Banca Sistema — Loan growth and ROE are positive features

BST reported factoring turnover 25% ahead of Q316 and factoring receivables outstanding were up by 19% to €1,255m. Factoring origination agreements with banks have contributed to this growth and BST has added a further three banks to the list, taking the total to 17, which now account for about 33% of turnover. Exhibit 1 shows the progression of factoring turnover and receivables outstanding since the beginning of 2015.

Salary and pension-backed loans outstanding (€423m) also showed very strong growth, nearly doubling compared with the prior year figure, and market share now stands at over 2%. The pre-tax profit for the third quarter was roughly double the prior year period. This was boosted by an increase in the rate of accrual of late payment interest, reflecting the accumulation of a longer run of data that resulted in a rise in estimated recovery rates (see further comments below).

We summarise other key points from the results announcement below, with comparisons between Q317 and Q316 unless stated.

Exhibit 2 shows the P&L for Q317 and the previous three quarters.

The group adopted a change in accounting for late payment interest (LPI) in June 2016. Previously, BST had only recognised this on a cash basis, but following the change it accounted for a portion of late payment interest accrued (in common with competitors). The policy reflected the adoption of a statistical model based on collection experience and is only applied where legal proceedings have begun. As noted above, with the Q317 figures BST changed the accrual rate for LPI. This reflected the collection of additional data points that feed into the statistical model. For receivables from the national healthcare system, the accrual rate is effectively unchanged at c 65% of qualifying late payment interest, while for other public sector receivables the percentage has increased from 15% to 31%. The average accrual rate of c 38% is still significantly below the 80% recovery rate that has been experienced. The change in rate resulted in the €9m increase in LPI mentioned above, of which €7.8m related to periods prior to the third quarter.

It is worth underlining that BST typically does not pursue a legal route to collect amounts owing and related interest, although it will do in some cases (c 10%). BST therefore differs from some competitors whose model is to focus on legal collection from the outset. However, because of the length of this collection process, the proportion of outstanding receivables under legal collection has risen to c 30% so LPI, both in the form of accruals and cash collection, has become an increasing part of total interest income.

Prospectively, factors that would affect the level of LPI in any period would include success in collection (positive for cash collection but reducing the outstanding level), the level of receivables entering legal collection, the mix of receivables in legal collection by sector and maturity, any sale of receivables, and any change in accrual rate generated by the output from BST’s statistical model.

Turning to look at trends in the overall yields in the factoring area, Exhibit 3 shows the average gross yield including interest income and fees and commissions. BST has calculated these numbers on the new LPI accrual basis and, while 2016 numbers exclude Beta Stepstone, they are broadly comparable. The yield has fallen slightly over a year but has been largely stable within the period shown, with the exception of Q117 which was affected by the lower level of turnover experienced at the end of Q416. Exhibit 4 shows the yield at the target collection date at the time receivables are taken onto the book. Here the yield again includes fees and commission. Actual returns are usually above the target level because collection is predominantly made before the date used when setting pricing. Where collection is delayed and moves into legal collection, accrual of LPI (and eventual cash collection of the amount of LPI outstanding) compensates for the delay enabling BST to generate a return similar to or above the original target return. As Exhibit 4 illustrates, the target yield has been within a narrow range over recent quarters. The funding period has stepped up significantly in the third quarter, reflecting a single healthcare transaction of significant size with a long duration (accounted for in the pricing).

Turning to loan loss provisioning, the third-quarter P&L charge was, as noted, down by 9% compared with Q316 and stood at 48bps of average loans versus 58bps on an annualised basis. The provisioning level does fluctuate between quarters driven by writebacks and other factors, so we assume a modestly higher level in our estimates for the next two years (50bps). The level of net non-performing loans as a percentage of customer loans was similar to Q217 and FY16 at 9.5% compared with 9.7% and 9.2%, respectively. As the majority of the past due segment relates to normal course receivables business, the more relevant percentages are for bad loans (1.8% and stable) and unlikely to pay (1.0%, lower sequentially).

The acquisition of 19.9% stakes in ADV Finance and its subsidiary Procredit for a combined consideration of €0.8m was announced with the results. This purchase will add to the origination network for salary and pension-backed loans and if ADV reaches its own stretching target could generate incremental loans of €50m. Completion is subject to authorisation by the appropriate authorities. Chief Executive Gianluca Garbi noted on the results call that there has been increased discussion of the potential reduction in the risk weighting applied to salary and pension-backed loans. If there were a reduction from the current weighting of between 60% and 75% to c 35%, for example, then it would become a more attractive business to expand and loans would be likely to be retained rather than securitised and sold (subject to market pricing). In these circumstances, BST might consider becoming involved more fully in the origination of these loans including increasing its stake in ADV. While this would involve an investment, it would be justified by an increased share of the overall margin generated.

The changes in our estimates are summarised in Exhibit 5, with a tempering of our assumed interest margin trimming operating income in FY17 and FY18. For FY19 we have also removed our previous assumption of a securitisation sale; this removes the related profit on disposal, which is partly offset by increased interest income. With limited changes in operating expense estimates there is a larger impact at the EPS level, with reductions of between 3% and 15%. The latter is for FY19 and excluding the securitisation sale impact, the reduction would be in line with that for FY18 at 11%. We note that it is possible that our interest margin assumption may now prove conservative with the potential for LPI collection/accrual to outpace our forecasts, while customer loan growth in the salary and pension-related area could accelerate if the capital weighting in this area is reduced.

We have updated the charts (Exhibits 6 and 7) showing our assumptions for customer loans outstanding and the yield on customer loans. The first shows slightly lower assumed growth for factoring balanced by increased estimates for salary and pension-backed loans including removal of assumed securitisation disposal for FY19. The overall yield on average customer loans is broadly stable on our estimates with the slight downtrend apparent in Exhibit 7 mainly reflecting mix change towards salary and pension-backed loans.

We have updated our comparative valuation table (Exhibit 8) showing BST in the context of selected peers involved in factoring, debt purchase, debt management and collection. BST trades on the lowest prospective P/E multiple and the highest yield within the group.

In terms of price to book ratio, BST is also well below the group average at 1.5x compared with 2.6x despite earning a return on equity of c 20%. Our next chart, plotting return on equity against price to book, highlights this cautious valuation. To position BST in line with peers on its current price to book ratio we would have to assume a return of equity between 10% and 15%, well below the returns we expect for FY17-19.

In our ROE/COE valuation we have used an assumed return of 20% (22% previously) to reflect the more conservative forecasts outlined in the previous section. Other assumptions are unchanged (including long-term growth of 4% and cost of equity of 12.4%). The resulting valuation is €2.80 compared with the €3.15 indicated in our last note. This is still 26% above the share price at the time of writing (€2.22).