Friday, October 26, 2012

Viren Srivastava

Recently, a reader of this blog asked I could provide some information about the late V.K. Srivastava, and the substantial contributions that he made to econometrics and to statistics generally.

I'm more than happy to oblige, as Virendra (Viren) was a good friend of mine, a treasured co-author, and a very caring and humble individual.

Viren was a statistician from Lucknow, in Northern India, and he spent much of his career in the Department of Statistics at the University of Lucknow. At the time of his premature death, in 2001, he was Chair of that department.

I first met Viren in 1981 when I hosted his visit to Monash University in Australia. Viren had approached us about a visit on the recommendation of Aman Ullah. The visit was a great success, and was followed by a second visit in 1985. By that time he and I were well into the writing of our book, Seemingly Unrelated Regression Equations Models: Estimation and Inference.

Subsequently I was able to organize a further period of time together, when Viren visited the University of Canterbury in 1989, while I was on faculty there.

During is career, Viren published extensively in statistics and econometric theory. He authored or co-authored three books and over 150 peer-reviewed papers. His research spanned the areas of finite-sample theory, simultaneous equations models, SUR models, and shrinkage estimators.

The details of his published work can be found on this web page, although co-authorship details are not provided there. In addition to the SUR book, I was lucky enough to co-author five published papers with Viren - these are referenced below.

Viren was an extremely hard worker, as all who knew him will attest. This work ethic has undoubtedly been handed down to his son, Shalabh, who is also a highly accomplished and well-published statistician. Viren was also a kind and gentle colleague, even when rebuking me.

The latter usually occurred when I could not immediately remember some mathematical result and was forced to look it up in some reference source. (This was in pre-internet days.) Then, Viren would chastise me, because he couldn't afford the cost of books, and so he had to commit everything to memory. And his memory was phenomenal! He'd come out with comments of the sort: "You should know that matrix result- it's set an an exercise on page xxx of the first edition of Rao's, Linear Statistical Inference and its Applications!

Viren's work and life were celebrated with the festschrift, Handook of Applied Econometrics and Statistical Inference, which was co-edited by Aman Ullah, Alan Wan, and Anoop Chaturvedi in 2002.

The dedication in that volume was admirably appropriate:
"To the memory of Viren K. Srivastava
Prolific researcher,
Stimulating teacher, 
Dear friend" 


Srivastava, V.K. and D.E.A. Giles, 1984. A pre-test general ridge regression estimator: Exact finite-sample properties. Australian Journal of Statistics, 26, 323-336.

Srivastava, V. K. and D. E. A. Giles, 1987, Seemingly Unrelated Regression Equations Models: Estimation and Inference. Marcel Dekker, New York.

Giles, D.E.A. and V.K. Srivastava, 1991.An unbiased estimator of the covariance matrix of the mixed regression estimator. Journal of the American Statistical Association, 86, 441-446.

Srivastava, V.K. and D.E.A. Giles, 1991. Asymptotic properties of the OLS estimator in simultaneous equations models. Journal of Quantitative Economics, 7, 273-276.

Srivastava, V.K. and D.E.A. Giles, 1991. Unbiased estimation of the mean squared error of the feasible generalised ridge regression estimator. Communications in Statistics: Theory and Methods, 20, 2375-2386.

Giles, D.E.A. and V.K. Srivastava, 1993. The exact distribution of a least squares regression coefficient estimator after a preliminary t-test. Statistics and Probability Letters, 16, 59-64.

© 2012, David E. Giles


  1. Its' very kind of you to write something like this recognizing the contributions of a former colleague. As someone with a number of family in the academic world, I can tell you this is not done by academics nearly as frequently as it should be.

  2. Dear Dave, thank you very much for remembering my great teacher Dr. VK Sivastava as we fondly called him. I was fortunate to be his student. It wss also the period while he wss working with you on SURE book. I still remember him showing me the hard copy of the publication when he got it. He was one of the finest human beings I have met in my life. A great scholar, and person with immense humility despite his high accomplishments. Few may know that he also had a great sense of humor. A person in love with his work and complete devotion to statistics. Unfortunately, he left this world too early. Seeing his son doing so well gives me immense pleasure. Thanks again for writing about this great person and my GURUJI (respected teacher). Regards, Kaushal Joshi.

  3. Dear Dave,
    Thank you very much for remembering my teacher Dr. V.K. Srivastava or ‘VK Sahab’ as we fondly called him. Apart from being an excellent teacher, he was one of the finest persons I have ever met in my life. Despite his high accomplishments, he was such a humble person. He truly loved statistics and completely devoted to his work – a very hard worker. Today seeing Shalabh doing well gives me a lot of pleasure. I fail to find words to express my gratitude to him – yes, I was fortunate to be his student. This was the time while he was working with you on SURE and I still remember he showed me the hard copy of the publication. Thanks again for remembering this great person and my GURUJI (respected teacher). regards, Kaushal Joshi

  4. Dear Dave,

    We have n't met and I have n't had any communication with you so far. But I already feel like addressing you by the first name: Reasons: 1.(Indian), I am older than you,2. I like the spirit of your Blogs,and 3. I like your write up about Viren, our common friend, and his son Shalabh. Let me add one thing to describing Viren. Once he wrote a small paper on proxy variables with his student and it was published in JQE. I found a mistake in it and published a small note in JQE. (I sort of specialized in writing small notes on others' errors as they count as line items in CVs for salary raises in USA, rather than writing any serious stuff. I now repent on that strategy)

    I revealed the error through pure intuitin first and then through a derivation. He felt very bad of not discovering that error made by his student. I also felt very bad, later, that I wrote that note on Viren's paper, having seen so many great papers of his in Econometrica and other econometrics journals. Very few Indians in those days had papers in Econometrica.He liked me for that note and I liked him for his great work in econometrics and we became close friends since then.I am a wandering Econometrician whereas he is a full-time devoted econometrician like you. I moved from mathematical economics and economic planning to econometrics to OR to economic consulting, managerial economics, general management, housing economics, poverty studies, development economics, etc (between India and USA and between academics and industry), whereas he and people like you are devoted to econometrics.

    Yes, CR Rao's 1952 book was like a Gospel to some of us, his students. When I first saw Chau's test in Econometrica in early sixties I said to myself, OMG this is what we derived from LR test in CR Rao's course as students years ago and it appears on page ... in Linear Models.

    Krishna T. Kumar
    Retd Professor
    ISI, Bangalore
    Rockville,MD. USA

    1. Krishna: Nice for you to comment on this. I think that a lot of us miss Viren very much.