Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable. ======================================================================== # Standardization and Expansion of Employment and Wage Records Authors: Erica L. Groshen, Julie Hatch, Chris Kurz, Frank Fiorille Discussant: None Video: https://www.youtube.com/watch?v=aXOe_NS3iVo&t=1065s ## Talk (00:17:45 – 01:20:18) [00:17:45] panel. Uh and if the panelists could come up and take seats at this uh at this table, uh I will u the panel I I don't think that the topic of the panel needs much introduction. Uh the [00:17:59] whole all of yesterday afternoon was a great introduction to the issues that are taken on. David Wilcox posed it pretty pretty starkly after the paper on the British experience. What would it [00:18:11] take to get uh to get to [music] to get here in in the US? And we're delighted to have uh multiple experts on on this topic who are working towards improving the statistical system in creative and [00:18:25] imaginative ways within the context of the US institutional arrangements. So I I I'm going to introduce the panel at at the start, then we'll just have short presentations and I hope there's a bunch [00:18:38] of back and forth among the panel and uh and uh and and the audience. So Erica Gian, former commissioner of uh labor statistics and now at Cornell University, Julie Hatch, currently at uh [00:18:51] currently at the Bureau of Labor Statistics, uh Chris Kurtz, Federal Reserve Board, and Mike Foli, uh paychecks. So this I'm sorry, Frank Feroli. [00:19:06] H sorry um uh the uh so why why don't we just take it in in in that in that order uh and have short presentations. Some of you might have slides, some may not. And [00:19:20] uh and I will uh chime in as as needed. >> You were gonna go first, right? Or >> Julie. Okay. Yes. Okay. We're gonna go have Julie. [00:19:31] Is this on? Okay, perfect. Um, so I don't have slides and partly that was because I did actually have notes walking into yesterday and then as people were talking I was furiously writing notes and scratching things out. [00:19:44] So, um, I'm pivoting a slight bit, but I do want to take a time to thank the conference organizers. I think this really is just a testament to how important these issues are and how many smart people are working on them from so [00:19:58] many different angles and I if it was easy we've already done it right like [laughter] these are complicated issues and that is in part because the US labor market is this huge growing dynamic organism that once you think you have [00:20:13] your arms around it it changes on you you know it it's like we didn't see co coming and that like was a whammy we didn't necessarily, you know, we heard about AI and the acceleration and then, you know, we think we had one story and then six months later we're like, well, [00:20:27] maybe it was telework, maybe it wasn't. So, it's it's, you know, we're really trying to um do the best we can with the data available. And to your to the point we've heard before, really push the frontier forward as much as possible. [00:20:41] Um, you've also seen I've worked with some really incredibly smart people and um, they're really what got me through the last year. So, thank you for that. [00:20:52] Um, there is a huge world beyond traditional survey data. And just to echo what the commissioner just said, survey data is here to stay. There are some things we simply can't get without it, but there's a huge amount of information we can get from other [00:21:07] sources. You've heard tons about QCW. Um, and if my QCW staff were here, they would have like a little heart because it really is the backbone of a lot of what BLS does. [00:21:19] And not only what BLS does, but a lot of other agencies. Uh, to your point about sharing C or BLS shares QCW extensively. [00:21:28] Uh, the Census Bureau, every quarter we ship that over. We share it with the uh, energy information uh, agency. Uh, we share it with BEA. So, so it is used quite extensively in addition to the own states itself, right? So, this is such a [00:21:42] rich dynamic data source and even though um BLS does fund the states to do a lot of the data processing and data quality, it it really is for the betterment of all of uh all of its fiscal agencies and [00:21:56] people who aren't even in sats just using uh data from that uh there. [00:22:02] And as much as I love QCW, I do want to take a pause and remind people it also revises. [00:22:09] The difference between the CES preliminary and final benchmark is because of QCW in the revision. [00:22:18] QCW also has to impute data. I know this might come [laughter] as a surprise, but sometime you know people say we're reporting survey data isn't on the top of business list. also reporting administer of tax data in UI insurance [00:22:31] isn't also uh necessarily happening uh on a timely basis in QCW land that imputation is very small uh it has grown a little bit and we've gotten a lot more sophisticated uh but when you're talking [00:22:44] about threading the needle that is CES that's huge um CES really is threading that needle there are millions of jobs added and lost every month and we are trying to report on that net change that [00:22:58] makes it reasonable for you guys and everyone else to understand. For instance, local education in July lays off over a million workers. Well, what would you do with that? You can't I mean, I don't know. Is that a lot? Is that a little? [laughter] [00:23:12] And that's where the seasonal adjustment comes in into play. But it's just this huge dynamic organism. [00:23:19] It also is under a lot of strain. Um I'm not even going to talk about budgets. [00:23:25] That's not what this conversation is. But for those of you who know, UI modernization is no joke. And oftent times the LMI, the people who we deal with are at the very end of the chain. [00:23:39] And if they don't have a good relationship with their UI office, they are the afterthought. And it has gotten very close sometimes where and in fact in one state we had to go public with it. Unless you give us the deliverables, [00:23:52] you're not only losing QCW, you're losing your CES, you're losing your national or your state and local unemployment rates, and you're losing your OEWS. And it only was when we had to actually pull that trigger and say, you know, I'm going to the governor's [00:24:06] office because he's not going to be thankful that he won't have a state unemployment rate to talk about, did we actually get invited to those conversations. [00:24:16] That takes up a lot of time and resources. It's also not uncommon uncommon they go live and all of a sudden they changed every UI account number without a crosswalk. [00:24:26] That that that's a little problematic. And if you think of, you know, why are there some some delays and revisions in in QCW, it's because they're just simply not adding up what people are giving them. There there is so much cleaning [00:24:40] that happens for people who deal with administrative data. It is messy. you know it's messy especially because it's you know not designed for statistical purposes. So so much time and energy uh is there to get it to the quality that [00:24:54] it is. Another recent phenomena remote work. When you talk about QCW it's at the state level. If you're an establishment and you have a remote worker in another state guess what you [00:25:07] get to open up a UI account there. Then all of a sudden the state LMI shops are trying to figure out the nakes code. [00:25:14] Where should we code them? We really don't want people's private information like their addresses as part of QCW. [00:25:20] That's lost to a lot of conceptual issues. Some states saw a lot of a boost of remote workers when COVID was hitting. Sometimes your state QCW staff were assigned to help with uh administer UI because that quite honestly was much [00:25:34] more important making sure people were getting their UI claims than than QCW. [00:25:40] I love it. There's a lot going on. Um, oh, and then you know, we also have the annual refiling survey and the multiple work site. So, there's lots going happen with QCW, [00:25:52] but there's so much energy and excitement about another part of the QCW world, the wage records and the enhanced wage records. And I honestly am I I think this is pivotal. Uh, obviously [00:26:06] wage records have been around for a while. That's what the whole LEHD is built upon. Um, but where this movement and energy and excitement is about the enhancement of those individual wage [00:26:20] records, where they're adding job titles, they're adding location, they're adding hours, and right now it's up to individual states to sort of chart it alone and BLS is partnering with those states. But there's also a lot of energy [00:26:33] and a lot of um stakeholders and people like Julia Pollock over there who sees the value. uh do uh chief economist sees the value and the potential of this [00:26:44] data. So having those uh cheerleaders, those stakeholders is really moving the path forward. Henry Mack, before he uh left, he was the uh ETA uh assistant secretary, I believe, who his official [00:26:58] title was. He's now back in Florida, but he was the first ETA person that I had ever heard talk voluntarily about enhanced wage records. like you could have pushed me over with a feather when he went up onto a national stage and [00:27:11] talked about enhanced wage records. And that momentum is growing. In fact, um I know I was in Milwaukee earlier this week. I know some other people were and at the the NASWA conference, the National Association of State Workforce Agencies and their planetary session was [00:27:26] about enhanced wage records. So this is a huge conference. It's not just the LMI folks, but it's a lot of the it's mostly actually for the UI folks, the people who are administrating and the they were everyone is in that room talking about [00:27:40] the benefits of enhancing wage records and really where we're seeing a lot of steam and to my colleagues over um from ONS, we're getting the traction because we're tying it back to the fundamental core of unemployment insurance and we're [00:27:55] tying it back to how it can improve the um benefits and how it can make that more efficient and then how we can help with fraud detection. Not you know me but how it enhanced wage records can help with fraud detection. So we're really getting the stakeholders who [00:28:09] ultimately is for interested in this too. So while we are definitely a beneficiary really where the traction is coming from are people who they can see okay this isn't just for you sort of the data nerds this is something that can actually help us administer the program [00:28:24] more effectively. So so much is going on there. Uh we have a couple different partners with our couple different uh projects with our state partners specifically looking at enhanced wage records. And this is why I'm excited. [00:28:38] Jolts job opening and labor turnover. We have this wonderful project we've had for a couple years with census looking at the LEHD and then comparing it to Jolts. And it is very obvious uh when you look at it that we're catching the same trends in Jolts, but the level it [00:28:53] they're just a different magnitude. And this has really helped us see like poke at it a little more. Now, some of that is conceptual. So, I'm not saying Jolts data is no, but I'm I I do see there's value where we can improve using if we [00:29:06] had access to administrative data, those wage records more timely, we could improve Jolts both at the national level, and this is really where I see it. We had to pause doing monthly state Jolts data. There's a value from that. [00:29:21] Labor is local. You know, we focus so much at the national level. It's really at the local level. So, the more we can push down those meaningful statistics that are current at the local level is really where you're going to get the bang for the buck. And as more states are joining the BLS wage records [00:29:36] program, I think we're almost up to 41. This is sort of what I'm telling them. [00:29:40] You give me wage records, I am seeing there's benefits for you that I can give you state jolts data that's a heck of a lot more quality than what we were giving you before. for the enhancement that are adding job titles. SOCK is less the standard occupation classification [00:29:54] system. I'm less thrilled about that, but give me some occupation information. [00:30:00] BLS has the huge OEWS occupation employment wastist program. If we can make that a hybrid system or more like QCW where we're still funding the states but instead of collecting survey data now qual or they're cleaning up [00:30:13] administrative data a much more data and if we get it quarterly as someone said earlier OEWS is not a time series but if we get that quarterly then you can start actually making that type of analysis on [00:30:27] a you know on a time series basis game changer. Some of them are adding hours worked, hours paid, productivity. They, you know, torture the data. They take CES and they look at CPS and figure out the differences and try to make that [00:30:42] meaningful distinction. If at the worker level they could get actually additional information about hours huge. You know, I don't work for productivity. I can say something there. Um, CES I think in terms of and speaking of hours and earnings, yeah, there's employment and [00:30:56] that's what a lot of energy is focused on. Our hours and earnings series have really taken a brunt of the hit. Uh you can tell that where we have two different aggregation structures and the one for hours and earnings is a much higher level and that's simply because [00:31:09] of data bargaining. CES, we're not asking you just to give us data one month. We're asking you to give us data three to five years. And quite honestly, if you're a certainty unit, we're never going to come down at your door. And occasionally we've had to data bargain [00:31:24] away especially if someone uh provides data through what we call our EDI center electronic data interchange that's where computers talk to computers very efficient if they report to the um multiple wage report uh we basically [00:31:37] sometimes will have to say well give us what you would give to the MWR just monthly but the MWR is employment and wages it's not ours so we can't use it for CES like that so a lot of there and you saw yesterday that subnational data [00:31:51] um picking up local events. Um not to mention other new products that we haven't even thought about yet and there's a lot of energy. I know AI is like the new buzzword. Um and we're doing a lot in that field as well. Uh [00:32:04] and for the private payroll data, we are we are so close to actually getting our hands on uh some of it. So I'm super duper excited. The delay right now is on Julie thinking that the hard part was doing the paperwork and I did not [00:32:18] appreciate the it. We're working on the IT, we're very close. Um, and that's going to let us start playing around and people like Steve Mance and his staff to really just see where can we get some of that utility um for you saw the the [00:32:31] chart um before where we're capturing, you know, the the localized events quicker. That's huge, especially if you're in those local communities. [00:32:39] That's what you want to know about. Um so I do see improvements to the whole ecosystem coming up. Um and it is at a very exciting time. Uh, so I think that's good. Lots of questions. [00:32:52] Okay, I've got I have some slides. [laughter] So while these are being pulled up, um [00:33:08] thank you to the organ to NBER to the organizers to NSF uh for organizing the conference and uh inviting me to speak. [00:33:18] Julie was talking about uh just how enhanced wage records in particular can um the impact of access on those to the statistical system. I'm [00:33:32] going to talk about the earlier stage in the production process. how it is that the administrative data itself can be improved because we've talked a lot about the problems and Julie talked [00:33:46] about all of uh all of the problems of actually cleaning the data and uh but what what also needs to be addressed is [00:34:01] why we need data standards to make sure that the data are actually better to use more fit for use and how to get those data standards in place. So, [00:34:14] so let me go there. So, why data standards? Why am I focused on this? [00:34:19] Well, all of what we've been talking about here is part of entering the information age and really getting the benefits of that. And this is the time when um when information becomes a [00:34:32] commodity that we use more quickly and more widely than ever before. [00:34:40] It um this administrative data though has limitations that we've been talking about. [00:34:48] And when you have data standards, you essentially make administrative data a little bit more like the surveys that we've come to know and love, right? It makes the data much more when you have [00:35:00] data technology standards then the data become interoperable more consistent more mergeable more designed for measurement purposes as well as for the transactions or activity that they were [00:35:14] um that they're originally created for. And that gives us benefits that uh that are very important lowering the burden for respondents and and for collectors and also provides the [00:35:29] opportunity to build in privacy protections from the very beginning. So data standards have to be part of entering this new world of using administrative data better and more. So [00:35:42] let's talk about wage records and the role that they play. our current workforce administrative data system. [00:35:49] And this is no reflection on the BLS, it's a mess, right? Because it wasn't designed for scratch um from scratch the way we would design it now. So, this is [00:36:04] a schematic showing you what it is. You've got employers, they are keeping their own administrative records, right? [00:36:10] And then they have a lot of reporting that they have to do to the federal government. [00:36:16] uh you know SEC, EEOC, UI, new hire data set, BLS, Sens blah blah blah. If you're a very large employer, you're talking about something like 175 reports that [00:36:28] you have to fill in, many containing same basic information, right? Uh the contents are similar but not the same. And you can think of these as mostly variants of the UI worker and [00:36:43] employer records. the employer records where the QCW comes from and the UI wage records. The UI wage records right now are really sparse. They don't have much information in them. They've got employer, they've got worker ID, [00:36:57] industry, quarterly earnings, time period. That's about it. And BLS has not had access to these, right? The Bureau of Labor Statistics has not had access to these. So where do we go? Well, in [00:37:11] context, think now. Let's go to ven diagram type ar um view of this. We have these employer administrative records. [00:37:20] Some part of them are being requested on a regular basis and they're all asking for some different combination of what these records look like. The then there are the core records in the in the middle. [00:37:33] We can go to a world where you have data standards that basically span most of these requests and that will make the reporting better and more consistent. Right? So that's [00:37:47] where we can go. Uh and there is an initiative for this uh the jobs and employment data exchange which is a public private project sponsored by the US chamber of commerce [00:38:01] foundation. So it is an an employer uh industry initiated effort to create these data standards and one of the leaders uh of this project Bob Sheets is [00:38:13] over there Jason was here earlier today yesterday um they're doing this uh by working with BLS census the payroll companies state [00:38:26] LMI offices philanthropies large corporations so they've they've got a lot of stakeholders at the table and the goals are to improve reporting starting with state UI records and this has three [00:38:39] parts data enhancement and standardization developing modern reporting systems like APIs, autocoding, privacy protections and then consolidating government [00:38:52] program and statistical reporting. And the idea is to provide more timely granular data for all of the purposes that uh that we understand and also providing better privacy protections [00:39:04] lowering burden and more and uh helping to promote a more efficient labor market. [00:39:11] So then we go from the old spaghetti view, all right, where we had the employers, all these different collectors, all the different users to one where the employers are reporting the same package of information for many [00:39:26] for multiple uses and the users can get what they need and what they're allowed to get on the basis of permissions. [00:39:35] So what's the progress so far? Jedex has already done the benchmark and standardization work. They've got a data dictionary. Uh they've used input from the statistical agencies, from employers [00:39:48] and from researchers on those data fields. They have a data package rec uh ready for pilot testing. They've put together the data collection tools and now they're pilot testing the pipelines [00:40:00] with simulated data in partnership with HR and the payroll companies with BLS and with two states and their next steps are refining these tools producing use cases and starting to release the [00:40:14] information to get uh companies to start uh standardizing. [00:40:21] So that's in the works. Combined with all the necessary work that the statistical agencies are doing, there's really a platform here for progress. The next question is whether the statistical [00:40:33] system as a whole is ready to be able to to make the changes and to evolve to be able to use this information. And there we have problems because to modernize [00:40:47] the statistical system as a whole has to become much more interoperable. And that's the other thing I want to mention. The American Statistical Association now has a roadmap for this modernization. [00:41:00] And it is a roadmap of 30 practical steps that the Congress and the administration can and should take to allow this modernization. So, it meets the system challenges that we've been [00:41:13] talking about today and yesterday. Declining response rates, aging workforce systems, fragmented data systems, and all of the barriers to combining and using modern data sources. [00:41:26] This modernization uh means that we've got to go to a world with more economies of scale and scope. [00:41:34] Um, we got to have more ways and fewer barriers to sharing data. [00:41:41] So, 30 recommendations here are the here are the the titles of all those recommendations and I've highlighted all of the ones that would need to be worked on to really implement enhanced wage [00:41:55] records and take advantage of them within the system. So, uh, let me just close by saying that, um, I encourage you to learn more about all of these initiatives and Jed and about Jed X. You [00:42:10] can talk to Bob and me about it. Um, and welcome any questions that you might have now and later about all of these opportunities. Thank you. [00:42:22] [applause] Okay. [00:42:38] So while my slides get put up, I just want to thank the organizers and uh everybody for attending. Uh also NBR for putting these together. Um I'm going to be sort of presenting a view of [00:42:51] individuals who've worked with uh private payroll data and uh try to kind of present some stylized facts and a little bit of the direction I think we need to move forward. Um, I'm also calling it I'll go full Bertold direct [00:43:05] here and say or the view of one of the guys in the garage. Um, because I like to look at what we did with the ADP data. Um, the private payroll data back, you know, starting about 10 years ago is just that it was really just five of us [00:43:19] working on this project. I will not I will not do the usher and wander around working working in a garage and and trying to solve some of these problems. And I think one of the important takeaways would be that um we [00:43:33] actually made it work. Um and it's also important to know that this is not the views of the Federal Reserve Board and it's really just one of the chaps in the garage presenting this. I got some feedback from the others, but uh uh they also wanted me to just talk about firms [00:43:46] and establishments and I'm going to talk about payroll data. Um so so I think I'm going to kind of go through a couple of things that we know and and I want to importantly point out that change is needed now. I think we're all kind of in that same space headsp space. Uh one [00:43:59] thing to think about is that uh why did the Fed start doing this you know more than uh 12 13 years ago and uh so what what I'm going to ask you is why did we create what we call the expanded measurement agenda and that was really [00:44:12] the Fed's attempt to leverage new data leverage new resources leverage new methodology and the uh father grandfather of this idea is here David Wilcox and it was really us trying to move forward in a world but why did we do that was it the case that we were [00:44:26] thinking that there was a problem with the statistical infrastruure structure at that time. Was it the case that we just had new techniques in our pockets at that time that we could leverage or was it the case that there was just new data to take advantage of? So, a lot of these questions I think hinge on the availability of this information before [00:44:41] you get to these ideas and questions about like low response rates. Um, since we're not going to be using alternative data, I'm going to call it PADD, which is private sector uh administrative data for payroll. uh and we've proven that it [00:44:55] is actually extremely useful for building labor market statistics. Um it's also important to note that you know this kind of idea that there's an exodus from sort of the statistical architecture that we're looking at has already started in some ways. uh if you [00:45:10] ask individuals who work sort of on an investment banks and hedge funds or Wall Street or if you ask individuals who are doing forecasting what fraction of their signal is being extracted from publicly available statistics as opposed to things they're getting from private [00:45:24] sector data that fraction has changed so much over time particularly over the past 20 years um an important thing to also keep in mind is that um you know there's a portion of the QCW that is already submitted by private Riv payroll [00:45:39] providers question is what is that portion? It's close to 40%. So when we're thinking about leveraging private payroll information for the CES, we should realize we're already doing that in an important part of our our statistics already. What are the [00:45:53] complications? We've been talking about these over the past day. Uh but I will kind of make the strong stand that without the micro data all is lost. [00:46:02] Being able to look at this information, digest its information, you need that micro data. It also is important to know that without you know the loss of a single data vendor shouldn't scuttle these plans. Um importantly this process [00:46:15] is iterative. I think people don't understand that kind of the combination of the micro data with this process is really important. You see the BLS responding to changes in the economy changing their methodology. The same is true for for our leveraging uh uh PADD [00:46:30] in in a way that like we had to change methodology when that methodology was being called into question due to circumstances outside of our control. Uh so flexible methodology for flexible people. Um and you definitely need a [00:46:44] birth death framework in this environment. You can't just look at a continuous sample or you're not going to have a true representation of what's happening in the economy. We know that a lot of the growth in in in employment is coming from those sort of new establishments and if those new [00:46:58] establishments are excluded, you're going to be in trouble. Uh you need sort of the dates of those business creation and you need some industry information. [00:47:06] But I would say that one of the interesting things we saw yesterday in the presentation of the uh the information on on a private payroll provider is that if you look at those trends, you'll definitely see something particularly over the past few years [00:47:19] where it diverges from the QCW. And I can tell you that that shows up in alter other data sources. So we continue to see that there's sort of a downward trend in a continuous sample and unless you fold in birth death, you're not going to be able to overcome that and tell what's truly happening in the [00:47:34] economy. So um and I kind of leave you with maybe what we'll call a rhetorical question but you know if if we're saying change is needed now we have to ask ourselves what would be the impact if policymakers move forward on their own without working with the statistical [00:47:47] agencies and we're in a situation where people ask that question. Um but it it definitely should be the case that we're kind of moving forward together. So just a couple of points here and kind of on the perspective of where I'm coming from is that the Fed, you know, the Federal [00:48:02] Reserve is not just a a consumer uh of of data. It's a producer of data. We have a long history and and Ryan mentioned this yesterday of producing statistics over 100 years of producing industrial production, produce [00:48:16] information on exchange rates, produce information on consumer credit. Uh so that expertise is there. uh and we also importantly have a long history of involvement in economic measurement. If it's not QCW, it's talking about price mismeasurement. It's talking about [00:48:30] understanding productivity and um and those things I think are our tool or sort of a group of information researchers that should be leveraged. Uh and and the hill that I will I will die [00:48:42] on is that policym and research is not they're not separate things. You can't say that we're just going to consume data for policymaking purposes without having the research to support and understand that. And one way of thinking that is to look at all the papers that [00:48:56] Ryan cited yesterday that went into the creation of our indices that leverage private payroll data. It took years of work. It took years of presenting at conferences. It took lots of feedback from our academic partners, from the [00:49:10] from say CRAW, from different conferences to actually come to a point where that was usable. So if you're going to say you can only use private sector data for the polic you know for policym you're actually missing an important component which is we need to [00:49:24] have that iterative process with the academic circles with presenting research uh and this is even more important when you think about this in the context of economic measurement so kind of what should be done where should we be going well I'm thinking absolutely [00:49:38] necessary for us to have the micro data if it's not just to be able to see what an outlier is doing to your output it's important to be able to just go in there and understand what those like large changes are doing. I mean, we've seen, you know, organizations because it's [00:49:53] tough to say whether or not they're a business or a firm change employment by a couple, you know, a couple thousand. [00:49:59] And if they have a large weight or if they're a small organization, that could have a huge impact on your actual measure for a particular month. So, being able to understand that is is going to require looking at the micro data. Another thing that we probably don't appreciate enough, but some people [00:50:14] here definitely understand is that, you know, it's you can come to different answers with the same data with different people putting those statistics together. And I think John has seen some of this when they put together the DISP statistics is that you can have a team from census, you can [00:50:28] have a team from the BLS, but use the same underlying data and come to two different measures of productivity. um you know when we were when we're looking at sort of the underlying micro data or we're looking at the the employment data you can have linked employee employer [00:50:42] data a panel over time and you can have extracts that just look at what the business is doing and you cannot come to the same employment measures per month. [00:50:50] So so there's this idea that you know once we all have some data we're going to come to the same conclusion but that's not necessarily true because the methodology the importance way to put on different aspects of that can lead to vastly different results. Another thing [00:51:04] that's important when you think about the micro data is that just being able to make comparisons with administrative data that exists, let's say the QCW is going to be super important to do to understand what your actual sample frame [00:51:15] is. So um I think we can easily say that uh the payroll provider data will complement and supplement the CES over time just like it does with the QCW. So what do I mean by that is that we already know that this private payroll [00:51:28] data being rolled into the QCW. we need to sort of switch that into the CES world and and the way I think about this is that we need these private payroll providers to work with the BLS in providing that information. So, how how [00:51:41] would I kind of design this if I could is uh it would be very similar to what we're seeing with uh the QCW and this would kind of perfectly tie in what with Erica's presenting with with Jed X and discussing is that essentially you [00:51:55] should be able to have private payroll providers giving a click button that allows that information to be submitted to the BLS. Um and Jedex could work through that, you know, how we actually contextualize that and maybe they're already in this position. Uh but [00:52:09] essentially just like you see individuals presenting information to the QCW through private payroll providers, the CES could do the same thing and all you need is a couple private payroll providers to move in that direction to make this possible. Um you know you have to be very specific [00:52:24] about NAKES codes and industries. You have to be importantly collecting information on birth debts, different types of payments because we want to make sure it's not just bonuses or corrections for paying for lunch that are being counted as payroll events. And [snorts] then you have to also [00:52:37] understand kind of beneath all of this there's large benefits for those payroll providers because they will be able to say better understand the industry that these individuals operated better understand the size class and offer services to them that might be a beneficial to both sort of the [00:52:52] collection of data but to those organizations and they would be able to save time for their clients. [00:52:59] So I would say lastly and maybe in my mind most importantly is that you really kind of want to let uh the Federal Reserve in this sandbox is as we're collecting information um and as we're kind of synthesizing that for the [00:53:12] statistical agencies you have to think that kind of the resources the desire and the precedent with working with these statistical agencies [clears throat] is already there and we would like to be in that sandbox but we can actually produce information that benefits. It's kind of like working [00:53:26] within the RDC produce information and benefits for these organizations uh maybe on a different tenor a different schedule because policym works on a different schedule you know it's not necessarily trying to make a statistical release but it's better understanding [00:53:39] that information so we can extract a signal about the economy uh and it's going to allow for more projects that are kind of more focused they're more policy related and and be able to answer sort of measurement questions that could be happening in the background is that as these agencies and organizations try [00:53:54] to meet a kind of a statistical deadline. You have some individuals that are more research focused in the background that could be able to point out measurement issues and make improvements that could be beneficial to all. That's it. Thank you. [00:54:08] [applause] Okay, can you guys hear me? Okay. Um, I'm going to go quick here because I think we're running a little late and I want to make sure that there's time for for questions. And um, so again, my name [00:54:23] is uh, Frank Fierelli. I'm a vice president of risk and compliance at Paychecks. For those who don't know, Paychecks is a big human capital management payroll company. Um I think the last I looked, we're like number 77th on the NASDAQ 100 as far as market [00:54:37] cap. Um so what I'm going to really talk about is is sort of the private data and specifically kind of tell you a little bit about um our small business index that we we built um about 12 years ago [00:54:50] um and have been reporting on. If you if you ever go on CNBC in the morning uh during jobs week on Tuesday morning, typically our CEO will go on and talk about our our results and you know it's it's like a barometer of what's going on [00:55:04] on Main Street. It's a um it's an index that really looks at just kind of the mom and pop small businesses under 50. [00:55:12] And um you know again we we've been around for 50 years and we've got a lot of really good data and um we think it's a it's one of the best uh data samples of of really what's happening in employment. And the reason why I say [00:55:26] that there's there's a number of reasons but you know we we are first it's like the front line right we we come out Tuesday of jobs week right before um you know ADB on Wednesday and the Friday non-farm payroll numbers. So, ours is [00:55:40] really sort of the first one that comes out on that Tuesday. Um, the other thing is, um, and we've been doing this for a long time, ours is is real payroll data. [00:55:50] This is not a sample. This is coming right off payroll journals. Um, so there's no revising it. So, we don't go back and when we come out with our number, our index number, there's no revising it because it's real payroll [00:56:02] actual data. Um, the other difference I think is we we have a really unique way. [00:56:08] we really cut off really quickly so that it's really fresh data. Um like I remember when um if you guys remember the in Houston when they had all those floods um we were really quickly able to [00:56:20] see like in the in the Houston metro um the impact on unemployment had because it was really right at that time that we actually cut off and reported on it. Um, and then it's a it's a really large sample like we we pay one out of every [00:56:34] 11 private sector employees every two weeks. So it's a it's it's a great sample of kind of what what is happening um in employment for for for that month. [00:56:45] Um and um the other thing what we did is we we enhanced it a while back too because right um employment is just sort of one indicator right what we did then is we supplemented it with wage data um [00:56:59] hours worked um and and then it's um you know you can drill down and look at every you know city state metro area um discipline or sector as well too so again I think it's a it's a really good again I would almost call it like a [00:57:14] barometer of you you know, really what's happening in the employment in in in jobs um for for that month. Um so what I want to do then is just talk about a little bit about kind of what we're seeing right now. I think it's also interesting not just how you get the [00:57:29] data but um you know really what's happening. And it's interesting we we've seen a um because not only do we look at uh businesses under 50, we've we've got big public enterprises as well too. And um it's really interesting and you know [00:57:43] what we do is it's an index, right? It's a um it's scaled to a 100. So over 100 you're seeing growth. Under 100 uh you're not seeing growth. Similar to some of like the ISM indexes. And um you [00:57:55] know we've been seeing um in the small business sector um even though it's been under 100, it's still very flattish. [00:58:04] It's it's it's you know small businesses are very resilient. Um they've been really creative in finding ways on how to um you know basically keep things going. Um, you know, the other really interesting thing that we've seen [00:58:18] recently is, um, that we track, and I know the um, the federal guys do this as well, is hours worked. We've really seen, uh, just in the past like five or six months, um, hours worked um, really [00:58:31] going up e every month. So clearly, you know, the signal there, right, is that these small businesses are really trying to squeeze out as much as they can from their current workforce before they have to go in and add another employee. So [00:58:45] again, I think it's not just the employment data, it's it's the hours worked, it's wages. Um, and we've been reporting on that, you know, the wages, you know, have been um pretty flat as well for for a long time after coming [00:58:57] off a boil um around the COVID time. Um we've really seen those kind of really steady down um as well. And um you know we also track you know job switchers and job stairs. Um you know the interesting [00:59:11] thing that we've seen on that is um you know a couple years ago you were seeing the job switchers get really really big pay raises. You're seeing that now really compressed now. It's really come come together and you're not seeing that anymore. And again something you you [00:59:25] probably probably would expect. Um, but what I was getting at is while you know the the smaller businesses are kind of flatish or maybe down a little bit from unemployment growth, the larger businesses have been really seeing some really really nice growth you know where [00:59:39] that with the index number around about 100.4 where you know some of the smaller businesses the mom and pop under 50 um have been around kind of the the 99.1 or 99.2 [00:59:52] o over the past few months. So, um it's just kind of interesting thing to kind of kind of watch with what's what's happening um with our index. You know, at Paychecks, we we again we think we got really really good data. We're we'd [01:00:05] love to to work with you all to to provide it and and support it. Um like I said, I think it's it's it's it's not doesn't get sort of um revised because it's real data. It's very fresh. You know, I heard some stuff about, you [01:00:19] know, kind of the the the job occupational stuff. We are really getting really good data now on um all our businesses and the employees in the business on job descriptions. So you can really start tracking and you know looking at kind of the the impact AI is [01:00:34] having on specific uh jobs which is another really really interesting thing. [01:00:38] So, you know, you know, we just are are really encouraged and really excited about sort of the role that we can kind of play in providing the data cuz, you know, I think you guys know, right? It's probably the the biggest most important economic indicator that comes out on a [01:00:52] monthly basis that everybody's really really watching it and it's just going to be interesting to to kind of see what what what that trend trend is. So, um, I think given that we want to provide time with with questions, I had a lot more information than I was going to provide, [01:01:06] but I just wanted to give you sort of a sense on on what's going on with with our index. Um, oh, the last thing I'll just mention, the other really interesting thing is it's really good about like providing sort of some stories on what's going on with with [01:01:20] various sectors as well as sort of uh geographical areas and and and kind of what events happen around those. and you can h you can really see sort of the the trend and the change in those numbers really really quick which is another [01:01:34] really thing that um we are really excited about uh being here in DC we actually provide our um indexing report to to all you know the folks on the hill so they can kind of see what's happening in their districts and their hill [01:01:47] they're very interesting in getting it and um again just like there's a lot of opportunities once we can really um you know kind of get this data and work with the rest of the partners to to really showcase it. So, I think I'll again I know we want to leave time for questions. So, I think I'll I'll end [01:02:02] there. [applause] >> Uh thank thanks to the panel. Um maybe [01:02:15] I'll I'll kick it off with a question which is we have we have kind of the life cycle of uh of a of the data sort of represented here. how it would flow [01:02:26] from uh from payrolls to uh uh to uh the payroll syndicators to the BLS to the to the analysts. Could I could we have a little dialogue along the vision that [01:02:40] Chris put forward and and it's it's uh and maybe I'll maybe I'll start with you Frank. Uh Chris was suggesting that uh having the payroll providers uh as a value added service to their customers [01:02:54] and as a benefit to the government and and as uh uh helping businesses fulfill their administrative obligations for for benefits uh would could be more involved uh starting perhaps with the harmonization initiatives of the of the [01:03:09] Chamber of Commerce and Jedex. Could what what are your reactions to to that part of the discussion? [01:03:15] >> Yeah. I know I that that that's great. Um so little bit a little fact here and maybe even a kind of an idea. So within the within the payroll um world, right, there's a thing called the National Payroll Consortium, which is really all [01:03:29] the big payroll providers, you know, you guys know all the names and we're included in that. And I think when you you add all those up, I think basically pay almost, you know, almost half of of all private sector employees. So if you [01:03:43] think about it, if we could just get those, so it's not like getting everything, just you get those, you'd almost get pretty much a sample size of almost half of all all workers. And um so I I would kind of go there and and you know, again, we're we're part of [01:03:57] that consortium would be would love if if they're not providing that data to to work with them. And I think that would be a really quick way to to to really get a a huge sample of the data right away. So that would be that'd be my idea. [01:04:10] uh other me members of the panel have thoughts on this on this general topic? [01:04:18] >> Well, I'll just say that uh Jedex has been working with payroll companies and with that consortium. So, >> good. [01:04:26] >> That's uh I don't think there's been any as far as I'm I think it's all been encouragement rather than resistance on their part. [01:04:36] And just to put a little flavor and why I don't think it will be a a BLS directed uh initiative, I don't even remember how many years ago, but we in just discussed internally like yeah payroll providers are are a main [01:04:49] thorough affair like can we work directly with them and in fact pay them to produce say a CES report where to your point you just click a button and our procurement people told us we couldn't target specific firms and we would probably have to be a small [01:05:03] business and put it aside and it was it was more of a procurement challenge. Um and once we realized that that idea kind of fizzled um and but you know different world whatever now but I do think [01:05:17] honestly the the business proposition and um where it's a betterment for everyone to the point where with UI they get something out of it if we can make it where it's clear uh that the payroll providers can actually sell this as a [01:05:31] service I think would go a little a little bit more than and say the statistical agencies by themselves trying to make that argument which is wonderful why um having Jedex, the Fed, other people uh invited to this conversation is so helpful. [01:05:46] >> Yeah, thanks. I think that's an important point. Thinking thinking of this as not as a procurement, but this is a a new mode for firms to fulfill their either mandatory or voluntary uh compliance [01:06:00] with uh statistical programs. Uh John So very interesting panel and and and I I was uh say very inspired Julie by your [01:06:15] your remarks of all the initiatives ongoing on all the issues we were about uh and so I'm going to in some ways paraphrase the same question that uh Katherine asked Brad as a question you could say how can we help uh and and by [01:06:29] that I mean I we globally or in this room but Emmery itself uh you know is is there room for collaborative projects Is there room for the academic community helping you get an an army of smart [01:06:43] graduate students to help you help work with the staff underneath this the supervision of the staff? I don't know, maybe you don't have an answer to that yet, but I think this is related to some things Chris brought up. We but we all know these initiatives you talk about, it's going to take time and research to [01:06:57] figure all this out, right? So, these are fantastic things. We'd like the the number of years that before we get there, not to be so large and and so we we'd really like to have a big push. How do we get the big push going? [01:07:11] >> Uh I don't have all the answers. Um, I think some of the things that we're doing, uh, as Chris said with the Fed, you guys actually have been instrumental in a lot of improvements that we've had, uh, throughout the years to twostep seasonal adjustment was because the Fed's like, "Hey guys, did you recognize [01:07:25] this improvements?" My last model are like, "Did you realize you had actually a lot about seasonal adjustment? You guys were [laughter] very much off it." Like, did you realize you had residual seasonal adjustment? I think the 10-ele1 day effect with earnings was because, you know, the Fed's like, you know, [01:07:38] we're seeing these certain finance industries that pay by monthly. we're seeing some funky things in the earnings depending on that. So, um, so that's been a a benefit. Um, I think also when we're having a lot of seed, uh, with our [01:07:52] state partners, I think just we're trying to seed to your point, we're trying to seed a lot of different projects. So, it actually isn't centrally controlled. Um, because I've been hiring like crazy and yet my net increase has been like two because [01:08:07] we're reshuffling the deck a lot. Um but I am so engaged with the quality of the people that we are bringing in with the people who are engaged in the academics this outside research to Frank's point [01:08:19] like there is so much out there that is informative and that can help and I think we're learning. So I think having things like this where we can learn from each other and say hey did you try this this seems successful. So I think [01:08:32] discontinuing that um and then I know where we are internally getting back more to the status quo and a good thing in terms of the outside researchers and um uh bringing graduate [01:08:45] programs and projects and things of that nature which we had to put on pause uh for a variety of reasons. Um, so, so, so that is I would say this has been Julie's like I'm positive like we we've been through a lot and now I am so [01:09:00] optimistic because even though we were through a lot overwhelmingly we didn't cut a lot. We were hitting our deadlines. We weren't I mean guys that shutdown was no joke and we came back out really close. And you might think, [01:09:14] oh, you just missed production. Well, the reason you're not getting CWS right away, I was hearing last night, well, that shutdown totally put a pause on that. So, when we came back, we had to figure out what to do with the missing months of CPS, which I'll tell you, the estimator and all the waiting was not [01:09:28] designed to miss the month. So, like time that we would have been spent working on that. We were working on other stuff. And then it's like, well, we can't do an annual average to do. I mean, there's people have strong opinions. Um, so there's just a lot of [01:09:41] going on. And then the DRP like we've survived but there are pockets like what we were seeing like on the census they got clamored on the the statistical methods. So this waiting this complicated waiting we've been going back and forth probably for over a year [01:09:54] now uh handing it off to CWS. So we're close. Um but but there are some things that we're still I would say the scarring effect. But that being said I am positive and we are uh looking for improvements. I am not looking to do the [01:10:08] same old same old. I've been challenging my staff to say, "What can we do more?" and really push that forward. [01:10:20] So, this is a question for Erica, and I'm kind of mildly deathly afraid of the answer, but [laughter] you know, do you have a a sense in at all in terms of kind of the number of laws that would need to be rewritten [01:10:35] and the number of agency data sharing agreements we would really need to get the streamlining that you're asking for in terms of how employers are reporting and just reducing that that administrative burden for So [01:10:57] um let's see actually I don't think it's mostly legal. [01:11:04] >> Okay. >> Uh uh but uh I think the the main place that that laws might help, right? It's not clear they're hindering, but would be a [01:11:19] presumption of access to the data produced by states in federally sponsored programs that so a law like [01:11:30] that that says that if you are running if you're a state running a federal program and producing data, the presumption is that the federal government would have access to the [01:11:43] data. that that I think would probably be the one thing that would be most helpful. [01:11:49] uh and that's probably just you know one law and uh I don't know how easy it would be to to write it and to get it passed but uh uh and then in you know [01:12:02] the list of 30 recommendations most of those things uh I mean the administration has the bulk of those but there but but there are some legal changes but I think that that's probably the the main one I think [01:12:17] that that that would make a because it also inhibits um for instance while the the ETA up until recently has always said we can't ask the states to [01:12:29] do that. It's their data and it's not clear that that's that that's really in the law but that was the way that the lawyers have been interpreting it for that many years. So having a law to clarify that would would make a [01:12:43] difference. Andrew Reamer, George Washington University. Uh Erica, very exciting to hear about the Jedex effort. Can you [01:12:57] give us a sense of the ex the um the number of states who because this is such is state-based um the data come from the states. So, how many states are kind of leaning [01:13:10] forward into this effort to enhance wage records, at least taking some action, being involved with Jedex, uh, making noises that are positive? What's your sense of the >> Well, uh, so one thing we didn't mention [01:13:25] that is, uh, in the background of all of this is the BLS's wage records program. [01:13:31] That is a purely voluntary program. Uh Julie knows much more about it than I do. Uh but uh it's up to 37 now and or [01:13:43] 39 now and and growing rapidly. The states are uh this is a voluntary program where the states put their wage records at the BLS and the BLS makes them interoperable. [01:13:57] Uh so I mean and that's uh maybe three years old, something like that. [01:14:02] Something like that. steam. [01:14:05] >> Yeah. And it's really picked up a lot of steam lately. So there's a lot So clearly there's a lot of interest in this. Uh and the number of states that are working on adding occupation to [01:14:18] their wage records is growing daily. Right? So I think it's something like 15 or 20 are in the works and more are thinking about it and some of them already have it. So uh so I think that [01:14:32] there is a lot of interest in doing that and the motivation is is actually local for for that information. So uh that's a wave to ride. I don't know want to add [01:14:45] no you're exactly correct in terms of states individually enhancing it is a little bit all over the place. Um and it depends on the specific element because there is no standardization currently. [01:14:57] uh not all of them are adding the same element the same way. Uh but to Erica's point, more states are interested and it isn't like a red blue uh like Texas is actually uh getting ready to I think implement enhanced wage records. Uh and [01:15:10] they're traditionally seen as a very employer friendly state and then you have places like Washington who has already been collecting it for a while, South Carolina, Nebraska. So there's a couple different states uh Illinois maybe that that have this information. [01:15:25] um and they're uh other states are jealous and they're working through different and states also have their own legal authorities to add things. So it's it's a complicated network. We're trying to add as much clarity as possible. Uh and them working with ETA. Uh and I [01:15:39] would say DO my hats off to this particular iteration of ETA. uh they have been much I would say compared to historically much more collaborative uh working with all the different facets of do and taking sort [01:15:53] of everyone's uh needs under consideration. So that's been really nice. [01:16:04] >> Julia, >> Julia Pollock, chief economist at Labor. [01:16:11] My question, you know, implementation is policy and um the big question keeping me up at night is how do we accelerate this whole process? So I would love [01:16:24] there to be a presumption that we have access to all the state program data. [01:16:28] The states push back and say this is an unfunded mandate. We are not designed for statistical you know data collection. We are here for program administration. Go away. So um how you [01:16:42] know h how best there are lots of things underway that I can't quite speak about because they're all working their way through various approval processes. But let's zoom out and just talk theoretical options here. [01:16:56] Should we give grants to each of the states to modernize their wage records to potentially work with Jedex or another similar solution? Um, let's say you give a million to$5 [01:17:11] million to each state and they do it all their own way, right? Does that not defeat the purpose? We want the data to be interoperable. We want them to enhance and standardize their wage records. We want um you know, so we [01:17:25] can't federalize this. We can't centralize it. There are all kinds of procurement reasons why we can't just go and buy the thing and enforce it on all the states. Uh we can't go tell every state work with this vendor. Uh so how [01:17:38] what you know what is the the path forward that could result in the best outcome the most quickly with the greatest employer buyin and state buyin. [01:17:50] That is what I'm trying to find. The silver bullet. [01:17:54] Help me find it. [laughter] Well I have a couple of uh ideas along those lines. One is that most of these programs have regular reauthorizations. [01:18:06] right? Written into the start a you know a a uh a process by which every reauthorization [01:18:17] includes in it that the data being produced to run the program should be available to the federal government. [01:18:27] Right? So that could uh when when the these programs being rethought thought start adding that in. Uh the other thing is to look at some of the models that work. I mean QCW is not mandatory but [01:18:41] because the states are supported uh and get some they get money and they get this valuable data back for doing it uh they're happy to participate right so [01:18:53] building on models like that I think I think really works and so I mean that's the kind of future I envision for the wage record program is that it becomes [01:19:05] QCW like that that that there start to be funding maybe from the UI trust fund, right? That you start using more and more of the UI trust fund [01:19:18] money to fund the information that the states need to run their labor market information and work and training programs efficiently. [01:19:31] [snorts] Anyone else on the panel follow? [01:19:40] >> I would just add that when we change NAKES codes, we don't spend years talking to each state on how they need to do it. [01:19:57] >> Thanks. Thanks. This is a great discussion and uh it it uh it I I hope it seeds further further discussions and further collaborations. We we now have a break. We'll return at uh 11:45. [01:20:11] >> What? >> 10 10 10:45. Sorry. [01:20:16] >> Nice to meet you. >> No, [applause] we should chat.