Random Analytics

Charts, Infographics & Analytics. No Spinning the Data. No Juking the Stats

Month: October, 2012

Random Analytics: Ebola!

***** This article has been superseded by an updated version which can be found at Random Analytics: Ebola! (2013) *****

I have been fascinated by Ebola Haemorrhagic Fever since reading ‘The Hot Zone’ by Richard Preston in the mid-1990s.

In recent weeks it has come to light that another outbreak has occurred, this time in the Democratic Republic of the Congo. On the 26th October 2012 the World Health Organisation:

“Update Ebola in DRC: of 52 cases, 25 people (12 confirmed, 13 probable) have died as of 23 October http://goo.gl/q2tp4

For those of you who are not aware Ebola Haemorrhagic Fever (EHF) is a severe acute viral illness often characterized by the sudden onset of fever, intense weakness, muscle pain, headache and sore throat. This is followed by vomiting, diarrhoea, rash, impaired kidney and liver function, and in some cases, both internal and external bleeding. Laboratory findings show low counts of white blood cells and platelets as well as elevated liver enzymes. People are infectious as long as their blood and secretions contain the virus. Ebola virus was isolated from seminal fluid up to the 61st day after the onset of illness in a laboratory acquired case. The incubation period (interval from infection to onset of symptoms) varies between 2 to 21 days. During EHF outbreaks, the case-fatality rate has varied from outbreak to outbreak between 25% and 90%.

To emphasise the case to mortality rates which are extremely high I have broken down the figures for outbreaks by year. As you can see from Figure 1 there is a higher density of red over green, reflecting the overall 66.5% mortality rate from the three confirmed lethal strains of Ebola Zaire, Sudan and Bundibugyo) plus the one confirmed case of Ebola Ivory Coast.

Figure 1: Ebola Mortality Rates 1976 – 2012 (provisional data for 2012). Data sourced from the World Health Organisation.

The latest outbreak of Ebola in the DRC reflects the burden that country has in both cases and mortality. The next graph reflects the almost half share of mortality that the DRC experiences of all recorded cases since 1976 but with only 41.8% of actual cases also confirms a higher mortality rate compared to other impacted countries. No doubt this is due to it being the least developed country in the world.

Figure 2: Ebola Mortality Figures by Country 1976 – 2012. Data sourced from the World Health Organisation.

The last graph shows a breakdown of mortality figures by virus strain. There are currently five known variants of Ebola. They are Ebola Zaire, Sudan, Bundibugyo, Ivory Coast and Reston although only the first three variants have been confirmed as being fatal.

Figure 2: Ebola Mortality Figures by Virus Strain 1976 – 2012. Data sourced from the World Health Organisation.

Although Ebola has never caused a confirmed fatality outside of continental Africa the Reston strain has been imported into the United States on a number of occasions (via macaque monkeys in 1989, 1990 and 1996) and Italy once in 1992. Given its very high mortality rate once infected it is one of the scariest of diseases.

There is no treatment or vaccine available so let us only hope that it continues to lay dormant and the current outbreak is contained as soon as possible by the local authorities supported by international agencies such as the World Health Organisation.

Update 1 (3/12/2012): A further outbreak of Ebola has occurred in Uganda, although at the time of this update I was unable to confirm which strain. The WHO announced via an internet update on 30 November that:

“As of 28 November 2012, the Ministry of Health in Uganda reported 7 cases (6 confirmed, 1 probable) with Ebola haemorrhagic fever in Luweero and Kampala districts. Of these cases, 4 died.”

Additionally, I was able to find an unconfirmed report (as it comes from a good news source, the South African Times but not from a primary source, i.e. via a Ministry of Health, WHO or CDC update) that the DRC outbreak which initiated this blog has ended. The Times reports that:

“The latest outbreak of Ebola in the Democratic Republic of Congo ends after claiming 34 lives, says Health Minister Felix Kabange Numbi.

According to revised figures, 62 people are believed to have been infected during the latest epidemic, which was declared in mid-August in the north-eastern Orientale Province and which officially ended on Friday, he said.”

Random Analytics: Romney’s Promise

Just two weeks out from the US election it may be worth noting a promise made by Mitt Romney during the first debate held in Denver on the 3rd October 2012.

He stated that “If I’m president, I will create, help create 12 million new jobs in this country with rising incomes”.

This is a big promise. Let’s put aside the promise about rising incomes (as the Economist recently noted US median incomes have reversed over the last 15-years after peaking in 1999) and unpack some of the data in relation to US employment.

If he were to win the US Presidency Mr Romney’s administration would need to create 250,000 new jobs per month which would equal 750,000 per quarter or 3-million per year.

Figure 1: US Employment gains & losses by quarter (please note that Q1 is Jan – Mar). Data sourced from the US Bureau of Labour Statistics.

Over the past 10-years and nine-months (129-months of data) there have been just 11-months where employment gains exceeded 250,000. To be fair April 2004 had a gain of 249,000 so let’s allow that one as a conceded pass thus increasing the total to 12. There have only been two periods where the employment increased at more than 250K in consecutive months, that is the first quarter of 2006 (during the height of the pre-GFC period) and the first two months of 2012. On the negative ledger the months where employment losses exceeded 250,000 also tally 12. Between August 2008 and July 2009 the US economy officially shed 6.874-million jobs over 12-months. Over the past 10-years no other month saw a loss of more than 250,000 except for that period.

In terms of quarterly data there have only been two quarters where employment increase by more than 750,000 (Q2-2005 and Q1-2006). On the negative side the period Q3-2008 through to Q3-2009 were five consecutive quarters where job losses exceeded 750,000. In the first quarter of 2009 the US economy lost more than 2.3-million jobs, more than three times the Romney target but moving in the wrong direction.

Figure 2: US Working Population 2002 – 2012 (updated with October 2012 data). Data sourced from the US Bureau of Labour Statistics.

To add to both Mr Romney and Mr Obama’s concerns are not just creating new jobs but also keeping pace with new entrants to the workforce which I would estimate could be between 300,000 – 900,000 per annum depending on how you might calculate the increase. For all the talk of the impending crunch time due to baby boomer exits there has been a sustained increase in the US working aged population over the past decade as shown in Figure 2. The red line reflects BLS actuals, an increase of 10-million workers in the past decade. The black polynomial trend-line I’ve added to highlight the incline-decline-incline impacts of the GFC. However way you cut it, even a minimum increase in new entrants adds another 10% to the annual job creation burden.

In economics we have a saying. Unemployment goes down the elevator but re-employment has to climb the stairs. Just looking at these two graphs highlights some of my concerns with Mitt Romney’s aspiration of creating more than 250,000 jobs over a sustained 48-month period.

Looking at the data, even in the best of times the task would have been incredibly hard. I still think the global employment situation has a lot of downside risk. Even with recent improved employment data I don’t believe the US is out of the woods yet, in fact I suspect we might see further soft, even negative employment data next year in the US.

For Mitt, if he were to get over the line I would suggest he might have promised in haste and will regret in leisure.

Update (6/11/2012): Given the recent discussion around the US unemployment figures for October, which showed a provisional increase of 171,000 non-farm jobs but an overall rise in the unemployment figure from 7.8% to 7.9% I thought an update of the US Working Population graphic would be useful. Between September and October 2012 the US Working Population (those in the civilian labour force) actually increased from 155,075,000 to 155,779,000 or an increase of 704,000. Interestingly, the available Working Population figure was at its highest in Jul 2008 at 156,300,000 so although this months figures may seem high to some there is still a reasonable amount of elasticity in this number (thus it doesn’t represent a truly high figure at this stage).

Random Analytics: A Story of two Economies: China vs. DRC

If you want to look at two economies who sit as far away from each other in terms of economic prosperity in the year 2012, then you need go no further than a comparison between the China and the Democratic Republic of Congo (DRC).

But it wasn’t always so…

Figure 1: GDP per Capita (PPP $) of China and the Democratic Republic of the Congo. Data sourced from the World Bank.

Just 30-odd years before,  Zaire (as the Democratic Republic of the Congo was then known) with its 35.6 million population actually boasted a GDP of $14.39Bn USD and a GDP per capita of $533. China on the other hand with its 1.139Bn citizens had a GDP of $189.4Bn which equalled a mere $193 per citizen.

Backed by the fastest industrial revolution in history by 2011 Chinese GDP would increase to $7.3Tn (the second largest for any country after the US) and per Capita income would explode from $193 to $5,430. For its 1.344Tn citizens, roughly a fifth of the world’s population that would be a 27-fold increase in GDP per Capita and an 84% decrease in those living on less than $1.25 per day. Living on less than $1.25 per day is a standard measurement of absolute poverty adopted by the United Nations as part of its Millennium Development Goals.

As for Zaire…

From being the second most industrialised African nation in 1960 and having rich agricultural and natural resource advantages Zaire would suffer throughout the 1980’s an exponential growth in corruption, known locally as le mal Zarois, or the Zairian sickness. According to Young, C. & Turner, T. (1985) by 1984 Mobuto Sese Seko had amassed $4Bn USD in personal wealth stolen from the state (effectively odious debt). Those billions of dollars stolen by the regime were not reinvested in the country’s infrastructure which had degraded significantly prior to the First and Second Congo Wars. These wars, aka Africa’s First World War, have cost the country and region more than 5-million lives since the commencement of bloodshed in 1996 along with enormous capital and infrastructure loss.

From an economic standpoint this has left DRC and it’s almost doubled population of 67.7 million a GDP in 2011 of just $15.64Bn. The additional $1.25Bn USD growth over the previous 32-years represents an 8.65% increase or just 0.27% per annum. From a GDP per Capita perspective the $231 for each citizen is only 43% of the 1980 figure.

For the DRC it is now a story of what could have been…

Random Analytics: Australian Unemployment: A Global Perspective (Oct 2007 – Sep 2012)

I had an interesting discussion today about Australia’s recent increase in its official unemployment rate which rose to 5.4% (seasonally adjusted). That percentile equates to 662,700 unemployed persons. It has been my consistent view that the current unemployment and participation rate data is softened by factors such as seasonal adjustment and under-reporting. My point here is that the real unreported figure could be higher due to severance economy issues and as Australia transitions from an overall boom economy to, at best, a more globally normalised economy.

Rather than focus on that rather negative sentiment I thought it might be useful to have a look at some global data to get a sense of perspective. Here is a look at the official (seasonally adjusted) unemployment figures for Australia, the USA, UK, Spain & Greece.

Figure 1: Unemployment (seasonally adjusted) for Australia, USA, UK, Spain & Greece. Data sourced from the ABS, BLS, ONS & EuroStat.

That gold line which has trended under or around 5% for the entirety of the Global Financial Crisis is Australia.

For the other countries represented in the graph the last five years have been dark days indeed. The US (dark blue), still the world’s largest economy has an official unemployment figure of 7.8% (or 12.1-million persons). The UK (in green) sits at 7.9% (2.57-million) trending at 8% or thereabouts since mid-2009. Then there are the current horror stories of Greece & Spain, both of which now have official Euro high unemployment figures of 25.1%. For Spain there are 5.77-million and Greece 1.26-million unemployed, a total of 7.03 million unemployed for both countries. Put in context this equals more than 60% of the total current Australian Labour force.

So things might not be great in Australia at the moment as we see layoffs across some sectors of the economy and our official unemployment figures nudge up.

I’m happy to bet that you would still rather be in Australia with this economy than in Greece or Spain with their economies?

Random Analytics: Mining Workforce Planning Environmental Scan (Jan – Sep 2012)

Around the end of last year I started to complete some research into the mining industry, specifically Workforce Planning as it impacts mining and energy. When I look at a new industry I generally poke around for some open source data and environmental scanning to get an understanding of where the sector is going. In the case of mining I was left bereft of any decent and non-biased scanning. Too much environment scanning and commentary was either completed by the sector itself (everything is great/rosy) or by those oppossed to mining (all things mining are evil and so on). So rather than complain about it I commenced putting some qualitative data together, namely Australian Mining stories, discarded any data that was not Workforce Planning and categorised those stories into 16 groups which has created (in my view) some interesting quantitative data.

Here is a first look at the Australian Mining Workforce Planning Environmental Scan (2012) updated to end of September.

Figure 1: Australian Mining Workforce Planning Environmental Scan 2012 (Jan-Sep). Data sourced from Australian Mining Newsletter & News Archive. Some stories have been verified against primary resources (i.e. ASX, commercial websites and other news agencies).

The standout feature from September has been the very high ratio of employment stories for the month which comprised 44.8% of all Workforce Planning related stories but a staggering 23.4% off ALL stories. So far only January 2009 had a higher figure when Employment stories hit a record high at 52.9% (remember, for Australia the impacts of the GFC started from November 2008).

What is also interesting is that Employment stories have dominated the scanning for three concurrent months post EOFY. June was low but this reflects a generally quiet month for employment activity due to companies focus on finalising EOFY and budgets. The Employment totals as a percentile since July have been:

  • July: 25.9% (61.9% negative, 38.1% positive);
  • Aug: 35.4% (71.4% negative, 28.6% positive);
  • Sep: 44.8% (76.9% negative, 17.9% positive and 5.2% neutral)

Already October is on the wane and although the Employment figures are not as high as September it is still the dominate category (35.7% as at 17/10). Given that this will then become a 4-month trend it will no doubt be the main point of discussion when I update the data & commentary for October in a couple of weeks.

Table 1: Australian Mining Workforce Planning Environmental Scan Sep 2012 data. Data sourced from Australian Mining Newsletter & News Archive. Some stories have been verified against primary resources (i.e. ASX, commercial websites and other news agencies).

Random Analytics: The Myth of Mining Employment

As a Workforce Planner I always get frustrated by almost every one’s insistence (especially government) that somehow the mining & energy sector will answer Australia’s unemployment concerns. Yet even after record FDI in recent year’s pre & post GFC these sectors continue to employ a very small percentile of Australians.

Here are some analytics that highlight the minor role that mining & energy plays in employing Australians. I’ve also added some analytics which show the recent exponential growth in employment which goes some way to explain why some get so excited about the mining and energy stories.

Figure 1: Total employed in the Mining & Energy sectors 1984-2012. Data sourced from the ABS, Aug 2012.

Mining has risen to a record high 275,200 (but just 2.4% of total employment) while Energy has risen to 153,300 (1.3% total employment). The increase in mining employment over the past 10-years has been significant, from 81,200 to 275,200 (an increase of 338.9%). Energy has increased but not to the same scale, from 83,900 to 153,300 (an increase of 182.7%). It should be noted that the Energy figures also include water & waste services employment. Although the mining data is cleaner it would also contain some infrastructure contamination, so the true mining employment figure would be lower but this won’t be quantifiable until after the current FDI deployments start to wind down (currently forecast to commence declining from next year but continuing until 2016/2017).

Figure 2: Changes in employment totals for the Mining & Energy sector 1985 – 2012. Data sourced from the ABS, Aug 2012.

I spent a lot of time travelling throughout north Queensland in the early 1990’s, when gold was under $300 per oz. and a lot of coal and gold mines closed. This is reflected in the negative numbers through most of the 1987- 1999 period. From 2000 through to 2012 the data has all been largely positive. Just this year the mining sector has put on an impressive 42,800 averaged over the first three quarters. Take a note of how Energy and Mining have increased employment every year since 2010. Mining has increased by 24,200 (2010), 33,500 (2011) and 42,800 (2012) as mentioned previously. Energy (and water & waste services) has increased but in a declining fashion with 13,300, 6,900 & 3,300 over the same period. Although not included in this data set I have been doing a Mining environment scan which has shown an increase in negative employment stories for both sectors. This should mean the first significant decrease in employment in these sectors come November.

Figure 3: Changes in employment percentiles for the Mining & Energy sector 1985 – 2012. Data sourced from the ABS, Aug 2012.

The last graph looks at the same data as represented by Figure 2 but looks at increases and decreases in sector employment as a percentile. This is one of the key reasons why so many politicians and policy makers have been sold the mining & energy sector employment myth. Even though the percentiles are impressive, with Mining averaging 11.9% increase over the past decade (13% if you remove 2009) it all comes off a very low base and is highly susceptible to commodity price movements as recent events have highlighted.

The main point here is that although mining and energy have been the boom sectors of recent years they still only employ 428,500 people or just 3.7% of the working population, although this figure is on the high side and would reduce once you factor data contamination and non-energy sector job categories. Given the recent commodity normalisation that has been occurring and other factors including (but not limited to) redeployment of FDI to other global suppliers, deferral of projects due to high CAPEX, overcapacity and a move to automation & augmentation in high-cost countries the mining & energy sectors are not going to supply the large scale employment increases many commentators, politicians and policy makers have suggested.