Another disappointing employment report for Arkansas:  The latest data from the U.S. Bureau of Labor Statistics and the Arkansas Department of Workforce Services show that unemployment was up again in August, but not enough to raise the unemployment rate from the 7.4% level of July.  Over the first 8 months of 2013, the unemployment rate has crept up by 0.3%, while the number of unemployed has increased by 2,710.  The number of employed from the household survey declined by 5,322 in August, contributing to a net loss of 27,490 since December of 2012.

Payroll Survey:
The report on nonfarm payroll employment for Arkansas was no more encouraging.  The net change in employment from July to August was -1,200 (seasonally adjusted).  Employment was lower in most sectors, including a declines of 600 in Construction and 800 in Wholesale Trade.  Two sectors did show month-over-month increases:  Financial Activities (+700) and Manufacturing (+900).  Over the past 12 months, we’ve seen healthy gains in Retail Trade, Education & Health Services, and Transportation & Utilities, but several sectors have seen declines.  Notably, the goods-producing sectors have continued to show net declines.  Compared to August 2012, payroll employment has increase by just 12,200 — approximately 1.0%.

The household data and payroll data continue to provide conflicting signals:  As shown in the figure below, payroll employment has continued to trend upward (albeit slowly and irregularly), while the household employment measure has been falling since the beginning of 2012.  It seems likely that the annual benchmark revisions to the payroll data will ultimately be negative, bringing the two employment measures into closer alignment.  With the release of the Quarterly Census of Employment and Wages next week, we are likely to see some evidence supporting this conjecture.

# # #

*Seasonally adjusted data for Arkansas nonfarm payroll employment, reported in a format compatible with the monthly news release from the Arkansas Department of Workforce Services, are available here:  Table – Seasonally Adjusted NFPE.

The latest report on state-level employment and unemployment suggests ongoing weakness in Arkansas labor markets.  The unemployment rate edged up to 7.4% (although the increase from June’s 7.3% rate was literally a matter of rounding error).  At 7.4%, the Arkansas unemployment rate is unchanged from a year ago and is now at the same level as the U.S.  This is the first time since the early months of the recession that the Arkansas unemployment rate has not been lower than the national average.

July’s uptick in the unemployment rate was primarily due to a decline in the number of employed persons (-4,764).  The number of unemployed was basically unchanged (+37).  Due to the fall in employment, the overall size of Arkansas’ labor force declined — resuming the trend that we’ve seen over the past year-and-a-half.  Since January 2012, the Arkansas labor force is down by nearly 40,000 — approximately 2.9%.

Payroll Survey:
Total nonfarm payroll employment was up by 1,900 in July (seasonally adjusted) — a fairly modest increase but a welcome change from the declines of the two previous months (revised).  Goods-producing sectors showed a collective decline of 1,600 jobs while service-providing sectors expanded by 3,500 — mostly in the areas of Health Services and Professional & Business Services.  Compared to a year ago, Arkansas employment is up by 13,600, mostly from increases in Health Services, Retail Trade, and Transportation & Utilities.

As with the household survey, the data from the payroll survey indicate that the pace of Arkansas’ recovery from the 2008-09 recession is lagging behind the U.S. average.  Although the recession did not hit Arkansas as hard as the rest of the country, employment at the national level has shown steady improvement while Arkansas’ employment growth has been irregular and slower, on net.  As of July 2013, cumulative changes in employment since the beginning of 2008 measure -1.5% for the U.S. and -1.8% for Arkansas.

#  #  #

The latest data from the Bureau of Labor Statistics and the Arkansas Department of Workforce Services show continuing stagnation in Arkansas labor markets.  The unemployment rate was unchanged at 7.3%, where it has held for most of the past 18 months.

With the number of employed down by 1,732 in June and the number of unemployed up by 628, the state’s labor force declined by 1,104.  Hence, the tentative signs of improvement in the Arkansas’ labor force participation rate that we had seen over the past two months did not persist.  Over the past 12 months, the household survey is showing a decline in employment of nearly 21,000.  The number of unemployed is down by 1,685 but most of that decline took place in the second half of 2012.  Over the first 6 months of 2013, the number of unemployed has shown a net increase of 1,935.

Payroll Survey
The payroll survey showed no change in employment for the month (seasonally adjusted).  Employment declines were recorded in Manufacturing, Financial Services, Professional & Business Services, and Education & Health Services.  Employment in the retail sector fell rather sharply (-1,700).  Other service sectors registered employment increases, led by Other Services (+1,900) and Transportation & Utilities (+1,200).  Construction employment was also up slightly.  Compared to a year earlier, payroll employment was up by 8,500.

The latest data from the household and payroll surveys continue to give conflicting signals.  Since mid-2012 the payroll survey has shown slow but positive employment growth while the household survey has shown employment declining sharply.

One indication of the eventual reconciliation of these trends is suggested by the most recent data from the Quarterly Census of Employment and Wages (QCEW).  The QCEW data are considered more accurate than the monthly payroll surveys, and are used to revise the payroll data during the annual benchmark revision process.  In the most recent release, the QCEW showed an employment increase of only 1,900 jobs during 2012 (December 2011 through December 2012).  In contrast, the current payroll data indicate job growth of 4,000 for the same period.  This suggests that when the payroll data are eventually revised, they will show slower job growth than current payroll statistics are suggesting — at least for the second half of 2012.

#  #  #

The pattern of sluggish labor market performance in Arkansas in recent months continued into May.  According to new data from the Bureau of Labor Statistics and the Arkansas Department of Workforce Services, the unemployment rate ticked up one-tenth of a percent to 7.3% from a revised 7.2% in April.  (The revision to the unemployment rate for April amounted to only 0.00004%, but that was enough to tip the rounding error upward.) The national unemployment rate also rose by one-tenth in May, to 7.6%.  As shown in the chart below, the gap between the state and national unemployment rates has continued to narrow over time.

There were a couple of positive tidbits in the report from the household survey:  The number of employed and the size of the total labor force were up for the second consecutive month.  Both of these indicators had been declining precipitously since early 2012.  However, the number of unemployed was up by over 1700, bringing the total number of unemployed back up to a level comparable to last October.

Payroll Survey
Data from the survey of employers was no more encouraging than the household survey:  Nonfarm payroll employment declined by 5,900 (seasonally adjusted*).  Many of the service sectors that have been supporting employment growth during the slow recovery were down in May.  For example, Professional & Business Services fell by 2,300 and employment in Leisure & Hospitality Services was down 1,100.  Even employment in Education & Health Services fell by 900.

Two bright spots in the payroll survey:  Although it remains lower than a year ago, construction employment was up 700 from April to May (seasonally adjusted*).  Jobs in retail trade also increased — up by 1,100.  Retail sector employment is now nearly 13,700 higher than at the employment trough of February 2010, and up 6,700 from pre-recession levels.  Although several service sectors showed employment declines for the month, overall employment in services is generally  higher than before the recession.  But with employment in goods-producing sectors down by over 41,000 since December 2007,  the overall recovery of jobs since February 2010 has been only 32,600 (compared to 57,100 job losses between December 2007 and February 2010).

Mixed Signals
Neither the household survey nor the payroll survey provided much in the way of good news this month.  But more disturbing is the conflicting employment trends indicated by the two reports.  Over the past 12 months, payroll employment has increase by around 3,000 jobs — not very strong growth, but positive.  Over the same period, however, the household survey is showing a decline of more than 23,000 jobs.  It is not unusual for the two surveys to give conflicting signals in the short-run, but rarely do they diverge this significantly over a longer period of time.

Generally speaking, economists consider the payroll survey to be more accurate, but it is also subject to considerable future revision as more detailed information becomes available.  It is quite likely, therefore, that the divergence between household and payroll reports will be reconciled by a downward revision to the payroll data.  Annual revisions will not be officially published until next March, however, so we’ll be monitoring incoming data closely in an effort to sort through the differences we’re seeing between the two surveys.

#  #  #

The Bureau of Labor Statistics recently published some preliminary statistics on 2012 state unemployment rates – broken down by gender, race, ethnicity and age.  A comparison of Arkansas with the rest of the country yields some interesting observations.

As shown in the table below, the overall unemployment rate in Arkansas was one-half percentage point lower than the national average in 2012, and Arkansas ranked near the median among the 50 states plus District of Columbia.  For the U.S. as a whole, the 2012 unemployment rate was slightly higher for men than for women, but there was no gender difference in Arkansas.

Sorting by race, unemployment rates were significantly higher for Black or African American residents than for Whites.  This is true both nationwide and for Arkansas; however, the degree of disparity is larger in Arkansas.  Among Whites, the Arkansas unemployment rate was 1.2 percentage points lower than the national average, but for Black or African American Arkansans, the unemployment rate was more than 2 percentage points higher than the U.S. rate.  Rankings among the states also show the relative disparity:  the unemployment rate among Whites was below the median while the unemployment rates among Black or African Americans were well above the median.  For Black or African American men, Arkansas had the 12th highest unemployment rate in the country.

In contrast, unemployment rates among residents with Hispanic or Latino ethnicity, unemployment rates were far lower in Arkansas (5.5%) than for the U.S. as a whole (10.3%).  For that ethnic group, Arkansas unemployment rates ranked among the lowest in the country.  For Latino or Hispanic women, the Arkansas unemployment rate was the fourth-lowest in the U.S.

Unemployment rates by age group break down similarly for both the U.S. and Arkansas, with rates significantly higher for younger age groups than for older cohorts.  Nevertheless, there are some notable distinctions.  Youth unemployment — 16-19 years and 20-24 years — was higher in Arkansas than the national average, with the state ranking 11th highest in both specific age groups.  Among older workers, unemployment rates in Arkansas are well below the national average. Of particular note is the 55-64 cohort, which has an unemployment rate of only 3.3% in Arkansas, compared to a 5.9% national average.

Among the various age groups, unemployment rates for men and women show some marked differences.  Among Teenagers (16-19), the rate for males is far higher than for females – and well above the national average.  Moving up just one age bracket, the pattern is reversed for 20-24 year olds.  There is also a sharp gender difference in the 45-54 category:  The unemployment rate for women is only 3.8% (compared to 6.2% nationally), while the rate for men is 6.4% — above the national average.

The unemployment rate is an important indicator of labor market performance, but my no means is it the best summary measure.  Labor force participation rates and employment-population ratios also provide important information about the employment status of various demographic groups that can help interpret the unemployment rate.  For example, we noted that unemployment rates among 55-64 year-olds is relatively low in Arkansas.  However, the participation rate for this cohort is only 51.6% in Arkansas, compared to 64.5% nationwide.  As a result, employed persons 55-64 years old comprise only 49.9% of the total population in that age bracket in Arkansas, but for the U.S. the employment-population ratio is 60.6%.  Hence, one reason that unemployment is so low for 55-65 year olds in Arkansas is that far fewer of the members of that cohort are in the labor force.

For those readers interested in looking at these figures in more detail, a supplemental table —  comparing the U.S. and Arkansas in terms of population shares, labor force participation rates, employment-population ratios along with unemployment rates – is available here (Excel spreadsheet)

The unemployment rate in Arkansas was unchanged at 7.2% in March.  The household survey suggests that the number of unemployed has stabilized at approximately 96,000, but ongoing decline in household employment continued for a 13th consecutive month.  Since February 2012 household employment has dropped by approximately 32,600 and the corresponding measure of the state’s labor force has dropped by 37,700.  If this entire decline in employment was classified as an increase in unemployment (instead of as a reduction in the labor force), the state’s unemployment rate would now stand at 9.7%.

Payroll Employment
The payroll survey continued to show greater strength than the household report.  Total nonfarm payroll employment was up by 2,400 in March (seasonally adjusted), following a revised increase of 3,000 in February.  As shown in the table below, job losses were particularly notable in Leisure and Hospitality Services and Manufacturing.  Without seasonal adjustment, employment in Leisure and Hospitality was up slightly (700 jobs), but that increase is far smaller than would typically be expected in March.  Sectors with notable increases in employment included Construction and Business & Professional Services.

The two measures of state employment — the household survey and the payroll survey — continue to provide conflicting signals.  The household measure of employment has now declined to a new post-recession low, while the payroll survey has shown positive growth (albeit slow and uneven) since February 2010.  Except for the most recent observations (which are still subject to future revision) the payroll figures are based on a very inclusive and comprehensive set of data.  This suggests that the anomalous reading are coming from the household survey.  However, we cannot rule out the possibility that the household survey is documenting significant weakness in Arkansas labor-market trends that have yet to be fully reflected in the payroll data.  Stay tuned.

#  #  #

The Arkansas unemployment rate was in February was 7.2%, unchanged from the previous month.  New data from the Bureau of Labor Statistics and Department of Workforce Services showed that the number of unemployed declined by nearly 600, partly offsetting the January increase.  However, the February household survey also showed a sharp decline in employment for a second consecutive month.  The two-month decline in employment with little change in unemployment implies a sharp contraction of the labor force.  This was the for 13th consecutive monthly decline in the size of the labor force, with an apparent acceleration of the trend during the first two months of 2013.  Since January 2012, the labor force has fallen by more than 30,000 — about 2.2%.

Payroll Survey
Nonfarm payroll employment was up by 2,800 in February (seasonally adjusted*).  For the first two months of the year, payroll employment rose by only 600, and is essentially unchanged (+800) from October of 2012.  As detailed in the table below, employment was up for the month in each of the goods-producing sectors and in most of the service-providing sectors as well.  The only monthly declines were in Information Services, Financial Services, and Other Services.   Employment in most sectors has increased over the past year, and most have shown positive growth since the post-recession employment trough of February 2010.  Notable exceptions are Construction, Manufacturing, Information Services, and Other Services.  Overall, employment is up 31,000 since the February 2010, a recovery of approximately 54% of the jobs that were lost during the recession.

The overall employment trends indicated by the household survey and the payroll survey have been diverging in the first part of 2013.  As shown in the figure below, the two data sources rarely provide identical signals about short-term employment changes, yet they generally show similar trends. However, the sharp decline in household employment in January and February clearly contrasts with the relatively unchanged employment profile suggested by the payroll survey.  Ordinarily, economists tend to rely more on the payroll figures as providing a more complete and accurate assessment of labor market conditions.  But as we have recently seen, the payroll employment statistics are preliminary and subject to future revision.  As additional employment data are released in the coming weeks and months, we’ll be closely monitoring the relative performance of the household and payroll employment statistics, seeking to reconcile the conflicting signals they seem presently to be providing.

#  #  #