Thoughts on today’s midterm elections

 

 - by New Deal democrat

My thoughts on how today’s midterm elections might shake out are expressed pretty well by this thread by MTSW on Bluesky:

Recall that poll forecasts themselves, in addition to all the problems he cites, *themselves* move the electorate, as do reports on who has been voting early. So I expect the polls to be wrong. The question is, how broadly and in which direction? There’s evidence that late deciders have been shifting towards Democrats, but that’s about as far as I trust those results.

But I think the above framing is a good one: today is likely to be “2018+.” The anti-T—-p hatred is hotter. Unlike past elections, his demand for total devotion means that no GOPer’s have been able to run campaigns based on their alleged “independence” or bipartisan credentials. In November 2018 consumer inflation was 2.5% YoY, and gas cost $2.86/gallon. T—-p hadn’t gotten the US involved in a war; and the 2018 tax cuts, as opposed to last year’s Big Bad Billiionnaire Bust-Out Bill, didn’t come into effect until after the midterm elections.

So while in my head I am very cautious, in my heart I am expectiing a 1974-style shellacking. We’ll see.

What I will be watching:

1. Per Dan Guild, Florida counts its votes very quickly. Frequently you will know how the vote has been breaking by 8:30 pm eastern time, and almost certainly by 9 pm. While the gerrymanders are likely to work, there is a governor’s race, where the average poll has the Democrat losing by about 4%. This in contrast to the 58% to 40% blowout in 2022, and the 52% to 47% loss in 2018. So I will be looking to see how the governor’s race appears to be panning out compared with those benchmarks. 

2. Two Senate contests on the East Coast can also give us some early information: Susan Collins in Maine (D+3% on average) and Lindsay Graham’s sister in SC (D-1% on average). If either of those gets called early - and in which direction - it will give us a good sense of how other races are likely to break.

3. Three Congressional races in the Virginia Tidewater: VA-2, -1, and -5. VA-2, which includes lot of military voters in places like Virginia Beach, has been perennially close, on the order of 51%-49% to 53%-47% in favor of whichever party goes on to win nationwide. I will be looking to see who wins, and whether there is a breakout towards the Democratic side. The other two districts have been rated “toss-up” (VA-1) and “lean GOP” (VA-5). But the former averaged a 10%-13% in favor of the GOP incumbent in 2018, 2022, and 2024. In the latter district, the lean has been in favor of the GOP by 6% in 2018, and about 15% in 2022 and 2024. If either of these flip, it will be a very bad night for the GOP. If both do, it will be catastrophic. Even if they don’t flip, it the GOP margin is close enough, it will be a long night for the GOP nationwide.

4. A few outliers: Did voters in Michigan who responded to polls hide anti-MOOSlim feeling? If so, El Sayed might lose. Have White voters in the Deep South been enervated by the economy and/or corruption, vs. Black voters burning with homicidal rage at the gutting of the VRA? If so, there could be upsets where Black Democratic candidates in Mississippi, Arkansas, and/or Louisiana pull off victories.

We’ll find out soon enough.

September JOLTS report: another indication of an improved labor market

 

 - by New Deal democrat

This morning’s JOLTS labor market report for September, which parses labor market turnover by hires, layoffs, and quits, among other things, continued the string of positive trends in this sector this year, as all four major components moved in a good direction.

The first graph below shows the “soft statistic” of job openings (blue), actual hires (red), and quits (gold), all normed to 100 as of just before the pandemic:

Both hires and quits have had a sideways trend for the past two years. As noted above, in September all of these improved. Hires increased 131,000 to 5.402 million, and were up 3.4% YoY. Quits increased 321,000 to 3.365 million and were up 4.9% YoY. The “soft” metric of openings, which has had a slightly increasing trend this year, and outside of recessions for the past 25 years, increased 343,000 to 7.523 million, and were up 4.9% YoY. Nevertheless, both hires and quits remain below their pre-pandemic levels of 2019.

Here is the same graph for layoffs, which declined -127,000 to 1.641 million, and were down -9.6% YoY:

Layoffs have been running at close to their lowest post-pandemic levels over the past ten months, and their declining trend since late last year is of a piece with what we have been seeing in the very low level of initial and continuing jobless claims each week.

Let’s also update the comparison of the quits rate (blue, right scale), which has been suggested to lead YoY average hourly wages (red, left scale):

Like the number of hires and quits, for the past year the quits rate has been close to completely flat. That argues that *nominally* average hourly wages YoY should be stabilizing at a 3.4%-3.5% rate. 

Last month I concluded that “the general takeaway from this report is a jobs sector that has normalized in a low hire, low fire trend.” This month I can add that the trend, like other reports from the labor market, has been improving.

In the goods producing sector, AI continues to soar, and everything else continues to languish

 

 - by New Deal democrat

As usual, we start out the month with reports on two important sectors of the economy, manufacturing and construction, via the ISM manufacturing report (for October) and the construction spending report (for September). And the economy continues to look as it has all year, with an inflationary expansion in manufacturing, and lackluster construction with the huge exception of AI data centers.

Let’s start with the ISM manufacturing report. Between this report and the ISM services report which will be released on Wednesday, we get an excellent snapshot of the economy at the beginning of each month. The former in particular has been around for almost 80 years, and has had an excellent record. 

The headline index rose +1.0 for the month to 55.5. For forecasting purposes, I average the last three months, which comes out to 54.9 (recall that the dividing line between expansion in contraction is 50):

The more leading new orders subindex rose +0.2 to 55.5, although the three month average declined slightly to 54.8:

The upshot of these two numbers is that the manufacturing expansion can be expected to continue at least through the end of this year.

Recently the employment subindex in this report has also improved into expansion, albeit at a subdued level. This also continued this month, as it declined -0.3 to 52.4. The three month average also declined slightly, by -0.1, to 52.1:

Finally, the negative new on inflation also continued this month, as the prices paid subindex declined -2.7 to 75.2. The three month average, however, rose +1.4 to 74.8, rising even closer to its post-pandemic inflationary peak:

This report reminded me of the regional Fed reports for the last month, which suggested that while the expansion is intact, the second derivative may be declerating, i.e., the rate of growth has been peaking; and further that its inflationary component is accelerating, which is not good for downstream consumer prices.

Finally, keep in mind that for forecasting purposes, I weigh manufacturing at 25%, and the other 75% from the ISM services report, which has been running quite positive for many months.

If manufacturing is in good shape (except for prices), construction is not in good shape at all - except for AI data center construction. Nominally total construction (blue in the graph below) declilned -1.9% in September, and is down -2.7% YoY. The more leading residential construction sector (red) also declined, by -5.4% for the month, and is down -8.8% YoY. Since the price of construction materials (not shown) rose 0.7% in September and was up 17.0% YoY, the declines were even more negative. Here’s the absolute values nominally:

And here are the YoY% comparisons:

The particularly important negative is that residential construction, a long leading sector, continues to worsen.

The news was “less dismal” in manufacturing construction spending, which rose for a change, up 0.2% for the month, but remains down -19.2% YoY and about 30% from its September 2024 peak. The “relatively” good news is that this number has been running almost completely flat for the last three months, and is only down -0.6% since June:

Both residential and manufacturing construction remain recessionary, plain and simple.

Finally, nonresidential construction, which includes but is by no means limited to AI data center related construction, continued to rise, up 1.1% for the month, and up 1.8% YoY:

But this is misleading. While unfortunately I can’t show you a graph, while total nonresidential construction spending nominally increased $7.647 Billion for the month, and $17.406 Billion YoY, but AI data center related construction spending rose by more, up 8.282 Billion for the month, and up $45.752 Billion YoY. In other words, take out the AI data center Boom and nonresidential construction is continuing to decline as well.

My conclusion this month is the same as it was last month: “Basically the goods-producing sector of the US economy includes the AI Boom, which is continuing, and everything else, which is somewhere between outright recessionary to slightly expanding. And the expansion is very inflationary.”

Weekly Indicators for October 26 - 30 at Seeking Alpha

 

 - by New Deal democrat

My “Weekly Indicators” post is up at Seeking Alpha.

The long leading indicators have been trending towards neutral at last since the start of the Iran conflict. But what is new over the last few weeks i that the short leading indicators are also beginning to trend in that direction. They’re still positive, but less so than even a month or two ago - although two components, stock prices and jobless claims, as I’ve been documenting here weekly, are still very positive.

As usual, clicking over and reading will bring you up to the virtual moment as to the state of the economy, and reward me with a penny or two for my efforts.

A mixed bag of long leading indicators in the Q3 GDP

 

 - by New Deal democrat

As promised, here is my look at yesterday’s GDP report - which importantly was supplemented by this morning’s report of the employment cost index for Q3. More on that below.

Since my emphasis is on forecasting rather than dissecting the entrails of the recent past, I don’t have much to say about the contributions of inventories, imports and exports. For my one and only “nowcast” graph, here is the quarterly change in real GDP (blue), compared with “core” GDP, a/k/a real final sales to domestic purchasers (red):

These are pretty good numbers, unsurprising for an inflationary expansion.

But since there are two items in the GDP report that constitute long leading indicators, real private residential investment, a proxy for housing, and proprietors’ income, a placeholder for corporate profits which won’t be reported until next month, let’s take a look at each of them along with one other long leading indicators, real retail sales per capita.

First: real private residential investment as a share of GDP tends to peak about 7 quarters before the onset of a recession. Below is the 30 year history of real private residential investment:

Since investment declined slightly, I don’t even have to deflate by GDP, which increased. The point is, the housing component of GDP is a negative. Which, by the way, is consistent with what we are seeing with the other (and best) long leading indicator of housing, single family permits:

Second: corporate profits deflated by unit labor costs are also a long leading indicator. Because they don’t get reported until the second estimate of GDP next month, I use proprietors’ income, which typically tracks corporate profits sometimes with a one quarter delay, until then. Nominally these rose 0.9% in the quarter:

Now let’s turn to the deflator in this series, unit labor costs, which were reported this morning to have risen 0.8% in Q3. Significantly, the wage component rose 0.7%, while the benefits component (think in particular medical benefits) rose 0.9%:

On a YoY basis, unit labor costs were up 3.4%. But wages only increased 3.1%, while benefits rose 3.9%:

You can thank the Big Bad Billionaire Bust-out Bill for that.

All of which means that after adjustment, proprietors’ income rose only 0.1% in Q3:

While certainly not a downturn, this suggests growth in profits may be stalling.

Finally, financial indicators of interest rates and money supply dominate the long leading indicators. But along with the two discussed above, the other “real economy” long leading indicator is real retail sales per capita. And unlike profits and housing, these have been on a pronounced upturn, although in September they only rose 0.1%:

The past 15 years have been potent evidence that these three “real economy” indicators need to turn down; relying on the financial indicators alone is not sufficient. So to recap: housing in Q3 was negative, profits in Q3 were flat to slightly positive, while consumption per capita was soundly positive.

This is a mixed bag which is not consistent with a downturn in the next few quarters, although it does suggest the economy will have no more than a moderate expansion in real terms for at least the first half of 2027.

A somewhat negative personal income and spending report for September

 

 - by New Deal democrat

As promised, let’s take a look at the second of this morning’s three significant data updates. As I typically note, along with the employment report, real income and spending are probably the most important monthly reports as to the health of the average American household. Earlier this year there had been a real split between the income and spending sides of that ledger, with the former looking recessionary and the latter very positive. Last month the negatives on the income side were revised away, due to incorporating the more comprehensive measures found in the QCEW (through the first quarter of this year) and income tax payments (through Q3).

This month the softness reasserted itself somewhat, as real incomes declined, while real spending was unchanged. Here’s the more detailed look.

Nominally income rose 0.2% in September, and nominal spending increased 0.5%. But after adjusting for the price deflator, which rose 0.5% as well, real income (blue) declined -0.3% and real spending was unchanged:

Real income less government transfers payments is one of the important coincidental indicators the NBER looks at in determining if the economy is expanding or in recession. Since even without government transfers real income was down, needless to say so was this metric as well, also by -0.3% for the month, although it too was higher 0.5% YoY:

Here is what both real incomes and spending look like YoY since the pandemic:

The former is up 0.5% YoY, which is weakly positive, while the latter is up 2.5%, a normal expansionary positive number.

As I’ve noted many times in the past, real spending on services frequently rises all the way through recessions. The leading indicator is real spending on goods (gold). Further, real spending on durable goods (red) historically tends to peak even before goods spending as a whole.  In September real spending on goods declined -0.3%, and real spending on durable goods declined a sharp -0.8%

But both remain higher YoY, the former by 2.5%, and the latter by 3.2%:

The difference between income and spending is saving. In September, the saving rate declined -0.2% to 3.6%. This is low historically, but not as low as the era of the housing bubble and in 2022:

Typically, as expansions go on, consumers become more and more risk-taking about spending, and the saving rate declines. An increase in the saving rate in the presence of price pressures is one of the signs of an imminent recession. Needless to say, whatever they are telling pollsters, in the aggregate consumers remain quite confident about spending, likely tilted towards those with AI-based stock market gains. 

Finally, this morning’s report also enables the update of real manufacturing and trade sales, one of the other important coincident markers used by the NBER to date recessions. These lag by one month. They were also flat in August, but their uptrend remains intact:

They are higher 2.5% YoY.

This was a slightly negative report. Real incomes declined, real spending was unchanged, and the goods component of that spending declined. Consumers in the aggregate dug into their savings to continue that spending. But the overall positive trend in these numbers remains intact. For me to change that opinion would require real spending on at least goods or durable goods to turn negative.

Jobless claims continue near 60 year lows; compelling evidence that the unemployment rate is headed under 4%

 

 - by New Deal democrat

As I wrote yesterday, this morning we got a deluge of significant data. I’ll write about personal income and spending later, and the details of Q3 GDP tomorrow. For now, suffice it to say that real GDP was very positive, up 3.4% annualized, and “core” GDP, a/k/a real final sales to domestic purchasers, was even more strongly positive, at 4.9% annualized.

But let me start out with my usual weekly update of jobless claims, which make up 1/2 of my “quick and dirty” forecast, along with stock prices. They have been excellent almost all this year, and this week was one of their “most excellent” of all. Initial claims declined another -1,000 to 194,000, and the four week moving average declined another -1,250 to 196,750. Any time either of these numbers are under 200,000, not only are they among the best since the pandemic, but among the best in the entire nearly 60 year history of the series. Meanwhile, continuing claims, with the typical one week delay, declined -21,000 to 1.701 million:

And the YoY% numbers got even better as well, with initial claims down -12.2%, the four week moving average down -13.2%, and continuing claims down -12.9%:

In short, if real GDP, one of the premier coincident indicators, is very positive, then jobless claims are telling us that should continue at least through this Quarter as well.

Finally, as we close out the month, let’s take a look at what jobless claims forecast for the unemployment rate (red) over the next few months:

The last time the monthly average of claims were this low, which was early in 2023, the unemployment rate in the immediate subsequent months, was several tenths of percent under 4%. The evidence is becoming stronger and stronger that that’s where the unemployment rate is heading between now and the end of the year.

 

The AI data center driven manufacturing rebound continues

 

 - by New Deal democrat

While we await tomorrow’s deluge of data (Q3 GDP, personal income and spending, and jobless claims) let’s take a look at the yesterday’s most recent update of manufacturer’s new durable and capital goods orders from yesterday.

This comes out twice a month, as a first estimate and then as a final number for each month. Yesterday’s report was the first estimate for September. And it continued the strong rebound we have seen for over a year.

On a monthly basis, headline manufacturer’s new durable goods orders (blue) increased another 1.8%, while the less noisy core capital goods orders (red) increased 0.5%, to new post-pandemic high in the case of the less-noisy, more-signal latter:

On a YoY basis, the headline number increased 9.9%, and core capital goods orders increased a very sharp 13.5%:

What is interesting about the YoY comparison is that, if we use a paradigm that AI adoption will follow a typical “S” curve of exponential acceleration in growth followed by a similar exponential deceleration in growth” (note, we are talking about the second derivative here), then the fact that the core capital goods growth, while remaining very strong, may have reached peak acceleration at about 15%, then the first sign of the diminishment of the AI story may be near at hand (i.e., still strong growth, but decelerating to under 10% YoY).

Also interesting is that this inflection point of the second derivative may also be operative in the regional Fed new orders indexes that I updated last week. Here again is that graph:

This is noisier, so I look at the three month average, which remains close to its 3+ year peak.

The bottom line is that the AI-data center driven manufacturing rebound is very much intact at present. This suggests that consumer spending and goods-producing employment will also continue to be strong (in the case of the former) or rebounding (in the case of the latter). The inflationary expansion continues.

New home sales and repeat home prices may both have sung their ‘swan song’ for recent trends last month

 

 - by New Deal democrat

Last month I closed with “a reminder that the Fed’s interest rate hike and the similar increase in mortgage rates suggests that the recent equilibrium in the market is probably going to be yanked to the downside.” We have already seen that with mortgage applications, which have fallen to close to their 5 year lows. Next in line is sales, and only after that, with a lag, will be prices.

With that in mind, let’s take a look at this morning’s reports on new home sales and median prices, as well as repeat home sales as measured by both the FHFA and Case-Shiller indexes.

Let me start by repeating that new home sales are perhaps the most leading of all the housing sector metrics. But they suffer from being very volatile and heavily revised. So their three month moving average is more reliable as a signal.

In September sales (blue) declined -4,000 annualized to 676,000. But as the graph below shows, this remained close to the high point for this year, and was in the middle of the range for the past 3+ years:

The above graph also shows inventory (red). Although I won’t bother with the historical graph this month, sales have *always* led inventory, which tends to turn down much closer to a recession, and generally bottoms towards or even after the end of a recession. So the sideways to slightly higher trend this year would normally suggest a recovering housing market.

Prices are not seasonally adjusted, so are best measured YoY. And there, they have been in a sideways to slowly declining trend, which continued in September as the median price for a new home declined -1.0% for the month (dark blue), and -2.3% YoY. The below graph also shows the quarterly averages (light blue), which are down -1.0% YoY:

The trend in absolute numbers indicates a general decline in the past four years. Even more importantly, as shown in the graph below whicg deflates the median price of a new home by average weekly earnings, normed to 1 as of September, indicates that the “real” median price of a new home (dark blue) is down to its range in the 1990s and after the housing bust around the Great Recession. This contrasts with the “real” similarly deflated cost of existing homes as measured by the FHFA repeat sales index (orange), which while lower than several years ago, is still very high in historical terms:

Before I sum up, let’s turn to this morning’s other housing reports. Although they are not an infallible indicator (nothing ever is!),house prices are an important component of the economy, particularly as they affect the official CPI measure downstream, not to mention their importance to especially younger people in terms of the affordability of purchasing a new or existing home. And the repeat home sales indexes, by S&P Case Shiller and the FHFA, are the best indicator of prices in the 90% of the market that is existing home sales. same.

Since the beginning of this year, prices in both indexes have appeared to be firming, with the FHFA (red in the graphs below) as it often does slightly leading the Case-Shiller National Index (blue). That continued to be the case in this morning’s reports through August, with both reports indicating a seasonally adjusted monthly increase of 0.3% in prices. [Note: FRED has not yet updated the Case Shiller data]:

As per the above comment, earlier this year there had been something of a divergence showing in the YoY comparisons of the two national indexes, as the Case Shiller national index had increased less than 1% YoY, while the FHFA Index had accelerated to a 2.0% increase. But in the past several months, the Case Shiller index has followed suit. As of this morning’s report, the FHFA continued to be slightly “hotter” overall, with a 2.4% YoY increase, vs. 2.1% in the Case-Shiller Index.

It continues to be the case that it will take considerably more inventory on the market to bring existing homes down to just their average affordability compared with the past 30 years.

Both of this morning’s reports discussed above may have been the last for the current trends. Higher mortgage rates have badly hit mortgage applications, and can be expected to his sales and permits for construction within a month or two. Prices will only react after that, and the inventory of new homes in particular can be expected to increase again. If there is a silver lining, it is that builders have been steadily making their new homes for sale more affordable as a multiple of average earnings.

‘Level two’ chaotic systems, polling, and election results

 

 - by New Deal democrat

There’s no important economic data today, so let me post something that I meant to write about over the weekend before I was distracted by real life.

In his book “Sapiens,” Yuval Noah Harari writes that 

[M]any people [ ] prefer history to be deterministic. Determinism is appealing because it implies that our world and our beliefs are a natural and inevitable product of history….

[But h]istory cannot be explained deterministically and it cannot be predicted because it is chaotic…. [H]istory is what is called a ‘level two’ chaotic system. Chaotic systems come in two shapes. Level one chaos is chaos that does not react to predictions about it. The weather, for example, is a level one chaotic system….

Level two chaos is chaos that reacts to predictions about it, and therefore can never be predicted accurately. Markets, for example, are a level two chaotic system…. [because the better the history of a forecasting system, the more it will become a ‘nowcast,’ as people front-run the prediction].

Politics, too, is a second-order chaotic system…. Imagine … some political scientist in association with a computer wizard have developed an infallible algorithm [that predicts the political future. A politician on the losing end of that prediction, for example, would take pre-emptive measures to ensure that the prediction does not come to pass….]

(Sapiens, pp. 239-241)

This is exactly what happens with political polling, and the fate of poll aggregators is a case in point. Nate Silver was initially very successful. But soon, players with other interests reacted to his system. First, and with increasing frequency, partisan pollsters ‘flooded the zone’ with junk polls, forcing him to weigh pollsters by accuracy, and even to develop a slew of alternate and inconsistent forecasting systems. But those junk polls also influenced voters to decide that either a race was competitive, or that one candidate was sure to win or lose. And now we have seen the rise of outright fraudulent polling, to the same end.

The net result has been that these poll aggregations have been increasingly off-target. For example, Sam Wang famously said in 2016 that “It is totally over. If Trump wins more than 240 electoral votes, I will eat a bug.” (and to his credit, he ate a honey-dipped cricket on CNN four days after the election. 

What happened in 2016, essentially, was that “everyone knew” Hillary Clinton was going to win. And because “everyone knew” that, some voters felt free to stay home. Others felt free to cast a protest vote for Trump. Clinton herself felt free not to campaign in the “safe” upper Midwestern States. And most notoriously, FBI Director James Comey felt free to hold a press conference announcing a renewal of a criminal investigation into Clinton while early voting was already underway!

Now in 2026 we are being told that Democrats are heavily favored to take the House, and more likely than not to take the Senate. We are being told that early turnout is heavily in favor of Democrats, and that especially Hispanic voters are turning out in those areas.

But we are not just the passive recipients of these messages. We should fully expect that both potential voters and likely non-voters are going to *react* to that message as well. For example, might some Democratic voters in swing States relax? Will GOP operatives flood likely GOP voters’ mailboxes and social media feeds with horror stories about the big early Democratic turnout, thus driving more tenuous GOP voters to the polls to counter that early trend? Or conversely might a significant chunk of the public decide that they want to show up to be on the winning side, or turn a wave into a tsunami? 

We don’t know. But what we *should* know is that the very act of publicizing all these polls and early turnout models is going to significantly affect the ultimate result. Ask Brazil, where reporting and polling after the first round of balloting that Bolsonaro was almost sure to be elected drove a massive outpouring of votes for Lula.

For my part, I am hopeful. But I am thoroughly discounting all of these polls and dissections of early voting.

A century ago, early industrial psychologists ran a study of factory workers. They found that, e.g., increasing the light on the factory floor made the workers more productive. But it turned out that, more than anything else, it was the act of *observing* the workers that was the chief reason that the altered their behavior. Because, as a fan of Star Trek, to put Harari’s observation in Ferengi terms, “Hoo-mans are a clever species. When you observe the hoo-mans, they always observe back.”

Weekly Indicators for October 19 - 23 at Seeking Alpha

 

 - by New Deal democrat

My “Weekly Indicators” post is up at Seeking Alpha.

While the usual recent troika of stock market valuations, low initial jobless claims, and strong consumer spending are holding up the short leading and coincident forecasting time frames, enough deterioration has occurred in the long leading time frame to turn that neutral.  This doesn’t mean that the economy ‘will’ stall in the next 12 months or so, but it does mean that the conditions are in place for a stall to be a significant possibility.

As usual, clicking over and reading will bring you up to the virtual moment as to the state of the economy, and reward me with a penny or two in my pocket for gathering and organizing the information for you.

No, we are not DOOOMED: real usual weekly earnings edition

 

 - by New Deal democrat

As widely reported in the past week, real weekly earnings declined a sharp -2.1% in Q3. For example, here is Menzie Chinn at Econbrowser:

And here, more histrionically, is Mike Koncxal on X:

Is this an important danger sign?

No. Why? To cut to the chase, because even after seasonal adjustment it is the noisiest wage earnings metric there is (light blue in the graph below), and specifically in comparison with average weekly earnings as measured in the monthly jobs report (dark blue), and even more importantly in comparison with the quarterly employment cost index (red). Here’s what they look like since before the onset of the pandemic:

You can easily see that the blue bars not infrequently have both upside and downside quarterly swings well in excess of average weekly earnings, and positively wild swings in comparison with the employment cost index. Note, for example, that real weekly earnings had an outsized increase in Q2 of this year that almost completely cancels out their Q3 decrease. Perhaps more importantly, while both usual weekly earnings and the employment cost index are median numbers rather than an average like the monthly weekly earnings figure, the employment cost index norms by type of employment. Thus, in the immediate wake of the pandemic, many more low wage earners were laid off compared with higher wage earners, causing the other indexes to increase sharply even as unemployment soared as well. Then, as the pandemic gradually subsided and the low wage earners were hired back, those series decelerated back to a more “normal” change. But the employment cost index, which accounted for the change in the mix of employment, remained fairly steady.

As a result, when we measure the changes YoY rather than quarter over quarter, the huge swings tend to even out:

The monthly metric of average weekly earnings for all workers, averaged quarterly, is higher YoY on a nominal basis almost by the same percentage as median usual weekly earnings, but with much less drama.

And the same holds true on a historical basis. Note that the above graph, as well as all those below, show *nominal* changes, to account for the fact that different deflators may be used in the normal reporting of the series.

Here is 1979-99:

And here is 2000-2019:

Again, the “usual median weekly earnings” series is much more volatile, as you can see from the wider swings in the blue bars compared with the others.

This is not to say that Q3 real wages for ordinary workers have been good. As I’ve reported for a number of months now, on a YoY basis, real earnings have been flat or slightly negative. But take the noisy Q3 report with multiple grains of salt. It is likely to revert to the mean compared with the other reports in Q4.