Devrup Ghatak writes:
I am a student of economics and recently read your review of Mostly Harmless Econometrics. In the review you mention that the book contains no time series. Given that your book on data analysis (Data Analysis using Regression) does not contain any time series material either, I wonder if you happen to have any favourite time series reference similar in style/level to the data analysis book.
I don’t know. The closest thing might be Hierarchical Modeling and Analysis for Spatial Data by Banerjee, Carlin, and Gelfand, but I don’t know of anything focused on time series that’s quite in the format that I’d prefer. This is not my area, though. Maybe you, the readers, have some suggestions?
Enders is the closest I imagine.
Wow—Amazon lists Enders’s book at $132.87. I guess that’s what it’s like if you’re an economist, you go and charge market value!
+1 for Enders, but keen on hearing other suggestions.
Probably off topic, but I loved the classic by Box & Jenkins. Don’t know if that’s considered dated now.
Rahul:
I enjoyed reading Box and Jenkins, but I do think it’s a bit old-fashioned. It’s a fun read in that it’s good to see top data analysts at work, no matter what tools they are using. But I can’t imagine these are the best tools to use, at least for the time series that I encounter.
I’ve heard Prado & West’s book recommended, though I don’t know the range of time series offerings well enough to vouch for it as opposed to, say, Enders. Prado & West is only $87 on Amazon, compared to Enders’s $132.
Not affiliated with the site, but I do see that alibris.com has a bunch of used copies listed in the ~$60 range. But pay attention to which edition they are selling. I expect that other used book sites might carry this as well.
Sorry — meant to say I was talking about Enders above.
Maybe too specialized, but I really like Diggle’s book: http://www.amazon.com/Time-Series-Biostatistical-Introduction-Publications/dp/0198522266
Such a book would be a huge benefit to astrophysics too.
The only book I have ever really used is Bretthorst’s now-public-domain monograph on identifying frequencies in time series probabilistically, which is limited in scope but very enjoyable go to this link and scroll down to the bottom where he links to his book.
I love that book. It’s not a cookbook which probably turns off many, but it’s incredibly useful in the right hands. I heard rumors in the 90’s it was being used by finance quants, but could never get any definite details on it.
While it’s not time series in the Arima tradition, I’ve found Wooldridge’s Econometric Analysis of Cross-Section and Panel Data to be enormously lucid and insightful — much more helpful than Greene. And given the kind of data I typically deal with, it’s much more practical as well.
I agree on Enders. I think is the best introduction to Time Series (particularly if you are an economist). Andy: I doubt Enders has a say on pricing. A second book I would suggest is Time Series Analysis by James Hamilton. Again this is a particularly good recommendation if you are an economist. Other possibilities are: Introduction to Time Series and Forecasting, Peter Brockwell et al, Forecasting: Methods and Applications by Rob Hyndman et al. Hyndman has a book online that your reader might want to check: http://otexts.com/fpp/
I’ve been eying http://otexts.com/fpp/ and Hyndman’s hts package, but I haven’t had the time to dig into those. How does that compare?
I am not sure about the style, but the following is a selection of good textbooks for time series (at least from Finance perspective). They are varied in sophistication and approach. I do not think there is one-size-fits-all book in this domain.
* Tsay
* Brockwell, Davis
* Durbin and Koopman (state space approach, good for macro/finance, 2nd edition is well written)
* Mills and (as of the latest edition) Markellos
* Hayashi (gen. econometrics)
I second the Hayashi book!
I was introduced to this book in a Time Series course as an econ grad student and I thought it was great!!
Hayashi is substantially more technical than the Mostly Harmless Econometrics book.
Is the Tsay book “Analysis of Financial Time Series”?
Of course, Hayashi is a graduate level book. Mills is even more technical to my mind. Tsay (indeed Analysis of Fin. TS) is the easiest to stomach.
I’ll just paste the literature list Gunnar Lischeid presented us for the course “time series and wavelets” at the university of Potsdam, Germany, focused on environmental data analysis:
Cowpertwait, P.S.P., Metcalfe, A.V. (2009): Introductory time series with R. Springer
Trauth, M.H. (2010): MATLAB recipes for Earth Sciences. Springer, 3. Auflage
Schlittgen, R. (2001): Angewandte Zeitreihenanalyse. Oldenbourg
Schönwiese, C.D. (2000): Praktische Statistik für Meteorologen und Geowissenschaftler. Gebrüder Bornträger, 3. Auflage
Nason, G.P. (2008): Wavelet methods in statistics with R. Springer
Hipel, K.W., McLeod, A.I. (1994): Time Series Modelling of Water Resources and Environmental Systems. Elsevier 1994 http://www.stats.uwo.ca/faculty/aim/1994Book/
http://www.gfi.uib.no/~nilsg/kurs/notes/course.html
Since time series analysis is mostly applied to macroeconomic data, most of the books deal with methods that have proven useful inn this context. If you prefer a Bayesian view on this sort of stuff, have a look at Gary Koop’s page (http://personal.strath.ac.uk/gary.koop/). Also Fabio Canova has a few neat chapters in his book “Methods for Applied Macroeconomic Research” (see chapters 9, 10 and 11 here: http://www.amazon.com/Methods-Applied-Macroeconomic-Research-Canova/dp/0691115044) or the Reading List of Chris Sims’ course (http://sims.princeton.edu/yftp/Times08/Times08List.pdf).
Hope that helps.
I am still puzzled by the implication that Gelman and Hill on the one hand and Angrist and Pischke on the other have similar style. Indeed reading Andrew’s review in the Stata Journal rather underlines that they don’t, really, unless it’s just being more informal and more friendly than average.
The enthusiasm of economists and econometricians for Angrist and Pischke seems to be based on the minor fact that they do have a sense of humour and the major fact that much of their disciplines’ literature is written in a very rigid and dry fashion, so any departure from norms attracts attention and creates some fans. But it seems unreadable to me unless you have been over the material about one and a half times previously.
That said, I will wave a flag for Chatfield’s introduction to time series, often revised but now seemingly stalled in Chatfield’s retirement: http://www.crcpress.com/product/isbn/9781584883173 It’s not rigorous, it is now longer up-to-date and it’s a bit of a dog’s dinner. So why I do like it? It’s short. It covers several methods briefly and is not bedevilled by the idea that there’s only one approach to time series worth taking seriously. (I still encounter people who discount anything not in Box and Jenkins.) It’s friendly. It shows a respect for real data and the need to do data analysis too. It includes a range of examples. Perhaps most of all, it is enthusiastic without overselling the subject.