Our Blog

https://sites.google.com/a/rockvilleanalytics.com/rockville-analytics-group/Our-Blog

August 18, 2021

I have gone through the pages of various journals in Animal or Veterinary Science. There is a recent focus on evidence-based clinical practice where a lot of clinical data from one large hospital chain, Banfield Pet Clinics is analyzed. In the process the knowledge gained from clinical practice is often labeled subjective and not based on credible scientific evidence. There seems to be a tendency to ignore clinical experience based decisions in favor of what the data analysis shows. But there is a better course. One can combine clinical knowledge through a structural model and then use the statistical evidence from Banfield data to buttress that model. This was we use both data as well as the clinical knowledge. Strangely enough this approach was developed in 1915-20 by Sewall Write, an animal geneticist.

Kris

December 2, 2015

We talk about analytics with the latest statistical techniques and big data playing a very important role in business policy. It is quite unfortunate that the same big data is available in global economic environment but the economic policies at the global level, like at IMF and World Bank do not seem to apply latest analytics tools. Let us look at the issue of which currencies should enter the basket of the hard currencies in IMF policy. IMF has with it data on international trade and finance, domestic finance and national economic statistics of various nations for various years. There ought to be some definite patterns in that big data that are stable in the long run. Can we use that pattern to develop global policies on what currencies should be in the basket of hard currencies? If we did this I am sure China's Yuan should have been picked up as a hard currency quite a few years ago reducing much of economic damage experienced by USA and China.

In this connection, I am tempted to reproduce below my comments on Yuan being recognized as a hard currency a few days ago.

I am remembering the forgotten Belgian born Yale economist-Robert Triffin who wrote the famous book in 1960(!) " Gold and The Dollar Crisis: The Future of Convertibility". Yesterday's news that Chinese Yuan has been chosen as one of the hard currencies is almost like 1971 situation when the IMF was driven by circumstances arising over a decade to change over from fixed exchange rate (linked to US$35 for an once of gold) to a flexible exchange rate. Robert Triffin argued way back in 1960 that a fixed exchange rate would result in a run for US Dollar that cannot be matched by the gold reserves it had. He recommended a flexible exchange rate. What happened then? IMF did not follow his advice. With Lyndon B. Johnson's Great Society agenda, with huge fiscal deficits, the value of dollar fell down forcing IMF to switch to flexible exchange rate in 1971, a decade later. This is because we can not have inflation at home due to budget deficits and claim a fixed exchange rate for dollar abroad (fixed at US $35 an ounce of gold)!

If I recall well Robert Triffin also said that the basket of currencies in the IMF reserve currencies must be based on contribution to world trade. So, I think that this new IMF policy is also two decades in waiting, after a considerable damage is done to the US economy! Remember that dollar value is artificially kept high and yuan artificially low as dollar is a reserve currency and yuan is not. If this were done two decades ago the dollar would have become weaker (due to its trade deficit) Chinese yuan would have become stronger (due to its trade surplus) the US debt to China would have been much smaller, US imports from China would have been much smaller, and so on..... Has anyone read my scribbling in these pages on US trade deficit and foreign debt, month after month, for several months?

If Robert Triffin lived beyond 1993 he would have written another book "Foreign Trade and The Yuan in Crisis" and would have said in 1990s that IMF should include in its basket all currencies that are needed in trading goods and services. In this connection I also recall a 1947 Econometrica article by first Nobel laureate in economics, Ragnar Frisch, in which he advocated that a nations' trade must balance in the medium to long run, which can only happen with flexible exchange rates and all major trading currencies being in the IMF basket.

Kris

December 1, 2015

I hope everyone had a wonderful Thanksgiving and getting ready for Christmas. Wishing all data analytics community a very Happy and Prosperous Holiday season and a New Year.

Data Analytics is picking up. More and more workshops, more and more applications, more and more data warehouses. more and more awareness of the usefulness of data analytics. Along with this also comes a concern if this craze for data analytics is really worth it as revealed by the benefits and costs. This concern is not to be brushed aside as insignificant by the data analytics professionals like us. If we do data analytics will have the same fate as statisticians working for businesses, playing only the second fiddle and loosing the prominence for data analysis in business it deserved before the data analytics revolution in 2006. There is a need to tell stories to the general public how data analytics generated higher dollar or yuan benefits (now that yuan is also a hard currency!) with much lower costs in the same currency units.

I have been in LinkedIn loop with Intuit, a small and an upcoming Data Analytics company with product line that helps small businesses. Their core group seem to be highly talented. Let us hope that it will become a major player in Data Analytics space. Here is a link to intuit:

http://www.intuit.com/company/

Kris

January 14, 2015

Time to rethink about the official statistics. In a rapidly changing data scene with big data that is available in the private sector official statistics are still frozen in a frame with small frequency data. It is also necessary to reexamine what types of data need to be collected by the "state" and what data collected by the private businesses must be regulated for quality and security. There is a very long history of collection of official statistics. From times immemorial Kings, small and large, collected land use data to help them assess taxes. Likewise they collected statistics on commodities traded to collect additional taxes and custom duties. Vital statistics were also collected by Kings. Starting from the industrial revolution they collected statistics on industrial output and employment. These were meant to aid the governments to collect corporate taxes and to protect employee welfare. Data can be collected by private sector or by the state. State starts collecting data when it notices that there could be a conflict of interest in how the data is used. Even today it is alleged that an Indian businessman has two books, one for himself and one for the tax collecting agency. Some data needs to be preserved for the posterity and the private sector has no incentive or resources to preserve data. The state collects reliable and impartial data and preserves it. Data collection through censuses and sample surveys was costly and hence it tended to be a low frequency data.

But the entire data scene had changed quite rapidly in recent years. The computer technology and the advent of the Internet has created "big data" and very high frequency online data. It also made personal and sensitive secret information flow through the information gateways. Information has value while the cost of its collection has almost dwindled to nothing. Security breaches through hacking have increased by leaps and bounds. Some of the low frequency data collected by the state had lost part of its utility value in view of the high frequency data that is available and because of the more sophisticated data analysis tools that use high frequency and big data. New data security issues arose with cyber attacks on data, often personal and sensitive data. New kind of regulations are needed on who can collect personal and sensitive information and what kind of security precautions they should take. A new thinking is needed RIGHT NOW on the role of the state and the types of partnership between the state and the private sector in collecting and preserving official statistics and in maintaining the security of personal and sensitive information.

We see in today's newspaper the President Obama is calling for a bipartisan effort to make new laws on cyber security. I hope one realizes that this is a global issue. Not only the US Government but all other governments and the United Nations should consider new ways of collecting and preserving data and its security through newly legislated laws and regulations to replace the old ones. The United Nations' Economic Commission for Europe seems to be seized with this problem and recently it published its report on this issue. See: http://www1.unece.org/stat/platform/display/bigdata/2014+Project

Kris

Apil 1, 2014

There is a crisis in macroeconomic modeling. It is openly admitted by people at te helm of affairs that the theories and models they based their judgments on to formulate economic policies were wrong. Shall we recall the famous congressional hearings abput the financial meltdown. Henry Waxman to Alan Greenspan: “In other words, you found that your view of the world, your ideology, was not right, it was not working."

Greenspan: “Absolutely, precisely. You know, that’s precisely the reason I was shocked, because I have been going for 40 years or more with very considerable evidence that it was working exceptionally well.”

See the video clip below:

http://www.c-span.org/video/?c3342718/waxman-greenspan-testimony

For my own views on Macroeconomic modeling, with a business analytics orientation, see my recent paper published in 2013, Vol 3, No. 2 Issue of Journal of Indiana Academy of Social Sciences. The abstract is here:

https://docs.google.com/file/d/0B0WaPggKV2mkUWJNb2drTTFMdTA/edit

Kris

March 31, 2014

I attended a session labelled Big Data at the Management Summit of University of Maryland on March 28. The discussion was on using big data. These days everyone talks about big data and using data to predict crises such as the missing Malaysian aircraft or why we could not predict the Financial crisis in 2008 etc. One more day is gone and still no definite clues on what has happened to the missing Malaysian aircraft and its passengers. People are asking how could this happen in a high tech era, that includes big data and data analytics or data science. Let us also look at what happened to Federal Bailout of Bears and Sterns, and General Motors later. Federal Government bails out GM and GM uses a faulty sparkplug to cause 13 fatal accidents. Is the Federal Government bailing out inefficient companies creating more probable environment for economic failures? Was the finnacial crisis in 2008 not predicted or was it allowed to happen? Are all these related? Has analytics, or data science, anything to say on this situation? Data Science is not only a science about the data we have, but it is also the science about the data we donot have. It is about the data generating process.

Yes. They are all related. Accidents are rare events. To predict or understand why rare events happen we need, not more information or big data, but different kind of information, including some information that is crucial but is not made available. Does the data generating mechanism have inherent mechanisms to suppress information? If so what can be done to go back to the data generating process to get the suppressed information? Without getting that unavailable data we cannot really model the rare events satisfactorily. No doubt the larger the data, more likely it is that we get some information on rare events. But there may be mechanisms in the data generating process that suppress the information needed to predict the rare events. In that event, however big is the big data, we will not have adequate information to model rare events. Many organizations have organizational hierarchy and people at lower tiers of hiearchy are afraid to report some information that their bosses do not like to hear. Even when the bosses get the information from their subordinates they do not make that information public and suppress it. It is these that quite often cause the accidents and rare events. I read a very interesting book on this topic by Chetan Dhruve, a friend of my son, titled Why Your Boss is Programmed to be a Dictator?. The detailed case study of the Challenger space accident (along with a few other accidents) makes it a very convincing argument. See the link: http://www.cvdhruve.com/bookhome.html

Kris

March 22, 2014

Analytics is very useful in solving a variety of real life problems we face. The most pressing problem we are facing today is the location of the lost Malaysian Flight.

In 2011Analytics was used through a Bayesian approch to locate the most likely position of a lost flight.

One may see the link: http://www.bbc.com/news/magazine-26680633

I thank my friend Vinod Vyasulu for bringing this to my attention.

Kris