factor) with absolutely no basis for making the choice. Of course, there is some average relationship between tax assessments and market values that can be calculated. But use of this figure to determine the market value of a specific property should be enough to make even the tax assessor giggle.

Amazingly, elaborate calculation schemes or algorithms (pronounced: Al Gore rhythms) have been devised to predict a market value from tax assessments. Although the mathematics are impeccable and some such schemes have even gotten favorable press nationally, the fact remains that when you do any calculation using basically flawed data, the result is inevitably flawed. Garbage in; garbage out. End of story.

Be alert to anyone who tries to draw a conclusion from just a few cases: This is logically as well as mathematically unsound. Also reflect on how long it has been since the tax assessor even visited your home. Never?

Zillow, and other such schemes that rely heavily on tax assessments, are just as inaccurate. Track down Zillow's own disclaimer: For our area, they report they are within 10% of actual market value 69% of the time. That means they are more than 10% off almost one-third of the time.

There is no substitute for a properly prepared market value analysis or an appraisal. Mortgage lenders do not rely on tax assessments, and neither should you. In short: The market price of a home is defined by the price paid by a buyer to a seller, neither being under duress, etcetera, etcetera. Anything else is an estimate, and some of them are way off!
[Click here to return to top]

Interest Rates

Interest rates are an important factor in a real estate transaction. Familiar to all, they do not need much explaining: Lower is better for everyone. It is striking to calculate that one can buy 50% more home at a 4% interest rate compared to 6%, at the very same monthly payment!

Whether you are a buyer or a seller, keep an eye on changes in interest rates. They can have an important effect on the real estate market, but sudden changes sometimes can have the opposite effect from that reasonably expected: For example, a sudden increase can sometimes trigger buying activity albeit temporary.

Also, real estate markets do not move in lock-step with mortgage rates: There can be a significant time lag between interest rate changes and changes in market activity. Finally, the great differences in home price appreciation and depreciation in different markets across the country are indeed notable, despite the fact that mortgage rates are nearly uniform.
[Click here to return to top]

Housing Affordability Index

Consider the interrelationship between interest rates, home prices, and personal income, or purchasing power. When one of these factors experiences a significant change, one or more of the other factors will also be affected. Mortgage interest rates can make significant moves in a short time. Personal income depends upon employment as well as wage and salary levels and therefore it moves more slowly. But personal income and interest rates change independently of each other and are influenced by macro-economic trends. Home prices are caught in the middle as the dependent variable.

For example, if mortgage interest rates move down this means improved affordability for real estate. Even if personal income remains stable, buyers can buy much more for the same monthly payment that they were undertaking previously. This results in upward pressure on home prices. But as we know, home prices do not move (inversely) in lock-step with mortgage rates. Home prices can be held in check by economic worries, a large resale home inventory, or economic malaise. On the other hand, they can increase quickly in hot markets without any change in mortgage interest rates.

So the housing affordability index is not fixed, but varies greatly with time and location. Examination of affordability in various markets across the US presents quite a diverse picture. What do you think the housing affordability index is in Honolulu? Hong Kong? Tokyo? London? Get the idea? The Index is an interesting but useless piece of data for home buyers and sellers. Further, it is way beyond our control.

So what is important? You have all the information you need to determine your own personal home affordability. An Index helps fill a news spot, but it is meaningless for you.
[Click here to return to top]

New Home Data

New homes account for roughly 10% of home sales, but the actual volume can expand or contract radically due to changing economic conditions. For you economist buffs, this is called elasticity of supply. Reliable data for this market segment is available from the National Association of Home Builders (NAHB). For the individual home seller, zero new home construction would be ideal. For buyers, the more, the merrier.

While most home buyers consider new homes as well as resale homes, individual home sellers need to be especially vigilant when considering their marketing strategy if comparable new homes are available nearby. National and regional data reported by the media include new home sales, housing starts, permits issued, etc. But new home construction is neither controllable by, nor significant for most buyers and sellers. It's just a "blip" on the curve.
[Click here to return to top]

Average vs. Median

Both average and median values can be calculated for almost any set of data, including, home prices, days on the market, and selling-price to asking-price ratio. Since both average and median values tend to rise and fall in unison, it seems needless to report both. It is suggested that only the average values need be reported, because the concept of an average needs no explanation.

It should be noted that median values for home prices and days on the market are generally lower than the averages for those data sets. For you statistics buffs, this is due to the skewness of the data. Median values for selling-price to asking-price ratio are usually higher than the averages, but sometimes the opposite is true. Such a reversal is most likely in an especially hot spring market, no doubt due to a reversal of the skewness.

There is another statistical measure called the mode. While it might be useful in some cases, it requires too much explanation for the casual reader. We therefore suggest that it not be considered, unless preceded by "pie ala-."
[Click here to return to top]

Methodology

Data is collected from the Realtors' MLS system in the morning on the 1st day of every month. The most important statistic is the Months Supply of homes on the market. To obtain this, the number of homes on the market is divided by the number of contracts that were entered in the previous month (sometimes called "pending sales"). The Report's focus is resale homes: Rental properties and new homes are excluded from the calculations as both are very different markets. Separate data are collected for detached homes, townhomes, and condominiums and for various geographical areas.

Over the years, some very interesting facts have been uncovered. The most important of these is that the best month for home sellers is March, not June to August as the others keep telling us. What they are saying sounds right, but the data prove otherwise.

Home buyers and sellers need the very best and most timely information anywhere. But many market reports are two to six months out-of-date because they are based on closings, not contracts entered and are generally not published in a timely manner. Some reports may be distorted by rental, and new home information. Further, some sources calculate Months Supply including pending sales along with homes on the market in the numerator of the fraction. This distorts the calculation since most agents and buyers consider that pending sales are not fully available and do not view them.

One government report on average home prices is heavily based on re-finance appraisers' opinions which clearly are not pertinent, and another widely quoted measure of prices employs an algorithm so arcane as to make its data useless! Finally, it bears repeating (again and again) that although data can be collected and averages can be calculated, there is no national real estate market. To know what is happening in our area, we must use data only from our area.
[Click here to return to top]

*   *   *   *   *   *   *

YFiRE
Your Friend in Real Estate, LLC
Arlington, Virginia, USA

EqualHousing

* Copyright © David Rathgeber *
* * All rights reserved. * *
* * * 2018 * * *