Will Tan; Investor Relations; Lexinfintech Holdings Ltd
Wenjie Xiao; Chairman of the Board, Chief Executive Officer; Lexinfintech Holdings Ltd
Zhanwen Qiao; Chief Risk Officer; Lexinfintech Holdings Ltd
James Zheng; Chief Financial Officer, Director; Lexinfintech Holdings Ltd
Zhuhan Wang; Analyst; Goldman Sachs
Alex Ye; Analyst; UBS Securities Asia Ltd.
Yada Li; Analyst; CICC
Operator
Good day, and thank you for standing by. Welcome to the LexinFintech fourth-quarter 2024 earnings conference call. (Operator Instructions) Please be advised that today's conference is being recorded.
I would now like to hand the call over to your host today, Mr. Will Tan. Please go ahead.
Will Tan
Thank you, operator. Hello, everyone. Welcome to our fourth-quarter 2024 earnings conference call. Our results were released earlier today and are currently available on our IR website. Today, you will hear from our Chairman and CEO, Mr. Jay Wenjie Xiao, who will provide an update on overall performance and the strategies. Our CRO, Mr. Arvin Zhanwen Qiao will then provide more details on our risk management initiatives and updates. Lastly, our CFO, Mr. James Zheng will discuss our financial performance.
Before we get started, I'd like to remind you of our Safe Harbor statement in our earnings press release, which also applies to this call. During the call, we may refer to business outlooks and forward-looking statements, which are based on our current plans, estimates, and projections. The actual result may differ materially, and we do not assume any obligations to update any forward statement except as required under applicable laws. Last, please note that all figures are presented in RMB terms and all comparisons are made on quarter-over-quarter basis, unless otherwise stated.
Please kindly note Jay and Arvin will give their whole remarks in Chinese first. Then the English version will be delivered by Jay and Arvin's AI-based voices.
With that, I'm now pleased to turn over the call to Mr. Jay Wenjie Xiao, Chairman and the CEO of Lexin. Please.
Wenjie Xiao
(spoken by AI) (technical difficulty) for joining us today for our fourth quarter 2024 earnings call. In the fourth quarter, we maintained a prudent operating strategy focusing on expanding high-quality assets and optimizing profitability. Driven by enhanced risk management system and advanced data analytics capabilities, we further reduced overall portfolio risk and delivered consistent profit growth.
As of quarter end, our outstanding loan balance stood at RMB110 billion. During the fourth quarter, our GMV was [RMB52 billion]. Revenue was RMB3.7 billion, and non-GAAP profit was RMB390 million. Performance has been improving for multiple consecutive quarters, and both revenue and profit have entered a clear growth trajectory.
Now I'd like to share a few key highlights of our fourth-quarter performance. First, new loans facilitated have consistently maintained high quality, resulting in a continued decline in overall portfolio risk and sequentially improving profitability compared to the third quarter, leading risk indicators for new loans, first payment default, FPD over 7 days improved by 8%, and FPD over 30 days decreased by about 9% on total loan portfolio. Day one delinquency ratio decreased by 4%. 90 days delinquency ratio decreased by 3%.
The improvement in risk performance is primarily attributed to our long-term and continuous investment in risk identification capabilities and risk management tools. In terms of risk identification capabilities, we introduced multi-dimensional third-party data, developed tailored data systems and identification models for segmented customer groups, strengthened real-time user risk identification, and leveraged the latest big model technology to improve model stability. As a result, the accuracy of risk identification improved by 15% compared to the previous quarter, while stability improved by 10%.
On the risk management tools front, we established a risk control laboratory for intelligent risk testing in the fourth quarter, creating a new paradigm of small scale experiments, AB testing, long-term observation, and dynamic strategy evolution. This approach ensures that every risk management decision and strategy iteration is grounded in data-driven insights and robust analytical support.
In the fourth quarter, we significantly enhanced our efforts in targeting and managing high-quality customer segments. We expanded customer acquisition channels and scenarios by developing CRL-based models and revamping the life cycle strategy framework, leading to continuous growth in new active users. Meanwhile, we upgraded and optimized the credit line decision-making system based on a rapid testing and validation approach significantly enhancing the accuracy of risk and credit line matching. Thanks to these initiatives, the competitiveness and profitability of our high-quality customer segments (technical difficulty) in the fourth quarter.
High-quality assets increased substantially, and the steady growth of our prime customer base further strengthened the stability and sustainability of our business. Our CRO, Arvin, will provide further details regarding our risk management initiatives later.
The second highlight is the improved efficiency and quality of our refined operations further strengthened our differentiated competitive advantage. During the fourth quarter, business lines in our ecosystem [consumer] finance, e-commerce, inclusive finance, and overseas business made notable progress.
For online consumer finance business, we deepened our exploration of customer acquisition and operations for segmented customer groups and developed tailored outreach strategies. For e-commerce business, we revamped our risk management system for installment e-commerce platform, leveraging real-time risk control, risk management, and an upgraded product supply chain to better meet consumers' needs for installment payments and hassle-free shopping. As a result, e-commerce profit entered a fast-growing track in the fourth quarter.
For offline inclusive finance business, we refined our sales management system focusing on serving small businesses' owners in lower tier cities. By enhancing one-on-one services for core customer groups, loans originated from fourth- and fifth-tier cities and below accounted for over 65% of the total GMV. Our inclusive finance business has now been profitable for three consecutive quarters.
For overseas business, we strengthened fundamental capabilities in risk management and mid- to back-office support while pre-monetizing localized operations and exploring new customer acquisition models. This led to a significant drop in new customer acquisition costs and improved operating continuity and stability.
While overseas business is still in the early phase, we remain committed to driving its steady growth. During the quarter, our Intelligence Credit Platform, ICP, gained further traction with its share of GMV continued to rise. This model has enabled effective collaboration with financial partners, leveraging complementary strengths to drive sustained growth in both revenue and net profits. Thanks to the consistent improvement in our overall asset quality, our assets have gained greater acceptance among financial institutions. This has diversified and strengthened our funding sources and structure while (technical difficulty) overall funding cost.
The third highlight is technology. We place a strong emphasis on research and development and its practical applications. In the fourth quarter, we invested RMB151 million in research and enhanced our industry-leading competitive edge.
A key focus has been on AI big models. We have completed our localized deployment of leading large models such as (technical difficulty) and developed our proprietary large model, Singularity. Now Singularity model is deeply embedded in our daily operations, enhancing efficiency across customer service, telemarketing, collections, coding, and data analysis.
In research and development, it's now fully adopted by our development teams, assisting and generating code 860,000 times monthly and offering 210,000 quality improvement suggestions in 2024, boosting coding efficiency by approximately 35%. Additionally, leveraging DeepSeek has allowed us to deploy private large models (technical difficulty) costs, putting new avenues for applications in risk management, operating refinement and workforce efficiency.
We believe AI holds immense potential to transform our core capabilities, and we will continue to invest in AI to maintain our competitive edge and drive greater value. In addition to the above mentioned highlights, to do more rights protection is always a core competitive advantage and key strength. In the fourth quarter, we further enhanced the digital and systematic development of consumer protection. Leveraging tools like AI large models, we optimize product service touch points, identified service gaps in real time, and refined communication mechanisms to improve the overall consumer experience, earning greater trust from our customers.
In supporting small and microbusinesses, we actively uphold the principles of inclusive finance. Through a continuous innovation in products and services, we have improved the accessibility and convenience of our financial services, helping small businesses address challenges in financing. Throughout the year, we facilitated over RMB30 billion loans for small and microbusinesses.
Looking ahead to 2025 amid the current macro and industry environment, we will continue to adhere to prudent operating strategy prioritizing risk management and driving further de-risking and asset structure optimization. We are confident in achieving significant profit growth this year. We will strengthen our differentiated offerings in credit lines and pricings and refine our operating systems to meet the diverse financial needs of our customers at all levels, delivering high-quality services throughout their life cycle.
Additionally, we will proactively broaden our business boundaries to foster consistent growth of our business performance. Starting this year, we will increase our dividend payout ratio to 25% of net profit. As profit continues to grow, we plan to further enhance dividends, consistently boosting shareholder returns.
Now I'll turn the call over to our CRO, Arvin. Thanks
Zhanwen Qiao
(spoken by AI) Thanks, Jay. Next, I will provide a review of our key initiatives and achievements in risk management for the fourth quarter. In the fourth quarter, we remain committed to our strategy of prioritizing asset quality, focusing on scale stability and profitability enhancement through three key initiatives, and have achieved solid results compared to the third quarter.
Leading risk indicators for new loans, first payment default, FPD, over 7 days declined by about 8% in the fourth quarter. On total loan portfolio, day 1 delinquency ratio decreased by 4%, and 90-day delinquency ratio decreased by 3% quarter over quarter.
The continued decline in risks was achieved by the following key initiatives we've taken. First, to enhance the accuracy and stability of risk identification, we have introduced new high-quality data sources while conducting deeper data mining and joint modeling with our existing core data sources to improve model performance. At the same time, we have ramped up the development of dedicated scoring models for different business lines, products, and customer segments.
Compared to general models, dedicated models use more targeted modeling samples, which significantly improve prediction accuracy. In terms of model stability, we have addressed uncertainties in model predictions caused by factors such as missing data and noise by employing algorithms to quantify the specific uncertainties, thereby, enhancing prediction stability. These automation measures have led to about a 15% improvement in risk identification accuracy and a 10% increase in model stability.
Second, we have upgraded our credit line management capabilities by adopting the test and learn approach, leveraging our strategy laboratory. We conducted credit line experiments across different customer segments to identify the optimal fit among credit line, borrower risk, and user conversion. This approach has enabled us to optimize credit lines for various user groups, improve the accuracy of credit allocation, and balance business growth and risk control, which has ultimately helped drive scale growth driven by a more competitive credit line for high-quality customers and mitigate risks from reducing credit line for high-risk customers.
Third, we have optimized and restructured our risk identification and decision-making system for API scenarios. We developed dedicated risk identification models for each core API scenario and enhanced risk screening upfront through joint modeling with API scenarios.
Furthermore, based on the customer characteristics and profiles of different channels, we have implemented a differentiated full suit of strategies covering admission, transactions, credit amount, and pricing. This enables us to meet the credit needs of customers from various traffic platforms while effectively keeping risk within our preferred range. Thanks to these upgrades, GMV of our API channels increased by about 23% quarter over quarter, while risk of new assets declined by 10% compared to the prior quarter.
Fourth, we implemented differentiated [per due date] reminder strategies tailored to groups with different probabilities of default. Meanwhile, for customers with repeated delinquencies, we launched a dedicated project focusing on optimizing repayment date settings, increasing the binding rate of frequently used bank cards, and improving the rate of auto-bid agreements. As the share of new assets from prime plus customers continued to grow, day 1 delinquency ratio of the total portfolio has continued to decline consistently.
Last but not least, in terms of risk control tool development, we have completed the construction of a risk control laboratory and fully applied it into our operations, advancing our risk management approach towards a combination of risk prediction and risk experimentation. The new risk control laboratory has established an end-to-end process covering experiment design, credit allocation, dynamic adjustments, results, evaluation.
With the laboratory, we can not only directly set experimental variables such as admission rules, tiering of credit line, and pricing models, but can also generate different strategies with one click, which significantly reduces deployment time. Also, the risk control laboratory supports intelligent traffic segmentation and dynamic sample isolation, enabling real-time, millisecond traffic distribution, as well as multi-layered experiment isolation at the user device and request levels. Furthermore, through dynamic bucketing [algorithm] based on user profiles and risk stratification, the laboratory ensures the independence of samples between experimental and control groups, effectively avoiding data contamination.
In 2025, we'll continue to improve our risk management capabilities comprehensively, covering risk identification, risk decision-making, and risk tool development. This will drive continued decline in risk, improvement in profit, and stable growth in scale.
Next, I will hand over to our CFO, James, to provide a review of the company's financial performance for the fourth quarter.
James Zheng
Thanks, Aaron. I will now provide a detailed overview of our fourth-quarter financial results. Please note that all comparisons are made on a quarter-over-quarter basis unless otherwise stated.
In the fourth quarter, we advanced our business transformation efforts, maintaining a prudent operating strategy while strengthening our risk management framework and driving business optimization. We are pleased to report the key performance metrics continued their upward trend from the third quarter, aligning with our expectations and delivering steady growth. These results underscore the effectiveness of our strategic direction and highlight the progress we've made in executing our initiatives.
During the quarter, driven by a decline in credit costs, including the provisions and the fair value changes of financial guarantee derivatives, our net income increased by 17% to RMB363 million even though the total GMV remained relatively stable. Net income increased by 54% compared to net income adjusted for the investment losses in the same period of last year.
The net income take rate calculated as the net income divided by the average loan balance increased from 1.09% in the third quarter to 1.31% in the fourth quarter, advancing by 22 basis points. We're on track of our profit margin expansion roadmap.
Before delving into the financial li items, I would like to share some highlights that have contributed to this sustainable and inline growth result. First, increased overall take rate due to the continued asset quality improvement. In the fourth quarter, we achieved revenue take rate of 6.22%, a 36 basis improvement from 5.86% in the third quarter even though the overall [APR] charge to users actually decreased by more than 100 basis points, as we focused more on high-quality customers.
The weighted average APR for loans now stands at 23.88%. This take rate was calculated as the sum of revenue from credit facilitation and take empowerment services, net of funding and the credit cost divided by the average loan balance.
The primary driver of this increase in take rate was continued improvement in asset quality. Our credit costs, which include all provisions and the changes in fair value or financial guarantee derivatives and loans at a fair value, decreased by 5% or RMB73 million, to RMB1.5 billion in the fourth quarter, reflecting enhanced risk performance.
This improvement stems from our risk management initiative initiatives previously highlighted by Jay and Arvin. All our key risk indicators showed continued improvement in the fourth quarter. Specifically on the loan balance side, day 1 delinquency rate declined by 4%, and the 90-day delinquency ratio declined by 3%.
The risk of performance of new loans aligned with our expectations, with the first payment default rate over 7 days decreasing by about 8%, and FPD over 30 days decreasing by almost 9%. Additionally, we shortened the loan duration from 13.24 months to 13.13 months.
Second, further decrease in funding costs. As another driver of our take rate improvement, our funding costs for new loans facilitated decreased by 26 basis points. Encouraged by our improved risk performance, our funding partners have been highly supportive, offering favorable terms in both funding cost and supply. We also expanded and diversified our funding sources with a number of financial partners growing to 63 in the fourth quarter. Looking ahead, we expect that the continued improvements in asset quality, deeper collaborations of our funding partners, and more diversified funding mix will lead to further optimization in funding costs, although may not be as significant as before.
Third, more balanced and a healthy revenue mix. Our revenue structure was optimized through several initiatives, including lower APR, increased capital-light loan volume, and the diversification of business lines. In the fourth quarter, as we continue to execute our strategy to optimize risk exposure, we focused on acquiring high-quality customers, which led to decrease in the APR for the newly originated loans and a corresponding decline in the credit facilitation service income.
However, this decline was offset by a 57% increase in tech empowerment service income, which represents income from our capital-light model and other services. The tech empowerment income accounted for 16% of our total income, up from 11% in the previous quarter.
The growth in tech empowerment income was primarily driven by increased the volume from our Intelligent Credit Platform, ICP platform. As an important component of our capitalized model, ICP was designed to match borrowers with a risk rating beyond our preferred range with the financial institutions and other platforms through a traffic redistribution platform. In the fourth quarter, the loan originations under the ICP model increased to 14% of the total new loan volume.
Furthermore, to enhance customer experience and provide more comprehensive services, we facilitated insurance products as well as certain royalty programs as a retention effort. Revenue from these initiatives also contributed to the growth in the tech empowerment service income. Last but not least, our installment e-commerce platform income, a complementary component of our core credit facilitation service, grew by 12% quarter over quarter and account for 9% of total income.
Fourth, improvement in custom acquisition efficiency. In addition to aforementioned take rate increase, funding cost decrease, and the revenue mix enhancement, we are committed to optimizing our sales marketing expenses by improving customer acquisition efficiency. By leveraging advanced risk identification and management systems, combined with our deep expertise in traffic distribution, we have strengthened our ability to target users more accurately, identify potential customers, and deliver better user acquisitions with higher approval rates.
As a result, new active users excluding the ICP business grew by 23% quarter over quarter, while the cost per active user decreased by 21%. We will continue to invest capital to acquire more users for the long-term sustainable growth.
Now, I will go through our key financial line items. On the revenue side, credit facilitation service income decreased by 9% quarter over quarter, mainly driven by the decrease in the new loan prices. The APRs for the new loans originated in Q4 decreased by more than 100 basis points.
The tech empowerment service income increased by 57%, driven by increased volume from our capital-light ICP and income generated from value-added services like insurance products and user royalty programs. E-commerce business revenues increased by 12% due to the increase in the GMV momentum.
On the cost and expenses side, credit costs including the provisions and the fair value changes of financial guarantee derivatives and loans at a fair value decreased by 5% quarter over quarter due to consistent improvement in our asset quality. Total operating expenses, which include processing the servicing costs, sales and marketing, R&D and G&A expenses, remained relatively stable at RMB1.3 billion.
Driven by the aforementioned factors, our net profit in the fourth quarter increased by 17% to RMB363 million. Our net profit margin as a percentage of total revenue increased from 8.5% to 9.9%.
For balance sheet items, as of year-end 2024, our cash position, which includes cash, cash equivalent, and the restricted cash, was approximately RMB4.1 billion. Shareholders' equity remained solid at about RMB10.7 billion. Our provision coverage ratio remains sufficient at approximately 255% at the end of fourth quarter.
As Jay mentioned, we are committed to providing sustainable values to our shareholders. We are pleased to announce the Board of Directors has approved a cash dividend of USD0.11 per ADS for the second half of 2024, equivalent to approximately 20% of total net profit for the second half of 2024.
As a reminder, as we announced the last quarter, effective from January 1 this year, our cash dividend payout will be raised to 25% of net income. The payout will be announced in August when we announced Q2 results. In the future, we are open to increase the cash payout ratio as appropriate to align with the growth of profitability.
Looking ahead, while our performance continues to show positive momentum, we remain prudent in light of ongoing macroeconomic uncertainties. Therefore, we expect Q1 GMV to be flat with Q4 also due to the Chinese New Year seasonality. For 2025 all year, we expect flat to single-digit year-on-year GMV growth depending on the macro, alongside a significant rise in net profit driven by profit margin expansion, underpinned primarily by continuous asset quality improvement and our overall business transformation.
This concludes our prepared remarks for today. Operator, we're now open to take questions.
Operator
Thank you. (Operator Instructions)
[Zhuhan Wang], Goldman Sachs.
Zhuhan Wang
(spoken in Chinese)
I will translate my question. The first question is what's our business plan for 2025? And the second is, what is our AI-related business layout and the specific applications of AI technologies such as DeepSeek, and what are our future plans about AI? Thank you.
Wenjie Xiao
(spoken in Chinese)
Will Tan
Let me translate for Jay. Thank you, Zhuhan. In 2025, our strategy remains prioritizing asset quality, focusing on profitability enhancement. With a priority on asset quality, we aim to profitability enhancement and the scale stability.
In terms of risk, we will continue to upgrade our risk management system. The [due] notes we facilitated this quarter, credit performance is in line with our expectation. And we will drive the continuous decline of key risk indicators in the future. In terms of profitability, we are committed to driving significant growth in net income by leveraging continuous enhancement in risk performance, optimize the funding structures and cost and improved operational efficiency.
In terms of scale, our goal is to achieve stable growth by improved efficiency of customer acquisition through our high-quality client engagement and enhanced synergies with our partners and platforms, offline inclusive finance and e-commerce business. We also will increase our investment in customer acquisition this year to further enhance the -- improve the efficiency of customer acquisition.
Despite the overall positive momentum, our performance may experience volatility due to macroeconomic headwinds and seasonality fluctuations. We will adjust our growth strategies in real time based on the evolving environment.
Wenjie Xiao
(spoken in Chinese)
Will Tan
Let me translate. As reported, Lexin is one of the first financial platforms in China to implement a DeepSeek model. Following the deployment of DeepSeek-V2 in May 2023, Lexin have recently upgraded to DeepSeek-R1. By leveraging over a decade of industry expertise and the data accumulation, we have conducted the pre-training and the localized the deployment on DeepSeek and developed Singularity AI, our own financial large model.
We have deeply applied AI technology to improve research and development efficiency, boost true innovation and business enablement. Our large model has been fully deployed in core operation workflows, including telemarketing, customer service, and collections. Through continuous optimization of dialogue flow trees and the user conversion, we have demonstrated substantial improvement in both operation efficiency and customer experience.
Also, we applied this advanced tech knowledge into our collection process. As Arvin just mentioned, we used this to improve the collection efficiency for delinquency customers. In the future, we will strategically intensify technology investment with a primary focus on advancing deployment of DeepSeek-R1. We will implement comprehensive process optimization across all business segments, explore its application in key areas of risk management, and leverage the tech knowledge to further enhance our risk management capabilities.
Operator
Alex Ye, UBS.
Alex Ye
(spoken in Chinese)
I'll translate for my question. So first question is about the company's ongoing investment in your risk management capabilities. Can you share with us some more color in terms of the latest progress on the achievement you have made in -- especially what's the current gap or differences versus your peers?
Second question is about the outlook for your risk management metrics. So what are the main targets that you aim to achieve in this year, and which are some of the most important indicators that you would suggest investors to check?
Zhanwen Qiao
(spoken in Chinese)
Will Tan
Let me translate. Overall, we achieved significant improvement of risk management capability for this quarter. Overall, our risk management capability has reached industry level and in specific technical aspects, we are already at the industry-leading position.
We have comprehensively restructured and upgraded key risk management process, including risk identification, decision-making, risk pricing, and the [post loan] management. These enhancements have significantly improved the accuracy and the stability of our risk management system. Meanwhile, we have upgraded our decision making methodologies such as test and (technical difficulty) and low and growth frameworks for credit line and pricing decisions.
And we also improved our risk tools such as dedicated risk control laboratories and the risk robots to validate and support our key risk decisions. As a result, our key risk indicators, including 90 days delinquency ratio and FPD 30 days ratio have improved for two consecutive quarters, which underscore the tangible benefits yielded by our risk management transformation efforts.
Despite these achievements, our overall performance still has some gaps compared to our peers, many dragged by legacy loans. However, as the proportion of high-quality new loans increases and the decrease of legacy loan, we expect the overall portfolio quality to further improve.
Zhanwen Qiao
(spoken in Chinese)
Will Tan
Let me translate. Our goal is still to prioritizing asset quality, focusing on scale stability and profit enhancement. Building upon the established risk of framework, we will further optimized as follows. In terms of risk detection, we will optimize asset structure by increasing the inflow of high-quality customers. Also, we will refine our collection strategies through differentiation and intelligence collection tools, ensure a continuous decline in risk for both new loan and loan balance.
In terms of scale, by enhancing specialized customer acquisition capabilities across all channels and improving offer competitiveness, we will drive the inflow of high-quality new customers, activate potential customers, and expand the credit admission through our e-commerce platform, thereby, promoting high-quality asset growth and strengthening the company's ability to navigate credit cycle.
In terms of profitability enhancement, we will upgrade the pricing strategies for customer segments with different risks, and we will further improve third-party data to further enhance our accuracy. We will improve the ROI of data cost and utilize collection tools such as intelligent case allocation and collection assistance to improve our collection costs.
Meanwhile, we will leverage AI and the large models to further enhance our efficiency and accuracy. On AI models level, we will strengthen our risk identification capabilities for different customer segments and scenarios. For example, by using customer retention model to predict customer [lapse], we can implement targeted retention strategies. By using competitiveness model to identify users key demand, we can improve -- offer competitiveness and effectively enhance high-quality customer acquisition and potential customer activation.
At the AI tool level, we will continue to develop our intelligent risk management capacity -- capabilities. For example, we will continue to leverage tools such as strategic robots and decision-making laboratories to enhance the accuracy and efficiency of strategic decision-making.
Regarding performance tracking, in addition to 90-days delinquency ratio and FPD over 30 days ratio, which we regularly disclose, we will also communicate the quarterly trend of FPD 7 days of new assets and the day 1 delinquency ratio of total portfolio, which could facilitate a more comprehensive understanding of our asset quality.
Thank you. Operator, we are ready for the next question.
Operator
Yada Li, CICC.
Yada Li
(spoken in Chinese)
Okay. Then I'll do the translation. So first of all, could you elaborate more about the trend of unit economics and the main drivers? Second, I was wondering how to view the OpEx in 2025, especially in sales and marketing expenses. Will the company become more active in customer acquisition in the following quarters? And the last, do you expect to deliver more value to the shareholders and any further plans? That's all. Thank you.
James Zheng
Okay. I will take the first two questions and then ask Jay to answer the third question. First, in terms of the unit economics, as we have been communicating with the market, if you look at the net profit margin of the company, it is calculated as a net income divided by the average loan balance. It will increase significantly to reach the industry average level in the next two years.
Obviously, the primary driver for the asset is the asset quality improvement, particularly for the new loans issued since the second half of last year. As an example, in Q4, if you look at the provisions, it was reduced by 5% compared to the previous quarter.
So as you know, the total loan portfolio is a mix of the old legacy loans and the better quality new loans. As we have more better quality new loans, and the old loans will mature and lapse, therefore, the overall asset quality will continue to improve, which will lead to sustained profitability improvement or net profit margin expansion.
Other factor obviously contributing to the improved profitability is the reduction in the funding cost. As our asset quality continues to improve, our assets receives a more acceptance from the financial institution partners and the funding costs will decline accordingly.
So as a demonstration of our profitability improvement, we can take a look in the net profit margin in the last four quarters in 2024. It started from 0.66% in Q1, 0.77% in Q2, 1.09% in Q3, and 1.31% in Q4. So we expect the net profit margin to continue to sequentially improve in the next two years to eventually reach the industry average level.
One reminder, of course, is that we may experience certain fluctuations in the degree of the profit margin improvement from quarter to quarter due to the impacts of seasonality, accounting rules, or any other timing factors. But we are very confident the overall net margin expansion trajectory will not change.
As for the second part of the question, the OpEx, basically, to continue to support the user acquisition and the business growth, i.e., the expanding new marketing channels, upgrading risk control systems, hiring top talents, and increasing AI technology investment, we do expect the absolute amount of the company's operating expenses to increase in 2025, although it will be at a slower pace than the overall company profitability improvement. The operational efficiency improvement is another factor that contributes to the marginal expansion.
So we will continue to work hard to balance the need of investing for the future and also the need to sustain the sequential profitability improvement. So that's the answers, if you will, for these first two questions, and the last one is for Jay.
Wenjie Xiao
(spoken in Chinese)
Will Tan
Let me translate. As we announced previously, our cash dividend payout will be raised to 25% of net income, effective from January 1 this year. The dividend will be announced in August when we disclose our second-quarter results.
We are committed to returning values to our shareholders. This year is our business and financial result turnover year. We expect our net income will increase significantly in 2025, and we are open to increase the cash dividend payout ratio as appropriate to align with shareholders' expectations. Thanks.
Operator?
Operator
Thank you for the questions. We have no more questions on the line. I would like to hand the call back to management for closing.
Will Tan
Thank you. This conference is now concluded. Thank you for joining today's call. If you have any more questions, please do not hesitate to contact us. Thanks again.
Operator
That does conclude today's conference call. You may now disconnect.
