Planning and Economic Development Information
Scott Bailey, Regional Economist

Labor Market and Economic Analysis Branch
Gary Bodeutsch, Director

Washington State Employment Security
Carver Gayton, Commissioner                                                                           October 1999

Each year, the Employment Security Department publishes the average monthly employment, total payroll, and average annual wage for all jobs covered by unemployment insurance¹. The average wage is available by industry and by county, and is often used to gauge the health of the economy. Like all averages, however, it can conceal as much as it reveals. For example, because the average is not adjusted for part-time workers, changes in the average may be due as much to changes in the average number of hours worked as to changes in the hourly wage. Further, changes in the average give no clue as to how the entire distribution of wages is changing. As is evident from data at the national level, a rising average income has not meant that all income levels have benefited equally.

This study attempts to deal with these two issues by tracking how the distribution of hourly wages has changed over time, specifically the 1990-97 period.

Methodology

The data for the study come from quarterly unemployment insurance reports submitted by all employers in the state to the Employment Security Department. Each record in the database includes the employer’s account number, the social security number of an employee, the number of hours worked by that employee during the quarter, and the total wages earned. Roughly 2.5 to 3.0 million records are in each quarter’s database. Because of the way data is reported, it is impossible to tell definitively whether a worker is seasonal, part-time, or in transition from one job to another or into/out of the labor market.

Each record was assigned the county and industry of the employer. For multi-county and multi-industry employers, an algorithm was developed to assign the county and industry. Unfortunately, there is no way to identify the age, sex, or occupation of the employee. For about 10% of the records, the number of hours worked was not identified; hours worked were estimated using similar records from the same industry. In addition, a small number of workers received room and board along with wages; in the database, they were assigned 1000 hours as a convention. These records were excluded from the analysis.

The average hourly wage for each employee was calculated by dividing total wages by total hours worked in each quarter. The average was adjusted for inflation using the U.S. Chain-Linked Price Deflator for Personal Consumption Expenditures². The cost of living, of course, varies by county.

Using the Consumer Price Index as a basis, it would be a fair assumption that prices increased faster than the U.S. average in the Seattle metro area, and in Clark County. This would mean that wages in these counties did not improve as much as this study suggests. The outcome is less clear for other counties; some rural counties, for example, experienced a run-up in housing prices during the 1990s, while housing prices were relatively slow-growing in others.

Exclusions

The Wage Files include all workers at Washington businesses covered by state unemployment insurance. Workers not included: federal employees, self-employed workers, railroad employees, 100% commission sales workers (such as most real estate agents and insurance agents), most corporate officers, and casual labor. Coverage was largely unchanged between 1990 and 1997, with one exception: workers employed by private households (SIC 88), such as nannies, those providing in-home care for the elderly, domestic servants, etc. Reporting requirements changed for these workers in the mid-1990’s, and as a result their numbers increased dramatically. Comparisons of 1990 and 1997 data in this report exclude this industry.

 
Definitions/How to Measure Wage Distribution

Of all statistical measures, the one most commonly used is the average. It’s a single measure, and is easy to understand. Yet it conceals as much as it reveals. If the richest individual in the world moved next door to you, the average income of your neighborhood would rise substantially—regardless of whether the income of your household or your other neighbors went up or down. To keep tabs on the monetary fate of the "average" person, ironically enough, tracking the average wage just won’t do. To better illustrate the wage trends, this report makes use of several other measures, depending on the question asked.

"How many jobs pay below the proposed new minimum wage?"

How many pay above $10/hour? How many family wage jobs are there? These and similar questions can be answered in a straightforward fashion. Graphs and tables throughout this report will present in detail the number and percent of jobs in various hourly wage categories.

"Have the rich gotten richer? Did the poor become poorer?"

To answer these questions, as well as to follow the fortunes of the "average" wage earner, we need a few more concepts. The median wage is the hourly wage at which half of all jobs pay more and half of all jobs pay less. In the example above about the new rich neighbor, the median income of the neighborhood would change very little if at all.

While the median can tell us about the average job, it doesn’t tell us much about the extremes. To do that, imagine sorting all jobs from top to bottom by wage. The percentile wage is the top hourly wage for a given percentage of jobs. In this report, jobs have been divided into ten groups, or deciles. The 10^th percentile wage is the median wage of the bottom ten percent of jobs. The 10^th percentile upper limit is the topmost wage of the bottom ten percent of jobs. Similar calculations were made for each decile. The median wage is the same as the upper limit of the 50^th percentile.

To calculate whether the poor are getting poorer, or more precisely, if low-wage jobs are paying any more over time, we can compare the 10^th percentile wage in different years. To determine whether the gap between high-wage jobs and low-wage jobs is increasing, we can look at three ratios:

• The 90/10 ratio, numerically, is the 90^th percentile wage divided by the 10^th percentile wage, and measures the span between top and bottom.
• The 90/50 ratio measures the gap between the 90^th percentile wage and the median wage.
• The 50/10 ratio similarly measures the gap between the median hourly wage and the 10^th percentile wage.

Following these three ratios over time allows a precise measurement of whether hourly wages are becoming more or less widely dispersed.

While percentile wages provide a much more in-depth look at the labor market, they don’t tell us what is happening at the very bottom or very top. On the bottom side, the minimum wage provides a limit. However, there is no ceiling on the hourly wages of the top 10 percent of jobs. To capture the trend of the best-paid jobs, the percent of total payroll for each of the 10 percentile groups will be compared over time.

Summary statistics

In 1997, according to published data, covered employment in the state of Washington averaged just over 2.5 million a month, with a total payroll of $77 billion and an average wage of $30,755 per worker. What these averages do not reveal is the dynamic mix of comings and goings, jobs beginning and ending, workers entering and leaving the workforce.

Analyzing the quarterly wages files gives us a chance to get considerable more detail. For example, the mix of full-time and part-time jobs can be converted to a full-time equivalency (FTE) basis of 40 hours per week and 2,080 hours per year. On an FTE basis, there were 1,927,371 jobs. The average workweek per job was 31.5 hours. The average hourly wage was $18.03. [Total wages in the wage files came to $73 billion, $4 billion short of the published total for the state. The difference was due to the absence of federal workers in the state database, and the exclusion of jobs coded to 1000 hours per quarter, which included room and board or other non-wage benefits as part of the working conditions.]

Over 3,000,000 different people worked for employers in the state that year, clocking over 4 billion hours. Some were employed only briefly; others worked full-time, year-round, or more. Those who worked at least 1,560 hours—the equivalent of working 30 hours or more per week, year-round—came to 46% of the total.

Hourly Earnings

The graph and table below show the percent of all jobs falling in each hourly wage category, which have been grouped by two-dollar increments for ease of presentation. The highest frequency (mode) came at $6 to $8 per hour; 12% of all jobs paid in this range. The percent decreases as the hourly wage increases. For convenience of presentation, hourly wages at the top of the scale have been grouped into three categories: $30 - $40/hour, $40 - $50/hour, and $50/hour and up.

There can be no definitive measure of a "family-wage" or "livable-wage" job. Families come in all ages and sizes, with differing numbers of wage earners and unique material needs. What constitutes a reasonable standard of living is subjective, and the cost of living varies throughout the state. Hourly wages in the database were calculated down to the penny; for presentation purposes, jobs were grouped into five categories:

• Below $8 per hour
• Between $8 and $12 per hour
• Between $12 and $16 per hour
• Between $16 and $24 per hour
• $24 per hour and up.

Median and Percentile Earnings

Does a rising tide lift all boats? Or, in the case of wages, does an increase in the average wage mean that workers across the board are benefiting? The easiest way to answer this question is to calculate percentile wages. The most commonly-used percentile wage is the median wage: the wage at which half of all jobs pay more, and half pay less. In the analysis that follows, the job pie was divided into ten parts, with the lowest-paying 10% of all jobs falling in the 10^th percentile, the median being 50^th percentile, and so on.

Percentile earnings are shown in Graph 2. The median wage was $13.96/hour. The 10^th percentile wage was $6.68/hour. The highest-paying ten- percent of all jobs paid above the 90^th percentile wage of $29.72/hour. One method of measuring wage dispersion is to calculate three ratios: the 90/10 ratio, the 90/50 ratio, and the 50/10 ratio. In 1997, those ratios were 4.20, 2.04, and 2.05. That is, the median wage job paid about twice the 10^th percentile job, and about half of the 90^th percentile job.

Median wages for assorted industries are shown in the table below. The average for all jobs was $13.96/hour. Industry medians ranged from $6.28/hour in the motion picture industry (includes theaters and video rental stores) up to $26.81/hour in petroleum refineries.

Jobs Paying Below $8 Per Hour

Over 350,000 jobs paid below $8/hour in 1997—18% of the total. One out of six of these lower-wage jobs was provided by an eating or drinking establishment—a fast-food or sit-down restaurant, a tavern, or a bar. Another 21% were located in other retail establishments. Almost a third were in service industries, and one out of ten were in agriculture.

The concentration of low-wage jobs varied greatly by industry, as shown in Table 5. Two-thirds of all jobs at crop farms paid below $8/hour. High levels of lower-wage jobs were also present in childcare services, motion picture production and rental, restaurants & drinking establishments, private households, and lodging. Relatively few jobs in the aircraft industry paid a low hourly wage. The same was true for utilities, aluminum, state government, heavy construction, and hospitals.

Jobs Paying $8 to $12 Per Hour

Another 440,000 jobs—almost one-fourth of the total— paid between $8 and $12/hour. Over one-quarter of these (28%) were in service industries, with health care and business services being the major sources. Another one in five was in retail trade, while one out of six was in manufacturing. Again, the concentration varied by industry, as shown in the table below. Industries with at least 40% of their jobs paying in this range included fruit & vegetable processors, residential care, janitorial services, millwork/plywood/structural members, and retail building materials. Low concentrations were found in industries such as aluminum, aircraft, utilities, logging, software, and heavy construction.

Jobs Paying From $12 to $16 per hour

Almost 360,000 jobs paid between $12 and $16/hour in1997. This pay range accounted for 18% of all jobs. The differences between industries were less pronounced than with the lower wage levels.

Summary

In 1997, over 1.1 million jobs in the state paid below $16/hour. This amounted to 59% of all jobs on a full-time equivalent basis. There were a number of industries in which fewer than 10% of their jobs met this wage standard.

Higher-Wage Jobs: $16 to $24 per hour

Looking on the more positive side, 790,000 jobs-- 41% of the total—paid more than $16/hour in 1997. A bit more than half of these (430,000) paid between $16 and $24 per hour. Industries with 40% or more of their jobs falling into this range are shown below.

Higher-Wage Jobs: $24 per hour and up

Over 360,000 jobs—19% of the total—met the "high-wage" standard in 1997. A handful of industries accounted for than half of these jobs. The most important industry, not surprisingly, was aircraft, which contributed 15% of all high-wage jobs in the state. Next came local government (county, city and other public agencies, along with K-12 education), also with 15% of the total. Transportation & utilities, wholesale trade, health care, and construction each chipped in 7% of all high-wage jobs.

Differences by County

The state of Washington has a remarkably diverse array of economies at the regional and county level. The political conversation about the "other Washington" has long contrasted the east and the west sides of the state. But the fractures are much more complex than that. Clark, Whatcom, and Yakima, for example, are all metropolitan counties, with distinct economies, as different from each other as they are from Seattle.

It’s no surprise, then, that there are wide differences in wage distribution at the county level. For example, the 10^th percentile threshold wage varied from a low of $5.53 in Okanogan County to a high of $7.47 in King County. Only five counties had a higher hourly wage than the state average of $6.68. The gap increased at the median, where Snohomish County had the high of $15.76 and Okanogan the low of $8.84. At the 90^th percentile, Benton County posted the top mark $33.16, while Okanogan again had the lowest percentile wage at $18.77.

The dispersion of wages, measured by the 90/10 ratio, also varied widely. Benton County, which had the highest 90^th percentile wage, had a 90/10 ratio of 5.51. In the San Juans, meanwhile, the ratio was only 3.14.

Differences by Size of Employer

The larger the employer, the higher the average wage, and this holds true throughout the entire distribution of hourly wages. For this study, employers were classified into eight categories based on their average employment in 1997. The accompanying table compares the percent of jobs in each size class paying in each broad hourly wages category. More than 55% of jobs at the smallest employers paid below $12/hour, while only 27% of the jobs in the largest size class failed to meet this threshold. Conversely, 11% of the jobs at small employers paid above $24/hour, versus 28% at the largest employers.

New Entrants to the Labor Force

Out of the 3 million different people who worked for a Washington employer in 1997, over 430,000 (about 14%) could be classified as new entrants—they did not appear in the 1990, 1995, or 1996 databases. They accounted for 92,600 FTE jobs through the year (about 5%). These new entrants might be:

1. People of any age, but most likely young adults, entering the workforce for the first time.
2. People re-entering the workforce after a long absence.
3. People who re-located to the state during 1997.
4. People who worked temporarily in the state during the year.

The hourly wages of these new entrants tended to be far below the average. While 18% of all jobs paid below $8/hour, 44% of the jobs held by new entrants were in this range. The median wage was $8.72/hour. Only about half as many of the jobs held by new entrants paid above $16/hour as for all jobs (23% vs. 41%).

The industry profile of new entrants differed from the profile for all jobs in several ways. Compared with the average worker, new entrants were much more likely to be found in agriculture, restaurants, temp agencies, and video rental stores; they were relatively scarce in manufacturing (especially aircraft, paper products, and aluminum), private utilities, banking, hospitals, state government, and local government.

1990 Overview

In 1990 there were about 2.1 million jobs covered by unemployment insurance in the state of Washington, with a total payroll of $48.5 billion and an average wage of $22,635 ($27,456 adjusted to 1997 dollars). The state economy was in near-peak condition, with job growth at almost 5% over the year.

Hourly Earnings

When compared with 1990, 1997 hourly earnings show a distinct improvement. There were fewer jobs at lower wage levels (up through $14/hour), and more jobs at higher wages ($24 and up). During this period, the effective minimum wage rose from $4.25 in 1990 to $5.15 in the fourth quarter of 1997—almost no change after adjustment for inflation.

Percentile/Median Earnings/Upper 10%

Earnings at the top of each of the ten deciles are shown in the accompanying table. The data show that, while earnings increased at each decile, the increase was greater in both absolute and percentage terms at higher wages.

Another way of expressing this dispersion is to compare the 90/10, 90/50, and 50/10 ratios in each year. Each increased slightly, with the 90/50 spread edging up by 4% and the 50/10 spread up by 2%. In other words, while the labor market as a whole moved up the wage scale, inequality increased.

The 90/10 ratio is limited in measuring inequality because it does not capture the fortunes of the top 10% of jobs. One way of overcoming this handicap is to look at the percent of total payroll accruing to the upper tenth of jobs. In 1990, the upper tenth earned 26% of the total payroll pie; by 1997, their share had increased to 29%. Meanwhile, the bottom 10% of jobs faced a slight decline in share from 3.48% to 3.15%. The share of the total payroll pie falling to each decile fell roughly three- to five-tenths of a point, with the exception of the upper tenth.

Hours Worked

No, it’s not just your imagination: you’re working more hours. At least, that hypothesis is not contradicted by this study. In 1997, the percentage of workers who toiled more than 2,080 hours—more than 40 hours per week—reached 19%. In comparison, only 13% of workers breached the full-time year-round threshold in 1990. This finding is in line with national studies showing a substantial increase in hours worked outside the home since 1960. Reasons for the increase range from financial need due to wage stagnation to financial want, in terms of yearning for a higher material standard of living. Another factor may come from the employer side, as managers find it less expensive to increase overtime rather than to hire, train, and pay benefits for new employees.

Industry Trends

Of 99 different industries analyzed, over half enjoyed improved earnings at every decile level. Two-thirds had widening inequality, as measured by the 90/10 ratio; in the most extreme example, the security brokerage industry (SIC 62), earnings at the 10^th percentile rose by a dollar, while those at the 90^th percentile jumped by over $26. Eighteen industries had increased earnings at the 90^th percentile and a decrease at the 10^th percentile, while six had the opposite trend. The software industry had the largest increase in median earnings (+ $7.18).

In the following sections, hourly wage trends in sixteen industries of interest are examined.

Aircraft

To state the obvious: the aircraft industry is very cyclical. In both 1990 and 1997, the aircraft industry (SIC 372) was at or near peak employment, so the comparison of wages is valid. In 1990, the industry employed 116,200 workers; in 1997, the average was slightly lower at 114,900. However, on an FTE basis, 1997 was slightly higher, indicating a longer workweek, and indeed, according to published reports, Boeing was paying a lot of overtime hours that year due to a rapid increase in orders.

Graph 6.  Aircraft Industry (SIC 372)
Percent of employment By Hourly Wage

The graph above shows a substantial positive shift in hourly earnings during the decade. Industry wages, already relatively high, climbed even higher. The percent of all jobs paying $16 to $24/hour fell from 43% to 32%; meanwhile, those paying $24 to $40/hour rose from 33% to 46%. The high wages and the positive shift can be attributed to the highly-skilled workforce, a high degree of unionization, a high level of profits, and the relative scarcity of labor during peak years. All helped to boost wages and promote overtime.

Agriculture—Farm Crops

In 1990, employment covered by unemployment insurance at farms in the state was just under 54,000. In 1997, the total was slightly higher at 57,000. On an FTE basis, however, there was a much more pronounced change—from 31,000 to 41,000. Jobs in the industry (SIC 01), which are highly seasonal, were lasting longer.

Graph 7.  Farm Crops (SIC 01)
Percent of Employment By Hourly Wage

Most of the additional hours worked were at low wages, however, and as a result, the overall wage distribution worsened. The proportion of jobs paying between $4 and $6/hour increased from 25% in 1990 to 28% in 1997. From the percentile point of view, wages at each decile declined.

Agriculture—Fruit & Vegetable Processing

Covered employment in fruit and vegetable processing was a bit over 13,000 in both 1990 and 1997, but as in the crop side of the business, jobs lasted longer; FTE employment rose 14%. Wages declined slightly for the bottom 20% of jobs, but improved modestly above that. The spread between the 10^th percentile wage and the 90^th percentile wage increased by 7%.

Graph 8.  Fruit & Vegetable Processing (SIC 203)
Percent of Employment By Hourly Wage

Lumber & Wood Products

Covered employment in lumber & wood products fell 12% from 1990 to 1997; on an FTE basis, however, the decline was only 5%. Logging employment was hit the hardest, but the number of sawmill and secondary wood product jobs also fell.

Between 1990 and 1997, the wage structure of the industry shifted substantially. The percent of jobs paying between $10/hour and $24/hour fell from 72% to 68%, while the share of low-wage jobs and high-wage jobs each gained about 2 percentage points. The loss in logging—the highest-paying sector of the industry—and a downward shift in sawmill and millwork/plywood wages were to blame.

Graph 9a.  Lumber & Wood Products (SIC 24)
Percent of Employment By Hourly Wage

Construction

Covered employment in construction grew by 14% during the study period, from 112,000 to 128,000. On an FTE basis, the gain was 23%. Construction workers enjoyed a generally positive shift in wages, with fewer jobs paying below $12/hour, roughly the same percentage in the $12 to $24/hour range, and an increase in the number and percent of jobs in the upper end.

Graph 9b.  Construction (SIC 15-17)
Percent of Employment By Hourly Wage

Trucking/Warehousing, Water and Air Transportation

Due to a shift in industry classification of a major employer between 1990 and 1997, these industries were combined. Employment trends were similar to the construction industry: a 14% increase in covered employment, but a 23% jump in FTE employment. Taken together, these industries suffered a mild decline in hourly wages, with a falloff in the percent of jobs paying more than $20/hour and faster growth in lower-wage jobs, especially in the $8 to $10/hour range. Water transportation actually had a fairly stable wage profile, with a slightly positive trend. Both trucking and air transportation shifted downward, however.

Graph 10.  Truck/Water/Air Transport (SIC 43/44/45)
Percent of Employment By Hourly Wage

Telecommunications

Telecommunications (SIC 48) has been one of the most dynamic industries of late, with employment cutbacks related to local phone service, and rapid growth in segments such as fiber optic carriers and cellular phone service. Overall, covered employment rose 27% to 29,000 jobs in 1997; FTE employment climbed even faster, at 41%.

The industry in 1997 has significantly fewer jobs paying below $10/hour. However, wages sagged somewhat in the mid-levels; hourly wages declined for the 20^th through 70^th percentiles. The 90^th percentile wage rose by 5%.

Graph 11.  Telecommunication (SIC 48)
Percent of Employment By Hourly Wage

Wholesale Trade

The wholesale trade industry added over 20,000 jobs between 1990 and 1997, an increase of 18%, reaching a total of 146,000. FTE employment grew by 25%. The wage profile shows an across-the-board improvement over the study period. Percentile wages rose from 1% to 2%, except at the upper ends: the 90^th percentile wage was 9% higher in 1990 than in 1997. Thus the disparity between low-wages jobs and high-wage jobs grew substantially.

Graph 12.  Wholesale Trade (SIC 50-51)
Percent of Employment By Hourly Wage

Eating & Drinking Establishments

Restaurants, taverns, and bars averaged over 140,000 jobs in 1990, and expanded to 165,000 jobs in 1997. Because of the part-time nature of much of the work, FTE employment was only 75,000 in 1990 and 95,000 in 1997. Hourly wages reported here do not include tips. With the exception of all but the lowest 10 percent of jobs, the industry paid slightly higher wages in 1997, with the median hourly wage increasing by a dime.

Graph 13.   Restaurants/Taverns/bars (SIC 58)
Percent of Employment By Hourly Wage

Finance

The finance industry—banks, credit unions, security brokers, and holding companies—grew slowly during the first seven years of the decade, due to an almost continuous wave of mergers, cost-cutting, and automation. Covered employment increased only 7% to 54,000 in 1997. The net effect, however, had a salutary effect on wages: there was a substantial drop in the share of jobs paying below $14/hour, and an increase in every wage range above that threshold. The improvement held for every segment of the industry. While wages improved across the board, gains were larger the higher the salary. Tenth percentile hourly wages rose by 6%, the median hourly wage by 14%, and the 90^th percentile wage by 23%.

Graph 14.  Finance (SIC 60, 61, 62 & 67)
Percent of Employment By Hourly Wage

Personnel Agencies

Employment at personnel agencies grew by 70% from 1990 to 1997 to a total of 46,000. At the same time, "temp" agencies became less temporary—FTE employment expanded at double that rate. While still primarily a supplier of lower-skill occupations, the industry has been adding more technical staff. Hourly wages went up across the board from 1990 to 1997. Much of the increase at the lower end reflects the tight labor market in Seattle in 1997—temp agencies are prime indicators of the market minimum wage.

Graph 15.  Personnel Supply Agencies (SIC 736)
Percent of Employment By Hourly Wage

Software

Software has been the fastest-growing industry in the state during the 1990’s, and is one of the highest-paying as well. Covered employment rose from 15,000 in 1990 to 40,000 in 1997. Hourly wages moved substantially upward during this period. The percent of industry jobs paying below $16/hour dropped almost in half from 39% to 20%. The median hourly wage rose by more than $7/hour, almost reaching $26/hour. Wages were boosted by strong demand for industry products and services, stock options, and a shortage of professional and technical labor.

Graph 16.  Software (SIC 737)
Percent of Employment By Hourly Wage

Health Care

Covered employment in health care grew by 46%, from 145,000 in 1990 to 174,000 in 1997. FTE employment rose at a slightly higher rate. Hourly wages in the industry underwent a substantial shift upward; the share of jobs paying below $12/hour fell from 45% to 37%. The percent of jobs falling in the high-wage category of $24+/hour climbed from 9% to 15%.

Graph 17.  Health Care (SIC 80)
Percent of Employment By Hourly Wage

Social Services

The fast-growing social services industry includes agencies assisting individuals and families, providing job training services, childcare services, and residential care. The industry median wage of $8.81/hour was one of the lowest; childcare ($7.03) paid especially low wages. The industry saw some improvement at the lower end, as relatively tighter labor markets in 1997 pushed up the market minimum wage.

Graph 18.  Social Services (SIC 83)
Percent of Employment By Hourly Wage

Engineering & Management Services

In contrast to social services, many jobs in engineering & management services require bachelor’s or professional degrees, including engineers, architects, accountants, management consultants, and researchers at commercial laboratories. High demand for these services from corporate customers helped boost wages in this industry. Fifty-five percent of the net new jobs created between 1990 and 1997 were in the $24+/hour range, and another 30% paid between $16 and $24/hour.

Graph 19.  Engineering & Management Services
(SIC 87)
Percent of Employment By Hourly Wage

K-12 Public Education

Public schools employ teachers, administrators, and a range of support staff. In comparison with some industries, public education’s wage profile changed little from 1990 to 1997.

Graph 20.  K-12 Public Education
Percent of Employment By Hourly Wage

Jobs Paying Below $8 Per Hour

The number of jobs paying below $8 per hour increased by 56,000 between 1990 and 1997; however, their share of total jobs fell from 19% to 18%. Net new low-wage jobs were concentrated in a few industries, as shown in the table below.

Jobs Paying $8 to $12 Per Hour

Over 70,000 net new jobs between 1990 and 1997 fell into the $8 to $12/hour range. However, the overall share of these jobs fell from 24% to 23% of the total. New jobs in this wage range were fairly evenly dispersed across industries, as shown in the table below, with a few exceptions. Manufacturing had proportionately more jobs in this wage range: while 6% of all net new jobs were in manufacturing, 12% of all net new jobs paying between $8 and $12 per hour were in factories. The same was true of retail trade. Finance and government had proportionately less net new jobs in this range, while state government actually had fewer jobs in this range.

Jobs Paying Between $12 and $16 per hour

Almost 75,000 net new jobs were added in the $12 to $16/hour pay range. In 1990 and 1997, jobs in this range made up 18% of the total. Again, net new jobs were fairly evenly spread across sectors. There were relatively few of these jobs created in farming, and somewhat fewer in services and local government, and proportionately more in manufacturing and state government.

High-Wage Jobs

Over 120,000 net new jobs created between 1990 and 1997 paid above $24 per hour. The percent of all jobs in the high-wage category increased from 16% to 19%. Industry detail shown in the table below confirms the importance of software and aircraft in the provision of high-wage jobs. Local government, health care, and engineering and management services also played crucial roles.

County Trends

Wage trends between 1990 and 1997 varied widely by county. Looking at the median wage, ten counties suffered a decline in hourly earnings, with Benton County having the sharpest drop at -$0.68/hour. Eight had increases of less than a quarter, eight between a quarter and fifty cents, ten improved between fifty cents and a dollar per hour, and three enjoyed larger increases, led by Garfield County at +$2.10/hour.

At the 90^th percentile, hourly earnings declined in three counties, while twenty-six counties had increases of more than a dollar an hour. At the 10^th percentile, eight counties eased downward, mostly by a dime or less, twenty-one inched up by a quarter an hour or less, and ten counties had larger increases.

Twenty-nine counties had increasing wage inequality, as measured by the 90/10 ratio; most spreads grew by five percent or less. Seven counties became more egalitarian.

Individual Trends

Of the 2.7 million individuals who worked in the state at some point in 1990, 1.8 million also worked in the state in 1995, and 1.7 million also worked in 1997. In analyzing their earnings over time, a distinction will be made between "full-time" workers and other workers. For the purposes of this report, workers logging at least 1,560 hours per year—the equivalent of working 40 hours per week for three-fourths of the year, or a 30-hour workweek year-round—will be considered full-time.

About 750,000 individuals worked full-time in both 1990 and 1995. Just over 700,000 workers met the same condition in both 1990 and 1997, while nearly a million worked at least half-time (1,040 hours) in both years. Over 600,000 worked full-time in the state all three years.

Most of these workers earned a higher hourly wage in 1997 than in 1990, but as shown in Table 23, about one in four suffered a drop in hourly earnings. The results are very similar for full-time workers, with slightly more enjoying a raise and fewer a decline. Graph 22 illustrates the curve for all full-time workers, with each data point a quarter-dollar increment. For example, about 2.9% of all workers gained between $1.25 and $1.50 in hourly wages.

Graph 21.   Hourly Wage Increase For Full-Time
Workers, 1990-97
Percent of Employment By Wage Increase

Table 24 compares all workers with those who earned below $8/hour in 1990. Those earning low wages did somewhat better than the general labor force, in part because there was so little downside potential. However, the improvement was generally not enough to boost their wages above $10/hour.

Table 25 shows that for full-time workers earning below $7/hour in 1990, 19% were still earning below $7/hour in 1997, 35% fell below the $8/hour threshold, 79% did not reach $12/hour, and only 6% topped $16/hour. The results were a bit better for workers making somewhat higher wages, but not greatly. In fact, there were a lot of similarities in the wage progressions for different starting wages. As the starting wage rose, the curve flattened on the wage gain side, and widened on the wage loss side. This is illustrated in Graph 22, which compares the wage gains for two sets of workers, one starting between $7 and $8 per hour, and the other earning between $15 and $16 per hour in 1990.