Tuesday, August 7, 2012

Fraser Valley Numbers

HERE

SFH Average is DOWN 2.3% YOY. Median is up a tiny 1.2% and yet benchmark is UP a huge 3.4%! 

That combination must lay at the outer reaches of statistical likelihood.

Apartments are DOWN for BOTH median and average and yet UP for benchmark YOY.

It really does not make sense to me. I think I am going to ignore the Real estate board generated HPIs and Benchmarks and concentrate on medians and averages from now on to gauge the market.

Okanagan Numbers

More sales than last year but median and average prices took a beating. However the sample is small and can't read too much into it. MOI running around 15. Actually higher volume with lower prices would be better for everyone except speculators.




20 comments:

  1. Averages and medians are susceptible to changes in product mix while the Benchmark is a statistical construct based on individual home characteristics. They do not have to agree month to month, though they should over time. Its not complicated.

    If you read the literature on measuring home prices, you'll find that hedonic price indices such as the MLS HPI or paired sales indices like Teranet/Scotiabank are much preferred to averages and medians.

    ReplyDelete
  2. oops - should be Teranet/Nationalbank.

    ReplyDelete
  3. Preferred by whom? The resl estate sell-side industry?

    ReplyDelete
  4. No. Preferred by economists, academics, etc. See this Kansas City Fed paper: http://www.kansascityfed.org/Publicat/econrev/PDF/2q07rapp.pdf

    ReplyDelete
    Replies
    1. Do you really think the re boards up here have armies of economists and staff and computer modeling To run proper comparisons?

      Especially in markets like the okanagan. Did you see what the difference in average between last July and this July was?
      $1millon less!
      I agree with fish the hpi diesntt make sense

      Delete
    2. ahh....., must be that HPI-X factor!

      Delete
  5. No I don't. But the MLS HPI was outsourced by CREA to Altus Group, who does have the resources. The Okanagan is not covered by either the MLS HPI or Teranet. For some markets, the average is fine, but in markets with major short-term distortions from changing product mix, it is going to be giving the wrong signal. This is all uncontroversial. It brings to mind a great Paul Krugman quote: "if you just can’t believe I’m saying the things I say, at least consider the possibility that you’re the one who just doesn’t get it."

    ReplyDelete
  6. Hi Anon

    Everyone who knows the tiniest amount of stats knows that if you want a true read of a group of numbers there are many ways to try and determine that. Average is the most blunt because a few outliers will skew things. However median is a very reasonable number, though of course in small samples it becomes a less accurate depiction of the underlying numbers.

    you can use standard deviation data to further refine your numbers like they do with research.

    An HPI is a subjective analysis, not an objective one as you have been alluding to. The averages and median are completely objective, accurate or not.

    An HPI is trying to compare apples with apples. Except who decides what aspects of the apple you should compare? The size , granite counters, garage, you get my drift? Of course not only is the modelling important but also the data entry. To make it reliable requires a lot of financial and human expenditure.

    furthermore in small samples the HPI or benchmark becomes meaningless , because there are not enough individual sales to compare, as you have conceded for the OK.

    For the OK apartment HPI to have gone up even as the average and median have gone down, means most of the apartments sold were 'below' the benchmark. Maybe the benchmark is set too high when most of the product sold is below it!

    ReplyDelete
    Replies
    1. So we agree that for large markets, like Vancouver, an HPI (whether Teranet's paired sales, or the MLS hedonic price index) is a better measure?

      Delete
    2. If the HPI is so good why did they have to change the way it was measured. And then change it again when they didn't like the way the numbers look. I agree that a price index is a very useful metric but ONLY when it is not controlled by people with a vested interest in making it go up (and the lack of morals to manipulate it thus).

      Delete
    3. They didn't change the way it was measured. There are two numbers, the HPI, which is an index, and the benchmark price which translates the index changes to dollar prices. The problem was that, as an artifact of how each was constructed, the growth rate in the HPI and the Benchmark price were very slightly different. Growth rates for Benchmark prices are now linked to their corresponding HPI Index. No change was made in the HPI Index. This is all publicly available info. If you choose to believe differently, then you are simply ignoring facts.

      Delete
    4. Even if you don't like the MLS HPI, what about the Teranet/National Bank Index? Is that number also being manipulated or perhaps is it more likely that well-known problems with averages are causing the average price to send the wrong signal (both on the way up and on the way down).

      Delete
    5. The pre-2012 HPI has been decreasing for five months before the REBGV introduced the new HPI which (oh what a surprise!) showed a steep increase. The new HPI was then recalculated in May turning what would otherwise have been a decrease into an increase. If you choose to believe Realtors are fine fellows only interested in statistical rigor that's up to you but it is simply ignoring the facts.

      Delete
  7. "So we agree that for large markets, like Vancouver, an HPI (whether Teranet's paired sales, or the MLS hedonic price index) is a better measure?"

    I wold place more faith in the Teranet numbers which come out with a two month delay than the REBGV HPI which comes out so soon after the end of the month.

    To bring numbers out so soon, you need to be inputting data as it happens, accurately, with good quality controls etc.

    Maybe they are, but the longer that we have low sales, high inventory, high MOI, lower average and medians and YET a stable or even rising HPI, the more credulity is stretched.

    The stock index is collated by S and P which does not sell stocks, would it be as reliable if Goldman Sachs was collating it?

    In the US we have the Shiller HPI, we have no similar disinterested party providing stats here

    ReplyDelete
    Replies
    1. Bravo! Sir, you're my hero. :)

      Delete
    2. So based on your faith in Teranet, you accept that prices in Vancouver are not actually falling yet? (or at least, were not falling as of June).

      I'd implore you to plot each price measure in a single graph (indexed since MLS HPI and Teranet HPI are index numbers). You'll see how obvious the distortion in the average price (and to a lesser extent the median) was starting in early 2010. If after see that chart you still think the average is providing a more accurate signal than the MLS or Teranet HPI, then there is no hope for you.

      Delete
    3. If you think house prices are still going up when there is a year and half of inventory in the OK in the Ok and will be between 9-10 months in Vancouver, and when the number of daily prices changes outnumber the sales by a large margin- then I would suggest that you have too much hope.

      Delete
    4. I didn't say prices were going up in the Okanagan. What I said was, according to Teranet's June numbers, prices were still rising in Vancouver (but falling slightly according to MLS HPI in July). I didn't even say that Vancouver prices would keep rising. I merely said that the average wasn't telling an accurate story over the past 18 months due to well known problems with averages.

      At current levels of supply, prices will very likely fall modestly in Vancouver in coming months. That is hardly controversial.

      What is with you bear bloggers anyway? Anything that doesn't taste like your favourite brand of kool-aid gets immediately spat out with all sorts of accusations of data manipulation and real-estate pumping. Its like trying to talk to a 9/11 truther or a gold-bug.

      Delete
  8. as prices were rising, averages were just fine???
    Dave#1

    ReplyDelete
  9. Very good post. I will be experiencing some of these issues as well.
    .

    Review my website: US real estate agents directory

    ReplyDelete