Reading The Data Through The Storm

When things get to this point, one has to believe that the worst of the possible outcomes is quickly becoming embedded in securities prices. By the way, the consensus now seems to be that we will indeed go over the fiscal cliff on the first of the year. Then after each side makes its points, some type of compromise will be reached within the first couple of weeks of the calendar year, before the impact of the fiscal cliff really kicks in. Personally, I think calmer heads will prevail, a limited portion of the cliff items will be implemented, and a majority of these issues will be kicked down the road until at least mid-2013. Given my desire for slow and gradual changes, that isn't the worst possible outcome.

Besides the fiscal cliff, markets had a lot of other headwinds this week. Earnings news, especially out of the technology sector, was less than wonderful, giving investors an additional source of worry. I also believe that fear of higher capital gains rates, and the reality of higher investment income taxes next year may have caused some investors to cash in their chips this week, potentially weighing on stocks. Additionally, the stream of poor economic news out of Europe continued this week, with news of a second quarter of negative GDP growth and falling monthly production data.

This week Hurricane Sandy ravaged the economics reports, with retail sales, initial unemployment claims, and industrial production all missing expectations. However, below all this economic noise, I continue to believe that the economy is showing slow but steady improvement.

Hurricane Sandy Leaves Economists Flying Blind
Unfortunately, Hurricane Sandy has dramatically complicated the interpretation of monthly economic data. The storm-affected area represents 20%-25% of the U.S. population. The storm had effects as early as Oct. 26, as forecasters correctly warned of an impending storm. The actual storm came ashore Monday, Oct. 29, and many businesses were closed through the entire week ended Friday, Nov. 2. Even the New York Stock Exchange was closed for two full days. Keep in mind that just one business day represents about 5% of monthly business days. Five days of complete shutdown would amount to 25% of monthly output regionally and 5% nationwide (20% of the nation times 25%). The fact that the consequences straddled two months and came at the critical month-end period for shipments further complicates the analysis. Given most metrics move just a few tenths of a percent in a month, the storm-related data will swamp the raw data.

Muddling matters even further are month-end survey issues. Every month, the data is based on surveys sent out by the government and returned by businesses. However, more than the usual number of proprietors were unable to get to their businesses, let alone mail in surveys. So what data we do have will be subject to major revisions as more data is eventually submitted.

The good news is a good portion of the business is merely being shifted from month to month and not lost. Clothing purchases planned for October will probably happen in November, and the same probably holds true for auto sales. Other service-related items like theater tickets, hotel rooms, airline seats, and even electricity generation cannot be made up for at a later date.

My Plan of Attack for Examining Data
My normal methodology of looking at most data on a three-month average, year-over-year basis will help reduce the distortions, but even here I will need to cut these indicators a little more slack than usual. The most critical data for my analysis of the economy has been weekly retail sales and inflation. My thesis has long been as consumers continue to spend, businesses will eventually have to follow suit. Inflation is exceptionally important, because it is almost always the final trigger for entering a recession. The inflation data for October that was released this week is still unequivocally in safe territory, and falling gasoline prices in November so far suggest that inflation will remain in safe territory, at least in the short term. 

Weekly retail sales have weakened some. However, even the un-averaged, year-over-year data remained in positive territory through the most affected week of the storm. Also, data for the week ended Nov. 9 looked better than the previous week but was still modestly below long-term averages. I am hoping the un-averaged data increases to above 2% year-over-year growth for this week (which will be reported on Tuesday).

I have generally been expecting some flattening in housing data this fall, but it really hasn't shown up just yet. I believe that the month-to-month data will flatten out some before accelerating again next spring. Since a lot of housing data is broken down by region, ferreting out the effects of Hurricane Sandy should be a little simpler than some of the other economic data.

Initial unemployment claims should also provide some clues on the employment side. That data is provided by states, with a lag of one week. This week there was a large jump in claims, as most of us had expected. However, we have to wait another week to get the state-level data to confirm that the big jumps were in storm-related states. I would really like to see those new claims below the 400,000 level within the next two weeks (although the Thanksgiving holiday could complicate that analysis). 

Separately, the monthly employment survey will occur this week, which is still likely to be subject to some storm-related issues. The claims data for this week, which will provide some early clues on the monthly report, will be released next week on Wednesday, a day earlier than usual, due to the Thanksgiving holiday.

Inflation Backs Down Just in Time for the Holidays
As expected, inflation as measured by the Consumer Price Index increased a very modest 0.1%. However, the report contained wide divergences by category, the largest I have seen in some time. Fuel oil, airline tickets, and electricity were all up big, while used cars, new cars, gasoline, and medical care all registered declines. The food at home category that I had been so worried about increased a relatively modest 0.3%. While meat and dairy items were up big, falling beverage and the "other" category helped offset a good part of the gains. 

The Producer Price Index for food was up 0.4% in October, suggesting that CPI food-related price increases for November could be a tenth or two higher than October's 0.3% gain. Hopefully, that will be just about the worst consequence that we will see from this summer's drought. Food away from home was nearly flat, showing just a 0.1% increase as competitive pressures in the restaurant arena continue to heat up.

In some bad news for consumers, airline tickets were up 2.4% following a 1.4% increase the prior month. Meanwhile, for those driving, auto insurance jumped 0.9% in October. Our insurance analyst, Drew Woodbury, noted that several insurers pushed through large price increases because of increased insurance fraud in no-fault insurance states.

Rent increases that we have seen and heard about anecdotally are finally turning up in the economic reports. Rents were up 0.3% in September and 0.4% in October. Combined, that annualizes to a rate of over 4%. While not good for consumers, it does help landlord incomes and helps to prop up the housing market. It should also help tip the balance for those renters contemplating a home purchase in favor of a purchase.

The year-over-year inflation rate increased modestly to 1.9% as expected. Timing of gasoline price declines explains some of the upward bounce. For the full year, I suspect prices will end up being about 2% higher than they were a year ago. Falling gasoline prices offset by food price increases should limit inflation in the last two months of 2012.

More Smoke Than Fire in the Retail Sales Reports
Everyone suspected the retail sales report would be weak this month as a result of soft auto sales (that were largely storm-related) and a pause in electronics sales as the iPhone sales bubble of September hurt October's results. However, overall sales declined even a little more than expected at negative 0.3%, as some Hurricane Sandy-related effects crept into the report. This was the first decline in four months. Although sales calendars of major retailers ended before the storm hit, other parts of the report, including restaurants and gas stations, did see some consequences from Hurricane Sandy. Gasoline sales were actually up in this report, moved by pre-storm hoarding. Restaurants, which were already seeing some minor weaknesses, were down 0.3% in October.

Even the year-over-year, three-month average data came in below where I like to see it at just 3.7% growth before inflation (that excludes autos and gasoline). Adjusted for inflation, the data was even worse.

Weekly Shopping Center Data Bounces After the Storm
I usually make it a point not to discuss individual weeks of the International Council of Shopping Centers data because of the volatility. This week I am making an exception. I wanted to show the individual weeks to make my point that the hurricane did not destroy the economy.

I was slightly fearful that sales growth, already not terribly strong, could fall into negative territory after the storm. Weekly sales did slow but never went negative. Retail sales for the week ended Nov. 3 were up 1.4% from a year ago (the storm hit on Monday of that week and this report goes through Saturday). For the week ended Nov. 10, the sales rate rebounded to 1.8%. I am hoping for a bounce to over 2% this week, which will be reported on Tuesday.

The rate has been running at 2.5%-3.5% for most of this recovery, so the storm is having an impact. Unfortunately, in the weeks ahead, it's going to be really tricky to determine if the ups and downs in the sales report are from fears concerning the fiscal cliff or lingering storm effects and rebounds.

Industrial Production a Surprise Casualty of the Hurricane
I was hoping for a nice surprise to the upside to the industrial production report due to some very strong auto production data. I had hoped that production would be 0.4% or more and instead was down 0.4%. The government pegged the effect of the storm as a whole percentage point. In other words, ex-storm, industrial production would have been up a more-than-adequate 0.6%.

I admit that I hadn't really thought about much about storm effects on production as the hardest areas, including New York City and the Jersey Shore, are not exactly major manufacturing hubs. However, I admit to missing the fact that with an unknown storm path, large portions of the East Coast, not just New York, were shut down for at least a day or two. And a day or two is 5%-10% of production. Unfortunately, the storm managed to pretty well scramble the data.

The auto sector, which I am guessing is located largely outside of the storm area, was one of the very few industries to manage a gain at all. Mining-related output, which is also mostly out of harm's way this time, was another category with a gain this month. Utilities, which I would have expected to be severly impacted by the storm, were down only 0.1%. I suspect this will be revised sharply downward the next time around.

Next Week Is All About Real Estate
Next week there are three major real estate releases scheduled, but not much else. The only other major release is initial unemployment claims, but the hurricane aftermath has rendered that series just about worthless, anyway. The markets are closed on Thursday, and no releases are scheduled for Friday. 

Turning back to real estate news, next week existing home sales, builder sentiment, and housing starts are all expected to be released. I am hoping that these data points will be virtually unaffected by the storm. However, if the numbers look disappointing, I'll make sure to check the regional analysis for any strong effects in the Northeast. I suppose that given that a lot of home closings happen near month-end, there is a chance that this data will see some storm-related impact. Ex-storm factors, I suspect existing home sales will remain relatively flat at 4.8 million units, primarily because of a lack of inventory in many markets.

Given worries about the fiscal cliff, I suspect builder sentiment could tick down 1 or 2 points from 41 last month. Housing starts are most likely to back off from September's rocket-like performance of 872,000 units, almost 80% over the bottom reached about a year and a half ago. Anything much over 840,000 units would keep me happy and maybe even fewer units if the softening was in the multifamily sector.

Consensus 3Q GDP Estimates Up to 2.8%, Due the Week Following Thanksgiving
The real economic fireworks come the following week with a sharp upward revision in third-quarter GDP growth and a decent personal income report potentially giving the markets something to cheer about after Turkey Day. A combination of a far-better-than-expected trade report last week, a better-than-anticipated construction report two weeks ago, and upwardly revised September retail sales reports have analysts scrambling to see who can raise their GDP predictions the fastest. The third-quarter consensus is now up to 2.8% compared with the first official reading of just 2% (when economists' expectations were a more meager 1.6%). I surmise that the economic truth lies between the dismal 1.3% level reported in the second quarter and the 2.8% forecast for the third quarter.

Some of the strength in the third quarter is also expected to detract from the fourth quarter, with fourth-quarter consensus GDP forecasts expected to fall back to 1.5% GDP growth. Hurricane Sandy effects are also hurting those expectations. Worries about the fiscal cliff probably aren't helping fourth-quarter expectations any, either. That said, I think fourth-quarter GDP growth is likely to be better than the consensus (because of auto production) but not as strong as the 2.5%-3.0% increase in the third quarter. Full-year 2012 GDP growth is likely to be very close to my original 2.0% forecast.