The yield curve became more inverted this week, with the negative differential between the 3-month and the 10-year at -49 basis points and a -76 basis point differential between the 10-year and the Feds fund rate. According to a Fed paper, that level of an inversion suggests there is now an over 40% probability of recession next year. This same model only predicted a 50% chance of recession in 2000, and as the paper authors acknowledge, the model probably understates risk in recent decades.
The yield curve and interest-rate data looks like this:
So the question for those who believe markets are future discounting mechanisms: Which market are you going to believe: Stocks or Bonds?
Meanwhile, Merrill Lynch's North American Economist David Rosenberg observes that: "Our recession-risk indicator puts the probability that there will be an economic downturn in the coming year at 51%. That's the highest level since the 2001 recession."
Now, its one thing for our little blog here to say a recession is becoming increasingly likely -- I'm more in the 60% camp myself. Indeed, back in the Summer of 2005, we identified that a late 2006/2007 recession was increasingly possible.
But Merrill Lynch manages a trillion or so dollars, and its worth noting when their economist foresees an increased possibility of recession.
Here are the factors Rosenberg identifies as influencing his thinking:
• "The Institute for Supply Management's [ISM] purchasing managers' index came in at a weaker-than-expected 51.2 for October, down from 52.9 in September. The consensus was 53.0."
Rosenberg notes that in the past, whenever the ISM index hit 51.2 while in a downtrend, it eventually broke below 50 -- the dividing line between expansion and contraction in industrial activity. Rosenberg adds that "once it does that, history shows that the Federal Reserve cuts rates within two months." Therefore, watch the 50 level on ISM to help forecast the next rate cut.
• "Our recession-risk indicator puts the probability that there will be an economic downturn in the coming year at 51%. That's the highest level since the 2001 recession."
This is a proprietary indicator, so I plead ignorance as to its components. Rosenberg does note that historically in all but 2 times, it accurately forecast a recession (knowing out of how many signals would be helpful). The average lead time between the penetration of 50% and a recession was 12 months, and the median was 10 months. That would put the ML recession (only 51%) somewhere in September-November 2007, and the market reaction sometime between March and May 2007.
• "We think GDP will grow at an annual rate of roughly 2% for the fourth quarter, and the risk is on the downside in light of the weak "handoff" in terms of macro momentum as the quarter began and what we know so far anecdotally about spending in October."
• "Our 2007 GDP growth estimate of 2% hasn't changed, but keep in mind that our forecast assumes that the Fed will cut the funds rate to 4% next year. Without that stimulus, growth could well be closer to 1.0% to 1.5%. That would be weak enough to be in the hard-landing zone that sees a sharp reversal in the unemployment rate and the emergence of credit problems"
This is premised on a few factors: No major new stimulus, externalities, oil shocks, etc.