John Early

Long/short equity, value, contrarian, investment advisor
John Early
Long/short equity, value, contrarian, investment advisor
Contributor since: 2012
Mark quit spreading misinformation. "All recessions in history have been preceded by an inverted yield curve." The four recessions between 1935 and 1955 were not preceded by an inverted yield curve.
The revision to shift counting research as investment rather than expense only changed growth rates about 0.1%. Data back to 1929 was revised. So comparing growth rates from previous decades to current rates is still comparing apples to apples.
Tom I think the volatility of earnings stems from the increased debt levels you talk about. More debt makes earnings rise faster in an expansion and fall faster in a contraction. The ratio of corporate debt to equity, to corporate cash and to GDP have been trending higher since the mid 1980s. CEOs can make a killing during an expansion even if they risk bankrupting the company if the next downturn is worse than expected.
Hi Brian
I'll start with a bit of context. I look at as reported earnings which unlike operating earnings doesn't exclude items like losses from discontinued lines of businesses. I am getting data from S&P Dow Jones indices ( I also use data from Cowles and Associates,
Common Stock Indexes which attempts to recreate what S&P 500 earnings and stock prices would have been back to 1870 (the source Robert Shiller uses).
S&P Dow Jones shows operating earning first being reported in 1988. At that time they were operating earnings was only modestly above as reported earnings. Over time the gap between the two has grown. I speculate this widening gap is an accounting fiction used to boost CEO pay. If I understand correctly CEOs only have to sign off on the accuracy of the as reported earnings for tax purposes not the operating earnings.
So a year ago the forward estimate for as reported S&P 500 earnings for the year ending September 2015 was 131.08 the actual number is going to be close to 90.81. The estimate was 44% too high.
If you take out inflation as reported earnings have a long term growth rate of 1.5%. They grow at a double digit pace during expansions, going from below trend to above trend and then fall back below trend usually during recessions. Once earnings reach a level far above trend they usually cease to grow at double digit rates. They may grow at trend for a while before starting to plunge.
A year ago real as reported earnings were farther above trend than they have been 98.5% of the time. The only periods further above trend were during the profit bonanza of WWI and that leading into the 2008 financial crisis. As reported earnings fell about 90% from those two peaks.
I have observed forward earnings to be estimating double digit growth going into the last two recessions when the actual result turned out to be 50% and 90% declines.
Currently, forward earnings estimate about 20% growth. I expect the best that can happen is trend growth of about 1.5% plus whatever inflation is. When the next recession comes I expect the eventual decline to be about 80% from the high a year ago.
In the last 10 years earnings growth has been several times more volatile than any period in history including the Great Depression and WWI. The leverage used to make earnings soar and give CEOs huge bonuses also makes earnings sink like an anchor during a contraction. I could be premature, but I think the earnings management techniques reached their upper limit a year ago and that the plunge has started.
Forward earnings have been pie in the sky for over a year. If you use actual trailing earnings (not the ex items) the S&P 500 earnings yield is 4.5%. Earnings yield and the so called Fed Model have weak correlation with future return over any length of period (1 year periods, 2 year periods, 3 year.. etc.).
Hi Go Lakers
Not so impressed with GDP much of the revision was higher inventory. The release also included profits and both measures I look at were down. On the other hand, personal income and durable orders that came out this morning certainly indicate the economy is stronger than I expected. Happy Thanksgiving.
Hi Go Lakers
You may want to take another look at real inventory. It has grown 50% faster than real GDP over the last 3 years. It has grown faster in 11 of the last 15 quarters and the last 4 quarters straight.
Hi Go Lakers
"not a single indicator" Really? So why is it that the peak month of industrial production was back in December 2014. GAAP or as reported earnings of the S&P 500 peaked in Q3 2014 and is down about 14.9%. It has fallen for 4 straight quarters. Profits ( with cost of capital and inventory adjustments also peaked in Q3 2014. Housing starts and permits both peaked in June 2015. The US stock market (S&P 500) is going on 6 months now without a new high. With inventories growing 60% faster than GDP in the last year not all is as rosy or sustainable as the idea that no indicator is pointing down.
David, thanks for the thoughtful comment. You are of course right not to look at single indicators when everything happens in the context of everything else. When I examined CFNAI I didn't find it had any meaningful leading indication on quarterly GDP. I viewed it as a coincident indicator, but will look at it again. Building permits as I interpret them show strength for Q4, but weakness in Q1 2016. I have not examined the STFSLI and look forward to doing so. Also will look deeper into temporary workers. You may want to reexamine the correlation between GDP and the yield curve during this period of ZIRP.
John "Too" Early its true. And it has gotten worse. I was about a year early on the 1990 recession, about 2 years early on the 2001 recession, about 2 and a half years early on the 2007 recession. Looking at about 3 and a half years now and counting on the next downturn.
In those previous times I came out ahead of the market over the full cycle. Not at all clear If I will come out ahead this time.
I attempt to look at actual correlation between variables and then determine if their is an actual influence between them or if perhaps some other variable influences one before the other. On a concurrent basis the change in the oil price has a significant correlation with the change in economic growth. It appears a rising price of oil corresponds to a stronger world economy. On the other hand a lower price of oil corresponds to faster growth in the future. It still surprises me that the strongest correlation for this is with oil leading 21 months. See the gold line in the chart.
The decline is directly related to the lagged effect of marginal tax rates and brackets on personal income. Back when the share of the pie going to take home pay for workers averaged over 50% the top bracket was over 100 times per-capita GDP now the top bracket is less than 9 times per-capita GDP. The higher the marginal rate on the last dollar of income is the lower the after tax marginal cost of hiring an additional dollar of labor.
It is hard to argue that Obama care is killing jobs. Since December 2009 when Obama care became law the economy has added 12.9 million private sector jobs. In the nine years prior to December 2009 the economy lost 4.7 million private sector jobs.
The Laffer Curve is about the relationship of the tax rate to government revenue. Maximizing government revenue is hardly a worthy policy goal certainly not a conservative one.
I have never seen a Laffer Curve plotted with actual data or even one that puts units on the axes. Does the axis about the tax rate refer to the average tax rate, the top marginal rate, the average marginal rate or the share of GDP going to tax? Does the axis about the revenue to Government cover Nominal revenue, real revenue or the share of GDP taken in the tax? If we had one tax rate that applied to all income the Laffer curve would probably make sense and be useful, but with the marginal tax rate system we actually have the Laffer curve as near as I can tell is a meaningless hypothetical used to argue for lower marginal tax rates.
If I tried to plot a Laffer curve with data through 2014 I would conclude a 39.6% top rate maximized revenue. With data through 1945 I would conclude a 94% rate maximized revenue. If through 1989 the conclusion might be a 28% rate.
On the other hand the rate of GDP growth and the level of the top marginal income tax rate can be plotted on a graph using actual historical data and the conclusion is consistent that a top rate of 50% is more pro growth than any top rate at a lower level. If we had a 50% top rate I think the bracket that optimized the balance between a low average tax rate and a high marginal one would be about $1.5 million. So only the portion of income above that would be taxed at 50%.
The great builders of industries usually have a steady capital expenditure program so that in any given year they take the depreciation for decades worth of investments. These captains of industry are not driven by the one year time frame you describe. Plus most of the expenses of building a business are fully deductible in the year they are made. The higher the marginal tax rate the more value there is to the deduction and the greater the reward of growing value within a business where it is only taxed at the discretion of the owner. The increased value of a business is untaxed unless it is drawn out as personal income, the business is sold for a capital gain or it is taxed in the estate of the owner. If the increased value is given to a charitable foundation it is never taxed.
The top marginal rate can be too high as well as too low both damage growth and investment. The top rate has its biggest influence on growth and investment 2 years later. So for example the top rate this year will have its largest impact on growth in 2017.
Since 1920 there have been 22 years in three different time periods where the top tax rate was at 35% or below. Using the 2 year lead time the 22 years influenced by the low top rates had an average GDP growth rate of 0.3%. All three periods had below average growth rates. For example, the 10 years from 2005 to 2014 influenced by a top rate of 35% annualized growing 1.6% about half the average rate.
Lunco while your argument sounds logical I don't know how you can back it up empirically.
You state, "Taxes take money from corporations/individuals that could otherwise be invested in new plant/equipment or in new enterprises."
The business revenue that is spent on real investment like productive equipment is deductible or depreciable. So those expenditures reduce the portion of revenue subject to tax. It is the profit or share of business revenue that is not spent on wages, equipment, marketing, research or the start up costs of new businesses that gets taxed. if all the revenue went to deductible expenditures or were part of a steady capital expenditure program there would be no corporate tax due or personal income tax due for the business revenue. Thus the statement that income taxes prevent investing in businesses is backwards. Plowing money into deductible and depreciable business expenditures prevents paying taxes.
Lunco if the economy worked the way you describe Clinton would have had dismal GDP growth rather than being better than Reagan and Bush 43 would have had stellar growth rather than being the first President since Hoover to loose private sector jobs while in office.
If the economy worked the way you believe the GDP growth rate during the 12 years of Harding, Cooldige and Hoover would have annualized a lot better than the 0.5% that occurred and the 12 years of FDR would have annualized a lot worse than the 9% that occurred.
The key to sustainable prosperity is for wealthy busnisess owners to have a low average income tax rate with a high marginal income tax rate. The average rate has to be low enough on a large enough amonut of income to make it worthwhile to run and or fund a business. The marginal rate has to be high enough that wealthy people avoid taxation by plowing money into growing the value of the businesses they own while pulling less out as personal income.
The industrial production index is a real index. It measures output not dollars of output.
So far industrial production and real retail sales show a peak in November 2014. To be in recession real personal income (RPI) and employment would also need to have meaningful declines. RPI did slip in March.
If we are entering into recession we will see a negative jobs number soon.
Before the NBER declares a beginning date for a recession often about a year after the fact, there will be a sustained downtrend in employment, real personal income, real retail sales and industrial production. They would want to see at least a 6 month downtrend and then a few more months in case there are revisions.
Industrial production and real retail sales may have had business cycle peaks in November 2014, but employment and real personal income have continued to make new monthly highs. If we are in a recession employment and real personal income would turn down soon.
bixbubba, there are 2 other quarters in the data period where the oil price declined by a larger percentage: Q4 2008 and Q1 2009. Obviously other factors are not as negative now as they were then.
The estimate of the model goes from 3.3% Q3 to -0.14% Q4. The 25% oil price decline for the quarter (69% annual rate) accounts for less than 1% (less than a third) of this 3.44% estimated decline. Even if oil had no impact on growth Q4 should still be substantially weaker.
I don't know of any reason why the relationship between growth and oil should be different this quarter than in the past 10 years, but that doesn't mean there is not one.
Thanks Bob. The standard deviation is 1.3 so it does seem probable that the number will come in below 1.2%.
bobdark, I think the biggest difference in this model and the mainstream outlook is the lead time for housing starts and the indication of oil. With the 9 month lead time housing starts have a very negative indication for the 4th quarter. Most people believe falling oil is good for the economy. While I think low oil prices are beneficial after a long lag, on a concurrent basis I believe falling prices are a sign of a weak global economy which will be a drag on the US.
The oil lead is longer than I expected when I calculate it with annual growth it is more in the 15 to 17 month lead range. I will keep an open mind about this lead time.
Don't know anything for sure. What I suspect is that when short term rates are below 2% or so the yield curve ceases to correlate with the GDP growth rate. For example US between 1930 and early 1950s. Also since 1995 there have been 6 recessions in Japan with no inverted yield curve. The Italian yield curve last inverted in 2008, but their economy recently went into recession.
If short rates are above 2% I think the yield curve is the best leading indicator of recession.
Most employment measures are lagging indicators. In a normal expansion growing above 3% a slowdown in growth has time to show up after a lag in faltering employment data before a recession begins. Since we are growing around 2%. I suspect a recession could start before the lagging indicator of employment show a decline. 2013 grew at 1.6%. 2014 is growing at about 2.1% through 3 quarters. I am expecting Q4 to pull full year growth down a bit.
Earnings? are you talking about corporate earnings or earnings of labor?
Corporate earnings often lead GDP about 1 quarter, but it varies.
In the last 7 years nominal GDP has annualized 2.7% (weakest 7 years since 1938) while as reported earnings have annualized 4.2%. I suspect the difference is leverage, financial engineering and accounting games. When the economy weakens just a bit more the leverage etc that has elevated earnings above GDP growth rate will likely shift to making earnings plunge. In the last few years "as reported" earnings have had a higher standard deviation than anytime in history. When the downturn comes earnings should fall fast and hard.
I don't have any smoking leading indicator pointing to an eminent recession. My baseline forecast of growth is 1%. In the last 7 years real GDP has annualized 1.1%. The last 5 years have grown faster than my estimated baseline. I perhaps have a bias toward expecting growth slowing toward or below the baseline.
What is your definition of a bear market that excludes the 30% plus decline in 1987. Also the four recessions between 1936 and 1954 were not preceded by an inverted yield curve.
Why don't you show what the SPX:AGG chart looked like in 2007 since the SPX peaked in October 2007 instead of starting your chart 9 months after the market peaked in July 2008.
levin70 Thanks for the input.
Historically the 2.2% is well below average, but I think it has been and continues to be a fulcrum between stall speed and escape velocity. You are right when growth has climbed above that level following a recession it has always strengthened further. On the other hand when it has been above this level and fallen to it a recession has always followed if the economy was not already in recession. It's curious growth has hung out at this level for two years. From opposite expectation we will both be watching to see which way it goes.
Salmo Trutta
Credit growth such as
lags quarterly GDP growth about 7 months. This does not tell you about a rebound in 2015.
There is a pretty strong correlation between interest rates and the ratio of the monetary base to GDP. Some charts on this are in this article.
"In the last 50 years the yield curve has a 100% accuracy rate for predicting recessions, with no false positives."
I See the data the same way and also note the yield curve inverted to forecast the first recession of The Great Depression. Yet I consider the possibility that the yield curve does not give warning when short rates are held below about 2%.
"which?" answer: all. There was no inverted yield curve prior to any of those four recessions. I did misspeak slightly: while the 3 month T-bill yield was continuously below 2% from 1933 through 1952 it was a bit above 2% for a few months prior to the 1953 recession.
How do you measure tightening for the recessions in 1937, 1947 and 1953 and how is it different than the 3 month T-Bill yield in this cycle bottoming back in 2011?
Yes, but the 4 recessions prior to 1954 all began with short rates below 2% and no warning of the yield curve inverting.
Thank you for your comments. We obviously see some things from a different perspective and I want to understand yours better.
You point out "a rising capital goods to consumer goods ratio that we find as recessions approach." I looked at a couple of possiblities of what you may be referring to. The one that I think comes closest was the ratio of "private non residential fixed investment" and "personal consumption". What the chart says to me is that investment is much more volatile than consumption and when the economy is growing it grows faster than consumption and when the economy shrinks it shrinks faster. I didn't really see any clear relationship between this pro-cyclical ratio and the long term growth rate. I believe the analysis of marginal tax rates and brackets finds correlations with the long term growth rate. A change in the marginal rates occasionally corresponds with the shift between expansion and contraction, but I don't think the correlations between marginal tax rates and long term growth are much help in timing a business cycle.
You state, "A lower tax rate represents a potential permanent lower cost on businesses..." Suppose a tax bracket on personal income were added to existing brackets where the portion of personal income above of $1 million dollars had a marginal rate of 45%. How do you see this effecting the business costs of wages, equipment marketing and the like?
One other area where we may see things differently. You say "Surplus spending would act to stimulate the demand side." What the data suggests to me is that if more business revenue is taken as personal income part of the increase on the demand side is taken up by increased imports. The positive correlation between marginal tax rates the balance of trade is another reason I believe business revenue spent on the business has a larger growth effect than business revenue taken as personal income.
Do not have any interest in being a hero to some political spectrum, but am curios if you have a preferred measure of "net business investment?"
In reference to comments further down we may have miscommunicated. The data I have shows nominal GDP did not make a new high till 1941, while real GDP surpassed the 1929 high in 1936. I am curious if you have a data source to show real GDP did not continuously remain above the 1929 level from 1936 on.