2006 Economic Outlook Roundtable:
Scenarios for the Next U.S. Recession

MR. BROWN: Now I'm going to turn to Art McMahon.

MR. McMAHON: Thank you very much, Rich. It's a pleasure to be here. I appreciate your invitation to join the panel. And I think my presentation will actually echo some of the things that you said earlier. You mentioned that we have a lot of discussions and share our views, and that may be reflected in some of the echoing that you're going to hear today.

And I'll also amplify some of the points that you made, including the idea of the rolling regional recessions.

I'm going to focus on three main things. First, I'm going to focus on where we are now and dig into net interest income. I would like to illustrate how important changes in margins can be and how they drove earnings in the last recession and right after the recession.

And then, obviously, there is the issue of housing. I think I have to talk about housing, because it has been the big driver for both bank earnings and assets. I will also take a look at what has happened in the past, picking up on the theme of rolling recessions that Rich mentioned in his opening remarks.

We have not seen a decline in national home prices for quite a while, but we have seen some regional corrections in home prices. And I'd like to take a look at some of the areas where we've seen that kind of correction. I will especially focus on California in the early 1990s, where we had a lengthy period with an impressive nominal decline in prices. I will take a look at what happened to banks that were operating in California at that time.

And then finally, since the theme is scenarios for the next recession, I want to spend a little bit of time talking about the main factors that are likely to be affected if we do see a significant slowing in the economy. And, again, I think you'll see that I will echo some of the points that Rich made.

Turning first to earnings, here is our chart illustrating commercial bank return on equity (see Chart 25). It is similar to the chart that Rich showed you before, though I would ask you to focus on the fact that earnings over the last decade or so have been extremely strong relative to the previous period. From the early 1990s on we've had a series of years of very strong earnings, consistently higher than the preceding 20 years.

Chart 25

d

[Chart 25] includes two green dots in the right upper corner. They represent our attempt to adjust the return on equity for the accounting treatment of two big mergers that happened in 2004. Those mergers affected the equity reported at some of the big banks, so they affect the aggregate numbers as well. So when you make that adjustment, you see that there's been some decline in ROE, but it's pretty modest. Earnings continue to be quite strong—just under 15 percent after that adjustment—for the system as a whole.

And in the little inset [for Chart 25] you see the quarterly numbers—the third quarter of this year compared to the third quarter of last year. And it's a similar story: a modest slowing, but continuing, very strong earnings. Obviously, that puts banks in a good position.

I want to spend a minute or two going through some of the dynamics of net interest income, because it is the single biggest item for bank earnings for most of the banks. I want to break down the components, one of which is the price—the net interest margin, measured on total assets.

We had a pick-up in the net interest margin in late 2001 and into 2002, and it stayed up throughout most of 2002. The economy was weak then, and the Fed's response led to a decline in short-term rates. This translated into a widening of spreads. A big increase in core deposits also helped to support margins. The resulting shift in the net interest margin was a major positive for net interest income.

Since that time, however, margins have been quite weak. What has offset the drop-off in margins has been very strong volume growth for commercial banks over the last three years or so (see Chart 26). The 20-year average rate of asset growth for the industry is about 6 ¼ percent. Commercial banks experienced a drop-off in asset volume growth to below-average levels but a pick-up in net interest margin when the economy slowed in 2001. This led to a big increase in net interest income, as the impact of the margin gain more than offset the deceleration in asset growth. As the economy subsequently strengthened, these forces reversed, and we've had an increase in volume and a drop-off in margin since then. The continued expansion of net interest income has largely been dependent on banks maintaining above-average levels of asset growth.

Chart 26

d

And as Kathleen mentioned, a lot of that volume has come from housing. Since the middle of 2002, residential real estate loans including home equity and single family mortgages have consistently grown at double-digit rates (see Chart 27). In contrast, the expansion of other loans has been less robust. I should mention that securitization does affect these numbers a bit, although it does not alter the fundamental story. If you combine mortgage lending with securitization activity, the main impact is to smooth the quarterly growth numbers out a little bit.

Chart 27

d

Given the importance of housing, both to the economy and to banks, I want to take a minute or two and talk about what past experience tells us about the housing market. First of all, as Kathleen I think implied, there is an expectation that most analysts expect a cooling of the market, but not a steep drop in nominal home prices. However, a number of analysts believe that regional housing price declines are possible. We're going to take a look at some declines we've seen in the pastespe—cially in California, where we've had a series of years when home prices fell. These declines don't occur in a vacuum; they tend to be associated with a contraction in the economy and some loss of employment. The home price decline tends to exacerbate the contraction once you have that kind of economic weakness develop. Once significant home price declines occur in an area, they tend to last for a while, and it takes some time to get back to the home price level that existed at the peak and for employment to recover.

We will also look at the experience of banks in California during this multi-year period of home price decline. What we will discover from a review of their experience is that it wasn't the residential real estate loan book that deteriorated the most. Rather, it was other kinds of lending. I think that speaks to the connection of housing to the overall economy. If the housing market is weak, the region is likely to experience softness in the job market. That means there's less small business activity going on, and there's more stress from C&I [commercial and industrial] loans and other parts of the portfolio.

Let us turn first to where we've been. When we look at the OFHEO home price index for the last 30 years, since the inception of the index, we see that we've never had a decline in nominal prices nationwide (see Chart 28). But we have seen periods in which home price growth has been positive but rather weak, as in the early 1990s. This is similar to what many analysts are projecting about home prices over the next few years.

Chart 28

d

For example, projections from Economy.com, which are consistent with the views of a lot of analysts, suggest a deceleration of national home price gains to 5 percent in 2006. In 2007, they expect price growth to drop to about 1.9 percent.

So the general expectation is for a flattening of home prices but not a steep decline in average nominal home prices. We know that even though we've never had a national home price decline, we have had a series of regional home prices declines. This relates to the concept of rolling regional recessions.

A number of major MSAs have had significant drop-offs in nominal home prices in the past. For example, both Houston and Los Angeles have experienced 20 percent drops in nominal prices (see Chart 29). In New York home prices dropped by over 10 percent in the 1990s.

Chart 29

d

Moreover, it took a long time for prices to get back to their previous peaks in those MSAs. For the three MSAs already mentioned plus San Francisco and Boston, which also experienced periods of nominal price decline, it took at least six years before home prices got back to their previous highs.

As I mentioned, home price declines tend not to occur in a vacuum. Typically, the regional economy is weak as well. That was certainly true in the case of all five of the MSAs that have been mentioned. All of them experienced sizable job losses at the same time that home prices were falling. In all five cases, there was a loss of over 7 percent of total jobs. Again, as with home prices, it took a while for those jobs to be completely recovered.

The message we take from this is that you can experience regional corrections in home prices even though you're not seeing a decline on a national basis. You could argue that California has been the poster child for this situation. I noted that both Los Angeles and San Francisco have experienced nominal home price declines. The problems were broader than that, and beginning in 1991 and lasting until 1997, nominal home prices declined statewide in California (see Chart 30).

Chart 30

d

Those home price declines were associated with a slowdown in the overall California economy. The state lost about 700,000 jobs, many of them the result of defense spending cutbacks. At the end of the Cold War, the federal defense budget was trimmed, and this particularly affected California.

Let's take a look at what that meant for banks in the state at that time (see Chart 31). We confine our attention to commercial banks in the state with assets under $1 billion. The idea is to avoid including banks that were operating in multiple states, for which it would be hard to accurately determine where their exposure was concentrated.

Chart 31

d

Let's look at the peak charge-off rate by loan type for the period of time when home prices were under pressure. One interesting finding is that residential real estate, which includes one-to-four family mortgages plus home equity loans, had the lowest rate of charge-offs. Charge-offs were much more significant in C&I loans, which for banks of this size are generally small business loans, as well as in construction, commercial real estate and consumer installment loans. These results reinforce our sense of the connection between the housing market and the overall economy and points to the likelihood that weakness in one is likely to spill over into the other.

Next, I wanted to talk briefly about the impact on banks if we do see a significant slowing in the economy. I must say, it was very comforting hearing your perspectives on things, Kathleen. It sounds like it's a pretty good environment for banks going forward.

MS. CAMILLI: Well, I'm usually an optimist.

MR. McMAHON: Just in case things turn out to be a little bit grimmer than that, let's take a look at what that might mean for banks. And again, I'll think you'll hear echoes of what Rich said. There are several factors that could impact banks in a slowdown, the biggest factor being deterioration of credit quality and the impact that has on provision expenses for banks.

You also could expect to see a slowdown in volume growth when you have weakness in the economy, and that would probably also hit non-interest income as well. But there are some positives, typically from softer short-term interest rates. Margins may well increase in this environment, as we discussed a few minutes ago. As we noted, during the last recession the improvement in margins provided a big boost to net interest income during the last slowdown, more than offsetting the impact of slower asset growth. You would also see gains on the sale of securities and other holdings that are sensitive to interest rates and rise in value as interest rates fall.

First of all, banks are now at all-time lows for levels of non-current loans (see Chart 32). So we're starting with a very strong credit quality position for the banking industry. The most recent recession bumped non-currents up a bit, but they remained well below levels seen in the early 1990s. It is also important to note the distinction between large banks and smaller banks with respect to the deterioration in credit quality in past periods of economic weakness.

Chart 32

d

Large banks tend to have much bigger swings, and the 2000 recession and deterioration of credit quality was heavily concentrated at the larger banks.

So we're starting in a better position than we were, let's say, in the late 1980s. That's the good news. The bad news is that banks have been able to hold down provision expenses because credit quality has been so strong (see Chart 33). That has been a key positive for them in terms of their earnings. That also means that loan loss provisions are likely to go up if there is any fall-off in credit quality.

Chart 33

d

In fact, in the last quarter, dollar provisioning rose for the first time in about three years. Many analysts believe we are at a turning point and provisioning will start to be a drag on earnings going forward, even if the economy continues to perform well. If there's a deterioration in credit quality, it will require more of an increase in provisioning, which will, of course, cut into net income even more.

However, there has also been a significant shift in the portfolio over the last 10 or 15 years that may act to benefit banks overall credit quality (see Chart 34). Not surprisingly, the importance of housing has increased.

Chart 34

d

Residential real estate lending is now about 30 percent of total loans for commercial banks. That is an increase of 10 percentage points. And C&I loans are now under 20 percent as a share of the loan portfolio. So there has been a shift away from what have been relatively riskier loan categories. If it continues to be true that residential real estate loans outperform business lending in an economic downturn, this shift will help limit the need for increased provisioning at banks.

And then, to echo a point that Rich made earlier, there does tend to be a drop-off in the volume of lending and asset growth generally when the economy slows (see Chart 35). This has been the pattern in recessions going back to 1950.

Chart 35

d

As I mentioned, there are some positives in the picture. If you do get a softening of the short-term rates, the yield curve will get steeper. Usually core deposits pick up in that kind of environment, and that allows banks to pay off some of their wholesale, more expensive borrowing. Both of those things tend to increase margins. As I noted earlier, the rise in the net interest margin provided a big boost to net interest income during the last recession and its aftermath.

And, finally, there are gains on interest-sensitive assets if interest rates come down as economic activity slows. That doesn't tend to fully offset the credit losses in most cases but does provide a cushion.

So, in summary, we're in a very strong position. Earnings have been high, and banks are in a good position now. Housing has been the big driver of growth. Analysts do think there's a possibility of some regional housing corrections. If so, the message is that it's not just the residential loans that banks need to worry about. It's other kinds of loans as well.

If the economy does slow sharply, which is certainly not the consensus here, we will see negative pressure on bank income. Fortunately, banks will be starting from a very strong position, and some of the impact will likely be cushioned by softer interest rates.

MR. BROWN: Art, thank you.