Part
01
of one
Part
01
Please get me as much information as possible about the Caixin China Manufacturing PMI. What are the survey questions, what kinds of companies are asked to participate, how many companies are asked, what companies are asked, how does it break down (can it be broken down by sector, industry, etc?).
Hello there! Thank you for your question on the Caixin China Manufacturing PMI. The short answer is that PMI is based on 12 questions, 30 company divisions, and anywhere between 500 and 3000 samples. The effect of PMI on the market can be monitored in the instances where PMI is wrong, having direct influences upon ETF's, commodities, and companies. Below you will find a deep dive of my findings.
METHODOLOGY
I began by familiarizing myself with the Caixin China Manufacturing industry. I looked over the previous request and additional notes in order to fully understand what success would look like to you. In respect to that note, I endeavored to use sources written by industry professionals or credible reports. I conducted a comprehensive search through academic databases, corporate websites, industry reports, government reports, and regulatory filings. In order to provide a full scope of information, I took a deeper look into trusted media sites, user forums and review sites.
FINDINGS
Below I will breakdown my findings into sub-headers for your ease of reading. Any additional notable insights will be detailed. When appropriate, I will specify the reliability of the source or writer.
SURVEY QUESTIONS
A total of 12 questions are covered. These can be noted to be in relation to: "production, new orders, export orders, existing orders, finished goods inventory, purchase, import, purchase prices, raw materials, inventory, employees, suppliers, delivery time, production and business activities expectation."
For each question, the diffusion index was calculated. This is defined as the percentage of positive answers in a number of companies plus half of the percentage in the same answers.
COMPANY TYPES
The survey contains 31 divisions of the manufacturing sector. These are found in the Industrial Classification for National Economic Activities. A total of 3000 samples are included. This information is based on the 2011 list, whereas the latest list only consists of 30 divisions. The panel covers small, medium and large sized companies.
For a complete list of the industries included see here, as an overview the categorization used is explained below.
• Primary Industries:
• Secondary Industries:
• Tertiary Industries:
COMPANIES INVOLVED
The selection method is described in detail: "The purchasing managers’ survey is a monthly survey, the data of the survey fluctuates very much for the influences of seasonal factors. The released PMI composite index and sub-indices are seasonally adjusted". This allowance for seasonal adjustment could justify the variation in samples per article. Some articles show a total of 430 sources or 500, whilst the source from which this quote originates has a total of 3000 sources.
SURVEY METHODS
The survey is broken down using divisions of the manufacturing sector as the selection strata. A sample size of each division is proportional to its proportion of the value, which is the addition of the division to the total value, followed by the addition of the manufacturing sector. Within the specified stratum, samples are then selected according to the probabilities proportional to their respective principal business revenues of the company.
FORECASTING METHOD
Who and how the information is forecasted is depicted here: "It is compiled by China Federation of Logistics & Purchasing (CFLP) and China Logistics Information Centre (CLIC), based on data collected by the National Bureau of Statistics (NBS). Li & Fung Research Centre is responsible for drafting and disseminating the English PMI report."
Each month questionnaires are sent to a total of 700 manufacturing companies across China. Data represented is compiled from company responses on purchasing activities and supply situations. It is advised to compare PMI to other economic data when using it as a basis for decisions.
It is noted that the forecast for China Caixin Manufacturing PMI was last predicted on June 20, 2017. The forecasts are projected using an autoregressive integrated moving average (ARIMA), this model is calibrated using analysts expectations. Past PMI behavior is modeled against historical data, in order to adjust the coefficients of the econometric model. Analyst assessments and future expectations are taken into further account.
Estimates are derived from analysis of current data, current events, and with respect to past data and PMI figures. There is a heavy emphasis on quantitatively analyzing past mistakes to better the estimation process. The people selected to determine estimates are considered to be industry professional, with no detectable bias on the matter. A notable insight on estimates can be seen here: "I can state categorically that neither the Chinese government or HSBC have any involvement or influence at all in the PMIs that we publish for China or for any other nations".
MARKET REACTIONS TO FORECAST
When the forecast is off the market reacts. In one case, crude prices recovered by 1% in Asia despite a lower than expected reading in private manufacturing PMI. It is further stipulated that China's manufacturing sector has been under a lot of pressure in May, a downward trend in trajectory is apparent.
An example of the effect on the market can be seen here: "Gold and copper drifted weaker in Asia on Thursday as a key measure of China manufacturing showed a dip into contraction in May. Gold futures for June delivery on the Comex division of the New York Mercantile Exchange fell 0.20% to $1,2769.47 a troy ounce. Also on the Comex, copper futures fell 0.50% to $2.537 a pound."
On the other, another example can be seen to have this effect: "The Aussie was quoted weaker in early Asia on Thursday ahead of a key China manufacturing PMI and domestic figures on capital spending. AUD/USD traded at 0.7430, down 0.48%, while USD/JPY changed hands at 110.77, down 0.05%. GBP/USD rose 0.26% to 1.2891. In Australia, first quarter private capital expenditure is due with a 0.8% gain seen quarter-on-quarter and retail sales for April are up with a 0.3% increase expected month-on-month."
I have found many other examples surrounding the different circumstances possible, here is a list of the links so that you can examine them at your leisure:
CASE STUDY
Here I will go into a more in-depth analysis of the full effect of market dips in China, on their respective companies and any quantitative information.
On page 11 of this document, there is a useful infographic that depicts the PMI level throughout and for comparison the consumer prices throughout. Through these graphs, we are able to determine the direct effect of PMI on the market. The time frame of the infographic is 2004 to 2016.
In December 2011 the PMI peaks 53.5, on a steady downward trend since then the market picked up in 2016 rising from 51.7 to 53.1. This depicts the strongest quarter in almost 6 years. Comparatively with the consumer prices, at a glance the market appears to follow the same trend. However, the troughs and peaks in consumer prices are much deeper and extended. This shows that the PMI figure is determining the market, the delay and extended reaction time support this statement. Notably, the manufacturing industry is more heavily impacted than other companies, this is due to PMI's effect on raw materials and commodities running the prices up.
Comparatively, on page 7 rates in the United States can be examined. In 2016 all 7 monitored sectors saw an influx. A dip in PMI surveys at the same time shows the effect of PMI on a global scale, despite market growth in the US. Market success is attributed to: "robust hiring, inventory building and improved confidence about the economic outlook". To bring it full circle, these attributes often come into play in the configuring of PMI estimates.
CONCLUSION
To summarize, I have taken a deep look into how the Caixin China Manufacturing PMI works, including the survey method and company demographic. I have found multiple examples of the effects of an inaccurate PMI on the market, accounting for all scenarios. I have further endeavored to take into account your latest requests on how estimates are calculated and the full effect of market dips.
Thank you for using Wonder! Please let us know if there's anything else we can assist you with.