Re-Imagining Government Procurement: A Manifesto
One thing upon which everyone seems to agree is that government procurement is chaotic.
Decades-old stories (see, for example, “The Pentagon Catalog”) of government over-payment for common products persist in the public mind:
“Other items offered in the catalogue include a $285 screwdriver, a $7,622 coffee maker, a $387 flat washer, a $469 wrench, a $214 flashlight, a $437 tape measure, a $2,228 monkey wrench, a $748 pair of duckbill pliers, a $74,165 aluminum ladder, a $659 ashtray and a $240 million airplane.”
Things have improved since the 1980s undoubtedly. But, still.
One of the first questions people ask me when I tell them that I am working on a startup related to changing the way governments purchase goods and services is, “Why can’t they just buy on Amazon like everyone else?”
What if we were to design government procurement with a clean sheet of paper? What lessons could we learn from such a thought experiment?
First, we would need to make explicit the objectives and constraints of the government procurement problem.
The function of government procurement should be to purchase the goods and services government agencies need to execute their missions in a timely fashion, obtaining the best value-for-the-money, while safeguarding the process against corruption, generally defined.
Sounds simple enough.
But, digging deeper into this, we see several choices and embedded assumptions.
Is procurement a task that should be separated from the functional execution of the agency’s mission?
How do we get the best value-for-the-money?
Is there an optimal way to divide a procurement into its components to obtain value?
Competition is a function of getting the largest number of relevant vendors to fight for the business. Is there any reason government should not have at least as much competition for its business when it comes to off-the-shelf products as large commercial buyers, like Fortune 100 companies? How do we assess the amount of competition? How does government know it is paying the right price?
What is the best way to minimize corruption? Is it better to develop rules and processes designed to anticipate and inhibit common types of malfeasance? Can we anticipate misbehavior? Or, would it be preferable to use leading edge analytic tools on fully transparent, granular reporting of transactions?
Explicitly, what are the trade-offs we make between speed, value, and compliance at every step in the chain?
What are the political exigencies for fairness?
The way to approach answering these questions is to develop principles of procurement for purchasing commercial off-the-shelf goods and services.
Here are mine.
Government has a duty to make it as easy as possible for companies of any size or in any location to compete to sell them goods and services.
The procurement process should not be so complicated that any potential vendor feels compelled to hire a consultant, or require third-party assistance, to help them navigate it.
All procurement communications should be written in plain language.
Wherever feasible, technology should be used by both vendors and government purchasing officers to improve the procurement process.
Procurement negotiations should be dynamic, not static; there should be no use of catalogs or schedules, for example.
Procurement procedures should generalize across agencies within the same level of government.
The procurement process should not distort the way in which government agencies decide to procure, or the goods and services government agencies elect to purchase.
Governments benefit from the consolidation of common procedures (e.g. vendor registration and qualification) or market investigations (e.g. RFP development).
There is a significant cost to delays in the delivery of government services.
Any action that elongates the procurement process must be justified quantitatively, or eliminated.
The government can only benefit from having more vendors responding to a specific procurement.
To attract the greatest number of vendors of commercial off-the-shelf products, the government cannot appear to favor any individual company ex ante.
Any government procurement process must be able to handle the relevant level of scale.
Traditional 20th-century reseller-distributors models constitute an unnecessary and unwarranted friction, unless there is no alternative.
It is never in the strategic interest of the government buyer to become reliant on an individual vendor or a small group of vendors for the procurement of commercial off-the-shelf products.
Every incremental rule constraining the procurement process loads the system with costs in terms of reducing the number of vendors contending for contracts, increasing the expense to vendors in responding to government solicitations, and delaying final delivery.
The government ends up paying all costs of any procurement process, as vendors either embed these in their bids, or decline to vie for government business.
For the procurement of commercial off-the-shelf products, any government agency should aim to achieve an outcome that is at least as good as those obtained by a commensurately sized private entity.
Individual agencies can only benefit from knowing what other agencies have paid in the past for similar goods and services that they seek to procure.
Individual agencies can only benefit from collaborating with other agencies looking to purchase contemporaneously the same goods and services they seek to procure.
Corruption cannot be eliminated, but it can be controlled and discouraged intensively.
Transparency and openness, not rules, are the principal tools to inhibit corruption.
If government procurement activity is to be used as a means of stimulus, either regionally or industrially, then such designated purchasing should go through a separate spending vehicle explicitly. Such spending can pollute the integrity of the procurement process, alienating vendors.
The more information government buyers have at their disposal about the past behavior and reputation of vendors, the better their procurement decisions will be.
With these axioms in mind, what would our idealized procurement system look like?
It would look remarkably like a futures exchange.
Membership: The ideal procurement system would consolidate common functions across and within levels of government, such as vendor registration and qualification.
Price Tape: A database showing what any buyer on the platform paid in a transaction would act like a tape for all to see. The more buyers, regardless of their level of government or nationality, the merrier.
Liquidity: The ability for buyers to develop ad hoc transaction-specific buying consortia to get improved volume pricing, particularly on commodities, is akin to the way liquidity builds on an exchange.
Analytics: Just as the Commodity Futures Trading Commission analyzes the database of transactions in futures exchanges for aberrant behavior, procurement officers here would function more as auditors, using data science to identify corruption. The City of Toronto used this approach to catch bid-rigging in city paving jobs.
Standardization: By making all RFP activity on the platform transparent, buyers can develop much of their market intelligence for a purchase by relying on the activity of other buyers. This cuts the time-to-market for the RFP, shaving months potentially from the process. Ideally, the process reduces to a hasty auction, conducted in days not months, with dynamic pricing refreshed every time there is a transaction. There is no catalog with stale pricing and some arbitrary discount to inflated list prices.
No Appeal: For the procurement of commercial off-the-shelf products using a standardized approach, with full transparency, vendors waive their right to appeal. The procurement auditing processes will catch any impropriety; losing vendors can flag deals for accelerated audit (but without holding up the procurement). This reduces the risk of executing the procurement and cuts the time to delivery potentially significantly.
Member-Paid Fees: Vendors pay the bulk of the fees, including a membership fee and a reasonable stamp fee for submitting responses to bid solicitations. They either pass these costs along to government, or lower the margin they are willing to tolerate to win government business.
Reputation Risk Assessment: Members can see information about the historical performance of individual vendors, as assessed by previous buyers. Think “Yelp-for-Government-Procurement”. Only vendors with sterling reputations end up as members, in a virtuous cycle.
On the face of it, this is what Section 801 of the National Defense Authorization Act for Fiscal Year 2018 does. It is an exciting experiment that reaches beyond the Department of Defense to span the entire federal government.
“This section would require the General Services Administration (GSA) to contract with multiple commercial online marketplaces for the procurement of certain commercial off-the-shelf (COTS) products.”
The proposal also promotes the collection and analysis of data which would provide better “transparency into the Federal Government’s purchasing and thereby enable more thorough oversight and accountability.”
Here is Congressman Mac Thornberry (R-TX), Chairman of the House Armed Services Committee, writing about his proposed markup of the National Defense Authorization Act:
“A recent GSA [General Service Administration] report revealed that for IT products alone, GSA rates were 13 percent higher than those on the open market. The bottom line is that the Department is rarely able to buy off-the-shelf items quickly and at a reasonable price.
“A solution is obvious to most consumers — allow the government to use online commercial sites like Amazon, Staples, or Grainger just as businesses do. These portals function like mini-marketplaces, ensuring that the buyer gets the best price without a lot of red tape.”
Note that he asks the same question as everyone else, “Why can’t the government just use Amazon like everyone else?”
It is a bold start, but there are some issues for the implementation that require further investigation. The Coalition for Government Procurement (the “CGP”) has written an opinion highlighting some of the issues to consider when implementing what they call one of “the most consequential procurement policy changes in a generation.” Keep in mind that these are the issues raised by an organization advocating for “commercial service and product companies selling in the federal market.”
First, the language in Section 801 refers to marketplaces, yet the Coalition worries that the GSA will end up selecting platforms, specifically platforms that have so much market power as to dominate the economics of any transaction, effectively subordinating sellers and discouraging sellers from participating.
“’Marketplace’ implies an open, neutral exchange between buyers and sellers. Current online service providers sought by Section 801 are really online platforms, not online marketplaces. These online platforms set the terms of participation for buyers. Moreover, it is our understanding that current online service providers can and do use the transaction data from the platform to compete directly against the suppliers they are servicing.”
Second, the CGP is concerned that the GSA, in implementing Section 801, will merely comply with the requirement to have a minimum of two portals, instead of seeking more platforms, referring to the language regarding commercial item procurement in the Federal Acquisition Streamlining Act.
Third, the CGP decries the language stipulating that “the standard terms and conditions of the online provider cannot be altered by the federal government.”
Some of the CGP language is certainly overwrought when it comes to online marketplaces because they have made the leap to conclude based on public commentary that the GSA will a) only select two vendors, and b) that these vendors will be platforms, not marketplaces.
In early July, the Commissioner of the Federal Acquisition Service, Alan Thomas testified before the House Subcommittees on Government Operations and Information Technology. Responding to the frustrations of one Member that the government had not used a commercial online platform, Thomas said:
“I think the bill, the NDAA bill, goes some ways towards addressing that. There are, as you know, some specific regulatory and policy concerns that the Federal government has that those of us who just buy as private citizens through Amazon don’t necessarily have to take into account. So, you want to make sure that those are accounted for in the appropriate way. But, I do think introducing commercial marketplaces, like that, into the government buying process should yield some savings and speed and offer some simplification.”
So, you’re telling me there’s a chance.
In terms of my framework, using Amazon or other platforms violates several axioms.
These large platforms potentially scare off competition from other vendors, especially smaller ones; they may also seize all the value in the transaction, with the concomitant policy implications for business development. (This is really what the CGP worries about, I suspect.)
There is no consolidation of common procedures.
The government may become overly reliant on a handful of vendors.
Can these vendors handle the scale of government procurement?
Some of these platforms are more like resellers or distributors and represent unnecessary transaction-level friction.
There is nothing here to promote inter-agency collaboration or information sharing; indeed, government buyers may become vulnerable to price discrimination over time.
But, don’t listen to me. The CGP points to the inconvenience of statute with respect to using Amazon to satisfy Section 801.
“Absent a rationale [sic] policy explanation, the statutory requirement for full and open competition should not be waived for the award of a platform contract …
“Organizational conflict of interest issues need to be understood and addressed. The proposed language does not prohibit the commercial online platform provider from selling its products in direct competition with the firms selling through the online platform …
“The impact of this platform model on small and medium-size[sic] businesses needs to be considered …
“If GSA is to administer the program, it must be given the authority to do so …
“Data security needs to be understood and addressed.”
For an even more detailed analysis of Section 801, check out an excellent article on The National Law Review, “Change is Upon Us: An Analysis of the Section 801 COTS Provisions of the 2018 NDAA,” by Jonathan Aronie.
In conclusion, there is a massive gap between a literal, straight-line interpretation of Section 801 and the position of incumbent players.
At EdgeworthBox, we believe that the solution we are building can bridge that gap.
We’re excited to be developing a transparent, technology-enabled marketplace and we look forward to telling you more about it as our development efforts proceed.