Oregon Office of Economic Analysis

This is interesting. We have three different measures (direct or indirect) of wages in Oregon and each is telling a somewhat different story. First, there are the Wage and Salary disbursements as reported quarterly by the Bureau of Economic Analysis (this is the data source for the quarterly updates on Oregon’s personal income, see this blog post from November for more information). Second, the QCEW, or Quarterly Census of Employment and Wages, has aggregate wages paid to employees covered by the unemployment insurance system (about 98% of all employees.) Third, the Oregon Department of Revenue collects withholding tax receipts on a monthly basis. While withholding tax is levied on more than just wages (pensions, bonuses, gambling winnings, etc) the bulk of these receipts do come from wages, so it is a very good, yet indirect, measure of Oregon wages.

The graph below illustrates all three of the measures on a year-over-year growth rate basis. Both the BEA wages and QCEW wages data is through 2011q3 while withholding is through 2011q4.

What’s going on here? First, withholding tax collections are the best, timely data we have however it is subject to tax policy changes. In 2007, the withholding tables were changed to withhold less money out of each paycheck than previously had been held, hence the growth was much lower than the other measures. In 2011, this change was reversed and so withholding growth has been stronger than underlying wage growth – strictly due to tax policy changes. The QCEW data is based on unemployment insurance records that employers must file with the state and due to the extremely large sample size (98% of employees), the data is typically very good but does come with a lag (third quarter data was just released this morning). It is our belief that this represents the best, true estimate of underlying wage growth in Oregon, particularly in real time. The light blue line, based on BEA data, is subject to revision as the BEA estimates data for all states based on limited sample information. The BEA data is the most cited as it is very detailed in terms of the components of income (of which wages are just one component), however it does tend to be revised more so than the other measures. In 2012 and beyond (after revisions to BEA data and after the effect of the withholding table change wears off) expectations are for the three measures to closely track each other like they have historically.