Skip to main content
Log in

Some Basic Issues in Time Series Modelling

  • Original Article
  • Published:
Journal of Quantitative Economics Aims and scope Submit manuscript


Modern time series analysis has been primarily concerned with an understanding (description and prediction) of time varying behaviour of economic phenomenon. However, the patterns of time varying behaviour have not remained the same over the years. The reasons for this could be several. Among others, mention may be made of: (i) the changing character of markets in the wake of liberalization of economies and globalization of trade and markets, and (ii) new and alternative ways of looking at time varying behaviour, given the availability of high frequency or longer time series data, computing algorithms and software, and new techniques and tests. In the backdrop of this, addressing the concern of understanding time varying behaviour as accurately as possible, poses a formidable challenge to the time series econometrician. Some clarity on the basic issues of time series modelling however, might provide some relief. Time series modelling may be carried out in time and frequency domains. In time domain, time is the unit of analysis. The dynamic evolution of a phenomenon over time is traced with the help of suitable statistical models designed for this purpose. In the frequency domain, the unit of analysis is frequency component. Under this approach a given time series is decomposed into frequency components and statistical analysis is undertaken to trace the behaviour of components. In both the approaches, what is studied is the inter-temporal dependence (association) among the past, present and future of a time varying phenomenon, from a given set of sample data. The only difference between the two approaches is that in the time domain the temporal properties of a variable are studied on the basis of realizations recorded at a pre-determined frequency, while in the frequency domain, the sample realizations depend not on a single frequency component but on several components. As a result, there is an absence of frequency based information in time domain, and vice versa. But both are equivalent approaches. Depending on the nature of the problem at hand, the choice of domain may be decided (Granger in Econometrica 37(3):424–438, 1969). It would however be more profitable if both time and frequency information are simultaneously available. This is possible in the time–frequency domain, known as wavelet analysis. The present paper highlights some basic issues of concern in time, frequency (spectral) and time–frequency (wavelet analysis) domains, by placing them in a comparative context. The paper is organised into three sections, one each for each of the domains of analysis. The first section deals with the basic issues in time domain, followed by the frequency and time–frequency domains in the next two sections.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others


  1. See Priestley (1981) for spectral estimation algorithm.

  2. Both DWT and MODWT based MRA are implemented via Mallat’s pyramid algorithm. See Gencay et al. (2002) for further discussion.


  • Bhandari, A., & Kamaiah, B. (2017). On the dynamics of inflation-stock returns in India. Journal of Quantitative Economics, February pp 1–17. Springer, New York.

  • Fishman, G.S. 1968. Spectral methods in econometrics. Santa Monica: R-453-PR. Rand Corporation.

    Google Scholar 

  • Gencay, R., F. Selcuk, and B. Whitcher. 2002. An introduction to wavelets and other filtering methods in finance and economics. San Diego: Academic Press.

    Google Scholar 

  • Granger, C.W.J. 1969. Investigating causal relations by econometric models and cross-spectral methods. Econometrica 37 (3): 424–438.

    Article  Google Scholar 

  • Hatemi, J.A. 2012. Asymmetric causality tests with an application. Empirical Economics 43: 447–456.

    Article  Google Scholar 

  • Masset, P. 2008. Analysis of financial time-series using fourier and wavelet methods. Fribourg: University of Fribourg, Department of Finance.

    Book  Google Scholar 

  • Nachane, D.M. 2006. Econometrics: Theoretical foundations and empirical perspectives. Oxford: Oxford University Press.

    Google Scholar 

  • Percival, D., and A. Walden. 2000. Wavelet methods for time series analysis. Cambridge: Cambridge University Press.

    Book  Google Scholar 

  • Priestley, M.B. 1981. Spectral analysis and time series. London: Academic Press.

    Google Scholar 

  • Ramsey, J. (2002). Wavelets in economics and finance: Past and future, Working Paper, New York University.

  • Shumway, R.H., and D.S. Stoffer. 2011. Time series analysis and its applications: With R examples, 3rd ed. New York: Springer.

    Book  Google Scholar 

  • Toda, H.Y., and T. Yamamoto. 1995. Statistical inference in vector autoregressions with possibly integrated processes. Journal of Econometrics 66: 225–250.

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Bandi Kamaiah.

Additional information

This paper is a slightly revised version of the presidential address delivered at 53rd TIES conference held at NISER, Bhubaneswar during 22–24th December, 2016. The author is thankful to Dr. Avishek Bhandari for extending help in revising this paper.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Kamaiah, B. Some Basic Issues in Time Series Modelling. J. Quant. Econ. 16 (Suppl 1), 1–21 (2018).

Download citation

  • Published:

  • Issue Date:

  • DOI: