PrefBlog

The Federal Reserve Bank of New York has released a staff report by Andrew Haughwout, Ebiere Okah and Joseph Tracy titled Second Chances: Subprime Mortgage Modification and Re-Default:

Mortgage modifications have become an important component of public interventions designed to reduce foreclosures. In this paper, we examine how the structure of a mortgage modification affects the likelihood of the modified mortgage re-defaulting over the next year. Using data on subprime modifications that precede the government’s Home Affordable Modification Program, we focus our attention on those modifications in which the borrower was seriously delinquent and the monthly payment was reduced as part of the modification. The data indicate that the re-default rate declines with the magnitude of the reduction in the monthly payment, but also that the re-default rate declines relatively more when the payment reduction is achieved through principal forgiveness as opposed to lower interest rates.

More specifically:

After reviewing relevant previous studies and describing our data, we turn to an analysis of the effectiveness of the modifications we observe. We find that delinquent borrowers whose mortgages receive some kind of modification have a strong tendency to redefault, but that different kinds of modifications have diverse effects on outcomes. In particular, while HAMP focuses on reducing payment burdens, our results indicate the importance of borrower equity — the relationship between the mortgage balance and the home value — a factor that has been stressed in the previous literature on mortgage defaults. We conclude with a discussion of the implications of our results for modification policy.

The authors conclude, in part:

Our findings have potentially important implications for the design of modification programs going forward. The Administration’s HAMP program is focused on increasing borrowers’ ability to make their monthly payments, as measured by the DTI. Under HAMP, reductions in payments are primarily achieved by subsidizing lenders to reduce interest rates and extend mortgage term. While such interventions can reduce re-default rates, an alternative scheme would simultaneously enhance the borrower’s ability and willingness to pay the debt, by writing down principal in order to restore the borrower’s equity position. We estimate that restoring the borrower’s incentive to pay in this way can double the reduction in re-default rates achieved by payment reductions alone.

Another distinction between modifications that reduce the monthly payment by cutting the interest rate as compared to reducing the principal is the likely impact on household mobility. Ferreira et al (2010) using over two decades of data from the American Housing Survey estimate that each $1,000 in subsidized interest to a borrower reduces the two-year mobility rate by 1.4 percentage points. Modifying the interest rate to a below market rate creates an in-place subsidy to the borrower leading to a lock-in effect. That is, the borrower receives the subsidy only if he or she does not move.

Seems to me that HAMP is poorly designed – just another piece of poorly thought out feel-goodism.

One thing the authors did not address that interests me greatly is the accounting treatment for modifications which are wholly and directly owned by a single bank – as is the general rule in Canada – and how a choice between interest rate reduction and principal reduction might be reflected on the books of the firm. I suspect – but am not certain – that principal reduction will affect current profit, while interest rate reduction will merely affect future profit and be amortized over the length of the loan, despite the fact that it may be presumed that the choices are equivalent on a present value basis.

Given all the current fooferaw over the pending apocalypse in Canada when the current crop of 90%+ LTV mortgages comes due and needs to be refinanced at a higher rate, it might be prudent to start examining – and, perhaps, changing – the bookkeeping implications now, rather than being surprised later.

But then – who cares? Interest rate reductions are easy to understand and get more votes – so why should our glorious leaders do anything?

Update, 2012-7-31: Rebuttal from FHFA.

Update, 2013-5-13: Redesign of HAMP in 2010:

The US Treasury Department, as it continues to revamp the Home Affordable Modification Program (HAMP), announced today an initiative to encourage principal write-downs.

The principal reduction plan is one of the changes to HAMP, to be implemented in coming months.

The changes will encourage servicers to write-down a portion of mortgage debt as part of a HAMP modification, allow more borrowers to qualify for modification and help borrowers move into more affordable housing when modification is not possible, according to a fact sheet on the improvements provided to HousingWire.

S&P Commentary 2013-4-26:

In June of last year, Standard & Poor’s Ratings Services contended that principal forgiveness was more likely to keep U.S. mortgage borrowers current than more commonly used modification tools (see “The Best Way to Limit U.S. Mortgage Redefaults May Be Principal Forgiveness,” June 15, 2012). Data gathered since then not only support this view but also demonstrate servicers’ growing adoption of this form of loss mitigation. (Watch the related CreditMatters TV segment titled, “Principal Forgiveness Remains The Best Way To Limit U.S. Mortgage Redefaults,” dated May 7, 2013.)

As of February of this year, more than 1.5 million homeowners have received a permanent modification through the U.S. federal government’s Home Affordable Modification Program (HAMP). Since the publication of our June 2012 article, there have been more than 400,000 additional modifications on outstanding mortgages (as of March 2013). This translates to roughly a 22% rate of growth in the number of modifications on an additional $2.4 billion in mortgage debt.

Under the HAMP Principal Reduction Alternative (PRA) program, which provides monetary incentives to servicers that reduce principal, borrowers have received approximately $9.6 billion in principal forgiveness as of March 2013. Interestingly, servicers have ramped up their use of principal forgiveness on loans that don’t necessarily qualify for PRA assistance. Indeed, among the top five servicers for non-agency loans, we’ve noted that principal forgiveness, as a percentage of average modifications performed on a monthly basis, has increased by about 200% since the latter half of 2011 (see Chart 1). We attribute part of this to the $25 billion settlement in February 2012 with 49 state attorneys general and these same five servicers: Ally/GMAC, Bank of America, Citi, JPMorgan Chase, and Wells Fargo). In fact, although principal reduction remains the least common type of loan modification among servicers, the percentage of non-agency modified loans that have received principal forgiveness has increased by 3% since June 2012 (see Chart 2). Since 2009, servicers have forgiven principal on approximately $45 billion of outstanding non-agency mortgages.

The Federal Reserve Bank of New York has released a paper by Todd Keister and James J. McAndrews titled Why Are Banks Holding So Many Excess Reserves?:

The buildup of reserves in the U.S. banking system during the financial crisis has fueled concerns that the Federal Reserve’s policies may have failed to stimulate the flow of credit in the economy: banks, it appears, are amassing funds rather than lending them out. However, a careful examination of the balance sheet effects of central bank actions shows that the high level of reserves is simply a by-product of the Fed’s new lending facilities and asset purchase programs. The total quantity of reserves in the banking system reflects the scale of the Fed’s policy initiatives, but conveys no information about the initiatives’ effects on bank lending or on the economy more broadly.

They quote a lot of commentary decrying the build-up of reserves, but state:

In this edition of Current Issues, we argue that the concerns about high levels of reserves are largely unwarranted. Using a series of simple examples, we show how central bank liquidity facilities and other credit programs create—essentially as a by-product—a large quantity of reserves. While the level of required reserves may change modestly with changes in bank lending behavior, the vast majority of the newly created reserves will end up being held as excess reserves regardless of how banks react to the new programs. In other words, the substantial buildup of reserves depicted in the chart reflects the large scale of the Federal Reserve’s policy initiatives, but says little or nothing about the programs’ effects on bank lending or on the economy more broadly.

One casual comment of interest is:

Note the important economic role of interbank lending in this example: it allows funds to flow to their most productive uses, regardless of which bank received the initial deposits.

This, presumably, is a justification for encouraging interbank lending in the BIS Capital Ratios (which allow one bank’s holdings of another bank’s paper to be risk-weighted according to the credit rating of the borrowing bank’s sovereign authority. I would dearly love to see this issue thoroughly discussed; to me, it seems to have had the effect of increasing contagion.

Anyway: in normal times, Bank A has lent to Bank B:

But in stressed times, the inter-bank market fails and the Fed lends to Bank B (via a credit to their reserve account), which repays Bank A by transfer of reserves:

The authors explain:

This simple example illustrates how a central bank’s extension of credit to banks during a financial crisis creates, as a by-product, a large quantity of excess reserves. Merely looking at the aggregate data on bank reserves might lead one to conclude that the central bank’s policy did nothing to promote bank lending, since all of the $40 lent by the central bank ended up being held as excess reserves. The example shows that this conclusion would be unwarranted. In fact, the central bank’s action was highly effective: it prevented Bank B from having to reduce its lending to firms and households by $40.

There are, as always, knock-on effects:

Actions by a central bank that change the quantity of reserves in the banking system also tend to change the level of interest rates. Traditionally, bank reserves did not earn any interest. If Bank A earns no interest on the reserves it is holding in Exhibit 2, it will have an incentive to lend out its excess reserves or to use them to buy other short-term assets. These activities will, in turn, decrease short-term market interest rates and hence may lead to an increase in inflationary pressures.

The Central Bank may therefore choose to sterilize its market action by selling interest-bearing bonds from its holdings to Bank A – or it can achieve a similar policy objective by paying interest on excess reserves.

The authors conclude:

We began this article by asking, Why are banks holding so many excess reserves? We then used a series of simple examples to answer this question in two steps. First, we showed that the liquidity facilities and other credit programs introduced by the Federal Reserve in response to the crisis have created, as a by-product, a large quantity of reserves in the banking system. Second, we showed that while the lending decisions and other activities of banks may result in small changes in the level of required reserves, the vast majority of the newly created reserves will end up being held as excess reserves. The dramatic buildup of excess reserves reflects the large scale of the Federal Reserve’s policy initiatives; it conveys no information about the effects of these initiatives on bank lending or on the level of economic activity.

We also discussed the importance of paying interest on reserves when the level of excess reserves is unusually high, as the Federal Reserve began to do in October 2008. Paying interest on reserves allows a central bank to maintain its influence over market interest rates irrespective of the quantity of reserves in the banking system. The central bank can then scale its policy initiatives according to conditions in the financial sector, while setting its target for the short-term interest rate in response to macroeconomic conditions. This ability to separate short-term interest rates from the quantity of reserves is particularly important during the recovery from a financial crisis. If inflationary pressures begin to appear while the crisis-related programs are still in place, the central bank can use its interest-on-reserves policy to raise interest rates without necessarily removing all of the newly created reserves.

The Bank of Canada has released a discussion paper by Etienne Bordeleau, Allan Crawford, and Christopher Graham titled Regulatory Constraints on Bank Leverage: Issues and Lessons from the Canadian Experience:

The Basel capital framework plays an important role in risk management by linking a bank’s minimum capital requirements to the riskiness of its assets. Nevertheless, the risk estimates underlying these calculations may be imperfect, and it appears that a cyclical bias in measures of risk-adjusted capital contributed to procyclical increases in global leverage prior to the recent financial crisis. As such, international policy discussions are considering an unweighted leverage ratio as a supplement to existing risk-weighted capital requirements. Canadian banks offer a useful case study in this respect, having been subject to a regulatory ceiling on an unweighted leverage ratio since the early 1980s. The authors review lessons from the Canadian experience with leverage constraints, and provide some empirical analysis on how such constraints affect banks’ leverage management. In contrast to a number of countries without regulatory constraints, leverage at major Canadian banks was relatively stable leading up to the crisis, reducing pressure for deleveraging during the economic downturn. Empirical results suggest that major Canadian banks follow different strategies for managing their leverage. Some banks tend to raise their precautionary buffer quickly, through sharp reductions in asset growth and faster capital growth, when a shock pushes leverage too close to its authorized limit. For other banks, shocks have more persistent effects on leverage, possibly because these banks tend to have higher buffers on average. Overall, the authors’ results suggest that a leverage ceiling would be a useful tool to complement risk-weighted measures and mitigate procyclical tendencies in the financial system.

The authors conclude:

Empirical analysis provides evidence that banks follow different strategies for managing their leverage buffers. Some banks tend to raise their buffers very quickly when a shock pushes leverage too close to its authorized limit (as might occur during a cyclical upturn). These adjustments are achieved through sharp reductions in asset growth and faster growth in capital. At other banks, shocks have more persistent effects on leverage, which could be explained by the fact that these banks tend to have higher buffers on average.

There’s not a lot of meat in the paper, but it’s a start. Lord knows, we’re nevery going to see any honest analysis out of OSFI!

The Federal Reserve Bank of New York has released a staff report by Morten L. Bech and Elizabeth Klee titled The Mechanics of a Graceful Exit: Interest on Reserves and Segmentation in the Federal Funds Market:

To combat the financial crisis that intensified in the fall of 2008, the Federal Reserve injected a substantial amount of liquidity into the banking system. The resulting increase in reserve balances exerted downward price pressure in the federal funds market, and the effective federal funds rate began to deviate from the target rate set by the Federal Open Market Committee. In response, the Federal Reserve revised its operational framework for implementing monetary policy and began to pay interest on reserve balances in an attempt to provide a floor for the federal funds rate. Nevertheless, following the policy change, the effective federal funds rate remained below not only the target but also the rate paid on reserve balances. We develop a model to explain this phenomenon and use data from the federal funds market to evaluate it empirically. In turn, we show how successful the Federal Reserve may be in raising the federal funds rate even in an environment with substantial reserve balances.

This issue has been discussed on PrefBlog before, two posts being Effective Fed Funds Rate Continues to Confuse and Effective Fed Funds Rate: A Technical Explanation?.

The authors state the problem succinctly:

Between the October and December 2008 FOMC meetings the average effective federal funds rate was 32 basis points, while the target was 1 percent; the interest rate paid on reserves was 65 basis points or higher for the entire period. This deviation from the theoretical prediction was surprising for even the most astute observers of the federal funds market (Hamilton, 2008, for example). Why did interest on reserves provide an imperfect floor for the federal funds rate, even after the policy was changed so that the interest rate paid on reserves was set equal to the target rate?2 Why would any financial institution lend out funds below the rate paid by the central bank? And even if that were the case, arbitrageurs would surely relish the opportunity of making a pure profit by borrowing cheaply in the market and placing the proceeds with the central bank, and by doing so, move the market rate toward the floor.

… and propose:

The explanation for the puzzling outcome is at least threefold. First, not all participants in the federal funds market are eligible to receive interest on their reserve balances. Government-sponsored enterprises (GSEs) in particular, which are significant sellers of funds on a daily basis, are not legally eligible to receive interest on balances held with Reserve Banks. This heterogeneity across participants created a segmented market with different rate dynamics. Second, banks’ apparent general unwillingness or inability to engage in arbitrage has produced a market structure in which those banks that are willing and able to buy funds from the GSEs have been able to exercise bargaining power and pay the GSEs rates below the interest rate paid on reserves. The lack of arbitrage possibly has been driven in part by banks seeking to control the size of their balance sheet more closely in part to avoid stressing regulatory capital and leverage ratios, as mentioned in Bernanke (2009a). Third, a combination of financial consolidation, credit losses, and changes to risk management practices has led at least some GSEs to limit their number of counterparties in the money market and to tighten credit lines. This trimming of potential trading partners has likely decreased bargaining power with the remaining counterparties. In addition, the GSEs have become a larger share of the federal funds market in recent history and hence have pulled down the weighted average federal funds rate.

The paper is organized as:

First, we briefly discuss the institutional details of the federal funds market. Second, we turn to the history and implementation of the interest-on-reserves regime. Third, we present our model of a bifurcated federal funds market with banks and GSEs. We also sketch out how the model can be extended to describe a trifurcated market in which some banks are slow to adopt the new policy regime. Fourth, we calibrate our model to federal funds market data and back out the relative bargaining power of the different market participants. Fifth, we explore the factors that affect our computed bargaining parameters. We show that the level and distribution of reserve balances, rates in other overnight funding markets, and the riskiness of the buyers all have significant predictive power in explaining the bargaining power of the different types of sellers. With this information, in our sixth section, we forecast the effective federal funds rate under a variety of exit scenarios from the current accommodative stance of monetary policy. The seventh section concludes and offers directions for further research.