Wednesday, March 29, 2017

“In Order That They Might Rest Their Arguments on Facts”

There is considerable concern among economists and social scientists that the Trump administration will take a hatchet to the government statistical programs. I find it difficult to argue for the value of government statistics, because their value seems so obvious to me that it is hard for me to imagine someone who both disagrees with me on this point but is potentially persuadable. However, Nicholas Eberstadt, Ryan Nunn, Diane Whitmore Schanzenbach, and Michael R. Strain have taken on the job a March 2017 working paper, "“In Order That They Might Rest Their Arguments on Facts”:The Vital Role of Government-Collected Data." The authors are from both the Hamilton Project at the Democratic-leaning Brookings Institution and the Republican-leaning American Enterprise Institute. 

They point out that total federal spending on statistics is about 0.18% of the federal budget--and just to be clear, that's not 18%, but rather a little less than one-fifth of 1 percent.  As the authors point out in detail, with examples, the potential benefits of this information are considerable. The federal budget is about $3.6 trillion, and of course the federal regulatory apparatus imposes additional costs. Information helps to direct government spending, taxes, and regulations, and it helps citizens to hold their government accountable. In addition, businesses and households often build on government statistics when making their own plans and decisions, thus allowing the economy to function more smoothly than if this information was only available, in partial chunks, from private providers.

But these arguments for the value of government statistics are old and well-known; indeed, they date back to the legislation involved in the first Census, back in 1790. Section 2 of the just-adopted US Constitution called for an enumeration of people to determine the number of members each state would have in the House of Representatives: "The actual Enumeration shall be made within three Years after the first Meeting of the Congress of the United States, and within every subsequent Term of ten Years, in such Manner as they shall by Law direct." But when the bill to enact the first Census came up in 1790, James Madison (then a member of the House of Representatives) argued that there was a great opportunity here to do more than just counting heads, and that it would be useful to gather more information. Our records of Congressional debates from that time do not quote exactly verbatim, but instead are paraphrased. The fuller comments attributed to Madison are below, but here's are some highlights of what he had to say on January 25 and then on  February 2, 1790:
"This kind of information, he observed, all Legislatures had wished for; but this kind of information had never been obtained in any country. ... If the plan was pursued in taking every future census, it would give them an opportunity of marking the progress of the society, and distinguishing the growth of every interest. This would furnish ground for many useful calculations, and at the same time answer the purpose of a check on the officers who were employed to make the enumeration ... And I am very sensible, Mr. Speaker, that there will be more difficulty attendant on the taking the census, in the way required by the constitution, and which we are obliged to perform, than there will be in the additional trouble of making all the distinctions contemplated in the bill. ... I take it, sir, that in order to accommodate our laws to the real situation of our constituents, we ought to be acquainted with that situation. It may be impossible to ascertain it as far as I wish, but we may ascertain it so far as to be extremely useful ...  If gentlemen have any doubts with respect to its utility, I cannot satisfy them in a better manner, than by referring them to the debates which took place upon the bills, intend, collaterally, to benefit the agricultural, commercial, and manufacturing parts of the community. Did they not wish then to know the relative proportion of each, and the exact number of every division, in order that they might rest their arguments on facts, instead of assertions and conjectures?"
The modern arguments for government statistics are pretty much all there. From my perspective, I would only add that it's very useful to have statistics that are publicly available, and where the methods are openly discussed. You might not like, say, exactly how the poverty line is defined, or how the government statistics draw the line between those counted as unemployed and those counted as "out of the labor force," but at least the method is clear and the same method is used over time. The process is isolated from politics in various ways. For example, the many tasks involved in producing government statistics are typically divided up so that no one person can tell what the results will be until the end. Groups of outside experts are called in on a regular basis to evaluate and critique. Politicians have limited input to the actual data process. For example, "In the case of the BEA [Bureau of Economic Analysis], political appointees have very limited access to the data until after journalists receive them—one hour before the estimates are made public— further ensuring the impartiality of the process .."

I'm all for managing every little sliver of the budget as effectively as we can. But to me, government statistical programs are part of what makes citizenship possible.

Maybe the best way to get an idea of the information available through the government is just to list some of the agencies involved, and to think about how government policies, business choices, and personal decisions would potentially be affected if substantially less information was readily and publicly available in these areas: 
"A substantial portion of our official statistics is produced by the 13 agencies that have statistical work as their principal mission. Excluding funding for the decennial census, approximately 38 percent of overall funding for Federal statistical activities provides resources for these 13 agencies. The principal statistical agencies are the: Bureau of Economic Analysis; Bureau of Justice Statistics; Bureau of Labor Statistics; Bureau of Transportation Statistics; Census Bureau; Economic Research Service; Energy Information Administration; National Agricultural Statistics Service; National Center for Education Statistics; National Center for Health Statistics; National Center for Science and Engineering Statistics; Office of Research, Evaluation and Statistics (SSA); and Statistics of Income (IRS). ... The remaining 62 percent of total resources allocated to statistical work in the U.S. Government is carried out by about 115 programs in the Executive Branch that conduct statistical activities in conjunction with another program mission, such as providing services (for example, medical care benefits for the elderly and the poor) or enforcing regulations (for example, with respect to the environment, transportation, or occupational safety). ... Additionally, there are other Federal agencies whose statistical activities are excluded because they are not part of the Executive Branch. These agencies include the Congressional Budget Office, which develops and applies projection models for the budgetary impact of current and proposed Federal programs; the Federal Reserve Board, which compiles the widely used Flow of Funds report and other monetary statistical series and periodically conducts the Survey of Consumer Finances; and the U.S. Government Accountability Office, which develops statistical data in evaluations of government programs." 
Afterword: Here is the fuller version of the comments attributed to Madison. On January 25, 1790, the House of Representatives began to consider the actual bill that would implement the census that was called for the by just-adopted Constitution.Here is the paraphrase of Madison's comments that day: 
Mr. Madison observed, that they had now an opportunity of obtaining the most useful information for those who should hereafter be called upon to legislate for their country, if this bill was extended to as to embrace some other objects besides the bare enumeration of the inhabitants; it would enable them to adapt the public measures to the particular circumstances of the community. In order to know the various interests of the United States, it was necessary that the description of the several classes into which the community is divided should be accurately known. On this knowledge the Legislature might proceed to make a proper provision for the agricultural, commercial, and manufacturing, interests, but without it they could never make their provisions in due proportion. This kind of information, he observed, all Legislatures had wished for; but this kind of information had never been obtained in any country. He wished, therefore to avail himself of the present opportunity of accomplishing so valuable a purpose. If the plan was pursued in taking every future census, it would give them an opportunity of marking the progress of the society, and distinguishing the growth of every interest. This would furnish ground for many useful calculations, and at the same time answer the purpose of a check on the officers who were employed to make the enumeration; forasmuch as the aggregate number is divided into parts, any imposition might be discovered with proportionable ease." 
And here is the fuller paraphrase of Madison's comments on February 2, 1790:
And I am very sensible, Mr. Speaker, that there will be more difficulty attendant on the taking the census, in the way required by the constitution, and which we are obliged to perform, than there will be in the additional trouble of making all the distinctions contemplated in the bill. The classes of people most troublesome to enumerate, in this schedule, are happily those resident in large towns, the greatest number of artisans live in populous cities, and compact settlements, where distinctions are made with great ease.
I take it, sir, that in order to accommodate our laws to the real situation of our constituents, we ought to be acquainted with that situation. It may be impossible to ascertain it as far as I wish, but we may ascertain it so far as to be extremely useful, when we come to pass laws, affecting any particular description of people. If gentlemen have any doubts with respect to its utility, I cannot satisfy them in a better manner, than by referring them to the debates which took place upon the bills, intend, collaterally, to benefit the agricultural, commercial, and manufacturing parts of the community. Did they not wish then to know the relative proportion of each, and the exact number of every division, in order that they might rest their arguments on facts, instead of assertions and conjectures? Will any gentleman pretend to doubt, but our regulations would have been better accommodated to the real state of the society than they are? If our decisions had been influenced by actual returns, would they not have been varied, according as the one side or the other was more or less numerous? We should have given less encouragement in some instances, and more in others; but in every instance, we should have proceeded with more light and satisfaction.