BWS Capital Partners, LLC

Investment Conclusion

In this week’s “Spotlight,” we posit that after six years of share price and P/E convergence our universe is potentially entering a period of divergence creating the opportunity for stock pickers to show their stuff. We look back at past cycles of divergence and convergence to see what factors have had an impact and which companies have moved the most as a result.

Summary

We look at three historical periods – two of price and P/E convergence (1992-1994, 2001-2005) and one of divergence (1997-2000). The first period was less of a stock pickers’ market as the group traded more closely together on industry themes. The next period saw a reversal of this as an industry call was much less profitable as investors awarded higher valuations to certain banks. In the third period, the value of stock picking appears to have diminished again. However, we have seen some evidence of widening in 2007 as business has been a differentiating factor. Looking out, we believe investors will pay more attention to competitive positioning and balance sheet strength.
The trend of banks’ P/Es has also tended to follow this trend. Most recently, the data show that multiples are currently near the narrowest band since 1995–96. We expect the market to continue to differentiate as the competitive environment remains challenging, loan demand wanes in some categories, and as asset quality inevitably deteriorates. Looking out, we expect multiples to diverge as performance widens.

In this week’s “Spotlight,” we posit that after six years of share price and P/E convergence our universe is potentially entering a period of divergence creating the opportunity for stock pickers to show their stuff. We look back at past cycles of divergence and convergence to see what factors have had an impact and which companies have moved the most as a result.

Stock Picker’s Market vs. Industry Call

We look at three historical periods – two of price and P/E convergence (1992-1994, 2001-2005) and one of divergence (1997-2000). The difference between the average stock price performance of the top quartile and the bottom quartile for the top 50 banks at the start of each year (including acquired companies) shows how in the first period the difference went from 75% down to 27%, reflecting a narrowing differentiation between banks. This period was less of a stock Pickers’ market as the group traded more closely together on industry themes. The next period saw a reversal of this as the disparity went from 30% at the end of 1996 to a high of 68% in 2000. In this period an industry call was much less valuable as investors awarded for picking individual stocks. In the third period we look at, the value of stock picking appears to have diminished again, as the spread between the top and the bottom quartiles has consistently narrowed since 2000. However, we have seen some evidence of widening in 2007 as business mix has become a differentiating factor with mortgage in particular under pressure. Looking our, we believe investors will pay more attention to competitive positioning and balance sheet strength.

We also divide our current universe into quartiles by forward P/E ratios (the lines), and examine the difference between
the top and bottom quartiles (the bars). The trend of this graph tends to correspond with the movements of share prices. Most recently, the data shows that multiples are currently near the narrowest band since 1995–96. We expect the
market to continue to differentiate competitive environment remains challenging, loan demand wanes in some categories, and as asset quality inevitably deteriorates. Looking out, we expect multiples to diverge as performance widens.
In the 1992 to 1994 timeframe, all companies saw P/E deterioration, but to varying degrees. The biggest erosion came from CYN, STT, FITB, NTRS, and BK, all of which saw a decline of more than 4 multiples. BOH, NCC, TCB, HBAN, and ASBC held up the best, both with less than 2.2 multiples. his period was characterized by rising loan demand, an overall flattening of the yield curve, improving credit quality with declines in NPAs and NCOs, muted M&A activity (factors described in more detail later). Of note, in this period the names that had the highest multiples going into this period came down a fair amount while single-digit P/E names improved. Fee income was rewarded as well due to its lower capital intensiveness, less credit sensitivity, and higher growth. Between 1997 and 2000, NTRS, STT, BK, CBH, and FITB all witnessed improvement of more than 10 multiples as several regained their premium valuations that were lost in the prior cycle. STI and BOH saw more than 2 multiples of deterioration. The divergence that marked this period had 22 banks generating multiple expansion (55%), while 17 saw multiple erosion. This period was marked by greater fluctuation in loan demand, flat NPAs, somewhat higher trending NCOs, and more variation in the shape in the yield curve. In the second period of convergence, 2000 – 2005, SIVB, UB, CNB, and BOH all witnessed improvement of 4 multiples or better as it appears better geographies were rewarded. NTRS, FITB, STT, SNV, and BK saw more than 10 multiples of deterioration, as less value was placed on fee income as the yield curve steepened. In fact, of the 24 banks that saw multiple erosion in this period, the vast majority (18, 75%) saw multiple expansion in the previous, divergent period. Of the 19 that had expansion, 13 (68%) declined previously. This last period was marked by a steep drop in loan demand for one year followed
by a few years of a rising trend. There was a steeper yield curve for most of this period with a flattening in the back-end, while M&A was more muted. Both NPAs and NCOs trended higher in 2001/2002 before dropping to historic lows.

Contributing Factors

Interest Rates – Not surprisingly, a contributing factor to the cycles of convergence and divergence is the interest rate/yield curve environment. Note, we use the spread between the 13-week and five-year treasuries as proxy for the yield curve. In period 1992-1994 (I), the Fed Funds rate ended up 150 basis points, while the yield curve initially saw a flattening followed some steepening and another spate of flattening. From beginning to end though the trend was a fairly significant flattening. Loan growth steadily moved higher, corresponding with a period of convergence as it was easier for the banking industry to book credits. All categories of loans grew in the period with consumer lending improving the most, followed by C&I then and real estate.
In period 1997-2000 (II), the yield curve saw major flattening twice and one extended steepening. Still, the overall trend was towards flattening. The Fed Funds rate was up 125 basis points in the period. Overall, loan growth improved. Consumer lending rose, while C&I and real estate were more flattish. Investors awarded different valuations to banks as the lending environment and interest rate positioning affected competitors to various degrees.
In period 2000-2005 (III), the Fed Funds rate was lower by 225 basis points as in 2001 alone it was cut 475 basis points.
Overall loan demand saw dramatic weakening in 2001 to near zero, followed by impressive growth to the double-digit level
thereafter. C&I saw dramatic loan growth improvement, while real estate was higher and consumer deteriorated. With rates going for most of this period, investors took a generally positive view of the banking group and differentiated less causing a convergence of valuations.
In sum, it appears that fee income is valued more in periods with a more challenging yield curve and less so when the yield
curve undergoes significant steepening.
Credit Quality – It seems that periods of convergence reflect improving credit quality trends as a better environment makes it easier for the industry as a whole. On the flip side, divergence is more prevalent when better competitors shine in a tougher environment. In period I, NPAs fell from 4.50% of loans to 1.50% and NCO went from 1.60% to 0.40% as convergence was the order of the day. In period II, NPA generally were flat, while NCOs rose from 0.57% to 0.68% providing an opportunity for differentiation. In period III, a period of convergence, saw NPAs rise from 1.15% to a peak of 1.21% before trending steadily down to a low of 0.49%. NCOs declined from a peak of 1.11% to 0.55%. These decreases in metrics aided the industry driving convergence.
M&A – Interestingly, it appears that periods of divergence are characterized by increased M&A thus rewarding stock pickers, while the opposite is also true. In the first period of convergence, 1991 to 1994, there were 7 bank acquisition of greater the $1 billion or an average of 2 per year. In the next period of 2001 to 2005, a period of divergence, there was a much greater 36 such deals for an average of 9 per year. The last historical period was another convergent, industry-call-oriented time and was characterized by a lower amount of M&A with 27 deals (5 per year). Figure 10-11 indicates this trend going back to 1991.
If this trend persists we may be entering a heightened M&A environment in the banking space over the next few years. In
fact, year-to-date in 2007 we have seen 13 deals closed or expected to close which is more that the average for both periods of convergence and above the 1997 to 2000 period of divergence. In our view, we expect to see this trend materialize.