The word “stagflation” became popular in the 1970s when stagnation of the economy and extremely high inflation (greater than 6 percent) were both evident during times of high unemployment. The head of
Mistakenly, most of us reserve stagflation to describe conditions seen during the late 1970s and early 1980s, which made conditions that the Exchequer described in 1965 seem benign. This error caused me to review conditions that existed in the mid- to late-1960s, when stagflation was first uttered, and to then employ them as a starting point for measuring factors that could quantitatively define stagflation.
The late 1960s were dire enough to cause
In 2006, I thought it was wise to measure conditions when stagflation was first used to define it, which led to the following analysis (updated through October 2009). This analysis also clarifies what Dr. Mohamed El-Erian of Pimco anticipates as a new normal of slow economic growth, high unemployment and low asset returns. He expects these conditions in the five-plus years possibly needed to recover from financial crisis.
Stagflation: Defined & Revisited
Stagflation is used more appropriately when it describes times when the following annualized readings are evident: the civilian unemployment is higher than 5.5 percent, RGDP is less than 2.5 percent and the inflation rate (CPI) exceeds 3 percent. Figure 1 shows the 54, 23 and 49 calendar years when these thresholds have been breached since 1890.
Stagflation is gradual and insidious. All instances took root after inflation rose above 3 percent. A stagflation episode was identified (defined) when unemployment rose above 5.5 percent and RGDP fell below 2.5 percent within two years after inflation exceeded 3 percent. Below the column labeled Annual Frequency in Figure 1, you will find “1, within 2” to depict this confirmation sequence for quantitative thresholds. CPI rises above 3 percent first, (1) with the others following within two years (2). For example, our current crisis was precipitated by a stagflation episode when the 12-month change in the CPI hit 3.5 percent in November 2007. It steadily rose to 5.6 percent by July 2008 and remained above 3 percent until November 2008. It has remained negative since, and in September 2009, the 12-month CPI was -1.3 percent. Stagflation often precedes and follows consumer price deflations.
Figure 2 groups economic factors in the top half of the table (blue highlight), and the investment performance of stocks and bonds in the bottom half of the table (green highlight). Median measures are employed for economic factors, while annual compound total returns and income yields are shown for investments.
Economic and investment samples employ 120 continuous years from 1890 through October 2009 to establish baseline standards for comparisons to our Stagflation Period Sample (28 chained calendar years, Figure 2, column 3). Our sample includes 28 calendar years out of the 120 years since 1890, which were stagflation years (a 23.3 percent frequency).
Investment performance chains 10 discontinuous stagflation periods into a single 28-year continuous holding period to arrive at total returns (Appendix, Figure A). Yields represent the averaged yields during the 28-year sample period.
Figure 3 displays performance results obtained from chaining the investment returns of stocks, gold and three bond categories. For example, $1,000 was invested into each investment at the start of 1909 through 1911, the first stagflation episode. By the end of 1911, there was $1,158, $1,000, $1,088, $1,122 and $1,092 invested in stocks (S&P 500 Index), gold, U.S. 10-Year Treasury Notes (10 Yr TSY), AAA-rated corporate bonds (Corp Bonds) and state municipal bonds (Muni Bonds), respectively. These sums were reinvested back into each category at the start of 1919. They remained invested until the end of 1921, the end of our second stagflation episode. This process was repeated for the remaining eight episodes.
There were wide variances in the duration and severity of each stagflation period. Gold and corporate bonds performed best. Equities were last and at their worst when inflation exceeded 6 percent. Bonds did best after stagflation gave way to inflations below 2 percent. Gold performed fairly well in all episodes, with the exception being the 1948 deflation period. Gold prices were fixed prior to April 1933, when our nation revalued its dollar/gold exchange value.
For the most part, gold bested stocks/bonds during past stagflations. All episodes begin with high inflation and end with declining inflation, leading to good corporate bond performance. Corporate bonds did best when inflation declined because the resulting rise in RGDP mitigated default risks. Treasury returns beat all assets when declines in inflation rates and CPI deflations are associated with debt and asset deflations.
Municipal bonds were the worst performers because stagflation causes state tax revenues to decline. The laden default risks inherent in munis are exposed (Figure 4). Past stagflation periods raised investor concerns about each state’s inability to honor its debts (unlike the federal government, which can print money to avoid default); consequently, munis were repriced for higher default risk.
I am dismayed when advisers tout the taxable equivalent yields offered by municipal (muni) bonds. Many advisers forget that there are no free lunches. High yields, taxable or not, are compensation for risk. Muni bond performance during stagflation years was 1.8 percent less than the inflation rate because nominal growth (NGDP) at 4.4 percent was too low to reward investors for higher default risk. RGDP growth at 1 percent annually canceled the price premium awarded to muni bonds (Figures 4). Muni bonds were no safe haven in 2008. They lost nearly 11 percent because nominal GDP was negative for the first time since the Great Depression (Figure 5).
Next month’s InPerspective will reveal why the final chapter of our ongoing financial crisis will culminate in a government credit crisis (state, local and federal). We learned in 2007-2008 that our bond rating agencies failed to properly rate toxic subprime debt, leading to a violent repricing of risk and subsequent liquidations. Municipalities have huge off-balance-sheet liabilities in the form of actuarially faulty financial accounting. Ryan ALM, Inc. and others estimate municipal and state liabilities as unknowable, but high enough to bankrupt most state/local governments. Embedded muni bond risk will come home to roost again in 2010 or 2011.
Next month’s, issue, The New Normal and Asset Class Cycles, Part II, will show why muni investors will suffer much pain during the later stages of our financial crisis. It will also demonstrate how investors can benefit from selling muni and other government bonds in favor of commodity-based securities. We will show how government revenue bonds funding global infrastructure development and the stocks/bonds of companies that produce and distribute commodities are relatively safer than state, local (tax obligation) and federal government bonds.
Stagflation’s Linkage To The Infrastructure Cycle
Nobel Laureate Simon Kuznets (1971) was the father of gross domestic product, but he also held a thesis that identified infrastructure expenditures as being the primary driver of a 15- to 25-year real/financial asset cycle. In the Kuznets Cycle, transportation, public service and utility structures are in periodic need of major revisions, replacement and/or new construction. These expenditures accelerate demand for basic materials and industrial supplies, which leads to higher commodity prices, increased labor costs and rising inflation. Economists/market participants recognize many cycles, one of which is the Kuznets Cycle. Here are others.
Recognized Cycles & Durations
• Real Business Cycle (inflationary effects): 4-8 years
• Kitchin Cycle (change in inventories): 3-5 years
• Business Cycle: 4-5 years peak to peak
• Juglar Cycle (macroeconomic activity): 7-11 years
• Jevons’ Solar Cycle: 10.5 years
• Kuznets’ Infrastructure Cycle: 15-25 years
• Market Value Cycle (price to earnings ratios): 30 years
• Tobin’s Asset Cycle (financial currency versus hard assets): 30 years
• Kondratieff’s Long Cycle (mini cycles of spring - green shots, summer - full bloom, fall - decline, and winter - contraction): 50-60 years
• Technical Asset Cycles (market prices): intra-day to years
I favor the Real Business, the Kuznets, the Market Value, Tobin’s Asset and Technical Asset cycles, which are factored into research and asset management decisions. For now, our focus is on the Kuznets Cycle because it is often associated with stagflation episodes.
Treasury and corporate bonds best taxable obligation municipal bonds when governments are spending heavily on infrastructure. Infrastructure expenditures that replace existing assets are a drag on GDP growth in the years following the expenditures. Consequently, the expenditure side of the cycle leads to cost-push inflation and high unemployment or stagflation. Stagflation years are bad for tax revenues because during these times, taxable entities pay less income/property taxes. Citizens are also not likely to support higher taxes, yet many infrastructure expenditures are necessary. Consequently, fiscal deficits rise.
In next month’s issue, The New Normal and Asset Class Cycles, Part II, we will compare and contrast real asset class returns (commodities, gold and direct natural resource investments) to financial asset returns during previous stagflation episodes as well as our current stagflation episode. Factors featured in Figures 1-5 are incorporated into the asset allocation decisions of the Arrow Insights (AI) 75/50 Portfolio.
In next month’s issue, we also will document 60 years of neglectful infrastructure maintenance. Civil engineers estimate that it will cost our nation about $2 trillion simply to fix existing structures. These costs come at a time when emerging markets are building virgin infrastructures, which puts our nation in a disadvantageous position.
We will soon make it clear that credit instruments secured by the assets of enterprises providing for basic needs and vital services have bested tax obligation munis because they are less likely to default. In addition, we will review how building virgin infrastructure enhances economic growth. New infrastructure often enables nations to ward off stagflations. We also will show how long/short commodity strategies factor into asset allocation decisions. For now, we are sticking to long-only alternatives for hedging inflation risk.
Strategic & Tactical Asset Positioning
The Arrow Insights 75/50 Portfolio was down 0.6 percent before dividends in October 2009. We added more defense during the first week of November via the purchase of the ProShares UltraShort Real Estate ETF (NYSEArca: SRS). Fundamentals support a more defensive posture and a more active tactical trading strategy because we may be in a 2- to 8-month topping process that may fail for the S&P 500 near its 1100 to 1200 price level.
Month-end readings are good indicators of trend changes. The bullish trend has not changed yet, but the market recently came close to a serious technical break. Strong bulls are hard to kill. I am expecting a 10-12 percent correction, a 15 percent or so rally followed by a topping process ending in December 2009 or sometime in May or June 2010. This view changes with price action. Fundamentals support a resumption of the secular bear.
We are becoming concerned about rising financial system risk stemming from the recent rapid rise in reserve balances at Federal Reserve Banks (FRB). Reserve balances rise more quickly when the FRB is rapidly printing money to speed up its purchases of toxic assets from its commercial banks, as it did when banks were failing or close to failing during the fall of 2008 and the winter of 2009. What caught our eye is the following month-end build in reserves in 2009 (Figure 6). Here is the sequence: June $692.591 billion (b), July $764.620b, August $842.256b, September $924.353b and October $1085.392b, which works out to be a 57 percent increase over four months equaling the rate of reserves added in November and December 2008, when fears of corporate bond defaults and the collapse of our financial system peaked.
As the FRB balances were skyrocketing, we began to pay closer attention to the financial system fear gauge, the TED spread ($TED). To understand $TED, you need to know its construction. It was originally constructed from the London Interbank Offered 3-month (m) Rate (LIBOR3), which reflects the credit risk of lending to commercial banks and Treasury Bills, which are close to risk free. Today the standard $TED is the difference between the 3m Treasury minus the 3m eurodollar (interest rate) spread. $TED is employed to gauge the degree of fear existing about insolvencies within our financial system. A high $TED is indicative of a growing perception of fears about financial system risks (real & imagined).
A standard $TED recorded when rates were high is not comparable to a $TED scored when rates were low. Normally, $LIBOR3 is about 4 percent and $IRX or the 3m eurodollar is a little less, at about 3.8 percent. $TED is then 0.20 (a very low risk reading). On Nov. 11, LIBOR3 was 0.27 percent, $IRX was 0.06, so $TED was 0.21. Given what we know about the financial system, are risks low? No! Figure 7 displays the standard $TED in green.
A better gauge divides $TED by $IRX (Figures 7-8, $TED:$IRX), which currently reads 0.36. Given that low risk readings usually range 0.10-0.15, recent fears are high. They are near levels when fears of a financial system meltdown were ebbing in summer 2008 and waning in winter/spring 2009. Most importantly, our $TED ratio (Figure 7, red and 8) has climbed dramatically since Sept. 12, 2009, when it was at 0.12.
Figure 7. The TED Ratio & Standard TED
Figure 8. The TED Ratio
Our normal times example and current readings both record standard $TEDs near 0.20. The current reading occurs while the FRB is hyper-actively saving the financial system. The standard $TED is meaningless when the FRB is supporting the financial system. When the FRB stops being a savior, interest rates will normalize and the standard $TED will be a sound fear gauge again.
I have another modified $TED. It also corrects for tracking error evident during low rate climates and times of artificially supported markets. It is the LIBOR to Bill Ratio ($LIBOR3 divided by the 3m T-Bill Discount Rate yield or $LIBOR3:$IRX). Figure 9 contrasts it to the S&P 500 Index since November 2008.
The LIBOR to Bill Ratio as of Oct. 30, 2009 was close to March’s month-end (Figure 10). Intraday comparisons also support the view that fears are rising. VIX, the S&P 500 Volatility Index, also confirmed the ratio’s low in September 2009 and its recent breakout to fear levels last seen in March 2009 (Figure 11).
Figure 9. The LIBOR to Bill Ratio & SPX
Figure 10. The LIBOR to Bill Ratio in Crisis
Figure 11. VIX Confirms Higher Financial System Fears
The Fed’s Easy Money (It) And Related Assets Bubble Will End When It Ends!
The FRB was easing in October and November 2009 with a new round of expanding bank reserves. There are three points that need to be made about the relationship between excessive reserves and asset prices.
1. Excessive easing in a financial crisis works best after the panic has passed, i.e., September 1929 - November 1929 and July 2008 to March 2009, when stock rallied about 50 percent and 60 percent after making market troughs.
2. Easy money flows into risky assets after a panic, such as from November 1929 - April 1930 and March 2009 to present.
3. The ability of easy money to lift assets wanes if enough speculators sell to take profits. Following the 1929-1930 rally, stocks declined in spite of easy money.
For example, FRB and government fiscal stimulus was 8.3 percent of GDP from 1930-1931, yet stocks declined 60 percent and were down 85 percent from September 1929 through March 1932 (credit James Grant).
So far, in 2007-2009, our monetary and fiscal stimulus has equaled 29.9 percent of our annual GDP. Much like
For now, growth in reserves is positively correlated to risky assets, but at some point, speculators might sell when the public or average person fails to support the uptrend. Then more traders take profits. Our current low volume market is the domain of speculators. It can revert to HELL with little warning.
The Chinese Renminbi (RMB) Or The Yuan
On March 25, 2009 in the Arrow Insights 75/50 Portfolio, we sold a 5 percent position in the WisdomTree Dreyfus Chinese Yuan Fund (NYSEArca: CYB) at $25.40. We sought the appreciation of the Chinese renminbi but our timing was not right. Recent events cited below have changed our view to one that is more patient for fundamentals to unfold. On Nov. 4, 2009, we added back the 5 percent exposure sold in March at $25.31. This trade repositions some of the sale proceeds from corporate bond sales made on Sept. 29, 2009.
The Chinese renminbi (RMB), or the yuan, is a currency pegged to the U.S. dollar, with minor revaluations over the last few years. We think the yuan-dollar peg will only last for a few more months, or at most, about 12 months because economic fundamentals support a higher yuan-dollar exchange.
Gold’s Breakout & Its Target Prices
Figure 20 was generated on Oct. 13, 2009, when we commented about gold’s recent breakout above its 1032 price resistance, which confirmed a continuation of its long-term uptrend. Gold made a weekly close above 1100 on Nov. 13. If so, I had thought that gold’s price could rocket to 1200 within a few trading days or weeks. I would not jump in with new funds to ride it, yet it would not be a surprise (Figure 12).
Even though we expect gold to hit 1300 within a few months, we will sell enough of streetTRACKS Gold Shares Trust (NYSEArca: GLD) and Market Vectors Gold Miners (GDX) to rebalance to our target allocations near 12 percent of portfolio gross exposures for each ETF. If so, we would accelerate the first half of our PowerShares NASDAQ 100 Index ETF (NasdaqGM: QQQQ) purchase with a 7.5 percent allocation (15 percent is the target). Gold has been high beta in 2009. The AI 75/50 Portfolio will have a better beta-balance if QQQQ is added after rebalancing gold positions.
Portfolio Exposures & Convictions
Figure 13 is composed of a focused list of indexes and assets classes and corresponding ETFs. Returns are through Oct. 31, 2009. The AI 75/50 portfolio weightings are labeled in the third column. Red highlights are for sectors that were shorted through 2x inverse ETFs (also red in Figure 14). In next month’s issue, we will add the Chinese yuan and real estate stocks, which were added in late October and early November 2009.
Our long equity ETFs are either overweight (OW) or neutral weight (NW) relative to the Dow Jones Global Index. The bond allocation is managed opportunistically and to enforce desired risk exposure without reference to a bond index.
AI Portfolio Performance (Objectives should be evaluated over 36-month periods)
Like the Hedge Fund Research Global Hedge Fund Index (HFRX), AI’s secondary objective is to provide an absolute return (consistently positive returns). As of Nov. 19, 2009, hedge funds and the S&P 500 are down -15.8 percent and -26 percent, while the AI 75-50 Portfolio is up 18.8 percent since the October 2007 peak. The AI 75/50 model has outperformed its benchmark, the HFRX and the S&P 500 during all periods shown in Figure 15.
AI 75/50’s primary objective is to capture 75 percent of the S&P’s upside and 50 percent of its downside, which requires us to hedge beta nimbly while maintaining core beta (equity and bond exposures). The portfolio’s 13.6 annualized standard deviation (ASD) is higher than that of the HFRX, but much lower than that of the S&P’s volatility since the market peaked in October 2007.
Recent Trades, Open Orders And Current Positions (Percentages are relative to gross exposures)
All positions and our strategy view for all holdings are listed in Figure 17, along with our beta/non-beta allocations and themes (see the February InPerspective for more).
We were down -0.6 percent for the month in October 2009. We are up 3.5 percent in November to date.
Our gross year-to-date (YTD) return as of Oct. 31, 2009 was 14.5 percent and 18.5 percent to Nov. 19.
In October 2009, we made seven trades. So far, in November 2009, there have been two trades (Figure 16).
Figure 16. ETF & Mutual Fund Trades
Open orders: We have orders to initiate a 7.5 percent position in QQQQ at $39 plus 7.5 percent more at $37. There are additional orders to buy 6 percent more of AMJ at $23.
We will sell 10 percent of TBT at $64 if it ever hits this price (rebalance).
We attempted to short real estate stock via the iShares Dow Jones US Real Estate (NYSEArca: IYR) and the iShares FTSE NAREIT Retail Index (NYSEArca: RTL) but could not locate short positions from a broker. As a substitute, we purchased a 7.5 percent position before the UltraShort Real Estate ProShares ETF’s embedded leverage and 15 percent after its leverage.