Wednesday, October 11, 2017

Wage Report Raises Questions

They say there are lies, damned lies and statistics. I don't subscribe to that, but with 12 years as a political economist in public policy, and a total of 17 years as a Ph.D. economist, I also know that people in influential positions often try to twist statistical facts into verifying their particular position. 

Responsibility for that is almost always on the shoulders of politicians whose desire to prove a point supersedes informed judgment (making life much harder for honest politicians who are simply trying to engage in civil, dispassionate discourse). However, as exemplified by my comments regarding the work of our state government's chief economist, an expert can make a big difference as well. 

In its July 2017 Wyoming Labor Force Trends report, the Wyoming Department of Workforce Services provides another example of the responsibility that rests on the shoulders of experts. The report claims that average weekly wages have increased this year:
During first quarter 2017 (2017Q1), Wyoming's total wages and average weekly wage increased over the year for the first time in two years, according to data from the Quarterly Census of Employment and Wages (QCEW). The last time the state experienced an over-the-year growth in total wages and average weekly wage was in 2015Q1. ... Total wages increased from $297 billion to $2.99 billion, an increase of $11.3 million, or 0.4%. ... Wyoming's average weekly wage increased from $924 in 2016Q1 to $954 in 2017Q1 ($30, or 3.2%). 
I went back and checked on these numbers, right to the source that the report's authors refer to. I found one error and one issue of questionable data selection that I will get back to in a moment. First, though, it is important to note that these numbers are at odds with what other sources say. That does not mean they are wrong, but they represent an incomplete picture - and therefore these numbers are of questionable value. 

On September 26 I explained that work-based income keeps falling in Wyoming. Based on the latest personal-income data from the Bureau of Economic Analysis, I noted that:
1. Our state's total personal income grew by 1.2 percent from the second quarter of last year to the same quarter this year; but

2. When we look more closely at where that increase came from, it turns out that it was not from work-based income;
3. In fact, wages and salaries declined by 0.5 percent, marking the ninth quarter in a row with current-price decline in wages and salaries earned in our state;
4. The component of personal income that increased the most was dividends, interest and rent, i.e., income from investments, which were up by 3.8 percent;
5. In addition, personal current transfer receipts - entitlements including but not limited to Social Security, Medicare and Medicaid - increased by 1.8 percent.
These are numbers for Q2 of this year; the report from the Department of Workforce Services refer to first-quarter numbers. However, the numbers from the Bureau of Economic Analysis are hardly better for Q1. Here is how total wages and salaries (please note, "wages and salaries") have developed over the past three years:

Figure 1
Source: Bureau of Economic Analysis

These numbers differ from the report by the Department of Workforce Services in two ways. First, it reports annual, seasonally adjusted data that in turn has been broken down by quarter. The annual-rate part is of less importance; more important is the fact that these numbers are smoothened out seasonally. On the one hand, this kind of "manipulation" of statistics puts a filter between reality and the statistics consumer; on the other hand, seasonally adjusted data is easier to use for the purposes of displaying trends. 

In other words, if our purpose is to highlight the long-term development of wages and salaries in Wyoming, we could easily rely on seasonally adjusted numbers. If we are just out to give a "snapshot" of reality, non-seasonally adjusted numbers are preferable. 

Since the Workforce Services report uses non-adjusted numbers, we could assume that they are less interested in the long-term trend than in the current state of the economy. Yet at the same time, their report concentrates its findings on one point: that wages are rising. There is an implication in this that somewhat contradicts the notion that the Workforce Services report only wants to give a snapshot of the economy: as Figure 1 above suggests, a comprehensive look at the compensation trends in Wyoming suggests that work-based earnings are not at all on the rise. 

This brings us to the second difference between the Workforce Services report and the numbers from the Bureau of Economic Analysis. The former reports only "wages", while the latter includes "wages and salaries". Since the data in the Workforce Services report are from the Bureau of Labor Statistics, it is their definition of wages we will have to rely on; as far as the form of pay - wage or salary - the difference is not significant from that which the Bureau of Economic Analysis uses for its "wages and salaries" report. However, the periodicity of pay can make a small difference, as can the fact that the BEA applies a standard definition across all states while the BLS conforms its state-by-state reporting to - admittedly small - differences between states in the definition of what counts as a wage.

In other words, there are small definitional and methodical differences that, over the short term, could create a discrepancy between the two datasets. Long-term, though, over a year or more, that difference should be of no consequence. Therefore, the main reason for the difference between the Workforce Services numbers and the numbers from the BEA are most likely due to differences in seasonal adjustment of the data itself. 

What does this mean for the usefulness of the data? The main point is that anyone who wants to use these numbers for policy decisions should avoid relying on the Workforce Services numbers. Since they are not tuned to report trends, it is difficult to say whether or not the uptick in wages is temporary or an exhibition of resiliency in the Wyoming economy. The BEA data, on the other hand, suggests reliably that no such resiliency exists. 

With all this in mind, I do not want to take anything away from the Workforce Services data reporting in and of itself. Raw data - which, among other things, is not seasonally adjusted - is essential for any direct understanding of what is going on at the ground level in the economy. What matters is how you use that data in your analysis. There, the choice of data is as important as is your theory and overall methodology.

There is, however, another point to be made about the Workforce Services data. Its numbers for the average weekly wage in Wyoming do not correspond with the source they refer to. They claim that the average weekly wage, as reported by the Bureau of Labor Statistics Quarterly Census of Employment and Wages (QCEW), the first-quarter number for 2017 is $954, up 3.2 percent over the $924 number for the same quarter in 2016. 

Here are the Wyoming numbers as published by the QCEW, for all industries and all establishment sizes:

Q1 2016: $852
Q1 2017: $880

For just federal government employees in Wyoming:

Q1 2016: $1,219
Q1 2017: $1,253

For Wyoming state government employees:

Q1 2016: $1,060
Q1 2017: $1,058

For Wyoming local government employees:

Q1 2016: 816
Q1 2017: 837

These numbers are lower than private-sector wages reported in the same publication (see below), a fact that seems to contradict my report on government vs. private employee compensation. However, those numbers included benefits, which in turn includes both pension and health benefits. With those accounted for, government employees at all levels make more than private-sector employees, just as I reported the other day.

Anyway. Onward and upward. Here are the numbers for the total government sector, all levels, in Wyoming:

Q1 2016: $904
Q1 2017: $923

For private-sector employees in Wyoming:

Q1 2016: $834
Q1 2017: $865

I am curious as to where the Workforce Services got their numbers from. Furthermore, it is worth noting that they do not separate their wage data into separate numbers for government and the private sector. For the purposes of devising legislation, such a separation is essential; at the end of the day, one would expect that a department under the executive branch spends money on producing statistical reports exclusively for the purposes of helping the governor make good policy decisions. For that purpose, it is essential to know the differences between wage trends in government and in the private sector - as well as the total costs of the government workforce, as I pointed to the other day.

Perhaps the problem with the Workforce Services report is not its design per se, but that there seems to be a lack of coordination across the state government's statistical publications. I have repeatedly criticized the reports from the Consensus Revenue Estimating Group and publications from the Economic Analysis Division for giving an incomplete macroeconomic picture of Wyoming. The Workforce Services report, while interesting as an isolated product, contributes to the confusion about what is really going on in Wyoming. It does so by not placing its data in a context where it can be used to inform economic policy legislation. 

Perhaps a future gubernatorial administration will reform the state government's production of economic statistics. In particular, it would be good if taxpayers' money could be used to explain our state's macroeconomic performance performs beyond variables that are of interest exclusively to the tax-consuming sector of our economy.  

No comments:

Post a Comment