Economics 1/8/10: Retail Sales data: to spin or not to spin?

The latest retail sales figures for Ireland highlight two interesting issues. One - deeply fundamental, another - deeply disturbing.

The first issue - the fundamental one - relates to the basic philosophy of 'reporting' the data. CSO's publication on RSI was headlined "Retail Sales volume index increases by 1.0%". The first paragraph of the 'analysis' reads (emphases are mine):

"The volume of retail sales (i.e. excluding price effects) increased by 1.0% in June 2010 when compared with June 2009 and there was a monthly decrease of 0.2%. If Motor Trades are excluded the volume of retail sales decreased by 1.3% in June 2010 when compared with June 2009 and the monthly change was -0.5%."

This, to me, as an example of the poor application of economics to what is essentially a purely economic data series. And it is also an example of poor statistical analysis. Here is why:

1. The series reported are monthly and seasonally adjusted. This means these series are first and foremost about monthly, not annual deviations (annual comparisons can be made unadusted for seasonal / monthly variations). Why does the CSO then elects to report an annual deviation headline?
2. The volume series of retail sales are secondary in importance to the value series. What matters to gauging the overall demand in economy is not the physical quantity of stuff traded, but the value of the sales. Imagine a situation whereby an economy is plagued by a recession (like Ireland). Country largest retailer goes out of business and has a firesale of its stocks. Suppose it sells lock stock and barrel in one month, but at a price of zero euros per item, i.e. it gives stuff away for free. What happens? Volume of sales goes up dramatically. Value of sales goes down. CSO records an increase in volume and reports a headline that implies demand is up, sales are up. Yet, economic impact of this transaction is nill. If anything, it shows that economy has no real demand underlying it. Exchequer returns are nill. Value of stuff sold is nill. Value of transactions is nill. Patient is as dead as it can be!
3. Monthly, not annual series show shorter term dynamics. And it is the dynamics of sales, not their absolute levels or longer term changes that should frame short-term policies, that are suited for a recession.
   

Of course, you might object, saying - hey, you should have read the first paragraph, mate. Not just the headline. Alas, our politicos making bullish noises about turnarounds can't be relied upon to do this much. "It's the good news, folks! Retail sales are up year on year".

CSO has more disturbing analysis presented in the latest release. Paragraph two, in fact, is about as manipulative, as the preceding text:

"A number of sectors showed year on year increases in June 2010, with the most
significant being: Motor Trades up 13.9%, Non Specialised stores up 1.4%, Clothing, Footwear and Textiles up 2.6%".

Now, let's take a look at CSO own data to decipher the spin in the above statement:

• Motor Trades up 13.9% yoy in volume, and 1.4% mom - good news (driven, as I've said before by a tax off-set for new cars - aka the scrappage scheme, and to a larger extent - by the vanity plates for 2010), but Motor Trades are up less significant 9.2% yoy in value and 1.3% in mom terms. So one might ask the question then - why is value of overall Motor Trades lagging behind the volume of these. Is it due to (a) rebates by the Government (VRT offset?) or (b) competition in the Motor Trade sector or (c) because people are buying lower quality, cheaper priced cars? CSO doesn't even attempt to provide an answer. My earlier analysis (here and here) suggests that all three might be at play. If so, Motor Trades figures for the entire 2010 are not exactly a shining example of economic turnaround.
• Non Specialised stores volumes up 1.4% yoy, but down 0.9% in mom terms. Values of these sales are down 3.5% yoy and 1.3% mom. Discounts, discounts, discounts. Selling cheaper doesn't really generate more economic activity, though it does benefit consumers. And this 'cheaper selling' in turn drives up not new demand, but induces a movement right along the same, recessionary demand curve. But wait, seasonally adjusted monthly changes are negative in value, which means that deflation is still there and demand for quantity is not exactly booming.
• Clothing, Footwear and Textiles up 2.6% in yoy volume terms. Really? Well, mom the same series are down 4.1%. In terms of value of Clothing, Footwear and Textiles sold in Ireland in June: yoy sales collapsed 8.1% and mom change was 4.1%. In a normal economy that should start ringing the 'Recession Alert' bells. In Ireland, for CSO this is bunched together with the aforementioned 'good news'.

Here is another good look at the CSO own data, not brought up to anyone's attention by CSO:

• All Businesses excl Motor Trades & Bars: Value down -1.3% mom and -3.9% yoy, Volume down -1.1% mom and -0.2% yoy. Some turnaround!
• All Bus. Excl. Motor Trades, Fuel & Bars: Value down -1.9% mom and -5.3% yoy, while Volume is down -1.1% mom and 0.3% yoy. No turnaround here either.
• Non-Food (Excl Motor Trades, Fuel & Bars): Value off -1.2% mom and -7.1% yoy, while Volume is off -1.7% mom and -0.6% yoy.
• Household Equipment (white goods stuff) Value down -3.1% mom and -6.3% yoy, Volume off by -2.6% mom and -0.1% yoy. Now, this category is important as white goods are subject to demand due to depreciation and new demand. We've had at least 2 years of collapsing demand for these goods, implying that things are so bad, people are reluctant to replace depreciated washers, dryers, dishwashers, fridges etc. Forget buying new jeans and coats...

So to do what I usually do on this blog - here are updated charts plotting actual data (no spin):
If you look closely at the last three months in the series, you can see continued deterioration pressures in both. But to highlight this trend - check out the chart below:
Monthly changes are now in the negative territory, and a positive annual volume bounce of the first quarter 2010 is about to be exhausted.

Removing motor trade:

Why wouldn't CSO just report data, plus charts and leave 'commentary' to others? At least they would be purely objective reporters of data, instead of playing the amateurish 'Spin Economics' commentators?