Friday, July 23, 2010

Weekend Weimar and Beagle

It's that time of the week again. Don't think about the market at all for the next few days. We'll be back on Monday.

Until then ...

Weekly Indicators: One tiny cheer for Housing edition

- by New Deal democrat

This week we finally got some good -- in the sense of being un-bad -- news about housing. Housing Permits for June actually improved slightly over the dismal May. Purchase Mortgage applications took a tiny step upward from the deep sub-basement they have fallen into. As expected, LEI's declined significantly for the first time in over a year.

In addition to slight improvement in purchase applications, refinance applications also continued to improve. It does seem that quite a few people are able to take advantage of the extremely low rates being offered. This bodes well for households continuing to shed debt, and to be able to save at the same time.

The ICSC reported same store sales for the week ending July 16 rose 4.2% vs. a year earlier, the best showing in over two months. They were also up 1.4% WoW. Once again, Shoppertrak did not issue a public report this week.

Gas stabilized at $2.72. It has been in this range for going on two months. The 4 week average of usage remians up substantially (about 2%) from last year.

The BLS reported 462,000 new jobless claims this week. This was up substantially from last week. This data series is particularly noisy and difficult to read now, with both the auto plant non-closures and filings by laid off census workers figuring into the mix.

Railfax gave up another poor report. Cyclical traffic remains substantially up from last year, but is declining whereas last year it was improving. Baseline traffic is actually slightly worse than last year. Only intermodal traffic, signalling imports and exports, continues to show improvement. Even compared with last year, motor vehilce loads have fallen off a cliff. By the way, they have premiered some excellent interactive new graphs that you should peruse.

Interestingly, the Railfax auto data is at odds with, which is forecasting that July bargain hunters will push auto sales up to a SAAR of 12 million, it's best showing ex-C4C in nearly two years.

The American Staffing Association failed to report this week.

There is unabashed continued good news in the
Daily Treasury Statement. July is continuing the pattern of improvement over last year's numbers, $95.9 B vs.$88.5 B last year, a gain of over 8%. For the last 20 reporting days, we are also up 11%, $131.0 B vs. $117.7 B. This series has been kicking butt on a YoY basis for the last seven weeks, even with about 75% of the census layoffs having taken place at about the same time. I continue to suspect that there is more strength in payrolls than has been showing up in the official numbers for May and June.

M1 was down 1.5% this week, but on a YoY basis for the month of July so far, is up 3% (meaning “real M1” is up about 2%). M2 is up 0.2% this week, or about 1.8% YoY for July so far (meaning “real M2” is up about 0.7%). A reminder that recessions have in the past coincided with a negative real M1 and real M2 less than 2.5%.

This was the best week we've had in a while for the short-term data -- but that is mainly because nothing awful happened.

The 2001 Recession: Employment Issues

We now know that the 2001 recession was caused by three events: a spike in oil prices, 9/11 and the stock market crash. Let's look at how this played out in the jobs market.

Notice the unemployment rate continued to increase after the recession, although it didn't increase to incredibly high levels. However, it hovered around approximately 5.75%-6.25% for almost two years after the end of the recession.

Above is a chart of total non-farm payrolls for 2000-December 2003. Simply eyeballing the chart, notice that total numbers decreased from approximately 132.4 million to a little below 130 million. That means total jobs lost were about 2.4 million.

Above is a chart of total durable goods manufacturing. Eyeballing this chart, notice that total employment fell from about 10.8 million to about 8.8 million -- or approximately 2 million jobs. That means that durable goods manufacturing job losses were the primary reason for the job losses of the early 2000s recession.

Above is a chart of total durable goods manufacturing employment for the early 2000s expansion. Notice that unlike the 1990s expansion, there jobs did not come back.

Above is a chart of productivity. Notice that despite the decrease in durable goods employment, productivity (output per man hour) continued to increase. In addition,

Above is a chart of total durable industrial production. Notice it started to increase at the end of the 2001 recession.

So -- the second jobless recovery was caused by a drop in durable goods manufacturing employment during and immediately after the recession which was not followed by a rehiring of durable goods employees. This is in direct contrast to the 1990s expansion when durable goods employees were eventually rehired. However, this decrease employment was not accompanied by a drop in manufacturing output. In fact, manufacturing output continued to increase, much as it had in the 1990s expansion.

The 2001 Recession

Let's continue with a look at another "jobless recovery" -- the recovery after 2001 which the NBER dates as 3/01 - 11/01. First, here are the charts that started this inquiry:

Initial unemployment claims remained at an elevated level for over a year after the recession ended. In addition,

The unemployment rate remained increased after the recession ended.

Let's take a look at the overall GDP picture:

The recession was actually very mild. There were only two quarters of negative growth -- the first quarter of 2001 where growth shrank 1.3% and the third quarter of 2001 when growth shrank 1.1%. However, notice that two of the quarters preceding the recession printed low GDP growt rates of 1.1% and .3%. A third quarter printed a growth rate of 2.4%. Also note the very weak growth rates after the recession. The median growth rate for the five quarters after the recession ended was 2.00% -- a subpar rate of growth. It is probably best to consider the entire three year period as an era of sub-par growth as the median GDP growth rate was 1.6%.

Let's note the causes of the recession:

1.) The stock market crash

2.) 9/11

3.) A spike in oil prices (note the spike on both sides of 2000):

Let's take a look at the various components of GDP both before and after the recession.

I've broken PCEs down into three areas. First, notice that in the four quarters before the recession, they were growing over 3%/quarter. During the recession (with the exception of 4Q2001) they were growing below 2%. Finally, notice that for the period of the recession and the five quarters after the recession, PCEs were growing but still at a subdued pace.

Gross private domestic investment was very weak for the four quarters before the recession. During the recession they numbers turned sharply negative, largely as a result of a large decrease in business investment. Investments increased for the two quarters after the recession ended, but then turned lower, advancing very little.

Exports and imports turned negative the quarter before the recession and turned sharply lower during the recession. After the recession, both increased at decent rates for the first two post-recession quarters, but thereafter started to weaken.

Government spending increased on a quarter to quarter basis during the recession, but then decreased afterwords.

Yesterday's Market

Industrial metals -- which had been trading in a range for the last month and a half, may have finally broken out. Ideally, we'd like to see a stronger candle on a move like this. However, prices have moved above the 50 day EMA for the first time in the last three months. We'll have to wait for follow-through, however.

Both the SPYs and QQQQs are in the same situation -- prices are right at resistance levels and the EMAs are in a tight range. The last time both averages were at these levels (right below the 50 day EMA), they retreated.

Notice how the long-end of the Treasury curve -- the TLTs are just hugging the trend line right now.

The 7-10 year part of the curve (the IEFs), however, are a long way from the trend line.

Gold continues to move lower, hugging its downward sloping trend line. Considering inflation is not an issue right now and Europe appears to be solving its debt problems, expect more downward pressure on gold.

Notice that oil closed above the 50 day EMA for the first time in nearly three months. Also note the last two times oil was here it retreated.

Thursday, July 22, 2010

LEIs decrease

From the Conference Board:

“The indicators point to slower growth through the fall,” says Ken Goldstein, economist at The Conference Board. “Two trends will have a direct impact on the pace of economic expansion. First, improvement in the industrial core of the economy will moderate as inventory rebuilding slows. Second, improvement in the service sector has been relatively slow, with little indication that it will pick up momentum.”

“The LEI decreased in two of the last three months, but its level is still about 4.5 percent above its previous peak before the recession began,” says Ataman Ozyildirim, economist at The Conference Board. “Moreover, the gains among the LEI components have been widespread, with the exception of housing permits and stock prices, pointing to an expanding economy, but at a slower pace in the second half of the year.”

Let's look at the data:

The chart clearly shows a slowing of the trend, but there is no sign of an imminent death spiral.

Note the primary reasons for the drop was a decrease in the workweek and supplier deliveries.

The Early 90s Recession: Employment Issues

First, let's go back to the charts that started this analysis.

Initial unemployment claims remained high for over a year after the end of the 1990s recession. In addition,

The unemployment rate continued to increase after the end of the early 90s recession. Let's look a little deeper into the employment statistics to see what jobs were lost.

Above is a chart of total non-farm employees for the years 1989-1994. Notice a few points.

1.) From an employment perspective, the recession was mild -- the economy only lost 1.621 million jobs.

2.) The recession ended in March 1991, yet total employment did not reach pre-recession highs until early 1993 -- nearly two years after the recession ended.

Let's divide the economy into manufacturing and service sectors.

Above is a chart of total manufacturing employment. First, notice it started declining in roughly early/mid 1989 -- a full year before the recession started in 7/90. Notice how it continued to decline after the recession ended in 3/31. Finally, notice that over three years after the recession ended, the level of employment was far below the pre-recession high. From the high on this chart to the low of this chart manufacturing lost 1.320 million jobs. By the end of this chart (12/94), total manufacturing employment was down 842,000 from its pre-recession highs.

Compare the total manufacturing job loss totals to the service sector job loss totals. At their worst, service sector employment dropped by 395,000 in the early 1990s recession. In addition, total service employment had returned to pre-recession levels by early 1992 -- a year after the recession ended.

Let's add two more charts: employment for durable and non-durable employment:

Remember that total manufacturing employment decreased by 1.3 million over the recession at maximum employment losses. Total employment of durable goods manufacturing contracted over 1 million, meaning it bore the brunt of the losses.

Non-durable manufacturing lost the obvious remainder of jobs. with losses totaling a little over 150,000.

Above is a chart of durable goods manufacturing employment for the entire recovery. Notice it took nearly the entire expansion -- 6 years -- before the durable goods sector returned to near the level it reached before the recession.

Above is a chart of output per hour of manufacturing employees. Notice that despite the decrease in employment that lasted after the recession ended, output per hour increased. This is one of the main reasons manufacturing employment did not rapidly increase at the end of the recession.

Above is a chart of service sector employment for the duration of the 90s expansion. Notice this is where a bulk of the job creation occurred. However, the recession ended in 3/91, year employment did not ramp up until the end of 92, over a year after the end of the recession.

S0, the first jobless recovery was primarily caused by a massive contraction in durable goods manufacturing. These jobs did not come back until nearly the end of the expansion -- over 8 years after the end of the previous recession.

Yesterday's Market

First, note how strong the 7-10 year rally is. The uptrend (a) has been in place for nearly three months. The EMAs (b) are incredibly bullish -- all are rising, the shorter are above the longer and prices are above all. The MACD is at worst neutral -- remember -- the line measures the difference between two moving averages. If both continue to move higher at the same rate, the indicator will flatten. The A/D line shows a slight move out of the security, but not in a big way. Simply put, the only thing constricting a rally is yield.

With both the SPYs and the QQQQs, notice that with the exception of one move up and down, prices are remaining in a range. With the SPYs, we're between 105.50 and 110 and with the QQQQs its 43.60 and 46.60.

However, the IWMs and IWCs are still clearly in a downtrend, indicating risk based areas of the market are still considered dicey.

Agricultural prices have clearly broken two major downtrends.

The EMA picture is also bullish (a) -- all the shorter EMAs are rising and prices are above all the EMAs. Also note that prices are using the 200 day EMA as technical support (b).

Wednesday, July 21, 2010

DeLong on One of the Basic Problems We Face

From the FT:

I see an economy in which there is enormous slack pretty much everywhere – empty retail storefronts in Berkeley just to my left, anyone? – in which even the US housing stock is no longer above its trend, and in which we are currently building houses at half the trend pace. If output in even our single-family residential-housing sector is significantly depressed below its steady-state growth value – if, economy-wide, 10 per cent of the spending that ought to be there is missing – then we need not policies that carefully create new jobs only in the appropriate sectors but instead policies that create new jobs pretty much anywhere.

Exactly right. High unemployment low capacity utilization mean little to no inflation and the need for demand side stimulus.

A Closer Look At the Early 90s Recession and Recovery

Before we move into the recovery, a commenter also noted that oil prices increased before the recession because of Iraq's invasion of Kuwait. In addition, with the then possibility of war, there was added tension in the economy which contributed to the slowdown.

First, here is a chart of oil prices with the spike outlined:

Notice the price spike the occurred at the same time as the beginning of the recession (7/90). Oil shocks are a very big issue in the US economy. For more on the topic, please read Professor James Hamilton's paper on the topic, which is located about halfway down the page. However, this does not explain the entire contraction, as oil prices increased sharply throughout the early 2000s expansion. Also consider that much of the contraction occurred in residential and non-residential construction, which was ultimately caused by the savings and loan crisis.

Yesterday, we looked at the recession, which was primarily one where investment dropped like a stone. Both residential and commercial real estate contracted in big ways, leading to a drop in PCEs.

However, the economy did not emerge from this drop in activity strongly. The NBER dated the end of the recession in March 1991, yet GDP grew 2.7%, 1.7% and 1.6% in the second, third and fourth quarter of 1991 respectively. And as noted in this article, initial unemployment claims and the unemployment rate remained stubbornly high for over a year after the end of the recession. In other words, there was talk of a jobless recovery.

Let's take a look at the three quarters following the end of the recession. I'm using the 2Q91 as the starting point because the recession ended in the first month of that quarter.

GDP grew 2.7% and PCEs added 2.01 to that. In other words, people spent money. Investment only subtracted .23 from the growth rate. Also note there wasn't a huge stimulus from government spending -- it only added .29 to the growth rate.

While non-residential investment dropped 10.5% from the 1Q91, residential investment increased 9.8%. That is the primary reason for the lack of negative impact from gross private domestic investment.

In the third quarter of 1991, GDP grew at a slower pace -- 1.7%. PCEs -- which only grew by 1.5% from the previous quarter -- added 1.01%. While non-residential investment dropped a whopping 23.5% from the previous quarter, residential investment increased 17.6%. In addition, business investment in equipment and software increased 6.6%.

in the fourth quarter of 1991, real GDP grew 1.6%. PCEs actually decreased .2% from the preceding quarter, but that was offset by a huge increase in investment. While residential investment increased 9.1%, non-residential investment only contracted 7.9%.

Finally, note that in 3Q and 4Q 1991, government spending subtracted from growth.

So. at the end oft he third quarter after the end of the recession, we have the following:

1.) PCEs are weakening, actually contracting in the 4Q91.

2.) Residential construction is expanding

3.) Non-residential investment is improving, although still negative (the rate of decline is improving).

4.) Initial unemployment claims are above 420,000

5.) The unemployment rate is holding a little below 7% (it would eventually increase to a little over 7.5% in 1992).

In other words, there's a lot of similarities between then and now.

Second Half Outlook: 3 important metrics

- by New Deal democrat

Here is my promised look ahead for the next 6-12 months. As before, the K.I.S.S. way of looking at things is simply to be guided by the LEI for the next 3-6 months, which suggest a stall (whether ultimately remaining positive or turning shallowly negative), and thereafter by the yield curve if we are confident of inflation, but if we aren't confident of inflation, then the fact of demand-driven deflation itself gives us the answer.

I see three items of particular importance.

1. The price of Oil and gasoline

Let's start with my January forecast, "2010: Gilded Recovery or Double Dip". Generally, I relied on the LEI to posit that the first half of the year would see a recovery stronger than most anticipated, but that Oil would determine the outcome in the second half (because the yield curve, which usually is reliable predicting the economy a year or so in advance, does not do so in a deflationary environment):

if consumers once again have to pay over $3 a gallon for gas (which ~$90 Oil would give us), it will have a psychological as well as economic impact on consumers, and I would expect them to cut back in other areas. This looks likely to be the case. We are just 6 months into renewed economic expansion, and at the seasonal low for Oil prices, and already Oil is at $80 a barrel. It is a near certainty that the expansion will continue for at least 3-6 more months....

Whether $90+ Oil will lead to a full-blown double-dip economic contraction, or just a slowdown later in the year, is almost impossible to gauge. It depends upon how far over $90 Oil shoots, and how long it stays there. If there is a dramatic overshooting a la 2007, there will be a double-dip. If there is a gradual increase over $90 that does not last that long before consumers cut back and the feedback loop causes price declines, then there may just be a slowdown, or if there is a contraction, it may be shallow and only last a quarter or two -- which is my best guess, and only a guess, at this point.
We now know two things:

1. Oil topped just shy of $90/barrel (or 4% of GDP) in April, and
1. From April on, we have started a shallow deflation.

Here's a graph from Prof. James Hamilton of UCSD showing that Oil just barely touched the level of 4% of GDP that historically has led to oil-price-induced recessions since WW2:

According to his analysis, each monthly increase or decrease in the price of Oil creates outsized effects that peak about 12 months later. That means that the effects of Oil's increase from $35 a barrel in January 2009 to $80 a barrel in June 2009 peaked this spring, and continue but are beginning to lessen. It was nevertheless a sufficient impact to cause both producer and consumer price deflation for the last 3 months.

This means, as I thought back in January, that the yield curve would not be a reliable forecasting device for the second half of this year. Secondly, by not going over $90/barrel, any economic slowdown brought about by Oil should not tip all the way over into a double-dip recession. The weaker the economy becomes, the more likely Oil continues to decline, setting up a rebound next year.

2. The cliff-diving of new home sales

The problem with Oil prices, however, has been exacerbated by the exaggerated decline in the housing market following the expiration of the $8000 tax credit. It seemed obvious there would be a slowdown, just as there was at the end of last year when the credit was initially slated to expire. But instead of slowing down, housing fell off a cliff. The much discussed "crash" of the ECRI leading index is apparently mainly due to the decline in purchase mortgage applications, which as of last week, have continued to crash:

What we know from the business cycle research of Prof. Leamer of UCLA is that housing declines typically lead recessions by about 5 quarters. Before the recession starts, other durable goods such as autos, also turn down. Yet here autos came in at their lowest sales number in 5 months, only one month after housing fell apart. Is this a feint, or is it because Leamer's research did not go all the way back to the deflationary busts that regularly occurred before the Second World War, which tended to come on quickly?

One hint as to what might happen is to look at the 2 other times in the last 5 years that Oil prices approached or passed $90/barrel, at the same time as housing was decreasing. That would be the second quarter of 2006, and the fourth quarter of 2007. Here's a graph of Oil prices per barrel, matched with quarterly GDP:

As you can see, in 2006 GDP slowed almost to a stall. In the first quarter of 2008, GDP declined, but only slightly.

The economy is sufficiently fragile at this point, and new home sales have fallen sufficiently long enough and deep enough, to make me believe that we will probably see one quarter of (slightly) negative GDP. But when?

A year and a half ago I took a very detailed look in five installments at theEconomic Indicators during the Roaring Twenties and Great Depression. I examined those indicators because our current situation more closely resembles those debt-deflationary downturns, as opposed to post-World War 2 inflationary recessions. That data from the Deflationary period of 1920-1950 showed that all of those deflationary recessions followed a pattern. The CPI declined from the beginning of the recession and its YoY rate of decline bottomed immediately before the recession's end. M1 money supply followed a similar pattern, sometimes coincidentally, sometimes leading slightly. In all 6 of those deflationary recessions, once M1 and CPI began to decline at a decreasing rate, the recession was about to end.

For example, looking at the Great Depression of 1929-32 (1= no change YoY, .9 = 10% decline YoY etc.):

we see that in this, the biggest of all economic contractions of the last century, like all other deflationary recessions, there was a clear pattern of M1 and CPI on the graphs --both money supply and inflation contracted at an increasing rate, then at a constant rate, before simultaneously or with M1 leading the way before CPI, both turn back up (meaning, prices and money supply are still declining, but at a decreasing rate) near the end of the deflationary recession. In other words, these deflationary recessions began to end once demand picked up. As demand picked up (recall Econ 101) inflation reappeared.

Turning our attention to our current financial crisis, which also features a debt deflation, here is the graph of the same indicators (M1 in blue, CPI in red as in the above graph):

The first thing that stands out is how "Helicopter Ben" intervened in the cycle by flooding the system with cash. The continuing good news is, M1 has not turned negative. "Real" inflation adjusted M1 is still positive, although at a decreasing rate. Remember, though, that pre-WW2, M1 money supply could be a more coincident than leading indicator, meaning if it declines, that would still signal a new round of recession. Assuming that happened, the monthly CPI readings do give us a clue as to when the maximum YoY deflation might be. Here is a chart of monthly CPI for the last five years:

If deflation persists for the remainder of this year, all the positive readings from the second half of last year will be replaced, and YoY deflation will intensify. But, absent a wage deflationary spiral (let's not even go there), there should be no further deterioration after early next year. In other words, if we do slide down into negative GDP for one quarter, it is likely to abate about in the first quarter of next year. The yield curve from this winter and spring was quite positive, and with a return to inflation, that would suggest renewed growth by next spring -- as is also indicated by the money supply and yield curve indicators from the 1920s and 1930s.

In summary, even if there is a brief "double-dip". it should not wipe out the advances in the GDP and Industrial Production we have seen since the bottom in June 2009. It will be a "recession within the recovery" just as 1938 was a "recession within the Depression" that did not return the economy to its former low point.

3. Income stagnation due to persistent high unemployment

One trend that stands out as deserving of more attention - including by me - is the affect of income stagnation. Here is a graph showing average (i.e., mean) hourly income for the last 5 years (blue), median wage growth (green). and inflation as measured by the CPI (red):

(Note: both mean and median income are included because the mean can be distorted by the top income earners. The median is the 50th percentile, but is measured only quarterly and not monthly)

This graph tells a story of income stagnation - increasing less than even tame inflation - leading up to the Great Recession, and also for the first 5 months of this year.

Shortly after I became convinced in April 2009 that the Recession was about to bottom, I penned a note stressing that there could be no sustained economic growth without real wage growth.. Here's a slice of what I said then:
with anemic wage growth to say the least, any incipient recovery appears doomed. For example, if our current (- 0.7%) deflation bottoms out and turns up within the next 2-4 months, how could there possibly be any significant expansion, if once again wages fail to keep up? Any such recovery would be short lived, strangled by the inflation caused by its own increase in demand. If the inflation rate agains exceeds wage growth, consumers will simply cut back again....
I went on to point out that in the last 3 decades, that statement had also been true, but had been avoided by consumers either cashing in appreciating assets or refinancing debt at lower rates -- and that the second possibility might still happen one more time. I concluded that:
if lower mortgage rates persist, there will be space for an economic breather ....
After that, the signs of a much stronger recovery, particularly in manufacturing, appeared, and GDP made up 3/4's of what it had lost in the "Great Recession." Jobs also began to be added much more quickly than following the 2001 recession. This was a strong, not weak, recovery, compared with the last two. But the nagging issue of wage stagnation remained. For example, in January, I asked:
If gas prices go up too much, or as the effect of last year's stimulus plan begins to wane, can gains be sustained?
And two months ago I reiterated that:
The most acute threat to the economic recovery that has begun is the same as the biggest chronic problem for the middle/working classes for the last 45 years: a lack of real wage growth.... In 2009, the bottoming of the Great Recession was helped by the fact that wage growth, although paltry, nevertheless was accompanied by a decrease in gasoline prices that gave consumers (the 85% or 90% who were employed) more disposable income.

That situation reversed in the last 6 months and now poses the most direct threat to the sustainability of the recovery.

The LEI have been slowly rolling over, and June's number will almost certainly show a substantial decline. This strongly suggests a stall if not an actual contraction within the next 3-6 months. Income stagnation -- that could not keep up with even a 2%+ YoY increase in inflation, and itself due to persistent high unemployment -- is a third potent reason why.

Let me conclude by making a few observations about the longer term:

1. With no structural assistance from Washington, Americans are solving their debt problems the way their forebears before 1932 did -- on their own, by saving more:

and paying down debt:

This process is likely to take another 2-3 years to undo the lion's share of the damage done since 1980.

2. While we are probably only a very short time away from the ultimate bottom in new home sales, price declines are likely for at least another 2 years until house prices return to their normal long term cost vis-a-vis average income. Yesterday's not good but un-bad news was that housing permits stabilized. If purchase mortgage applications also do so, that at least will mean we've reached the bottom.

3. Energy prices are another structural concern. We could get lucky and as new deepwater sources of Oil, e.g., off the coast of Brazil, come online, Oil prices could decline. But it was a fundamental miscalculaton of the Obama stimulus not to get to work on mass transit and power generation issues that weren't immediately "shovel ready" and bring them to fruition over a course of several years. But for now, proponents of the theory of "peak Oil" have been winning the argument.

4. Despite all of the above, there is no desire inside the Beltway for demand-side economics, building up the income and wealth of average Americans. Campaign contributions for both parties are dominated by wealthy banks and other corporate interests. Until Washington ceases to be Versailles on the Potomac, the structural problems facing 90% of Americans will not fundamentally change either.

To sum up, the idea of a reverse-square-root recovery, with an initial V that is incomplete may be the most fitting. Certainly with "peak oil" at least temporarily true, wage stagnation continuing, and several years until the bottom in housing prices, a recovery that is both strong and sustained will be hindered by powerful headwinds.