States are getting more comfortable with the reporting requirements for funds doled out from the American Recovery and Reinvestment Act, according to a recent report from the U.S. Government Accountability Office.
But last year, the states initially were thrown for a loop when the federal government asked them to report data in more detail than they had ever done before. States scrambled to improve their back-room financial systems to meet the demands.
Pennsylvania was fortunate to have a head start. By adding onto an existing SAP BusinessObjects suite it had previously acquired, the state was able to adapt to the stimulus by spending less than $2 million, according to Naomi Wyatt, the state's secretary of administration.
Since October 2009, Pennsylvania has submitted more than 1,000 reports to the federal government, and Wyatt believes the state was the first to file its stimulus data with the federal government last year.
Wyatt talked with Government Technology about how Pennsylvania was able to make it happen so quickly.
Last year, when the federal government first outlined what be the procedures for reporting data from the economic stimulus, a.k.a. the Recovery Act, how tough was it to assemble a workflow for those requirements?
It felt like we got hit by a tidal wave in February 2009. It was a great relief to all of the state budget officers, but in terms of the dollars, the flow and the reporting and all those things, it was a huge set of requirements that just sort of came out of nowhere. For at least the first month or two, we spent a great deal of time just trying to figure out exactly what would be coming our way in terms of: Would the dollars be flowing through the state government, or just go directly out into communities? How much would we have to report on? To what level? Would we just have to report what we gave somebody? Would we have to report them on behalf of that somebody? Do we have to report on behalf of the "somebody's somebody?" It was a real learning curve for everybody.
It was a lot of back and forth of the federal government in terms of what kind of data they were going to be asking from us to what level of detail. How specific would we have to be? How much would we be on the hook for the bad data that we got from other people? I remember being on conference call after conference call with lists of questions to figure out those parameters.
How did your state decide to handle it?
Pennsylvania determined pretty early on the set of tools that we were going to use, so we had a bit of an advantage. We instituted SAP in 2002, so we were already using that for our procurement, financials and budget. So we didn't have to set up an entirely new system just to deal with [the stimulus]. What we did have to do, though, is combine the system we already had with some legacy systems that we had in our individual agencies. For example, everybody uses SAP for finance. But our transportation agency also has its own project management tools, so that when it's doing a road project, it's tracking the dollars it's spending with each individual vendor and tracking the deadlines.
So you had to bring those systems together so you had a single view?
Right. Once we've started nailing down with the feds what kind of data elements we would need, some of those elements were well outside SAP and were in other systems. What we had to do was create a system that would pull
all of the data from the various subsystems that we had, and present it in one report to the feds in the way that they wanted it. For that we used another set of SAP tools from their BusinessObjects suite to do that work. And we were lucky because we purchased that set of tools in September 2008 for a totally different reason because we wanted to do better reporting out of SAP on the things that we were already doing. It just so happened then when we needed to do the federal reporting, we could use those tools to do that task.
We purchased the BusinessObjects suite in September 2008, and we really purchased it to get better reporting out of the SAP system we already had. So for example, we used SAP for HR. And to run reports about payroll or any of those things, we need sophisticated people who know how to do queries -- and the reports that come out are hard to read and they are not visual. So we purchased this tool to make that kind of day-to-day reporting that we do a lot more user friendly and a lot more acceptable by our managers. And then when this whole 1512 stuff kind of came, we're like, "Wow we can really use this tool to pull data from all these other kinds of system that we have, and get it out there in a very public facing way that's easy for people to see, understand," and that then has kind of spurred us to think about how we can get out more of the information we have in ways that citizens can respond to it.
So how much data does this system pull in for Pennsylvania?
We have 19 different state agencies that are reporting through our tool; we've got 3,500 vendors, grantees and subgrantees that are all filing data and information into our system that gets then sent to the feds. So it's a giant net that goes out and pulls all this information in. And I believe the state is responsible for somewhere in the realm of $12 billion that's flowing through the state -- the amount that we're responsible to report on.
We have an electronic war room with eight or 10 computers in it. About a week before the [stimulus] report is due, those 19 liaisons [one from each state agency] come to the room and start entering data. They do cleansing data there, they collect information from their subgrantees. We put them all in the same room because lots of questions come up. It was a lot easier to have them together as they are working through those issues instead of having them out in their various agencies.
Did Pennsylvania find any way to streamline that data collection process?
One of the things that we did to make sure the data that we were presenting was accurate is that we pre-populated as much of the information in the forms as possible before sending them out to our agencies and our subgrantees, which we were able to do because we had a lot of the information -- the entity's name, address, tax ID number, etc. -- in our key system or other systems. We would populate all of those cells so there was less chance of us getting back incomplete data that we would then have to verify and go back and forth a bunch of times. Because with a tax ID number, if somebody doesn't put the first two zeros in or whatever it may be, you just end up with junk data you have to spend time cleaning up.