Weekly Market Update - September 8, 2016

Dodge Starts Non-residential Construction Starts Relationships: Figure 1 shows Dodge non-residential construction (NRC) starts in inflation adjusted dollars per square-feet as ($ per Sq-Ft.). This data is charted for both 3 and 12 month moving averages. Not surprisingly the 12 MMA smooths out the volatility of the 3 MMA. The magnitude of volatility has been more pronounced since the recession of 2009. Despite the inflation adjustment to constant dollars (which from the fit of the Sq-ft. line looks reasonable), the installed price (12 MMA), of NRC increased from approximately $100 per Sq-Ft. in the 1994 through 2002 period, jumping to the $150 per Sq-Ft. range in the 2003 to 2007 period. At that point the price per Sq-ft. soared, topping out at $268 per Sq-Ft. in October 2010. The price per Sq-ft. rapidly declined in the period between 2010 and 2013. In the two years since, the price per Sq-ft. rose through mid-2015 and has been steadily falling since reaching $159 per Sq-ft. in June. The price per Sq-ft. for NRC is not only impacted by commodity prices, labor rates and other inflation or deflationary factors, it is also influenced by apparent demand which referencing recent starts data (Figure 5), looks to be in decline.

Figure 2 compares annual Sq-ft. totals with apparent steel supply. The correlation coefficient (C.C.), between what AISI documents as annual structural steel supply and the total yearly Sq-ft. of NRC is a near perfect 0.990. We were somewhat surprised that the relationship is so strong. This demonstrates that the structural steel supply pipeline is efficient at adjusting to changes in demand, and is a marked improvement from the 1990 to 1997 period where the C.C. was 0.832.

Figure 3 compares Dodge NRC starts with net job creation. Historically job growth has lead NRC starts. This makes sense, create jobs people need offices to work in have money to spend at retail outlets. In addition larger government tax receipts increase pave the way for increased spending on institutional buildings and infrastructure. There is a great deal of “noise” in the data, but generally the relationship held true. That is until recently. Beginning in 2011 the two lines began to diverge (Figure 4), and the relationship fell apart. The C.C. was 0.727 in the period between 1990 and 2012, falling to 0.021 since then. So what was the cause? One thing that comes to mind is the consumer shift to buying goods and services online. There has been a huge upswing the construction of warehouse space at the expense of commercial construction over this time-frame. Another fact is that public construction spending suffered a long lag after the recession ended and remains relatively flat compared to privately funded ventures. A “wait and see” attitude Vis a Vis the upcoming presidential election have also been referenced numerous times in the media. Time will tell whether or not the long term tend regresses to the mean.

Figure 5 examines the “peak to peak” cycles of Dodge NRC starts from 1990 to present. During this time interval, there have been three peak to peak cycles: from 1990 to 2000 = 10 years, from 2000 to 2007 = 7 years and from 2007 to 2015 = 8 years. The average for the past three cycles has been 8.33 years, which is approximately where we are at currently. The chart suggests that we may be at the inflection point of the next down-cycle. However, it remains a possibility that the decline in starts activity (down 10 consecutive months y/y), may just be a temporary stall and NRC will pick up again after the presidential election and into 2017. This theory is supported by the fact that the July American Institute of Architects consensus (of seven firms including Dodge), forecast for NRC called for growth of 5.6% in 2017, with a range of between 3.5 to 9.3% y/y over 2016.

Figure 1

Figure 2

Figure 3

Currency Update September 7th 2016: The latest data for the Broad Index value of the US $ from the Federal Reserve is September 2nd. Figure 6 shows the history of the daily index since January 2011. In the 7 days prior to September 2nd the dollar strengthened by 0.7%. This was the day that the August employment and ISM manufacturing data was released. Wall St. Breakfast had this to say on the 5th; The weakness in the manufacturing ISM index, the softer than expected employment data, and now the weaker services ISM makes some economists fret about Q3 GDP and lower the odds of a rate hike by the FOMC later this month.

For the purpose of analyzing the strength of the dollar against the currencies of 16 steel trading nations we use the historical exchange rates published in the Oanda Forex trading platform. This data is daily, the latest of which was Wednesday September 7th. Table 1 shows the number of currency units that it takes to buy one US dollar for each of the 16 nations and the change in value of each currency in one year, three months, one month and seven days. The difference between the Fed data in 7 days prior to the 2nd and the Oanda data in 7 days prior to the 7th is striking. The Oanda data shows that in 7 days prior to the 7th the dollar weakened against every one of the 16 except for Ukraine. This means that trader sentiment has discounted the likelihood of a rate increase by the Fed later this month.

Figure 7, 8, 9, 10 and 11 show the history of the currency value in US dollars for Canada, Mexico, the Euro, Turkey and Brazil.

On September 2nd Mark Chandler Global Head of Currency Strategy at Brown Brothers Harriman wrote, “I still believe before this cycle ends, the euro will sell off further, and I still target the old lows of 82-85 cents back in 2000. The euro had a 35-cent move, from $1.40 to $1.05, and since then has been consolidating between $1.08 and $1.16 and in a band that is still wide enough for traders to lose a lot of money. With this type of broad sideways trading, investors should be asking whether this is a base (for a move higher) or a top (for a move lower). And I think this will prove more likely to be a top for the euro. The ECB is still easing while the Fed is tightening. Before the ECB is done easing, the Fed will tighten a few more times. Combining the health of the European banking system with the political environment and uncertainty from elections in Italy, France and Germany all point to a weaker euro.”

Figure 6

Table 1

China Steel Production and Exports: In the seven months through July, China removed 21.26 million tons (mt) of the 45 mt of excess capacity it pledged for removal this year. This issue was addressed by the director of National Development and Reform Commission in an industry conference last month. Director Xu Shaoshi said that 8 of the 46 provinces targeted for capacity elimination had met one-fourth of their annual closure targets for 2016, while another 10 had not commenced their cuts. The industry minister reiterated how “imperative” it is that these 18 provinces step up efforts to meet targets for elimination, if possible. However, these provinces were behind schedule due to industry priority focused on the negative impact to local economies, over the national edict to eliminate capacity. There are now less than four months left in the year and China has much further to go before it’s pledged elimination targets will be met. The director of national development and reform commission stressed that these targets must be met or fines will be issued to the offending companies.

Based on data reported by China Metals Weekly, China produced 67.3 mt of steel in August, up 0.7% m/m. Steel output data is now reported in 10 day increments, shrouding exact steel production tonnage, (Figure 12). China steel global long product exports have risen dramatically since 2009, (Figure 13). Long products include steel classified as wire rod, rebar, angles, channels, and steel beams. In July, China exported 4.7 mt, down 15% m/m, yet brought year to date long product exports up to 31.2 mt, a 19.6% y/y increase.

Chinese steel prices have been a roller-coaster over the last 5-6 months, rising and falling dramatically, pressured by global steel over capacity. During February, March and April mandated idling of steel production for Hebei Horticultural Festivals, in China’s largest steel producing region, eased pressure on domestic steel prices, (Figure 14). Figure 15 charts China’s utilization of the total steel making capabilities. Steel production fell to 66% of capacity, as seasonally slow mills in the rest of the country also idled production. Price growth was healthy through spring, peaking mid-April. As prices rose, previously idled production, and seasonally idle production were encouraged to resume in March when capacity utilization reached 77.1%. As the glut of global steel inventories continued, China’s domestic steel prices fell, losing almost all price gains made during the first half of 2016. Since bottoming in June, prices have grown, peaking again in August. In spite of the highs and lows, Chinese domestic steel prices for week ending September 7th were up month over month and year over year, (Figure 14).

While China made some progress shuddering some of its massive excess capacity, this has not reduced its raw steel output. On the contrary, raw steel volume and export levels have increased significantly. China is simply operating its remaining capacity at higher utilization rates resulting in lower operating costs. This does nothing to address the underlying problem of producing far more steel than its needs and exporting the huge excess.

Figure 12

Figure 13

U.S. Jobless Claims: Data released today from the US Department of Labor showed that US jobless claims fell to 259,000 for the week ending September 3rd, falling 4,000 from last week. The four week moving average stands at 261,250, its lowest point in five weeks. Figure 16 shows that the total initial jobless claim figure has dipped into the green “sustained economic growth” region which is positive news on a go-forward basis for the US economy. The economy is nearing full employment per Moody’s which portends that ongoing job growth will moderate. Businesses are having an increasingly hard time finding suitable workers. The solution will necessitate higher wages, increased training and resorting to hiring workers that would not have made the cut previously. Looking forward, job gains are expected to average around 180,000 per month with approximately 100,000 per month required to absorb new entrants.

Figure 16

Steel Demand Indicators: Table 2 is a snapshot of the market situation on 9/8/2016. Indicators updated since we last published two weeks ago are shaded beige. In most cases this is not August / September data but data that was released in those months for previous months, the actual month to which the data relates is shown in the second column. Of the 27 indicators under consideration, the present situation of 4 are positive by historical standards, 8 are negative and 15 are neutral. This was a decrease of one positive and an increase of one neutral since our last update on August 26th. The change that occurred was to our view of the ISM manufacturing index which fell below 52.0 on a 3MMA basis in the September 1st release. The 3MMA of the latest value for this leading index of manufacturing was 51.7. There were no changes in our perception of the other 26 indicators.

In our trends analysis, most of the values reported are three month moving averages to avoid the knee jerk reactions that are characteristic of most economic reports in the press. Please note that there is nothing subjective about this trends analysis. The numbers presented here are the facts available as of September 8th 2016. The number of indicators trending positive in this latest analysis was 13 with 12 trending negative and two unchanged. This was a net decrease of one positive and an increase of one unchanged. The changes that occurred were as follows. The price of rebar CIF Houston which had declined in our August 24th report was unchanged at $340.50 / net ton on September 2nd based on the daily Platts Steel Price Report. Service center inventories of long products trended negative in the MSCI report for July. Inventories increased as shipments decreased. Finally the ISM manufacturing Index trended negative as mentioned above. This was the first time for the ISM index to decline on a 3MMA basis since February. There were no other changes in the direction of trends.

We have separated the leading indicators from the main table for a closer look, these are shown in Table 3. Of the twelve leading indicators eight are trending positive, and four negative. This was an increase of one negative since our August 26th update. The change was in the ISM manufacturing index as mentioned above. In summary the present situation is historically weak but trends are good which leads us to have confidence in the long products business environment into the 4th quarter. (Explanation of Indicators).

Table 2

Contributors this week include; Laura Remington, Peter Wright and Steve Murphy