In our view, such fears are misplaced. Some of these factors are statistical, some are fundamental, but we believe that all are temporary. As a result, they do not change our view that the recovery is sustainable. And they will probably set the stage for a powerful 4% spring snapback in the economy - one that could surprise with its force and for which markets are ill prepared.
Reviewing the case for sustainable growth. It's worth quickly reviewing the four factors we believe are promoting sustainable growth through 2011: 1) Monetary policy has fostered improving financial conditions; 2) The impact of fiscal stimulus will last through 2011; 3) Strong growth abroad will lift US exports and earnings; and 4) Economic and financial excesses are abating. In our view, most of these are playing out according to script. Markets are functioning, although the demand for credit remains weak. Spring tax credits should support consumer spending, and infrastructure outlays are only starting to show up. Exports are booming, earnings are beating expectations, and production is still catching up to final sales, while companies are reducing capacity and inventories.
To be sure, headwinds to growth remain significant: Housing imbalances, though smaller, persist, and 2 million foreclosures are coming; job gains are still a forecast; and policy uncertainty clouds the outlook. The balance between those headwinds and tailwinds has kept our forecast for real growth in 2010 at 3.25% (4Q/4Q basis) for more than a year.
Classic consolidation. However, incoming data lately have weakened: New and existing home sales continued to slide, construction outlays and ‘core' durable goods orders and shipments turned down, and one measure of consumer sentiment tumbled in January. Vehicle sales slipped in February, and initial jobless claims have risen appreciably since the start of the year.
In our view, this disappointing string of data partly represents a classic consolidation in the pace of recovery following the late 2009 surge. It does not represent the onset of a slowdown, much less a double dip. Recoveries never go in a straight line; even strong ones are characterized by surges followed by pauses in their early stages. The economy, especially at turning points, is far more volatile than any of our smooth forecasts anticipate. The culprits include the vagaries of the inventory cycle, the fact that traditionally lagging components of demand, like capital spending, are still contracting while others are growing, and the uneven effects of stimulative policies employed to promote recovery.
For example, we illustrate the contribution of inventory swings to the growth in real GDP in the early stages of past US recoveries. While on average, such swings show a time-honored cyclical pattern - they account for about a third of the advance in GDP in the first two quarters - the variation from one episode to the next is striking. To be sure, some of those differences over time reflect structural change in the economy - the birth and death of industries, the adoption of ‘just-in-time' inventory management techniques, and the advent of global supply chains. More importantly, the variation across cycles reflects cyclical factors like the extent and nature of the shocks that triggered the recession and whether inventories were top-heavy or lean going in.
In addition, if anything, the economy is actually more volatile than current estimates show, because they are based on extrapolations of indicators that are only available with a lag. Successive revisions of GDP and associated aggregates tend to be more volatile than the preliminary estimates, because the Bureau of Economic Analysis substitutes the actual data in place of the trend extrapolations initially assumed. For example, BEA calculates that the standard deviation of ‘advance' estimates of quarterly GDP from 1993 to 2006 was 2.7%, but in the first annual revision, that variance rose to 3.9%.
Policy paybacks. Two policy changes temporarily boosted demand last year: ‘Bonus depreciation' investment incentives and a first-time homebuyer tax credit. When these ‘use-it-or-lose-it' policies expired, demand slipped below what it otherwise would have been. These factors add volatility to the results: Just as the transitory demand increase represented an overshoot, the expiration of these incentives will trigger a payback when demand undershoots the underlying pace.
Bonus depreciation incentives may have contributed to a 4Q surge in capital spending, especially in trucks, at the expense of 1Q. Real capital spending on light and heavy trucks accounted for nearly 40% of the 4Q surge in real equipment and software outlays, and the payback in 1Q is depressing growth. Fortunately, the infrastructure spending mandated by the American Recovery and Reinvestment Act of 2009 (ARRA) started to kick in late last year, and contractors have begun to order and replace construction equipment and trucks to meet that coming demand.
The first-time homebuyer tax credit is more complex. Despite its extension/expansion, a payback in housing demand seems to be underway, at least judging by the slide in sales and pending home sales through January. The impact of the credit was most prominent in existing home sales, which soared through November and then plunged in December and January. But gauging the payback requires a counterfactual baseline to calibrate the impact of the credit on demand. That is difficult, because the US$8,000 maximum tax credit represented only about 5% of the price of homes typically purchased by first-time homebuyers and because there is scant basis for comparison - there is only one example of a similar credit (in 1975, and that was for new homes only).
According to the National Association of Realtors (NAR), about 350,000 of the 1.8-2 million buyers who claimed the credit last year would not have purchased a home without it. But it is uncertain how much of that was genuine additional demand and how much was simply brought forward. Traditional measures of affordability soared, courtesy of the plunge in home prices and in mortgage rates. Last year lenders demanded bigger down-payments, and with the credit not available until the deal closed, down-payments and credit availability remained hurdles for many buyers, especially first-time ones. Those terms of lending have since eased a bit, judging by the improvement in the Fed's Senior Loan Officer Survey.
Another complication in calculating the payback from the expiration of the initial homebuyer tax credit is the impact of the second tax credit that was enacted in November. Congress passed new legislation that extends the credit for first-time buyers and expands it to cover current homeowners purchasing a new or existing home (up to a maximum credit of US$6,500) between November 7, 2009 and April 30, 2010 (the purchaser will have until July 1, 2010 to close). Current homeowners must have used the home being sold or vacated as a principal residence for five consecutive years within the last eight. Married couples with incomes up to US$225,000 are eligible for the maximum credit, higher than the US$175,000 under the old credit. We assume that the new credit will promote a healthy pick-up in sales as the April 30 signing deadline approaches (and sales of existing homes will likely rise with the approach of the June closing date).
Severe weather. We knew when Punxsutawney Phil saw his shadow back in early February that the rest of the winter would be tough, but this one has been unusually harsh. That's been especially the case in the Northeast and Mid-Atlantic states, which account for 24% of US GDP and 19% of non-farm payrolls (measured on a sum-of-states basis). The back-to-back storms that hit the Northeast and Mid-Atlantic states on February 4 and February 9 appear to have had a major impact on jobs, hours worked and income. As Dave Greenlaw notes in a companion piece, we estimate that, absent special factors related to the census and the weather, payrolls likely would have risen by more than 100,000 (see Harsh Weather Masks Improvement in Jobs and Hours, March 5, 2010).
Snowstorms probably hobbled sales of light vehicles and retailing activity in February as well. Here the impact may be smaller, because the sales figures measure activity over the entire month, in contrast to the snapshot of payrolls taken in the week of the storm, so some bounceback during the month is possible. Nonetheless, industry sources suggest that weather trimmed vehicle sales by 0.75 million annualized in February. And our retail analysts estimate that, although sales were strong in the month, they might have been 100-200bp stronger but for the weather. Not surprisingly, the effects on outdoor activities like construction and on industrial production have also been significant. Based on the decline in construction payrolls and workweek (hours tumbled 2.4% on the month), we estimate that housing starts sagged last month. And the 1% plunge in manufacturing hours worked suggests that industrial production declined temporarily in February (plant shutdowns for safety recalls at one manufacturer also depressed production).
Spring rebound coming. The important point is that the three factors depressing 1Q growth have largely been temporary. Fundamentals still point to sustainable growth and are gathering pace. Consequently, as the temporary factors dissipate, we believe that a spring rebound is coming and could surprise with its force. Indeed, we'd argue that the weaker is 1Q, the stronger will be the snapback.
Underneath the weather, a stronger-than-expected report. February's payroll report showing a loss of 37,000 jobs came in better than our estimate (and the consensus) of -65,000. This is especially notable because the negative weather-related influence on the February data appeared to be even more powerful than we had expected. We had guessed that harsh weather depressed payrolls by about 100,000 and the workweek by 0.2 hours. But, as noted below, it appears that the weather impact was significantly larger. Moreover, the add-on from new census workers was somewhat smaller than anticipated. Absent special factors related to the census and the weather, we believe that February payrolls would have risen by more than 100,000.
February's data confirm our belief that labor market conditions are improving on an underlying basis. In particular, the Challenger survey of layoff announcements, released March 3, shows clear signs of a significant underlying deceleration in the volume of layoffs. Meanwhile, the ISM surveys point to improved demand for labor in both the manufacturing and non-manufacturing sectors. And the ADP survey is also showing steady progress towards improvement. Admittedly, the jobless claims figures have shown some surprising elevation of late, and if sustained this could represent an important signal. However, as we have noted in recent weeks, some strange gyrations in California - which seem related to claims-processing bottlenecks - and the unusually severe winter storms in the East appear to have played a role in the recent upward drift in jobless claims.
It now appears that the unemployment rate peaked in October. The unemployment rate held steady in February at 9.7%, as the household survey showed employment growth of 308,000. Up until this point, we had thought that a rebound in the labor force participation rate would trigger a move back above 10% for the jobless rate. However, we are now inclined to believe that the rate peaked at 10.1% back in October. While the rate could tick slightly higher as discouraged workers again search for jobs, we believe that the jobless rate will begin to drift somewhat lower - in an irregular fashion - over the course of coming months.
Weather impact exceeded our estimate. The "not at work due to bad weather" series contained in the household survey - our favorite proxy for weather-related influences on the payroll data - soared to 1.1 million in February. This is the highest February reading on record (the data stretch back 1976). And for any month over the course of the last 28 years, it is surpassed only by the 1.9 million posted in January 1996 (the so-called ‘Storm of the Century'). We were expecting a reading closer to 650,000 in February and believed - based on some simple historical correlations - that this translated into about a 100,000 subtraction for February payroll employment. Thus, it appears that the subtraction was even larger than we had anticipated - perhaps as much as 150,000.
Corroborating the weather impact. To cross-check this estimate, we looked to a US Commerce Department compilation of severe winter storms stretching back over the past 55 years, and checked job losses in episodes similar to February's. Government statisticians have developed ‘The Northeast Snowfall Impact Scale' (NESIS) to gauge the severity of storms. This measure takes into account both the amount of snowfall and the size of the affected population (more information is available at http://www.ncdc.noaa.gov/snow-and-ice/nesis.php).
We have added our own rough approximation of the weather-related impact on private payroll employment to the table in the full report (see the column labeled ‘Payroll Impact'). These figures are derived by simply subtracting that month's reading for the change in private payroll employment from the average of that month and the succeeding month. The basic assumption is that payroll growth is stable over the very short run and that all the jobs impacted are recouped in the very next month. Private payrolls are used because of the potential for short-run volatility in the government category. Also, we show the change in employment rather than the percentage change because we suspect that the degree of weather sensitivity has diminished over time. For example, a payroll impact of -100,000 in 1960 might translate to about the same impact today if we assume that the rise in the level of employment has been about offset by reduced sensitivity to the weather. An ‘na' in the table indicates that the storm occurred too far outside of the survey period to have a meaningful effect on employment. Finally, the asterisks in the last column designate the events that we view as most comparable to February 2010.
In choosing the events that are most relevant to the current situation, we leaned towards selecting those that occurred in February (although we included one in January) and those that hit shortly before or during the survey period (note: the survey period is the week that includes the 12th day of the month). The average NESIS of the seven similar events is 5.75 and the impact on payrolls is approximately -100,000.
We believe that the impact of the February 2010 storms was even larger, because two separate events - ranked 20 and 25 in the table - likely had an impact. While the average NESIS of the selected events was higher than the index associated with either of the February 2010 storms, the combined potential impact of the storms is one factor that makes the current situation somewhat unique and implies a larger impact on payrolls than we originally assumed.
Caveats. It's important to keep in mind that we are looking at historical data that have been smoothed by countless revisions over the course of the years. For example, the estimated employment impact of +21,000 for the February 2003 storm would be more like -100,000 if we had used the originally reported data. We checked a few other observations, and most were reasonably close to the current calculated value, but it's still important to recognize that revisions can be significant and the originally reported data might overstate the effect of the storm.
The average workweek in February slipped only a tenth of a point (to 33.8 hours). However, the modest dip partly reflects the changeover to a new series that the BLS introduced last month which includes managers. The old series, which covered only production and non-supervisory employees, declined 0.2 hours in February. In contrast, the series dropped 0.4 hours in the much more severe storm in January 1996 and recouped all of this move in the following month. So, it appears that the length of the workweek may be starting to rise on an underlying basis.
Temp help employment jumped another 48,000 in February - in line with the trend that has been evident since last October. Increased demand for temp help workers and a lengthening workweek are two key leading indicators of labor demand.
Census impact less than expected. The government added 15,000 census workers in February - versus our expectation of +35,000. However, we should see a sharp ramp-up in census hiring over the next few months. In fact, based on past patterns, we expect more than 100,000 new census workers to be added in March.
We believe that policy uncertainty - particularly related to healthcare - continues to act as a restraint on hiring. However, it appears that some of the uncertainty may be addressed relatively soon. We note that the Senate-passed version of healthcare reform is not particularly onerous in terms of its requirements for employees of small businesses.
In all, we suspect that labor market conditions will show further progress towards improvement in coming months and look for average monthly payroll gains (ex-census) of 150,000 or so over the balance of this year.
Bottom Line
From our standpoint, the inability to gauge the impact of weather-related distortions with any real degree of precision means that Friday's employment report is less important by itself than February and March taken together. In that context, next month's report could be shaping up as one of the most important releases in quite some time, since it will allow us to better gauge the impact of special factors and thus get a clearer sense of the underlying performance of the labor market.
The prospects for the Mexican economy have been steadily improving in recent weeks, as indicated by ongoing upward revisions to growth forecasts for this year. The finance ministry upgraded in mid-February its 2010 GDP outlook to 3.9% from 3.0%, citing a "significant pick-up in non-oil exports, auto production, other manufacturing activities, commerce and transportation". And when Banco de Mexico released its monthly expectations survey last week, the consensus of economists had upgraded growth expectations sharply to 3.9% from 3.3% a month earlier - the largest monthly revision since the central bank began surveying 2010 growth expectations.
Despite the consensus' mounting optimism about the Mexican economy for 2010, the trend of GDP growth upgrades is likely to have much further to go, providing support for investor sentiment ahead, in our view. But how much higher will revisions go? We would not be surprised to see the economy rebounding over 5% this year; indeed, we are revising our forecast to 5.2% from 3.8% previously, making Mexico the fastest-growing country among the seven-largest economies in the region. Our upgrade in great part reflects a more favorable comparison base after a stronger-than-expected 4Q GDP, during which the economy expanded 2.0% sequentially (non-annualized), nearly double the 1.2% guidance by the finance ministry. Meanwhile, incoming data - which is showing that industry is gaining ground at a rapid clip and that activity in domestic-focused sectors has finally stabilized - remains consistent with our positive outlook that the strong externally driven upturn will translate into a broader recovery over the course of 2010 (see "Mexico: The Industrial Fiesta", This Week in Latin America, January 19, 2010).
Deepest Slump, Sharpest Rebound
After experiencing a wrenching 6.5% contraction last year, Mexico is likely to lead the regional rebound in 2010, based on our newly revised forecast. To be fair, a great part of the 2010 rebound reflects a very easy comparison base. Mexico's recession diverged from the regional trend as the slump intensified to -24.9% annualized in 1Q09, after the sharp contraction in 4Q08 (-7.5% annualized). By contrast, the 1Q declines in Brazil and Chile eased to less than half the four-quarter pace, while Colombia actually posted positive growth in the first three months of 2009. The deep contraction in early 2009 combined with the V-shaped recovery in auto production and broader industrial activity in 2H09, in turn, mean that if GDP were to stall sequentially during 2010, the economy would still be 2.8% higher compared to 2009, according to our calculations. Accordingly, the consensus' current GDP view (3.9%) implies a sequential growth rate of just 0.4% non-annualized per quarter over the course of 2010 - an anemic pace that seems at odds with the growth prospects for US industrial output and the global economy.
A deeper look at the market's cautious assumptions for growth this year also seems to indicate some degree of disconnect with today's improved economic conditions. The consensus currently expects industrial output to grow 5.0% in 2010; to reach the GDP growth forecast of 3.9%, therefore, services - which represent nearly two-thirds of the economy - would have to grow near 3.3%. Due to the depressed 2009 base, we estimate that the statistical carryover into 2010 for services is 2.8pp, not far from the 3.3% required growth to reach the 3.9% GDP growth forecast. While it is possible that some service areas will stall during 2010, around 30% of services - including commerce, rail and truck transportation, which last year knocked off some 3.0% from total GDP growth - are closely linked to external trade and thus should continue to mirror the sharp improvement in industrial activity. Accordingly, the consensus seems to assume virtually no growth in services outside trade-related groups - an overly pessimistic assumption, in our view.
The Industrial Fiesta
Incoming data from Mexico's industrial sector indicate that recovery in production has been following a V-shaped trajectory, echoing the pick-up in US manufacturing since mid-year. The most recent December production report extended industry's run of sequential gains to four consecutive months, led by soaring manufacturing production, which rose at a sequential 24.2% annualized pace in 4Q09 - the strongest three-month jump since the post-Tequila normalization. Industrial exports, once seasonally adjusted, were more than a quarter higher in January than at their mid-year trough and just 12% below the mid-2008 all-time high. While the auto sector continues to lead the way - vehicle exports as well as light-vehicle output in January were back to historically high levels - other manufacturing sectors are catching up as well. Importantly, since the July bottom - nearly coinciding with the first signs of a pick-up in exports and output - hiring in industry has accelerated to over 9% annualized in the three months through January 2010. Not only has the current industrial recovery been accompanied by renewed hiring, but it has also coincided with important share gains for Mexico's exports in the US import market: by December 2009, Mexico's market share stood at the highest level since early 2002 and just shy of the most recent peak from late 2001.
The rebound in US manufacturing, importantly, seems to be gaining momentum more quickly than our US team had expected. Indeed, our US economists Ted Wieseman and David Greenlaw point out that growth in manufacturing is looking typically V-shaped coming out of recession as the rebound in the auto sector since mid-year that started the recovery continues to support factory activity across a number of sectors. Meanwhile, strength in exports and a continued catch-up in production to stronger final sales - as inventories continue to plunge - are also helping to make the growth more broadly based. The most recent payrolls data, as Ted points out, reinforce industry's positive outlook: manufacturing payrolls gained 1,000 in February - with a likely significant negative hit from the weather - on top of a 20,000 gain in January. Rising manufacturing jobs has been an infrequent occurrence since the late 1990s and they reflect how strong the recent growth has been. In addition to permanent gains to factory jobs in the past couple of months, a good portion of the record surge in temporary employment since October has probably been attributable to the manufacturing sector. Our US team expects industrial output to grow at an average sequential 7% annualized pace over the four quarters of 2010, which would be one of the best annual performances in the past 30 years.
Is the Recovery Finally Broadening?
While Mexico's recovery so far has been industry-centric, this strong external-led upturn is likely to translate into a broader recovery in domestic-focused sectors over the course of 2010, in our view. Nothing reflects the two-tier nature of the ongoing recovery better than confidence indicators: whereas optimism among the captains of industry is back to levels last seen in 2Q08, consumer sentiment remains deeply depressed. Our view that the recovery is likely to broaden depends heavily on the trajectory of labor markets: unlike the largely jobless recovery that followed the 2001 recession, today's manufacturing rebound is leading to immediate job creation with positive potential implications for income and consumer spending.
There are already some signs suggesting that the recovery is starting to become more broadly based. For example, the initial recovery in jobs since August took place exclusively in manufacturing. By November, however, sectors outside of manufacturing and construction began to add jobs as well, although at a modest pace and with temp workers having a disproportionate impact. This trend echoes evidence from surveys: in February the employment gauge from IMEF's non-manufacturing index moved back into expansionary territory (51.7), while the manufacturing hiring index has been above 50 since September.
Import demand has also posted a modest improvement, suggesting some firming in domestic demand. Though still at depressed levels, imports of capital and consumer (ex-fuel) goods, for example, were in January around 10% and 25% above their 2Q troughs, according to our calculations. And, even though consumers are unlikely to see positive real wage growth this year - due to the inflation shock caused by higher taxes and administered prices - our measure of wage mass has moved higher in recent months, reflecting gains in employment. The evidence available to date does not point to a sharp turnaround in total services - in fact, our conservative assumption is for a sequential expansion at about half the average pace of 2003-07; however, trade-linked areas such as commerce and transportation should continue following manufacturing's solid recovery in coming quarters. In addition, domestic-linked services appear to have already stabilized and there are some hopeful signs of improvement, suggesting that the consensus may be too cautious on this area of the economy as well.
Bottom Line
The prospects for the Mexican economy have been improving and the trend of upward 2010 GDP growth upgrades is likely to have much further to go, in our view, supporting positive investor sentiment. With Mexico's industrial recovery following a V-shaped trajectory and some signs that the strong externally driven upturn is starting to broaden into domestic-focused sectors as well - in a context of positive job creation - the Mexican economy is likely to expand over 5% this year, representing the strongest rebound among the region's major economies.
When the central bank last week announced its plan for US$1.26 billion in dollar purchases for the first half of the year, it was responding to significant appreciation of the Colombian peso in recent weeks. Indeed, the accompanying central bank statement noted that the exchange rate was "misaligned". However, while the central bank has additional policy tools at its disposal to try to address this ‘misalignment', we suspect that there are fundamental factors driving exchange rate strength. And while there are questions whether we will see any meaningful policy shift from the next administration - concerns that have arisen following the recent court decision ruling out a third term for President Uribe - the currency strength may be a signal of just how favorable the external environment facing Colombia has become.
Headwinds from Venezuela
Our view that Colombia faces a favorable external environment may come as a surprise, given that many Colombia watchers remain concerned about the headwinds from the deterioration in trade with one of Colombia's main trading partners - Venezuela. Indeed, over the past six years Venezuela has been Colombia's second most important export destination. And exports to Venezuela fell sharply last year, declining over 30%Y. But we suspect that rather than the bleak historical data, what worries Colombia watchers more is the prospect that there is further pain ahead on the export front as Venezuela's economy continues to deteriorate and as the impact of Venezuela's January exchange rate devaluation filters through.
While we acknowledge that Venezuela is unlikely to provide much cheer to Colombia's economy this year, we suspect that after last year's impact on Colombia's exports it has become a sideshow. Indeed, while exports to Venezuela fell near 30% last year, the impact was much more severe in the final months of the year. By December 2009, Colombian exports to Venezuela had fallen nearly 78%, making Venezuela largely irrelevant as a trade partner for 2010. Indeed, while Venezuela started 2009 as Colombia's second most important export destination, accounting for 16% of all exports, it ended the year with its share down to only 4.6% of all exports. Indeed, given the magnitude of the fall in Colombian exports to Venezuela last year, further declines this year may have a limited impact on activity.
Export Diversification
And while exports to Venezuela have suffered, we suspect that this year's export growth outlook is much improved as Colombia continues to diversify its export destinations. Indeed, Colombia's exports to the US - its main trading partner - as well as Latin America ex-Venezuela and ex-Ecuador, the Caribbean and non-Japan Asia, particularly China, have been on the rise (sequentially) over the past eight months. And given that the call from our colleagues in the global economics team remains for a sustained global economic recovery, we expect the benefits of Colombia's diversification of export destinations to continue to support both its economic rebound and currency strength.
Read Full Article »