Tuesday, November 14, 2006


October 2, 2006

The U.S. housing sector directly contributed more than $2 trillion to the national economy in 2005 and accounted for one-quarter of real GDP growth. Don’t expect a repeat performance any time soon. Data over the past five months show that the housing market is in the midst of a correction. The supply of detached resale homes has hit a 13-year high, affordability has eroded to late-1980s levels, and all three major housing markets – new, existing and construction – have absorbed double-digit declines in activity relative to last year.

The question is not whether the housing market correction will dampen U.S. economic growth over the next year. It will. The cliffhanger is whether it can single-handedly tip the economy into a recession. Indeed, the current housing cycle is already mirroring trends leading into past recession cycles. So why would it be different this time?

To evaluate the risk of an economy-wide recession, we cannot look at the housing sector in isolation. Other factors also enter into the equation, such as interest rates, inflation, labour market conditions, inventory overhang and the overall health of Corporate America. These influences bear little resemblance to patterns seen in prior pre-recession cycles, and we believe this tips the scale in favour of an economic slowdown rather than a recession.

The importance of housing in the American economy

The real estate sector has been punching above its weight in the American economy since the housing boom gained traction in 2002. More and more jobs, incomes and consumption have become leveraged to the performance of the housing market over the past four years – leaving little doubt that a housing correction will have knock-on effects to the broader economy. In fact, housing-related indicators alone leave the impression that a recession is just around the corner.

There are three main ways in which the housing boom has weaved its way through the economy. There is the direct link of residential investment, which accounts for 5.5% of real economic activity, a share that has risen a full percentage point in just four years. This sector has consistently contributed about half a percentage point to real GDP growth since 2002, twice its historical norm. Over the next year, however, the opposite is expected to be true, with residential investment shaving half a percentage point from annual economic growth. On its own, this would be a barely audible hiccup in the economic expansion, but residential investment has been a heavyweight in influencing recent labour market conditions.

Construction and real estate jobs are to credit for one-fifth of all American job growth in the past four years. This is remarkably disproportionate to its size in the labour market. For instance, the construction industry accounts for less than 5 per cent of all jobs. Not surprising, regions that had the greatest gains in home prices during the boom also experienced the most robust demand for construction workers. New England can thank the construction sector for more than one-third of all job growth over the past four years (August 2002-August 2006). The respective shares in California and Nevada are similarly high at 28 and 24 per cent. A pull back in housing, therefore, presents a clear and present danger to employment, and hence incomes and consumption across America.

And the impact from housing doesn’t end there. The third and biggest influence has come through two arteries of consumption: direct housing-related purchases and the wealth effect. The former includes the likes of furniture, appliances and other expenditures related to household services and operations. These purchases accounted for just over 18 per cent of the real economy in 2005. Meanwhile, housing wealth effects have been fueled by the rapid appreciation in home prices coupled with record levels of refinancing and home equity withdrawals. Unfortunately, the peak in refinancing has already long passed and if current declining trends continue through this year, refinancing activity will have slumped over 60% from 2003. Meanwhile, cashed-out home equity is projected to fall by 40% in 2007 alone!

The significance of this should not be overlooked. In the past three years, we estimated that housing wealth effects accounted for, on average, 2 percentage points of the annual growth in real consumer spending each year. The U.S. is now facing a situation where the unwinding of housing wealth effects will drag consumption growth. The second quarter of 2006 presents some preliminary evidence that this process is already underway. According to our proxies for capital gains and cashed-out equity, quarterly declines in both marginally detracted from real consumption growth. Even so, the overall wealth effect continues to support spending growth because the biggest driver is real estate assets valued at nearly $17 trillion in inflation adjusted terms. But even here cracks are forming. This measure expanded by just half a per cent in the second quarter, marking the slowest pace since 1997. As overall wealth effects continue to reverse course, we estimate that it could shave at least a full percentage point from real GDP growth over the next 12 months.

How close is the U.S. to a recession?

With all these dire predictions, would there be an outright contraction in U.S. economic activity? There is significant risk of a recession in 2007 or early 2008. However, the most likely outcome remains a mid-cycle slowdown. Although a number of housing indicators are mirroring the path of past recession cycles, the data provide an incomplete picture and can often send false signals. Interest rates, inflation and employment are also material in shaping the economic landscape.

Starting with the housing market backdrop, the news is somewhat glum. There is a strong link between real estate assets and recessions. The annual growth rate in inflation adjusted real estate has contracted in three of the past four recessions. On average, asset growth hits a trough of -3% following several quarterly declines that either preceded or coincided with the start of the recession. Currently, this seems a long way off since real estate assets were still expanding by a healthy 7% annual clip in the second quarter. However, the slowdown in the quarterly growth rate is a red flag. In addition, the annual pace of growth for detached real home prices moved deep into negative territory (-5.3%) in August, which does not bode well for the wealth measure in the third quarter. But, readers should bear in mind that there are also instances when price growth dramatically slows or turns negative that does not coincide with recessions. The mid-1990s offers a perfect example of repeated false signals, though a mid-cycle slowdown did ensue.

Sharp double-digit declines in the annual growth of housing starts in August also put this indicator in bad company for either a recession or cyclical slowdown. But, the data is inconclusive on which one it would be because single-family housing starts contract steeply in both cases. The only distinguishing feature is that activity during a recession cycle remains in the red for a longer period of time. For this outcome, we’ll have to wait and see, but what we do know is that construction activity in the most recent quarter has already contracted to a greater extent than the initial backslide leading into an average recession cycle.

Another possible signpost that the U.S. is on the cusp of a recession is if purchases of big-ticket housing items begin to contract on a quarterly basis. Falling expenditures on items like furniture and household equipment tend to precede a recession by 1-to-3 quarters. This has yet to occur in the present economic cycle, but we would not be surprised if it did. However, the bigger economic determinant is if a contraction in spending for non-housing items follows in toe. When this occurs, a critical threshold of consumer confidence has been breached.

Consumers hold the key to economic expansion…

In past recession cycles, consumer spending was undermined by three critical variables: inflation, interest rates and unemployment. There is good news on all three fronts in the current economic cycle. The graph below shows that present inflation behaviour is more akin to periods of mid-cycle slowdowns than recessions. U.S. core inflation typically accelerated ahead of recessions and was at least double current levels. Because movements in inflation and interest rates are highly intertwined, acceleration in the former usually prompted the central bank to respond with higher interest rates – often causing monetary policy to overshoot to the point of choking off domestic demand. This is evident in the graph on the following page, where the real fed funds rate climbs, on average, about 2 percentage points over the course of four quarters preceding a recession. The current cycle has produced an equivalent amount of tightening, but it has been gradual by comparison, occurring over eight quarters. Equally important, the level of real interest rates remains a full 2 percentage points below peak pre-recession periods. So in both cases, interest rates appear to be only tapping on the brakes, rather than driving the economy to a full stop.

This may account for why household credit is still showing little sign of stress, even at this late stage in the monetary tightening cycle. There has been much made of recent up-ticks in foreclosures and delinquencies, especially in states that have a notorious reputation for the use of riskier debt products, such as California and Nevada. Yet, the national level of delinquency rates remains low. And, in the subprime market, California and Arizona still ranked as the lowest delinquency states in the U.S. even though both have a disproportionate amount of non-conventional mortgages. In the first quarter of 2006, negative amortization loans represented 9 per cent of new mortgages in the U.S. In Nevada and California, those respective shares were 22.5 per cent and 21 per cent.

With this in mind, we fully recognize that the high use of ‘exotic’ mortgages does present a near-term risk to the economic expansion. In 2005, nearly 40% of new mortgages were either negative amortization or interest-only products. What’s more, it is estimated that $1 trillion of adjustable rate mortgages will adjust in 2007, or about 8% of outstanding mortgage loans. However, once again we take comfort in the inflation-interest rate link. Without the shackles of high inflation in the current cycle, we believe the Fed will react quickly to an economic slowdown and will cut interest rates by 100 basis points in the first half of 2007. This should help alleviate the interest rate strain on some of the more vulnerable credit holders.

The labour market backdrop is the third factor supporting the outcome of a mid-cycle slowdown. The final scenario graph below has two useful insights. First, the current labour market looks to be paralleling the path of a cyclical slowdown, referring to the fact that there has not been a material deceleration in job growth that typically precedes a recession. Meanwhile, growth in real wages per employee has recently accelerated, which is opposite to pre-recession behaviour. Second, the amount of new jobs relative to total employment is much lower in the current cycle than either of the scenarios of a recession or mid-cycle slowdown. This is good news because less job buildup going into a slowdown provides some insulation against the risk of mass job layoffs.

This possibility is further enhanced when we consider that the job boom that typically takes place in post-recession periods did not occur after the 2001 recession. Rather, corporate restructuring scratched nearly 3 million jobs from payrolls between 2001 and 2003. And, in some sectors, employment restructuring has yet to cease. For instance, the manufacturing sector continues to shed jobs five years after the official end of the recession.

Current lean employment structures provide some reassurance that aggressive job cuts are not waiting in the wings. This is a vital component to any economic cycle, because mass layoffs cause a sudden interruption in household income that is highly destabilizing to an expansion. Total wages in the economy amount to over $4 trillion, which is more than five times the amount of mortgage equity withdrawal.

…but Corporate America is in driver’s seat

But, the security of job growth ultimately hinges on the response by Corporate America to an economic slowdown. Even here we find a number of key positive features in the present cycle. For one, the ratio of prices to unit labour costs is typically falling heading into a recession period. This is definitely not the case in the current cycle, plus the ratio is at a historical high reflecting the fact that businesses have successfully constrained unit labour costs to lift profits. Second, corporate liquidity and savings have never been better. If the U.S. economy was nearing a recession, growth in retained earnings would normally be slowing, but the opposite has occurred. Given that retained earnings are at a record level, firms should not find themselves in dire straights as demand growth tapers off, especially if the slowdown extends a short period of one year, as we predict. And, as we already mentioned, job restructuring occurred only three years ago, so there’s not much fat to cut off these bones.

The final corporate variable that provides comfort is the historically low inventory-to-sales (I/S) ratio. Inventory swings have proven to be highly disruptive during an economic downturn and can often hasten job layoffs. When consumer demand growth trails off, I/S ratios tend to rise, causing firms to aggressively liquidate on-hand stock. The Fed estimated that this inventory adjustment process accounted for, on average, one-quarter of the overall slowdown in real GDP growth during postwar recessions. This is quite a powerful influence from a sector that represents a mere 0.5 per cent of GDP.

In the current economic cycle, I/S ratios across all sectors – retail, wholesale and manufacturing – are already flirting with record lows. In fact, these ratios have been extremely lean since 2004, the start of the Fed tightening cycle. So, it’s quite possible that firms have long been bracing for an economic slowdown. Low I/S ratios relative to past norms should limit the severity of an economic downturn, while also mitigating some of the negative risks that naturally flow to the job market.


We are not blind to the obstacles faced by the U.S. economy, such as a negative savings rate and high debt service costs. If the consumer back finally breaks, look for preliminary signals to emerge in the form of steep contractions in housing-related expenditures and escalating debt defaults. But, we remind readers that there is no single variable that can push America into a recession. The health of the economy is dependent on a confluence of many factors. Although recent trends in the housing sector are closely paralleling events leading into past recessions, deteriorating wealth effects and falling home prices alone do not portend a recession. The environment for interest rates, inflation and corporate balance sheets bears no resemblance to past recessions. We believe these factors will dominate, preventing the mass job losses that would compromise income growth and the ability for households to meet debt obligations and future expenditures. So, to return to the question we posed in the introduction as to whether the housing correction will lead to an outright recession, the likely answer is “no”. The economic indicators are precisely in line with previous mid-cycle slowdowns, such that the U.S. economy is more likely to bump along the 2 per cent mark for a one-year period, than contract outright.

Beata Caranci, Senior Economist

No comments: