presentation file

Download Report

Transcript presentation file

Measuring the impact of securitization
on imputed bank output
Adam B. Ashcraft and Charles Steindel
Federal Reserve Bank of New York
May 13, 2008
All views expressed are those of the authors only and not necessarily those of the
Federal Reserve Bank of New York or the Federal Reserve System
Overview

Description of the mortgage securitization process.

Measurement of imputed nominal output from assets on a bank’s
balance sheet.

Measurement of imputed nominal output from securitized assets
off a bank’s balance sheet.

Implications of including off balance sheet imputed output on the
banking sector’s output.
2
The Securitization Process

Cash flows from underlying assets are tranched into classes of
securities.

Trust servicer (often a bank subsidiary) acts as intermediary,
collecting and paying interest (arranging swap agreements, etc.,
etc.).

Many similarities to a commercial bank financing a loan portfolio
with deposits.
3
Bank Lending and GDP Accounting: I

A commercial bank is assumed to produce imputed output in the
course of lending.

Such imputed output is estimated to equal the spread between
loan interest earned and deposit interest paid.

The difference between the “reference rate” and deposit rates is
used to compute “imputed interest paid” to depositors.
 This

is part of service consumption and thus aggregate GDP.
The difference between lending rates and the “reference rate” is
used to compute “imputed interest paid” to borrowers.
 This
interest presumably represents monitoring, service costs, etc.
provided by the lending institution.
 This adds to bank output, but reduces the output of borrowers.
4
Bank Lending and GDP Accounting: II

The “reference rate” is that earned by banks on Treasuries.

Wang-Basu-Fernald contend the proper reference rate should be
a risky one.
 Hence,
bank imputed interest paid to borrowers is overstated, and
imputed interest paid to depositors is understated.
5
Traditional and Securitized Lending

Bank subsidiaries servicing securitized loans provide many of the
same services as banks servicing on-balance sheet loans.

But spread income of such subsidiaries (over and above explicit
fees, such as servicing) are not counted in bank output.
 Addition
to profits would be offset by reductions in net interest paid by
banks.

What difference would it make to treat servicing subsidiaries like
banks?
6
Bringing Securitized Assets on the Balance Sheet

Use data from Schedule RS of the Call Reports.

Compute on-balance sheet effective lending rates and reference
rate; apply to securitized assets, deduct explicit servicing fees.

Our estimate is that 2006 imputed interest paid to owners of bank
securitized assets was about $40 billion, equal to 13.5% of
imputed interest paid to traditional borrowers.
 Share
has been growing gradually.
 Contribution from credit card asset-backed somewhat larger than that
from home mortgages sold.
7
Some Implications for Understanding the Macro Economy

Distribution of, and contributions to growth.
 Share
of banks (and other financial institutions) may be understated,
and nonfinancial share, overstated.

Potential distortions of aggregate GDP
 If
purchasers of securities backed by loans were assumed to earn and
consume imputed interest, nominal and real GDP would be higher.
 More importantly, aggregate GDP would then not be affected by sales
of loans into securitized structures.

May be of some interest in the current environment.
 Such
treatment may increase short-run inflation volatility, due to
volatility of the reference rate.

But actual issue would be covariance of reference rate and asset-backed
security rates.
 Another issue would be computation of relevant reference rate.
8