All right, hello everyone. I'm Andrew Steinerman. This is the Ultimate Services Conference. This is the Info Services Track. This is the Equifax team, John Gamble, Mark Begor. I'm also going to point out for those nerds in the room, my Information Services Data Book, which is my industry primer that we update quarterly since 2013. You can pull that up on the way out. Mark, why don't we just start by talking about the government segment and EWS? You've been excited about it. Do you feel like this is still a large opportunity ahead? How do you feel like kind of revenue has done so far in government, and what sort of needs to happen in terms of kind of maybe usage ahead driving more revenues for EWS?
Yeah, thanks, Andrew. It's great to be here. Appreciate you hosting us at the conference. Andrew's referring to the government vertical inside of our workforce solutions business, which many of you are familiar with. It's the largest business in Equifax, a little over $2.5 billion this year. It's the fastest growing business over the long-term for Equifax. It's got very attractive EBITDA margins at 50%. It's a unique business. This is a business where we have a data set of income and employment data that we collect from, at the end of the quarter, almost 6.5 million companies giving us data in the United States every pay period around their employees. We use the data in many verticals, in financial services, in mortgage to verify income and employment when someone's applying for a mortgage and an auto loan. It's used somewhat in credit cards.
It's also used very heavily in personal loans. It's used in the background screening space. One of the data attributes we get is someone's job title. So we have a digital resume on individuals. So we sell it there. What Andrew's referring to is the fastest growing part historically in the future of workforce solutions, which is the government vertical. The government vertical is where the data is used in social service delivery. I think, as everyone in the room knows, close to 100 million people in the United States get some level of social services generally from the federal government, whether it's food support, rent support, childcare support, healthcare support, Medicaid, Medicare, et cetera. That's all needs-based. That delivery of that social service is based on your income. If you make less, you get more social services.
When we define the TAM for the business, we define it as about a $5 billion TAM for the income and employment verification of those social services. That is done principally at the state level, which is where social services are delivered across all 50 states, but also in some cases at the federal government level. It is a business that is about $800 million today inside of workforce solutions. Over the last five years, it has grown kind of 20% CAGR until the last 12 months. There was a pause in the growth because of a change in CMS data costs with the states where the states had to start picking up some of those costs.
Going forward, we expect it to be a business that will be our fastest growing vertical inside of workforce solutions and faster than the rest of Equifax, meaning we expect it to grow double digits because of that large TAM that we have inside of the space. Today, less than half of the agencies across the United States use our data. If they're not using our data, they're doing either manual income verifications of an applicant that comes in, or they're using state wage data, which they have, or they're using IRS data, which is typically aged. Our data is more accurate going forward. The big change in 2025 is really around the current administration's, the Trump administration's focus on improper payments on social services.
You've seen the same reports that we've seen is that the federal government characterizes, because as you know, the federal government pays for the social services principally that are delivered by the states. Think about healthcare support, Medicaid, Medicare, SNAP, TANF, food support, some of the other social services. The federal government characterizes that as $160 billion a year of improper payments. There's a real focus in this administration to really address that. On July 4th, the president signed the OB3 Bill, which had all kinds of things in it. There was relevant to workforce solutions, our business, there was a lot more teeth put into the requirements for the states to deliver social services. We think that's going to drive even more engagement by the states around social services using our solution, that delivery. We deliver really three value props to the states.
One is speed. If someone's applying for social services, they generally need them today. We deliver kind of an instant verification. That is a real value prop of using our solution versus manual where the applicant might have to go home and get more pay stubs, more paperwork. Second is we deliver productivity to the state from an agency standpoint. You have a case worker that is going to do the adjudication of someone's paperwork to see if they qualify for that social service. When it is done using our instant data, it is done very quickly. We deliver productivity. Of course, we deliver the accuracy. With OB3, there was a lot more teeth put in, as I mentioned, around the requirements for social service delivery to try to bring down that $160 billion of improper payments.
Some of those, for example, include stricter requirements about the data validation that's done when it's delivered. There's also an element in the healthcare costs around an audit threshold for the verifications. If a state is over that threshold, in late next year, they'll be subject to paying for a portion of those social service costs. We quantified at current levels of verification, about $14 billion would move from the federal government to the states if they don't improve their audit error rates. We're working collaboratively with all the states around using our solution to deliver that. There's another new change that happens in some of the healthcare service delivery where you have to show that you're willing to look for work or go to school or volunteer. In certain age categories and certain types of households, typically single households, there's now a work requirement.
We have hours worked in our data set. We are collaborating, and that is a new product where we provide, in essence, a monitoring solution for someone on that social service and use our data to verify are they working 20 hours a week. We have access to education data in our partnership with National Student Clearinghouse, so we can validate that someone is going to school part-time, which would allow them to qualify to continue to get those social services. We are going to build a solution that integrates those two data elements along with an ability to upload volunteer documentation because you can also volunteer for 20 hours a week. That is a new solution that we are working on, and we expect to roll out in 2026 in advance of the new requirements from OB3.
I suppose just ask, what's the timing of Equifax benefiting from OB3? That's one beautiful bill.
Yeah. First off, we expect to have continued kind of core growth in the government vertical into that $5 billion TAM. What Andrew's referring to is a lot of these requirements going to affect late next year. The states are now having to prepare for that in the federal government going forward. We just see opportunities in kind of our core growth in government social services from income and verification because of the value prop we deliver. We see new product opportunities and new revenue opportunities for us around the delivery of some of the requirements with OB3. The other change that they've made is today there's 12-month redeterminations of someone's eligibility that now moved in certain services to six months. It's going to be done more frequently going forward.
The last vector, and I'll pause Andrew, is we see a lot of opportunities at the federal government level on programs we've been working on for many years, whether it's with the IRS, Department of Labor on unemployment claims. We're using our data that can just reduce error rates that happen in those services. One large one that we're focused on, and we have been for many years, is the Earned Income Tax Credit with the IRS. This is where if you make under a certain amount of income, you get a payment from the IRS as an Earned Income Tax Credit. The IRS characterizes about $15 billion a year of improper payments there. We think using our data can bring that $15 billion down. OB3 also has a new requirement where tips and overtime for certain individuals are not taxable.
We think we can deliver a solution that can help in really validating that for the IRS. We just see a whole bunch of opportunities. Going back to that $5 billion TAM that we've got to really resize because it's larger now post-OB3 against our $800 million of revenue today, we just see a lot of opportunity for growth. Again, we expect this to be our fastest growing vertical going forward because of that big penetration opportunity and the value we deliver in the social services delivery around speed, productivity, and accuracy.
Twenty twenty-six will be a good year for this, right?
We expect government to be much stronger in 2026 than certainly in 2025. Over the long-term, we expect our government vertical to outgrow EWS and outgrow Equifax.
Now let's talk about talent solutions, which is another vertical in EWS. I'm surprised that more background screeners don't use EWS. I think it's only about a third. You could tell me if that's about right. Use Equifax for verifications. Are they not doing VOEs? I don't quite know what they're doing. My question is, is a lot of the talent solution growth going forward going to be new client acquisitions for you guys?
Yeah, it's really all of the above, Andrew. Just putting a size on it, this is one of the verticals that's a fast-growing vertical for us where we deliver verification of historical employment. As I pointed out earlier, every payroll record we get, we get a job title with it. We keep all those records. Think about it, we have a digital resume on the average American going back 20 years. Obviously, as we add records, we build out that even further. One of the things a background screener does is verify past employment. Depending upon the job, it might be verify last job worked. It might be where'd you work in the last 12 months. In jobs in this room, it might be verifying your last five years of employment.
We're able to deliver that instantly versus the background screener doing that more manually by calling the company to validate, did Mark work at Equifax? In my case, did I work at General Electric? We can do that instantly. Our goal is to be the data provider to the background screening space around all the data they use in a background screen and deliver that really on an instant basis. We have a partnership with National Student Clearinghouse. Another verification that's done is where did you go to school as a part of your resume? An employer wants to validate that that's accurate. We're able to do that with our partnership with National Student Clearinghouse. Andrew remembers that we made an acquisition, I guess it's four years ago, of a company called Aperis Insights that's the only data set on incarceration data.
Another verification that's done as part of the employment process, not to deny employment, but to allow the hiring manager to have a conversation with the prospective employee is around were you incarcerated in the past? That's very common. More than 90% of background screens do that. We're the only data set on incarceration data. We also have healthcare credentialing data. If you're a doctor, nurse, dentist, healthcare professional, that's another data set that has a lot of verifications and then also re-verifications that happen every 12 months. Our goal is to continue to build out the data sets. To Andrew's question about growth, that's about a $4 billion TAM of the data requirements for background screening. We don't want to be a background screener. They do other things like drug tests and other verifications that take place, but we want to be the data provider.
A $4 billion TAM, our business now is about $400 million. The growth levers, as Andrew points out, is penetrating to more of the background screeners and converting them from doing manual verifications of those kind of elements on the background check to using our digital instance solution. It's also adding more data assets either through partnership or acquisition, like our Aperis Insights incarceration data set acquisition to add more data elements to our data set. It's also products, developing new products that really meet the requirements. For example, we're working now on a product for the hourly workforce that's usually a thinner background check than one that would be for white collar and trying to get the right data elements that really meet the requirements of those employers.
That would be an example of a new product that we would roll out that would drive really penetration into the TAM.
We also drive growth by growing records, right?
Yeah, of course.
Records, across workforce solutions, we've had tremendous success in growing our record base, not just of current and active records, but also of historical records. Obviously, historical records are critical in background screens. We have a record on almost everybody in the United States, not quite, but pretty close in terms of current or historic. We've done a very good job of continuing to grow our current records, up about 10% this year. Historically, we've been growing consistently over 10%. We're making tremendous strides in filling out that database. It's probably the most consistent long-term driver of growth in workforce solutions, is really record growth.
Maybe just, Andrew, to put a point on records, I think everyone in the room knows there's roughly 250 million income-producing Americans in the United States. Think income-producing, that could be pension retirement income, defined benefit pension income. It's W-2 income. There's about 170 million W-2 non-farm payroll employees in the United States. Then there's about 40-50 million self-employed 1099 employees in the United States. Think doctors, dentists, lawyers, but also think Uber drivers, DoorDash drivers, self-employed landscapers, etc. Against the 250 million, as John said, we've been growing our records about 10% a year. It's not our long-term growth rate. It's been quite strong in the last four or five years, as Andrew knows. Against the 250 million at the end of the quarter, we had about 100 million.
We have a lot of growth potential to add more records either directly with employers. As I said earlier, about 6.5 million companies today contribute payroll records to us every pay period. That is growing. There are a lot more companies to go and a lot more individuals. What is unique about our business model is obviously we have integrations with mortgage lenders, auto lenders, credit card issuers, personal loan lenders. We talked about background screeners and government social services. When we have that commercial relationship, they send us every inquiry to Equifax, and we fulfill the records that we have. When we add records at any point in time, we are able to monetize them immediately because we are already getting the inquiries from all those different sources. That record growth is very valuable to us in growing our top line and bottom line.
We have a dedicated team that really all they do is add records.
It's definitely the gift that keeps on giving. I'm going to ask a more annoying question, John. This is about EWS mortgage volumes, which in the third quarter were still down. Other non-Equifax data, industry data about mortgage, suggests mortgage is up. I know there's been a gap for a while. All these data sets, if we look at MBA or Fannie, they're all particular, let's call it that way. I really don't know which data set I should be looking at, but there does seem like there's a gap. My question is, what do you think of it?
We focus very hard, obviously, on the data we have ourselves. We look very closely at USIS hard mortgage credit inquiries. They were down in the quarter, right? We disclose that every quarter. We provide long-term trends. We gave a perspective on what we expected to occur with USIS hard mortgage credit inquiries for the full year, indicating they would be down high single-digit percentage rate. I think if you take a look at USIS hard mortgage credit inquiries and try to correlate it back to Twin inquiries, they are not exactly the same, but they are generally in line, right? We think those are two pieces of information that are current and real, right, that we can correlate together to try to see perspectives. Correlating to MBA data or other survey-type data, we think is very difficult to do.
They do not include full market coverage. We do not think it is that relevant to compare. That is why we focus so heavily on the data we can see every day. Again, as a reminder, to close a loan, you need to pull credit, right? Hard credit inquiries, we think, are a very relevant measure of what is really occurring in the industry.
As you know, for 10 years plus, we have used hard mortgage credit inquiries from USIS as our definition of market. We think that is the best indicator we have because, as John said, every mortgage pulls all three credit files. We are seeing every mortgage.
I agree with that point.
We think that's a good indicator. The reason USIS revenue in 2025 is up, even with the mortgage market down, is really because of the FICO pass-through. We're pleased with the EWS performance versus that market that is down year- over year as they roll out new products and bring penetration and record growth and pricing to the marketplace.
This is a thesis that I have had about the gap, that more of it is getting done manually because mortgage lenders are not so busy, let's call it yet. They could do it manually.
There's probably some element of that, but I think it's small. We track very closely our customers' utilization. There's very high utilization of the Twin. Maybe just so everyone knows, what we're talking about is that in a mortgage application process, you pull credit and you also verify income and employment. That's required for all conforming federally guaranteed mortgages. You can do that verification either manually with paper pay stubs or you can do it using our verified solution. The vast majority of the market uses our verified solution, instant solution, because it provides speed, productivity, and accuracy versus a paper pay stub. We've seen very high adoption in the mortgage market of using our solution because of really the value it delivers.
Okay. Looking into next year, in 2026, I wanted to talk about USIS's mortgage business. I see there's basically three subjects to talk about. One is lower mortgage rates. The second one is the potential commercialization adoption of Vantage in mortgage. The third is anything going on around pre-qualification period, pre-approval period. Is that increasingly being used by 1B or 2B and has Equifax's solution for that with income flag moved the needle? How do we feel about 2026 mortgage market across those three points?
I think we need another half hour in our meeting. For that question, Andrew, it's a loaded one. Let me try to do it quite quickly. I think those that are close to the space know there's a lot going on in the mortgage space, particularly around credit files. First on the market, I think everyone knows the market's been down substantially over the last four years of rates that more than doubled from where they were during the COVID timeframe. In the case of Equifax, that's over $1 billion of revenue that came out of our P&L as the market went down over the last four years. As John mentioned earlier, in 2025, as we define the market, the market's down kind of high single digits in 2025.
As we look forward to 2026, we would expect, and we're seeing signs kind of in the second half of the year that the mortgage market will bottom. If you believe rates are going to be stable or down from here versus going up, that's really where you think about the mortgage market bottoming. We've sized for our investors when we look forward, we look at 2015 to 2019 inquiries as being normal, kind of pre the COVID refi boom that happened. And we're 50% below that today. We believe there's a very sizable mortgage market recovery, not with rates going back to two and a half, but with rates coming down from what would be 20-year highs of where they are today once inflation comes under control. I'm not going to handicap that.
What we tried to size for our investors is the opportunity at Equifax when rates come down is very substantial. We've sized it as being over $1.2 billion of incremental revenue with a mortgage market recovery coming back to normal, about $700 million of incremental EBITDA and $4 of incremental EPS going forward. When we look at the future, we're not going to handicap when that's going to happen. We want to size for our investors that when that does come through, we're not going to invest that incremental EBITDA, free cash flow, or EPS in more people. We're not going to invest in more CapEx. We're not going to do more bolt-on M&A. We're doing the right amounts today of investing in Equifax.
That incremental revenue from a mortgage market recovery and EBITDA is going to show up in free cash flow and it'll end up in dividend and buyback. That's where that's going to go from a market standpoint. I think your second question was around what's going to happen with the change in the conforming mortgages with Fannie and Freddie. FICO had a 30-year monopoly, I think as people know. In July, that was removed and Vantage was accepted as a credit score. I think the room knows that the three credit bureaus own VantageScore. And we see a real opportunity to migrate the mortgage industry from using FICO to using Vantage. FICO came out about a month ago with pricing for 2026, and they doubled their price from $5 to, it's actually $4.95- $10.
About a week later, we announced that we're going to sell the VantageScore for $4.50, meaning half of the FICO pricing. We think that there's an opportunity with that pricing umbrella to drive conversion of FICO to VantageScore that's going to provide value for the mortgage industry, meaning real cost savings and hundreds of millions of dollars across all three credit bureaus, and also for the consumer with a lower price.
USIS '26 question. Is it going to move the needle in 2026?
Yeah, I think that's harder to handicap. It's going to take time for that to roll into 2026. Over the long-term, we sized for our investors that it's a new $100 million-$200 million profit pool for Equifax. We'll give guidance in February when we do our fourth quarter earnings of what we think 2026 might look like. We'll also provide some scenarios of upside, downside versus that. We're seeing quite a bit of momentum in the industry because of that shock of the costs of $5 going to $10 around the desire for an alternative. I think as many of us in the room know, Vantage is used in other verticals quite substantially. We just think there's a real opportunity. I think your third question is about what are we doing around differentiating our credit file.
Andrew highlighted in mortgage and prequal, a few months ago, we rolled out a plan to add an income and employment, we call it Twin. The Work Number is Twin, indicator on our credit file that's used in the prequal process. If you're applying for a mortgage today, generally, frankly, because of FICO pricing, there's only one credit file pulled by a mortgage originator to check your credit score in that shopping process. They've got a very large digital funnel, and they're trying to evaluate Andrew versus Mark versus John about which mortgage do I put them in? Is this someone who's going to qualify, or are they not going to qualify from a credit standpoint? That's why they pull a prequal or shopping credit file. They're blind to, is Mark working? And they're blind to my income levels.
We're adding to the Equifax credit file to drive share gains at no cost to the mortgage originator, some indicators on our credit file that Mark's working, who Mark works for, which is quite important in a mortgage process. That's one thing that has to go into the mortgage application process. And then an indication of Mark's historical income. Not his actual income, but Mark's a $150,000 a year employee. We believe that's going to differentiate our credit file when they're only pulling one versus our competitors who don't have the ability to add that income and employment indicator. We're going to do the same thing in auto and cards and P loans to differentiate our credit file going forward. We think that's going to drive share gains because we're not going to charge for it.
We really want to differentiate our credit file when there's a $1 billion pull. John, go ahead.
Just real general comments around 2026, just to give some people some perspective on some items that it's hard for them to predict, right? Just looking at some below-the-line items, looking at depreciation and amortization, excluding acquisition amortization. We expect our D&A to be up about $50 million in 2026 versus 2025.
That's really from our big investments in the cloud transformation that we're now substantially complete with.
Absolutely. Yep. In terms of interest and other expense, as we continue to execute acquisitions and then also execute our share repurchase program that we started in the second quarter, we would expect interest expense to probably be up on the order of $25 million. As we look at our effective tax rate in 2026 relative to 2025, likely higher by on the order of 50 basis points. Again, I'd say interest expense is higher, consistent with what we've, if you take a look at what we did in the second quarter and third quarter, very significant share repurchases that we executed, buying back, for example, in the third quarter over $300 million worth of stock, about 1% of our outstanding. We are going to continue to execute against that capital return program that we indicated, which is what's driving interest expense higher.
Let me just make sure I get that. I think you said $25 million of additional interest expense. Is that on the share buyback that was done, or does that contemplate more?
That's 2026 versus 2025. That would contemplate our capital return and acquisition programs in 2026 as well.
Oh, it would. I gotcha. Okay, great. Open for questions. Oh, come on, you guys are so chatty usually.
Do you find it to be poor shopping in Vantage or in FICO? Are people going to pull those and see if that's kind of better LTA?
Yeah, I think clearly the regulators do not want that, meaning the agencies definitely do not want score shopping to take place. One of the things that we announced we are going to do, and we are doing it as we speak, is we announced the $4.50 versus $10 score for Vantage, meaning half the price of the FICO score. We also said we are going to lock it in for two years. Our customers have some certainty around that score pricing as they think about converting. We are also offering, and customers are taking it up on us as we speak, a free Vantage score with every paid FICO score so they can really evaluate the Vantage score going forward.
No, the score is going to be very similar from an actual score standpoint, and there will be a band of pricing levels for FICO that exist today, and those same ones will be created for Vantage. I do not see that being an issue.
Yeah, so we only have one minute. I do want to ask this question. Your Equifax AI strategy, what should we know about it? Here's a key question on that. For those clients that are more AI leaning, does that kind of cohort of customers end up consuming more Equifax data than the control group?
Yeah, so AI for us is really, we're on offense with AI. I think what's really unique about our space is we have proprietary data. So we're the only ones that can deploy AI against our data. What's also unique is you have to have explainable AI, which is really hard, explaining which data element drove the decision. So we've been investing for seven, eight years around explainable AI. We've got over 300 patents in explainable AI. We added 12 in the first half of the year around explainable AI. Where AI is going to orchestrate itself for us is in our scores, models, and products being higher performing because we're able to ingest more data.
We're seeing meaningful lifts in the performance of our products when we use AI and our differentiated data at scale in order to deliver those, which means we're going to be able to sell a higher ROI solution to our customers, which is either going to result in share gains or higher price because we're going to charge a higher price because it's higher performing. We're also using AI inside of our operations. We've got a lot of opportunity in our back office to use AI to drive productivity. That's kind of a new chapter, kind of post-cloud for us. Our principal focus is really deploying AI to deliver a higher performing product using more data. You have to use AI when you have large data sets and multiple data sets in the kind of product you're delivering to marketplace.
All right. Great place to end. Mark and John, thank you for joining us. We.