What is an often-misunderstood economic statistic quoted routinely in news stories? It’s the unemployment rate.

The Salem area’s February unemployment rate was 4.1 percent, and approximately 8,870 persons in Marion and Polk counties (the Salem Metropolitan Statistical area, or Salem MSA) were unemployed.

Both the rate and the number of unemployed are estimates, not exact numbers. And, being unemployed has nothing to do with collecting unemployment insurance benefits. There’ll be more on both these points.

What does it mean?

How the information for the unemployment rate is produced and interpreted, and how some 100 or so Marion and Polk County households play a part in this process, will answer this question, and provide a bit of insight into the political side of the topic as well.

The definition of the unemployment rate is: those aged 16 and older who are unemployed, divided by the labor force, and expressed as a percentage.

A person is considered unemployed if he/she has used an active job search method to look for a job in the last four weeks.

The labor force is those unemployed plus those employed.

And finally, a person is considered employed if they; worked during the last four weeks for at least one hour as a paid employee; worked in their own business; worked at least 15 unpaid hours in a family business; were temporarily absent from work because of illness, bad weather, etc.; or were waiting to be recalled back to work after being laid off. See more extensive definitions here.

The definitions are official and precise, the “official” being the federal office of the Bureau of Labor Statistics (BLS), the executive agency responsible for them.

Precision is important because the information used to calculate the unemployment rate comes from a monthly survey of households, conducted jointly by the BLS and the Census Bureau, called the Current Population Survey.

Household members aged 16 and older answer questions about work experience or lack thereof. Responses are put into precise categories, so that the information is consistent from month to month.

Precise answers became important during the first months of the pandemic when an unprecedented number of workers lost jobs, some as a result of business shut-downs to slow the spread of Covid.

During those early months, some survey responses were difficult to categorize. For example, one of the questions to a person laid off from a job is “do you expect to be called back to work?” Who knew the answer to that one in April of 2020? So, extra survey questions were added to elicit as precise a response as possible and ensure survey consistency.

There’s more to the story.

The number of households contacted monthly across the U.S. is 60,000, out of a total of 127 million households, a good-sized sample. The number of Oregon households in the survey is approximately 1,000 of 1.7 million households, 100 or so households out of 66,000 for Salem. These are both small samples, but are assumed to be representative of the population they are drawn from.

How reliable is an unemployment rate calculated from small samples?

The Oregon and Salem MSA unemployment rates are best analyzed over a number of months. Here’s why – the shorter the time period under consideration, the bigger chance there is that there might be “noise” in the sample.

What does that mean? When there’s something out of the ordinary about the information produced by the household survey, it’s possible that some households in the survey aren’t as representative of the population as they should be. This may skew the numbers, and is called “noise.”

Here’s an example. Ordinarily, the U.S, Oregon and Salem MSA unemployment rates track together well. (see graph)

From 2014 to 2022, this was pretty much the case. Then in 2022 and 2023, the Oregon and Salem unemployment rates rose noticeably above the U.S. rate. It’s possible that this out-of-the-ordinary occurrence was due to noise in the sample.

But here’s the bottom line: While we can’t say with certainty that Salem’s unemployment rate was exactly 4.1 percent, and that 8,870 were unemployed, we can say with certainty that the Salem’s unemployment rate has been historically low over the last few years.

Let’s return briefly to those folks collecting unemployment insurance benefits.

The Oregon Employment Department collects taxes from employers to establish the unemployment trust fund, pays out benefits to those eligible, and the tally is published monthly. In February 2024, there were 2,598 individuals in Marion and Polk counties collecting benefits.

This number is in line with the household survey. The survey tells us that more than half of the 8,870 unemployed were entrants into the labor force – that is, it’s their first time looking for work, or, first time looking for work after being outside the labor force for some time.

Another third or so were those laid off and looking for work – these individuals are generally eligible for benefits and would be approximately 2,661 persons, close to the number actually collecting. The remaining percentage of unemployed are those who’ve quit jobs to look for a different or better one.

The household survey also provides demographic information that includes gender, race, Latino origin, educational attainment, veterans’ status, and presence of disabilities, but most of these only at the national level. At the Oregon level, only gender, age, race, and Latino origin are published – that small sample again. Demographic information isn’t published for the Salem area unemployed.

Why is the number of households in the Oregon and Salem MSA samples so small?

Here’s where politics come in. It costs money to do the surveys. The Bureau of Labor Statistics and the Census have budgets that, like all federal agencies, are subject to the vagaries of the Congressional budget process.

Besides budget constraints, another aspect of how statistical information is gathered in the U.S. is how decentralized it is. There are 125 different agencies or parts of agencies producing some type of statistical information, and BLS and the Census are two of these.

This contrasts with other developed countries, where producing this kind of information is highly centralized, and likely more efficient and less costly. On occasion various U.S. commissions have recommended more centralization with few results. Agencies like control over their own information.

To conclude: the unemployment rate published monthly for the Salem area is an estimate from a small sample of Marion and Polk County households, and it is best to view it over a number of months to deduce a trend

And last but not least, the 100 or so households deserve thanks for their considerable and voluntary commitment to helping produce some of the most important descriptors of the area’s economy.

Pam Ferrara of the Willamette Workforce Partnership continues a regular column examining local economic issues. She may be contacted at [email protected].