Two housing reports in the week just demonstrated, yet again, economists are not infallible.
On Tuesday, the National Association of Realtors (NAR) reported existing home sales for December: 4.94 million against a consensus forecast of 5.1 million. The forecast would have represented an increase of 1.2 percent over the 5.04 million sales pace (these data are reported as a seasonally adjusted annual rate). Instead, the report showed a drop of 1 percent from the revised November sales rate of 4.99 million and an almost 2 percent drop from the original report.[IMAGE]
Then on Friday, the Census Bureau and HUD reported jointly 369,000 new homes were sold in December compared with a consensus forecast of 388,000. The forecast would have represented an increase of almost 3 percent from the 377,000 November sales. Instead, the forecast was a drop of 2.5 percent from the revised November sales of 398,000. The actual sales report was a 7.3 percent decline.
If your head is spinning from the numbers, wait, here's more.
Economists of all stripes will insist their forecasts for both existing and new home sales are correct.
We shouldn't--according to one school of thought --be looking at one month's data set but rather quarterly numbers--although they are reported, and forecast, monthly.
The two home ""sales"" reports measure different activity--the new home sales report measures contracts, the existing home sales report measures closings, and neither is a perfect gauge of the housing market. Indeed, according to the Census-HUD joint release, the new home sales report is ""estimated from sampling surveysÃ¢â‚¬Â¦subject to sampling variability as well as nonsampling error including bias and variance from response, nonreporting and undercoverage.""
The release includes a percentage range. In the case of the December report, the 7.3 percent drop in sales could have been plus or minus 15.3 percent: anywhere from down 22.6 percent to up 8.0 percent. ""If a range does not contain zero, the change is statistically significant. If it does contain zero, the change is not statistically significant, that is it is uncertain whether there was an increase or a decrease.""
In other words, sales may have gone down, they may have gone up, but we really don't know.
The NAR report contains no such disclaimer, but rather than focusing on the month-month change when the change is negative, NAR highlights the year-year change as it did in its December report. If NAR had indeed looked at quarterly numbers, it would have seen the fourth quarter sales averaged 4.9 million per month compared with 4.7 million in the third quarter and 4.4 million in the fourth quarter of 2011.[COLUMN_BREAK]
NAR is, of course, an advocacy group trying to put housing in the best possible light. We have a right to have different expectations from a government report.
NAR stresses the year-year rather than month-month change in the median sales price. The month-month change in the sales price would logically affect sales. During 2012, sales dropped month-month five times and in four of those months the median price rose; the median price fell in four of the six months in which sale increased (sales were unchanged in May). While the median price moves not only on supply and demand, but on the mix of homes sold in a given month, to suggest there is no or limited correlation between prices and the number of homes sold would be unrealistic.
The cautionary tale is to watch not only what the numbers are, but what analysts say they are.
There are several important housing related reports due out next week, but they will take a backseat to Wednesday's report on fourth quarter GDP and Friday's report on the employment situation.
The NAR will report its Pending Home Sales Index Monday, its version of the New Home Sales report. The index has improved for four of the last five months and in its purest form, is an indicator of the existing home sales report two months forward. (Monday's PHSI for December should ""predict"" existing home sales in February). That said, existing home sales rose in four of the last six months, precisely what the PHSI forecast--except that the months were different. Using the PHSI as an indicator, sales should have dropped in August and October; they fell month-month instead in September and December. The forecast for the December PHSI is for another modest 0.3 point gain, far below the average increase of 2.4 points for each of the last three months. If the consensus is accurate, it will add a dash of reality to the unbridled optimism surrounding housing.
Standard & Poor's will report Tuesday on the Case-Shiller Home Price Index for November. The comparable Federal Housing Finance Agency index showed a 5.7 percent year-year gain, supporting the forecast for a 5.3 percent year-year Case Shiller improvement. If the forecast is accurate, it would be the strongest year-year increase in the 20-city index since August 2006 but still leave the index 29.5 percent below its July 2006 peak.
Hurricane Sandy's toll on the economy--reported in bits and pieces in other data releases--will show up in Wednesday's report on Gross Domestic Product--a 0.8 percent annualized growth rate, which would be the second weakest quarterly gain of the recovery. While consumption appears to have held up, net exports will be the biggest drag, with a drop off in government spending also holding back growth. Government spending fueled growth in the third quarter, powered in part by end of fiscal year defense spending.
Friday's employment situation report, despite the sharp drop in initial claims for unemployment insurance, is expected to show the unemployment rate remained flat at 7.8 percent as the labor force grew with re-entrants. Job growth is expected to hover around 155,000, which appears to have become the new norm for the recovery.
_Hear Mark Lieberman Friday on P.O.T.U.S. radio, Sirius-XM 124, at 8:45 am eastern time._