3
$\begingroup$

I want to test index models that are applicable to macroeconomic data to test my hypothesis in R or some other statistical software (I have most of them).

The properties of most of the macroeconomic variables have trend--going up or down over time and cyclic--A cyclic pattern exists when data exhibit rises and falls that are not of fixed period. The duration of these fluctuations is usually of at least 2 years. Also the quarterly dataset has level data and growth data e.g. for EU GDP 2013--$17,578.4 and unemployment rate--10.8%.

Which econometric indices are best for macroeconomic variables? Thank you for you time. Regards.

$\endgroup$
10
  • 1
    $\begingroup$ I'm voting to close this question as off-topic because it is asking for code. $\endgroup$ Feb 10, 2015 at 0:47
  • 1
    $\begingroup$ @gung I read it as asking for "examples", which may include code, but may instead be examples in papers. I don't see how merely mentioning code as one possible way of giving an example makes this off topic. $\endgroup$
    – Glen_b
    Feb 10, 2015 at 1:28
  • 1
    $\begingroup$ Cross-posted on SO and closed there. stackoverflow.com/questions/28420901/… Please don't cross-post! Unfortunately in my view this question fits neither forum well. SO expects code-specific questions, not open-ended requests for code. CV expects specific statistical questions. Requests for code are off-topic here; there is also a flavour of asking for references. There may be scope for increasing the statistical element of this question. Is "e.g." intended by "i.e."? $\endgroup$
    – Nick Cox
    Feb 10, 2015 at 1:32
  • $\begingroup$ How is it an open ended question? I am asking for examples or documentation of conducting macroeconomic forecasts using diffusion indexes. I have gave a brief explanation of my hypothesis. If its an issue of citation, no worries I will cite accordingly. Thanks $\endgroup$ Feb 10, 2015 at 15:00
  • 2
    $\begingroup$ Open-ended meant, and means, that on Stack Overflow questions of the form "please show me code" do not go down well. I have to guess that you would be better off in a macroeconomics/econometrics forum, but I don't know which or even whether they exist. $\endgroup$
    – Nick Cox
    Feb 10, 2015 at 16:07

2 Answers 2

2
$\begingroup$

A couple of indices that may be useful for macroeconometric modelling are the Purchasing Managers Indices (PMI) and various indices based on Consumer Sentiment. There are many different indices (some being proprietary), so basically, any high-frequency survey data may be useful if your interest is forecasting or nowcasting macroeconomic variables.

The PMI data gets regular attention in the media. For example, see this Bloomberg snippet.

One could imagine that, say, the new exports orders index could be useful in the context of trying to build a leading-indicator econometric model or nowcasting model for a country such as Taiwan; a small open economy dependent on exports.

Consumer Sentiment indices also get regular attention in the media. For example, see this news report from Ireland.

This kind of data, usually available on a monthly frequency, are sometimes referred to as "short-term indicators" and play a role in "nowcasting", which involves trying to obtain early estimates of key macroeconomic variables that are available only on a quarterly basis.

There's a lot of literature on the usefulness of this kind of data and it has become more common to use this kind of data in a mixed-frequency setting.

For examples of publications in this area, see below and the references therein.

  1. The value of hard and soft data for short-term forecasting of GDP - by Mary Keeney, Bernard Kennedy and Joelle Liebermann (Central Bank of Ireland).
  2. Survey-based nowcasting of US growth. A real-time forecast comparison over more than 40 years. - by Antonello D’Agostino and Bernd Schnatz (European Central Bank).

If you want to do some empirical investigation, but do not have access to the proprietary data then you could take a look at indices that are available publicly, such as retail sales, industrial production, residential property prices, and so on.

I hope this answers your question.

Edit: I read in the comments that you'd like to create your own index. If this is the case, a book called Machine Learning for Hackers contains an example in R on how to do this. Refer to the literature on principal components analysis and latent factor models.

$\endgroup$
0
1
$\begingroup$

Start with reading the literature on the subject. In econometrics this will save you a lot of time, because anything you'd come up with was already tried. If you don't know where to start then look at here, it's monthly recession index series according to: Chauvet, M., "An Economic Characterization of Business Cycle Dynamics with Factor Structure and Regime Switching," International Economic Review, 1998, 39, 969-996.

Below is the graph of the index. enter image description here

Once you read all the relevant literature, you can start thinking of your own index.

$\endgroup$
1
  • $\begingroup$ Thank you. I just want to test the current indexes that are applicable to macroeconomic data, not make a new econometric indexing procedure to test my hypothesis. I appreciate you showing me the Chauvet paper. I will continue looking for peer-reviewed journals. Thanks again. $\endgroup$ Feb 12, 2015 at 23:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.