Good afternoon, everybody. Thank you for being here at day one of our 24th Annual Financial Conference. My name is Manav Patnaik. I cover business and information services for Barclays. We're happy to kick off the afternoon session here with Equifax. We have Mark Begor, CEO, and John Gamble, CFO. Mark and John, thank you for being here.
Thanks for having us.
Thank you.
Mark, maybe—
This is two weeks in a row we're with you.
Yeah, that's right. It's always a good time. A lot of information to digest. Maybe just high level, let's just start off with the state of the consumer from the data that you're seeing. I think resiliency was a word you had used last—
Yeah.
—week as well. But there are also concerns with oil. The gas is not cheap. Rates at seven.
Gas, diesel fuel.
How do you balance this for the rest of the year?
Yeah. We should probably separate what's happening in the mortgage market related to higher rates. Obviously, with where the 10-year went, it touched 5% today. That's going to push mortgage rates up at 7% or north, and that's clearly having an impact on mortgage activity, particularly in refis, but also in the purchase side. Your question is more around the consumer, which I always think about, and I think it's a great indicator. If consumers are working, which they are, unemployment is low, which is very good. Employment is very high. That's always a good environment for the consumer. They have the capacity to repay, and that means our customers are still out there originating. I don't see a change in that. Clearly, at the lower end, and you can go into mid-market, as you point out, inflation, particularly with fuel, is having pressure.
That lower end subprime consumer has been challenged for quite some time. Really post-COVID, there's been an inflation that's been higher than anyone would like, which has clearly pressured that demographic. That is one where the subprime lenders two years ago kind of reset some of their originations, but it's fairly normal now. Clearly, with where inflation is, it's not good. It's not helpful for the economy. It's not helpful for interest rates. But broadly with the consumer working, I think we're in good shape. The other side is that our customers are still strong. Whether it's a bank or a financial institution or a fintech, they have strong balance sheets. They're managing themselves well. I think they're operating at a fairly strong level. They're obviously still originating, as they would. That's what their business is.
We haven't seen any change of our customers changing cutoff scores, pulling back around thinking there's a change coming in the economy. Now, mortgage clearly with where rates are now, that's more challenging. That's one where we started the year expecting a slightly down mortgage market. When we got to July, it was getting a little bit weaker through this first quarter and second quarter. Clearly that adds some pressure with where rates are now.
Got it. John, maybe if you could remind us in the context of the guide that's there for the rest of the year, mortgage rates have gone higher from the last time you spoke. Gas prices. Is it still within the context of your guidance ranges that you had assumed?
Well, the guidance we gave in July is kind of as Mark indicated, right, that we were expecting to see overall, we're going to see a decline in the mortgage market in terms of originations, right? Obviously we haven't updated our guidance since then. But obviously you've seen rates move up meaningfully since when we gave guidance in July, right? We've seen that impact on transaction volumes in the market certainly over the last month.
Got it. Mark, maybe a little bit of the longer-term picture on mortgage. We've been waiting for this recovery. It hasn't come. Sounds like it's at least delayed for the foreseeable future. Can you meet your long-term guidance targets on the top line without—
Yeah.
—a mortgage recovery?
Yeah, I think mortgage is down in the ZIP code of 50% from historic levels, mortgage activity. The good news is that even at these lower levels, there's still mortgage transactions happening. So we're building what I would call a population or a backlog of higher interest rate mortgages when inflation does come under control, when rates do come down some, there's going to be a tailwind from mortgage activity. We saw that, you may remember, pre the war in the Middle East. We saw that in the March timeframe. Rates came down slightly. We saw an uptick in refis. It only takes about a quarter point and John, there's what? 15 million mortgages now above 6% and there's almost 10%, right, over 6.5%, right? So there's a large population out there and obviously we'll see what the Fed's going to do on Wednesday.
I think the expectation is clearly the 10-year has moved up expecting the Fed is going to do a rate increase. But at some point the war's got to get resolved. At some point, inflation from really fuel and oil should come under control. There should be an opportunity for rates to come down. We've been very clear that when there is a mortgage market recovery, it's all going to flow through to shareholders. We've sized that at kind of today's levels, a normal mortgage market when it comes, and as you point out, it's harder to see today with where the tenure is, but at some point, whether it's 2027 or 2028, there's a $1+ billion of incremental revenue available to Equifax and very high incremental margins. So think $700+ million of incremental margins and that'll flow through to our bottom line.
That'll flow through really to EPS, to dividend increase and to buy back that excess free cash flow. So to your question about do we deliver our long term framework in a flat mortgage market? The answer is yes. We've been very clear about that. We have a long-term framework to grow 7%-10%. That includes a couple of points of GDP. So think about normal increases in economic activity across all of our verticals. We have a lot of confidence in our ability to deliver that 7%-10% in a call it a flattish mortgage market, which we haven't seen in a long time. It's been on the other side. It's been declining really since COVID as rates continue to move up now with the impact from the war. So we have a lot of confidence in that.
With that comes 50 basis points of operating leverage in our margin expansion, which we think is quite powerful. Very high cash conversion. Our capital allocation plan that we put in place a year ago is our intention to grow our dividend in line with the earnings. Think mid-teens, excluding a mortgage market recovery. Our excess free cash flow we will use for both on M&A, but predominantly to buy back stock. That is really our capital allocation model going forward.
Got it. Somewhat sticking to the mortgage market, not a lot of activity on the mortgage industry side, but a lot of Twitter activity going on. I want to touch on that a bit.
I think there is 39 tweets from Director Pulte in the last 10 days.
You can add one for today as well.
That'll make it 40.
It's at 40. Just curious.
I did one on Friday.
Okay.
I got to put a tweet out.
All right. Well, maybe just to address that and just overall, what do you think he's getting at, and I don't know if you've directly met him recently, but I know you said you were meeting the team broadly.
Yeah.
What is your kind of impression of what's going on here?
Yeah. Our view and the dialogues that we've had with him and his staff is that he's been very frustrated around the FICO price increases. I think everyone knows it was a little over a year ago in July of last year when Director Pulte of the Federal Housing Finance Agency said they were going to allow lender choice around scores, and adopt VantageScore. Later, really early this year, he and Secretary Turner said they were going to start accepting VantageScore, which is owned by the three credit bureaus. Today's FICO price is roughly $10. We have a $1 VantageScore price out there. So he's very positive around the VantageScore adoption. I think you saw really a week ago Friday, he came out, I think it was Thursday night, and said he was going to accelerate to full VantageScore adoption.
They had been phasing it in over time with lenders, and I think they were up to like 30 lenders were able to use FICO or VantageScore for agency mortgages. As effective Friday, it's now all lenders. So that's positive. That from our perspective is around driving that lender choice around using VantageScore or FICO. I think as everyone in the room knows, the score is really used just in the pre-qual process to really give an early indication to the consumer in the marketing flow what their pricing would be off of the Fannie Mae and Freddie Mac pricing tables. The score really helps deliver that. Once the application goes in, the score is not really used. The credit data is used from the three credit bureaus. So there's been a big focus I think by the Federal Housing Finance Agency around driving to activate VantageScore, and now that's fully activated.
Our dialogues with customers, the mortgage customers, at $10 versus a dollar, it's a huge difference. We've gone to the industry and said we're going to maintain the dollar through 2027, so to give some visibility to drive adoption. That delta, when you look at the cost of a credit file with the VantageScore versus a credit file with a FICO score, it's 45% lower with the VantageScore. That's a ton of savings for the mortgage industry, where somewhere in the neighborhood of six, seven, eight loans out of 10 don't close. They start in the process and don't close, so that's breakage. For the consumer. The total is over $1 billion. Our dialogues with the Federal Housing Finance Agency have been focused around how do we support the implementation of VantageScore. That's continuing.
I'll be in D.C. in a couple of weeks for more meetings with the regulators in Washington. I go there quite regularly, we'll continue our dialogues with them.
Got it. Just on the pricing front, maybe a two-parter. You're keeping $1 through 2027, but longer term, how should we think about that? Then the second part is your data file costs.
Yep.
How do we think about the pricing on that?
Yeah. We don't give long-term guidance to our customers or to the Street. We have a long-term framework. But when I think about pricing of the credit score, the VantageScore credit score, we're never going to have larger increases like FICO's been doing. Just not our model. We're in this for the long haul. We're in this to support our customers. We don't think about pricing in that fashion. And when we think about pricing of the credit score, we're certainly going to keep it flat in 2027. We'll decide what we do post-2027, but we want to give real visibility to our customers so they can really drive adoption of the VantageScore in mortgage. With regards to the credit file, we do modest price increases there.
We'll continue those modest price increases, really reflecting our long-term framework, which is kind of a little above mid-single digit growth, 6%-8% growth in U.S. Information Solutions. We're not in this to really drive price, and price is not our only lever. We have lots of levers at Equifax around new products, around new solutions, going into new verticals. Price is only one, and it's not one that we use as a strong one.
Got it. And it seems like even though, as you said, it appears he's obviously not happy with FICO, it seems like you guys have been caught in the storm.
We're in the blast radius for sure.
Yeah. I guess he's now revisiting the idea of tri-merge to bi-merge. Just curious, I know we talked about this extensively last year as well.
Yeah, I'm not sure if he's revisiting or if it's just still on the table for them to analyze, but we're collaborating. We have been, and we're continuing to collaborate more so around why is tri-merge so important. It's really quite basic because there's meaningful differences between the three credit bureaus' credit files. You look at the data which is out there's 10 million U.S. consumers only on one of the three credit bureaus. If you were to go to a bi-merge, they'd never get approved if you don't pull that file. If you look at most of us in this room that are kind of near-prime, prime, I'm going to say prime. I'm going to say most of this room is prime as opposed to near- prime.
Most of us in this room, if you look at your credit score at Equifax, TransUnion, Experian, it's likely 40, 50, 60 points difference for the average consumer. Think about 60 million U.S. consumers that are in the core of financial services. Why? Because not every financial institution contributes data to all three credit bureaus. You could have a bank or a fintech where it's only going to one of the three, but not to the other two. Meaningful differences. Now, what does that mean in an application process? If you were to go to a bi-merge, those consumers might not get approved. There's cutoffs. If you only pull the two low credit scores for that consumer and not the higher credit score that has more data in it, they may not get approved. It'll certainly result in price differences.
If you're only pulling the two, it's not going to cover the top two are typically used versus the bottom two if they were selected, that consumer could pay a higher price. When you think about federally guaranteed mortgages, the purpose of those is to promote homeownership in the United States and to provide access to government guaranteed mortgages. If you're going to exclude people, that doesn't resonate well politically, and it's not a positive. The flip side is safety and soundness. If you're excluding some data from the underwriting, it results in having a more risky loan if some of those trade lines or data that's included is the bad trade lines. You have a loan that's more risky than is realized because you don't have the full picture on the consumer.
Those are the reasons why we think tri-merge is so important, and why we think it's here to stay. When we meet on the Hill, we meet with Treasury, we meet with all the constituents involved, they all understand that very clearly. We'll keep collaborating with FHFA around why tri-merge is important. The other point I'd make, Manav, is if you look at the most sophisticated lenders in the U.S. outside of mortgage, or even in mortgage, if they're balance sheeting a mortgage loan, they're pulling tri-merge. Safety and soundness, approval rates. They spend a lot of money marketing. If you go in non-mortgage, where there's no requirements around whether it's 1B, 2B or 3B, the most sophisticated lenders pull a tri-merge because they get a more complete picture on the consumer. They're able to approve more at lower losses because they have more data.
I think tri-merge is kind of fundamental because of the differences in the three credit bureaus' data, and it's one we're just going to be a little more deliberate around sharing the facts around that, why it's so important.
Got it. John, maybe if I would bring you in here, just even last week you talked about kind of the breaking down the mortgage revenues by the FICO pass-throughs and other items. What's the real exposure here on tri-merge to bi-merge?
Sure. If you take a look at the portion of the Equifax credit file that have sold into tri-merge, so our trended credit file, it's about 30% of total U.S. Information Solutions mortgage revenue, which is, think, ± $900 million. Substantial amount, but on Equifax that's $6.7 billion, it gives you some perspective on the size that you're talking about here.
Got it.
I would also add, maybe we will get to this in your questioning, but I just want to make the point. We are also investing heavily to make sure we differentiate our credit file when it is used in a 1B or a 2B environment. I think we know that the mortgage credit file process has changed because of FICO pricing. If you go back three years ago, the pre-application or pre-qualification credit file pulls and score pulls were predominantly tri-merge. As FICO pricing went up, that has moved to more of a 1B pull, and then there is a 3B pull at application as required by the Fannie Mae and Freddie Mac and the Federal Housing Finance Agency. Same thing in non-mortgage. Most of that is a 1B pull in auto, card, and P-loan. I talked about the more sophisticated lenders.
We have been investing over the last six, nine, 12 months to try to differentiate our credit file versus our competitors when it is used in a 1B pull. We are adding, I think everyone in the room probably knows, we are adding income and employment attributes that we have from the TWN dataset to our credit file in mortgage. We are adding that Mark is working, that Mark, in my case, works for Equifax, and we are adding an average of Mark's income last year. Remember today, for 40 years, in an application process in mortgage, all you are looking at is the credit score. You have no idea of the applicant's income during that marketing phase before application. All you know is Mark's credit score is 750, 680, or whatever. You do not know if he is working.
You do not know if his income meets the debt- to- income DTI ratios that are requirement. There is no visibility in the historical process. We are adding that information, which we think is super valuable for the mortgage lenders to better manage their marketing funnel. Like which consumer should I lean into, where I am spending money on to get to an application and then get to closing because I have confidence that they can close, and open up that visibility around income and employment. We are adding that information for free in order to drive share gains. We also have a very unique data set on cell phone utility trade lines. Think about if you pay your streaming bill, your cell phone bill, your electric gas, water bills on time, those are very valuable attributes to add to the credit file because that data is not in the credit file.
We are adding that to our mortgage credit file, 54 different attributes, again for free to differentiate our credit file in that pre-qual process. Then what I described around income and employment data, we are also doing on our auto, card, and P-loan credit files in order to differentiate those going forward. These are examples of things that we can do post-cloud that was super complex for us to do before we had really leading technology. Second, it is a great example of the differentiated data sets that Equifax has that really give us a lot of levers versus our competitors on how we go to market.
Got it. Just one more on this topic. The MBA has been, I guess, pushing the idea of a single file and seems to have caught Director Pulte's attention as well. Is there anything to that? I guess it is quite common in the non-mortgage side of the equation, right, to do 1B, but just curious your thoughts there.
Yeah. It is one where I would remind you that the MBA represents mortgage originators, not consumers. Consumers are the ones that are really impacted by this. There are no consumer advocates that are suggesting 1B versus the current 3B. I think you have to put it at its face of the MBA's view. As a reminder something like six, seven, or eight loans that are started in the pre-application process do not close. That is breakage costs for the lenders, and that is predominantly been fixed in the pre-qual by going to a 1B and then by the adoption of VantageScore with the huge cost savings that come from Vantage. There is another lever where the industry is going to be able to pick up upwards of $1 billion worth of cost savings through that 45% cost savings of the credit file and score from VantageScore versus FICO going forward.
We think that answers the question for the mortgage industry.
Okay. Let's stick with government, but to better aspects of it.
Yeah.
Within your EWS business, your government business, last quarter you gave us some new disclosures around the ACV and renewals.
Yeah.
Can you just remind us of those numbers and just put it into perspective of how we should interpret that?
Yeah. Just for everyone in the room knows, I think this is our Workforce Solutions business where we use our income and employment data for government social service delivery. As you know, there's almost 90 million Americans that get some form of social services that's all needs-based and income verified. So if you make less, you get more social services or you qualify. That's delivered by the states, but funded predominantly by the federal government. Almost $1 trillion a year is used in the delivery of social services. Think about Medicaid, Medicare, food stamps, SNAP, TANF, rent support, childcare support, all kinds of social services. The average recipient gets over five different services, so all income verified. This is about an $800 million business for us. It's about a $5 billion TAM.
Most of the states, and I think as I mentioned earlier, the federal government pays for most of the social service dollars. The states distribute it and are responsible for doing the verifications under the requirements of each social service of each of the applicants around their income and employment. Very attractive business for us, one that we've been growing quite rapidly. As Manav points out, we had a very attractive three, four months in the second quarter of commercial activity. What's really changed in the last year or so is the big focus on the current administration around the integrity of government social services. I mentioned that there's about $1 trillion of payments that go out. The government has quantified just under $200 billion of improper payments. Improper payments being someone's receiving the social services that no longer qualify, and they shouldn't be getting those.
That's the $200 billion. I think everyone knows, last July, OB3 was passed. Inside of OB3, there was all kinds of tax stuff and everything else. There was also some additional requirements, principally around food stamps and Medicaid on the states to increase the income requirements of what you use to verify the eligibility. We've seen a very large increase in our deal pipeline. In February and again in July, we shared that our new business pipeline, so this is our commercial pipeline and our government vertical doubled year- over- year. As you asked at the front end of your question, in July, we shared that we landed in, really in the second quarter principally, $100 million of new ACV, meaning new contracts with new customers.
Think about that $5 billion TAM versus the $800 million we're penetrating into states or at the federal level, in this case, principally states. So $100 million of new ACV that benefits principally 2027. That's when the contracts kind of start. There's some of that in the fourth quarter, but the vast majority is in 2027. We also shared that we had $200 million of the existing business renewals, and it was just a lot. Government can be lumpy. I'll remind, Manav knows this, but a year ago, April, in April of 2025, we had a large new contract with Social Security Administration that we landed. So that was a large contract. Those can happen kind of episodically throughout the year. I think the most important point is that our engagement at the federal and state level has never been higher.
Having a pipeline that's up 2x year- over- year, those are commercial opportunities that we're working on is a great indicator. Obviously the $100 million of new ACV is a very positive setup for 2027.
So I guess two parts. Is this something you will give us regularly, I guess? And then the $200 million of renewals, how should we think about just retention rates—
Yeah.
—overall, just to get some perspective on that.
Yeah. The $100 million, are we going to give it to regular? You want everything. You are like most of our investors. We thought it was large enough that we shared it. Whether we share it every quarter or not, we will make the right decision on that. But $100 million is quite a bit, given the size of the vertical, and we thought it was appropriate to share. And the retention is very high. Retention levels are very high. You get very sticky once you get into workflows. It is unusual for us to have a state that either pauses for funding reasons or any other change like that.
Got it. And maybe, John, if you can just keep us grounded here on the numbers. I think you said second half government will be better than the second quarter, but second quarter was, I think, negative.
We also said back to growth.
Back to growth. Sorry, better. So just some perspective on maybe some help on what that means. Then I think last week on the deck, you also said expect accelerated growth in 2027. So just some framework before we get too carried away there as well.
Sure. I think you covered it in your question, right? So back to growth means exactly what it sounds like. We expect to see growth in the second half, then we expect to see improvements in 2027 relative to what we've been delivering here in 2026. Based on a lot of what Mark already described—
The $100 million.
—the $100 million and the renewals, right?
Yeah.
We expect to see a government business that's improving as we go through the rest of this year, and then continues to improve in 2027.
Given the size of the TAM, we've said this quite consistently. I think everyone knows, I mentioned it earlier, we expect Equifax to grow 7%-10%. We expect our U.S. Information Solutions business to grow 6%-8%. This is over the long term. So long term. This isn't this quarter, next quarter, it isn't next year. So over the long term, Equifax 7%-10%, U.S. Information Solutions 6%-8%, International 7%-9%, Equifax Workforce Solutions low double digit. And we expect government to really be at the top end of that. Given the size of the TAM and the market opportunities, we expect government to be one of the larger and fastest-growing verticals inside of Equifax Workforce Solutions, and obviously helping power their growth.
I don't know if you want to touch on talent is performing exceptionally well.
Yeah, I was going to move to talent. Yeah. I guess mortgage and government are your top two, right?
Yep.
And then talent next.
Yep.
So you've been posting some pretty good results there. So what's driving that? Because employment base—
Totally.
—doesn't seem to be that great.
Yeah. As Manav points out, the hiring market is still strong, but down. I think it's a remarkable number. I think on a normal level, 70 million people a year change jobs. I think we're in kind of the mid-60s, probably something like that, million changing jobs every year. And then for us in the background screening industry, each of those job changes result in some form of a background check. And we have a business, I think everyone knows, it's our, I guess our third largest vertical now is our talent vertical. We call talent where we sell data to background screeners to help them do their background checks.
One of the core data elements we have is everyone's job title. When we get payroll data from a payroll company or directly from an employer, we have almost 6 million companies delivering data to us every pay period. We get over 50 attributes. One of those attributes is an individual's job title. We have a digital resume on the average American. We sell that. One of the things that's checked in a background check is your prior employment. If a background check was being done on this room, you'd check five years or seven years of employment to make sure your resume wasn't fabricated. Did you really work for Chase, Citi, whatever the company is?
Five years worth of, seven years worth of job history. We can do that instantly because we have that digital record every pay period of the job titles. We also sell incarceration data. You'll remember, I think four years ago, maybe it's five now, we bought Appriss Insights, the only data set on incarceration. One of the checks that's done in a background check is were you incarcerated previously? Not to deny employment, but to allow the hiring manager and the HR manager to talk about that. That's another valuable data set. That's been growing for us. We have education data. We have a partnership with National Student Clearinghouse. Another thing that's checked in a background check is to make sure your education is accurate, that you're not rounding up or changing where you went to school. We do that check in there.
We're also rolling out a bunch of new products. We're adding a new hourly solution. If you think about this room, highly professional financial services jobs, a lot of data is used in your background check. If someone's an hourly worker at a warehouse, restaurant, retail, they might check last job work, they might check last 12 months. We've now got a product just for that. We've seen some growth there. We've rolled out a product with our incarceration data set that'll do monitoring of employee bases for incarceration after employment, right? To make sure that that's understood if there was some kind of incarceration dependent upon what the job is. That's been another positive. Record growth, as you know, we've been growing our records. We were up 10% in the second quarter, 10% for the half, and more records result in higher hit rates.
That's benefiting the business. We had some element of price. What did I miss, John?
I think you covered them all.
And penetration.
Yeah.
Just adding new clients. That's about a $5 billion TAM also, and we've got a business that's got a lot of room to grow.
Got it. John, maybe just, we get a lot of questions on the margins for Equifax Workforce Solutions, and it's kind of been ± 50 plus, but roughly there. It sounds like—
Where was it minus?
Not minus. That's why it changed from there. I guess the point is, I think Mark pointed out a lot of partnerships, including the payer-provider value investments. Is that why? Does it limit to those margins intentionally, perhaps?
We like the margins at just over 50%, and we've delivered them very consistently for a long time, and we're very specifically investing in Equifax Workforce Solutions to expand product sets. Some of them are through partnerships, some of them are through expanding data. Much of the investment also is around expanding The Work Number database, so we continue to grow it effectively. We think holding the margins at that level while investing in new product, investing in new sales channels, investing to broaden the number of white label employer services products that we bring to our partners, so we can expand those relationships and build increasing record contributions to Equifax Workforce Solutions, we think is the right way to manage the business so that we can deliver the growth rates Mark's talking about.
We like those margins, and I think we're being very specific to make sure we invest to be able to continue to deliver the growth at that margin level.
Got it. I think just one more on talent marketing. You talked about reinvesting back to manual verification.
Yeah.
I know a couple of years ago you guys got out of that business. Just to help us appreciate the difference.
Yeah.
Why that has now—
What I am referring to is that when you think about every one of our verticals, whether it is mortgage, auto, card, P- loan, we deliver dependent upon the demographic set 50%+ hit rates. Just remember the data set we have. There is about 250 million income-producing Americans in the U.S. We have roughly 110 million of them in our data set. We are growing that every quarter. So when a customer sends an inquiry to us for a mortgage application, an auto loan, a background screen, a government social service verification, we will deliver back the TWN data set, but then they still have 30%, 40%, 50% of their transactions they have to do something else with to verify the income because we do not have the records.
And we've had customers come to us and say, "Hey, can you do the whole thing?" And we did it before, now we've invested more in tech. It's a place we're investing more, to make it more efficient. And in mortgage, in background screening, in government, we're rolling out solutions where we can do the complete verification, and we would, in essence, do the manual or use some of our AI technology to do it for those records we don't have. And we think that's a real positive value add. So we're in the marketplace talking to our customers about that.
Got it. One last one on Equifax Workforce Solutions. Every now and then, the question around competition keeps popping up.
Yeah.
I think since the government introduced that ME program and you have SteadyIQ making noise, Experian, who you talked about as a priority. C heckr, who treat you as a minority investment.
They will be the same soon.
Yeah. I guess just the question on, are you seeing any changes? How do you—
Yeah. We watch the competitors. Obviously, it is a competitive marketplace. I think it really starts with records. If you can maintain the records and keep growing your records, that is really a very valuable part of the equation. And any way you cut it, our 110 million, we have 170 million active. The delta is people with two jobs. Really remarkable when you think about our data set. You have 60 million people that have two jobs in our data set. It is really remarkable. So records is really a very important area for us that we want to keep adding. We have added in the last five years, something like 50 partners. And remember, we get our records two ways. We get them through partnerships. Think about payroll partnerships, payroll processors, HR software companies. New area for us is benefit administrators.
A newer area for us is pension administrators, because remember, when you break down the 250 million income-producing Americans, non-farm payroll, which is W-2, is around 170 million people. There is 50 million- 60 million 1099 or self-employed individuals. That is obviously a lot of gig workers, but it is doctors, dentists, lawyers, private equity executives, lots of really high-paid people, dentists, et cetera, that are self-employed. Then there is another 30 million defined benefit pensioners. Think about legacy companies like IBM, General Motors, GE, that have legacy pension payments to their prior employees. Also think State of New York, City of New York, firemen, teachers, police officers, federal government, federal employees still get defined benefit pensions. That is income.
We have a multifaceted strategy to go after the records, and I think our record additions speak for themselves, that we just have very scaled data set. We continue to add records. We have a dedicated team focused on it. We are continuing to drive top-line growth.
Got it. In the five minutes we have left, I guess let us touch on capital allocation. There has definitely been a noticeable shift in the balance you have had.
Yeah.
So maybe just remind us of your priorities—
Sure.
—to date and how we should think of it in the context of leverage as well, please.
Yeah. Just to be clear, the capital allocation plan is not new, but it is fairly new. We put it in place last April. Through last April, I think everyone knows we were putting most of our capital towards completing the cloud. We felt that to be a great data analytics company, we had to have the very best technology. We spent $3 billion incremental on our tech. That is behind us. CapEx is coming down. Last April, a year ago April, we announced our capital allocation plan going forward as we substantially completed the cloud and really laid out that our intention is to grow our dividend in line with earnings. So think about mid-teens dividend growth going forward. That is going to be use of our cash.
Our free cash flow and leverage from growing EBITDA, we are going to use for bolt-on M&A and to return cash to shareholders through buyback. We have been doing a substantial amount of both. As you know, about two months ago, we announced the Mexico acquisition, Círculo de Crédito. We are super excited about that. A $740 million acquisition, fast-growing Mexican market, very strategic in its connection to the U.S. Growing middle class, lots of fintechs, and this business has been growing that we are acquiring, kind of 20%+ revenue growth rates at mid-40s EBITDA margins. So very attractive acquisition. We intend going forward to do what I use very clearly, I call bolt-on. Bolt-on acquisitions is how John and I and the board and the team talk about it to strengthen the core of Equifax.
Our excess free cash flow and leverage that comes from growing EBITDA, our intention is to maintain a strong investment-grade balance sheet. That excess free cash flow is going to go back to shareholders. In the last kind of trailing 12 months, we brought back $1.4 billion of stock. We did $500 million in the fourth quarter last year, where we've been told our investors if we're not doing bolt-on M&A, which we're going to be measured about, we're going to buy back stock. To make our last point on it from the capital allocation plan, we've also been clear, I said it earlier in our comments, when that mortgage market recovery comes, that's all going to shareholders. Said differently, we're not under-investing in Equifax because the mortgage market is constrained.
We're investing the right amounts today, so when that mortgage market comes back, we're not going to add more people. We're not going to invest more in CapEx. We're going to really deliver that through dividend and buyback. I would add one more point. We think we're one of the few info services companies that now has an AI productivity goal out there, and maybe that was going to be one of your questions that relates to capital allocation because it's going to expand our margins. As we said earlier, our intention is to grow over the long term, 7%-10% on the top line delivers 50 basis points of operating leverage. We laid out in February, as our first step, some of the AI productivity we expect to deliver inside of Equifax. Think about agents taking calls instead of people.
Think about agents doing paper processing from consumers, which we get a ton of. We have a couple thousand people in our operation center. A lot of productivity there. Technology's our largest workforce. We have a lot of technology coders and operators. We're seeing a lot of productivity there. In July, we increased our productivity goal from $75 million to $150 million. This year, against the 50 basis points operating leverage long-term framework, we set out a guide of 75 basis points, so 25 basis points higher, much of that from that AI productivity that we're delivering across Equifax, inside of Equifax, and at the half, we're 110 basis points. So we're really seeing a lot of momentum around the use of AI, not only for products, models, and scores with our customers, but also inside of Equifax.
We believe the fact that we did the big cloud investment and that we have, we think, the most advanced tech in the marketplace, it's enabling us to deploy AI more quickly for growth, but also for margins and productivity.
Got it. Well, we're almost out of time, so it's a great place to end. Thank you, Mark and John, for being here, and thank you, everybody, for coming.
Thank you for having us. Thanks.
Thank you.
Yeah, of course.