A Tale of Two Depressions*: An Empirical Zero Bound Policy Analysis

[following was published in SumZero]

Abstract

According to Robert Prechter, a crisis will take on the characteristics of the one before the previous crisis. In order to entertain this line of reasoning we look past the inflation crisis of the 1970s and the boom of the postwar period, to the deflationary crisis of the 1930s. Both periods (2008-current, 1929-1946) were marked by concomitant endemic deflationary and artificial (policy-driven) inflationary pressures. A comparison integrating the following variables will show that policy decisions and real outcomes are similar to the 1930s: GDP (inflation adjusted), total outstanding credit (public, private and household), inflation (CPI, T break-evens, etc.) and interest rates (market and policy determined). As debt has been borrowed and denominated in USDs the long US government bond will benefit from capital appreciation and currency strengthening due to further deleveraging. A short-term target (6-12 months) for US 30-YR (USGG30YR) is set at 4.1% (as of writing, it is 4.55%)

Profs. Barry Eichengreen and Kevin O'Rourke have compiled relevant figures that will be used in the following exercise

World Industrial Production, 1929- (yellow), 2008- (blue) [Credit: Eichengreen/O'Rourke]

Broad Parallels: Total Credit, Policy Rates and the Yield Curve

Not only is total credit outstanding several magnitudes above the previous period, but the current period's policy rate environment is standard deviations more severe -- compare zero-bound policies today to +150bp in the 1930s. Traditionally expressed through the Treasury's short end, this policy mechanism broke down in 2008; QE was initiated thereafter -- longer-dated Treasuries and paper were purchased as a greater elastic measure to depress long rates and to liquefy credit markets, respectively.

7 country average [credit: Eichengreen/O'Rourke]

Yield curve dynamics in the current US environment is similar to the Japanese deflationary experience (1990-current) and the previous US depression in the 1930s-40s.

The Japanese yield curve for nearly the past two decades has been consistently normal (upward sloping) and steep; while concurrently facing massive deflationary headwinds (despite counter-cyclical policies).   The only exception was the the summer of 1987 to the winter of 1991 when the Japanese yield curve was inverted.  Since early 1992 the curve has been upward sloping and steep with yields falling in nearly every duration.  In the U.S., in the early 1930s, also after a brief period of inversion, the yield curve became upward sloping, quickly steepened and yields dropped progressively in nearly every duration. On the long end, yields on the long-term 30-year US Government bond (and high grade equivalents) drove lower until 1946 or 47. In terms of nominal/real rates, 1926-1927 was the cross-over point, with real rates rising above nominal, and did not cross back over until 1937/38.  These dynamics describe a 'bull steepener', which is bullish for bonds, and bearish for equities and risk-assets. A bull steepener is typical of  depression/deleveraging cycle.  Recently, the yield curve's dynamics fit the above bull steepener profile. The general trend is a depression in yield across all maturities, while the curve steepens. The short-end is policy; the long-end, deflation.

Comparable Policies

The outstanding difference in policy in terms of rates is that the zero bound has been approached more rapidly in this period versus the previous period. In the previous cycle of bust following credit excess, policy rates were lowered in relatively gradual fashion compared to the current period. Tracking the rediscount rate of the Federal Reserve Bank of New York, beginning in 1914, 1920 marked the highest average annual rate of 6.48%. At the year of the stock market crash, 1929, the year's highest rate was 6.0%. By the following year, the average annual rate collapsed to 3.0%. But it was not until 1938 for the rate to average the absolute bottom of 1.0%. In the intervening period, rates fluctuated between 3.5 and 1.5. Compare to the near immediate response of reducing the same discount rate to 0.50% (since February 2010, the rate has been 0.75%) and the fed fund's rate to 0.25, in the current cycle; the FOMC settled on the zero bound on both measures as soon as December 2008, or within the same quarter as the severest portion of the market crash. (see chart)

Fed Rediscount Rate, 1920-1945 [Credit: ACP Capital Advisors, LLC]

Though the current combinations of measures are unprecedented, the question is if the speed of policy response and consistency of zero-bound policy coupled with longer maturity purchases of debt and paper (quantitative easing) cause a difference in outcome. It is worth pointing out that though the response in the 30s was slower, less consistent and less drastic, the collapse in rate was 91% to the eventual lower bound (1.5% of 1.0%, from 6.0%) within two years (1931); policy, therefore, was not as delayed as generally perceived. One can argue whether a 1.0% target relative to the current 0.50% would have made a difference, but considering the credit market debt increase in the previous periods respectively, it is hard to imagine that a 50bps difference would have ignited appetite. The following evidence will demonstrate that current policy measures pursue the same string pushing regime with respect to further credit creation.

Policy Efficacy

In addition, fed policy makers and observers alike have noted the decline in the past quarter century in the effectiveness of policy tools in shaping longer term rates. In terms of manipulation of short maturity rates in order to shape the yield curve, the relationship has changed dramatically since the late 1980s, according to Daniel Thornton of the St. Louis Fed. Thornton hypothesizes that the marked change in the relationship is due to the FOMC switching from using the funds rate as an operating instrument to using the funds rate to implement monetary policy.  

If the Fed's standard monetary policy tool has been losing its effectiveness, controlling money's velocity may be out of policy's hands, save for an experiment called QE. In fact quantitative easing may be the Fed's only recourse to contain rates and prevent a deflationary spiral and a collapse in broad economic activity. The same unconventional methods also place the economy closer to the risk of hyper-inflation from the direct monetization of debt -- as they are the only remaining walls separating the economy from uncontrolled deflation. Having in recent memory the inflationary crises of the 1970s, senior managers remain fixated on the latter outcome. Money velocity trends and broad debt accumulative patterns paint a different picture.

 Underlying Dynamics

Structural dynamics underlying the economy have not changed.  Cycles of credit/debt expansion and destruction are unavoidable; resulting in boom periods (1950s/1976[currency float]-2000) and subsequent deleveragings (2000-[postponement period; 'Greenspan put']-2008-current). Note that though the initial crash began in 1929, the Great Depression did not begin officially until 1932, or three years afterwards – in terms of aggregate output and unemployment.  A similar outline of events may be unfolding, with some non-minor differences.

Total Credit Market Debt as % of GDP, 1920-2008 [Credit: Ned Davis Research]

Total Nonrevolving Credit Outstanding YOY %Change [Credit: St. Louis Fed] (Shaded regions indicate recession) [note: violation of zero-bound]

With respect to credit expansion as a percentage of GDP prior to the crash (see total credit figures) fiscal and monetary tools used in the 30s were similarly accommodative as the current period.  Chairman Bernanke dropped policy rates at a breathtaking clip as early as 2007 from 525bps to 25-0bps, and initiated the unconventional debt purchases, but given the relative explosion in credit relative to GDP, the measures taken seem much less out of proportion to the previous period.  A zero-bound interest rate policy, plus QE and other expansive fiscal policies, may be approximately as accommodative as the low-mid single-digit policy of the 1930s, though not in absolute terms.

Fiscal and regulatory policies will likely continue to mirror the 1930s, when government intervention prevented market forces to resume organic growth, despite low rates and deficit spending. Cole/Ohanian's study of wholesale government intervention and its adverse affects on wages and prices in the 30s is a troubling reminder that applies today. Where in the 30s the emphasis was on a policy-induced artificial differential in prices in labor and industry, the current crisis' emphasis is on the differential in paper and real estate assets on corporate and household balance sheets. Market pricing mechanisms are being retarded by bad policy; the effect on the real economy may be as distortionary as in the 30s.

Comparable Data

Some absolute figures are as disconcerting as those seen in the 1930s and comparable with respect to their counter-inflationary implications.

With consumer spending in the U.S. accounting for over 70% of GDP and a current real unemployment rate (U6) of 16.7% (and over 20% if using pre-Clinton era BLS metrics), the depression in real wages will figure onerously against the rising threat of policy induced inflation. Just like the 30s when rates were eased and fiscal policies were expanded, endemic deflationary considerations did not allow monetary and fiscal policies to achieve their aim.  Signs of a repeat of history, like the collapse in the M1 multiplier (keeping money velocity in the system low) are evidence of such endemic deflationary concerns despite counter-cyclical policies.

Since the crash in late 2008, the credit transmission mechanism has broken down; the credit created at the Fed, through asset purchases, emergency loans (Bloomberg, LP has sued the Fed for disclosure) and longer-end Treasury purchases, is failing to reach the broad economy.  Excess reserves of deposit taking institutions have gone parabolic.  Add to this, the deflationary effect of private sector unemployment is further compounded by the imminent reduction in public sector employment, which constitutes a considerable 12% of U.S. GDP, given the $125+ billion deficit facing local and state governments.

Excess Reserves of Depository Institutions  "Perfectly parabolic"  [credit: St. Louis Fed]

Federal Reserve Balance Sheet (Total Assets) 8/07-12/11

M1 Multiplier Drops Below 1.0 [Credit: St. Louis Fed] (Shaded regions indicate recession)

And the fact that average wages figure so prominently in the economy is not helped by the observation that all of the debt has already been created (mostly via securitized instruments in this cycle).  Though before the crash, debt creation was accelerated by unprecedented leverage fueled by artificially low rates, % change credit figures (see charts) show that the increase in monetary infusions and direct asset purchases by public monetary and fiscal entities are replacing (not creating) the huge sums of notional value actively being eroded during the deleveraging/depression process. At least $12 trillion has been already written down, out of the $60+ trillion in outstanding debt (global GDP is $58 trillion), not counting the additional trillions that need to be written down if covers and sensible mark to market accounting rules had not been introduced and suspended, respectively. In any case, these numbers are staggering in contrast to the low single digit trillion QE figures.

Bridgewater Associates' Ray Dalio** likes to emphasize the long and drawn-out nature of deleveraging/depressionary cycles, which are also characterized by rolling waves of secular and intermittent countertrend market movements. The counter-cyclical policies are attempting to at least allow an overwhelming tide of debt outstanding from being too rapidly expired. Total non-revolving and revolving credit outstanding have been declining and declining at a high rate. It is unclear if Fed and Treasury policies are achieving their intended result of keeping markets liquid and stable. Market participant reactions have been worryingly inconsistent with past stimulative counter-cyclical policies.

Inflation Metrics / Concerns

Currently, the five-year five-year forward breakeven rate sits at 2.97%, with a short term high of 3.28%.  While five-year TIPS - five-year Treasuries = 2.31%.  As recently as October, core CPI had its lowest year-over-year print since 1958, when records began.  The Commerce Department's PCE index excluding food and energy rose 0.8% year-over-year in January after increasing 0.7% in December for the smallest advance since 1959, when records began (source: Bloomberg).  Although core CPI and  PCE weigh housing prominently (49% for CPI), its significance to consumer spending and output cannot be understated in a post-2001 credit bubble environment, when housing overtook equities as the main driver for individual net worth.  In terms of their pass-through effect, the exclusion of food and energy costs in these metrics, though seemingly disconnected with the public, has been essentially non-existent for decades (source: Reinhart/Bloomberg).

These inflation expectation meters, though higher since the depths of the current downturn, are still signaling stronger signs of disinflation, and potential deflation, than outright inflation.  Keeping in mind QEI's completion and QEII's (over) halfway execution, resulting in over two years of nearly uninterrupted currency creation; in light of these prints, perhaps it is more worrisome that even with the debasing tendency of the Fed's QE policies inflation has failed to be meaningfully produced.

Trimmed Mean Inflation Index (Dallas Fed) [credit: Dallas Fed, Bloomberg]

Total Revolving Credit Outstanding (Billions of USD) [Credit: St. Louis Fed] (Shaded regions indicate recession)  [note: uninterrupted positive first derivative and subsequent negative reversal]

Post-bubble Debt Appetite

Following a mass credit expansion, market participants (excluding policy makers) are reluctant to bear debt all over again -- at least not right away. If history is a guide, they will act reversely and logically delever their respective balance sheets; government fiscal/monetary stimulus notwithstanding. None of this is helped by the fact that the pre-2008 credit expansion was the largest in recorded human history (due to poor arbitrary monetary policies and controls, current account imbalances, fiscal profligacy and relaxed accounting standards for corporates and sovereigns).  At this point in the debt cycle and rate of monetary expansion, attempts at policy-induced credit expansion will continue to 'push on a string.'

To conclude that all further policy action will result in immediate and inexorable calamity in the direction of the most extreme perceived outcome for a given course of action, simply because of failed monetary policy leading up to the recent turmoil, is not the result of clear-headed thinking.  Short-sighted extrapolations can lead to problems not unlike the pre-bust era case of underlying home prices and future valuations of CDOs-- when only post-war data was used in predictive models, which reflected nearly consistent positive first derivative prints.

In the U.S., overwhelming consensus of economists, forecasters and buy/sell-side analysts and many portfolio managers is that deflation is no longer the primary risk, and many others believe inflation is imminent on a relatively short-term basis.  Comparisons to Japan style deflationary malaise are generally rejected, citing U.S. savings tendencies and significant foreign ownership of U.S. debt.

Average monthly CPI and CPI Forecasts [Credit: Bloomberg Survey]

Economists forecast by Bloomberg (Bloomberg, L.P.) show overwhelming votes for relatively high-moderate inflation (see figure).  They have been shown to be consistently wrong (overshooting significantly) in the previous two years, compared to revised figures.  However, despite their lack of success, revisions have been minimal.  Their expectations remain high for inflation.

Future QE Considerations

Consensus irregardless, considering the change in total credit and inflation expectation figures, the Fed will continue and finish QEII as planned, despite a lack of unanimity within, and will most likely announce QEIII and QEIV, largely shadowing policy decisions of Japan's BOJ for nearly two decades.  This intention has already been signaled to the market as soon as Bernanke/FOMC decided to undertake Quantitative Easing (the arguments for a zero-interest-rate-policy trap are convincing).  The net effect on inflation will remain muted if not negative, or deflationary.  A first hand observer of the policy reactions of the BOJ to the credit bubble implosion shares the same view.

DJIA Sep. 1929-July 1932 [credit: Elliot Wave International] [note: initial decline -49%, subsequent rally +52%, and crash ~-85% (from top)]

Consumer Loans (blue) and Commercial/Industrial Loans (orange) [credit: Fed]

The Federal Reserve (whose shareholders include publicly traded banks) though not a traditional profit-seeking entity, shares a characteristic in common with the banks that they are mandated to regulate.  The Fed has strong incentive to keep its monetary and employment mandate, and guarantee that its practicing institutions remain in tact; they will not act to destroy the currency and effectively go out of business – unless under truly extraordinary circumstances.  The Fed will only resort to hyper-inflationary policies out of sheer last resort in order to avert a deflationary crisis that is multiple standard deviations more severe, measured in total debt contraction.  The pace of debt destruction must be more comparable to the total debt outstanding.  This means printing some multiple of GDP greater than 0.75-1.  In all other scenarios, the Fed will tread the line carefully, as it could spell its own end along with the currency.

As the Fed continues to provide excess liquidity surplus credit is clearly failing to penetrate an iron curtain of excess reserves held at the Fed. In addition to loss reserves related to a rising number of NPLs, the availability of credit-worthy entities is causing these reserves to build up without entering the broader economy. In addition, the previous buildup of debt on balance sheets of both corporate and household borrowers is causing appetite to wane. If minimal loss reserve requirements are raised, or if another correction precipitated by a fall in asset prices wipes out large chunks of equity in banks, excess reserves may very well have to be increased further -- which, in turn, will further restrict credit to the economy.

Government Debt / Revenue (2010) Government Interest Expense / Revenue (2010) [Credit: Moody's, Hayman Capital Management, L.P., JMF] [note: Despite large debt (top), U.S. benefit: w/ low relative rates and high revenue - (bottom)]

Dollar Index Spot (DXY) (2006-2010) [note: rally in '08-'09 (orange) and subsequent higher lows (red)]

Endemic Deflation

Given these concerns, deflation is the current greatest probable risk to market stability.  The base case remains to be the overlooked explanation: the inflation, which the majority is expecting, has already occurred. The previous multi-decade decline in the major trade weighted dollar, despite the Fed's devaluation, continues to mark higher lows and lurches upward during broad-market sell-offs. Technically, based on five-year view, a case can be made for a base-building process in the USD. Weaker trading partners may accelerate the strong/stabilizing dollar trend.  Keynesian Endpoint discussions become urgent with the Eurozone and Japanese sovereign debt concerns.  A comparison of Revenue/Interest Expense ratios of both peripheral and some core -- France and Spain -- Eurozone states and Japan, to the U.S. illustrates the disparity.  As the least worst case, the USD stands to gain in purchasing power intermediately in synergistic conjunction with the deflationary pressures of debt repayment.

Gold price (2007-2011) [note: sell-off in '08-'09 highlighted in orange]

Though in the background, there is a systematic currency devaluation to help manage the debt load.  The relative strong performance in the price of gold and other precious/monetary metals is a corroborating factor.  The undercurrent of destabilization of major currency pairs has been a theme tracing back to 1971, but recent aggressive monetary policies exacerbated the problem (policy rates veritably mirror a roller coaster in the past two decades).  Gold's strength in the face of competitive devaluative policies is no surprise.  But as the liquidation that occurred in 2008-09 illustrates, such being the nature of deflationary cycles, even the considerations of inflation via gold's price were swept to the side as the price of gold lost nearly 30% of its value in a matter of weeks.  The current retail euphoria does not help its short-term-intermediate case; it is susceptible to a non-minor correction.

The best performing sovereign debt in the past decade has been JGBs.  The failure of market participants to discriminate between real and nominal yields/returns has been behind many failures to successfully short this asset.  In nominal terms, 10-year JGBs have recently yielded just above 1%.  In real terms (real return), with the strengthening Yen, due to their endemic deflation, the yield is several hundreds of basis points above the nominal rate. Given the real yield, Japanese patriotism and high savings rates alone are not the cause for nominal yields staying low. Having experienced the same, the Japanese are increasing their UST overweight.

Jeffrey Gundlach**** relates the tug-of-war between endemic deflationary and artificial inflationary forces to a, "a room with a heater and humidifier both running at the same time." As long as one is running at full force, the other, if also running strong, will counter-balance the effect to some degree.  As soon as one is turned off, the humidity, or lack thereof, will certainly be unbearable.  The Federal Reserve will likely continue QE and its forthcoming iterations, and not cease until natural inflation is produced -- in the U.S., that means real and sustainable employment growth.

Avg Duration of U.S. Unemployment > Standard Unemployment Benefits [Sources: U.S. Dept of Labor, Bloomberg]

Employment

This excludes intermediate tampering by the BLS with the unemployment calculation from pressure by the Administration, as was done recently (inflating the denominator in unemployment rate calculations).  Considering the current trend of the growing long-term unemployed, which recently reached new levels -- where the average duration of unemployment is now longer than the length of standard unemployment benefits (source: bloomberg) -- the single most important determinant of inflation seems ever distant to recover.

Interestingly however, given his track record for accurate foresight -- [paraphrased] "The subprime problem will be largely contained, and its risk, in scope to the rest of the mortgage market and economy, limited" (Bernanke, 2007), Chairman Bernanke, in his latest FOMC speech, described the risk of deflation going forward as, "negligible."  Again, if history is a guide, considering the state of employment and credit growth, this statement could be interpreted as an inverse indicator.

Total Credit Market Debt of U.S. Issuers as % of GDP [Credit:DoubleLine Capital LP, Federal Reserve, Commerce Dept]  [note: circa 2008 = 1st reversal in otherwise consistent positive 1st derivative in 2+ decades]

Recommendation In the face of large fiscal deficits brought about by war, transfer payments and stimulative policies, the 1930s (till past mid 1940s) were characterized by yields on the long end that continued ever lower, as yields at all maturities rode a general decline. The decline in short maturities can be explained by concerns of protection of principal and a general lack of liquidity in a deflationary environment. The concurrent steady decline in long maturities can be explained by investors' appetite for yield amid an environment of low growth and inflation. To the benefit of the current long bond, if inflation continues to print lower and growth expectations resume a downward trend, yield curve dynamics may mirror the prior deflationary example. As pointed out by Irving Fisher, the dollar's value, or purchasing power, will rise during deflationary times.  The U.S. dollar stands to benefit in intermediate phases of deleveraging, as already seen in 2008-09.  Given that the vast majority of outstanding debt globally has been issued and therefore denominated in USDs, the resulting deleveraging (or paying back of the debt) will create a demand for the currency that is in proportion to the unnerving rate of growth of debt during the pre-2008 period.  Crises of liquidity denominated in this currency will continue.  To meet debt obligations, risk-assets will sell off, including equities, commodities, and investment-grade/high yield corporate and municipal/local GO's. Discrimination between high and low qualities within any particular risk-asset class will prove, as seen in 2008-09, to be largely immaterial.  Cross-asset class correlations go to 1.0.  Because of their safe-haven status, long-term U.S. Government bonds will stand to benefit. Given low inflation or deflation, they should return a relatively high realized compound yield versus low inflation. More aggressive short term traders will seek the capital appreciation benefits of zero-coupon and discounted (off-the-run) U.S. Treasury bonds. In a period of debt deflation, markets become liquidity-driven; individual asset fundamentals are a distant secondary concern. Along with inflation and a return to pre-crisis growth rates, the market will weigh the risk involved with a long bond issued by a government that is running record fiscal and current account deficits. Given the record US Federal budget deficits, and the unavoidably large unfunded liabilities, a $113 Trillion worth ($1.13E14), or 909% of U.S. GDP, it is reasonable to assume that the first sign of weakness in markets from a perceived lack of confidence in the future value of the security will precipitate a wave of selling in Treasuries.  However, as always, the question is in the timing. The forgone conclusion of many participants is that it will be within months or a few short years. These participants overlook that officials will very likely place further temporary covers and/or guarantees while compromises are made on the future liabilities, thus staving off a sell-off in Treasuries for at least several more years, despite these deficits and imbalances. Moreover, a 10-year yield above 4% (and 30-year above 5%) will be deflationary with respect to burden on total interest expense, and rates will not rise further. Recent UST30YR trading behavior does not confirm a bearish scenario; there is a persistent bid at key technical levels. As total credit outstanding is still multiple standard deviations above the mean, Treasuries will rally intermediately as the dollar strengthens from further deleveraging. Total deleveraging, at least, will take several more years. Because of the fluid situation, however, we advocate an abbreviated time frame of 6-12 months; 30-years seem over-sold within the current range (2.5-4.75%). As debt as been borrowed and denominated in USDs, the long US government bond will benefit from capital appreciation and currency strengthening due to further deleveraging.  A short-term target (6-12 months) for US 30-YR (USGG30YR) is set at 4.1% (as of writing, it is 4.55%). -- Appendix:

[Credit: Eichengreen/O'Rourke]

M2 Index YOY% Change [Credit: Fed]

Commodity Index, Reuters Jefferies CRB [credit: Bloomberg]

--

*Title in homage to the original article written by Eichengreen/O'Rourke

**Ray Dalio is President/CIO and founder of Bridgewater Associates, the world's largest hedge fund, managing over $80 billion.  They specialize in currency/interest rate overlay strategies.

***Mr. Gundlach was a runner-up for Morningstar's Fixed-Income Fund Manger of the Decade.