Feedback: Object Class

[bsweger]

As of May 8, 2015, the data element below has been finalized and can be found here.

• Object Class

[JimHarperDC]

Thanks for the links just above. Handy!

The type of data element and format appear to be missing from the definition. I'm sure I don't have the terminology exactly right, but I believe the definition should specify that it's a number consisting of two to three digits.

[HerschelC]

The definition isn't clear and specific to the data element required. The OMB Circular specifies object class codes, titles and descriptions. I think this should be clear in that you're requesting the object class code as defined in the circular. Unless you're making a conscious decision not to get into that level of detail. (Though, why not?)

[halsteinberg]

The Object class classification structure has been developed piecemeal over many years. It is no longer a logical structure. This causes errors in its use. The assignment of object codes to transactions is now performed to fulfill a compliance requirement, rather than to accumulate useful information. The question, therefore, in developing the object code standard is whether OMB/Treasury is willing to use the DATA Act as an opportunity to create a 21st century object code classification structure; make minor adjustments based on defined criteria; or make no changes which means agencies will continue to assign the existing object codes.

The answer to this question can be based on a decision as to the purpose of assigning object codes to appropriations, obligations, and outlays. For instance, is the purpose to manage better, e. g., recognize that the amounts of monies spent on contractual services to provide a service would not be as effective as spending the monies on personnel compensation? Or is the purpose to control costs, e. g., issuing an edict that for a period of time, no monies shall be spent on purchase of equipment? Or is the purpose to let citizens know how the monies are being spent, e. g., $xxx,xxx has been spent on loans?

Examples of shortcomings in the current Object classifiation structure are:

• 11.3-Other than full time Personnel Compensation and Benefits is a catch-all. There is no opportunity to separate part-time salaries from all other than full time personnel compensation and benefits, even though use of part time employees is a major HR option.

• 11.5-Other Personnel Compensation is also a catch all, for which the opportunity to separately capture overtime compensation, annual leave, and bonuses, all types of spending in which there is considerable interest, is not available.

• 30-Acquisition of Assets, which assumedly is property, plant, and equipment, includes Investments and loans, which represents an entirely different way of delivering services.

• 40-Grants and Fixed Charges does not provide for dividing grants into formula and discretionary grants, or categorical and block grants, each of which represent significantly different ways of proviing financial assistance.

  *There is now clear indication of where Loans shoul be classified.

A logical object code classification for governemnt would have the following at the top level
Personnel services
Contractual services
Materials an supplies
Property, plant, and equipment
Investments
Grants
Loans

The definition of the lower levels, the level at which the information for management, cost control, and accountability is generated, shoul be developed with an interagency process.

[MJB-NIH]

The proposed definitions is sufficient given an acceptance of the existing framework of major object class title, (two digit) codes, and definitions outlined in latest version of OMB Circular A-11.

That is not to say that there would not be some advantages to revisiting the hierarchal relationships of groups (2 digit zero trailing), major (2 digit) and minor (three-digit) classes, as noted in a previous commenter's thread. But being an "old hand" with 30 years experience as a budget analyst, and 20 years as a payroll budget subject matter expert, I would advise we keep the existing structure -- as cobbled up as it is -- to expedite implementation of the Act and rationalize the burden on the enterprise financial systems. Only so much change can be absorbed and still maintain a modicum of quality control in this effort; there will be lots of remapping and cross-walking as it is.

The more we drill -down into "detail codes" -- beyond the 2nd digit -- what were labeled Major Object Class (MOC) codes when I worked with Treasury bureau budget shops, is where agencies will find disagreement. The 3 digit codes -- called Budget Object Class (BOC) because that was where funding buckets were established in bureau funds control scheme -- will likely yield a fair measure of agreement between agencies. I would propose that the government-wide strategy avoid (or defer) requiring reporting at the 4-digit level. This is where agencies become very proprietary in coding schemes - the Sub Object Class (SOC) in the hierarchy -- is granular to provide information relevant to analysis. Meshing 4-digit codes government-wide would be -- a very steep hill to climb.

I request that IAC and BOAC consider level of detail to be the key issue surrounding Object Class; there is no debate it is mandatory and we have high-level, consistent references per A-11. What requires clarification is whether reporting by object class will be accomplished at the 2, 3, or 4 digit level--Congress was silent on that aspect.

As we progress from 2 to 4 the level of difficulty in achieving comparability between agencies increases geometrically. I would recommend 2 digits initially, e.g. at 11, 12, 13, 21, 22, etc. and 3 digits, e.g. 11.1, 11.3, 11.5, 12.1, 12.2 etc. after successfully consolidating summary-level 2-digit outputs from the agency enterprise systems. Having worked for 5 different agencies - I can share that there is substantial comparability at 2 digit MOC and essential commonality at 3 digit BOC. Consistency falls apart at the 4 digit SOC tier of OC code hierarchy, where there can be dozens, even hundreds of codes and definitions that go along with code, some all numeric and some even alpha-numeric, e.g. 113B after running through 1131-1139 options.

We can agree on 2-digit

[HelenaSims]

AGA established a work group charged with responding to inquiries by the federal government, regarding implementation of the DATA Act. The following response is being submitted on behalf of AGA's work group, which is comprised of state, local, tribal and university practitioners from across the country. In commenting on specific data elements, our work group thought it important to indicate that - while the work group supports the standardization of definitions and reporting across federal agencies, - we are uncertain as to whether data standards will apply to entities receiving federal funds. Section5 (b)(7) of the DATA Act states that the Director has a year after submitting a report on the pilot program to "...issue guidance to the heads of Federal Agencies as to how the Government-wide financial data standards shall be applied to the information required to be reported by entities receiving Federal awards..." This provision makes it appear that, after the pilot program, some financial data standards may not be applied to information reported by entities receiving federal awards.

In addition, in order to clarify the reporting and posting responsibilities of recipients, and which data elements will be reported at the awardee and subawardee level, it would be helpful for the federal government to provide the data source for each data element. This information was included in the data model for FFATA reporting in Appendix C of the OMB Guidance for Federal Spending Transparency: Subaward and Executive Compensation Data Reporting Requirements, dated August 27, 2010.

Comment on Object Class
It would be difficult for entities receiving federal funds to consistently apply this definition. Section 4(b)(3) of the DATA Act requires that the data standards "include unique identifiers for Federal awards and entities receiving Federal awards that can be consistently applied Government-wide." Because this data element is only used for federal budget purposes - and is not a universal financial data category or classification required by any accounting (GASB/FASB), or other standards-setting body - it is unlikely that this data element could be applied consistently by non-federal entities.

[HerschelC]

The definition needs to be specific as to what level of detail is required. @MJB-NIH makes a practical comment regarding a lack of consistency in the lower levels of detail. I would agree that this could be a heavy lift to change today. But I would disagree with the premise that this data shouldn't be reported now because the deeper levels of detail aren't consistent across Government. Go ahead and publish the detailed, four level deep data with some understanding that on Day 1, the lowest levels may not be consistent across Government.

I don't think any reasonable person expects that on Day 1 all the data will be of 100% quality with 100% consistent definition across all stakeholders for all levels of detail. That's OK. It is fine if it isn't - so long as people understand the quality of the data that is presented, the fitness for purpose. By publishing all the data, different groups within Government can then see what other groups are doing. Analytics can be used to identify clusters of definition consensus as the Government works through iterations to improve the quality and consistency of the data at all levels. So when we go live on Day 1, Object Classes are consistent to the 2nd level of detail. Just be clear that they are not consistent to the 3rd and 4th levels. Having the data out there will facilitate the process of establishing this consistency.

Regarding @halsteinberg 's comment about a better system for reporting - the optimist in me agrees. If we can't make a complete rework, can we at least take a step in that direction? Also, "Other" categories are absolutely horrible to have in a reporting system. Whatever is in "Other" should be processed via the data governance process. If it truly doesn't fit into a category, then add a new category. This should be a fairly easy effort to do now - simply do some analysis on all the "Other" reported in the past and address via new categories. The use of "Other" should be a red flag and an exception condition requiring specific approval for use.

[MJB-NIH]

Object Class: Furnishing the entire body of object class data from the detailed (4- and 3- digit) to summary levels (2-digit) in a single push supported by@HerschelC -- with "caveat emptor" warnings regarding the cross-agency consistency at detail level -- has the advantage of complete transparency and publication simplicity. There is more variability than we would like to acknowledge. Even within agencies, the 2nd and 3rd tier sub-units that manage programs and code the obligating documents can possess differing interpretations of what 4-digit object class applies. Where @HerschelC sees opportunities for agencies to correct (partially) inconsistent data after examination, the public and stakeholders also have ample evidence that Federal object class accounting "accuracy" is flawed. The human tendency to make a single data point a trend should not be underestimated. Stakeholders would have particular anxiety and make the argument that: If it is flawed in one place, what is to prevent it from being flawed elsewhere? Will they understand that detail transaction records can be coded inconsistently yet those errors are not material once all the records are rolled up to a major object class level? My concern remains that many users and stakeholders will not respect the warnings given and find the glass "half empty" hampering our drive towards credibility. We will need large bolded fonts to underscore "quality of data" limitations.

[bsweger]

Thank you for your feedback on the data element object class. We will consider your feedback as we continually assess how we can improve data definitions. You may still contribute the discussion for other data elements. Check http://fedspendingtransparency.github.io/dataelements/ to find data elements with open feedback pages.