Since the market decline hit the front page of every major news paper in the land, the war between Case-Shiller, the National Association or Realtors, and the FHFA has become a blood bath. The battle ground? Who produces a better model for tracking the decline in home prices and whose numbers can you rely upon? The winner, well there isn't one and their won't be! Here at Atlanta Marketing Data, we pride ourselves on looking at all three objectively and letting you, the consumer, have all the facts and making the best decision for you!

All the major players in the game have admitted that their indexes have major flaws but they all claim to be tops in the numbers. How can this be true? We have spent several years in the field, looking at the numbers in good times and bad.  The one thing that they have gotten right, the fact that they are all wrong. We will take a look at the major flaws in each index and then propose a better way of thinking.

Case-Shiller
We start our case against Case-Shiller. There are two major flaws that just can not be ignored, plus many minor flaws that skew the data. First major flaw is that the reporting within the index lags 4 months in report but up to 18 months since the market change. How can you work off of numbers that are 18 months old? The second flaw is the vague areas in which Case-Shiller tries to produce numbers. While they have made an attempt to correct this major flaw, Case-Shiller still only reflects a small portion of the market activity out there. First, their major indexes only view 10 or 20 of the major marketing areas, or MSA. For a national index, this is too small of a sample size to project any meaningful numbers. But it must be good in the areas that it does cover, right?

In some cases these "local" MSA areas can be very different from each other and scattered internally. For example, Shiller labels his areas after cities but they are really reporting on MSAs (Metropolitan Statistical Areas). Let's take the Manhattan/New York MSA according to Case-Shiller. This MSA covers Manhattan, Putnam County - two hours north, Bergen County - in New Jersey, and Long Island. Even if you are not familiar with New York City and the region, you can imagine that this is too broad of an area for statistical analysis of the area. Would the person who can afford to live in the city buy a mobile home two hours away? How many single family homes are there next to Time Square? So we are not dealing with cities or even areas with the same motivations and likely buyers.

Then there is the old reported flaw that Case-Shiller reports did not break down the MSAs into price points so they were skewed to the low end. This is a valid case and one that the index attempted to fix. They now offer a low, middle and high price point in most of their MSAs. However, let's take the example of San Fancisco's MSA. The low is set at anything below $512,000 the mid is anything between $512,000 and $750,000 and the high is anything over $750,000. I am not an expert in the San Francisco marketing area, but it has been widely reported that these ranges are too broad. What about the homes that are $250,000? Granted that might be the very low of the market but they are compared to a home twice their value? Would a buyer looking at a $250,000 home in the market also be influenced by what is happening in a home twice its value?

Now let's take a look at what Case-Shiller is looking at and what they are not. To make it on the Case-Shiller list to be analyzed, you have to be a house that has sold at least once prior. This means that they do not calculate in new construction sales. They have to have data on your prior sales, so it must have sold within the prior 10 years. The home must not have deed transfers within 6 months of each other at any time. It must be a single family home, no condos or multi-family. And the list goes on. Most people who have the time to look at the numbers agree that Case-Shiller is only looking at about 0.5% of the market sales at any given point because everything else has been excluded. So they expand their MSAs so they have enough data to even get to what they can call a sample size while watering down the market in such a way that it makes it a mute point. Could this be why Case-Shiller indexes are all over the board in their analysis?

Getting back to what they do use, Case-Shiller does not measure the change in house prices from last year or last quarter or last month or the peak of the market. Case-Shiller measures a home from the last time that group of homes sold. So, by definition, it is more of a "sold house index" then a housing index. Granted, they TRY to weigh a home based on how long ago it sold vs. now and plug that into an algorithm that is weighted to TRY to adjust for any changes that the house might have undergone or changes in consumer tastes that might have occurred, but how accurate can it be. Without getting into too much boring detail, a home that sold 7 months ago would be given a weight factor of 1 and a home that sold 10 years ago would be given a weight factor of 0.8. What do these numbers mean? Where are they gauged from and what do they translate to in the market? No one, not even Case-Shiller can explain this in terms that anyone can understand. I applaud them for trying to adjust for such difference but if you can not explain what or where that adjustment is from, it is not a good adjustment.

Most importantly, the Case-Shiller index does not tell you what has gone on between the sales of a home. They try to draw a straight line between two data points and imply that the market must have dropped XX% in that time. If a home sold for $100,000 in 2001 and then $150,000 in 2009, Case-Shiller is going to report that it had an increase in value of 50%. But what does that mean? How much at any one point did that house have that gain? Did it have a total gain of 75% at its peak but now had a 25% loss in the decline? They do not know and can not analyze that data. The way the index works any gain is a positive and any decline is a negative. So in areas where people hold on to their homes for 5 years, the rate of decline would have eaten up any increase those properties had before the index would show such a market in decline. This is important to realize when reading the data. In our example above, the home would be reported as an overall increase while it might have been declining for several months or even years. The net effect of the data is a positive trend and that is all that Case-Shiller would report on. As long as the gains out number the declines, the index will be upwardly biased, and vice versa.

The way Case-Shiller wants markets to fit into a box is not how markets work.  By forcing your data into a box, you ALWAYS are skewing the results! Within each market, there is a sub-market which we already know they ignore. Within those sub-markets there are factors such as seasonal variances that they do not calculate. Pull up the most recent Case-Shiller data out there. Go on, I will wait. What do you see for the 2Q2009? Most marketing areas rose in value vs. 1Q2009, right? The news outlets jumped on this is the "downfall" of the market had reached its bottom and home prices are on the rise again. But what is really going on here? Could it be that the second and third quarters in the housing market are normally higher than the first and the fourth? I do not know about New York but I can tell you that a 10 year study of the general Atlanta Marketing area would show that more homes are sold in the second and third quarters of a year than any other time. So are we seeing real recovery or are we just seeing what happens every year within the market. If we do see this increase every year, the important question is how did the increase from 1Q2009-2Q2009 compare to the increase from 1Q2008-2Q2008? What about the cycle before that? Can conclusions be drawn from that data? Case-Shiller is unable to tell us that.

I could stay on the case of Case-Shiller until your eyes gloss over. But I promised that we would be fair and balanced and spend time on the other two major players in the game. So we will close with one last point on Case-Shiller. Let me say that I am not accusing Case-Shiller of anything sinister but as a fraud investigator, I always look at the motivations of what someone says and how they act. Case-Shiller knows what type of data their report produce and what reactions they get from the media and by the consumers. They know the flaws in their report but insist that there is nothing that needs to be fixed. It is important for the people that rely on Case-Shiller indexes to know that Case-Shiller is not only an index firm but also could make profit off of the doom and gloom that they report. When any market free fall, many investors, traders, banks, etc, want to hedge themselves against the price declines by buying insurance against that free fall. Case-Shiller reports are used for the Chicago Mercantile Exchange (CME). The CME uses his reports for hedging housing futures values. The more hedging futures and insurance they sell, the more money gets deposited with Dr. Shiller's bank account. So, there could be a relationship and financial incentive to scare the market. I would hate to think that anyone with ethics would do such a thing, but we have seen worse in the past. Rely on this data at your own risk!

FHFA Home Price Index
The second major player in this three way war is the FHFA, formally the OFHEO, Housing Price Index. It is important to note right off the start that the FHFA is a government agency. So if Case-Shiller can be looked at more closely because of its financial relationship to the CME, the FHFA must be looked at closer in insure that no political bias exists in their number reporting. While Case-Shiller has tried to make the claim that the FHFA is producing numbers that show the markets better than they are, little evidence has been produced to show such a claim. But it is always good to be mindful of where your data is produced and what motivations might be behind such data.

The FHFA produces data more closely to the Nation Association of Realtors than it does Case-Shiller. We will get into the NAR data later but there are some unique factors that the FHFA numbers get into that are worth explaining on their own.

The FHFA utilizes numbers reported by Fannie Mae and Freddie Mac and only data from those sources. We can already see how this could be a problem. What about the areas that are heavily funded by FHA loans from HUD? In some Atlanta markets, that can be more than 50% of the sales inventory that is being ignored. Further more, they only track mortgages less than $417,000. So you have restricted your analysis to the middle ground, excluding the higher end homes and the lower value FHA homes. As does Case-Shiller, the FHFA numbers do not include new home sales or condo sales. So now you only report the middle of the market that does not include any new home sales or condo sales. Do you see a pattern here? What data ARE they using?

So we have a small data pool of houses that may not be representing the entire market but yet they claim it to be reliable data. Let's take a look at the data they do use and maybe we can find a bright spot to hang our data hat on. They use FNMA transactions. What type of FNMA transactions? Well I am glad you asked. The answer is any type, new mortgages, refinances, cash-out transactions, whatever FNMA will lend on they use as data. Why is this a problem? Well, for one, a refinance price is NOT what the house will or did sell for, not even close. Talk to any industry professional and they will tell you that what you might be able to get a refinance appraisal to state is NOT what the house will sell for. It should be, but it rarely ever is. So you are using potentially inflated refinance appraisal values to calculate your housing index? Let's make the problem even worse. In many high fraud areas, and Atlanta has some of the worse areas for mortgage fraud, what type of data are you going to get if you use refinance appraisals as true market value? I can point you to the zip code of 30310 if you would like to see an area that was/is driven by mortgage fraud and where no less than 80% of the transactions done in that area were fraud related. By opening their data up to these transactions, with no checks and balances to evaluate the fraud data, the FHFA has skewed their already limited data pool.

The critics of the FHFA Housing Price Index also point to the fact that the index uses median sale prices to calculate their index. They state that this creates skewed results when one part of your data pool is heavy on the low or high end. While there is merit in this argument there is also an easy way to fix this problem; Instead of evaluation one whole city, state, MSA, or region at a time, concentrate your data on the sub-markets that make up the total market. To do this, the FHFA index would have to be produced with local knowledge of each region that they report on and report the findings for the sub-markets. Unfortunately, the man power to do this for the entire country would be huge and the amount of funding that would be needed would be bigger than the GDP of a small country. The problem is that the FHFA is a government agency and doing anything that is logical or takes time is out of the book of most government agencies.

The other major problem with the FHFA index is the lag in reporting. FHFA already has about a 7-9 month delay in the reporting of data figures. For example, the offers made in February - May, closed in April - June, were released in September. So offers made in February were reported in the second quarter report that was not available until September. That is quite a lag!

National Association of Realtors
The last major player in the war on housing prices is the National Association of Realtors (NAR) Housing Index. In the sense of a traditional way, you can think of the NAR and the FHFA of tentative allies against the perceived greater foe, Case-Shiller. It is true, the methodology that the NAR and the FHFA uses are very close to the same. Both calculate median house prices. Because of this, the NAR has the same issues with skewed data results that the FHFA does. Neither one has broken down their numbers into sub-markets in an attempt to produce reports that apply to sub-markets. Take the example of San Francisco market and assume that Case-Shiller's low, mid high, ranges are correct. You would show a 32% decline in the low range sub-market, a 21% decline in the mid ranged sub-market and a 6% decline in the high ranged sub-market. If there were 5 times more homes that sold in the low range than in the mid range, your median data totals are going to be skewed to the lower range and thus produce a rate of decline for the ENTIRE marketing area that is skewed to a more drastic decline. Now, this is one example, in one specific market, at one specific point in time. Does it create the same problem everywhere? The simple answer is no. You have to evaluate your data pool versus the marketing area and the effects it is going to have in your analysis. Problem is that both the NAR and the FHFA reports claim to be the de facto source with the "correct" numbers. That is just not the case.

The Solution to the Problem
It is nice to spend some time picking on Case-Shiller, the FHFA, and the NAR, but all this means little if we can not try to improve the manner in which reporting and analysis is done. All the major indexes have problems and they always will.

There is just no good way to take broad data about a region that has many sub-markets and paint it with a broad brush. It is like trying to tell you that the average nationwide temperature will be 59 degrees today. It does not tell you much about the temperature where you live. The solution is firms, like Atlanta Marketing Data, that are local to the marketing area and understand the sub-markets. Firms that understand marketing trends, data, statistics, buyer's motivations, socio-economic influences, etc and can produce data that reflects these factors in the market. You need to know:
     1. What has sold and for what price?
     2. Why did it sell for that price?
     3. Who lives there and who is moving there?
     4. What's the current inventory?
     5. What are the relevant trends?

So to do this we need to go beyond admitting that home price data is flawed, because it all is inherently flawed. We need local home sales and price data. We need to be able to slice and dice the market by existing vs. new, price point, home features, home type, home size, and home age.

At Atlanta Marketing Data, we can do just that. We can fix the list of flaws below while extracting the good merit points of the three major indexes to produce a complete picture of the market. Contact us today about how we can produce customized reports that paint a complete and detailed Atlanta market picture for you.