The reason:  “Grocery sales, as a percent of total sales, dropped sharply in June.  This could actually be a positive development.”

The preliminary estimate for second quarter growth of Arkansas Taxable Sales (ATS), announced in a previous post, was based on information from the July 2010 General Revenue Report from the Arkansas Department of Finance and Administration (DFA).  With the recent release of more detailed information in the DFA publication Arkansas Fiscal Notes, a revised estimate of ATS shows a somewhat smaller increase in the second quarter of 2010:  Sales rose by 3.0 percent, rather than the 4.0 percent previously reported.

The downward revision is disappointing.  But as shown in the chart below, it doesn’t really change the qualitative interpretation of second quarter data:  Sales accelerated in 2010:Q2 and have risen significantly since the apparent trough of 2009:Q3. 

Sources: Arkansas Department of Finance and Administration and author's calculations.

Investigation of the reason for the downward revision in ATS reveals an interesting observation:  Grocery sales, as a percent of total sales, dropped sharply in the June.  This could actually be a positive development.  

Preliminary estimates of ATS are based on gross receipts, which primarily reflect total sales tax collections.  But total sales taxes represent a mix of taxes on groceries (at a current tax rate of 2 percent) and taxes on non-groceries (at a tax rate of 6 percent).  Changes in the mix of these two categories alter the effective tax rate on total sales.

The revised figures for ATS are based on data on receipts from the Arkansas Conservation Tax.  Because it is written into the Arkansas constitution, the Conservation Tax was not altered by reductions in the tax rate on groceries in July 2007 (from 6 percent to 3 percent) and in July 2009 (from 3 percent to 2 percent).  Instead, the lower sales tax on groceries is reflected in proportional reductions in the other three components of Arkansas’ Sales and Use Tax (the General Sales Tax , the Educational Adequacy Tax, and the Property Tax Relief Tax).

When groceries decline as a fraction of total sales, these three taxes show a disproportionate increase in overall tax receipts (because a larger share of sales are taxed at a higher rate).  The preliminary ATS estimates (which are implicitly based on the assumption of a constant share of grocery sales) would therefore overstate the increase in total taxable sales if this were the case.

This appears to be exactly what happened in June.  By comparing changes in tax receipts from the conservation tax with receipts from the other sales tax components, it is possible to estimate the share of grocery sales as a proportion of total sales (see technical note).  The results of this estimation are shown in the chart below.  Before the national recession began at the end of 2007, groceries constituted about 6 percent of total sales.  This proportion rose during the early stages of the recession in 2008, and increased sharply at the end of the year when the financial crisis intensified the economic downturn.  Estimates for June 2010 show a sharp decline, from about 12 percent to just over 8 percent.

Sources: Arkansas Department of Finance and Administration, and author's calculations.

Economic theory distinguishes goods by the responsiveness of their demand to changes in income.  Some basic grocery items are considered “inferior goods,” reflecting the fact that fewer are purchased as income increases (in favor of higher-quality goods).  In general, groceries are considered “normal goods,” for which demand increases when income rises, but typically less than one-for-one.   In contrast, some  goods are considered “superior goods” or “luxury goods,” for which demand increases in greater proportion to income.

The observation that grocery sales increased as a percent of total sales during 2008 and into 2009 is therefore consistent with a downturn in household income during the recession.  The recent decline in the grocery sales suggests that households are increasing their purchases of superior goods relative to normal and inferior goods.  This provides indirect evidence of higher household incomes (or at least expected incomes).

This observation is supported by data for only one month, so it should be taken as conjectural.  However, it suggests that the downward revision in ATS for the second quarter is not unambiguously bad news.

#  #  #

The Arkansas Taxable Sales series is calculated by IEA to serve as a timely measure of Arkansas retail sales.  The series is derived from sales and use tax data from DFA, adjusting for the relative timing of tax collections and underlying sales, changes in tax laws, and seasonal patterns in the data.

A spreadsheet of the most recent data is available here:  Arkansas Taxable Sales data (Excel file)

Information on residential real estate markets has dominated the news over the past few days.  When it comes to evaluating the housing markets in Arkansas, recent reports from various sources show a consistent picture of recent trends, but the outlook is very uncertain.

Home Sales

The Arkansas Realtors Association (ARA) released statistics this week showing that June home sales were up slightly over the previous year, but down from the previous month.  As shown in Chart 1 below, housing sales follow a recurring seasonal pattern, with the summer months tending to have higher sales than winter months.  July is often the peak sales months of the year, but so far this year, sales peaked in April and have declined since then.

Source: Arkansas Realtors Association

Disentangling the cyclical and seasonal effects is crucial to understanding recent trends.  Chart 2, below, shows seasonally adjusted data that are averaged over calendar quarters to smooth out some of the month-to-month variability (the data are seasonally adjusted using the technique described in a previous article).

Sources: Arkansas Realtors Association, National Association of Realtors

With these refinements, the cyclical pattern of home sales is more evident:  After declining throughout 2007 and 2008, sales steadily increased over the course of 2009.  The first quarter of 2010 saw a sharp decline, followed by a rebound in the second quarter.  Recent data for Arkansas home sales from the National Association of Realtors  (NAR), also displayed in Chart 2, confirm the recent quarterly pattern in the adjusted ARA data.

The sharp decline in the first quarter of 2010 can be partly attributed to weather.  Sales are usually low in the winter months, and this past winter in Arkansas was particularly ill-suited to house shopping.  Sales recovered in the second quarter to about the same pace as in the fourth quarter of 2009.

But when it comes to assessing the outlook for home sales, the expiration of the home-buyer tax credits is the elephant in the room.  Sales in the fourth quarter of 2009 were boosted by last-minute purchases under the original tax credit program for first-time buyers that expired at the end of October.  Weather might have kept activity low in the first quarter, but the impending deadline for taking advantage of the extended and expanded home-buyer tax credit on April 30 was undoubtedly a factor in the second quarter resurgence.

The recent tax-credit deadline required that a contract be in place by the end of April, but the closing date could be later.  As a result, recorded sales in May and June included some transactions that are associated with last-minute contracts to qualify for the credits.  The crucial question is whether the second quarter increase reflects sales that would have taken place later in the year in the absence of the tax-credit deadeline – implying that we should expect a sharp decline over the next few months — or whether it reflects a sustainable improvement in market conditions.

Preliminary information suggests that we will see a drop-off in the third quarter.  New data on existing home sales and new home sales nationwide showed sharp declines in July.  Information on home sales in central Arkansas for July show that Arkansas housing market experienced this downturn as well.  For example, sales in Pulaski county were reported to be down 30.2 percent from the previous month and down 31.5 percent from July 2009.  When the statewide sales figures for July are released by the ARA next month, they are likely to show a sharp downturn.  

Arkansas home sales over the remaining summer months will probably be relatively weak, as the market returns to an equilibrium undistorted by government subsidies.  On the other hand, mortgage rates remain low and house prices have fallen (see below), so the market continues to be favorable for buyers.  This factor should help sustain sales later in the year.

House Prices

In addition to reports from the ARA and NRA, new data on housing prices were released this week by the Federal Housing Finance Authority (FHFA).  Since the housing downturn began in early 2007, the FHFA purchase-only price index has shown a distinctly lower rate of appreciation than the all-transactions index, which includes appraisal data from refinancings.  Ultimately, it is the market sales price that matters, so the purchase-only index provides more accurate information on current market conditions.  However, refinancings are more likely to be prevalent in areas where houses are retaining their value, so the all-transactions index might give a better indication of long-run home values.  [See also, Comment:  Differences Among Home-Price Indices]

As shown in Chart 3, below, the two measures of house prices have diverged considerably since the market downturn began, with   the purchase-only index showing considerably more variability.  Over the past four quarters, the purchase-only index has experienced quarterly ups and downs, but has changed little on net.  The all-transactions index has fallen by 2.4 percent.

Source: Federal Housing Finance Authority

Under either method, house prices in Arkansas have not experienced declines as large as the nationwide average.  Over the past four quarters, for example, the FHFA all-transactions index for the U.S. showed a 5 percent decline compared to the 2.4 percent decline in Arkansas. 

Chart 4, below, compares the two FHFA house price indexes for Arkansas with the average sales-price data from ARA.  The series are normalized to take on a value of 1.0 in the first quarter of 2007, so the data presented in the chart show cumulative price changes since that time.

Sources: Federal Housing Finance Authority, Arkansas Realtors Association, and authors' calculations

 Not surprisingly, the average price data from ARA shows some similarity to the FHFA purchase-only index.  Both measures are based on prices of actual sales.  Both series show a sharp decline in the first quarter of 2010.  This drop is associated with the decline in home sales discussed earlier.  With fewer homes selling in general, the segment of the market including foreclosures and other distressed sales constituted a larger share of total sales, depressing average prices.  With the pickup in market activity in the second quarter, the price spike was also reversed.

Cumulatively, from 2007 through the second quarter of 2010, house prices in Arkansas have decline modestly.  All three measures of home prices show prices down over the period, but by magnitudes of only about 1 to 3 percent.

The FHFA also releases all-transactions indexes for metropolitan areas.  As shown in the table below, these data show that house price changes vary considerably across Arkansas’ metro areas.

Source: Federal Housing Finance Authority

Northwest Arkansas has experienced the largest and most persistent declines in house prices.  Earlier in the decade, the Fayetteville metro area showed the greatest appreciation in the state, so declines more recently serve to offset some of those earlier gains.  Hot Springs has recently shown fairly large declines as well, but compared to five years ago, prices are up about 14.6 percent.  Data for Jonesboro, Texarkana and Little Rock show that house prices increased in the second quarter, with prices also higher than a year ago.  In fact, Jonesboro serves as a counterexample to the experience of Northwest Arkansas:  Rather than rising sharply in the early part of the decade and subsequently declining, house prices in Jonesboro have experienced a relatively slow but steady upward trend.

Source: Bureau of Labor Statistics

Analysis of the payroll survey is complicated by seasonal patterns.  Without seasonal adjustment, the data show employment down by nearly 11,000 in July.  However, as shown in the chart below, July is typically a low-point in the recurring seasonal cycle diplayed by employment data.  (See Seasonally Adjusted Unemployment Rates for Arkansas MSAs.)  After seasonally adjusting the raw data, the statistics from BLS show that payroll employment rose by 3,600 jobs.  Since the end of 2009, payroll employment has increased by nearly 16,000 jobs.

Source: Bureau of Labor Statistics

The increase was broad-based, with increases in every major sector except financial services and government.  Employment in manufacturing showed its 7th consecutive monthly increase.  Trade, transportation and utilities was up by 900 jobs, and business and professional services rose by 1,400.  Education and health services continued to expand, adding approximately 2,400 jobs in July.  The decline in government employment can be primarily attributed to layoffs of temporary census workers.

Source: Bureau of Labor Statistics

Typical summer doldrums aside, the July employment report provides additional evidence that labor markets in Arkansas are steadily improving.

#  #  #

In fact, newly calculated statistics show that Arkansas Taxable Sales rose by 4.0 percent in the second quarter (preliminary*) following a revised 2.6 percent growth rate in the first quarter of the year.  The latest reading represents the third consecutive quarterly increase.

Source: Calculated by the UALR Institute for Economic Advancement using data from the Arkansas Department of Finance and Administration

The improvement in Arkansas Taxable Sales growth in the second quarter contrasts with national retail sales data, which showed a slowdown in growth.  After increasing at a 2 percent rate in the first quarter, U.S. Retail Sales rose by only 1 percent in the second quarter.  As illustrated in the chart below, Arkansas Retail Sales have increased by a total of 7.6 percent over the past three quarters of growth, while U.S. Retail Sales have increased by only 6.8 percent over four quarters of growth.  The turnaround in Arkansas sales growth lagged the rest of the nation, but the most recent data on Arkansas Taxable Sales suggests that Arkansas is rebounding more quickly.

Sources: U.S. Census Bureau, Arkansas Department of Finance and Administration, UALR Institute for Economic Advancement.

The Arkansas Taxable Sales series is calculated by IEA  to serve as a timely measure of Arkansas retail sales.  The series is derived from sales and use tax data from DFA, adjusting for the relative timing of tax collections and underlying sales, changes in tax laws, and seasonal patterns in the data.

A spreadsheet of the data is available here:  Arkansas Taxable Sales data (Excel file)

* Data are preliminary until the release of the DFA report, Arkansas Fiscal Notes for July 2010, and will be updated when information becomes available.

The map of metro areas from BEA, displayed below, shows that income growth in Arkansas’ MSAs was mixed, but generally better-than-average.  Pine Bluff was in the highest-growth quintile, with Little Rock in the second quintile.  The remainder of the state’s MSAs ranked in the third and fourth quintiles, with none finishing among the bottom fifth.

Table 1 provides more detail, comparing personal incomes of Arkansas’ MSAs to the total metropolitan portion of the U.S.  Personal income declined in 5 of the state’s 7 MSAs, but increased in the Pine Bluff and Little Rock.   Income growth in each of Arkansas’ MSAs out-performed the nation’s metro areas as a whole.

Source: Bureau of Economic Analysis

Table 2 shows figures for per capita income in Arkansas MSAs.  In per capita terms, only Pine Bluff experienced positive income growth in 2009;  Its growth rate of 1.6 percent puts it in the top 25 among the nation’s metro areas.  Fayetteville’s per capita income declined slightly more than the national average, and Jonesboro’s decline was about equal to the total metropolitan portion of the U.S.  The other 5 MSAs in Arkansas experienced smaller losses than other metro areas across the nation.

Source: Bureau of Economic Analysis

The last two columns of Table 2 show per capita income in Arkansas’ MSAs as a percent of per capita income in the metropolitan portion of the U.S.  The latest data show that Arkansas’ metro areas generally remain below-average.  Fort Smith, Jonesboro, and Pine Bluff are in the lowest quintile of MSA per capita personal income.  Fayetteville, Hot Springs, and Texarkana are in the fourth quintile.  The MSA with the highest per capita income, Little Rock, ranks in the second quintile.  Despite the fact that Arkansas’ metro areas outperformed other parts of the country in 2009, they gained little ground terms of per capita income.

The 2009 comparison of personal income in Arkansas’ MSAs to the rest of the nation’s metro areas is affected by two factors.  First, the recession in Arkansas was not generally as severe as in many parts of the country.  This accounts for the fact that all of Arkansas’ MSAs experienced above-average income growth.  On the other hand, evidence from statewide income growth and Arkansas taxable sales suggest that Arkansas reached a trough in economic activity in the third quarter of 2009 — one quarter later than the national economy.  This tends to depress measured growth rates for the year compared to those parts of the country that entered an economic recovery phase earlier than Arkansas.  As a result, the annual income statistics for Arkansas’ MSAs in 2009 reflect more weakness than do current economic conditions.

In the spirit of the Arkansas-centric focus of the Arkansas Economist, we ask the question:  “How does this issue affect families in Arkansas?”

The Tax Foundation just issued a report that provides a state-by-state comparison of the tax implications for a “middle-income” family.   The report compares the  tax change for an average family in the middle 20% of each state’s income distribution.  The comparisons illustrate some interesting features of the tax changes that would result if the tax cuts are allowed to expire.

Table 1 shows the Tax Foundation figures for Arkansas compared to the U.S. as a whole.  For the average middle-income American family, the expiration of the tax cuts would mean a tax increase of $1540.  For a middle-income family in Arkansas the increase would be $1418–a little bit less than the national average.   In percentage terms, however, the difference is remarkable:  the national average represents an increase of 45% while the Arkansas increase is over 100%.

Source: The Tax Foundation

Table 2 shows a 50-state comparison of the Tax Foundation data, along with some additional calculations to show percentage increases and relative state rankings.   Arkansas ranks 43rd in terms of the dollar-value impact of allowing the tax cuts to expire, but it ranks number 2 (2nd only to Mississippi) in terms of the percentage increase in taxes that it would imply.

Source: The Tax Foundation

The state rankings reveal a clear pattern:  Those states in which middle-income families would bear the largest dollar-value increase in tax bills (e.g. Alaska, Connecticut, and New Jersey) are the same states that show the lowest percentage increases.  Those that have the smaller dollar-value changes (e.g. Mississippi, Arkansas, and West Virginia) are the states with the largest percentage changes.

This pattern is related to the structure the Bush-era tax cuts.   The tax cuts were across-the-board, but the largest cuts–in dollar terms–went to the wealthiest Americans.  In percentage terms, however, the largest reductions went to lower and middle-income Americans.   As a state with a relatively low median income, Arkansas has many families that benefited from higher exemptions, the reduction of the lowest tax rate, and larger tax-offsets like the increase in the child tax credit.  As a result, a complete expiration of these modifications to the tax code would result in large percentage increases in the tax burdens of middle-income Arkansans.

The course of legislative action on the issue is uncertain, but President Obama has proposed to extend the tax cuts for all taxpayers except married couples with incomes over $250,000 (or single people with incomes over $200,000).  Table 3 gives an indication of how many households this limited extension might effect:  It reports data from the Census Bureau’s 2008 American Community Survey (ACS) on the number of households in each state with an income greater than $200,000.  In Arkansas, over 20,000 households fall into this category– about 1.8% of the total.  This is the second-lowest percentage among the 50 states.

Source: U.S. Census Bureau, American Community Survey

An Excel file of the data in this post is available here.

June data for metro area employment and unemployment came out earlier this week (see Metropolitan Area Employment and Unemployment Summary from the BLS).  According to the raw numbers—not seasonally adjusted—unemployment rates were unchanged or higher in each of the state’s metropolitan statistical areas (MSAs).   However, June is typically a month in which unemployment ticks upward for purely seasonal reasons.   One factor:  teachers and students are off for the summer, and students in particular are reported as being unemployed if they are seeking, but fail to find summer jobs.** 

After accounting for these recurrent patterns, seasonally adjusted statistics paint a much brighter picture.  Unemployment rates fell in every one of the state’s MSAs.  In many cases, the declines were substantial.

Source: Bureau of Labor Statistics

Employment statistics for June (seasonally adjusted by the BLS) showed that payrolls fell in most of Arkansas’ MSAs.  Employment was down by 0.1% in Fayetteville, down 0.9% in Fort Smith, down 0.6% in Little Rock, down 0.4% in Memphis and down 1.1% in Pine Bluff.   Employment was unchanged in Jonesboro, but was up by 0.3% in Hot Springs and up 0.2% in Texarkana.

Cumulative job losses in the state’s MSAs since the start of the recession have varied considerably.   The chart below shows the percentage decline for each MSA since December 2007.

Source: Bureau of Labor Statistics

Jonesboro has lost the fewest jobs:  total losses reached a peak of 2.2% in April but have recovered over the past two months.  Similarly, Texarkana has come back from a cumulative loss of 3.3% in December of last year to a total loss of only 1.2% as of June.  Hot Springs losses totalled 6.7% in April, but have bounced back to within 3.8% of the December 2007 level.  Employment in the remaining MSAs is at or near recent low points, having yet shown little sign of recovery.

#   #   #

**The June 2010 issue of Arkansas Labor Market, from the Department of Workforce Services, includes a feature article describing some of the reasons that employment data show seasonal fluctuations.

The unemployment rate dropped by two-tenths of a percent, falling to 7.5 percent in June.  The number of unemployed fell by 4,100 to 100,943.  This is the lowest reading since the number of unemployed first topped 100,000 in June of 2009.   The unemployment rate would have fallen further had it not been for a contraction in the size of the labor force.   The labor force declined by more than 8,000 in June, following three months of previous declines.  The most obvious explanation for this phenomenon is an increase in the number of “discouraged workers” who have given up looking for employment–at least for the time-being.   Overall, however, the news from the June employment report should serve as a reason for more encouragement going forward.

Source: Bureau of Labor Statistics

The payroll survey for June showed a healthy rate of job-creation:  Nonfarm payroll employment increased by 6,000 jobs (seasonally adjusted).  Since December of last year, payroll employment has risen by 11,400 and it is up by 4,200 from a year ago.  From the low-point for employment measured in February 2010, the number of jobs has increased by more than 15,000.

Source: Bureau of Labor Statistics

Recent data have shown employment gains in some of the key sectors for the Arkansas economy:  Manufacturing employment was up by 1000 jobs, and has now increased for five consecutive months.  Employment in Trade, Transportation and Utilities edged up by 800 jobs in June, recovering the job losses from earlier in the year.  Employment in Education and Health services resumed a healthy pace of job creation, increasing by 4,500 jobs in June.

Source: Bureau of Labor Statistics

Interpreting the growth in public sector employment is complicated by seasonal patterns and changes in the Census Bureau’s employment of temporary workers.  On a not-seasonally-adjusted basis, government employment was down by 4,700 jobs, with an 1,800 decline in Federal Government employment largely attributable to the winding-down of Census Bureau employment.  However, much of the decline in state and local government employment was related to typical fluctuations in public school employment.  After taking account of this school-year phenomenon, seasonally adjusted employment by government entities increased by 1,800  (with federal government employment down by 2,100 jobs and state & local employment up by a total of 3,900 jobs).

#  #  #

First, The Business Cycle Dating Committee at the National Bureau of Economic Research has laid out clear criteria for defining a recession.  They cover a broad range of economic measures, including production, income and employment–not just quarter-to-quarter movements in GDP.  In fact, the Business Cycle Dating Committee has never defined or declared that we’ve seen a double dip recession.  The chair of the Committee, Dr. Robert Hall has recently referred to the recessions of 1980 and 1981 as the closest thing he’s seen to a “double-dip,” yet the Committee decided at the time to define two separate recessions  (See A ‘Double-dip’ Recession Defined).

Another reason that I question the current danger of a double-dip recession is that most economic indicators continue to show a moderate rate of recovery in the economy.  The buzz about a double dip recession seems to have intensified following the release of the June employment report, which was somewhat weaker than expected.  It is never wise to put too much emphasis on one month’s data, and in this case it can be misleading.  As shown in the chart below, private-sector employment has shown increases in each of the past six months. 

Source: Bureau of Labor Statistics.

Total employment has been subject to fluctuations in Government employment due to the hiring of temporary workers by the Census Bureau.  When the report on May employment came out, many media sources pointed out that the strong monthly showing was entirely due to Census hiring, with private sector employment falling.  Subsequent revisions to the data now show that private sector employment did increase in May, albeit only slightly.

One of the reasons that the June employment report looked so discouraging was that the Census Bureau laid off many of the workers it had hired in earlier months, so that total employment declined, on net.  Yet private sector hiring showed its sixth consecutive month of increase.

Evidence from previous recessions–particularly the two most recent recessions–shows that employment tends to be a lagging indicator, picking up only slowly during the early stages of an economic expansion.  So far, we’ve seen total growth of about 900,000 jobs during 2010.  This is not consistent with a robust rebound, but it does indicate a clear sign of a turnaround in the job market. 

Meanwhile, the other key indicators that are considered by the Business Cycle Dating Committee–industrial production, real income growth, and total sales–are all showing patterns of renewed strength since the middle of 2009.  The economy is recovering, but it’s looking like this recovery will follow a pattern similar to the last two “jobless recoveries.”

Source: Bureau of Labor Statistics

Nonfarm payroll employment report also contained positive news:  the number of jobs was up by 2800 in May (seasonally adjusted)*. 

Source: Bureau of Labor Statistics

Gains were particularly notable in some of Arkansas largest sectors:  Manufacturing employment was up by 900; Trade, Transportation and Utilities was up by 2100; and Education and Health Services was up by 900.

Source: Bureau of Labor Statistics

Government employment was also up sharply +4200 jobs.  Mirroring the pattern seen in the national data for May, many of these jobs were related to temporary hiring by the Census Bureau.  Federal Government employment in Arkansas was up by 4600 in May.

Since December of 2009, employment is up by 6300 jobs (including 6000 in manufacturing).  4000 of those jobs are government-related, with an increase in private-sector employment of 2300.

#  #  #