360 DigiTech Inc

NASDAQ:QFIN

Earnings Call Transcript 2024-Q1

[Operator]

Ladies and gentlemen, thank you for standing by, and welcome to the Qifu Technology First Quarter 2024 Earnings Conference Call. [Operator Instructions] Please also note today's event is being recorded.At this time, I'd like to turn the conference call over to Ms. Karen Ji, Senior Director of Capital Markets.

[Karen Ji, executive]

Thank you, operator. Hello, everyone, and welcome to Qifu Technology's first quarter 2024 earnings conference call. Our earnings release was distributed earlier today and is available on our IR website. Joining me today are Mr. Wu Haisheng, our CEO; Mr. Alex Xu, our CFO; and Mr. Zheng Yan, our CRO.Before we start, I would like to refer you to our Safe Harbor statement in earnings press release, which applies to this call as we will make certain forward-looking statements. Also, this call includes discussions of certain non-GAAP financial measures. Please refer to our earnings release, which contains a reconciliation of the non-GAAP financial measures to GAAP financial measures. Also, please note that unless otherwise stated, all figures mentioned in this call are in RMB terms.Today's prepared remarks from our CEO will be delivered in English using an AI-generated voice. Now I will turn the call over to Mr. Wu Haisheng.

[Haisheng Wu, executive]

Hello, everyone. Thank you for joining us today.Starting in the second half of 2023, we have adjusted our business strategy in a timely manner, focusing on quality growth and improving profitability as the company's primary goals. Over the past few quarters, we have strictly managed risks and enhanced profitability by adhering to a prudent business strategy. All these efforts have enabled us to demonstrate stronger resilience in today's challenging macro environment and deliver solid performance to the market.Over the past quarter, we continued to expand the coverage on both ends of our platform, empowering 159 financial institutions to provide credit services to over 52 million credit line users on a cumulative basis. In response to macroeconomic headwinds, we further tightened our credit standards and streamlined our business structure to enhance the overall health and sustainability of our operations. We also optimized our profitability model through refinements made to our product offerings, risk management, fund structure, user acquisition and asset distribution capabilities.Revenue during the quarter increased by 15.4% year-over-year to RMB 4.2 billion, while our net take rate increased by 54 basis points to roughly 3.5%. Non-GAAP net income increased by 23.4% year-over-year to RMB 1.2 billion and non-GAAP net income per diluted ADS increased by 28% year-over-year to RMB 7.58. ROE reached approximately 22% in the quarter, significantly outperforming the industry peers. Under more stringent credit standards, total loan facilitation and origination volume across our platform came in at RMB 99.2 billion in Q1 with further improvements to risk indicators for new loans.Now let's move on to our key initiatives and the progress we have made in this quarter. Our top priority in Q1 was improving our asset quality. As we tightened our overall credit standards, we further iterated risk strategies across the loan facilitation, credit operation and post credit processes to improve risk metrics. We also revamped our strategy framework to integrate risk segmentation and introduced external data sources from leading Internet platforms for joint modeling and scoring, improving our ability to identify and intercept high-risk customer segments.With regard to loan collection, we actively expanded and optimized line resources to increase connection rates. By refining collection strategies and enhancing incentive schemes, we gradually boosted our overall collection efficiency. As a result of these measures, our expected vintage loss for new loans in Q1 decreased by roughly 15% sequentially. Additionally, D1 delinquency rate and 30-day collection rate of the overall loan portfolio improved by 14 basis points and 19 basis points respectively. We expect to make further gradual improvements to our risk metrics in Q2.With ample liquidity in the financial system during the quarter, demand from financial institutions for consumer credit assets remained robust. Our industry-leading risk management capabilities placed us in a competitive position in collaboration with financial institutions. Leveraging our stable asset performance, we stepped up ABS issuance efforts and actively worked with financial partners to reduce funding costs.During the quarter, we issued roughly RMB 5.3 billion worth of ABSs, representing an increase of 130% year-over-year with issuance costs falling by roughly 150 basis points year-on-year. Notably, we issued the first domestic exchange-traded ABS with a AAA international rating valued at RMB 1 billion. The ABS program attracted subscriptions from global professional investors, which significantly expanded our funding channels globally. Driven by both a higher percentage of low-cost capital-light ABS in our funding mix and a further reduction in funding costs for loan facilitation, our overall funding cost decreased by over 70 basis points sequentially during the quarter. We expect to maintain our advantage in funding costs throughout the remainder of the year.We also adopted a more prudent marketing strategy, further optimized customer acquisition channels and bolster acquisition efficiency of major channels. In the quarter, our acquisition cost per credit line user decreased by roughly 12% sequentially. The percentage of new users with approved credit lines from our embedded finance business increased to 36.4% from 34.9% last quarter. We continue to maintain our edge across leading embedded finance channels in terms of user conversion rate and loan volume by leveraging our user identification and risk control capabilities. Through differentiated operations, we have continued to optimize risk and unit economic models. In Q1, the credit performance and operational efficiency of the embedded finance channel were further optimized and the ROA of new loans from this channel increased by 115 basis points from Q4.As we improve the accuracy of user identification and profiling, we have been able to onboard a more diverse pool of financial institution partners, strengthening our ability to serve various loan asset segments. By aligning assets with the risk appetites of different institutions, we improved asset allocation efficiency and increased overall returns on our loan portfolio. Through a more precise match between loan assets and funding partners, we achieved better risk performance and overall profitability.During the quarter, the percentage of our on-balance sheet loan volume increased to 28%, while the percentage of our loan facilitated under the ICE model increased to 21%. Meanwhile, the take rate of ICE model increased by 76 basis points compared to the same period last year. Our extensive user base has always been the bedrock of our operations. To cater to users' diverse needs, we have offered differentiated value-added services through a loyalty program to boost user retention and engagement. Going forward, we aim to further enrich the value propositions of our product offerings and we'll implement differentiated user operations to enhance user satisfaction and drive long-term growth in LTV.We continue to invest in cutting-edge technologies with a strong focus on expanding the application of AI and large language models in the fintech sector to elevate user experience and improve operational efficiency. We integrated large language models into our core capabilities and developed a standardized Qifu AI Co-Pilot System that has been deployed across key segments of our business, including risk management, telemarketing, loan collection and customer service.The system enables intelligent human-computer interaction through automatic speech recognition technology or ASR. It has currently achieved a recognition accuracy rate of 97% in our own collection scenarios, leading the financial industry standards. Additionally, through the use of voice print recognition capabilities, we have achieved a remarkable 95% accuracy rate in identifying blacklisted customers, helping us effectively prevent asset losses and malicious complaints. Finally, we also rolled out an AI development tool, UGAI, and applied it across various stages of our development cycle, including requirement communication, solution design, coding and testing. With an adoption rate of 20% for AI-generated codes, we achieved a 30% improvement in development efficiency in the applied fields.Our technology solutions business continued to make steady progress. During the quarter, we entered into partnerships with 2 additional financial institutions, bringing the total number of financial partners for our end-to-end technology solutions to 7. These partnerships cover different categories, including Internet, private and municipal banks as well as consumer finance companies. Through our end-to-end tech solutions, daily average loan volume reached RMB 11 million in April 2024. As financial institutions take on an increasingly prominent role in the consumer credit market, we remain committed to assisting financial institutions in advancing digital transformation and sharing the benefits of their long-term growth.Moving on to the outlook. Despite the marginal improvements in our risk indicators and initial positive signs of a macroeconomic recovery during the quarter, we will remain patient and continue to prioritize risk performance and operational efficiency until we see clear signs of a recovery in credit demand. In the meantime, we also recognize the vast market potential there is with a substantial base of unmet user needs and inefficient connections between financial institutions and end users.With more than 52 million cumulative users with credit lines, we have developed deep user insights and industry-leading capabilities in online customer acquisition and profiling. Moving forward, we will actively explore a more open platform model. Starting with user needs, we aim to facilitate more efficient connections between users and financial institutions and work with financial partners to offer a broader spectrum of products that address user credit needs throughout the life cycle.Since 2024, we have significantly optimized capital allocation by stepping up share buyback efforts, while ensuring stable returns through a dividend policy. The USD 150 million share repurchase program announced in June 2023 was successfully completed at the end of March this year, 3 months ahead of schedule. Starting on April 1, 2024, we have been actively executing our new share buyback plan of up to USD 350 million. We have full confidence in the long-term development of our company. Through ongoing buybacks and dividends, we aim to further boost capital allocation efficiency, optimize shareholder structure and enhance long-term shareholder returns.With that, I will now turn the call over to Alex Xu.

[Zuoli Xu, executive]

Thank you, Haisheng. Good evening and good morning, everyone. Welcome to our first quarter earnings call. Despite the still uncertain macro environment in the first quarter, we made good progress to optimize our operations and further train exposures to underperforming assets and deliver solid financial results. Total net revenue for Q1 was RMB 4.15 billion versus RMB 4.5 billion in Q4 and RMB 3.6 billion a year ago.Revenue from credit-driven service capital-heavy was RMB 3.0 billion in Q1 compared to RMB 3.2 billion in Q4 and RMB 2.6 billion a year ago. The year-on-year growth was mainly due to growth in on-balance sheet loans and contribution from other value-added services, partially offset by declining in offset -- in off-balance sheet loans. On-balance sheet loans account for around 28% of the total loan volume. Overall funding costs further declined over 70 basis points sequentially and over 100 basis points year-over-year with the help of our strong relationship with financial institutions, partners and record high ABS issuance.Revenue from platform service capital-light was RMB 1.1 billion in Q1 compared to RMB 1.2 billion in Q4 and RMB 969 million a year ago. The year-on-year growth was mainly due to strong contribution from ICE and other value-added services, substantially offsetting the decline in capital-light loan facilitation. As we try to strike an optimal mix between risk-bearing and non-risk-bearing assets in the uncertain macro environment, we are also gradually cutting back loans that generate marginal returns. In Q1, we saw continued sequential improvement in revenue take rates for both cap-heavy and cap-light operations.During the quarter, average IRR of the loans we originated and facilitated was 21.5% compared to 21.3% in prior quarter. Looking forward, we expect pricing to be fluctuated in a narrow band around this level for the coming quarters as we further optimize our loan portfolio in response to the macro uncertainties. Sales and marketing expenses decreased 25% Q-on-Q and 2% year-on-year as we intentionally control the pace of user acquisitions in an uncertain environment. We added approximately 1.45 million new credit line users in Q1 versus 1.7 million in Q4.Unit cost to acquire our new credit line users decreased significantly Q-on-Q to RMB 285 from RMB 326 mainly due to our more disciplined approach and the Chinese New Year seasonality. We will make timely adjustments to the pace of the new user acquisition based on macro conditions from time to time and further diversify our user acquisition channels. Meanwhile, we will continue to focus on re-energizing existing user base as repeat borrowers historically contribute the vast majority of our business.90-day delinquency rate was 3.35% in Q1. This ratio was calculated by dividing outstanding balance of on and off-balance sheet loans that were 3 months past due with the total outstanding balance of on and off-balance sheet loans across our platform on March 31. During the quarter, we purposely cut our exposure to certain risk-bearing assets and reduced the total outstanding balance of on and off-balance sheet loans by approximately 16.5% sequentially. As such, the 90-day delinquency rate was mathematically inflated by roughly 16.5%, which is somewhat misleading.Furthermore, as we always know, this metric is backward looking in nature and provide little value to help investors understand our asset quality trend. We strongly recommend investors focus on key leading risk indicators such as day 1 delinquency and 30-day collection rate. In fact, we started to see modest improvement in asset quality in Q1. Day 1 delinquency was 4.9% in Q1 versus 5.0% in Q4. 30-day collection rate was 85.1% in Q1 versus 84.9% in Q4. The improvement was more noticeable among new loans issued in Q1 as timing risk management measures start to show benefit in the quarter.As Haisheng mentioned, expected vintage loss from new loans issued in Q1 declined by roughly 15% sequentially and 30-day collection rate further recovered to nearly 86% in April. We have further optimized our risk management model and applied more restrictive standards to new applications -- on new applications to mitigate potential risks throughout the quarter. We also proactively adjust our business mix to further reduce our exposure to higher risk assets. Although economic conditions remain uncertain, we believe overall risk performance of the loan portfolio should gradually improve throughout 2024.As macro uncertainty persists and credit quality fluctuates, we will continue to take prudent approach to book provisions against potential credit loss. Total new provision for risk-bearing loans in Q1 were approximately RMB 1.4 billion versus RMB 2 billion in Q4 and the write-backs of the previous provisions were marginal in Q1. The significant sequential decrease in new provisions was mainly due to the substantial Q-on-Q decline in off-balance sheet capital-heavy loan volume, while the new provision booking ratio remained relatively stable. The decline in write-back was due to expected risk of existing loans remain stable and micro uncertainties persist.Provision coverage ratio, which is defined as total outstanding provisions divided by total outstanding delinquent asset-heavy loans -- loan balance between 90 and 180 days were 414% in Q1 compared to 481% in Q4. The provision coverage ratio was still well within our historical range. Non-GAAP net profit was RMB 1.2 billion in Q1 compared to RMB 1.15 billion in Q4. The effective tax rate for Q1 was 23.3% compared to our typical ETR of approximately 15%. Net profit and ETR was negatively impacted by RMB 130 million withholding tax provision related to significant cash distribution from onshore to offshore for dividend payment and share repurchase program during the quarter.With solid operating results and higher contribution from capital-light models, our leverage ratio, which is defined as a risk-bearing loan balance divided by shareholders' equity was 2.5 -- sorry, 2.8x in Q1 at historical low. We expect to see leverage ratio fluctuated around this level in the near future. We generated approximately RMB 1.96 billion cash from operations in Q1 compared to RMB 2.35 billion in Q4. Total cash and cash equivalents was RMB 8.3 billion in Q1 compared to RMB 7.8 billion in Q4. Non-restricted cash was approximately RMB 5.3 billion in Q1 compared to RMB 4.2 billion in Q4.As we continue to generate healthy cash flow from operations, we believe our current cash position is sufficient to support our business development and to return to our shareholders. On June 20, 2023, we announced a share buyback program to repurchase up to $150 million over a 12-month period. In Q1, we bought approximately $16 million worth of ADS in open market under the 2023 repurchase plan. As of March 28, 2024, we have completed substantially all of the $150 million 2023 share repurchase plan.On March 12, 2024, we announced a new share repurchase plan to purchase up to $350 million worth of ADS over a 12-month period starting April 1, 2024. As of May 17, 2024, we had in aggregate purchased approximately 3.4 million ADS in the open market for a total amount of approximately $65 million, inclusive of commissions at an average price of USD 19.3 per ADS under the 2024 share repurchase plan. The pace of the repurchase is faster than time scheduled.The proactive execution of a share repurchase plan further demonstrates management's confidence and commitment to the future of the company and the management intends to consistently use share repurchase plan to achieve additional EPS accretion in the long run. With the full execution of the new share repurchase program and the dividend plan, we are generating highest combined yield on the recurring basis among Chinese ADRs to our shareholders.Finally, regarding our business outlook. We will continue to focus on enhancing profitability and efficiency of our operation and the current macro conditions. For the second quarter of 2024, the company expects to generate non-GAAP net income between RMB 1.22 billion and RMB 1.28 billion, representing a year-on-year growth between 6% to 12%. This outlook reflects the company's current and preliminary view, which is subject to material changes.With that, I would like to conclude our prepared remarks. Operator, we can now take some questions.

[Operator]

[Operator Instructions] We will now take our first question. This is from the line of Chiyao Huang from Morgan Stanley.

[Chiyao Huang, analyst]

[Interpreted]So basically 2 questions. One is regarding the loan volume outlook, especially given the demand and the risk situation we see right now. How is management's view changing from beginning of the year? And second question is on the outlook on take rate improvement. Just wondering how confident is management about improving take rate for the full year 2024? And has the view changed compared to beginning of the year?

[Haisheng Wu, executive]

[Interpreted]Okay. I will do the translation. We do understand everyone is concerned about our credit mix risk and loan growth. Since the beginning of 2024, consumer credit demand is due year to recover, especially since late March, whereby demand has been weaker than we expected. This is reflected in the user initiation rate on our app. Excluding the impact of the spring festival, user initiation rate in the first 4 months of last year was relatively stable. However, the initiation rate in March and April this year is slightly lower than the previous 2 months. So the user credit demand is weaker than the same period last year.The changes in our user demand are basically consistent with this year's macroeconomic trends. In the first quarter, CPI still maintained a relatively low growth rate and the balance of household short-term consumer loans decreased by RMB 271 billion quarter-on-quarter. The social financing scale in April also decreased by RMB 199 billion sequentially, marking the first leveraged growth on a sequential basis in the past 2 years. This data also confirms that credit demand is yet to recover.In terms of credit risk, we have taken a lot of actions in the past 2 quarters, including tightening the approval rate, optimizing credit limits and contracting long-term assets of over 24 months. On this basis, the risk performance of our new loans has been continuously improving since November 2023 and the vintage loss in this quarter is expected to decline by about 15% compared to Q4. At the same time, it can be seen that day 1 delinquency rate and 30-day collection rate of our overall loan portfolio has also improved marginally in this quarter and the momentum will continue in April and May. At present, the risk optimization work is on track, so we will maintain current credit standards, which will be largely stable going forward.Here, we would like to emphasize that since second half of 2023, we have been very clear that our company's strategy is to pursue quality growth. Under this strategy, we will not take the overall growth rate of our loan volume as the primary goal, but to pursue quality growth as our goal. For ineffective loans, our loans with negative or marginal returns, we will optimize those kind of loans. For business with healthy profit, we will continue to invest for growth. For example, the embedded finance will continue to be a focus of our growth this year.The loan volume of the top 2 channels for embedded finance in Q1 increased by 8% and 12% respectively sequentially, which is far higher than the overall loan volume growth. At the same time, the ROA of our embedded finance model also improved by about 1 percentage point in Q1 sequentially. We will continue to deepen cooperation with the quality channels and replicate embedded finance model to more [Technical Difficulty].Through this structured growth the loan volume [Technical Difficulty], our profitability will be steadily enhanced and we are confident in fulfilling our profit guidance. Recently, we have seen the government introduced a series of policies to support the real estate industry. We believe it will play a positive role in stabilizing the real economy and it also helps virtue the gradual recovery of users' confidence. So we will continue to observe the trend of macro economy and user demand and adjust the loan pace in a timely manner.

[Yan Zheng, executive]

Okay. I will do the translation. Regarding the take rate, I want to share some color with you. Since the second half of the last year, we have made adjustments to our business strategies, emphasizing more on the overall profitability of our business and have achieved good results. Our take rate has increased from 3.2% in Q4 to 3.5%. If we exclude the impact of the withholding tax on dividends and buybacks, the increase in the operating profit margin is even more.Next, we expect the take rate in Q2 to be further optimized with the main driving factors being the first one is the risk optimization. By cutting back business with lower or negative margins, we enhanced the profitability of the overall loan portfolio. The linkage loss of new loans in Q1 is expected to be roughly 15% optimized compared to Q4. In addition, we will continue to improve the efficiency of the collection process and we expect further optimization of the risk indicators in Q2.Second one is the funding cost. This year, market liquidity is still ample and we have increased the issuance of the ABS. This quarter, we issued 5.3 billion ABS. At the same time, our funding costs for cap-heavy loan facilitation continues to decline. With these 2 factors combined, our funding costs in Q1 have decreased by about 70 basis points sequentially. We expect to further reduce our funding costs by issuing more ABS and optimizing our funding structure going forward.The third one is about asset distribution. By introducing more financial institutions and matching different assets according to their risk appetite to enhance the overall commercial rate, while also improving our own profitability. This quarter, the loan volume of ICE has further increased and the take rate of ICE has also been optimized by 76 basis points compared to the same period last year. It is expected to maintain at its level in Q2 and going forward.We also empowered the business through artificial intelligence, such as our AI development tools with a code adoption rate of 20%, which can improve our development efficiency by roughly 30%. We have also used large language models to empower our staff in the collection and telemarketing operations to improve the efficiency in user communication. We have seen some benefits in these testing errors and we will continue to invest in this direction in the future to continuously improve our operational efficiency. Based on the work we have been doing so far, we expect the take rate to be further optimized on the basis of 3.5% in the future.

[Operator]

We'll now take our next question. This question is from the line of Emma Xu from Bank of America.

[Emma Xu, analyst]

[Interpreted]So I have 2 questions. The first one is about your asset quality. So some of your leading risk indicators have already stabilized in first quarter. So can we continue to see such trend or even more significant improvement in the second quarter? And about the 90-day delinquency rate, you explained earlier that the significant increase is partly driven by lower the loan balance. However, even excluding these factors, the 90-day delinquency rate still increased quite significantly. So when will we see the improvement in these metrics?The second question is about your buyback. You mentioned earlier that you execute the share buyback expense at a pace faster than the [ time's ] value. So do you expect to continue to maintain as such repurchase pace?

[Yan Zheng, executive]

[Interpreted]Okay. I will do the translation. As Haisheng just mentioned, that vintage loss has declined by roughly 15% in Q1 sequentially. As for the overall loan portfolio, the 30-day migration rate of days past due 30 plus for Q1 has been -- has seen a 4% reduction compared to Q4. With our continuous efforts in April and May, it is expected to be further optimized by more than 8% in May on the basis of Q1.The main work driving the optimization of our early risk indicators, including the following 3 aspects. First one, the risk strategy and the credit line optimization. By building the model in conjunction with the third-party, we have improved the performance of the risk model to identify high-risk transactions.Second, in the post loan operation, we established a new self-operating principles and procedures and a collection partner management method in March. We optimized the case collection algorithm and commission mechanisms for different collection teams. Promoting the internal and internal collection teams to invest in higher quality collection resource to improve their collection rates for both front and back end. It is expected that the 30-day collection rate will improve by 80 to 100 basis points in April and May compared to 85.1% in Q1.Third, we're focusing on upgrading the post loan repayment infrastructure. We have made a lot of optimizations in both the front and back end, including externally expanding new deduction channel trends and a method to improve the coverage and the success rate of the repayment deduction. And internally, we have also optimized the polling algorithms and the [ time limit ] of the repayment deduction. In this way, we have ensured better collection efficiency, while maintaining our customer experience.In addition, we have also achieved the Group results in the back end collection, especially in the application of legal collection methods, such as litigation, property preservation and lawsuits, which has resulted in an actual recovery amount of additional RMB 170 million in the first 4 months of this year compared to the same period last year. We have further increased investment in legal collection resources this year.Regarding the increase in the 90-plus overdue rates this quarter, the core reason is that in the process of derisking, we will guide the tail-end customers out, which leads to a 16.5% decrease in the denominator of the statistics. The decrease in the denominator is directly [indiscernible]. Since this indicated mismatch and the delay in time, it doesn't reflect the real-time trend of our risk performance and we will recommend that we can pay less attention to it.If we continue to do this action, it will lead to a further fluctuation in the denominator and the 90-plus overdue rate will fluctuate as well. But the actual result is we are doing the derisking and our risk performance is improving. So we suggest focus more on indicators such as the day 1 delinquency rate, 30-day collection rates and 30-day migration rate that can truly reflect the current risk situation, which are all continuously improving.

[Zuoli Xu, executive]

And regarding the share buyback, as we mentioned in the prepared remarks, we are ahead of the schedule. If you do the math, $350 million buyback program over 250 trading days for a year, which averaged about USD 1.6 million per day -- per trading day. So far, we are running at about close to USD 1.9 million per trading day. So about 20% ahead of the schedule right now. And the reason we do that more proactively is because we still view this as a very attractive valuation. We still believe this is a good investment for our cash. And at this point of time being, we will continue to maintain a relatively faster pace than the time schedule.Next question, please.

[Operator]

We'll now take our next question. This is from the line of Alex Ye from UBS.

[Xiaoxiong Ye, analyst]

[Interpreted]My first question is on the funding cost outlook. So management has mentioned that there is clearly a room for improvement in the second quarter. So I'm just wondering, in terms of the market here, how much should we expect the funding cost to further improve in the second quarter?And then secondly, in terms of the early prepayment ratio, it increased quite a bit in Q1 last year. And now given the credit demand appeared to be relatively weak at the moment, so does it have any impact on your prepayment ratio as well? How is the current prepayment ratio running compared to last year and last quarter?

[Haisheng Wu, executive]

[Interpreted]Okay. I'll do the translation. First is about our funding costs. As I discussed earlier, the demand for assets from financial institutions remained very strong this year and we'll remain competitive at funding side. We expect that the cost of funds in Q2 will continue to decline due to one hand our funding cost for cap-heavy loan facilitation and ABS issuance will continue to decrease. And on the other hand, the proportion of ABS in the South will be further increased. Second is about early repayment. This year, the Central Bank has placed a greater emphasis on guiding the balanced allocation of credit, enhancing the stability and the sustainability of the overall growth of the credit. So the market environment has a better impact on the prepayment than the same period last year.We have also taken many measures to control the early repayment ratio. First one, we control the issuance of turnover for active users in terms of our operations. Second, we also predict users' early repayment tendency based on our algorithm and enhance our credit offering on a timely manner to manage the early repayment ratio. Therefore, the 7-day and 30-day repayment rate in Q1 this year are basically the same as the Q4 last year, but has decreased by approximately 15% compared to the same period last year. And we expect our early repayment ratio will maintain stable going forward.

[Zuoli Xu, executive]

Okay. Operator, let's take the final question for the day.

[Operator]

The final question is from the line of Yada Li from CICC.

[Yada Li, analyst]

[Interpreted]I will do the translation. My question today is regarding the loan structure. By the end of this year and going forward, can you give us more color on the breakdown of volume percentage for capital-light and ICE in platform services? And furthermore, how to view the mix change for guaranteed and self-funding model in credit-driven services respectively? In addition, what are the causes for the potential change? And that's all.

[Haisheng Wu, executive]

[Interpreted]Okay. I will do the translation. I want to share some of our consideration on the asset mix. Regarding the asset mix, we do not have a specific target, but instead, we want to balance our risk and profitability by optimizing the asset mix. At the same time, we want to enhance our take rate by improving the efficiency of asset distribution. At a different stage and under different market conditions, we choose different asset portfolios. Currently, we maintain a relatively balanced asset structure, which makes our business healthier with better risk performance compared to the company that are 100% asset-heavy and better profitability compared to the company that are completely asset-light.This quarter, we increased the proportion of on-balance sheet loans, mainly driven by more ABS issuance. Our ABS grew by 130% compared to the same period last year and the cost of funds for ADS is significantly lower than the capital-heavy loan facilitation. Therefore, increasing the portion of on-balance sheet loans is beneficial for improving our overall take rate.Additionally, our ICE model has also increased this quarter, mainly because we introduced more financial institutions and optimized the efficiency of asset allocation. Our take rate from ICE has increased by more than 70% year-on-year. So the increase in ICE proportion has positive impact on our profitability. Going forward, we will continue this strategy for asset allocation, improving operational efficiency under different models and a better balance risk and returns by dynamically adjusting the asset structure.

[Zuoli Xu, executive]

Okay. Thank you, everyone, to join us for this conference call. And if you have additional questions, please contact us offline. Thank you very much. Have a good day.

[Operator]

Thank you. That does conclude the conference for today. Thank you for participating. And you may now disconnect.[Portions of this transcript that are marked [Interpreted] were spoken by an interpreter present on the live call.]