Update: Major Trends in New Jobless Claims

What's the real story for the overall trend in new jobless claims in the U.S.?

It has been several months since we last featured an update of our statistical analysis of the trends in new jobless claims in the U.S., which is mostly by design on our part.

When we last left off, the overall trend, which we identified as Trend L, was largely flat, but characterized by extreme volatility, with unusual outliers in the data popping up with an alarming frequency.

Some of that was understandable. We saw the state of California mishandle its processing of new jobless claims, causing their reported number to plummet one week, but soar in the next as the state caught up to their mistakes. We saw the impact of the "superstorm" Sandy, which caused claims in the northeastern part of the U.S. jump well above the trend, then slowly fall back toward it, then fall below it, before rebounding back to the otherwise flat trend level established early in the trend as that region of the country recovered from the event.

And then, several weeks after that, we started seeing volatility to the downside. The data would suddenly fall outside the statistically expected range for one or two weeks, but then be right back up in it the next.

That was more volatility than we wanted to have to keep addressing on weekly or biweekly basis, so we opted to let the data settle down on its own and pick up the pieces later.

And that brings us today, because it turns out that what was happening is that Trend L was breaking down and being replaced by a new trend, which we'll unimaginatively call "Trend M".

What appears to have happened is that the number of initial unemployment insurance claims being filed each week "stepped down" from the higher level defined by Trend L's mean level of 367,915 to a lower level defined by Trend M's mean level of 344,402.

Part of why we waited so long to discuss this shift is that we wanted to have enough data points to be able to describe the new trend that was establishing itself. Typically, it takes at least six to ten data points to do that, but the appearance of an outlier for the week ending 30 March 2013 led us to choose to hold off until we had at least 10 consecutive non-outlier data points. After that, it was just a matter of waiting for an opening in our schedule of posts.

Aside from the one outlier, which we recognize to be an artifact of the Department of Labor's recently revised seasonal adjustment factors for the week in question, Trend M has largely been flat since it established itself after 23 February 2013. While media reports have indicated that the trend is improving, that assessment is based on a four week moving average, which really capturing the data becoming less volatile. While that is an improvement, it hasn't affected the overall trend, which is largely just moving sideways.

Speaking of which, Trend M is characterized by a surprising amount of variation, which we typically have not seen in the data for our residual analysis, and which is the result of relatively wide swings in the data with respect to the general trend from week to week. We should also note that the current flat trend is still well elevated above the 317,000-318,000 new jobless claims filed on average each week in the years preceding the December 2007 recession.

The following chart shows each of the primary trends that we have identified in new jobless claims since 7 January 2006.

The next chart shows the residual distribution of the trends in new jobless claims over the same period (the first chart in this post focuses on the most recent trends identified in this chart.

Finally, since we've marked the formation of a new trend, we've updated our table describing each of the ones we've covered since 2006, along with more discussion of our thoughts on what's driving them:

Timing and Events of Major Shifts in Layoffs of U.S. Employees
Period	Starting Date	Ending Date	Likely Event(s) Triggering New Trend (Occurs 2 to 3 Weeks Prior to New Trend Taking Effect)
A	7 January 2006	22 April 2006	This period of time marks a short term event in which layoff activity briefly dipped as the U.S. housing bubble reached its peak. Builders kept their employees busy as they raced to "beat the clock" to capitalize on high housing demand and prices.
B	29 April 2006	17 November 2007	The calm before the storm. U.S. layoff activity is remarkably stable as solid economic growth is recorded during this period, even though the housing and credit bubbles have begun their deflation phase.
C	24 November 2007	26 July 2008	Federal Reserve acts to slash interest rates for the first time in 4 1/2 years as it begins to respond to the growing housing and credit crisis, which coincides with a spike in the TED spread. Negative change in future outlook for economy leads U.S. businesses to begin increasing the rate of layoffs on a small scale, as the beginning of a recession looms in the month ahead.
D	2 August 2008	21 March 2009	Oil prices spike toward inflation-adjusted all-time highs (over $140 per barrel in 2008 U.S. dollars.) Negative change in future outlook for economy leads businesses to sharply accelerate the rate of employee layoffs.
E	28 March 2009	7 November 2009	Stock market bottoms as future outlook for U.S. economy improves, as rate at which the U.S. economic situation is worsening stops increasing and begins to decelerate instead. U.S. businesses react to the positive change in their outlook by significantly slowing the pace of their layoffs, as the Chinese government announced how it would spend its massive economic stimulus effort, which stood to directly benefit U.S.-based exporters of capital goods and raw materials. By contrast, the U.S. stimulus effort that passed into law over a week earlier had no impact upon U.S. business employee retention decisions, as the measure was perceived to be excessively wasteful in generating new and sustainable economic activity.
F	14 November 2009	11 September 2010	Introduction of HR 3962 (Affordable Health Care for America Act) derails improving picture for employees of U.S. businesses, as the measure (and corresponding legislation introduced in the U.S. Senate) is likely to increase the costs to businesses of retaining employees in the future. Employers react to the negative change in their business outlook by slowing the rate of improvement in layoff activity.
G	18 September 2010	2 April 2011	Possible multiple causes. Political polling indicates Republican party could reasonably win both the U.S. House and Senate, preventing the Democratic party from being able to continue cramming unpopular and economically destructive legislation into law, bringing relief to distressed U.S. businesses. Fed Chairman Ben Bernanke announces Federal Reserve will act if economy worsens, potentially restoring some employer confidence. The White House announces there will be no big new stimulus plan, eliminating the possibility that more wasteful economic activity directed by the federal government would continue to crowd out the more effective economic activity of U.S. businesses.
H	9 April 2011	26 November 2011	Rising oil and gasoline prices exceed the critical $3.50-$3.60 per gallon range (in 2011 U.S. dollars), forcing numerous small businesses to act to reduce staff to offset rising costs in order to prevent losses. The trend ends when average motor gasoline prices in the U.S. fall back below the $3.50 level in the week between 5 November 2011 and 12 November 2011 - the corresponding improvement in business outlook shows up in the data with the next full pay cycle (2-3 weeks later, or rather, the week ending 26 November 2011!)
I	3 December 2011	11 February 2012	With average gasoline prices in the U.S. having fallen below the critical $3.50 per gallon level, employers respond to the improving business outlook by reducing the number weekly layoffs at a faster rate, as both businesses and consumers benefit from lower transportation and fuel costs, while consumers gain more disposable income. Trend I ended shortly after gasoline prices rose back above the $3.50 per gallon mark in late January 2012.
J	18 February 2012	23 June 2012	With average gasoline prices continuing to be a high levels through the spring and summer, the pace of layoffs in the U.S. steadily increased until June 2012, when the national average price of gasoline in the U.S. finally dropped back below the $3.50 per gallon mark.
K	30 June 2012	15 September 2012	Trend K began with a sudden shift downward in the number of new jobless claims as gasoline prices fell below the $3.50 per gallon mark in June 2012, and although the average price of gasoline in the U.S. has since risen back above that level, there has been no sudden upward shift in new jobless claims. Instead, the number of initial unemployment insurance benefit claim filings has been rising at a faster rate than at any time since on the onset of the 2007 recession.
L	23 September 2012	16 February 2013	A surging housing market beginning in July 2012 followed by the Fed's action to aim its QE efforts directly at the industry on 13 September 2012 helped averted a developing recession in the U.S. at the end of 2012 as it stabilized the number of new jobless claims filed each week. Still, Trend L was characterized by a having a tremendous number of outliers as compared to previous trends, primarily due to the impact of Hurricane Sandy.
M	23 February 2013	Present	Trend M began with a sudden shift downward in the number of new jobless claims beginning in mid-January 2013, which became a definitive shift by mid-February. With gasoline prices elevated at the time, our thinking is that other factors, such as the Fed's amping up QE 4.0 on 12 December 2012 to offset the negative effects of tax hikes they believed would take effect in 2013, are responsible for the downward shift in the number of claims. Here, we think the Fed's monetary policy succeeded in offsetting the fiscal drag from the tax hikes that did take effect in 2013, which is a good thing because the alternative would have been the onset of a full-bore recession, where we would be watching the trend in new jobless claims rise instead.

Forecasting 2013-Q2's GDP

Now that the BEA has released its massive revision of GDP data going back to 1929, it's now worth our time and trouble to forecast where real GDP for 2013-Q2 will be finalized.

To do that, we'll be taking the data that the BEA has reported for the three most recent quarters for which it has finalized its estimates of the U.S.' real gross domestic product and applying our "modified limo" technique for forecasting the nation's GDP for the second quarter of 2013. While the BEA has already put forward an initial estimate of real GDP in 2013-Q2 of $15,648.7 billion in terms of chained 2009 U.S. dollars, that's really just their first estimate for the figure. What we're really looking to forecast is the value that will be recorded in the BEA's third estimate, which won't be reported sometime late in August 2013.

Since our modified limo method only incorporates finalized data in its formulation, and not initial estimates, the value of the first estimate that the BEA has already reported will neither affect nor alter our forecast value. If we had known how the BEA was going to revise GDP to incorporate the relatively intangible elements of the economy such as research and development activities in the private sector, we could have obtained the results we'll present today over a month ago, long before the BEA published its first estimate for real GDP in the second quarter of 2013.

That's because our modified limo forecasting technique is really an inertia-based method for projecting what real GDP will be in the future, which is based upon the very recent performance of the economy. It is most accurate when the economy is growing at a steady rate, at least, as measured over two-quarter long periods of time, and even provides insight in those cases when it is off target, as that indicates a situation where the economy is turning a corner, so to speak, either positively toward expansion or negatively toward contraction.

Let's do some math, shall we? We've entered the data for the three most recent quarters for which the BEA has revised its data at least three times already (2012-Q3, 2012-Q4 and 2013-Q1):

Previous Quarters Inflation-Adjusted GDP Data
Input Data	Values
Three Quarters Ago (Billions of Chained U.S. Dollars)
Two Quarters Ago (Billions of Chained U.S. Dollars)
Most Recent Quarter (Billions of Chained U.S. Dollars)

Next Quarter's Inflation-Adjusted GDP Estimate
Calculated Results	Values
Modified Limo Estimate (Billions of Chained U.S. Dollars)

From that data, we find that the midpoint for our forecast range for real GDP in 2013-Q2 is $15,609.0 billion in terms of chained 2009 U.S. dollars, which is nearly $40 billion, or 0.25%, below the BEA's first estimate for the quarter, as the early indication that the U.S. economy performed slightly better in 2013-Q2 than a simple inertial trajectory would project.

We find the following probabilities for where GDP will be finalized for the quarter:

• There is a 68.2% probability that real GDP will fall between $15,437.8 and $15,754.3 billion.


• There is a 95.0% probability that real GDP will fall between $15,292.3 and $15,925.6 billion.


• There is a 99.6% probability that real GDP will fall between $15,133.9 and $16,084.0 billion.

And of course, there is a 50% probability that it will be finalized at a value above our midpoint forecast value of $15,609.0 billion. And actually better than that, as the BEA's first estimate for the quarter indicates it is above that value.

The chart below visualizes our forecast, and gives the first hint of what 2013-Q3 might look like, if 2013-Q2's GDP is finalized at its currently estimated value of $15,648.7 billion:

Although the early data favors real GDP being recorded above our forecast midpoint value, that's really not saying very much, as the overall level of economic growth suggests that the U.S. economy is experiencing recessionary conditions, and has been since the third quarter of 2012, as confirmed by our GDP growth rate temperature gauges shown at the top of this post!

But then, while that may be a surprise to people who believed the BEA's pre-revision estimates of GDP through 2013-Q1, which didn't indicate much potential of recession, it's old news to our readers, who know that we have other, more reliable methods than the federal government has for telling when the pace of economic growth has become too cold.

Dividends: U.S. Economy Exiting Recession

The number of public U.S. companies announcing that they would be cutting their future dividend payments fell from 20 to 14 in July 2013. While this level of dividend cuts announced in a single month is still consistent with recessionary conditions being present in the U.S. economy, it does represent an improving situation as it appears that the U.S. economy is beginning to exit the microrecession that began back in July 2012.

By our definition, a microrecession represents a period of slow or negative economic growth for the nation that is either too limited in scope, severity or duration to qualify as an official period of recession, as might be determined by the National Bureau of Economic Research.

We'll note that the Bureau of Economic Analysis' recently revised GDP data now confirms what our dividend indicator has been communicating in real time - the pace of economic growth in the U.S. has been in the "cold" recessionary zone, as indicated by our two-quarter GDP growth rate temperature gauge:

Prior to its revision, the BEA's GDP data had indicated that the U.S. economy was growing at a somewhat faster pace.

As best as we can tell, it appears that the microrecession peaked in the fourth quarter of 2012. The U.S. economy would now appear to be exiting this period of microrecession, which has now lasted for over a year.

References

Standard and Poor. Dividend Action Report. [Excel spreadsheet]. Accessed 5 August 2013.

U.S. Commerce Department Bureau of Economic Analysis. Current-Dollar and "Real GDP". [Excel Spreadsheet]. Accessed 31 July 2013. - Note: Inflation-adjusted figures are actually presented in terms of chained 2009 U.S. dollars (not chained 2005 U.S. dollars).

Summer Jobs 2013: Zero New Jobs for Teens

The most remarkable thing was documented in the July 2013 employment situation report. There was absolutely no change in the number of employed teens in the United States from June to July 2013.

Zero. Nil. Zip. Nada. Zilch.

Meanwhile, the number of young adults Age 20 through 24 who were counted as having jobs increases by 49,000, while the number of adults Age 25 or older rose by 178,000, for a total month-over-month increase in the number of Americans with jobs of 227,000, which was enough to lower the official unemployment rate in the U.S. to 7.4%.

That increase puts the total number of employed Americans at 144,285,000. Of these, 126,162,000 are Age 25 or older and 13,654,000 are between the ages of 20 and 24.

Meanwhile the number of teens with jobs in July 2013 is 4,469,000. Which just happens to be exactly the same number of U.S. teens that were reported as having jobs in June 2013.

It would seem that President Obama has once again failed to help teens get summer jobs, as both he and a large number of Democratic Party mayors across the nation who partnered with the President this past April appear to have been completely impotent in their efforts to create any notable improvement in summer job opportunities for U.S. teens in their communities this summer.

That 11 April 2013 discussion would appear to be the last time that President Obama paid any attention to the stagnant employment situation for U.S. teens, as he certainly isn't considering teens at all in his latest pivot to jobs.

Which means that zero change in the number of teens with jobs this summer fully reflects the level of effort that President Obama has exercised on their behalf, which in turn tells us quite a bit about the priority he places upon their plight.

If only the President cared about American teens more....

S&P 500 Hits 1700 A Month Late

On 1 August 2013, the S&P 500 closed above 1700 for the first time in its history.

That was almost exactly one month later than it should have. Let's revisit one of our now rare public predictions for stock prices, from back on 10 June 2013:

Overall, it was a really boring week, but at least it set up a fun, seeming paradox for our favorite chart. Given the nature of how the moving average of the change in the growth rate for stock prices works, we'll soon see the 20-day moving average for the S&P 500 top out then fall back toward the green line (where it really wants to be), even as the value of the S&P 500 itself rises.

In the relative absence of noise in the market, that would create the potential for the S&P 500 to reach 1700 before the end of June 2013.

But then, noise is always present in the market. It's its volume that's what's random and unpredictable!

A little something we like to call The Bernanke Noise Event intervened in that forecast for the future, throwing the market off track for a whole month, the recovery from which was then drug out a little longer by a poorly considered trial balloon put out by the White House proposing Larry Summers as the leading candidate to be nominated to run the Fed after Ben Bernanke's current term expires in January 2014.

But that new noise began to dissipate when the White House gauged the market's negative reaction and began to float the name of an alternative candidate on Wednesday, 31 July 2013, as the Fed also continued its efforts to repair the damage done by Bernanke back on 19 June 2013. By the next day, enough of the noise of Summers had decreased to allow the S&P 500 to finally close above the 1700 mark.

The chart below shows where we're at today. Mind the comments in our chart above and note the differences from previous versions (hey, if the Fed's Open Market Committee can get by on tweaking the otherwise same statement from meeting to meeting, so should we be able to get away with doing the same thing on our S&P 500 stock prices and dividends per share acceleration chart!)

Previously on Political Calculations

Unless it becomes relevant as events develop, this is probably the last time that we'll feature the following chain of analysis in considering the major market drivers this summer.

• How Stock Prices Work - The basic theory for why "Which Future?" is the most important question for investors to answer.


• The Math Behind How Stock Prices Work - Just as advertised. Welcome to chaos!


• Whither Dividend Futures - we tell you exactly where we get our data. Between this and our math (linked above), you can do this kind of cutting edge analysis too....

Next, that electronic trail of analysis we've provided throughout the event:

• The World Investors and the Fed Live In Now - Our snapshot of the market right before the event, in which we note that investor concern about the future of QE was growing and remark that there will be a market reaction in response to the outcome of the Fed's two-day meeting later that week.


• The Bernanke Noise Event - as the Summer of 2013 shall ever be known to investors....


• Now Is It Time to Sell? - according to statistics, a quaint branch of mathematics that only works to describe how stock prices vary with respect to their trend when order is present in the market. The problem with it is that the market goes in and out of order, so it's periodically pretty useless....


• The Fed's Real QE Mistake: Timing - We explain how Bernanke really screwed up.


• Now What Will You Do? - the statistical line is crossed! We look at everything that we see screaming "sell", without actually saying it's time to sell.


• The Fed Attempts to Walk It Back - we anticipate how the Fed will respond to Bernanke's error, and we determine if it will work.


• The GDP Multiplier for QE - Not about investing, at all! Instead, we explain why sustaining QE at current levels is so important to the U.S. economy at present.


• "Never Bet Against the Fed" - we visually illustrate that the Fed's response to repairing the damage from Chairman Bernanke's blunder is working and recap why fears of stock market doom, despite signals to the contrary, were really overblown.


• Bernanke Closes the Gap - we mark the end of the Bernanke Noise Event.


• The Noise of Summers - we note the beginning of a new negative noise event....

Inflation and the Batman

This is cool - via Core77, Mashable has determined the cost of living for a number of superheroes in the year they first hit the scene and today, and presented the information in the form of a comic book cover:

Also featured: Superman, Spider-Man, The Hulk and Wolverine in addition to Batman!

Visualizing the GDP Revision

How did the overall level of reported GDP change as a result of the BEA's finally including research and development (R&D) and also arts and entertainment (A&E) production as part of U.S. GDP?

Here, both R&D and A&E are now being accounted for by the BEA as fixed assets rather than as intermediate components of producing goods and services, which effectively boosts the reported level of GDP.

To find out, we first converted the pre-revision data from being in terms of inflation-adjusted, chained 2005 U.S. dollars to be in terms of inflation-adjusted, chained 2009 U.S. dollars, which we've presented as the green line in our chart below. We next charted the newly revised real GDP figures being reported by the BEA as the black line, then drew lines between the various quarterly data points to help visualize how much the data changed from the previous revision.

In this chart, which presents all of the BEA's quarterly data for GDP from the first quarter of 1947 through the first quarter of 2013, we find that R&D and A&E production had very little impact on GDP up until 1983, when it finally reached a share of 1% of real GDP. Ten years later, that share of GDP had doubled, and through the first quarter of 2013, has now reached a share of 3.5% of the real U.S. economy.

Our next chart focuses on the data since the first quarter of 2000.

The main news for this period is that with the revision, the 2007 recession was not as deep as previous data indicated, and the recovery was somewhat stronger.

But perhaps the most surprising revision in the data applies for the first quarter of 2011, when GDP appears to have contracted, as the data had previously indicated slightly positive growth. White House economist Alan Krueger blamed the economic shrinkage on the disruption to world supply chains stemming from the Japanese tsunami.

That particular event occurred on 11 March 2013, just 19 days before the end of the first quarter of 2011. In reality though, most of that effect wouldn't show up in the U.S. economy until the second quarter of 2011, since the transit times between the U.S. and Japan for cargo are such that any trade between the two that would be recorded in the economic statistics for the first quarter in 2011 would already have been en route at the time of the tsunami and would only need to be re-routed to non-damaged port facilities in Japan, an observation that is backed by trade data recorded for both March and April in 2011. Meanwhile, plants shut down in Japan as a result of the event would not affect production at U.S. facilities for several weeks following the event, and many U.S. facilities of affected Japanese companies, which would be the most likely to be impacted, continued production without any stoppages.

Since most of the adjustment related to the BEA's data revision affects GDP data since 1980, we thought it might be fun to present large-scale versions of our GDP temperature gauge charts. The first of these charts shows the one-quarter real annualized GDP growth rate for each quarter from 1980-Q1 through the first estimate of the GDP growth rate for 2013-Q2:

Our next chart presents the annualized two-quarter real GDP growth rate for each quarter from 1980-Q1 through 2013-Q2. Since this data is much less subject to volatility, we use it for forecasting, as it provides a more accurate picture of the pace of economic growth in the United States over time:

These charts provide an indication of how strong the U.S. economic growth rate is, which can range from a "hot" red (growing at a rate greater than 5.5%) to a recessionary "cold" purple (growing at a rate less than 1.5%).