In order to make progress on doing macroeconomics in India, one weak link has been business cycle measurement. This, in turn, requires access to a wide range of seasonally adjusted time-series. In most countries, the infrastructure of seasonally adjusted data is produced by the statistical system, but in India this has not come about.
Seasonally adjusted series are particularly important in tracking current developments in the economy. The familiar year-on-year change is the moving average of the latest twelve monthly changes. In order to know what is happening in the economy, it is better to look at recent months, rather than looking back 12 months. The familiar y-o-y changes are a sluggish indicator of what is happening. Month-on-month changes are more informative: but this requires seasonal adjustment.
We have initiated some computation and release at cycle.in
At present, we have a dataset with seasonally adjusted levels for a few time-series. We will be updating this every Monday. At the above URL, you get a sense of what is happening with month-on-month changes of seasonally adjusted data in these series.
In the spirit of creating public goods, we make it easy for you to embed these graphs into your work products. We also have a .csv file with data for levels which can be the foundation of further work.
This will be useful in tracking current developments in the economy, and also make possible research in macroeconomics, which critically requires seasonally adjusted data.
We hope this is useful. Please use the comments on this blog post to give us feedback.
Thanks Ajay, It will definitely be helpful to all of us to get a clear look of what changes are happening recently.
ReplyDeleteThanks,
Madan Kumar Rajan
Thanks for starting this. I had tried to do this for WPI and posted it on my blog. You could check it out at http://abhikush.blogspot.com/2009/06/seasonality-in-indian-inflation-wpi.html
ReplyDeleteDear Ajay,
ReplyDeletethanks a lot for creating this website. I have one request. I wonder if it would be possible for you to develop a Composite of Leading Indicators on the lines of that released by OECD. The OECD data is released quite infrequently. Often the leading indicator data for any given month is released 2 months after the month, making the exercise pointless.
Considering the no. of variables you are already tracking, it should simply be a process of evolving the model and assigning weights to each parameter. After this initial startup-the LEI should calculate itself with some periodic re-adjustments to the individual parameters and addition/deletion of variables
takethetimes@gmail.com
ReplyDelete