Wednesday, May 9, 2012

French business difficulties

Via INSEE, we have an interesting evolution of the main difficulties of the french manufacturing industry. In the quarter ending in July 2011, the share of businesses reporting difficulties of supply only or difficulties of demand only was the same, at 30%. In the last quarter, the shares are respectively 15% and 45%. The ugly graph below has the share of businesses reporting difficulties of supply and demand, supply only, and demand only, over the last four quarters:

Add caption

In the long run, it looks like the other graph below, with the quarter on the horizontal axis starting from January 08 to the last quarter, and the vertical axis being the share of businesses reporting specific difficulties. Yellow bars show the share of businesses reporting difficulties of demand only, orange bars report difficulties of supply only (purple bars in 09 is related to difficulties in asset liquidity).

Yellow bars are shares of businesses reporting difficulties of demand only, orange report difficulties of supply only. The purple in  09 report difficulties of liquid assets

Related, KPMG has a quick survey out here from March for small and medium firms showing another evolution between June 2009 and March 2012, with small firms increasingly reporting difficulties because of high suppliers cost (from 29% 3 years ago to 65% today) and a falling share of firms reporting difficulties due to a fall in turnover (from 63% to 38%).

Now there must be a way to square those two reports...

Sunday, May 6, 2012

Predictions and Results of the 2nd round

Interestingly, the final score of Francois Hollande will be between the last two predictions I posted. As of now, CSA has 51.8, Harris 51.7, TNS Sofres 52, and the rest in between.

The two predictions only differed by the inclusion of a pollster quality index.
Without quality index

With quality index

Obviously, this is more a stroke of luck than anything. Basically, those models managed to capture the trend (down) of the last few days and the last poll numbers. It's good to know that polls are informative, though (not only in levels, but also in trends).

Here are an incomplete and quick list of things to do for the next elections to improve the use of polling data:

  • Change the weights: I added the function that (I think) Nate Silver is using for the poll's closeness to election day (exponential decay with a half-life of 30 days), but I didn't know what to do with the poll size and with the very rough indicator of pollster quality
  • On the latter, we need to have a better measure of voter quality. The main problem is that I don't think we have polls region by region, so the data to compute this quality measure will be quite weak.
  • Adding the approval rate of the incumbent would probably be a good thing. 
  • I used the simplest thing I could find in R but there must be better ways to do the estimation given the nature of the data (e.g. the span of the non-linear regression I guess?)
  • Other stuff, but you know, I am watching TV

Friday, May 4, 2012

Last election graphs?

Update: Using the great Ceteris-Paribus (who has probably done the thing more precisely than I did), we can include a rough indicator of pollster quality using the results of the first round (we could also use this post from Ceteris-Paribus to include an indicator from 2007, but not all the 2012 pollsters were there). I just added in the weight the fraction pollster_error/min(pollster_error). Results: 

We won't have any more data from polls, so I updated the graph and added a couple of bells and whistles. I use the loess function, and weigh the data by sample size and how recent the poll is (for the latter, following Nate Silver, I use a weight of (1/2)^[(today-poll day)/30] . For the sample size, I put a concave function of pollsize/max(pollsize)) .

The downward trend for Hollande is pretty clear, and a rough forecast base on recent polls has him at 51.5% on Sunday. 

We can do similar exercises with the vote transfers and intrade data, again, with a huge handful of salt. Sarkozy opened the gap in the Le Pen and Bayrou vote transfers.

Hollande closing price in red, Sarkozy's in blue.
Blue is transfer to Sarkozy, red to Hollande, grey is abstention, null and blank