Statistics and Other Untruths

 

by Paul J. Updike

INTRODUCTION

The purpose of this article is to argue that many macroeconomic measurement statistics like GDP are approximations at best. At times, these approximations reflect reality, but sometimes they miss reality by a country mile. One of my economics professors suggested that some economists in the field of economics considered econometrics to represent the ‘rubber-ruler’ approach. The colorful image in my mind’s eye of a straight-edge that twists in the wind in order to connect data points is unforgettable. There is a natural human tendency to ascribe more accuracy to numbers that are in print than those numbers may deserve. Please remember that human tendency when employing econometrics and other macroeconomic tools.

In order to make meaningful economic-based comments, I often end up quoting GDP statistics to buttress this idea or sell that point. However, it is important to remember that neither access to nor use of computers have changed human nature much from that found in the quote the economist David Colander puts at the beginning of Chapter 7 of his 4th edition book, Macroeconomics -- "The government is very keen on amassing statistics … They collect them, add them, raise them to the nth power, take the cube root and prepare wonderful diagrams. But you must never forget that every one of these figures comes in the first instance from the village watchman, who just puts down what he damn pleases." Sir Josiah Stamp - head of Britain’s revenue office in the late 19th century. (Colander, p160)

If one is looking for absolute truth, economics is probably the wrong field. Meaningful macroeconomic analysis depends on statistics. If you are searching for truth, try looking within the world’s religions. Truth might well be found there. (The search for truth is a much different discussion). However, the point of this writing is to propose that one may find approximate statistics that are close enough, such that meaningful economic conclusions can be drawn.

By the way, the main reason an abundance of economic data actually reflects reality is because of taxes. This lecture will discuss what kinds of economic statistics are more or less likely to be meaningful and why.

COMMENTS REGARDING GATHERING STATISTICS

Trying to measure GDP in a highly industrialized country that has well-developed markets is hard, but do-able. Measuring GDP for an economy in a less developed country with immature markets that suffers from serious transportation and communication infrastructure problems is much more difficult.

But such measurements abound all around us. One time I performed a Goggle search on the Internet for ‘Nations and GDP’ and got 317,000 resulting sites. May I comment on just two of these sites? When I went to Mr. Dowling’s Interactive Table of World Nations site, all kinds of fascinating statistics appear, listed by individual nation. But may I point out below significant problems with these numbers? When I went to the United Nation’s site, United Nations Statistics Division - Common Database, I found hundreds of records and links, a real plethora of numbers. Do any of these measurements reflect reality? Can these numbers be used for meaningful economic analysis? Good questions.

At the bottom of his World Nations chart, Dowling admits that the "literacy rate figures in each country are only a rough estimate" [emphasis mine] and that "nations often have unique methods to define literacy". How rough is rough? Plus or minus two percent, plus or minus 20%, or plus or minus 50%? Yet Dowling includes these questionable statistics in the table like they are facts, even though he knows that literacy numbers for many countries can be relatively meaningless. Dowling cites the CIA World Factbook as his source. Many people wonder how good the CIA is at spy-related intelligence work, their main task. Why would we conclude that CIA people are better at collecting meaningful economic statistics?

Is it possible that many of the nations included in Dowling’s list also have "unique methods" for measuring GDP? Besides import and export trade with various sophisticated countries that have better measuring tools, do nations use ‘rough estimates’ of their own country’s economic activity? What if they are mostly command economies? Who is going to argue with the strongman running the country? Do the ‘worker bees’ in the United Nations want to lose their jobs over an argument with a particular country’s GDP numbers because the worker bees are not happy with the collection methodology? I would doubt it. When exports of commodities to industrialized nations are the biggest contributor to their GDP, then perhaps GDP can be reasonably estimated. But that still makes GDP an estimate.

SOME PROBLEMS WITH NUMBERS

When serving as a controller for several companies in Arizona, I personally answered many government requests for financial and employment information about my company. These requests never rose to a high priority. We had a business to run and profit to make! A recent Phoenix Chamber of Commerce reminder I received said that the government information request forms needed to be returned for all businesses with five or more employees or the company faced a $600 fine. A $600 potential fine gets higher priority than no fine, but what about firms with one to four employees? Do they count too?

The government's reach into local and statewide markets and most companies in highly developed countries, as well as the public disclosure of financial information for many companies ensures that there is rhyme and reason to most of the economic numbers in America. In Arizona, firms needed to file financial information like income statements and balance sheets with the state government every year. Yet if the company is privately held, perhaps some of those numbers filed with the state are not accurate. (The Enron, WorldCom, Tyco and Global Crossing debacles in the early 2000s demonstrate that public companies can misstate financial numbers too.)

Have you ever tried to call a state employee to ask for help filling out state government forms? After a thirty-minute hold (who has 30 minutes to spare in a ‘for-profit’ business?) someone finally picks up and says, "I don’t know the answer, you need to talk to so and so to ask". Whenever this happened to me, I never called back, but did the best I could without assistance and sent the form back to the state government. Then those numbers I struggled over, the numbers I knew were not very accurate were added together with all the others. When I saw GDP numbers that included the ones I sent to the government, I said "Oops!"

GLOBAL OIL USAGE NUMBERS

When numbers get large, many people, even economists, can make mistakes. May I illustrate? There are approximately 70 to 75 million barrels of oil produced and consumed in the world every day in 2001. (See link below – That total has grown to approximately 85 million barrels daily consumption in 2007.) http://www.eia.doe.gov/pub/oil_gas/petroleum/analysis_publications/.htm

This approximates a 25,000 million barrels (25 billion barrels) annual usage rate. In the world in 2001, there are approximately 1 trillion barrels of proven oil reserves including approximately 600 to 700 billion barrels of proven oil reserves in the ground under the Middle East. (See link below.) That means that if the global oil usage rate stays approximately the same, the world has several decades of oil available – doomsayers to the contrary. http://www.eia.doe.gov/pub/oil_gas/petroleum/analysis_publications/.htm

The course textbook by David Colander (4th edition) is perhaps the best overall, and most readable book on the subject of macroeconomics I have ever perused. Yet even careful economists can misuse numbers. On pages 92-93 of the 4th edition, Colander uses the following ‘shift in supply versus movement along the curve’ example, "At $15 dollars a barrel, producers supplied 1,500 million barrels of oil a day, and at $36 a barrel they supplied 1,750 million barrels a day." These are not real oil consumption numbers!

If you do the math extension with Colander’s numbers, i.e., 1,500 million barrels of oil a day for 365 days a year at $36 per barrel, you arrive at a dollar amount that is twice the size of the total American economy in 2002. This is a rare example of an error with numbers by Colander. Yet, and more importantly, his supply curve example in Figure 4-6 is not affected at all by the relative size of the numbers. But it is instructive to realize that we cannot be too careful with statistics, which are the lifeblood of macroeconomics.

POPULATION COUNTS

We have a constitutionally mandated census in America every 10 years. Immediately afterwards, the politicians argue about the correct statistical estimation method to use to include the one, two or three percent of people excluded in the actual count. Do two or three percentage points matter? The United States official census result (with estimates included, but not always noted boldly) is published, then estimated again each year for the next ten years until we officially count again, followed by estimation again. We can be confident that the total is close. But how does one know which yearly population measure is being used for per capita GDP measurements? Are the population counts (estimates?) and the GDP measurements (estimates?) calibrated for the same 12-month period? We can only hope that both the collectors and users of these statistics are careful when they match various numbers.

What about lesser-developed countries in Asia or Africa or Central or South America with serious communication and transportation issues? Is an economist for the World Bank, the International Monetary Fund or the United Nations, who resides in a nation with safe drinking water and plenty of flush toilets, actually going to visit a nation that has unsafe drinking water and few flush toilets long enough to participate in the population count? Or might it be easier to use the numbers that a government employee of that country in question gives you regarding the population in that nation’s rural areas and not ask too many questions?

Estimation, using ‘tried and true’ statistical methodology from one’s desk is often much cleaner, faster, and safer. Besides, there is a reason to believe that native who represents their government. All governments are very interested in establishing the exact population in their country. It is for a simple, but universal reason. If you count your citizens, then you know who they are and where they live and you can tax them much more easily.

TAXES

One of life’s certainties is that your government will attempt to collect taxes from you. The government’s bias is to collect as much in tax revenue as possible. Therefore the government makes all kinds of measurements and gathers all sorts of statistics to assist themselves in that quest. If a nation’s income tax compliance is high, then corresponding government statistics drawn from tax information tend to reflect reality. If income tax compliance is low or poorly enforced, then we ought to consider questioning whether or not some of the economic statistics emanating from that country are meaningful.

Therefore, we need to know how taxation works within a country. What is the source of government revenues? Because fiscal policy has a hand-in-glove relationship with tax collection, our economic analysis is improved by our knowledge about taxation in a country. There are usable economic statistics collected alongside tax revenues.

When I refer to taxes, I do not simply mean income taxes, but also include customs, tariffs, excise taxes, license fees, payroll taxes, sales taxes, access fees, city taxes, county taxes, state taxes, corporate taxes, property taxes, estate and gift taxes, etc., etc., etc., (there are literally hundreds of taxes most of us pay). However, not all market activities are taxed, even in a highly developed country. If GDP measures market-based activity, what about significant market-based activity that often happens outside official (taxed) channels?

The so-called underground economy or black market is alive and well in most countries. There are two main reasons: 1) they are trying to hide the production and/or distribution of illegal goods and services like illicit drugs or prostitution; 2) they are involved in the legal production and/or distribution of goods and services that they choose not to report. People try to avoid paying taxes, if they can get away with it, everywhere, which is why the government wants as accurate statistics as possible. Sometimes government officials help people avoid paying taxes as long as the beneficiary is willing to pay something to someone. One meaningful word that can be found in most languages around the world is the word ‘graft’.

When I lived in Taiwan in the 1970s, I could buy American products with American money from merchants that did not officially report or pay taxes on their black market-related activity. Shipments of goods from America were never secure because they first passed in front of corruptible government officials. Government corruption is not the sort of activity that is usually reflected in GDP statistics. How could it be? Because there is no official record, the activity did not actually take place. :-) Is black market activity equal to 2%, 5%, 10% or 25% of a country’s economy? Who knows?

ECONOMIC STATISTICS

From personal experience I can tell you that oranges cost many times more on the streets of Tokyo than on fruit stands or in grocery stores in Arizona. Economists apply what is called Purchasing Power Parity (PPP) to GDP measures for various countries to normalize currencies to determine the same value for the same basket of goods in each country. Dowling includes PPP in his World Nation's per capita GDP column cited earlier in this paper. But exactly what basket of goods is used to calibrate Dowling's GDP numbers? Does he use nominal or real GDP numbers consistently throughout the table? Are population counts used in calculating per capita GDP from the same year as GDP numbers?

Theoretically, by adjusting for PPP one can make comparisons of per capita GDP figures from different countries, apple to apple. The theory is quite reasonable, and allows for meaningful analysis in macroeconomics. Yet the idea of one basket of goods for every country can be problematic. If one major food in the basket of goods is not available in another country, what do we substitute instead? Are substitutes for foodstuff in the basket of goods the same in nutritional value? What if haircuts always include a wash, massage and a shave in one country, but most people never get massages and shaves when they get haircuts in another country? Should one add up or subtract down to equalize the basket of goods and services used to measure PPP?

Year-to-year statistics are much better barometers of reality, if they are collected the same way, by the same people, in the same country, etc. (Ceteris paribus – other things being equal). This is important because it means that yearly data within the same country is often reliable and can be fruitfully analyzed. Comparing measures of growth within an economy from year to year, or even quarter to quarter are meaningful analyses! However, once we leave one country and move to another, we need to remember that just like the changing appearance and value of the paper currency, economic statistics change too. Economic statistics may be called a similar name, but may not be measured the same way in a different country.

CONCLUSION

There are several ways to ensure that you are using meaningful statistics. One way is to relate your numbers to the taxation numbers being used within that country. If that country is highly industrialized (i.e. most market activities are taxed) then you are more likely to find numbers that approximately reflect reality.

Another way is to find multiple sources that have the same or very similar information. Ignoring the possibility that several different sources all used the same original source, then by comparing and contrasting the numbers, you develop a meaningful range of information. If the range is relatively tight, then you likely have good information. To use economic statistics well, it is always best to define and stay within the relevant range.

Yet another way is to check the footnotes for the year the data comes from for information how the data was collected. Be wary of questionable sources that do not include assumptions and explanations on how the numbers are collected and calculated.

Try to avoid situations which insist you compare the same-named numbers for different countries. When you have to compare different countries in the same sentence, identify known or likely problems with the data. Economic tables that show comparison economic data from year to year are the most likely tables to have good data. Then you can use your own analysis to generate your own measures of growth or other economic data.

Finally, beware of the political agenda of the people who collect the data, then present that collected information as fact. A political agenda can cause enough bias in the data collector that important corollary information can be overlooked or even altered. Sometimes that corollary information is intentionally overlooked to fit the political bias of the data collector. It is not that unusual for collected economic data to be intentionally altered to fit a political agenda. Always remember to use economic information in the same context in which it is found and cite the source.

One of the best economic statistics web sites I have found is at the following link: http://www.bea.doc.gov/bea/dn1.htm.

(All links cited in this article worked at the time I first used them, but may no longer work.)

REFERENCES:

Colander, D.C. Macroeconomics, 4th edition, Irwin/McGraw-Hill, 2001

Global Oil Consumption Report:

http://www.eia.doe.gov/pub/oil_gas/petroleum/analysis_publications/.htm

Global Oil Consumption report:

http://www.eia.doe.gov/pub/oil_gas/petroleum/analysis_publications/.htm

Fortune magazine, September 10, 1990, Special Oil Report, pp 35-56

Robert D. Hershey, Jr., Counting the Wealth of Nations, New York Times, December 19, 1995

US Overview: Shifting views on Inflation www.senate.gov/~budget/republican/analysis/archive/eb4.pdf

Mr. Dowling, Interactive Table of World Nations

United Nations Tables and Graphs, United Nations Statistics Division - Common Database