Last month, when new home sales fell 33%, I warned investors that “This data series is notoriously noisy, and you are must better off using a 3 month moving average than reading too much into any single month.”

This month, we see that New Home sales jumped 24%. The same caveat applies: Ignore the swings, look at the moving average to smooth out the volatility.

In fact, the series is so noisy that way back in 2005, we had to to a study to figure out what to make of it. Our conclusion?

“Looking back over the past 15 years of data, we see that a mean regression has followed nearly all double digit monthly gains. The subsequent month’s data was significantly lowered — flat to negative in nearly every case.”

I have yet to do study on the negative double digit months, but I have a suspicion it would be the same. Bottom line: Anytime you see a big gain or drop in this data series, you should expect a reversal the following month.


click for larger chart

Courtesy of Calculated Risk


New Home Sales Data: Don’t rely On It Either (November 30th, 2005)

New Home Sales Plunge 33% (June 23rd, 2010)

U.S. Department of Commerce, Manufacturing and Construction Division APRIL 23, 2010 AT 10:00 A.M. EDT
Erica Filipek or Stephen Cooper

Category: Data Analysis, Real Estate

Please use the comments to demonstrate your own ignorance, unfamiliarity with empirical data and lack of respect for scientific knowledge. Be sure to create straw men and argue against things I have neither said nor implied. If you could repeat previously discredited memes or steer the conversation into irrelevant, off topic discussions, it would be appreciated. Lastly, kindly forgo all civility in your discourse . . . you are, after all, anonymous.

10 Responses to “New Homes Sales: Mostly Noise, Little Signal”

  1. wally says:

    Much like the local baseball team.

  2. X on the MTA says:

    30 thousand new homes were sold (NSA). This is a new record low for June.

    May was revised down by 33k to 267k

    – Peter @ MacroNotes

    So I guess one these is annualized? Anyway, if the revisions are over 10% of the figure, that’s reason along to not pay any attention to numbers and instead focus on 4 month trends (i use 4 months, so that you get a full three months of revised data)

  3. I think Calculated Risk figured all this out. Bottom line: the report was awful but because it it increased over last month Wall Street celebrated. Just look at the melt-up today.

  4. realgm says:

    The 24% “gain” was based on a revised down figure of the previous months which was a 33% loss (revised down to 39%.

    The number is suspicious. Did the government intentionally cook the figure?

    I wonder if this 24% “gain” will be revised down next month.

  5. Mike in Nola says:

    I don’t see where the May or June 2010 are that out of line with the corresponding months of 2009 when there was no panic to get in under the tax credit deadline. I suggest it is approximately the new normal til the existing home inventory is worked off because new homes are competing with existing homes.

  6. realgm says:

    oops. It was a revised down to 37% loss for May.

  7. Mike in Nola says:

    realgm: If they were going to manufacture a number, it probably wouldn’t be the all time lowest June sales figure. It was pretty pitiful despite the CNBC pumping (assumed, since I don’t watch them any more unless I know someone honest is going to be on.)

  8. louis says:

    The VIX on Shadow inventory points ^

  9. [...] digesting the new home sales report, keep in mind the data point is notoriously noisy, Ritholtz says. June’s 24% monthly increase comes on the heels of May’s 33% drop. “Ignore the [...]

  10. philipat says:

    On a moving average basis, the aberations appear to be March and April, rather than May and June. But I bet the NAR will make hay with this month’s data? Can’t wait to read their spin, always good for a laugh!