[Winstonm]

Our corporate media keeps reporting as news stories supplied to them from vested interests - and we have to keep asking ourselves "Is it news or an infomercial?"

We are forced to turn to blogs to avoid influence peddling.

The latest example is the NAR (National Association of Realtors) approved message of "improvement in housing starts" as reported by the mainstream financial media.

But turning to commentary by Barry Ritholtz of "The Big Picture" blog, we get a completely different description:

Quote

    BUILDING PERMITS: Privately-owned housing units authorized by building permits in June were at a seasonally adjusted annual rate of 563,000. This is 8.7% (±3.0%) above (revised) May rate, but is 52.0% (±3.6%) below the June 2008.

    HOUSING STARTS: Privately-owned housing starts in June were at a seasonally adjusted annual rate of 582,000. This is 3.6% (±11.3%)* above the revised May estimate but is 46.0% (±4.3%) below the June 2008.

Quote

What can we tell from this data?

Nothing about monthly change in Starts (data points less than the margin of error are not statistically significant); We can say that permits were up month to month, although how much of that is seasonal is hard to decipher.

The year-over-year data is much clearer: New Starts down 46%, Permits down 52%.

Quote

Incidentally, much of the media reportage on this was simply innumerate — the numerical equivalent of illiteracy. Not just a little wrong, but totally, embarrassingly incorrect.

    WSJ: “Housing starts increased 3.6% to a seasonally adjusted 582,000 annual rate compared to the prior month, the Commerce Department said Friday.”

    Bloomberg: Housing starts in the U.S. unexpectedly rose in June as construction of single-family dwellings jumped by the most since 2004, signaling the market is stabilizing. The 3.6 percent increase brought starts to an annual rate of 582,000.

    Marketwatch: Housing starts rose 3.6% to a seasonally adjusted annual rate of 582,000, the highest figure November.

    Reuters:  New housing starts and permits jumped more than expected in June, propelled by a rise in single-family homes, a government report showed on Friday. Housing starts climbed 3.6 percent to seasonally adjusted annual rate of 582,000 units, from May’s upwardly revised 562,000 units, the Commerce Department said.

Quote

No, that is not what they said at all –  plus 3.6% with a margin of error of 11.3% = YOU DON”T KNOW.

I know, this is a pet peeve of mine — but still, it makes you wonder if these people can count to 21 unless they are naked.

Exactly - first off because of seasonality the essential numbers for new housing starts are year-to-year comparisons not month-to-month, and when a set of numbers is within the margin of error, there is no way to know what really occurred - still, our media dutifully reports the NAR talking points as if it were news instead of advertising.

These are the telling statistics left unreported by the cheerleading from the media: The year-over-year data is much clearer: New Starts down 46%, Permits down 52%

To find someone with a comprehensive understanding, one has to go outside the regular media.

Is this individualized ignorance or organized audacity? Or are the rest of us who put up with this media crap simply too stupid to know or care?

Really, this should matter. It doesn't help to pretend things are better than they are, just as whistling in the dark doesn't stop the mugger in the alley from slitting your throat. The essential component in taking wise action is a precise understanding of reality - to carry a gun or avoid the alley altogether.

Seems our corporate media would have us whistle while they keep us in the dark, though.

[PassedOut]

In our businesses, we always compare a given fiscal month with the same fiscal month for the year before (and previous years). The same goes for quarterly and yearly comparisons. It seems to me that comparing this month with last month has meaning only when a business is immune to seasonal effects (whatever business that might be).

But it can be helpful to compare the data for a month against a rationally derived projection for that particular month, which might be based in part on trends identified in the data for the immediately previous months.

I agree with you that the web is a strong defense against misinformation in the media (which understandably always has a strong corporate bias). However, posters on the web often have strong biases also, and trustworthy sources that get the facts right are true gold.

[Winstonm]

You bring up a good point about web-bias. One certainly has to determine if the web author has an agenda to promote or is simply trying to add to rational understanding.

Of course, I am old enough to remember when news departments were based on the integrity of the reporting - a totally foreign world from today.

[onoway]

maybe when they count new housing they are including the tent cities?

[barmar]

I agree that normally it's more important to compare year-over-year results. But in this case, I can understand why they're not emphasizing that. We're coming back from a really bad decline, and no one really expects this year's numbers to compare well with previous years. Since they're reporting that we seem to be at the beginning of a recovery, it makes sense to compare recent results.

Consider if you were looking at a graph of this. If you've been experiencing a steady decline for a while, good news would be when the graph turns towards the horizontal and then starts to move up. But that trough would still be well below levels from the previous year, so you wouldn't be able to tell anything by comparing year-to-year.

As for the margin of error, is it similar to the margin in previous reports of the same statistic? In that case, I think it's not unreasonable to compare them. It would certainly be nice if we had more accurate data, but in absence of that you have to make do with the best estimate you have.

[Winstonm]

barmar, on Jul 19 2009, 10:56 AM, said:

I agree that normally it's more important to compare year-over-year results. But in this case, I can understand why they're not emphasizing that. We're coming back from a really bad decline, and no one really expects this year's numbers to compare well with previous years. Since they're reporting that we seem to be at the beginning of a recovery, it makes sense to compare recent results.

Consider if you were looking at a graph of this. If you've been experiencing a steady decline for a while, good news would be when the graph turns towards the horizontal and then starts to move up. But that trough would still be well below levels from the previous year, so you wouldn't be able to tell anything by comparing year-to-year.

As for the margin of error, is it similar to the margin in previous reports of the same statistic?  In that case, I think it's not unreasonable to compare them.  It would certainly be nice if we had more accurate data, but in absence of that you have to make do with the best estimate you have.

Nice try but swing and a miss. You seem to either be apologizing for the media or falling into the reasoning trap that the state of the recovery should be included in whether or not the statistics are reported accurately. That is irrelevant. It doesn't matter if we sold 3% more houses this month than last if the normal increase for this month over last month is 3%. It becomes even less important when 3% falls within the margin of error - then the statistic becomes irrelevant.

Housing starts could have been down -7% and still be within the margin of error. The statistics show that whatever happened last month was statistically meaningless. That did not stop the media from reporting the meaningless as the meaningful.

These annoying facts have not stopped our illustrious media, though. They have been making the same type of misleading and inaccurate comparisons now for months on end - this isn't the first time the media has made unfounded claims about what the housing statistics show.

Bloomberg's report is especially troubling as they used fake statistics to draw a positive conclusion:

Quote

Bloomberg: Housing starts in the U.S. unexpectedly rose in June as construction of single-family dwellings jumped by the most since 2004, signaling the market is stabilizing. The 3.6 percent increase brought starts to an annual rate of 582,000.

Sorry, but the job of the media is not to boost moral with false claims or to hide facts that are unpleasant. By the simple act of tuning them out, we should hold these news outlets to a higher standard than they seem willing to produce.

[barmar]

Winstonm, on Jul 19 2009, 01:34 PM, said:

Nice try but swing and a miss.  You seem to either be apologizing for the media or falling into the reasoning trap that the state of the recovery should be included in whether or not the statistics are reported accurately.  That is irrelevant.  It doesn't matter if we sold 3% more houses this month than last if the normal increase for this month over last month is 3%.  It becomes even less important when 3% falls within the margin of error - then the statistic becomes irrelevant..

If they've already made the seasonal adjustment in the numbers, doesn't that include the "normal 3% increase", which means the actual increase was 3% above normal?

But even if not, we're not in a normal year. Suppose normal years have sales increasing steadily from Feb to June, but this year they've been dropping from Feb to May. If they actually increase from May to June, that's surely a sign that a recent bad trend has ended.

Regarding the margin of error, isn't there a bell curve involved, so the value is much more likely to be +3% than -7%? And with or without the bell curve, it's just as likely to be +13% as it is -7%. If last month's increase was +1% +/-10%, and this month it's 3% +/-10%, that seems better, since the entire range has shifted up (this month's worst case -7% is better than last month's -9%). But if it was previously +1% +/-5%, that's not as good