The economy seems to be improving a bit, and the one thing almost everyone can agree on is that small businesses will be a significant part of the recovery. Those small businesses need working capital to grow, open new locations, hire more employees, buy inventory, advertise and initiate the other investments that make them the primary engine of our economy. At the same time, banks are faced with new regulations that threaten to cut or eliminate many of their traditional revenue and profit lines.
They want to grow their loan books to generate the fee and interest revenue, the cash management revenue and the merchant processing revenue a growing small business loan portfolio provides. This seems like a perfect scenario; millions of customers looking to banks for loans the banks are eager to provide them. But something is wrong. Banks are not making enough loans to satisfy the need; small businesses are getting angry and sharing their anger with their elected officials, and small businesses are turning away from their banking relationships and looking for alternatives. Why isn’t the system working the way we need it to?
Risk is the reason things are not working; specifically, risk assessment in underwriting and risk management through the loan’s life cycle. This can be explained by the fact that the environment has changed. Today it is more dynamic, with new impactors greatly affecting small business risk scoring models that were originally influenced by commercial lending models built to assess larger companies and then adapted for small business lending. What were believed to be traditionally strong predictors of risk are now less strong and increasingly misleading, as evidenced by the increase in second-look approval processes.
Underwriting uses owner personal credit score and ratio modeling as risk indicators, yet they don’t predict risk as well as other indicators. But, these indicators and models are new to underwriters, and new means risky. Instead, the typical response to risk is to tighten up existing indicators — like years in business or owner credit scores. Servicing models use monthly billing — net 30 and no adverse action until net 60+, which give them too little data to adequately manage their risk levels. The banks are forced to decline more applications to achieve the risk profile they need. Small businesses are stuck looking elsewhere for capital that was needed yesterday.
Banks need to move faster. They need to change the basics of the way they look at the small business risk world and invent new solutions to risk assessment and risk management problems. Banks are faced with a hostile regulatory environment, loss of interchange fees, falling property values and a tough economy. Those that figure out how to lend to small businesses in greater numbers without increasing charge-offs will — in a word — win. Those that do not will lose.
What is needed is a way to say yes with confidence. If a bank says yes with confidence to a small business applicant, it gets not only a good, performing asset, but most of the time also will take the merchant processing business, the DDA and even personal accounts. The converse is also true. If some other bank says yes, your bank probably loses the rest of the business as well.
In a world where social media has put us all in thousands of constant conversations 24/7, it is not surprising that the solution to this underwriter’s dilemma — how to approve more, but lose less — is right in front of us. We all know people who don’t feel the day is complete without getting 25 pieces of data (“tweets”) from Charlie Sheen. Marketing departments are exchanging data with customers and prospects every day through Facebook, Twitter, LinkedIn, email and blogs. Instead, traditional risk management sends a single bill per month and hopes for a data response (payment) that may or may not show up — and this is with people who already have the bank’s money.
The environment has now changed enough, and there is such a substantial body of evidence that alternatives to traditional bank approaches are working, that now is the time that banks need to act.
The Solution: Daily Remittance Model
The daily remittance model is a new way to look at asset structure, risk assessment and risk management. Daily remittance assets are structured to amortize continuously based on payments more frequently than monthly (for these purposes, daily, weekday, weekly or bi-monthly payment or remittance schedules would all be daily remittance). Right away, it is easy to see how risk is lowered as amortization happens more continuously rather than with periods (between traditional monthly statements) of no amortization activity followed by spikes (monthly payments) of monthly amortization.
However, the amortization schedule is not the biggest value to daily remittance. It is the data, and the data works two ways.
First, the data enables the configuration of highly sensitive risk management workflows – fine-tuned to SIC-based, seasonality, weather and any number of variables. Such workflows not only create the ability to assess loan performance on a daily basis, but enable machine-driven determinations of when trouble arises, and machine triggered interactions with troubled borrowers. The stimulus (distressed behavior) and response (servicing intervention) cycle times are driven down to days, rather than months.
Second, the continuous flow of sales, deposit, merchant processing and payment data points enables the development of dozens of very granular risk indicators — both actual and synthetic — that can feed dynamic risk assessment models that are seasoning risk variables on a daily basis, and shifting risk weightings dynamically in response to what is actually happening in the highly volatile small business markets. The result is risk assessment based on risk indicators based on actual sales inflow data — a much more predictive basis for risk decisioning than an owner credit bureau score.
Continuous, Data-Driven, Active Risk Management
Banks accumulate a wealth of data every day. Deposit data, merchant processing data, payment data and more flow into the bank every single day from every single customer. This huge river of data is broken out into separate databases within the bank’s sprawling and often disconnected networks, and there it sits — largely unused. Banks need to harness that data, accumulate it and use it to build asset management workflows that respond to the changes the data show every day.
To illustrate, let’s examine two examples: First, let’s say the bank made a loan to a dentist on Main Street. The dentist has just received his second monthly statement and made the payment on the same day. The next day he decides he isn’t going to make the next payment — we will call that DD day (for default decision). On DD day +30 the bank sends the third monthly statement. The dentist decided 30 days ago he was not paying and does not. At DD day +60, the third monthly statement is probably in some grace period and the bank issues the fourth monthly statement. The dentist does not pay the fourth statement. At DD day +90 the fifth statement goes out, the third statement has now moved to active servicing and the fourth is in grace period. This loan is going down, and the bank has two data points (failure to pay third and fourth statements).
Next let’s look at daily remittance: Let’s say the bank made a loan to a chiropractor on Main Street, but used a daily payment structure tied to the chiropractor’s credit card volumes (a fixed percentage of each day’s credit swipe goes directly to the bank to amortize the loan). We have learned in underwriting (from the merchant processing division) that the chiropractor batches out his credit card terminal every Monday and Thursday like clockwork. About two months into the loan term, on a Saturday, the chiropractor decides he isn’t going to make more payments on the loan and switches credit card processors in violation of the terms of his loan agreement (DD day). The next Tuesday is DD day +3, and the workflow engine takes the daily upload of processing records and does not see a payment from the chiropractor. The bank has configured the system to allow for a missed payment without action to account for the fact that many chiropractors are sole proprietors who miss days for illness, bad weather, school plays, etc. The following Friday is DD day +6 and still no payment is seen from the chiropractor. Now the system automatically populates a risk manager’s workflow with the account and the manager calls the chiropractor with a “soft” call. The chiropractor realizes the bank is watching closely, confesses his loan violation and returns to the bank’s processing fold. In less than one week, the bank has received six data points, the system has analyzed two anomalies and risk management has intervened. This loan is saved.
Alternative finance providers have begun to incorporate today’s real-time data streams into their risk scoring models and risk management workflows. And they are doing it with much less data than the banks already are receiving every day. But they have designed models and workflows that reduce risk substantially while saying yes much more often.
Building a Better Mouse Trap: Dynamic, Data-Driven Risk Scoring Models
Banks can use their vast data warehouses to create better risk scoring. But they need to put those data assets to work every day. On an individual applicant basis, the amount and frequency of bank deposits, or merchant processing information on any given small business seems like a slim reed on which to build a robust, machine driven risk assessment engine. And it is. However, when tens of millions of such data points on tens of thousands of merchants are combined into a machine-learning environment, the result is a highly precise prediction of small business viability and relative level of activity (growing, shrinking, staying the same).
This article is not about the design of risk scorecards, but risk officers will recognize that the systematic analysis of daily cash inflows from tens of thousands of borrowers employing dozens and dozens of actual and synthetic variables for each quickly gains critical mass enabling a highly predictive scoring model. If the bank takes the next step, and can teach its scoring algorithms how to learn based on the daily input of data and volunteer seasoning, deployment and weighting strategies, then an incredibly valuable asset has been created for the entire institution with ramifications for all areas — not just small business lending.
This kind of next generation risk modeling is already in use. Its outputs are startling to traditional risk managers. In one such deployment, businesses with owner credit bureau scores of 500–600 are routinely approved, while those with owner credit scores of 750+ are often declined. And those same 750+ declines are getting auto-approvals from bank credit card departments within 30 seconds of an inbound phone call or Web application.
How Can a Bank Implement a Daily Remittance Model
We know that banks already have the data needed to power a daily remittance model; they just need to start using it. In order to put it to use, they need to develop first the will to use it, and then the systems to use it.
It is commonly said that banks, especially large banks, tend to operate in functional silos. And those of us who know banking know that there are some good reasons for that, which tend to get short shrift — especially among those who have left banking and are looking backward at their old institutions. But all successful institutions have leaders who have the responsibility and authority to get the bank into new markets, products and ways of seeing the world.
The small business lenders at a bank might see value in daily remittance, but cannot interest compliance and IT in the exploration of the model. Or merchant processing or DDA might see value in the cross selling and customer retention that comes from giving a client a loan, but cannot get anyone in small business lending on board. The unit leaders are under intense pressure from every direction. The will to shift paradigms might be there, but they need senior leadership sponsorship and focus.
Once the will is in place, banks can partner, build or buy a daily remittance model or product.
One alternative is to refer declines to a third party; this would enable banks to offer customers needed capital and, at the same time, retain the customer and revenue from other services. This partner option has the advantage of getting to market quickly, and can be done on a co-branded, white label or arms-length basis. It can be done on the bank’s balance sheet or the partner’s or a combination determined through criteria established out at the outset of the relationship.
Banks can build a daily remittance platform themselves. A successful build will require data entry and automated underwriting, configurable and automated risk management/collection workflows, daily amortization accounting, a dynamic risk scoring module and the ability to operate off of a single data platform that pulls data from all relevant systems into a single data architecture to ensure data and output integrity. It will also need to provide all of the risk, accounting, treasury and operational reporting that daily data makes possible. Building the system will require time, labor and cash. Then, once built, the bank will need the will to experience the losses necessary to season the risk scoring indicators, and develop the critical mass dynamic risk modeling requires to become an ever-appreciating asset of the institution.
Banks can find a platform and buy it. This option has the benefits of speed to market and control. However, there are not many systems on the market that deliver the risk management and risk scoring benefits of daily remittance. Most focus on boarding loans or binary-decision underwriting filters. This option obviously requires a financial investment, but could offer speed of implementation.
As the economy improves, banks want to make loans and small businesses need loans. Small businesses will be key to thriving banks, especially in the wake of the current regulatory overhauls touching so many revenue streams of the industry. Traditional risk-scoring and risk-management models will not get the job done for banks and a new paradigm is needed. Daily remittance is one such paradigm allowing small businesses to get the capital they need and banks to make the loans that will drive their bottom lines and customer relationships. Banks that solve the underwriter’s dilemma earlier will win the battle for small business banking. And the datasets that the banks will build will continue to provide value for the next 25 years and beyond.
Glenn Goldman is president and CEO of Capital Access Network Inc., Scarsdale, N.Y. Contact him at www.capitalaccessnetwork.com.
Copyright © September 2011 BankNews Media