What the employer's payroll survey tells us


Chih Kwan Chen

March 18, 2011

1. Introduction

A distinctive character of 2007 - 2009 recession induced by financial crises was the massive job loss. Since the fourth quarter of 2009 when the recession gave way to the recovery only a small fraction of the lost jobs has been regained. To understand the current economic situation and to peek into the future course of the economy, the reason and the consequence of this anemic job recovery must be confronted. Especially important is to explain the obvious contradiction between the sorry job market and the full recovery of consumer spending. In this article the problems are approached by analyzing one by one major sub categories of the whole job market.

Employer's payroll survey, published monthly by U S Bureau of Labor Statistics, is used exclusively as the source of data about the job markets in this article. An overview of the payroll survey is given in Section 2. Section 3 is devoted to the analysis of private service providing jobs. Goods producing sector is studied in Section 4 and government jobs in Section 5. The last section is reserved for some concluding remarks.

2. Overview

Employment payroll survey is based on the survey of 200,000 businesses supplemented by the estimate of workers of sole proprietors. The result is published monthly by US Bureau of Labor Statistics. This survey breaks down the payroll numbers to various sub fields. The major component that we will concentrate our attention on is the "total nonfarm payroll number", simply called "total nonfarm" here. "Total nonfarm" is further broken down to "private nonfarm" and "government". In the first graph posted at the right, "private nonfarm" is plotted as the blue curve, and "government" is plotted as the red curve. During the recent boom and bust cycle, "private nonfarm" reached the peak at January, 2008, and hit the bottom at March, 2010. From the peak to the bottom, "private nonfarm" sector had lost more than 8.8 million jobs. In the recovery phase, from March, 2010 to February, 2011, this sector has regained about 1.4 million jobs, a small fraction of the lost jobs. In contrast during the 2000 - 2001 recession, totally 3 million jobs were lost in the sector of "private nonfarm", but in the recovery phase from July, 2003 to June, 2004 1.6 million jobs were created, representing 53% of lost jobs. This shows how severe this recession has been. "Private nonfarm" payroll number is further broken down to "private service providing" sector and "goods producing" sector. The former will be discussed in the next section and the latter in Section 4.

The red "government" curve has two spikes, one at 2000 and the other at 2010. Those spikes are due to the hiring of temporary census taking workers. Detailed analysis of "government" sector payrolls is postponed until Section 5.

3. Private Service Producing Jobs

The blue curve in the second graph at the right represents monthly values of private service providing jobs excluding healthcare and education service. The red curve depicts the jobs in the healthcare and social assistance sector, and the green one is for the jobs in the education service category. The behavior of the red and the green curves are very different from the all important blue curve so we will discuss them separately. We will first discuss the red and the green curves, and reserve the remainder of this section for a rather detailed analysis of the blue curve.

The red healthcare and social assistance curve has risen steadily since 2000. The jobs in this sector is increasing about 2.9% per year in average. Further breakdown into sub categories reveals that the steady increase is rather across all the subfields. For example, physician's office sector has seen an average annual increase of 2.2% per year, hospitals +1.6% per year, social assistance +3.6% per year, and so on. The green education service curve covers all the jobs in various kinds of private schools. Though the green curve has more structures than the red healthcare curve, the green curve can also be classified as "steadily increasing" through the years; it has risen 3.2% per year. Again the steady rise in the green curve is manifest across most of its sub fields, like +3.1% per year in average for private colleges and universities, +4.0% for technical and trade schools, +1.4% for private secondary schools, +13.7% for education supports and so on. Since we are interested in the analysis of the recession and the recovery here, the steadily rising red and green curves will not be discussed further.

The blue private service less healthcare and education curve peaked at January, 2008 and bottomed out at January, 2010. During those recession years 5.3 million jobs were lost. Since January, 2010 bottom only 870,000 jobs has been recovered. This trend is not so different from the trend of private nonfarm jobs discussed in the previous section since private service providing jobs less healthcare and education is the largest contributor to private nonfarm jobs. The subfields within private service providing jobs less healthcare and eduction are retail jobs, wholesale jobs, transportation and warehousing, utilities, information, financial activities, professional and business services, leisure and hospitality and other services. Looking at those sub fields of the blue curve, it is apparent that the number of jobs in the private service providing sector less healthcare and education is closely related to the "Real Consumer Spending Less Healthcare and Education". Here "Real" means adjusted for inflation. "Real Consumer Spending Less Healthcare and Education" has already recovered almost to the pre-recession peak reached in December, 2007. Thus we are encountering a contradiction whether the anemic recovery of the private service providing jobs excluding healthcare and education or "Real Consumer Spending Less Healthcare and Education" is reflecting the real situation of US economy at present. To entangle this mystery we need first to ask why such a job loss in this sector and then to look into the details of the compilation of Real Consumer Spending.

The last recession was triggered by massive consumer default on mortgages and other consumer credits, an inevitable result of the perpetual runaway US trade deficit as explained in article 13. With such defaults the spending power of consumers was curtailed. Less spending power means less final sales and thus many businesses fell, causing the loss of jobs. Many surviving businesses had also cut back the number of employees to save cost. Such actions created large unemployment that in turn eroded consumption power further that led to less final sales for businesses. Thus a vicious cycle was formed to result in massive job loss. To fight the severe recession US Government pumped out huge amount of liquidity in the form of the direct or indirect injection into commercial banks, investment bankers, insurance companies and auto makers, whereas also handed out huge amount of money to consumers directly through the social welfare programs as more and more consumers has gotten into trouble. Thus the spending power in the form of current Dollar returned. Since many businesses had disappeared, consumer buying is concentrated into remaining businesses. However, those remaining businesses have refrained to expand their work force in a significant way but have forced their existing employees to shoulder heavier workload. That was the reason of severe job loss in the recession and the anemic recovery of the blue curve in the second graph. In order to resolve the mystery of contradiction between the anemic job market recovery and the bouncing back of Real Consumer Spending, we need to examine how Real Consumer Spending is compiled next.

The first step of the compilation is to sum up all the dollar values of consumer spending. The result is called "Nominal Consumer Spending". Next is to compile a price index for each kind of items that consumers buy. To compile such price indexes, a reference year needs to be defined. Currently year 2005 is used as the reference year in this country. For example, a cast iron skillet was sold for $20 in 2005 and for $22 in 2010, the price index of the iron skillet is 1.0 in 2005 and 1.1 in 2010 since the price of the skillet has gone up by 10%. When price index for each consumer good is compiled, a weighting is assigned to each price index according to the total sale amount of the good compared to the total nominal consumer spending. A weighted average of all price indexes is then calculated; let us call this weighted average of price indexes the price index of consumer spending. Real Consumer Spending is then calculated as

Real Consumer Spending = Nominal Consumer Spending / Price Index of Consumer Spending.

The calculation of price index actually is not so easy. Let us take an auto as an example. Suppose a car was sold for $20,000 in 2005 and a similar car was sold for $20,000 in 2010. We may be tempted to say that the price index of the car stayed at 1.0 in 2010. Unfortunately it is not so. The model of 2010 usually contains more accessories in the price than the 2005 model. Thus with the same price of $20,000 consumers in 2010 are getting a better car than the deal in 2005 so the price index of the car in 2010 should be set below 1.0. Such an adjustment of the price index to below 1.0 is called the improvement of a product. In principle if the car of model 2010 becomes inferior than the one in 2005, then the price index in 2010 should be set above 1.0 to reflect the negative improvement, but we believe that such negative improvement is rarely applied by the compilation agency.

The way to calculate the price index as outlined in the previous paragraph has serious problems. The first one is the subjective manner to apply product improvement. Bureau of Economic Analysis, the agency that compiles Real Consumer Spending (officially called Real Personal Consumption Expenditure or just Real PCE), applies the product improvement to durable goods more aggressively than US Bureau of Labor Statistics that compiles the familiar CPI index. Thus CPI is usually 0.5% to 1.0% higher than PCE price index. The arbitrary nature of the improvement of tangible goods is not the factor to explain the contradiction between anemic job growth and the strong growth of Real Consumer Spending, but the lack of negative improvement in the service sector is. Let us see why it is so.

Only a small fraction of the purchase price of a consumer product is attributable to the true cost to manufacture the product. The majority of the price that the consumer pays is called "overhead" cost that is actually the cost of various services rendered during the course for the good to reach the consumer. The anemic recovery of the blue curve in the second graph means one service worker now must handle more customers compared to the pre-recession era. This means that the quality of service has declined. For an unbiased compilation of price index, this negative improvement in the service sector should be taken into account but not. If the negative improvement of services is taken into account, the PCE price index in 2011 will be higher than the published index so Real Consumer Spending will be lower than the published value, and the gap between Real Consumer Spending and the anemic job growth will be buried. Let us see how much increase of the PCE price index will do the job of resolving the contradiction. Real Consumer Spending less Healthcare and Education bottomed out in April, 2009. In January, 2011, it has risen by 3.8% in 21 months. Translated into the annual rate, it becomes 2.2%. This means that if PCE index is raised by 2.2% per year, all the recovery of Real Consumer Spending less Health and Education will disappear. Of course, we do not need to raise PCE price index by 2.2% per year, just the raise of the order of 1.5% per year due to the negative improvement of services will bring the picture of Real Consumer Spending closer to that of the anemic job growth. From this discussion we can see that the anemic recovery of the blue curve in the second graph reflects accurately the current situation of the economy, but not the inflated Real Consumer Spending based on biased compilation of PCE price index.

4. Goods Producing Jobs

Goods producing jobs are divided into three categories, manufacturing, construction, and mining and logging. The number of jobs in the mining and logging sector is rather small compared to other two sectors so will be ignored here. Manufacturing is further broken down to durable goods and nondurable goods. In the third figure at the right manufacturing jobs are depicted as the blue curve, durable goods as the purple curve, nondurable goods as the green curve and construction jobs as the red curve respectively.

In the construction sector the red curve shows that slightly more than one million jobs were added during the housing bubble, and than more than two million jobs were lost as the bubble burst. Currently the red curve is crawling at the bottom with scan sign of rebound. It will take many more years to get back to the level of year 2000.

The green nondurable curve shows a steady decline. During the recent recession the trend of decline has only accelerated a little, then it has stabilized but the decline has resumed recently. At the beginning of 2000 this sector had 6.5 million jobs, but now only has 4.5 million jobs. With this rate of decline, in another 10 years the number of jobs will decline to 2.5 million. The reason of the decline is, of course, the outsource of manufacturing to foreign countries.

The purple durable goods curve is showing somewhat different patten compared to its nondurable cousin. The sharp drop of the purple curve is synchronized with the recessions, and it stabilize in the recovery and the bubble phases. Currently it is entering the stabilization phase. At the beginning of 2000 there were 11 million jobs in this sector, now there are only 7.2 million jobs left in the durable goods sector. The reason of the decline of jobs is again due to the outsourcing of the manufacturing to foreign countries. In general the outsourcing is mandated by Wall Street. If a manufacturer refuse to outsource, its stock price will drop and the management team will be replaced by a team that is eager to outsource. Eventually only an insignificant manufacturing ability will be left and the ability of US to continue as an economic super power will be in serious doubt.

5. Government Jobs

Government jobs are divided into federal government jobs, state government jobs and local government jobs. Federal government jobs are further divided into US Post Office jobs and federal government jobs excluding US Post Office. US Post Office jobs have been declining steadily since 2000 due to automation so will not be discussed here. State and local government jobs are divided into educational jobs and jobs excluding education respectively. In the fourth graph at the right, federal government jobs excluding US Post Office, local government education jobs, local government jobs excluding education, state government education jobs and state government jobs excluding eduction are depicted by the green, blue, red, purple and black curves respectively.

The two spikes in the green federal government job curve, one at 2000 and the other at 2010, are due to the hiring of temporary census workers. Beyond that federal government has increased hiring of 250,000 workers for public works since the onset of the recession.

State government workers excluding education, local government education workers and other workers all peaked around 2008 and all have declined since due to budget woes. Only the number of workers in state colleges and universities had plateaued during the recession and are growing again as the recovery comes. Probably the growth of workers in state colleges and universities is in the rank of administrative staff instead of from teachers and supporting staffs, just as their private counter part as discussed in Section 3.

6. Concluding Remarks

In Section 3 it has been pointed out that the actual situation of the economy is more close to what the job situation depicts rather than close to the rosier inflated consumer spending picture based on biased price indexing without taking into account of negative improvement in the service sector. This point is not just an academic interest but has real consequence in the decision of monetary policy. In the case of published inflation gauge without the consideration of the negative improvement of service, the inflation rate at the early stage of recovery will appear exceptionally low. This deceptively low inflation rate will often mislead the monetary authority into believing the onset of deflation and prompt it to over stimulate the monetary condition. When businesses start to hire more workers and the cost of labor increases, the inflation rate will shoot up. Only then the monetary authority will tighten the monetary condition belatedly, causing unnecessary boost to the bubble and then abruptly pierce the bubble. If the negative improvement of service is taken into account, the inflation rate at the early phase of recovery, like at present, will be sufficiently higher than the published rate so that the monetary authority will be restrained to do over stimulation. As businesses start to hire more workers, the negative improvement in the service sector disappears and a factor to boost inflation rate also dissipates. Thus the increased inflation from the increase of labor cost will be canceled out to give a picture of stable inflation for a prolonged period than the biased published inflation rate. As the consequence, the monetary authority does not need to give the monetary policy a wide ride as happened during 2003 to 2006 and what is happening now.

The sorry picture of the job loss in the manufacturing sector as discussed in Section 5 indicates the unavoidable conclusion of the self-destruction of an economic super power. As has been pointed out in articles 12 and 13, such wonton destruction of manufacturing jobs as the consequence of the outsourcing to foreign countries is dictated by the short term interest of Wall Street, and it can only be stopped when the current ad hoc globalization scheme self-destructs.