Wednesday, September 19, 2007

Where else can I get BLS' CPI information?

In response to an inquiry I received today, there are several places on the BLS website where the Consumer Price Index can be accessed. On CPI release day, At 8:30 EST (translation, way too early for me on the West Coast), National CPI data are transmitted to everyone in a press release format. Some journalists, given advance access under a strict lockdown arrangement, are then released from captivity to wire the information to their respective news outlets.

At 8:30 EST, local data are also released on the publicly available database. However, press releases for local data are updated regionally and are not posted until later, typically on the same day.

Here are the top level numbers for the San Francisco Region.



For more detail, access the database link to get line items.

Thursday, September 13, 2007

Open source, mashups, and data competition

I ran across an interesting meld of public data and private use today. One of many trends in websites is the use of mashups, sites built upon data from sources utilizing a standard Application Programming Interface (API). Cynergy Systems provides an interesting visual interface for Census data. However, they do not get their data from Census. From what I gather, Cynergy Systems receives census information in a readable format from StrikeIron. StrikeIron takes advantage of the lack of Census' interface, and makes the database readable for programmers.

How Cynergy Systems makes money, I do not know. Perhaps Cynergy Systems uses this site as interactive billboard for consulting services. Or, in this day of "if you build it, they will come," perhaps the next phase after attracting viewership is yet to come.

P.S. While on StrikeIron's site, check out their press release on synergistic collaboration between the firm and BEA.

Tuesday, September 4, 2007

Refuting the Mommy Money Myth

The Monthly Labor Review (MLR) from BLS recently released an article on earnings by gender. To a large extent, the article does not expose new revelations. Women are paid less and a sizeable gap exists.

At the risk of alienating my female readers, I have always been a believer that the primary cause of the gender gap is self-selection. That is, women -- particularly those with young families -- are more selective of their working conditions, finding some situations unacceptable in exchange for remuneration. Men, on the other hand, may not find those same conditions objectionable, or may feel as though lucrative remuneration weighs more heavily on their decisions.

To this end, the MLR article refutes at least a portion of my theory, particularly with regard to family situation. The article computes a dispersion index indicating the range of wages paid. Dispersion index P90/10 indicates the spread between earnings at the 90th percentile and 10th percentile. An index value of 5.0 indicates that at the 90 percent level, participants earn five times what those at the 10 percent level do. Suprisingly, women with children at home posessing a bachelor's degree or higher, have a higher dispersion index than similarly situated women with no children at home. Could that mean that bargaining power plays at least as strong a role as free choice? Do mothers posess skills that the workplace is willing to barter for at higher rates?

This does make sense at an intuitive level. A combination of understood free choice and high skill set would result in employers' willingness to negotiate more generous pay in exchange for services. Importantly, the quantitative analysis of this article refutes the commonly held belief that women give up salary for children. Why women in general give up salary -- if indeed it is a conscious choice -- remains unanswered.
-------------------------------

Earnings by gender: evidence from Census 2000
http://www.bls.gov/opub/mlr/2007/07/art3full.pdf

Wednesday, August 22, 2007

SRDS, alas

I intended to enter a brief hiatus after producing a thoughtful piece on how advertising rates have moved over time. Alas, the Alameda County libraries no longer carry the Standard Rate and Data Service (SRDS) ad rate references because the series became too cost prohibitive.

My hypothesis is that quoted ad rates have stagnated over time -- the result of evolving advertising media and an explosion of suppliers. Quoted rates are likely flat, but actual rates (analogous to the discounted price of DSL, versus what providers claim you would pay for a competitor) are most likely falling.

Case in point, Google's bid service for major newspaper ads. It's like Priceline for air travel. Put in a bid, see if it's accepted by the Los Angeles Times, San Francisco Chronicle, or whichever rag you've chosen. Ditto for radio. In fact, the radio series has caused a stir with broadcast companies concerned if the FCC will label the lowest bid as the rate that all political campaigns receive.

Just to keep this post BLS relevant, I ran a quick list of producer price indexes (PPI) to see how prices have changed. According to BLS, cable network advertising services' prices have fallen in price. In July 2007, prices were down 11.9 percent from the previous year. Radio station time sales' prices were down 2.6 percent. So far, the hypothesis seems to hold. BLS does ask respondents for price adjustments such as discounts.

For Daily and Sunday newspaper advertising sales? Up, 2.5 percent from the previous July. Wha?!! In fact, newspaper ad prices have been on a perpetual price climb in the 4 to 5 percent range for ten years now. Granted, the 2.5 percent represents a deceleration, perhaps a harbinger of things to come. In light of revenue losses, does this mean that the great media moguls do not have a sense of price elasticity? Are they raising prices and losing clients?

I doubt that is the case. Hearst did not build an empire on poor economic logic.

Instead, I'd look more closely at the process of collecting PPI data. BLS sends data collectors out with government forms to be filled out by newspaper personnel on a "voluntary" basis. Newspaper staff, fearful of their company's solvency, are less inclined to put a lot of effort into any paperwork that does not fulfill the ultimate mission of profitability. Further, they are now understaffed, having lost co-workers to attrition, or worse. Trading places, how much effort would you put into figuring out how much the company has given to advertisers in discounts?

Like SRDS, how real are BLS figures on ad rates? As a business person, every time I've declined an advertisers' pitch, I've been offered another rate -- sometimes discounted by 50 percent. It's a wild world of advertising out there, looking for some type of arbitrage.

Monday, August 20, 2007

A personal note to readers

We all gravitate toward familiarity, reluctant to release ourselves from our past. This blog has been a wonderful way to bring users closer to the data I've heard requests for over the past five years, and a good way for me to share what I have learned.

This week however, I'll close with research from Standard Rate and Data Service (SRDS), the de facto source of advertising rates, then take a two week break to ramp up for the seasonal rush of the school year. I'll also work on my memoir, a piece that has begun to take on a life of its own, moving from a sense of civic duty to a cultural circumspection -- a tragicomedy with an emphasis on maintaining humor throughout.

When I return, little will have changed in the broadest sense of the world. I will return, however, a more astute businessperson, with thoughts on whether BLS data affect microeconomic decision making or only macroeconomics. Stay tuned.

Sunday, August 19, 2007

Two-handed economics of housing

On one hand, more than 1/3 of homeowners do not have mortgages (table 8). On the other hand, for all consumers, shelter expenditures rose more than 10% from 2004 to 2005 (table A).

Consumer Expenditure Survey 2005
http://www.bls.gov/cex/csxann05.pdf

Wednesday, August 15, 2007

Consumer Price Index Metro Comparison

I've always wanted to publish the CPI in a table like the one below, because I think it makes sense.

I've addressed the issue of the CPI basket of goods in an earlier discussion about ACCRA v. CPI . For this exercise, we are comparing apples to apples in discussing the average household's consumption price increase or decrease, not a change in line item prices.

Journalists are interested in who's on top. In this case, the Tampa, Fl area tops the metro list at 5.2 percent, more than twice the rate of the national average. From an acceleration standpoint, Tampa also tops the list, increasing 1.7 percentage points from the twelve-month average ending 2006.
============================================================

CONSUMER PRICE INDEX -ALL URBAN CONSUMERS - ALL ITEMS
12-mo avg
Percentage Point Change
First half 2007
Second Half 2006
U.S. city average2.52.6-0.1
Size Class A greater than 1,500,0002.72.8-0.1
Size Class B/C Between 50,000 and 1,500,0002.32.30.0
Size Class D less than 50,0002.72.8-0.1
METROPOLITAN STATISTICAL AREAS RANKED BY FIRST HALF 2007 RATE
Tampa-St. Petersburg-Clearwater5.23.51.7
Honolulu5.05.8-0.8
Seattle-Tacoma-Bremerton3.94.1-0.2
Miami-Fort Lauderdale3.64.2-0.6
Phoenix-Mesa3.62.80.8
Washington DC -Baltimore3.43.30.1
Portland-Salem3.42.50.9
Los Angeles-Riverside-Orange County3.33.5-0.2
San Francisco-Oakland-San Jose3.33.30.0
Minneapolis-St. Paul2.81.81.0
Atlanta2.82.20.6
New York-Northern New Jersey-Long Island2.73.5-0.8
Chicago-Gary-Kenosha2.61.11.5
Pittsburgh2.52.6-0.1
Denver-Boulder-Greeley2.53.4-0.9
Cincinnati-Hamilton2.42.9-0.5
Kansas City2.42.30.1
San Diego2.33.0-0.7
Philadelphia-Wilmington-Atlantic City1.93.4-1.5
Detroit-Ann Arbor-Flint1.92.2-0.3
Boston-Brockton-Nashua1.82.4-0.6
Milwaukee-Racine1.81.80.0
St. Louis1.81.50.3
Cleveland-Akron1.60.51.1
Anchorage1.52.2-0.7
Houston-Galveston-Brazoria1.42.3-0.9
Dallas-Fort Worth0.71.7-1.0