The 1996 Mortgage Survey provides evidence that the effects of the Savings and Loan crisis on New York City's multifamily lending market in the early 1990s have fully played themselves out. The years immediately following the recession ushered in vast changes in lending, including tightening lending standards, careful scrutiny by Federal regulatory agencies, institutional mergers, lenders exiting the lending market, and mounting delinquent and defaulted loans. Towards the end of 1993, the lending market for multifamily mortgages showed signs of improvement. Borrowers were no longer defaulting in large numbers, lending standards and loan volumes stabilized, and interest rates declined, reaching a 15-year low of 8.6% in 1994.
While 1995 Survey results were mixed - interest rates rose by 1.5%, though lenders increased the volume of loans underwritten - this year's Mortgage Survey shows continued growth in multifamily lending. Interest rates fell back to 8.6%, a drop of 150 basis points, and additional lenders entered the mortgage market. Likewise, the Federal Home Loan Mortgage Corporation, or Freddie Mac, infused $113 million into the New York City secondary market in only its second full year of operation following a temporary shutdown that began in 1990. Lending institutions are responding to almost non-existent loan delinquencies and to anticipation of continued low interest and inflation rates by allowing lower interest rates, longer loan terms, more fixed rate mortgages, and higher loan-to-value ratios.
Section 26-510(b)(iii) of the Rent Stabilization Law requires the Rent Guidelines Board to consider the "costs and availability of financing (including effective rates of interest)" in its deliberations. To assist the Board in meeting this obligation, the RGB research staff conducts an annual survey of financial institutions which underwrite mortgages to multifamily properties in New York City.
During the past year, RGB staff made several improvements to the Mortgage Survey. Last year's survey sample was updated to include only those institutions that still offer loans for multi-unit buildings in New York City. In addition, since so many of the lenders surveyed in the past have merged or discontinued offering mortgages for multifamily properties, staff combed newspapers, trade magazines, the yellow pages and other sources for lenders to include in the sample. We more than made up for institutions lost last year by adding ten new lenders to our sample, reaching a total sample size of fifty-six institutions.
In response to requests from RGB Board Members, staff made a few additions to the Mortgage Survey questionnaire. New or enhanced questions include whether the change in the Major Capital Improvement program (MCI) has affected the level of non- performing loans; the percent of refinanced mortgages accounted for by small buildings; and a distinction between vacancy and collection losses. Finally, staff added two new sections to the Mortgage Survey Report. One is a longitudinal perspective of those institutions completing the 1994-1996 Surveys, the other is a retrospective of the multifamily lending market in New York City during the past decade which combines data from RGB Mortgage Surveys and other sources.
Graphic Charts:
Twenty-one of the fifty-six financial institutions surveyed completed the 1996 Mortgage Survey, furnishing the RGB with details about the current multifamily mortgage lending market. Unlike past years when we found several lenders had stopped underwriting mortgages for multifamily buildings, two institutions recently created separate multifamily mortgage divisions and are currently developing lending standards. And contrary to previous years' spate of mergers, not one lender in our sample merged with another this year, though three institutions responded that they have too few outstanding loans for rent stabilized buildings to respond to our questionnaire.
Thirteen of this year's respondents also completed last year's Mortgage Survey and eleven completed the previous three surveys. Given this strong response in multiple years, we added a separate longitudinal section to this year's Report allowing us to distinguish between differences due to changes in the lending market and those due to changes in institutions responding to RGB surveys.
Interest rates for multifamily mortgages dropped this year, averaging 8.6% for new and refinanced loans. This is a decrease of 150 basis points from last year's survey when we reported an average rate of 10.1% and marks only the second time since 1980 that mortgage rates averaged below 9%.
Because lending institutions take their cue from the Federal Reserve, it is not surprising that mortgage rates declined last year. Following its strong anti-inflationary stance throughout 1994 and early 1995 when the Federal Reserve raised rates seven consecutive times, the Federal Reserve reversed this trend in mid- 1995 reducing the federal funds rate twice by a total of .5%. At its first committee meeting in 1996, the Fed lowered short-term interest rates by an additional quarter of a percent to reach 5.25%. Similarly, the discount rate is now 5%. Rate cuts by the Federal Reserve spur large banks to decrease their prime lending rates, leading to similar reductions for mortgages, home equity loans, small business loans, and credit card balances.[1]
Points, terms and types of loans for both new and refinanced mortgages have remained relatively constant in recent years. Points, or service fees, currently charged by lenders range from 0 to 3, the same as last year, but the average service fee for new loans is now 1.32% versus 1.25% one year ago making the initial outlay for these loans somewhat more expensive. Average points charged for refinanced loans are once again lower than for new loans, averaging 1.21%, about the same as last year.
Since survey respondents normally provide a wide range of term lengths rather than a single number, it is difficult to know where within the range banks choose to lend. With this caveat, it appears that mortgage terms increased since a year ago for new and refinanced mortgages. Though the length varies between 5 and 30 years for the last two years, more lenders providing single numbers indicated 15-year mortgages this year, while respondents have indicated 5 or 10 years in the past. This longer term may signal increased competition among lenders and an improved economic outlook.
Lenders are also offering terms that exhibit more flexibility. For example, more banks are now offering loans with fixed rates for the beginning of the term and adjustable rates thereafter, as well as mortgages with longer amortization schedules than the loan's term.
Last year, approximately one-half of lenders offered fixed rates and the other half supplied adjustable or balloon mortgages, perhaps anticipating that low mortgage rates would not persist in the long run. This year, two-thirds of lenders offer fixed mortgages while the remainder offer adjustable or balloon type mortgages. Many respondents report they offer all three, and one lender lets the customer decide. An adjustable rate mortgage is usually rescheduled after 3 years for shorter term loans and after 5 for longer term loans.
With the relatively large reduction in interest rates from 1995 to 1996, came a corresponding increase in refinancing activity, reaching levels similar to 1994. This year nine lenders (43% of those responding to this question) indicated a portion of their mortgage portfolio was refinanced at lower rates, six of these institutions refinanced more than 10% of their outstanding loans.
A new question on this year's mortgage survey reveals that 55% of the mortgages refinanced at lower rates are in buildings that have 20 or fewer units. This is partly because about half of the lenders reporting high levels of refinancing activity typically lend to small buildings. Thus, this survey shows that small buildings are also benefiting from lower debt service payments resulting from refinanced mortgages.
The volume of loans underwritten by financial institutions declined slightly throughout 1995 despite decreases in interest rates. Nearly 30% of respondents reported a one-third reduction in the number of loan applications received, and two other institutions report decreases in the rate of application approvals. This was offset somewhat by 15% of institutions underwriting more loans due to both increasing applications and approvals. Qualitative reasons for decreasing loan applications provided by respondents suggest heightened competition among lenders, a sentiment which may indicate why banks are reducing their standards. Perhaps lenders are not only reacting to a better market outlook, but they are also attempting to attract more business.
As mentioned in previous Mortgage Survey Reports, lenders developed increasingly cautious lending criteria in the early 1990s responding to rising loan delinquencies and defaults and to pressure by Federal oversight agencies. The Federal Deposit Insurance Corporation (FDIC) closed several financial institutions and took control of others, and the Resolution Trust Corporation (RTC), established by Congress in 1989, restructured the thrift industry and worked to minimize the effects of the costly S&L scandal. The proportion of lenders claiming they implemented stricter standards dropped remarkably after 1993 to 15% and 10% respectively in 1994 and 1995, and fell to nearly zero this year. Only one lender in 21 mentioned tightening its standards by using more stringent approvals and monitoring requirements. This was in response to increased delinquencies by landlords in the past and an increase in opportunities to sell loans on the secondary market. Those banks reporting more stringent standards last year mentioned these same two factors. Again, we attribute the decline in the number of banks tightening their standards this year to enhanced standards implemented in the early 1990s, and since maintained, leading to low delinquency and default rates, as well as to better economic conditions.
A second set of questions relating to lending standards requests institutions to furnish other requirements including loan-to- value ratios, debt service coverage, and building characteristics. The mean dollar amount respondents are willing to lend based on a building's value (loan-to-value ratio, or LTV) increased in 1996 by 1% to reach 71%. Standards for LTVs range from 50% to 80%. This is the second year that the average standard LTV ratio increased one percent, indicating a slight loosening in mortgage financing standards.
The debt service ratio (net operating income divided by the debt service) measures an investment's ability to cover mortgage payments using its net operating income. Currently, lenders' standards for debt service ratios vary from 1.15% to 1.4%. The mean debt service coverage is 1.24%, slightly less than the average 1.25% reported last year. The 1.15% standard falls somewhat close to the "risky" level where available net operating income is only 15% higher than the debt service. Some lenders reported the same requirement for debt service coverage last year, though, and have not indicated the presence of defaulted or non-performing loans.
Requirements regarding mortgage levels and physical characteristics of buildings have not changed much since last year. Three respondents have minimum loan values ranging from $500,000 to $1.25 million and one bank offers loans of no more than $36,000. These figures are in line with last year's responses. This year's survey also yielded similar responses in terms of number of units, building age, location and level of maintenance. Almost all lenders require buildings to be in at least good condition, four lenders have building-size requirements (minimum of 5 to 10 units), two specify location, and three consider whether a building has the potential to convert to a co-operative or condominium. Unlike last year, no respondents consider building age or whether the owner lives in the building in their lending criteria. Two institutions mentioned additional requirements not listed on the survey; one looks at the environmental aspect of the building, and the other reviews buildings' management.
This year's responses to the non-performing loan section of the Mortgage Survey are even more encouraging than last year's results which showed that the recession of the early 1990s had finally stopped reverberating through lenders' portfolios of outstanding loans. Last year, three lenders reported decreases in non-performing loans and four claimed their level of foreclosure proceedings declined substantially. No lenders reported increases in non-performing loans or defaults. Once again this year, not one survey respondent experienced an increase in non-performing or defaulted loans. One institution reduced its non-performing loans and foreclosure proceedings by 100% and attributed these results to the improved rental market.
At the end of 1994, the New York State Court of Appeals capped Major Capital Improvement (MCI) increases at 6% and allowed them to become part of the base rent. Formerly, temporary increases up to 12% were allowed but were not added to the base rent. The new ruling caused concern among owners that the reduced return would inhibit building repairs and therefore would cause buildings to deteriorate over time. Since no institutions responding to the 1996 Mortgage Survey experienced an increase in non-performing loans, none responded to the question containing the change in the Major Capital Improvement (MCI) program. This leads us to believe that the effects of changing this program have not been overwhelming or that the effects will manifest in the long term and are not yet visible in owners' balance sheets.
RGB Mortgage Surveys also ask lenders how they resolve foreclosure actions against rent stabilized buildings with delinquencies. Again, many respondents did not answer the question since they currently have no non-performing loans, though some institutions did provide answers. Of those who responded, most (six out of seven) seize the building or restructure the outstanding debt. Others reported resuming regular debt service and arranging financing with another financial institution, while one lender reported working out any problems with the building owners. These results do not differ from last year.
A number of questions on the Mortgage Survey ask about characteristics of buildings currently in lenders' portfolios including building size, vacancy and collection losses, loan-to- value ratios, and operating and maintenance costs. Similar to last year, over half (57%) of lenders in our sample typically provide loans for buildings with more than 20 units, the most frequently cited size being 50 to 99 units. The next most common building sizes are 11-19 and 20-49 units respectively. Two lenders typically lend to buildings with fewer than ten units and one mainly lends to buildings with 100 or more units. Again this data does not vary from responses in previous years.
A change in the Mortgage Survey instrument allows us to distinguish between relinquished rental income due to vacant apartments versus lost income caused by delinquent rental payments. The combined vacancy and collection losses reported by respondents declined considerably since last year when the mean was 4.6%. This year's average is 3.7%, similar to 1990 when the average was 3.5%.
Last year, nearly three-quarters of respondents had vacancy and collection losses of 5% or more. This year, only half of respondents reported losses this high and one-quarter (4 of the 17 institutions responding to this question) claim combined losses of 1% or less. We can only surmise this is due to the overall improvement in New York City's economy.
It is unlikely that this substantial decline in vacancy and collection losses stems from the change in the survey instrument, since the question regarding combined vacancy and collection losses is the same on both surveys and precedes the question requesting respondents to separate the two types of losses. This breakdown shows that, on average, 2.9% of this year's total losses are attributable to collection problems, while just under 1% is due to vacancies. Given an overall vacancy rate of 3.4% in New York City's housing stock, this figure appears low. However, such low vacancy and collection losses are not unprecedented - RGB Mortgage Surveys from 1988 to 1990 found combined losses of around 3%.
Though the RGB did not request lenders to separate vacancy and collection losses in the past, in 1993 RGB staff conducted a survey which requested vacancy information of owners who have buildings in tax arrears. The survey found that "almost 20% of the average building's potential rent roll remains uncollected due to [vacancy and collection] losses. A 6% loss derives from vacancies and an additional 13.5% from an inability to collect rent from tenants." These results show a similar proportion of losses due to vacancies (approximately 30%) and those due to unpaid rent (roughly 70%), though buildings with tax arrears relinquish far greater amounts of their rent roll than most buildings in lenders' portfolios.
The loan-to-value ratio (LTV) on mortgages currently held by respondents averages 65%, or the same as last year. Though the average is the same, one-third of lenders (7 out of 20) reported typical LTVs of 70 or higher, twice as many as last year. Apparently, some lenders are beginning to lend up to their maximum LTV standard which they have refrained from in recent years. LTV standards have also increased in each of the last two years by 1% and now average 71% as mentioned earlier in this Report.
The Mortgage Survey questionnaire traditionally requests typical operating and maintenance (O&M) expenses of buildings with outstanding loans. Because lenders' answers are extremely varied, we do not present average or modal values. Lenders' responses are more a reflection of the type of building, whether luxury or basic and the buildings' conditions, for which the lender underwrites mortgages rather than a guideline of costs involved in operating New York's rental housing. Nonetheless, such responses are valuable in determining what type of buildings currently hold outstanding mortgages. For example, a response of $3,000 in monthly operating and maintenance expenses indicates the institution lends to highly-staffed and well-maintained buildings with large units. More than half of 1996 responses range from $240 to $550 per unit per month, while two respondents indicate O&M costs of $3,000 or more. Further, reported O&M costs range from 30% to 60% of gross income according to this year's survey respondents, similar to previous years.
The differences between an institution's current lending standards and the characteristics of its overall portfolio point to changes in that institution's formal or informal practices and possible exceptions to its standards when choosing to underwrite individual loans. The loan-to-value ratio data provides concrete evidence that a subset of lenders are sufficiently comfortable with the economy to ease their lending practices even if they have not officially changed their underwriting standards, as none report doing since last year.
With so many of the same institutions responding to the 1994, 1995 and 1996 Mortgage Surveys, we decided to add a longitudinal perspective to the Report. In this section, RGB staff compare responses from lenders who replied to all three surveys (longitudinal group) with the data from all institutions providing responses in these years (cross sectional group). This comparison will help to determine whether the changes we have seen in the last two years reflect changes in the lending market or differing Mortgage Survey respondents.
The terms offered by institutions consistently responding to our surveys (longitudinal group) differ slightly from those of all respondents (cross sectional group). For example, interest rates for new mortgages were 8.2%, 9.7%, and 8.3% respectively in 1994 through 1996, which is slightly less than the respective 8.6%, 10.1%, and 8.6% we reported for all lenders in these years. Though interest rates were lower, service fees are higher for respondents in all three years averaging more than 1.4 in the longitudinal group as opposed to roughly 1.25 for the cross sectional group. Loan lengths and types in the longitudinal group are more consistent with the cross sectional group. Overall, financing terms are not very different for the two groups.
Similarly, refinancing activity was fairly consistent for lenders who responded in all three years, except that in 1994 a larger proportion of cross sectional lenders reported an increase in refinancing activity. Thus, the percent increase in banks refinancing a sizable portion of their portfolios in 1994, 42% of all lenders, may have been overstated; likewise for the proportion of lenders experiencing increases in loan volumes. This does not change the trend for these years, rather it merely alters the year in which the refinancing activity and loan volume changes occurred.
Some of the changes in lending practices we have reported since 1994 may have been overstated or have occurred in different years from those reported because of differing respondents to the Mortgage Surveys. We noted in previous Reports that acceptable loan-to-value ratios have been increasing over the years (by a total of 2% since 1994), a finding that the the longitudinal data confirms, though the increase for longitudinal respondents was about 1% and occurred between 1994 and 1995.
Further, the longitudinal debt service coverage data, as well as the longitudinal data for the LTV ratio of outstanding loans, supports our finding of relaxing standards. Likewise, the reduction in vacancy and collection losses reported in the cross sectional data also is evident in the longitudinal data. The average losses reported in 1996 are 3.4%, or nearly 1% less than in 1995. Four lenders out of six who responded to this question on all three questionnaires report fewer losses due to delinquent rental payments and vacant apartments. Though caution must be exercised whenever using so few responses, we present the longitudinal data since it mostly corroborates the findings of previous cross sectional studies.
Another optimistic finding is that almost all institutions responding to RGB surveys in multiple years (longitudinal group) report decreases in non-performing loans and foreclosure actions. Those lenders not reporting declines had no delinquent loans to report. This backs up the findings in our cross sectional studies that delinquencies have in fact declined or were minimal for several years.
Savings banks make up the vast majority of respondents to annual RGB Mortgage Surveys with commercial lenders and savings and loans providing the rest. However, the percent of each type of lending institution fluctuated from 1994 to 1995. In 1995, the RGB received the highest percent of responses from savings banks - nearly three-quarters of all respondents. This year, slightly more than half of returned surveys were from savings banks, with commercial lenders picking up the slack and savings and loans remaining constant. There are distinct differences among types of lenders and lending practices. Specifically, saving banks' average interest rates are usually lower than those charged by savings and loans and commercial lenders. Because most of the lenders in the longitudinal group are savings banks, this explains why longitudinal interest rates average less than the cross sectional data, with neither group - longitudinal or cross sectional - necessarily reflecting the "true" value.
Though the small number of institutions responding to a question in all three years renders the data unreliable on its own, the longitudinal perspective is useful if presented in conjunction with the more abundant cross sectional data. With noted exceptions, the longitudinal perspective confirms that the multifamily lending market has improved considerably since the recession in the early 1990s and has continued to loosen in the past three years. Interest rates and rental losses are down, lending standards have relaxed, and outstanding loans are remaining current. With lower costs of borrowing and abundant mortgage availability, perhaps demand for lending services will pick up in the coming years.
Though RGB staff provide two- or three-year perspectives on multifamily lending practices in its annual Mortgage Survey Reports, the vast changes in this market in the last decade or more call for a review of the mortgage lending market that is longer in term and broader in scope. We draw data from RGB Mortgage Surveys, from nationally collected statistics regarding housing construction and from participants in the secondary lending market.
Mortgage Survey respondents report altering their lending practices in recent years to conform with required standards of the secondary mortgage and mortgage insurance markets, particularly programs of the Federal Home Loan Mortgage Corporation and State of New York Mortgage Agency. Though it is difficult to assess the impact of these two groups in fueling the local single and multifamily lending markets, it is also hard to overlook their importance.
Since 1978, the State of New York Mortgage Agency (SONYMA) has provided mortgage insurance for construction and rehabilitation of single family and multifamily housing as well as for community development projects. As of December, 1995, the Agency provided additional credit to build nearly 32,000 dwellings in New York City, 85% of which are in buildings with five or more units, worth approximately $500 million. SONYMA issued commitments for an additional 10,000 apartments in New York that have not yet been completed.
The Federal Home Loan Mortgage Corporation (Freddie Mac) has traditionally been a strong force in the New York area where much of the multifamily secondary market is located. The corporation, established by Congress in 1970 to provide a continuous flow of funds to mortgage institutions, purchases mortgages from lenders and packages them into securities to sell to investors. These purchases lead to more available funds for the lenders to make additional loans.
Freddie Mac shut down its multifamily loan program in October, 1990 to minimize its losses when a large part of its assets were distressed due, in part, to the bottoming out of the real estate market. By 1994 Freddie Mac had fully re-entered the secondary lending market after spending 1990-1993 refinancing some of its portfolio and restructuring its lending and organizational procedures. Since then, Freddie Mac has purchased a total of $135 million in mortgages, $22 million in 1994 and $113 million last year, from multifamily lenders in the New York City. Though these figures are below the average amounts purchased prior to the 1990 shutdown, the corporation expects more volume in 1996. Other signs of Freddie Mac's growth are its recent decisions to create new programs including an affordable housing pilot, a single loan program for mortgage-backed securities, and a 5+5 program where interest rates are fixed the first 5 years and adjustable thereafter.
Since lenders have tightened their standards in response to New York's real estate crisis, most do not have to further tighten their standards to participate in the secondary mortgage market. Given recent trends toward heightened participation in the secondary market and the creation of additional Freddie Mac programs, more opportunities are expected for lenders to participate in secondary lending thereby creating additional mortgage resources.
The most striking change in the lending market over the years has been the steady decline in interest rates for both new and refinanced loans on multifamily properties. Likewise, while rates for both types of loans are down considerably, refinanced loans are no longer at interest rates that are almost twice the rate of new loans, as experienced in the early 1980s. In other words, owners who had balloon mortgages in this period were forced to refinance their mortgages a few years after origination at much higher rates inflating their debt service payments. By the late 1980s, refinanced loans were in line with those for new loans and in the past several years have been nearly indistinguishable.
Since lending terms are comprised of points, terms, and types in addition to interest rates, it is important to review how all of these components change when assessing the stringency of lending standards in any one year. In the 1989 Mortgage Survey Report, the RGB stated "it appears that the long-term fixed rate mortgage has largely disappeared. Only two banks responding to the survey offer fixed rate loans of 15 years or more." This year, in contrast, we found that institutions lowered interest rates and offer longer loan terms and more fixed rate mortgages. This change provides additional evidence of much looser lending practices resulting perhaps from major changes in the outlook for multifamily financing.
Surprisingly, the continued decline in mortgage interest rates since the mid 1980s has not sparked more multifamily housing development in the Metropolitan area. While multifamily housing starts have rebounded in other areas of the country, especially in the South and Midwest, residential building throughout the Northeast remains sluggish.
Similar to the trend in multifamily housing, conventional mortgages rates are at their lowest point in several years but are not spurring single family development or purchases in the area. Data from the U.S. Bureau of the Census shows that despite uncharacteristically low conventional mortgage rates, single family housing construction in the Northeast reached the lowest number of starts in two years but could cite no specific reason for the decline. The fact that housing construction in both the single and multifamily sectors has not rebounded supports the notion that construction activity is more a reflection of the region's economic performance rather than a response to national monetary policy or to local housing practices.
The relationship between interest rates and housing construction in New York City is relatively strong, but other factors probably override the decline in mortgage rates to thwart new housing construction in recent years. These may include rising unemployment, reductions in government subsidies at all levels, and growing concerns over an economic recession, however mild. With continued pessimism about the economy's performance and mounting employment insecurities, a reduction in mortgage lending costs and enhanced loan availability may not be sufficient to pull New York's housing construction out of its slump.
1. Christopher Drew, "Federal Reserve Trims Key Rates To Spur Economy" New York Times, February 1, 1996.
