As discussed earlier, Treasury yields and stock prices are down sharply since the December FOMC meeting on news that indicated greater odds of a recession and large writedowns at financial institutions. As shown by the blue line in the top left panel of exhibit 6, the fall in stock prices pushed up the ratio of trend forward earnings to price. The difference between this ratio and the real Treasury perpetuity yield, shown by the shaded area and plotted to the right, is a rough measure of the equity premium. As you can see, this measure jumped in the past few months and is now at the high end of its range of the past twenty years. In the corporate bond market, the spread on high-yield corporates, the black line in the middle panel, widened sharply, and investment-grade spreads, the red and blue lines, also rose. Forward spreads (not shown) rose especially in the near-term, suggesting particular concern about credit risk in the next few years. In the forecast, we assume that the equity premium and bond spreads will recede some from their recent peaks as the risk of recession recedes and activity picks up, but they will remain on the wide side of their historical averages. As shown in the bottom left panel, our most recent indicators suggest that the OFHEO national purchase-only house-price index, the black line, fell 2¾ percent in the fourth quarter; we project further declines of about 3¼ percent in both 2008 and 2009. In some states with many subprime mortgages— such as California, Arizona, Nevada, and Florida—house prices, the red line, began to fall earlier and have declined by more. Reports of spectacular writedowns from some financial firms may also have caused investors to assign greater odds of tighter financial conditions. As noted in the bottom right panel, financial firms took writedowns and loan-loss provisions of more than $80 billion in the fourth quarter. Most of the reported losses were from subprime mortgages and related CDO exposures, but many banks also increased loss provisions for other types of loans. In response, financial firms raised substantial outside capital and cut dividends and share repurchases. Still, the risk remains that writedowns and provisioning will grow larger if house prices or economic activity will slow more than currently anticipated or if financial guarantors are downgraded further. Moreover, many of the largest firms are still at risk of further unplanned asset expansion from previous commitments for leveraged loans and their continued inability to securitize non-agency mortgages. Consequently, these firms are likely to be cautious in managing asset growth.
Your next exhibit focuses on business financial conditions. As shown by the black line in the top left panel, top-line operating earnings per share for S&P 500 firms for the fourth quarter are now estimated to be about 23 percent below their year-ago level, dragged down by losses at financial firms. For nonfinancial firms, the green line, earnings per share are estimated to be up 10 percent from a year ago. Analysts’ estimates of Q1 earnings for nonfinancial firms were trimmed a bit last week but suggest continued growth. Robust profits since 2002 have put most businesses in strong financial shape. As shown in the right panel, loss rates on high- yield corporate bonds, the black line, have been near zero for more than a year as very few bonds defaulted and recovery rates were high. However, we project that bond losses will rise gradually in the next two years as the nonfinancial profit share slips from its currently high level. In commercial real estate, the middle left panel, the net charge-off rate at banks, the black line, was low in the third quarter of last year despite a slight tilt up mostly from troubled loans related to residential land acquisition and construction. We project that this rate will rise fairly steeply, reflecting weakness in housing and expected softening in rents for commercial properties.
A similar outlook may lie behind the tighter standards for business loans reported in the January Senior Loan Officer Opinion Survey. As shown by the orange line in the middle right panel, the net percentage of domestic banks reporting having tightened standards on commercial real estate loans in the past three months reached 80 percent, a notable increase from the October survey. In addition, one-third of domestic banks tightened lending standards on C&I loans in the past three months. Large majorities of the respondents that tightened standards pointed to a less favorable or more uncertain outlook or a reduced tolerance for risk. Despite wider spreads, borrowing rates for investment-grade firms are lower than before the December FOMC meeting. As shown by the red line in the lower left panel, yields on ten-year BBB-rated bonds, the red line, fell about 25 basis points, and rates on thirty-day A2/P2 nonfinancial commercial paper, the blue line, have plummeted about 200 basis points since just before year-end. In contrast, yields on ten-year high-yield bonds, the black line, are up and now are close to 10 percent. Net borrowing by nonfinancial businesses, shown in the right panel, is on track in January to stay near the pace of recent months. Net bond issuance, the green bars, has been sizable in recent weeks with most of that issuance by investment-grade firms. Unsecured commercial paper, the yellow bars, rebounded after substantial paydowns ahead of year-end.
Your next exhibit focuses on the household sector. As shown in the top left panel, delinquency rates at commercial banks for credit cards, the blue line, and nonrevolving consumer loans, the black line, edged up in the third quarter, as did rates for auto loans at finance companies through November. Some of the recent rise in delinquency rates for credit cards is in states with the largest house-price declines, and could represent spillovers from weak housing markets. As shown to the right, delinquency rates on subprime adjustable-rate mortgages, the solid red line, continued to climb and topped 20 percent in November, and delinquency rates on fixed-rate subprime and on prime and near-prime mortgages also rose. Looking ahead, we expect delinquency rates on consumer loans to rise a bit from below-average levels as household resources are strained by higher unemployment and lower house prices. These developments have spurred lenders to tighten standards on consumer loans. As noted in the middle left panel, responses to the January Senior Loan Officer Opinion Survey indicate a further increase in the net percentage of banks tightening standards on credit cards and other consumer loans in the past three months. Banks also reported substantial net tightening of standards for subprime and prime mortgages, with the latter up considerably from the October survey. In addition, spreads on lower-rated tranches of consumer auto and credit card ABS jumped in January amid news that lenders were increasing loan-loss provisions. That said, interest rates on auto loans and credit cards, not shown, are not up, and most households still appear to have access to these forms of credit. As shown to the right, issuance of securities backed by these loans was robust through January. Securitization of nonconforming mortgages, the grey bars in the lower left panel, was weak in the fourth quarter of last year, and there has been little, if any, this month. But agency-backed securitization, the red bars, was quite strong in the fourth quarter and appears to be again in January. Moreover, as shown to the right, interest rates have fallen appreciably. Rates on conforming thirty-year fixed-rate mortgages, the blue line, and one-year ARMs, the red line, fell, and offer rates on prime fixed-rate jumbo mortgages, the black line, are also down. The six-month LIBOR, the rate to which most subprime ARMs reset, plunged in January, although, even at this level, the first payment reset will still be substantial for many households.
The next exhibit presents our outlook for mortgage defaults. The top left panel shows cumulative default rates for subprime 2/28 ARMs by year of origination. A default here is defined as a loan termination that is not from a refinancing or sale. The default rates for mortgages originated in 2006 and 2007, the red and orange lines, respectively, have shot up, and for mortgages originated in 2006, about 18 percent will have defaulted by the loan age of eighteen months. This rate is higher at every comparable age than for mortgages made in 2005, the blue line, and the average rate for loans made in 2001 to 2004, shown by the black line, with the shaded area denoting the range across years. Softer house prices likely played an important role in defaults on 2006 and 2007 loans because borrowers had little home equity to tap when they lost their jobs or became ill, or they walked away when their mortgages turned upside-down. These mortgages have not yet faced their first interest rate reset. As shown to the right, we expect a sizable number of borrowers to reset to higher payments, about 375,000 each quarter this year, if these mortgages are not prepaid or rates are not reduced. While many borrowers still have time to refinance or sell before the first rate reset, lower house prices and tighter credit conditions are likely to damp this activity. As noted in the middle left panel, to project defaults on subprime ARMs, we use a loan-level model that jointly estimates prepayments and defaults. The model considers loan and borrower characteristics at origination, subsequent MSA- or state-level house prices and employment fluctuations, interest rates, and “vintage” effects. As shown to the right, with data for the first three quarters in hand, we estimate that defaults in 2007 about doubled from 2006 and predict that they will climb further in 2008 and stay elevated in 2009. These estimates imply that 40 percent of the current stock of subprime ARMs will default over the next two years.
An important source of uncertainty around our projections is how borrowers will behave if falling house prices push their loan-to-value ratios above 100 percent. As shown in the first line of the bottom left panel, we estimate that 20 percent of subprime borrowers had a combined loan-to-value ratio of more than 100 percent in the third quarter of last year. If we assume that national house prices fall about 7 percent over the forecast period, as in the Greenbook, an estimated 44 percent of subprime mortgages would have combined LTVs above 100 percent. A similar calculation for prime and near-prime mortgages, shown in the second line, indicates that a not-inconsequential share, 15 percent, of these would also be upside-down by the end of 2009. While prime borrowers likely have other financial assets upon which to draw in the case of job loss or sickness, such high LTVs pose an upside risk to our baseline projection of defaults. Another source of uncertainty—this one on the positive side of the ledger—is how loan modifications can reduce defaults or loss of a home. We have limited information, but recent surveys indicate that loan workouts and modifications were modest through the third quarter of last year but likely accelerated in the fourth quarter. Servicers are strained working on the large number of loans that are delinquent before the first reset. One survey indicated that servicers assisted about 150,000 subprime borrowers in the third quarter, which would represent about 15 percent of those with past-due accounts, but were not addressing current accounts with an imminent reset.
As highlighted in the top panel of your next exhibit, we summarize our projections for credit losses in the next two years for major categories of business and household debt. These projections rely on the paths for house prices, unemployment, interest rates, and other factors from the Greenbook baseline. We also present projections based on the Greenbook recession alternative with the additional assumption that national house prices fall 20 percent. In this alternative scenario, real GDP growth turns negative in 2008, and the unemployment rate rises above 6 percent in 2009. As shown in the first column of the bottom panel, if we use the loss rates over the past decade or two as a guide to approximate losses under average economic conditions, total losses, line 6, would be projected to be $440 billion over the next two years. Such losses could be considered what might be expected by lenders of risky debt in the normal conduct of business. But conditions over the next two years are not expected to be normal, even under the baseline scenario. As shown in the second column, losses under the Greenbook baseline are expected to be considerably higher than average and total $727 billion, given our outlook for only modest growth. These losses might not greatly exceed the amounts that investors already have come to expect given signs of slowing activity. The above-average losses are especially large for residential mortgages, line 1, including those for nonprime mortgages, line 2. In contrast, losses for consumer credit, line 3, and business debt, line 4, are only a touch higher than normal. In the alternative scenario, in which business and household conditions worsen further, losses are projected to rise even more, not only for mortgages but also for other debt. Losses of this dimension would place considerable strains on both households and financial institutions, creating the potential for more-serious negative feedback on aggregate demand and activity than is captured by our standard macroeconomic models. Nathan will continue our presentation.