The Greater Moderation
“The record of failure to predict recessions is virtually unblemished,” was the IMF’s conclusion following a 63-country study. In contrast, a scrutiny of ECRI’s track record prompted The Economist to declare – prior to our 2008 recession call – that “ECRI is perhaps the only organization to give advance warning of each of the past three recessions; just as impressive, it has never issued a false alarm.”
In line with the old adage, “never say never,” our September 2011 U.S. recession forecast did turn out to be a false alarm. As we’ll explain in a moment, we’ve good reason to have waited to say this, but there’s a more important question. In light of our track record, does that false alarm suggest that recessions cannot be forecasted?
We’ve examined the issue at length, and come to the conclusion that recessions can indeed be predicted. The reason that the 2012-13 cyclical downturn turned out to be the worst “non-recession” in half a century – rather than a full-blown recession – is because of something that’s unlikely to be repeated.
And no, we aren’t referring to the Fed’s actions, which reflected our assessment of cyclical risk. We’re referring to what one might call the “Greater Moderation.”
Windows of Vulnerability
But first, please recall that ECRI’s recession forecasting approach is quite different from standard practice. Because our methods aren’t based on back fitted models, no re-jiggering of models is called for as the economy evolves. This is why we’re able to use our framework for economies around the world that may be quite different in structural terms.
Since most economists think recessions are caused by shocks that then propagate through the economy, they often focus on “nowcasting,” i.e., being alert to the economy’s descent into recession following the shock. In contrast, our framework, based on three generations of research, postulates that endogenous cyclical forces periodically open up windows of vulnerability, as detected by multiple leading indexes.
Once the economy enters such a state of cyclical vulnerability, an exogenous shock can easily tip it into recession. Because such shocks tend to arrive sooner or later, an economy’s entry into a susceptible state – as signaled by the endogenous cyclical forces subsumed in our leading indexes – is almost always followed by recession. It’s in this sense that our leading indexes allow us to “predict” recessions.
A recessionary shock, in principle, can arrive from any source. For example, in modern Japan, they’ve been both natural, like the Tohoku earthquake in 2011, and man-made, like the 1997 and 2014 sales tax hikes. Indeed, the continuing efficacy of ECRI’s approach was showcased by our 2014 recession call in the face of Abenomics (ICO, August 2014), following our earlier warnings, in stark contrast to the near-universal surprise at Japan’s relapse into recession.
The U.S. stock market crash in 1987 was another exogenous shock, but in the weeks following that crash we publicly ruled out a recession since the U.S. economy wasn’t in a window of cyclical vulnerability. Similarly, with many economists expecting Hurricane Katrina to trigger a contraction, we had to explicitly declare that the economy was “unlikely to be tipped into a new recession” (USCO, September 2005).
Why did we wait to acknowledge that our recession call had been a false alarm, when the consensus had already come to that conclusion? For starters, the consensus had disagreed with every one of our recession calls – however accurate in retrospect.
One key reason we waited is that economic data, including GDP, tend to be substantially revised years later. For instance, in July 2013 – more than five years after the fact – U.S. GDP growth for Q1 2008 was revised down to -2.7% from -1.8%, having been initially reported as +0.6%. Similarly, in August 2014, German GDP was revised to show back-to-back declines in both Q4 2012 and Q1 2013; and in December 2012, revised data showed New Zealand GDP declining in Q3 and Q4 of 2010, belatedly vindicating our real-time recession call. Given the proven robustness of our approach, and the substantial data revisions that routinely occur with multi-year lags, we had expected that such revisions would validate our U.S. recession forecast.
And there was another important reason: we wanted to understand why our approach to business cycle forecasting – having enjoyed unrivaled success for decades – had trouble, in this instance, distinguishing between a serious cyclical slowdown and a business cycle contraction.
In the fall of 2011, our leading indexes correctly showed the economy to be in a clear window of vulnerability. So what happened to the exogenous shocks?
In recent decades, U.S. recessions have typically been triggered by oil shocks, often in the wake of Fed rate hikes. That the Fed wouldn’t hike rates was clear in 2011, but we had no reason to believe that other kinds of shocks would be absent. While not predictable, such shocks would naturally amplify the cycle volatility engendered by endogenous cyclical forces, thereby resulting in cyclical downturns of recessionary magnitude.
Prior to the Lehman collapse, we had observed (USCO, August 2008) that U.S. trend growth had been stair-stepping down at least since the 1970s, and we expected this pattern to persist following the Great Recession. Certainly, that’s evident in terms of GDP growth (Chart 1). But also, given the resurgence in cyclical volatility following the end of the “Great Moderation” of 1985-2007, we expected the years following the Global Financial Crisis (GFC) to see at least moderate cyclical volatility which, in combination with low trend growth, would more easily take growth below zero.
But, unexpectedly, the cyclical volatility of U.S. economic growth – as measured by the three-year centered standard deviation of year-over-year (yoy) growth in ECRI’s U.S. Coincident Index (USCI) – plummeted below 0.6 by late 2011 from a three-decade high the previous year. It has now stayed below that threshold for two straight years – the longest stretch on record (Chart 2), having dropped in Q4 2012 to a low surpassed only in 2005. This goes way beyond a return to the Great Moderation, all the way to the Greater Moderation – and it’s important to understand why.
The critical clue comes from crude oil price volatility which, as Chart 3 (top line) shows, had collapsed by 2013. As the former head of BP’s global economics team noted last year, “the oil price has been above $100 for three years in a row, the highest … such period ever, but extremely stable, the lowest three-year volatility since 1970,” when prices were fixed. While there were supply disruptions, he said, “[t]he cumulative level of these disruptions over the last three years is balanced almost one by one, almost barrel by barrel, by the increase in tight oil production in the U.S. So it’s an absolute fair statement to say [that] if we had only had the disruptions … you would have seen oil prices shooting up.”
Thus, despite a fairly normal pattern of supply disruptions, oil price volatility fell to a four-decade low in 2011-13. This happened, in large part, because of what’s been called the fastest ramp-up in oil production in history, creating an unusual period devoid of oil shocks.
Meanwhile, the Fed launched the open-ended program of monetary easing dubbed QEternity in the fall of 2012 – as GDP growth, we now know, neared zero, following Operation Twist, announced a year earlier on the same day that ECRI made its recession call in September 2011. The Fed’s actions effectively expressed its agreement with ECRI’s recession forecast – in deeds, if not in words – first in September 2011, and then when going all out a year later. Additional international monetary easing launched by the ECB in December 2011 added to the torrent of global liquidity, as did Abenomics a year or so later.
With the trillions in international QE amounting to not just a “Fed put,” but a “global central bank put,” for equities, stock price volatility hit a six year low in 2012 (Chart 3, bottom line), consistent with our observation early that year that “the world’s major central banks are flooding the financial system with liquidity, and monetary easing is often greeted with bullish moves in risk assets” (USCO, January 2012). No wonder stock prices kept climbing despite the very real cyclical downturn in U.S. economic growth.
Because of the singular absence of oil shocks, abetted by the drop in stock price volatility, economic cycle volatility plunged, and stayed at extreme lows for the longest period on record. This happened even as trend GDP growth clearly downshifted further, validating our 2008 recognition of slowing trend growth that underpins our “yo-yo years” thesis.
Even though endogenous cyclical forces propelled the economy into a window of vulnerability, the conspicuous absence of exogenous shocks meant that the period around late 2012 to early 2013 – which we later characterized as the “epicenter of recession” – is unlikely to be classified as a recession because the key coincident indicators that define recession failed to decline in the pronounced, pervasive and persistent manner characteristic of a business cycle recession.
Yet, the downturn we forecast did turn out to be the worst “non-recession” in over half a century. Indeed, the smoothed growth rate of the USCI – a more comprehensive measure of economic activity that subsumes those indicators (Chart 4) – fell in January 2013 to a low never matched away from a recession before the Great Recession, except during the 1952 and 1956 steel strikes. Separately, a Gross-Domestic-Income-based measure of the economy’s “stall speed” introduced by a 2011 Fed paper dropped below 1% by Q3 2012, having never before fallen below that threshold away from a recession (not shown).It’s a testament to the enormity of the Fed’s mounting concern about deteriorating cyclical conditions that it kept ratcheting up its QE during the 2011-13 period.
But key to avoiding recession in 2012 was the fortuitous plunge in oil price volatility that resulted in a highly unusual absence of oil shocks, despite ongoing supply disruptions. With the “central bank put” squashing stock price volatility, as desperate central banks around the world took turns opening up the QE spigots as never before, exogenous shocks emanating from the stock market were effectively kept at bay; and still the worst non-recession in half a century took its toll on the U.S. economy.
Can oil shocks continue to be avoided in the coming months and years? History doesn’t encourage such a belief. And whether or not the “central bank put” is still in place, the nascent uptick in oil price volatility (Chart 3, top line) raises the risk that the Greater Moderation in the volatility of economic growth will prove fleeting.
Forewarning Beats Foreboding
Because past Fed rate hikes’ lagged impact slowed the economy to a point where endogenous cyclical forces could more easily open up a window of vulnerability, recessions have often eventually followed. But, with or without Fed rate hikes, progressively lower trend growth has now made it easier to open up a window of vulnerability. Incidentally, the U.S. experienced 28 recessions in the 117 years before the Fed was created, and 11 recessions in less than half a century thereafter, i.e., just about as often.
The critical implication – with secular stagnation or some variation thereof now being widely acknowledged – is that most major developed economies, including the U.S., are operating closer to the edges of their windows of vulnerability. So, unless the Greater Moderation ushered in by rock-steady oil prices becomes a reliable feature of the economic landscape, we may be closer to the next recession than many believe; the only question is when it will arrive.
This makes it more important than ever to judge the timing of their entry into those periods of heightened cyclical risk. While we’re all painfully aware that there’s no perfect forecasting method, ECRI’s approach to predicting cyclical turning points is better than any other we know about.
What counts in this fraught environment is not foreboding, but forewarning: specifically, forewarning of the next time the cyclical window of vulnerability begins to open up.