Never Believe Any Chinese Economic Statistic For The First Quarter Of The Year

11
942

A useful and basic rule of thumb about international economic statistics. Never, but just never, believe nor pay attention to anything about the Chinese economy for the first quarter of the year. No, this isn’t because our inscrutable bretheren dissemble more or less at this time of year, it’s not because their statisticians spend January drunk or hungover (unlike our own), it’s because the Chinese New Year obeys its own little calendar.

The Islamic calendar is a lunar one, Ramadan thus falls in different months over time. We’re just into it right now in fact. The Christian, what we now call western, calendar is a solar one. Thus the same events occur at the same time each year. Well, OK, except that 364 days isn’t quite right, every 4 th year we throw another one in and then umm, well, isn’t it every century end divisible by four we don’t?

But, you know, it’s a solar calendar. The only major date which shifts around is Easter plus all the associated bits and bobs like Lent, Fat Tuesday and so on, none economically important events these days. No, we’ll not go into how to determine the date of Easter. Except to point out that determining the Chinese New Year works in much the same manner:

The first day of the New Year falls on the new moon between January 21 and February 20.

Well, OK, so if this was a western country that really celebrated the New Year (say, Scotland) then everyone would be back at work 48 hours later. However, the Chinese New Year is also the start of the two week holiday. Sorta a mixture between American Thanksgiving (you WILL eat at your mother’s table or a close simulacrum of it) plus a Wakes Week (English industrial towns would shut every single factory so that all could get away to the beach for a week. Well, beach not so much, Skegness maybe). The combination of the two means that near every factory in the country shuts for a couple of weeks as the largest migration in history takes place. All those migrant workers heading back to Mom’s dumplings.

If this all took place at the same time each year then our economic statistics would take account of it just fine with our seasonal adjustments. Just like we do with Christmas. We know very well that hundreds of thousands get hired for temporary jobs packing and delivering just before, get laid off immediately afterwards. We don’t see that reflected in the unemployment numbers because we’re not interested. We want to see trends, not known seasonal variations. So too with output and all that – many European factories do close in that week after Christmas. We don’t measure a drop in GDP then because we know about it therefore ignore it.

If the day itself and the two week vacation moves around the months then we can’t do that. At which point, this:

There’s a seasonal pattern in China’s economy that exports tend to be lower in the early months of the year. Assuming the usual rise in China’s exports later in the year, China may well end the year with a trade surplus, but it will be quite modest in size.

Yes but no but yes but no but…..

There is a seasonal variation there, of course there is. But it’s also not a predictable one. Therefore it’s not really a seasonal variation in the manner we think of with our own economic statistics, the ones we make adjustments for and thus ignore. Note what this means – we cannot extract trend from China’s first quarter stats. We know that New Year effect is vastly greater than any year on year variance. Just as Christmas is a hugely greater effect than any recession or boom in the US. And we’re really not sure how we should be trying to back out the New Year effect to find trend. Exactly unlike Christmas in the US.

The correct answer here is simply not to believe any first quarter Chinese economic statistic whatsoever.

11
Leave a Reply

avatar
9 Comment threads
2 Thread replies
4 Followers
 
Most reacted comment
Hottest comment thread
8 Comment authors
SpikeBaron_JackfieldMoosealotjghBaron_Jackfield Recent comment authors
  Subscribe  
newest oldest most voted
Notify of
Spike
Member

Irregular, but not unpredictable, nor impossible to correct for.

Spike
Guest
Spike

Irregular, but not unpredictable, nor impossible to correct for.

jgh
Guest
jgh

364?

jgh
Guest
jgh

364?

Moosealot
Guest
Moosealot

@Spike – normally the way its done is by comparing $_STATISTIC against the year-ago-quarter to give a year-on-year change which is reasonably useful as most of the things we worry about are annualised. Given that the new year and its associated holiday will always fall entirely in the Jan-Mar quarter, it’s actually a different effect that we observe, which is this… 5-10 years ago, a worker from rural China would go to the coastal industrial areas for 5 years to earn enough money to buy a farm ‘back home’. Since then, there has been a huge disparity in growth between… Read more »

Spike
Member

That is sophistry. When the Western calendar deals us a 53-week year, we don’t compare the year’s economic statistics to the previous 53-week year; we compare 52/53 of the figures to last year’s.

Moosealot
Guest
Moosealot

@Spike – normally the way its done is by comparing $_STATISTIC against the year-ago-quarter to give a year-on-year change which is reasonably useful as most of the things we worry about are annualised. Given that the new year and its associated holiday will always fall entirely in the Jan-Mar quarter, it’s actually a different effect that we observe, which is this… 5-10 years ago, a worker from rural China would go to the coastal industrial areas for 5 years to earn enough money to buy a farm ‘back home’. Since then, there has been a huge disparity in growth between… Read more »

Moosealot
Guest
Moosealot

The difference between a 53-week year and a 52-week year (or a 14- vs 13-week quarter) is — as you point out — easy to correct for because it’s generally safe to assume that the ‘extra’ week is ‘typical’, so a 52/53 or 13/14 correction factor is easy to use. The issue with China is not that there is some known difference in quantity of time but that there is an unknown difference in quality as productivity drops off a cliff at some point in the middle third of Q1 for two weeks and recovers at a pace that is… Read more »

Spike
Member

The cliff of underperforming because you just returned from two weeks off is surely not as steep as the two weeks off itself. The result is that Chinese statistics are wobbly. They are not wobbly enough to suggest trends that are nonexistent. They are meaningful, just not perfectly precise, as no statistics on any economy are; close enough that an asterisked footnote is all they need.

Baron_Jackfield
Guest
Baron_Jackfield

Er… Slight Point of Order Mr Chairman, there are 365 days in a normal year. (365.254 or thereabouts).

Baron_Jackfield
Guest
Baron_Jackfield

Er… Slight Point of Order Mr Chairman, there are 365 days in a normal year. (365.254 or thereabouts).