Increasing hours worked versus increasing hiring

The current recovery has been characterized by increasing production and sales without an associated expansion in employment. Part of the explanation for the lack of hiring has to do with increased productivity of workers (output per hour worked)—either by improved production methods or simply requiring more effort from staff per hour worked. Another reason why firms have been relatively slow to hire is that, in addition to slashing payrolls during the recession, many firms also cut the work hours of the remaining staff to levels well below prerecessionary norms. As a result, these firms have some scope to increase the hours worked by their current staff before hiring additional workers. This fact is evident in the often-cited increase during the recession in the number of people working part time for economic reasons (see here and here, for example). That number has remained relatively stable at around nine million people over the last year, but it is still more than twice its prerecessionary average.

Another perspective on the part-time issue can be gleaned from data on average work week obtained from the U.S. Bureau of Labor Statistics (BLS) Current Population Survey. Chart 1 shows the pattern of average weekly hours (not seasonally adjusted) for all nonfarm wage and salary workers during the period of January 2008 through May 2010. For ease of comparison, the chart is scaled to be relative to the 2002–07 average. Compared with prerecession levels, average hours worked declined during the recession although they really didn't begin falling until the second half of 2008. As of May 2010, average hours worked were still about 1.5 percent below the prerecession average but have been trending higher in recent months. (Note that the sharp drop in September 2009 is a quirk of Labor Day falling during the survey week and hence cutting the work week one day shorter than usual.) The fact that average hours worked has moved higher is an encouraging sign for employment growth going forward if the historical norm is any guide. Of course, a firm may need to hire new workers even when hours per worker are below average. For example, the decision to start an additional manufacturing production line will probably require hiring new staff even if existing staff on other lines are working fewer hours than usual.

(enlarge)

The aggregate picture in Chart 1 masks considerable variation across industries. For example, Chart 2 shows the normalized average weekly hours reported by workers in the education and health services industries and in the financial industry. For these workers, although average hours worked per week declined mildly during the second half of 2009 weekly hours worked have since returned to prerecessionary levels. This performance suggests that, other things equal, additional demand for hours of work in these industries is likely to be met by additional hiring.

(enlarge)

Chart 3 shows the evolution of average weekly hours reported by workers in the manufacturing and transportation/warehouse industries. In these industries, average hours worked began to decline in the fall of 2008, but they have recovered much of the decline in recent months and are now about 1 percent below their prerecession averages. As with the aggregate picture, the fact that average hours worked has been trending higher recently is encouraging news for future employment growth in these industries.

(enlarge)

In contrast, Chart 4 shows the pattern of average hours reported by workers in the construction industry, the wholesale and retail trade industry, and the leisure and hospitality industry. For these workers, average weekly hours started to decline in the fall of 2008 and have shown no clear signs of recovery—still sitting some 3 percent to 4 percent below their prerecession averages and not trending higher. Thus, there appears to be more scope for firms in these industries to increase hours without necessarily having to hire additional workers.

(enlarge)

This analysis does have some caveats. For one thing, it is based on worker-reported data about hours worked drawn from the BLS's Current Population Survey. An alternative would be the employer-reported measurements in the BLS's Establishment Survey.

Probably more importantly, this analysis uses the prerecession history as a guide to what is "normal." If, for example, firms decide to keep average weekly hours lower by increasing the use of part-time workers, then the fact that average hours are below prerecession levels does not imply that firms won't hire when demand increases. Some industries already make heavy use of part-time employment. For example, in May 2010 the reported average weekly hours by workers in the leisure and hospitality industry was 33.3 hours compared to 42 hours in manufacturing. Absent an offsetting increase in wage rates, a permanent shift toward increased part-time employment would lower a worker's income relative to full-time employment and probably result in an increased propensity for multiple job-holding by individuals and households.

By Amy Ellingson, economic analyst at the Atlanta Fed

Another view of the structural versus cyclical unemployment question

One of the key functions of labor markets is matching firms looking for workers who have particular attributes (or skills) with individuals looking for work who have those attributes. What economists have been worrying a lot about recently is the potential for a substantive mismatch between the skills of those looking for work and the skills that firms want. This type of labor reallocation friction is one of many potential structural problems affecting the U.S. labor market at present (see, for example, here, here, and here).

A 2003 New York Fed article by economists Erica Groshen and Simon Potter examined the issue of structural rigidities in labor markets during the recovery from the 2001 recession. Their idea was to identify the share of employment in industries that had either continued to lose or gain jobs on net after the recession versus the share of employment in industries that had responded cyclically (gaining jobs after having lost them during the recession or losing jobs after gaining them during the recession) to the recession. The New York Fed researchers used industry of employment as a proxy for industry-specific skills, though it's not a perfect measure. For example, the skills of construction workers are generally different from the skills of health care workers. The more often that employment is accounted for by industries that are continuing to gain or lose employees, the more the potential exists for skill mismatch going forward.

Using the first 12 months of the recovery as a basis, Groshen and Potter found that in the 1974–75 recession and the recessions of the early 1980s the share of employment in industries continuing recession employment trends was around 50 percent. That share increased to 57 percent for the 1990–91 recession and rose sharply to 79 percent for the 2001 recession. The researchers took these findings as evidence of structural change playing a more significant role in influencing the labor market recovery from the 2001 recession than earlier recessions saw.

Visually, this observation can be presented as a four-quadrant "bubble chart" that measures job growth during the recession on the horizontal axis and job growth in the first 12 months of recovery on the vertical axis (the size of the each bubble reflects the relative employment size of the industry). We replicated Groshen and Potter's work with minor data definitional changes and find that for the first 12 months of recovery from the 2001 recession 81 percent of employment was in industries continuing recession employment trends (the top right and bottom left quadrants in the chart).

(enlarge)

Using the same approach as Groshen and Potter, how does the 2001 recession compare with the most recent recession? To make that determination, we used data available from the 11 months of recovery coming out of the most recent recession (assuming the recession ended in June 2009). We calculate that 65 percent of employment is in industries either still losing or gaining jobs. This share is less dramatic than the 2001 experience but a bit more than the 1990–91 experience.

(enlarge)

The positioning of certain industries within the four quadrants is not too surprising given the nature of the most recent recession. For instance, construction and related industries are deep in the continued job-loss quadrant. In contrast, the temporary help sector has behaved procyclically. Jobs in federal government and health care have continued to grow, with the former boosted by temporary hiring of census workers. Of the 79 industries examined, about a third of them have landed in a different quadrant compared with the 2001 recession.

Of particular interest is the share of employment in industries that are continuing to lose jobs. For unemployed workers from those industries, there is less prospect of being reemployed in that industry and hence a greater chance that skill mismatch will be an issue for those workers. Interestingly, the share of employment in industries experiencing continued net losses is similar to that seen during the 2001 recession (45 percent versus 41 percent).

This most recent recession was especially deep, and the large share of unemployed workers reporting they were permanently separated from their employers suggests that many of those jobs in all likelihood will not come back. If new jobs come with different skill requirements, then skill mismatch could become a significant factor once labor demand increases. However, the relatively disappointing May private-sector payroll jobs numbers released last Friday and the improving but low level of job openings reported in the JOLTS data for April are reminders that weak labor demand is still the dominant factor inhibiting the overall employment recovery.

By Menbere Shiferaw, senior economic research analyst, and John Robertson, vice president and senior economist, both in the Atlanta Fed's research department

The recovery: Job rich or job poor?

Gross domestic product (GDP) is growing, supported by strong labor productivity numbers and modest employment growth. Even after today's downward revision to first quarter labor productivity measures, growth in labor productivity in the business sector has averaged 5.6 percent over the last three quarters—more than twice the long-run average. Okun's law has underpredicted the rise in unemployment, and some commentators call for a jobless recovery. Today, Federal Reserve Bank of Atlanta President Dennis Lockhart gave a speech on the prospects for the future of employment and labor productivity that posed the question:

"… how long can firms ride this productivity growth before having to yield to new hiring to support greater activity?"

To provide some context, below is a chart that shows a decomposition of GDP growth during post-WWII recessions. This decomposition does not explain growth, but it does show the relative contribution of various factors such as average labor productivity, average hours worked, the unemployment rate, and other labor market variables.

(enlarge)

Note: The decomposition is based on the identity.

Where Y is GDP, H is total hours worked from nonfarm payrolls, E* is employment from the nonfarm payroll survey, E is employment based on the household survey, L is labor force, and N is working age population.

As this chart shows, relatively high labor productivity growth during a recession is not a phenomenon isolated to the 2007–09 and 2001 recessions (for present purposes, the end of the most recent recession is identified with the trough in GDP in the second quarter of 2009). All recessions from WWII through 1970 also featured sizable growth in labor productivity. Notice also that these recessions experienced large declines in the employment rate (rise in the unemployment rate). Hence, the productivity gains were attributable to outsized reductions in hours rather than gains in output. The same also holds true of the 2007–09 recession. As President Lockhart noted:

"Many employers reacted to the downturn by aggressively cutting their workforces, reorganizing remaining workers, and cutting other costs."

Of course, the big question centers on the recovery. Specifically, does the outsized contribution from labor productivity during the last recession portend continued outsized gains (and hence subdued employment prospects) during the recovery and expansion as well?

Perhaps history can provide a guide. The next chart provides the same GDP decomposition, but this time for the first three quarters of the recovery (we chose three quarters to allow a common comparison across recoveries). This chart reflects at least two salient features. First, the fastest way to employment growth is faster GDP growth—recessions that are immediately followed by very strong GDP growth also tend to have strong employment growth. Second, productivity remains an important contributor to growth, and that contribution has been especially large during the early phases of the past three recessions.

(enlarge)

President Lockhart's interpretation of these observations is that many employers:

"… have reacted to the upswing by holding employment at or near recession levels, seeking efficiencies in supply chains, investing in labor-saving automation, and generally tweaking their business models to operate more efficiently than before the recession. We've heard this story frequently in anecdotal accounts of our directors and business contacts across the Southeast.

"As long as efficiency and productivity gains can be achieved in this way, employers may remain hesitant to hire."

Going forward, President Lockhart concludes that:

"I do not expect the recent outsized productivity growth to continue indefinitely and become a new, permanently higher trend rate. Some degree of 'wait and see' behavior is at work and is no doubt reflected in the productivity numbers. With growing economic momentum, deferral of hiring will become riskier.

"Some employment gains should result as labor productivity levels out and falls back over time to something resembling the historic trend rate. But the pace of hiring is likely to be gradual. Current data on the use of part-time workers suggest that businesses have some scope to increase hours without hiring new full-time employees. And there are other, more structural obstacles to the rapid reemployment of the jobless. Some jobs in the construction sector and certain manufacturing industries are likely permanently lost, requiring some amount of migration of workers to other sectors. And, for a time, skill and geographic mismatches may frustrate employers willing to hire.

"Also, the weight of uncertainty about the future business environment makes a gradual pace of employment progress a reasonable assumption. I hear often from members of the business community that uncertainty regarding federal, state and local fiscal fundamentals and regulatory rules-of-the-game are feeding reticence to pull the trigger on new ventures, new hires and new investments. The recent European sovereign debt and banking pressures have added to uncertainty in financial markets.

"Sizing all this up, I expect recovery in the medium term to be neither jobless nor job rich."

The current Blue Chip consensus for second quarter GDP is about a 3.2 percent annual rate. Payroll employment expanded by an estimated 290,000 in April, and with an employment report out Friday that will likely feature very sizable gains (some forecasts calling for an addition of more than 500,000 jobs), the relative contribution of labor productivity to GDP growth in the second quarter is almost certain to decline sharply. However, most of the surge in May's number will be temporary U.S. Census workers, and this increase will reverse itself in pretty short order. Hence, it will probably take awhile to see how President Lockhart's forecast of continued modest employment growth pans out.

By John Robertson, vice president, and Pedro Silos, research economist and assistant policy adviser, at the Atlanta Fed

The inventory question

Mark Thoma asks a very good question:

"I hadn't looked at this for awhile—should I interpret the return of the inventory-sales ratio to near normal levels as good news?"

Here's the picture Thoma was looking at, updated to incorporate today's U.S. Census Bureau release on February manufacturing and trade inventory and sales:

(enlarge)

Tim Duy has taken a look at these data and comes to this conclusion:

"Increasingly, the recovery looks sustainable—sustainable in the sense that a double dip recession looks unlikely. As Bloomberg reports, this is the message of the inventory cycle, which appears to have largely run its course. Inventories surged as the recession intensified, leaving firms scrambling to bring output in line with the new level of sales. Now, firms have inventories under control."

I have been pondering those data as well, ever since the advance fourth quarter gross domestic product report indicated that 3.4 percentage points of the then-reported 5.9 percent annualized growth rate was accounted for by a slowing in the pace of inventory decumulation. (The numbers have subsequently been revised to 3.8 percentage points of a 5.6 percent growth rate.) It certainly appears that inventory-sales ratios have reverted to the prerecession norm, justifying Duy's sense that inventories will not be a big part of the economic story as we move through 2010.

That conclusion does rest, of course, on the likelihood that a downward trend in the ratio truly did break in the middle part of the decade. As the chart shows, the same pause in the trend occurred in the mid-1990s, only to commence its southward trek on the other side of the 2001 recession.

But the situation is even more curious than that. If you dig a little deeper, you find that not all inventory-sales ratios tell the same story. In particular, inventory-to-sales ratios at the retail level look very lean relative to prerecession levels while manufacturer's inventories still appear to be relatively bloated.

(enlarge)

What, exactly, is that chart trying to tell us? Does it represent some shift in supply-chain management, with inventory holdings being pushed down from the retail level to manufacturers? If not, can we expect some resurgence in retail inventories (as the Duy-cited Bloomberg article suggests), coupled with continued decumulation at the manufacturing level? And what would be the net effect of such developments on aggregate inventory levels?

Those are good questions, too. If you have any insights, I'd love to hear them.

By Dave Altig, senior vice president and research director at the Atlanta Fed

Bad by any measure

A few weeks back, The Economist published a story touting the well-known fact that the "American economy just had its worst decade since the 1930s." Whether looking at gross domestic product (GDP), consumption, income, or nonfarm payrolls, the decade from 2000 to 2010 generally looks bad from an economic perspective. During this decade, of course, the nation has experienced two recessions—the latter being the most severe since the Great Depression. (See the graphs below, reproduced from the February 25 article in The Economist titled "Back to The Crash.")

(enlarge)

But given that decades are rather arbitrary economic demarcations, why not examine other time periods? So here's another approach: Using the yearly trough-to-trough periods according to National Bureau of Economic Research (NBER) recession dating, the charts below have replicated the ones shown above from The Economist. (Note: The NBER only designates troughs by quarter. So for a trough ending in first or second quarter, the calendar year in which the trough falls is designated as the "trough year." If the trough falls in third or fourth quarter, the following calendar year is the trough year. For these calculations, we use annual data instead of quarterly because pre-1947 quarterly data are unavailable.) This simple exercise sheds some interesting light on the recent experience of the U.S. economy—namely, that it was bad by any measure.

Looking at real GDP growth over these periods, the 2002–09 era looks very weak, with only 1946–49 having a lower average annual rate of growth (in these years, GDP averaged an annual decline of 2.01 percent). Average annual real GDP growth was 1.72 percent for the 2002–09 period, much lower than the average of 3.97 percent for the previous 10 trough-to-trough periods.

(enlarge)

Similarly for real consumption and income growth, the 2002–09 period is also bleak. Average annual consumption and income growth had averaged 3.81 percent and 3.79 percent, respectively, going into 2002. But during this recent trough-to-trough period, income growth was very weak at 1 percent, with only the 1946–49 period doing worse (–1.09 percent). But consumption growth in 2002–09 was the lowest on record, averaging only 2.12 percent growth annually.

Another interesting observation is the spread between average annual consumption and income growth. The 1946–49 and 2002–09 periods are where it's the largest, at 5.9 percent and 1.1 percent, respectively. These large imbalances could possibly reflect growth in household debt and/or lower saving rates, as consumption growth far outstrips income growth. Indeed, debt grew and savings declined notably during 2002–09.

(enlarge)

Lastly, a look at the nonfarm payroll growth confirms the most recent trough-to-trough period as one of extraordinary weakness. Given that data prior to 1939 are unavailable, the previous eight bottom-to-bottom periods saw average annual growth of 13.5 percent in payrolls. But for 2002–09, average annual payroll growth of 0.44 percent reaffirms the so-called "jobless recovery" from the 2001 recession and the large decline in payrolls during this current recession. No other previous period comes close.

(enlarge)

By Andrew Flowers, economic research analyst, at the Atlanta Fed

A look at the income-side estimates of growth

Last week, a post in the New York Times' Freakonomics blog on Okun's law made note of the statistical discrepancy between the two methods for calculating national output:

"…there are two measures of output growth—the usual measure, which adds up total spending in the economy, and the alternative, which adds up total income. In theory, the two should be exactly the same. In practice, they have been very different during this recession… These GDI [gross domestic income] numbers suggest that output growth actually declined much more sharply than had been widely understood."

Indeed, the recession looks deeper and the recovery seems much less pronounced, looking at the income-side data in this chart.

(enlarge)

There has been a good deal of coverage about the discrepancy between the income and expenditure sides of gross domestic product (GDP) calculations in the past couple of years. Jeremy Nalewaik, at the Federal Reserve Board, is often cited for his work arguing that GDI may be a more reliable measure for delineating recessions than GDP (see 1, 2, 3). In fall 2008, Jim Hamilton noted the relatively weak behavior of GDI toward the beginning of the current recession: "It is interesting that while GDP indicates sluggish growth over the last three quarters, GDI looks much more like a recession, with 2007:Q4–2008:Q1 satisfying the traditional rule of thumb of two quarters of falling real output."

But apart from recession dating, how seriously should we take these income-side numbers?

One issue with using GDI data is that they lag the GDP data by a full quarter. That stated, a 2006 study by Fixler and Grimm at the Bureau of Economic Analysis (BEA) argues that GDI data contains valuable information.

"There is evidence that income-side measures contain information about revisions to estimates of GDP. National income is statistically significant in explaining revisions from the final current quarterly to the latest estimates of GDP. Conversely, there is no evidence that product-side measures contain information about revisions to GDI and national income."

In other words, history suggests that when these two measures of national output disagree, GDP tends to get revised in the direction of GDI and not the other way around. So, if this relationship holds, it would be prudent not to dismiss the latest divergence in the two measures because it suggests that the decline in national output has been more protracted, and the recovery (through the third quarter 2009) more modest, than what is being reflected in GDP.

If true, this pattern could raise questions about current levels of productivity and associated labor cost measures. The decline in unit labor costs over the recession has been remarkable by either measure. But the recent drop in unit labor costs by way of the expenditure-side estimate was roughly 1.5 percentage points larger than the labor cost estimate that would be computed from the income side of the accounts. If revisions going forward continue to favor the income-side estimates, then maybe downward wage pressure—while probably still large—may be less than many believe.

(enlarge)

By Laurel Graefe, senior economic research analyst, and Jacob Smith, quantitative research analysis specialist, both at the Atlanta Fed

Competing histories

If your economic forecast for the coming year embeds something like robust growth in consumer spending, last Friday's Federal Reserve report on consumer credit should probably give you pause.

(enlarge)

At least some folks look at that picture and see a slow slog ahead. Calculated Risk sums up the concern:

"Consumer credit has declined for a record 11 straight months—and declined for 14 of the last 15 months and is now 4.8% below the peak in July 2008. It is difficult to get a robust recovery without an expansion of consumer credit—unless the recovery is built on business investment and exports (seems unlikely to be robust)."

At Angry Bear, the question is a little more pointed:

"Remind me again why all those banks were 'bailed out?' Wasn't it supposed to be to kick-start the economy again?"

Well, here's the thing. That consumer credit picture embeds both the supply of credit and the demand for credit. Though both tighter credit standards and weak loan demand are certainly at play, it is does seem that, at the moment, weak demand is the factor most responsible for slow loan growth in the United States. Recall, for example, this information from the Federal Reserve's January Senior Loan Officer Survey:

"The January survey indicated that commercial banks generally ceased tightening standards on many loan types in the fourth quarter of last year but have yet to unwind the considerable tightening that has occurred over the past two years. The net percentages of banks reporting tighter loan terms continued to trend lower. Banks reported that loan demand from both businesses and households weakened further, on net, over the survey period."

As regular readers of macroblog know, our own Atlanta Fed surveys (here and here) are indicating that soft customer demand, not credit access, is a significant story in business capital expenditure and expansion plans.

Of course, we don't really know whether credit availability will become a more significant problem when demand begins to recover. This uncertainty is behind what is the real back story at this critical point of the recovery. As we peer ahead, we essentially have two competing, and contradictory, economic histories as our guides. First, there is the statistical regularity that deep recessions in the United States have in the post-WWII period been reliably followed by rapid recoveries. But second, there is the Reinhart-Rogoff statistical regularity that recoveries from financial crises are slow and difficult.

A Wall Street Journal interview with Carmen Reinhart provides reasonable arguments as to why slow and painful is a sensible bet. On the other hand, one could argue that the advance fourth quarter gross domestic product figure is consistent with the sharp bounce-back scenario. (If you are looking for that argument, Brian Wesbury and Robert Stein oblige.)

One thing is certain. At least one history is going to be revised.

By Dave Altig, senior vice president and director of research at the Atlanta Fed

Southeast businesses offer insights on capital spending plans

With last week's capital orders data giving signals of renewed growth in business fixed investment in equipment and software in the fourth quarter, the question turns to whether this growth will be sustained. In early January 2010, the Federal Reserve Bank of Atlanta reached out to our contacts in the Southeast through our Regional Economic Information Network as part of our monetary policy information-gathering efforts to inquire about businesses' capital spending plans. We received responses from 320 businesses across Alabama, Florida, Georgia, Mississippi, Louisiana, and Tennessee. (I want to note we were helped significantly in this effort by colleagues at the Kennesaw State University Econometric Center, who conduct a monthly PMI survey of manufacturers in the Southeast.)

As with our recent small business finance survey (discussed in this macroblog post), readers should be cautious about the results because of the tendency, for example, to sample established, relatively successful firms. That said, we still believe the results are instructive. Of note, 36 percent of respondents indicated that they planned to increase spending over the next 6–12 months relative to actual spending over the past 6–12 months. Another 42 percent said they would leave their spending at about the same level (unchanged), and 22 percent indicated that their spending would fall. The difference between those planning to increase spending and those planning to decrease spending equals a net positive of 14 percent. Across industries, construction firms were the group least likely to increase spending, while retailers were the most optimistic group. Our manufacturing contacts and the "other industries" group of firms (firms across a myriad of industries such as transportation, healthcare, and business services) expressed intentions similar to the overall response (see the chart).

(Enlarge)

For those who planned to increase spending on new plant and equipment, the most commonly given reasons (respondents could select more than one reason) were that they expected growth in sales to be high (37 percent of those respondents), or they needed to replace information technology equipment (37 percent of those respondents). Also, 61 percent of those planning to increase spending indicated that at least some of that spending reflects investment that had been postponed because of the recession. Not surprisingly, for those who did not plan to increase spending, the most commonly cited reasons were the expectation of low growth in sales (cited by 47 percent of those respondents) and heightened economic uncertainty (cited by 39 percent of those respondents).

Interestingly, cost and availability of external financing were among the least frequently cited reasons for either increasing or not increasing capital spending (cited by 9 percent and 15 percent of respondents, respectively). This theme is consistent with the findings of our recent small business survey, as well as the trend in the National Federation of Independent Business (NFIB) survey of small businesses. According to the NFIB, "finance" was reported as the number one small business problem by only 4 percent of respondents in December 2009. The number one single factor was poor sales.

The NFIB survey also found that while plans to increase capital spending by small firms rose modestly in December 2009 to a net 18 percent, they remained near historic low levels—in December 2007, the net percentage stood at 29 percent and 17 percent in December 2008. This performance suggests that our finding of a net 14 percent of firms planning to increase rather than decrease spending on plant and equipment should not be read too encouragingly.

So where does that leave things? Probably with more questions than answers. For instance, the fact that about two-thirds of the firms that are planning on increasing spending are doing so because they had postponed capital expenditures during the recession would be consistent with some bounce in capital spending by businesses following the most recent recession. But how sensitive are firms to changes in economic conditions? Currently, we hear a lot anecdotally that cash is a high priority on firms' balance sheets as a precaution against economic uncertainty. If sales were to increase more than expected, how fast would firms rethink their investment spending plans? The answers to these types of questions are important, and we are consequently planning to conduct a follow-up survey in due course. As always, we'll keep you posted.

By John Robertson, vice president in the Atlanta Fed's research department

The demand and supply of bank credit: A small business snapshot from the Southeast

In his recent speech, Federal Reserve Bank of Atlanta President Dennis Lockhart highlighted concerns about the linkage between commercial real estate loan problems at banks and small business financing during the economic recovery:

"The overall commercial real estate debt in the financial system is smaller than residential, but it is disproportionately concentrated in small and regional banks. Smaller banks are a significant source of credit for small businesses, and in most recoveries we look to small businesses to generate a significant number of jobs."

President Lockhart also referenced the results of a survey of small business finances the Atlanta Fed conducted late last year.

"A recent small business survey performed by the Atlanta Fed suggested that business loan demand was down primarily because of weak sales and modest revenue prospects. The credit availability picture was mixed. No surprise, construction-related firms and manufacturers had the most trouble obtaining credit during the last six months. But others did well in having their credit needs met. Of more than 200 respondents, nearly half did not look for credit at all, mostly citing weak sales or sufficient cash reserves."

The survey President Lockhart was referencing was conducted in early December and included responses from 206 small businesses across the Sixth Federal Reserve District (the states of Alabama, Florida, Georgia, Louisiana, Mississippi, and Tennessee) regarding their access to credit. The intent of the survey was to include some additional small business perspectives to supplement our other monetary policy information-gathering efforts.

The firms in the survey were contacts established through our Regional Economic Information Network. In that sense, the survey is not based on a pure random sample of firms. However, the industry distribution of respondent businesses was reasonably representative of the industry mix of the Sixth District (see the chart). The average firm size in the survey was about 22 employees, with around 40 percent of respondents having between one and nine employees.

So, how did businesses surveyed respond? Slightly more than half the respondents said that they had sought to obtain a loan or line of credit from a bank in the last six months. The primary reasons given by those seeking credit were to replace an existing loan (cited by 50 percent of those respondents) and/or to obtain additional working capital (cited by 45 percent of those respondents).

The degree of difficulty firms felt they had in obtaining credit was mixed, with about 60 percent of respondents saying they were able to obtain all or most of the bank credit they sought. The small size of the survey (206 respondents) limits the accuracy of any sector-by-sector comparisons. However, it is interesting to note that construction firms stood out as the business type that had the greatest difficulty having their demand for financing satisfied, with 70 percent of them saying they were unable to obtain the funding they sought. That percentage compares with 50 percent of small manufacturers surveyed and 25 percent of retailers responding they were unable to obtain the funding they desired.

Of those businesses that had not sought credit during the last six months, the dominant reason given was poor sales/revenue (cited by 55 percent of those respondents). Other reasons for not seeking additional credit included sufficient cash reserves.

Slightly less than half of respondents expected to try to obtain a loan or line of credit from a bank during the next six months. The reasons given for seeking credit (businesses could give more than one reason) included the need to replace an existing loan (cited by 43 percent of those respondents), the need for additional working capital (cited by 44 percent of those respondents), and the need to purchase equipment (cited by 21 percent of respondents). Among firm types, construction firms anticipated a higher demand for credit than others.

For respondents who were not expecting to seek credit over the next six months, the anticipation of poor sales growth was the most frequently cited reason (cited by 49 percent of those respondents).

There are plenty of caveats that should be applied to these results. For example, the survey respondents represent established, relatively successful firms. We could not, with this effort, capture the experience of firms that have recently failed (perhaps for lack of credit). Nor can we ascertain the businesses that were never formed because they could not obtain start-up funding.

Still, we believe the results of our survey are instructive. To the extent that the firms in our survey are representative, it appears most going concerns have been able to obtain all or most of the credit they need. What they don't have are customers.

Of course, this is a snapshot of current conditions, and things may change as the economy picks up, demand expands, and credit needs grow. And it would be very useful to know what the story is with those firms that have failed or were never created. We are consequently planning to conduct a follow-up survey as 2010 progresses. We'll keep you posted.

By John Robertson, vice president in the Atlanta Fed's research department

Jobs and the potential commercial real estate problem: Still keeping us up at night

In the season of good cheer, it is certainly gratifying to know that some in the economic forecasting business are actually feeling cheerier:

Reaffirming last month's call that the Great Recession is over, NABE [National Association for Business Economics] panelists have marked up their predictions for economic growth in 2010 and expect performance to exceed its long-term trend. "While the recovery has been jobless so far, that should soon change. Within the next few months, companies should be adding instead of cutting jobs," said NABE President Lynn Reaser, chief economist at Point Loma Nazarene University.

While we at the Atlanta Fed agree that the recession has likely ended, we wish we could feel as optimistic about the current jobs outlook. We've catalogued those concerns before—here, for example—but we continue to look for reasons to believe that our pessimism is unwarranted.

As was noted in a recent speech by Federal Reserve Chairman Ben Bernanke, weak bank lending remains one potentially significant headwind impeding the jobs recovery:

"… reduced bank lending may well slow the recovery by damping consumer spending, especially on durable goods, and by restricting the ability of some firms to finance their operations."

Among the factors restricting lending, "… with loan losses still high and difficult to predict in the current environment, and with further uncertainty attending how regulatory capital standards may change, banks are being especially conservative in taking on more risk," Chairman Bernanke said.

One area where bank loan losses are potentially high and uncertain is commercial real estate (CRE). As highlighted in a macroblog post from October, if the CRE problem falls disproportionately on financial institutions that also finance small business activity, we will be all the more worried that "the post-recession employment boost [small] firms typically provide may be less robust than in previous recoveries."

In fact, as Atlanta Fed President Lockhart noted in a speech last month, as of mid-2009 the banks with high exposure to CRE (relative to tier 1 capital) accounted for about 40 percent of commercial and industrial (C&I) loans to small businesses.

Underneath that statistic are a couple of additional facts that also have our attention:

1. Over time, CRE loans have become increasingly concentrated in those banks whose CRE lending activity is high relative to their available capital. As of June 2009, banks with CRE loan books more than three times their Tier 1 capital level accounted for 52 percent of the $1.6 trillion of CRE loans in bank portfolios. Though this is lower than the 2008 peak of 59 percent, it compares to just 17 percent in 1993.

2. Small businesses that rely on bank loans for credit are much more likely to be affected by a bank's CRE exposure than in the past. In 1993, banks with CRE loan books more than three times their Tier 1 capital accounted for just 11 percent of total small business C&I loans. But this share increased to 42 percent in 2008 and stood at 38 percent in June 2009 (of a total of $281 billion of C&I loans to small businesses).

The following chart summarizes these two observations.

Thus, both commercial real estate loans and small business C&I loans are much more concentrated in banks with relatively lower levels of capital than has been the case in the past. Combined with our previous observation that a relatively high fraction of small business loans sit in banks with significant exposures to commercial real estate, these facts do not strike us as a case for optimism regarding the near-term outlook for growth in small business borrowing.

Perhaps today's job summit will result in additional ideas to counter what we see as a serious drag on job creation in the near term. And, of course, tomorrow's employment report could show signs of improvement in labor markets. That would be good news.

By David Altig, senior vice president and research director, and John Robertson, vice president, both in the Atlanta Fed's research department