The non-farm payrolls survey is released by the Bureau of Labor Statistics on the first Friday of each month. The data is eagerly appreciated by the markets and economists due to its great value in establishing the current unemployment situation, and also because of the forward-looking nature of some of its sections. Apart from private individuals and academics, the non-farm payrolls data is also closely scrutinized by the Federal Reserve as part of the process leading to the interest rate decisions. Due to the dual mandate of the institution, the Federal Reserve often chooses to reduce its main rate in case that the non-farm payroll data is showing a labor market under pressure, with obvious consequences for the forex, bond, and stock markets.
Among the major national economies in the world, the U.S. economy presents a rare picture due to the crucial role of consumers. In Japan, the export sector is paramount to economic growth. In the European Union, the manufacturing industries are central to economic performance. In developing economies as well, the manufacturing sector is of great importance in determining economic trends. In the U.S., on the other hand, the consumer accounts for more than two-thirds of economic activity due to the general propensity to borrow, and also because of the high-labor costs incurred by manufacturers in the American economy, and the popularity of outsourcing. Throughout the years of this decade, for instance, the manufacturing sector has been losing jobs even as the overall economy has been creating numerous new positions. Thus, the health of the American consumer is extremely important for growth, and the both the central bank and the financial markets attach great value to labor statistics as a result.
The value of the non-farm payrolls survey is born of its timeliness, scope, and depth. The report covers the entire U.S. economy, with the exception of farming sector jobs which form a tiny fraction of the overall labor market. Unlike the other surveys like the PMIs the sample size is large, and in contrast to the weekly jobless claims data, the non-farm payrolls survey is released only twelve times each year, with less volatility, and reliability, although it does have its disadvantages.
If the American economy is creating jobs at a healthy pace, labor costs may rise for firms, which may be reflected in higher wages, and then if the firms pass these wages onto prices, inflation will result. A fully-employed economy also consumes a lot more goods, putting pressure on the price of raw materials that are in short supply, granting producers greater control over pricing. In time, producers will increase the supply of goods and services to meet demand better, but in the meantime, it is up to the central bank to fight inflationary pressures by raising rates, and contracting credit. The role of the labor market in determining and clarifying these pressures is clear, and as such, authorities keep a close eye on these statistics, and in time react in a predictable fashion to labor market trends, creating opportunities for traders who are up-to-date with this release.
Now let’s examine the various pieces of data in this release.
The establishment survey is one of the two main components of the non-farm payrolls release. This survey is conducted by questioning some 400,000 firms across the United States. The actual numerical change in the number of the unemployed (the headline value) is derived from this survey.
The first part of the report gives the net value of job losses or gains across sectors in the United States economy, along with an average of monthly changes, and then compares these values to the last year’s numbers. These numbers are useful for traders for a number of reasons. Apart from the obvious value of the statistics for guidance on the present state of the economy, we also get an idea on how well the economy is performing compared with the same period of the previous year. If there were no natural disasters or similar events affecting production and labor demand at that time, this kind of comparison can be very useful in determining the health of the labor market, as it eliminates seasonal factors. This section also provides a net total of job gains or losses since the beginning of the present phase of the cycle (that is, a growth phase, or a recession).
The establishment survey then continues with a discussion of the employment changes in the various sectors of the economy. The report breaks the unemployment data into the following sectors.
Construction: The vigor of the construction sector reflects the overall strength of the economy. Since much fixed investment comes in the way of construction, and as investment reflects the confidence and expectations of managers and business owners, if the construction sector is adding jobs, it is likely that the overall economy will continue to grow in the medium term.
Manufacturing: This category states employment changes in the various U.S. goods producing industries. This part is then further broken down into motor vehicles and parts, machinery, mining, fabricated metal products and others.
The manufacturing payrolls are important since they are some of the worst hit sectors in a recession. The history of the manufacturing sector is older than that of the services sector, so it is easier to make comparisons with historical data, and consequently, to reach conclusions about the future direction of the economy in light of this analysis.
Retail Trade: One of the most important categories in the overall labor market survey is retail trade. This and the following categories belong to the services sector, which is the largest section of the American economy. Changes in this category have important implications for overall economic activity, consumer spending and investment across sectors.
Retail trade is the business category that caters directly to the consumer. While most of the other sectors work for firms, professionals, or act as intermediaries between various businesses, the retail trade sector is the distribution arm of the economy, passing goods and services to the end-user. Not only do products like computers, automobiles, and groceries, and cosmetics, and myriad other goods, but also the countless different kinds of services are also distributed to the consumers by businesses in the retail trade segment.
Generally businesses are late to begin discharging and hiring workeers in response to economic trend changes, and as a result, the behavior of the consumer adapts slower to economic changes, which is reflected in the belated reaction of the retail sector to economic developments. But although this category lags trend shifts, its hiring or layoff policies are crucial for the health of the economy, due to its huge size and significance in the overall scheme of events.
Sometimes recessions pass without causing major shifts in the breakdown of labor in this category. In those cases, the recession is likely to be mild, and recovery easier. In other cases, the retail segement itself undergoes large transformations due to chronic inflation, contracting of credit, rising unemployment, and other factors that affect long term consumer behavior. In such cases, the retail trade segment may adjust slowly, with more job losses, and a slower recovery.
This category states both the retail trade, and wholesale trade segments, providing comparison with past data, and information on the existing trends.
Professional and Business Services: Apart from stating the changes in professsional and business services categories, this item also provides information on the state of the temporary help business. This sector is a good forward indicator on the labor market, because in the uncertain period leading to a recession or recovery, businesses often prefer to hire temporary workers instead of opening permanent positions. Temporary help positions reflect the anticipations of businesses as well. If the manager anticipates that activity will be subdued, there is no need to keep temporary workers on payrolls. Thus, employment in this category will change faster than in the overall labor market, providing timely indications on the future dynamism of the economy.
Transportation and Warehousing: This item states the employment changes in transportation and warehousing jobs, such as the number of truckers, drivers, and the managers and workers at the various types of warehousing businesses. This item is a lagging indicator on the overall jobs market. However, the transportation industry is one of the earlier parts of the economy in feeling the effect of a recovery or recession.
Financial Services: Employment changes at banks, mutual funds, credit unions, financial publishing and telecommunications firms, and the many different kinds of financial services businesses are stated in this category. This item is usually a lagging indicator of the overall labor market, because the financial sector usually reacts to overall trends in the economy before making decisions about new employment. Of course, if it is the source of the problem, as it is now (2009), the situation may be reversed.
Health Care: Along with the education sector, this category is thought to be recession-proof, in that even in a recessionary economic environment people have to pay for health care and send their children to schools. Layoffs and hirings among doctors, nurses, and similar positions are summarized in this item. Following the heath care category, the report examines the employment status of the leisure and hospitality business which includes restaurants, movie theaters, hotels, and similar businesses.
Stock traders often use the non-farm payrolls survey as a way to gauge the health of the various sectors in the American economy. If a segment of the economy is shedding jobs at a high rate this is often thought to warn on present and future weakness. Naturally, hirings are thought to signal that a segment of the economy is doing well, justifying investment, while layoffs are thought to imply future weakness in profitability.
Bond traders mostly focus on the headline number and the unemployment rate. Since the Federal Reserve is expected to reduce rates in response to continued weakness in the labor market, the bond market is very sensitive to labor market statistics, often reacting with sharply to unexpected releases. The currency market often acts on cues provided by these two market segments. Apart from the role of forex as the source of fund flows into both of these markets, carry traders, and short-term traders like to capitalize on the short-term signals generated by this report for profit. Separately, fundamental traders like to focus on the long-term implication of the release, analyzing trends, and taking positions as justified by the data and the price action.
The establishment survey data is conducted by questioning a number of U.S. businesses in the private and public sectors. The household survey is conducted by questioning 60, 000 U.S. households according to statistical methods. In the long-term these two pieces of data converge on each other, but at times they can diverge significantly, making the derivation of trade signals somewhat harder.
The household survey states the number of unemployed in the economy, followed by the unemployment rate, and various demographical statistics. In general the establishment survey is more relevant to traders, and details such as the racial and sexual breakdown of the unemployed are of very little consequence for traders of all categories. Here we’ll limit our discussion to categories with significance for long-term fundamental analysis, or headline value.
The number of long-term unemployed, while often not a headline release, is still exceptionally important for analysis because it acts as a measure of the speed of turnover in the labor market. If layoffs and hirings are equally fast, for example, even a large number of unemployed can be easily absorbed into the payrolls of U.S. businesses over the long term. On the other hand, if the number of the long-term unemployed is rising, the implication is that the newly unemployed are joining a queue of workers who are actively looking for jobs over a long time, with no positive results. This situation is usually disinflationary, as it implies that there is a large pool of workers in the market ready to work as soon as jobs are available, keeping wage pressures under control.
Changes in the employment-population ratio item reflect a number of demographic and economic factors. One of those is the age breakdown of the population: as the nation grows older, the participation rate also falls. This ratio also reflects the economic situation of families. In case that a household perceives itself comfortable and secure financially, some members may leave the labor force in order to concentrate on domestic activites. Finally, as more and more workers become discouraged and stop looking for jobs, the employment-population ratio will fall.
The household survey also includes an item termed the marginally attached labor force stating the number of workers who are available for work, but had not looked for a job in the preceding four weeks. This category, with its various components, is not added to the total count of the unemployed, leading some to criticize the payrolls release for understating the actual number of the jobless.
The Establishment survey includes the headline total hirings or layoffs in the U.S. economy, including government hirings. The other headline in the release is the unemployment rate. The unemployment rate is an important gauge of recessions and recoveries, although it may lag both. It is calculated from data included in the household survey.
The unemployment rate can be recorded as a benchmark value. Usually an unemployment rate below five percent is regarded as a sign of a healthy economy. It is thought that there’s a lower limit of around two percent to the unemployment rate due to the natural flow of workers changing jobs because of their own demands, and without being compelled to do so by pressures in the labor market. An unemployment rate above 6 percent is usually a sign that the labor market is sluggish, and that the employers have the advantage in setting wages due to the slight scarcity of jobs. Finally, an unemployment rate above ten percent is uncommon, and coincides with significant difficulties for the overall economy.
Along with the headline job losses, the unemployment rate is a lagging indicator of overall activity. The rate will peak sometime into the recovery, and as such, it is difficult to use it for generating signals about the future health of the economy.
Although the unemployment rate is often secondary to the headline number, it can snatch the spotlight from the actual number of job losses or gains if it is a significant surprise. It is important to remember that the unemployment rate is derived from the household survey, while the headline number is acquired from the establishment survey. These two are completely different, and can diverge from each other at times, emitting conflicting signals, and causing market turmoil.
The establishment survey also states the changes in the average workweek at U.S. private non-farm businesses. At economic inflection points, business owners prefer to satisfy demand by adjusting the work week, and hiring temporary help instead of hiring new labor outright, since it is harder to discharge workers than adjusting these factors. As a result, the workweek at U.S. firms reflects the outlook on the U.S. economy as perceived by business owners and managers.
The work week also has a significant role in determining the potential future income of workers. In fact cutting down on the workweek is equivalent to laying off workers stealthily. Since the total number of work hours reduced by the employer is equivalent to the labor shift of a number of worker, we could consider such changes as another name for layoffs in any business sector.
This section of the report also provides data on factory overtime, and separate information on average workweek across manufacturing industries.
Average Hourly Earnings
One of the most important pieces of data in the payrolls release is the average hourly earnings of production and non-supervisory workers on private non-farm payrolls. This item states the changes in labor compensation across U.S. industries on an hourly basis, and along with the previous item, establishes the spending power of the consumer on a long term basis.
At this point, it is important to consider a simple formula
Total Unemployment Change = Total Change in (E x W)/ average weekly earnings + H
E = average hourly earnings
W= workweek in hours
H = Headline number
What does this formula mean? The total potential contribution of an employted worker to the economy is limited to his net earnings, which is determined by the multiplication workweek and hourly compensation. Thus, the real change in employment is better measured by the change in the earnings per person of U.S. workers, added or subtracted from the headline number, depending on the direction of the change. So if the headline number is 100000, and the average worker makes $20000 per week, while the total reduction in weekly compensation across U.S. firms is $ 200 million in the past month, the total effective reduction in payrolls would be 90000. These numbers are of course arbitrary and somewhat exaggerated, but our aim is only showing that the stealthy changes in weekly compensation equate to hidden layoffs for the U.S. labor market.
The non-farm payrolls release is useful in many ways, for many different types of traders. It is a good summary of labor market conditions, and it provides adequate data on the hiring policies of different firms across the economic sectors. As we mentioned before, the data has important implications for overall economic dynamism, the interest rate decisions of the Federal Reserve, and the inflation rate over the longer term. It is clear that this release is one of the most precious pieces of information available to traders and analysts.
But that is not the complete picture. In spite of its many advantages, a few negative aspects of the non-farm payrolls release demand that some caution and conservatism be applied while it is being evaluated. In this final section of our examination we’ll take a look at these disadvantages, and their implications in analysis and study.
When making seasonal adjustments to the data the BLS makes some assumptions that may lead to the overstating of job losses, or understating of gains depending on the phase of the cycle. The Bureau uses a statistical model for estimating the “birth and death” of U.S. businesses across the country and since the BLS cannot adjust it for recessions until the NBER has officially declared a recession, the actual number of job losses is often understated due to the inflated assumptions about new businesses being created in the middle of a downturn. The same factors also have a role in possible understatement of labor market recoveries as the economy emerges from contraction. As such, the seasonal adjustment factors have to be considered with a little bit of skepticism, and perahps keeping a closer eye on the raw numbers at these times is a good idea.
An important problem faced by traders while evaluating this extremely useful and important release is the frequency of revisions. Especially during periods when large seasonal changes take place in the labor market (such as during labor strikes, or in August, as teenagers leave their jobs to go back to school) large distortions in the data gathering process can lead to great errors which are later corrected through big revisions. Revisions reaching up to 100, 000 are not unusual, and sometimes releases that were reported as job gains can be revised to show losses across sectors.
For short term traders, this issue hardly matters; as they attempt to trade both the initial release and the revision for immediate gains, with varying levels of success. For long-term traders, however, it is a very important subject since it adds a large dose of uncertainty to analysis performed in the immediate aftermath of the release.
In general, then, a surprising number should not be regarded as a guideline for future trading decisions by long-term traders. The effects of changes in the labor market are felt over a very long term, along with the interest rate decisions of the Federal Reserve, and the lending policies of banks. As such, it is a good idea to wait until the data has been revised a couple of times, so that there is a sufficient degree of reliability in our conjectures about the future.
The Bureau of Labor Statistics consistently underestimates job gains at the early stages of a recovery, and job losses in the beginning stages of a recession. The result of this situation is that the trader should be somewhat more aggressive in his analysis of the labor market at these incipient phases of economic activity. This implies that when the economy is coming out a recession, it is wise to expect a faster and larger improvement in labor market health, and to invest in stocks, bonds, or currencies in accordance. And conversely, when the economy is heading into a recession, it is a good idea to be more bearish on the labor market than what is suggested by the BLS surveys.
Establishment survey is more reliable
Both the trader and analyst communities, as well as the Bureau of Labor Statistics itself regard the establishment survey as being more reliable than the household survey. This is mostly because of the larger sample size of the establishment survey; while the household survey questions about 60,000 households across the United States, the establishment survey has a sample size of 400,000, about seven times larger. Still, over the longer term these two values converge in each other, and the differences are usually temporary.
The main consequence for trading decisions is the preference for establishment survey data in the formulation of strategies. It is in general a better idea to attach greater significance to the statistics contained in the second section, including changes in the workweek, and the average hourly compensation. As such, the headline number of the hirings or layoffs is also more valuable than the unemployment rate declared. But although these are valid considerations for level-headed analysis of the release in the aftermath of trading, during the release period itself market participants will react to any piece of data with equal aggressiveness if it turns out to be a significant surprise.
Discrepancies between the sums of states’ data, and the non-farm payrolls release.
A short time after each month’s non-farm payrolls release, the BLS also releases the sum of states’ payrolls surveys, and the numbers may contradict each other strongly. A very large discrepancy would indicate that the data will see a major revision in some short time, and would invite caution in the decisions made on the basis of the unrevised release.
In sum, the BLS release provides a vast amount of data in a concise format, and includes the most important statistics relating to the labor market. For the labor market itself, the non-farm payrolls release is not a leading indicator. But for the economy at large, the performance of the labor market is crucial. Hence the great value attached to this piece of data by the markets, policy makers, and commentators.