Docstoc

Porteous Session final Version KfW Entwicklungsbank

Document Sample
Porteous Session final Version KfW Entwicklungsbank Powered By Docstoc
					Assessing the potential of financial markets:
    Connecting low income borrowers




                Commissioned by KfW
    for 2006 KfW Financial Sector Development Symposium
             Financing Housing for the Poor:
         Connecting Low-Income Groups to Markets
                November 9th and 10th, 2006
                       KfW, Berlin


           V5.0 Conference Draft not for quotation
                      31 October 2006


                    David Porteous



                   www.bankablefrontier.com
V4.0 Draft




Abstract
The volume of housing finance has grown at unprecedented rates in many developing
countries in recent years. However, the majority of people in these countries still lack
access to formal housing finance—whether by mortgage or housing microloan.
Gathering data from a diverse sample of eight developing countries, this paper first
compiles measures of mortgage market size, penetration and completeness. The
minimum lumpsums to buy a house are calculated for each; and the access frontier
approach is applied to understand better the current dimensions of access to
mortgages. Then, drawing on evidence from recent surveys of poor households in
several developing countries, the paper considers how poor people do in fact finance
their housing expenditures. Information gathered from a range of leading housing
microfinance providers about their client profiles locates access to housing
microfinance within a common framework. Using the lens of access, the paper aims to
bridge the two worlds of mortgage finance and microfinance, which have hitherto
often been distinct, but are increasingly blurring. More of both types of housing
finance is required to expand access to the unserved 60-90% of households in
developing countries.


Executive Summary
1. This paper is intended to provide a context and a framework for discussion at the
   KfW Housing Finance Symposium. The paper collects and analyzes data, which
   provide indicators of the extent of access to housing finance in a sample of
   developing countries, with a special focus on low-income borrowers. The paper
   frames the issue of how low-income households connect to formal housing finance
   markets by taking, first, a top down view on the reach of the mortgage
   instrument; and second, a bottom-up view on the current known experience of
   low-income households.

2. The volume of housing finance has risen sharply in many developing countries in
   recent years. Much of this increase has come from the development and
   liberalization of mortgage markets. At the same time, microfinance has come of
   age as an approach for providing financial services to households which were
   previously regarded as ‘unbankable’. While mortgages and microloans have
   traditionally been strongly differentiated on the basis of the collateral required
   and the nature of the provider, these differences are blurring as both loan types
   are used for diverse purposes, and are increasingly provided by the same entities.

3. The mortgage has long been regarded as the best financing instrument for house
   acquisition, but its reach is limited. How far does access to mortgage finance in
   fact extend in developing countries? The paper reports on a range of measures of
   the depth, affordability and completeness of mortgage markets in a sample of
   eight countries. Although mortgage markets are very new in some and minute in
   others, many are expanding at historically unprecedented speeds. The mortgage


                                                                                      2
V4.0 Draft



    product set available is increasingly converging towards a minimum 20-year,
    variable rate mortgage with a loan-to-value ratio of 80% or more. Applying an
    approach designed to measure access rather than usage in the few countries
    where the data permits, the paper finds that at most, a third to 40% of households
    in middle-income countries have access to mortgages on current terms. This
    number is much lower in low-income countries.

4. Most home-owners improve their houses over time regardless of their legal tenure.
   Housing microfinance to support home improvement is increasingly hailed as a
   vital solution for those whom mortgages cannot reach. How do poor households
   actually finance their housing? And which entities currently provide formal
   housing finance, connecting poor households to the retail financial services
   markets? Drawing on new analysis of data from in-depth ‘financial diaries’ of poor
   households in Bangladesh, India and South Africa, it is clear that own savings
   remains by far the most important source of finance for housing. It is clear that
   formal housing microfinance is relatively new and small in scale. At present,
   formal housing microfinance has very limited reach, mainly among the urban poor
   and near poor. Information from Interviews with and recent research on leading
   providers in the different categories suggests that, despite a range of obstacles,
   housing microfinance is starting to reach early scale. It is questionable, however,
   whether specialized housing-only micro-financiers can survive competition from
   larger lenders with a broader base; and there is evidence that community shelter
   loan funds that do not rely but which draw funds from government and/or donors
   perform poorly over time.

5. There is a growing need and demand to collect and analyze regular, consistent
   information about housing finance markets and about the performance of housing
   finance portfolios. In this way, changes in access over time can be tracked.

6. While a housing finance system alone cannot engineer away problems of lack of
   income or of affordability, the success of a system should be measured not by
   volumes of lending alone but by the extent to which it expands access across the
   population.




                                                                                    3
V4.0 Draft




Foreword
My thanks are due to the many people who provided helpful assistance, especially at
country level: both country correspondents and institutional survey respondents: Jim
Hokans (Ghana), Kecia Rust (South Africa), Alan Elizondo (Mexico), Mehmood Bughio
(Pakistan), Mher Yedigaryan (Armenia), Noah Sawyer (Honduras), Ljubica Gelev
(Serbia), Dev Goel (India).

In addition to my fellow speakers at the symposium and various KfW officers,
especially Cerstin Sander, who helpfully commented on versions of the paper, several
others were very helpful with contacts or information: Bertrand Renaud; Bob Buckley;
Friedemann Roy, Tim Eliott, Asif Dowla, Daryl Collins and Stuart Rutherford. People
whom I interviewed from specific entities were also extremely helpful in providing
information on their portfolios and experience: Christy Stickney (Habitat for
Humanity, LAC), Beth Rhyne & Nino Mesarina (Accion International), Harish Khare
(HDFC), Mariana Balestrini (BCEI), Smbat Nasibyan (Conversebank, Armenia), Frieder
Woehrmann (ProCredit Holdings) and Mirjana Zakanji (Procredit, Serbia). Anne-Marie
Chidzero and Andrea van der Westhuizen of FinMark Trust were helpful in accessing
FinScope data for Zambia.

My thanks are due to all of these people.




                                                                                  4
V4.0 Draft




List of Contents

1. Introduction .....................................................................................6

2. The changing frontiers of mortgage markets ..............................................9

   2.1 Empirical measures ........................................................................9

   2.2 Going deeper: access frontiers ......................................................... 17

3. The poor and their housing finance: a bottom-up view................................ 25

   3.1 Which housing finance instruments do poor people use? ........................... 25

   3.2 Who connects low income populations to formal housing finance? ............... 29

   3.3 Connecting the pieces ................................................................... 33

4. Conclusions .................................................................................... 35

References ........................................................................................ 38

Annex A: MOW mortgage market completeness measure.................................. 42




                                                                                                    5
V4.0 Draft




1. Introduction
“In fact, because many of the countries where housing finance has developed in
recent years are so populous…a majority of people in developing countries, if not a
majority of the countries, now have access to market-based mortgage credit.”
                                                    Buckley and Kalarickal 2005:247

International housing finance is coming of age as a distinct field of knowledge and
endeavor. At least, this may be our conclusion if we judge by the number of
international conferences held on the topic in 2006: this KfW Symposium is the fourth
of which I am aware. The conferencing activity is in part both cause and result of
international housing finance slowly emerging from the ‘twilight zone’ between urban
development and financial sector development to assert its claim to be a distinct
field.

The volume of housing finance has grown rapidly in many developing countries in the
past five to ten years. However, do the majority of people in these countries yet have
access to formal housing finance instruments, such as mortgages or housing
microloans, as the quotation above implies?

This paper seeks to explore key dimensions of access to formal housing finance—
especially how to understand and measure it. To do this, the paper applies the
insights and some tools developed in terms of the emerging ‘access to financial
services’ approach. This approach is premised on two main insights. First, not all
people will want to use a particular product or service; there can be particularly weak
demand for credit products among groups that have a strong cultural resistance to
borrowing. Policy measures cannot expand usage on their own but should rather seek
to expand effective access to desirable products, allowing usage to find its own level.
Second, to the extent that they have access, people typically use a range of
instruments and providers for their financial service needs. Understanding retail
financial markets through narrow provider categories only—for example, clients of
microfinance institutions only—are less useful than considering access and choice from
an integrated client perspective.

The access to financial services approach has emerged over the past decade following
new evidence at micro and macro levels. First, at a micro level, Stuart Rutherford
(2000) demonstrated in his seminal work, The Poor and Their Money, how important
the question of accessing sufficiently large lump sums is in the lives of poor
households. The core task of housing finance may also be understood in these terms:
how to reduce the large lumpsum required to buy a house to lowest possible levels
through spreading it over time. Second, at both micro and macro levels, various
survey-type techniques have given new insight into low income households’ usage of
financial instruments, and their attitudes towards them. The ‘Financial Diaries of the
Poor’, undertaken in at least three countries to date, provide an in-depth view of the
use of financial instruments by poor households, including for their housing needs.


                                                                                     6
V4.0 Draft



Drawing on data collected by nationally representative financial surveys such as
FinScope in several African countries, the access frontier approach provides a means
of interpreting the state of access to financial services. Applying these approaches to
housing finance enables us to understand better the current situation of access across
the population.

By using the lens of access, the paper therefore bridges two strong currents of
financial sector development over the past ten years: the growth of the traditional
instrument of housing finance, the mortgage, a long-term obligation secured by legal
attachment to a property; and the rapid expansion of the microfinance sector,
offering largely short-term unsecured loans. On the surface, the two sectors are very
different: certainly, the main credit instrument of each differs in fundamental ways—
such as term, collateral and interest rate. Furthermore, the development of the
mortgage market requires a relatively high degree of legal certainty and of funding
capacity; while microfinance has often thrived in countries with low legal certainty
and without developed financial markets. However, in practice, the distinctions
between the two are increasingly blurring as traditional mortgage lenders start to
downscale by adding unsecured loan products; and as microfinance institutions have
become regulated and have offered longer term, larger home improvement loans, and
even in some cases, mortgages. Even the purpose of the two types of loans may not
be that different in practice: after all, even though secured by a house, mortgages
are used for a variety of purposes, including working capital for small businesses in
developed countries. And twenty to thirty percent of traditional microloans have
often been used for home improvement by borrowers.

The paper is structured around two perspectives on the question of access to formal
housing finance. First, in what may be considered a ‘top down’ view, Section 2
compiles and analyses some dimensions of the mortgage market across a diverse
sample of eight developing countries (see Box A below). Even as the volume of
mortgage finance has expanded in these countries, the section asks how does
mortgage finance contribute to reducing lumpsums required to buy a house; and how
far does access to mortgage finance extend in developing countries?

Second, the paper takes a ‘bottom up’ view: most, if not all, home owners improve
their houses over time, whether or not they have formal title to the land or buildings.
Most improvements are in fact financed out of formal and informal savings. However,
housing microcredit is increasingly regarded by many observers as a critically
necessary solution to speed up the incremental improvement of housing conditions,
especially among poor people in urban areas. There is now considerable focus on slum
upgrading in line with Millennium Development Goal 7, Target 11: “Have achieved by
2020 a significant improvement in the lives of at least 100 million slum dwellers.”
Microfinance itself has gone through a rapid evolution over the past decade as it has
commercialized and scaled up, and it will be considered in more depth in another
paper to be presented. However, for the purposes of understanding non-mortgage
access to formal housing finance, how do poor households actually finance their
housing? And who are current providers serving? These questions will be addressed in


                                                                                     7
V4.0 Draft



Section 3, drawing on new data from Financial Diaries in three developing countries;
and on the client profiles and experiences of several leading housing microlenders,
representing different categories of provider—mortgage lenders like HDFC in India,
specialist microfinance groups like ACCION or ProCredit; or non-mortgage housing
lenders like the clients of the Rural Housing Loan Fund in South Africa; Patrimonio
Hoy in Mexico and international housing NGO Habitat for Humanity.

                               Box A: Developing country sample frame
Given the difficulties of assembling relevant data across developing countries and the resource
limitations of this paper, we focused on collecting data on a sufficiently large set of developing
countries to achieve the following objectives:
• Cover middle- and low- income countries
• Cover most geographic regions; and in the process
• Cover most types of developing country housing finance systems as identified by Renaud (1999)
The objectives must be balanced against the constraint that in each case, a local contact was
necessary in order to obtain and/or validate country-level data, which was obtained from or verified
with local correspondents.

The result was a sample of 8 countries: namely Armenia, Ghana, Honduras, India, Mexico, Pakistan,
Serbia and South Africa.

Some basic background indicators for each country are given in Table 1 below, together with a
comparison in the right-most column to an average for OECD high income countries.

Table 1: Country background profile
                                                     Hond-                            Pakist      South                  OECD
    Country                Armenia       Ghana                   India    Mexico                             Serbia
                                                      uras                             an         Africa                 High
                             Central                             South                 South                  Eastern
     Region                   Asia
                                           SSA         LAC
                                                                 Asia
                                                                            LAC
                                                                                       Asia
                                                                                                   SSA
                                                                                                              Europe
                                                                                                                              na


 GDP per capita--
                     $        1,620         840       2,788       728      7,403        728        5,277       3,322         36 715
 current $ 2005
 Population 2005    Mil         3           21          7        1,079      104         152         46           8
 Homeownership       %         93           57          80        87         78          78         72          89           43-80
 % Urban             %         64           48          56        29         76          35         57          52            77
 M2/GDP              %         15           29          59        63         27          45         66          na            79
 World Bank                                                                              Low                  Lower
                             Lower                    Lower      Low;      Upper                   Upper
 country                                 Low; IDA                                     income;                 middle;         Na
                             middle                   middle     blend     middle                  middle
 classification                                                                         blend                 blend
Sources:
GDP, population, % urban, M2/GDP: World Development Indicators, mainly 2005, except for M2/GDP (OECD) which is 2000.
Country classification: World Bank, via
http://web.worldbank.org/WBSITE/EXTERNAL/DATASTATISTICS/0,,contentMDK:20420458~menuPK:64133156~pagePK:64133150~p
iPK:64133175~theSitePK:239419,00.html
Homeownership: various years from official sources, accessed via list compiled by Andrew Gall
http://www.housingfinance.org/Content/ContentPage.php?interest=100370; except for South Africa: Porteous & Hazelhurst 2004




                                                                                                                        8
V4.0 Draft



These two perspectives are then brought together to provide a generic developing
market profile. In the conclusion, proposals are made to develop the analysis and
push out the access frontiers of the different segments of the housing market.

2. The changing frontiers of mortgage markets
“Only a quarter to a third of households in most emerging markets can afford a
mortgage to purchase the least expensive developer built unit.”
                                                                Ferguson 2004

Bruce Ferguson expresses above a commonly held view in housing finance: that
mortgages reach a minority of the population in developing countries. Is this
statement still true, given the rapid growth which has been observed in mortgage
markets in many developing countries? How can one benchmark the levels of
development of mortgage markets across countries, and what do we learn by this
process, especially for the issue of access to housing finance. Section 2.1 reports on a
variety of empirical measures of the size, depth and completeness of mortgage
markets across the sample of eight developing countries to help to answer these
questions. Using recent data, Section 2.2 extends the analysis of access in more depth
to two countries—South Africa and Zambia.


2.1 Empirical measures
There are at least three categories of currently available empirical measures of
mortgage markets, which will be applied as consistently as possible across the
developing country sample: measures of depth and growth; of affordability; and of
completenesss.

2.1.1 Depth & growth

The conventional measures of the depth of mortgage markets are:
      • Value of residential mortgages outstanding per capita; and
      • Value of residential mortgages outstanding to GDP.
Growth in the mortgage market may be measured by the change in the latter ratio.

In our sample of countries, mortgage balances outstanding per capita vary
considerably from $5 in Ghana to $1300 in South Africa. In most of the countries, the
number of mortgages (and even the number of households in some cases) is not known
with any certainty; the average number of mortgages per household is calculable only
in some: this average varies from less than 0.03% in Ghana to 10.9% in South Africa.
This compares with almost half (47.9%) of all US households in 2004 who had a
mortgage secured by their primary residence. Among US homeowners only, the
percentage rises to over two thirds (69.3%).1



1
    Federal Reserve Bulletin 2006, p.A26


                                                                                      9
V4.0 Draft



Figure 1: Residential Mortgages Outstanding to GDP



                        90                                                                                         78
      % Mortgage/ GDP



                        80
                        70                                                                                      59
                        60                                                                              45
                        50                                                                             36
                        40                                                           27
                        30                                                          16
                        20                     78       46        8 10
                        10   00      00                                    01                 11
                         0
                            M ia




                             Se a
                           Pa ico
                           on a

                                     s
                             G a




                                                                                                          SA
                          ut tan




                                 EU
                                   ia
                                 ric
                                 ra
                          H an
                                   i




                                  d
                                en




                                rb
                              ex
                               In




                                                                                                        U
                              du




                        So kis

                              Af
                              h
                             m




                            h
                           Ar




                                                  Prior year                Most recent
Note: years used in each case: Armenia: 2005/2002, Ghana 2005/2003—numbers so small as to round down to 0;
Honduras:2006/2000; India: 2004/2002; Mexico: 2006,2000; Pakistan: 2006/na; South Africa: 2006/2000; Serbia: 2005/2004; EU:
2005/ ; USA:2005/2001
Sources: developing country sample: see country references; EU; USA: M. Lea (2006)




Figure 1 above shows the ratio of residential mortgages outstanding to GDP for the
most recent available year (2005 in most cases) and a previous year, which varies by
country depending on data availability as reported in the note below the figure. The
numbers from the developing country sample are benchmarked against numbers for
the US & EU. Within the sample, there is a considerable range: from South Africa,
where the ratio approaches the levels of EU members in Eastern Europe, to Ghana and
Armenia where, at less than 1% of GDP, mortgage markets are nascent. However,
underlying some of the smallest numbers are rapid increases: in Serbia, for example,
the figure has doubled in a year to reach 1.4%, while in India, the ratio increased 55%
between 2002 and 2004. As reported in Box B below, financial analysts in the three
larger markets at least have noted these historically high rates of increase, which
show little signs of slowing.




                                                                                                                         10
V4.0 Draft




             Box B: Recent and projected breakneck growth in mortgage finance

India 2006: “..India has huge potential to ratchet up its property financing. It is therefore hardly
surprising that banks are turning increasingly to this market. Mortgage debt as a percentage of GDP
could plausibly rise from barely 5% at present to around 15% by 2015.” Deutsche Bank Research
2006:20,21

Mexico 2006: “From the beginning of the present decade, the strength to channel resources to the
financing of housing has increased. In 2005, the system financed 72% more dwellings than in 2001… At
the same time, the improvements in the conditions of finance have permitted a growing number of
families to have had access to a dwelling.” Translated from Spanish, Bancomer Situacion Immobiliaria,
August 2006:2

South Africa 2006: “In spite of the two interest rate hikes this year, mortgage advances by the
financial services sector increased 30,2% year-on-year last month, the highest growth on record since
1965, according to data released yesterday by the SA Reserve Bank.” N Wilson, Business Day, 30
August 2006




2.1.2 Mortgage affordability & penetration

Struyk (2005) outlines three typical measures used to assess housing affordability:

    •    Average house price/ average income ratio—this is a very commonly used
         measure, but it says little about the means of financing;
    •    The Housing Affordability Index (HAI) as used in Australia and by the National
         Association of Realtors in the US considers the relationship between the income
         to afford a representative house and representative income; or alternatively, a
         target income with target house price;
    •    The Housing Opportunity Index (used by NAHB in the US) measures the share of
         homes within a specific market that a typical household earning median income
         can afford to buy.

Struyk demonstrates the sensitivity of results to using only one indicator; and cautions
that there is sometimes too much focus on mortgage affordability and a consequent
failure to consider other costs of home ownership.

To assess affordability in the access tradition, I have sought to concentrate on the
lump-sum costs in each case associated with:

    (i) closing the purchase of a mortgage-financed dwelling, including the deposit
         required by the mortgage lender as well as any additional taxes and fees paid
         on transfer; and
    (ii) paying the monthly mortgage installment thereafter, which is a function of the
         percentage of house price to be financed, the term and rate.




                                                                                                  11
V4.0 Draft



For the purposes of this study, data was collected to the extent available in the eight
countries on the value of (a) an average house financed by mortgage; (b) the smallest
formally constructed house widely available; and (c) the smallest mortgage loan
generally available (which may not be for home purchase).

To enable cross-country comparison of the resulting nominal values of lump sums in
each case, I normalize the country’s lump-sum numbers, expressed in current US
dollars, by current (2005) monthly-equivalent GDP per capita in US dollars. 2

The measure provides an indication of the relative lump-sum financing burden across
countries: a higher number implies that relative to GDP per capita, the lump-sum cost
threshold is higher (and therefore less affordable) for more people. The results are
shown in Table 2 below.

Table 2: Mortgage penetration measures

                                                             Hondu                                 South
                                        Armenia     Ghana              India   Mexico   Pakistan            Serbia
    Installment on:                                           ras                                  Africa
 1. Average mortgaged house/
                                   %       371       1,079       217   778      58       1,608      420      220
 GDP pc (monthly)
 2. Smallest formal new house /
                                   %       NA         634        Na     0       22       1,891      160      87
 GDP pc (monthly)
 3. Smallest available
                                   %        74        507        109    72      12        197       12       24
 mortgage / GDP pc (monthly)
 4. Upfront lump sum required
                                   %      1,912      1,184       Na    1,636    27       2,335      237      815
 (incl deposit)/ GDP pc (monthly)
Sources: Calculated from data obtained from country references


Clearly, the first two rows of numbers are very sensitive to the average price of
housing assumed in each category—mortgaged housing stock as a whole; and the
smallest formally built house; as well as to the financing features—in particular, the
maximum term of loan and the maximum loan-to-value (LTV) ratio available.

Relative affordability is highest in those mortgage markets—such as Mexico, Honduras,
South Africa, India—which are relatively more developed; and in particular, those like
Mexico where a large subsidized mortgage system for formal employees extends the
reach of housing finance to lower income groups. Serbia is an interesting exception
since mortgage finance is relatively new, but it appears relatively affordable. By
contrast, the under-developed mortgage systems in Ghana and Pakistan are affordable
only to borrowers with much higher incomes than national averages.

Seen in this way, the eight countries provide an interesting cross-sectional view across
different stages of mortgage market development. Figure 2 below traces this
development from an early stage, where term and LTV is low, represented best by
Armenia today where mortgage markets are very new; moving to a second stage in
which loan terms are longer (over 10 years but less than 20) and LTVs rise to 80%,
such as in Pakistan; to a third position in which loan term is stretched to the

2
 Clearly, mean household income would have been a preferable measure but is not available for all
countries.


                                                                                                             12
V4.0 Draft



conventional maximum of 20 years (Serbia, Ghana, Honduras). From this third
position, some markets have stretched the boundaries further: for example, in
Mexico, low-income loans may be available for 25 years; and in South Africa, where
terms are limited to 20 years, but where LTVs may exceed 100% due to the ability of
borrowers to capitalize closing costs.

Of course, there is another key financing dimension not shown here: the interest rate
on the mortgage. Both its absolute level, determined primarily by capital market
conditions, inflation expectations and the competitiveness of the mortgage market,
and whether it is fixed or variable, affecting how rate risk is allocated between
lender and borrower, matter.


Figure 2: Mortgage Market Features (low income)


                    120                                                4a
                                                                                South Africa
                    100
                                                   2                                   Mexico
                                                                                Serbia
      Max LTV (%)




                                                             India
                    80                                      Pakistan            Ghana 4b
                                                                       3        Honduras
                    60
                                          Armenia
                    40
                                    1
                    20

                     0
                          0           5             10          15         20        25        30
                                                       Max term (years)

Source: General loan characteristics: country respondents




The progression observed in Figure 2 is summarized in Table 3 below, which shows the
four basic clusters of mortgage market features in stylized form as the basis for the
lumpsum calculation which follows in Figure 3. Figure 3 calculates the lump sum
required to acquire a house costing $42 600 (which is the simple average of mortgage
housing stock across these countries) with a mortgage loan based on the features in
the cluster indicated. Lump-sum costs here include both the one-off sum required at
closing (deposit plus any costs and taxes) and the first month’s installment, expressed


                                                                                                    13
V4.0 Draft



relative to a monthly income level at which the installment at stage 1 is considered
affordable (25% of income--$1983). Note that the change across the stages is not
expected to be linear for the options 4a, 4b or 4c, since they represent different
vectors of movement from stage 3 of one loan characteristic, different in each case.

Table 3: The progression of mortgage markets: assumptions

                                   0              1                 2               3            4a           4b          4c
                                                                                                                         More
                                                                 Lengthen
                              No mortgage                                        Lengthen    Maximize     Maximize    competitive
            Description                     Undeveloped          term; up
                                market                                             term        LTV          term       (4%-2.5%
                                                                   LTV
                                                                                                                        margin)
  Other costs/house
                                       5%              5%               5%              5%         5%          5%            5%
  price
  Interest rate                        Na             14%           14%             14%           14%         14%         12.5%
  Max LTV                              Na             50%           80%             80%          100%         95%         100%
  Max Term (mos)                       Na               60           120             240           240         300          240

                                                                              Ghana,
                                                Armenia          Pakistan                        SA        Mexico
  Country example                                                            Honduras




Figure 3: Housing finance and reduced lumpsums


                       25.0

                       20.0
      Lumpsum/salary




                       15.0

                       10.0

                        5.0

                        0.0
                               0            1                2               3              4a           4b          4c
                                                  Stage of development

Source: Lumpsum/ salary is the multiple of upfront costs (deposit plus assumed extra costs) plus the monthly installment due on
a house of $40,000 under the various product assumptions in Table above, relative to the monthly salary level at which the
installment at stage 1 is affordable at a typical 25% installment to income level.




                                                                                                                                14
V4.0 Draft



Figure 3 clearly shows that the biggest gains from smoothing lump-sum costs
connected with acquiring a house are to be found in the early stages of development—
moving from having no mortgage to stage 1 reduces the financing burden by 50%; and
it halves again between stages 1 or 2. Beyond this, the progression reflects tradeoffs:
rising LTVs reduce the initial deposit required but increase the monthly installment.
Beyond the point at which 95%+ LTV loans are available for 20 years or more, there is
relatively little impact on affordability from further changes: 4a,4b,4c are quite
similar. Even squeezing lender margins through competition from 4% to 2.5% (levels
which are common at the higher ends of the larger developing markets already), that
is, to position 4c, has relatively few effects at these levels.


2.1.3 Market completeness

The preceding analysis of affordability suggests that mortgage markets in developing
countries may be converging towards similar features of the basic mortgage product,
with many variations around this. The availability of a range of product features is
one aspect of the completeness of mortgage markets. In their landmark 2003 study of
various European mortgage markets for consulting firm Mercer Oliver Wyman (MOW),
Low, Dübel and Sebag-Montefiore have developed the approach further to incorporate
some access dimensions (which groups can receive mortgages) as well as considering
the choice of distribution channels available. Their methodology is summarized in Box
C below, and available in detail in Annex 3 of their full report via www.hypo.org. In
order to preserve comparability with the earlier work on EU countries, this
methodology was substantially followed for the developing country sample, with some
minor changes which are described in Annex A of this paper.

The completeness characteristics were assessed in each case by a country expert,
usually located in the respective country or at least with in-country support. The
underlying characteristics which generated the scoring for each country are reported
in Table 1 of Annex A. Figure 4 below compares the overall scores for the developing
country sample, using two EU countries as a benchmark with the caveat that their
scores were calculated for 2003 and therefore not strictly comparable.

After the preceding analysis of sequential development, it is hardly surprising that
Figure 4 shows a wide variation in completeness scores, from 20% in Armenia, where
mortgage lending is new, to India, where a large number of lenders offers a wide
range of products, resulting in a score of over 80%. What is more surprising is that
even very small mortgage markets, such as in Ghana, which until June 2006 had only
one provider, score comparatively highly. Although the methodology may lend itself
to some subjectivity in assessing availability, more importantly, this suggests that
mortgage markets may have a wide and diverse product set, which may be open to a
range of groups, but remain very narrow in their outreach. ‘Completeness’ alone,
while a useful indicator of relative sophistication, provides little guide to the depth or
penetration of a system: other indicators such as the depth and affordability



                                                                                       15
V4.0 Draft



measures introduced earlier are necessary to provide multi-dimensional perspective
on the state of development of a national mortgage market.


                 Box C: MOW Measure of Mortgage Market Completeness

In 2003, authors Low, Dübel and Sebag-Montefiore, for consulting firm Mercer Oliver Wyman (MOW),
published the results of a study of mortgage markets in EU countries, commissioned by the European
Mortgage Federation, with the aim of assessing the potential benefits, costs and obstacles associated
with integrating European mortgage markets. The project involved the collection of extensive
comparative data on prices, costs, product range and profitability of mortgage markets across 8 EU
countries.

As one part of the exercise, MOW developed a measure of mortgage market completeness. This
measure involves scoring the product and channel attributes of a national mortgage market against a
hypothetical full list of desired attributes in categories such as:

•   Risk tolerance (weighted 35%) —which borrowers can access a mortgage;
•   Product Range (weighted 50%)—what range of products is available;
•   Distribution (weighted 10%)—how easy it is to access the mortgage product; and
•   Availability (weighted 5%)—how easy it is to access information and advice on mortgage products.

The index has a maximum value of 100. In 2003, European markets surveyed were found to have score
ranging from 86% (UK) to 47% (Portugal). Annex 3 of the MOW report provides in-depth breakdown of
the completeness methodology. The MOW report notes (p.23):”..the completeness index identifies the
extent to which there are gaps in an individual market’s product range, distribution or range of
borrowers served relative to those available in other countries. We do not assess here whether there is
a specific need for each product in the country or whether the need is provided outside the mortgage
market.”

The MOW completeness scores have since been cited in various official documents, including CGFS and
Miles Report on the UK Mortgage Market. Although the product list arguably reflects European
experience, with scores obtained for example for mortgages on overseas holiday homes, the
completeness measure nevertheless provides a starting point for consistent cross-country analysis.
Hence, we have applied same methodology, with a few minor exceptions, to calculate the equivalent
scores for the developing country sample. Since the score represents the current (2006) position in
these countries, it cannot be directly compared with 2003 European measures; these are likely, if
anything to have increased in the intervening three years. Nonetheless, the EU measures provide rough
developed market benchmarks, and the 2006 results are at least comparable across these countries.




                                                                                                    16
V4.0 Draft



Figure 4: MOW Completeness measures across country (max=100)


     100
      90
      80
      70
      60
      50
      40
      30
      20
      10
       0
                                                a




                                                                                                       3)
                                                                an




                                                                                                        )
                         na
                  a




                                                                 o




                                                                                     ia
                                   as




                                                                a




                                                                                                      03
                                             di
                   i




                                                              ic




                                                            ric
                 en




                                                                                   rb

                                                                                                     00
                                                             st
                       ha


                                  r

                                           In




                                                                                                    20
                                                           ex
                                du




                                                                                 Se
                                                          ki


                                                          Af
                m




                                                                                                  (2
                       G




                                                         M




                                                                                                 l(
                                                       Pa
                              on
             Ar




                                                        h




                                                                                               K

                                                                                              ga
                                                      ut
                            H




                                                                                       U

                                                                                           rtu
                                                    So




                                                                                         Po
Source: UK, Portugal: MOW (2003); others: calculated from data obtained from country sources.




2.2 Going deeper: access frontiers

The preceding section has analyzed various cross-country measures of mortgage
depth, affordability and completeness. However, these numbers alone still give little
sense of the extent to which mortgages are accessible across a population and in
which segments. Those with access include those who use a product, by definition,
but access typically exceeds usage, since non-users may have access to a product but
choose not to use it, or could be denied access on a range of grounds—such as product
eligibility criteria or income. The access frontier approach, described below in Box D,
seeks to separate out the underlying components of demand and supply in order to
ask the questions: what are the limits of access in the current market situation, and
correspondingly, how might the frontier, or current outer limit, of access be pushed
back over time?




                                                                                                            17
V4.0 Draft




                              Box D: The access frontier approach

Two recent papers seek to provide a tool with which to analyze how markets provide access to goods
and services over time. Beck & de la Torre (2006) provide rigorous micro-economic foundations to
develop an ‘access possibilities frontier’ for the provision of financial services. The frontier is shaped
by the identification of different demand and supply constraints for a set of state variables. This allows
one to differentiate between outcomes below a constrained optimum and outcomes where the
constrained optimum is too low and where the outcome is too high. Using a similar approach without
the micro-econometric foundations, Porteous (2005) describes the access frontier as the maximum
proportion of a population which can access a particular good or service using available products and
technology. He uses the device of a ‘market map’ to indicate the trajectory of usage in a market over
time, as well as to capture the different zones of provision. The key insight of the tool is to distinguish
the different categories of access to a good or service that are possible. Both papers use similar
nomenclature to delineate a current and potential market into several zones:

•   Current market zone—which includes current users;
•   A market enablement zone—which includes those who do not currently use the product but are
    potentially eligible to access the product, and do not obviously self-select not to use it;
•   A market development zone—which includes those who cannot access the product now because of
    structural features (such as location or characteristics)
•   A supra-market zone—which comprises those who are denied access by virtue of their income
    alone, and may need non-market intervention such as subsidies if they are to participate in this
    market.

In order to map a market into these zones, it is necessary to have access both to supply-side
information about product criteria (such as eligibility) and demand side information about users and
potential users, including information about why they currently do not use a product. The FinScope™
surveys, developed by FinMark Trust in South Africa, are examples of surveys that provide necessary
demand side information for access frontier-type work. FinScope surveys use a comprehensive face-to-
face interview on the financial service needs, usage and attitudes of a statistically representative
sample of adults in a country. Note that, as a survey based on respondent knowledge, answers to
questions like “Do you have a title deed?” could be inaccurate or unknown (an allowed category). To
date, FinScope surveys have been completed in South Africa (2003, 2004, 2005), Botswana (2005) and
Namibia (2005), where the costs are syndicated in whole or in part with financial sector institutions;
and Zambia (2006) where the costs are financed by donors. Further surveys are underway in 5 further
countries (Tanzania, Ghana, Uganda, Kenya, Pakistan), with results expected by 2007 (see
www.finscope.co.za). This data will enable further detailed analysis to be performed across these
countries.

Meanwhile, the access frontier concept has been applied empirically to the market for bank accounts
in South Africa (Porteous 2005), insurance in Ukraine and Georgia (Matul 2005) and mortgages in South
Africa (Meltzer 2006).

This approach will be applied in this section to better understand the question of
‘how low can mortgage markets go’?

Access to a mortgage is a function of:

    •    borrower criteria, which define the risk of default and commonly include
         source and proof of income, age, level of income (to define affordability) and
         previous credit record; and


                                                                                                        18
V4.0 Draft




    •    property criteria, which define the loss given default and commonly relate to
         the valuation of the property, linked to its physical condition and how easy the
         value is to realize. The latter is a function of the tenure and court system.

To assess the effect of applying typical borrower criteria requires a nationally
representative database in which various borrower characteristics can be cross
tabulated with financial product usage and reasons for non-usage. To date, I am
aware of this information being available on a consistent basis only for one of the
eight countries, South Africa, although more relevant data will be forthcoming for
other developing countries through FinScope surveys in 2007. Apart from South Africa,
FinScope data for Zambia, a low-income country, have recently become available3.
Since South Africa has a stronger than usual mortgage market for a developing
country, the inclusion of Zambian data helps to provide a low-income country
counterpoint where the reach of the mortgage system is currently negligible (less
than 0.1%/GDP or less than 50c (US) per capita in Zambia today)4.

To assess the application of the property criteria, ideally an inventory of housing
stock with details of tenure and condition is required, together with the ease of
valuation. To my knowledge, such databases are not available on a national level, at
least in the sample countries. Nonetheless, household surveys may serve as a proxy
for this data. Access to a mortgage is limited first and foremost to property owners;
and among them, to those owners who have ‘mortgage-able’ tenure. Those with
informal tenure do not have access to a mortgage, even if they claim to own; nor do
renters, although they could become owners. Within the group with mortgage-able
tenure, the property must meet further physical criteria to assure the lender that the
structure will survive the term of the loan, and more importantly, will at least
preserve its value as collateral for the loan.

Using the tenure perspective as the first level segmentation, the national market may
be divided into four main categories:

    1. Current mortgagees, who are assumed de facto to meet all criteria for
       borrower and property;
    2. Owners with formal title to their property but no current mortgage;
    3. Claimed owners without formal title; and
    4. Renters.

Using FinScope data, Figure 5 below shows the relative size of these categories for
South Africa and Zambia as a percentage of all adults. The table which follows then
compares various characteristics of each category.



3
  Note: Central Bank of Zambia which is the owner of FinScope Zambia database will launch it publicly in October
2006; findings using the data can be used publicly only after this date.
4
  Source: Building Society consolidated balance sheet for December 05, reported by Central Bank of Zambia,
www.boz.com.


                                                                                                              19
V4.0 Draft




Figure 5: Mortgage market access lens




                        Zambia 1.4 15.7                              64.4                          19.9
      % of adults




                    South Africa 7.4                    45.0                    23.5             24.2




                                   0%           20%           40%            60%           80%          100%

              1. Mortgageholder                                  2. Formal title, no mortgage
              3. Informal title                                  4. Renter
Source and definitions: SA: FinScope 2004. Q refers to relevant question number in SA database; Current mortgage holder: Q49;
Ownership status: Q48a,b,c;
Zambia: FinScope Zambia 2006, data extraction by Christian Keulder of FinScope

Not surprisingly, the percentage of people with mortgages and the percentage with
formal property title are much higher in South Africa than Zambia, reflecting higher
average incomes and urbanization level. More surprisingly, the level of renting is
similar. Table 4 decomposes the four groups further by other characteristics.




                                                                                                                           20
V4.0 Draft



Table 4: Characteristics of tenure lens groups: South Africa & Zambia
                                                                                              3. No
                                                            1.Bond &        2. No bond,       bond,
            % of total in each column group                formal title      formal title   informal   4. Renters
 Of each column category:
 1. Location: % urban
      SA                                                      98.5               57.1        53.4        87.3
      Zambia                                                              64.1               20.8        73.6
 2a. Housing situation: living in a formal (SA)/ brick
 (Zambia) house
       SA                                                     99.8               74.7        62.5        89.2

      Zambia                                                  100                80.1        46.8        79.3
 2b. % with water, sewage in house and cook with
 electricity (SA); use electric stove with oven (Zambia)
         SA                                                  95.4        40.8           25.9       71.8
         Zambia                                                    46.5                  4.8        28
 3. Attitudes
 3a. % who would invest in home improvement
         SA                                                  25.1        25.4           19.5       24.6
         Zambia                                                     28                   29         26
 3b. % who see home as tradable asset
         SA                                                  78.2        34.4           13.1        0.0
         Zambia                                                    83.1                  39        0.0
Source: as for Figure 5 calculated from FinScope SA (2004); FinScope Zambia (2006), in addition:
2a & 2b. House quality: Type of house Q46; services in house: Q35
3a. Home improvement: answer to question 14a, l 21: “would consider investing in improving own house”
3b. Tradable asset: Answer to Q48d


Several observations may be made on the above table:

     •    The desire to improve one’s house (line 3a) is consistent across the groups and
          across both countries at around a quarter of people - i.e., it is not correlated
          with tenure, even extending to renters.
     •    Believing that one’s house is a tradable asset is unsurprisingly correlated with
          tenure, but particularly with having a mortgage.
     •    While current bond holders (Group 1) are the smallest of the four groups in
          both countries, they are also the wealthiest and most urban.
     •    Group 2—with formal tenure, but no bond—is the largest single group in South
          Africa, and is relatively large even in Zambia (15% of the population). Within
          this group, only around 70-80% actually lives in a formal house; and less than
          half have a house with basic facilities such as running water inside the house
          and sewage, which conventionally define the limits of mortgage-ability.
     •    Group 3—owners without formal tenure—are more rural than other groups as
          expected; and this is the largest single group in Zambia today. Titling is
          therefore a bigger issue there.

In order to define the limits of access to mortgages in these countries today, one
clearly has to analyze the characteristics of borrowers in Group 2 further since the
mortgage access frontier must lie inside this group. Those who own but without
formal tenure (Group 3) are clearly beyond the reach of the mortgage instrument, as
are renters by definition unless and until they opt to purchase a house. In fact,


                                                                                                                    21
V4.0 Draft



renters in both countries are relatively better off on average than Group 3 members,
hence some may indeed be able to access a mortgage when and if they choose to buy.
In South Africa, some 3.1 million people (11.5% of the population) have incomes,
which suggests they could do this.

In fact, the minimum household income necessary to support the smallest available
mortgage can be calculated as being the income at which an installment and other
housing-related costs can be paid, leaving sufficient income to cover expenditure
above a nationally defined expenditure line. That income threshold is today
considered by government and banks in South Africa to be around $220 per month,
which is the lower limit of the Financial Charter market targeted by banks in terms of
their social commitments to government.5 A similar exercise to define minimum
thresholds has yet to be performed in Zambia. This income threshold alone reduces
the eligible number in Group 2 in South Africa from 11.2 to 4.9 million adults.
Furthermore, of this number, 1.6 million are over 55 years old and would be unlikely
to qualify for a mortgage. If one applies the minimum income and maximum age
eligibility criteria, together with the requirement that the house be formally built and
have basic services (such as water and sanitation), only 2.1m people, or 17% of Group
2, are considered to presently have access to mortgage finance, though they do not
presently use it. This is equivalent to almost 8% of adult South African population.

In access frontier language therefore, some 7% of adults in SA are current mortgage
holders, while a further 8% of Group 2, and potentially 11.5% from the renter group,
(that is, a total of 26.5% of adults), may be deemed to lie within the current access
frontier. Others among the renter group are likely to have access: Note, however,
that because FinScope data is weighted primarily to individual adult level rather than
household, these are not household numbers. For mortgages in particular, household
numbers are more relevant.

This, however, is only a preliminary indication of how levels of access to the mortgage
market may be assessed. Recently, Meltzer (2006) has applied the access frontier
approach in more depth, using all available household survey data on households and
housing types, to analyze only the so-called ‘Charter target market’ for mortgages in
South Africa: households with monthly income between $220 and $11006. A third of all
SA households, some four million, are in this income range, which itself lies between
two other groups:

    •    Upper income households: some 20% of households which earn more than $1100
         per month, for whom mortgage finance is generally accessible based both on
         income and place and quality of residence; and
    •    Poor households: the balance of almost half of households earning less than
         $220 per month, who are considered by agreement between banks and

5
  South African banks committed to make some $6 billion in relatively low income mortgage loans over ten years,
on a total current mortgage book of around $60 billion. For full Charter wording, see www.banking.org.za.
6
  Exchange rate: 1US$=6.8ZAR


                                                                                                              22
V4.0 Draft



         government to be too poor to afford formal housing finance offered by formal
         institutions, and to whom government subsidies are targeted.

Separating out the various access criteria for this middle group, Meltzer defines the
various access zones as in Figure 6 below.

Figure 6: Access frontier in the Charter Target Group: SA 2005
                      Market                                       Market                                 Market
       Current
                    enablement                                  development                            redistribution
       market
                       zone                                         zone                                   zone




             5%     12%          8%                                53%                            1%        20%

 Number of
 households




             Currently have and use the product                      Have access but do not use
             Do not want a mortgage                                  Do not qualify
             Qualify but cannot physically access the product        Too poor



Source: Meltzer (2006:3)



Meltzer’s analysis suggests that at most, 25% of the Charter target group (the first
three bands from the left in Figure 6 above) can potentially access a mortgage,
although only 5%, or one-fifth of those with access, currently has a mortgage. Eight
percent, or close to a third of the group, are judged not to want a mortgage on the
basis of factors such as age, leaving just 12%, or 480,000 households which comprise
the prime target group. Of the rest of households in this range, just over half (53%) do
not qualify for mortgages based on current lender criteria although they may have
access if these criteria change. Fully a fifth of households are considered too poor, in
that even if there income falls within the range, household-level indicators of poverty
(such as regularity of going without food) suggest that the additional burden of a
mortgage would not be sustainable.
Loosening finance eligibility criteria cannot solve a general problem of lack of income
– indeed, it may exacerbate it. This access frontier type of analysis adds this
perspective to the question currently debated in some of the sample countries of
whether mortgage finance is constrained more by conservative loan criteria or by the
unavailability of affordable properties to buy. The Banking Association of South Africa
recently released research on the supply of affordable units (Settlement Dynamics et


                                                                                                                        23
V4.0 Draft



al 2005). On average, fewer than 18,000 new housing units affordable to the target
market have been formally completed and registered each year (as required by law).
Deeds office data reflects that there were around 72,000 transactions in which
properties with a theoretically affordable price changed hands in 2004. The report
concluded that the problem was primarily one of supply, and that there was a need to
promote faster development procedures to reduce the price of new stock, as well as
to promote the development of the resale market.
In Meltzer’s assessment, there are around 1 million households (close to 10% of the
national total) in the Charter target group within the current mortgage access
frontier. Combined with the existing 20% of households which are upper income and
currently have or have access to mortgages, just under a third of South African
households have access to mortgages today.
As a counterpoint in another middle income country with a relatively developed
mortgage market: SHF in Mexico has calculated that only households earning above
$550 per month can afford a mortgage (Babatz 2005:16). This is 40% of the total
number of households in Mexico. However, this number is based on borrower income
alone; and would also have to be assessed against property characteristics to be
comparable to the number above.
However, in general, Ferguson’s norm of mortgages reaching only a quarter to a third
of households appears to hold in these middle-income countries; certainly, it is well
below a majority. In low-income countries, the percentage is far lower—in Zambia,
the maximum percentage with access based on having formal tenure alone would be
perhaps 8%, and perhaps lower if borrower characteristics were also considered.7
This limited reach of the major instrument of housing finance in developing countries
underlines the need for careful analysis of the art of the possible for mortgage
finance. Nonetheless, the access frontier can be pushed outwards over time through
measures that improve tenure for those without formal tenure; and improve the
process of risk assessment and management among the growing group in urban areas
who may have formal tenure but no mortgage (Group 2). Even where mortgages are
legally possible, they may have limited value as collateral where house prices are
uncertain or cannot be realized. In such circumstances, mortgage lending assumes
more of the characteristics of unsecured lending or micro lending, where the lender
has to manage its relationship with borrowers more intensively.




7
 Calculated as the 1% with mortgages plus the proportion of households with formal tenure (15%) who have
mortgage-able houses measured by basic services (46%).


                                                                                                           24
V4.0 Draft




3. The poor and their housing finance: a bottom-up view
“As understanding of the need for and methods of housing micro finance grows,
however, it is becoming clear that knowledge of its use at household level is lacking.
There are almost no studies of how poor households 'turn money into house'. This is
problematic in two ways. On one hand, those who favour conventional mortgage
models for the poor tend to assume that households are incapable of utilising loan
funds effectively, leading to a bias in favour of 'developer-driven' construction
systems in which households are passive 'beneficiaries' of housing, even though they
are expected to repay the loans that finance it. On the other hand, proponents of
micro credit for incremental housing development often assume that poor households
have the skills and opportunities to use such loans effectively - 'turning money into
house'. In both cases, the focus is on finance rather than the end use of that
finance.”
                                                     Baumann in Kuyasa 2005:7

If mortgage finance at best reaches much less than half the population in developing
countries today, how are the housing finance needs of the majority met? This section
compiles evidence from the demand and supply side of housing finance specifically for
poor households in developing countries, in order to help create a platform for other
sessions at the Symposium. Housing microfinance is here understood to mean the full
range of financial services offered to and used by low and moderate income
households. This client-based definition reflects the growing understanding, expressed
for example in the recent CGAP publication Access for All (Helms 2006), that, first,
microfinance is about more than microcredit only, but also includes inter alia micro-
savings and micro-insurance; and second, that microcredit is not specific to the usage
of the loan instrument. While traditional microcredit focused on providing working
capital to micro-entrepreneurs, borrowers often used their loans for other purposes
too. Today, microfinance institutions provide an increasingly wide range of loan
instruments, unsecured and secured, to meet the needs of their clients; even
mortgages. Housing microfinance is therefore not only about unsecured home
improvement loans, although this restriction is still common.

As microfinance integrates into the mainstream of low-end retail financial services,
the definitional boundaries are therefore increasingly blurring. While this integration
is welcome, it also means that microfinance has begun to lose its ‘halo’ effect as a
special poverty-alleviation tool.


3.1 Which housing finance instruments do poor people use?

Even though, as Fay & Wellenstein report from studies among the urban poor of Latin
America in particular, “housing microfinance is growing (among the urban poor) and
looking very promising” (2005:108), there is as yet little firm published evidence of,
in Baumann’s words above, “how poor people turn money into house”. Fortunately,


                                                                                    25
V4.0 Draft



the “Financial Diaries of the Poor” methodology enables a finer grained understanding
of the financial instrument usage of poor households. The Diaries methodology was
first developed by Stuart Rutherford in Bangladesh and has subsequently been refined
through various iterations. It is based not on one-off surveys, where respondents are
more likely to under-report certain categories (such as credit) and/or misunderstand
survey questions, but rather on repeated interaction with the survey households over
a prolonged period, usually a year. Diary-type exercises have now been completed in
three countries—Bangladesh, India and South Africa, as reported in Box E below, and
are being considered in others such as Brazil and Zambia.

From her analysis of the financial diaries of poor households in these countries,
Collins (2006) shows that, of the large8 lump sums raised by these households, 13%
were used for housing purposes. Housing was the second largest usage category for
these sums: in India and Bangladesh, business investment in livestock came first,
reflecting the mainly rural profile of the households there. In South Africa, almost
half of households had acquired their houses via incremental building processes
financed by savings. In rural areas, inheritance was a relatively common route to
ownership. However, very rarely did households acquire a home using credit from
formal (bank or retailer) or informal (family or money lender) sources.

During the Financial Diaries year in South Africa (2004), even half of the very poor
group managed to find funds for some sort of expenditure on housing, spending on
average 4-6% of monthly income on housing. This percentage was similar even for the
relatively wealthier group. However, it is still far below the average percentage of
consumer expenditure on housing in India (11%) and SA (12%) as a whole, which in
itself is below the norms of developed country which range from 18% (US, UK) to 28%
(Sweden, Denmark).9 The lower housing expenditure in developing countries is both a
cause and an effect of undeveloped housing finance there: one the one hand, the
figure is low because few households have to meet the expense of bond repayments;
on the other hand, few people have bond repayments in part because any additional
housing investment above these low levels would have to displace household
necessities from the budget.

The Financial Diaries also found evidence that a number of urban households make a
main, or at least significant, housing investment in rural areas. In Bangladesh, for
example, 15% of households squatted or rented in the city but owned land in their
home (rural) village. 24% of Indian households in the urban Diaries sample raised lump
sums for home construction, mainly in their home villages. The desire of urban
workers to improve their rural houses is the basis of the business model of most of the
lenders supported by the Rural Housing Loan Fund in South Africa10. Financial security
was not the strongest motivation for this pattern of investment, but rather a sense of
place and belonging close to relatives. As links to rural areas become more remote

8
  Where large is defined relative to average monthly income
9
  Euromonitor, World Consumer Lifestyles DataBook 2005, Table 3.122. for 2004: housing expenditure as % of
total consumer expenditure.
10
   For more information, see www.rhlf.co.za


                                                                                                             26
V4.0 Draft



among generations of urban migrants, it will be interesting to see how this investment
pattern changes over time.

                              Box E: Financial Diaries of the Poor

“The financial lives of the poor are complex. Household membership and sharing arrangements are
ever changing and often ambiguous, incomes come from a variety of sources and livelihoods and cash
flows are tiny and irregular. The first step to address the challenge of providing appropriate financial
products to the poor is understanding the financial arrangements in which a household is already
engaged.

The Financial Diaries studies aim to fill that gap by continuously tracking a small number of households
across an extended time period. The first Financial Diaries study took place in Bangladesh in 1999.
Forty two households were interviewed across two research sites – one in the slums of Dhaka and the
other in a rural village. Households were seen every fortnight for one year and asked about their
financial transactions over this time.

The next study, in India, took the lessons learned from the Bangladesh Diaries and applied a more
rigorous framework, particularly in the area of livelihoods. This study used 48 households across rural
and urban sites and took place from September 2000 to September 2001.

Most recently, the South African Financial Diaries took a leap in rigor, using a specially built relational
database to track daily cash flows across 152 households from November 2003 to December 2004. This
study took place in three different areas: Langa, an urban township; Lugangeni, a rural village; and
Diepsloot, a peri-urban township. This report draws largely on the larger and more quantitative
dataset resulting from the South African Financial Diaries, using supplementary information from the
Indian and Bangladesh Diaries.

In all three countries, the sample was drawn by means of a participatory wealth ranking. This
methodology has been shown to be robust in identifying poor households in many countries…This
method of sample selection gave a broad selection of households across different housing, wealth
levels and neighborhoods in an area.

The richness of the Financial Diaries allows us to delve deeply into the financial decision-making of the
poor, to understand a bit more about what drives their economic behavior. In all three countries, the
quest to own a home or a piece of land was a key factor in their financial aspirations. The respondents
we interviewed in the three countries spanned a number of different types of homes. In the rural
areas, most respondents owned their homes (which varied in terms of strength and comfort of
construction) and the parcel of land on which they stood. The luckiest would have a large piece of
land to farm (in the case of India and Bangladesh) and a larger compound (in the case of South Africa).
In the urban areas, household might live in a mud hut, or a shack, or even (in some cases in South
Africa) a permanent brick home. In the urban areas, there were also more households who rented a
home. We learned, however, that respondents with shaky tenure in the urban areas could
simultaneously be building a home in the rural areas.”

Source: Daryl Collins, from a background paper “Housing and the finances of the poor” commissioned
for this paper using data from the three Financial Diaries projects



One would expect that households with secure tenure would be more willing to invest
in their housing, since they can expect to capture all the benefits. Interestingly,
however, even shack dwellers with insecure tenure among Diaries respondents in



                                                                                                        27
V4.0 Draft



South Africa spent on average 3% of their income on housing, about the same level as
those in rural areas, but less than that of people in formal areas.

There is therefore evidence of considerable housing-related expenditure and activity
among the poor; but is this always, or even usually, a good thing for them? Fay &
Wellenstein (2005) have raised an important question about the underlying rationality
of promoting home ownership among the urban poor given illiquid housing markets.
Liquid assets are vital to urban poor as a means of dealing with frequent income
shocks; but because much urban housing of the poor is not readily saleable, housing
may have poor investment characteristics relative to household needs and objectives;
in this sense, poor urban households may be over investing in their housing, whether
or not they have tenure. Fay and Wellenstein did not have sufficient evidence on hand
to answer their question about over-investment in housing; but clearly, working resale
markets in poor urban areas are highly desirable and necessary for improved economic
welfare. Greater housing market liquidity would also reduce the risk to mortgage
lenders since collateral becomes easier to value and realize.

A major research project entitled “The Workings of Township Property Markets”
(TRPM 2004, available via www.finmarktrust.org.za) has considered the state of
resale markets in poorer urban areas in South Africa in some detail. The project
identified four distinct types of urban property market in township areas, in which
low- and moderate-income black South Africans were empowered to hold title to
houses in urban areas only within the past twenty years. The turnover of houses in
these areas was considerably below the level in older areas with similar income
profiles, although there were signs of an upward trend in more formal categories. This
suggests that in part, it is simply a matter of time for first-time homeowners to enter
the resale market as sellers; and for a mature market to develop in new housing
developments in which first time buyers are concentrated. The recent residential
property boom in suburban areas has also started to trickle into township areas, as
buyers have sought to locate more affordable existing housing stock, and lenders have
been more willing to finance houses in these areas as the value and potential liquidity
has improved.

Interestingly, the TRPM household survey found evidence of an informal resale market
more active than that in certain formally tenured segments. Almost a quarter of
households who live in informal (shack) areas where no formal tenure was available
reported ‘purchasing’ their dwelling during the preceding five year period; and they
perceived that they had high levels of security of tenure, regardless of their legal
status.

That there is informal transaction activity at the low end of the market bears out one
observation underlying the so called ‘dead capital’ hypothesis of Hernando de Soto
(2000). It also leads to the key conclusion of The Mystery of Capital: that, although
informal markets are often vigorous and may replicate formal markets, because they
lack formality, they bar homeowners from participating in wealth accumulation. The
proposition that secure land titles are a necessary condition for sound mortgage


                                                                                    28
V4.0 Draft



finance at least is self-evident; but it is also becoming clearer that titling alone is far
from sufficient to enable a working housing market to develop.

Indeed, in a large scale econometric study of homeowners in Peru who benefited from
the large scale titling process starting in the 1990’s, Field & Torero (2006) find little
evidence that commercial banks increased their lending to households who obtained
their title through the titling program, controlling for other borrower characteristics.
However, there is evidence that households with tenure borrowed more under
mortgage from the state owned retail housing bank, Materials Bank. Field and Torero
suggest that one reason why titling does not seem to improve access to private credit
may in fact be a perverse incentive effect from tenure: newly entitled households
showed less fear of losing their property in case of default since their title was
conferred as part of a politically significant process. This attitude would cause private
lenders to place less reliance on the presence of title as realizable collateral. It is
clear therefore that title in itself has no magic effect on access to credit—rather, it is
necessary insofar as it allows and encourages the development of an orderly resale
market in which home owners, and their secured creditors, have sufficient security to
be able to realize value when they need or choose to.


3.2 Who connects low income populations to formal housing
    finance?

Across the various Diaries surveys to date, one of the striking findings is the very low
incidence of usage of formal and semi-formal11 financial providers. Even in
Bangladesh, where microfinance penetration is relatively high, Rutherford (2002)
found that 88% of lump-sum transactions by poor households in his sample were
accumulated in the informal sector (i.e. among family, friends, informal lenders). The
proportion of usage of the microfinance sector was highest (at 16% of transactions) for
the upper poor group. In South Africa, Collins found that, for a composite household,
70% of financial transactions were informal.12 This is in line with the Honohan’s (2004)
observation that, even in the countries like Bangladesh or Indonesia with the most
active microfinance sectors, the penetration of microcredit is low: measured as
percentage of population, it does not exceed 15% and is usually much lower, or as a
percentage of total domestic credit, 3%.

Although the reach of formal microfinance may be low in absolute terms, it has been
growing fast in many places. This is in part due to a widespread process of
commercialization which, in a number of countries, has turned leading NGO
microfinance providers into regulated entities, able to raise funding more easily and
offer a wider range of products than small working capital loans only. In particular,
although microfinance institutions, such as Grameen Bank, have had housing loan

11
   Rutherford uses this distinction to apply to non-regulated MFIs in Bangladesh; in other places, the two categories
are hardly distinguishable.
12
   See www.financialdiaries.com/highlights; Financial Institution Synopsis sheet.


                                                                                                                   29
V4.0 Draft



programs since the 1980’s, housing microfinance has started to attract significant
attention, and to raise significant expectations, in the past decade. This section seeks
to understand who the clients of housing microfinance providers are, using available
borrower profiles. These profiles are available only at an individual institutional level,
and not consistently measured there, so the full picture must be tentatively pieced
together through the lens of representative institutions.

Housing microcredit is today typically provided by three categories of lender:

     •   Traditional microfinance providers (MFIs), which started providing loans to
         microbusinesses and have more recently added housing microfinance. In this
         category are the Grameen Bank, which started its housing lending as early as
         1984, ACCION affiliates such as MiBanco in Peru and ProCredit subsidiaries,
         such as ProCredit Serbia, which have started more recently (generally from the
         late 1990s).

     •   Traditional mortgage banks or commercial banks, which serve the
         mainstream retail market but have downscaled by introducing low-end housing
         finance products, and may or may not have any interest in or control over the
         use of loans by clients.13 HDFC in India is an example of a successful mortgage
         lender, which introduced a housing microfinance program in 1987, providing
         wholesale loans to NGO’s to onlend to their members.

     •   Specialist housing microfinance providers (HMFIs), which focus only or
         mainly on housing microloans; and may be divided into several sub-categories:

             o Commercial housing microlenders, such as the clients of the Rural
               Housing Loan Fund in South Africa and more recently in other developing
               countries. This group has usually been separately identified from
               consumer lenders in general only when there has been a special source
               of funding, such as a state apex, which has encouraged or even
               subsidized a housing-focused approach and placed restrictions on its
               clients;
             o NGO housing microlenders or housing providers, such as Habitat for
               Humanity, which have traditionally provided mortgages on houses built
               under its program, but often on preferential terms (no or low interest,
               although some adjustment to the real cost of construction is often made
               during the repayment term);
             o Building material suppliers who sell on credit, of which the highly
               structured savings to credit program Patrimonio Hoy of Cemex in Mexico
               is an example (see for example, the case study in Prahalad (2004) or
               more recent article by Segel and Meghji (2005)).

13
  For example, among large commercial microlenders, targeting moderate to low income salaried clients but with
no restriction on usage of small, short term unsecured loans, borrower surveys have consistently shown that around a
third of borrowers used loan proceeds for the purposes of home improvements (ECI 2004).


                                                                                                                30
V4.0 Draft




Table 5 below summarizes what is known of the profile of clients of a leading
institution in each category.


Table 5: Client profile of borrowers
   Category        No. of HMF        Product/s                    Borrower               Source
                      clients,                                     income
                      Country                                   classification
1. MFIs
Accion affiliates     $117m               Unsecured            Mainly upper poor,   Accion Insights
(Peru, Bolivia,       outstanding;        microloans for       near poor            5,8,13
Haiti)                42,000 borrowers    home
                      at present; 6% of   improvement; also
                      total loan          mortgages in some
                      portfolio of        cases
                      network
2. Traditional mortgage lenders
HDFC                  $10m outstanding;   Microloans through   Economically         HDFC Innovation
(India)               140,000 clients     NGOs for shelter     weaker sector;       Case Study;
                      cumulatively        and home             income <$80 per      Correspondence
                                          improvement          month
                                          purposes
3. Specialized housing lenders
3a. Commercial        50,000 households   Unsecured micro      Near poor; and       Pearson & Greef
HMFIs: RHLF           cumulatively;       loans; 1-4 years     non-poor. Almost     (2006)
clients                                   term                 all h/holds
(South Africa)                                                 income>$300pm;
                                                               67% < $850pm
3b. Housing           $70m portfolio to   7-15 year            Minimum Wages        Interview with
NGOs—Habitat for      date; average       mortgages on         (MW) 2-5; poor to    Christy Stickney,
Humanity              6800 new loans      houses built         near poor            Director LAC
(LAC)                 p.a.                through assisted
                                          community build
                                          process
3c. Building          145,000             Savings to           MW2-5 (household     Segel & Meghji
material suppliers:   cumulative          unsecured loans      income $280-$900     (2005); Meeting
Patrimonio Hoy,       customers to date   through supervised   pm); poor to near    with Israel
Cemex (Mexico)                            construction         poor                 Moreno, Director
                                          project cycle                             General

Even though some of the lenders and programs named above are well known and
considered leading examples, the second column of the Table makes clear that even
the cumulative number of microfinance clients served to date is very small—less than
400,000 in total. However, the housing microfinance portfolios of most (but not all)
lenders are now growing fast—in some cases, faster - than the rest of their loan
portfolios. In addition to unsecured home improvement loans, several lenders above
offer low-end mortgage products, although usually on newly constructed houses.
There is little evidence yet of flexible standards, which would enable incremental
improvements to enable a house to become mortgage-able (if sold to a new buyer)


                                                                                                      31
V4.0 Draft



over time. Encouraging such standards could be an important part of building the link
between housing microfinance and conventional mortgages.

In addition, Table 5 above shows that housing microfinance in most cases reaches only
borrowers who can be described as the upper poor (income above 50% of national
poverty lines) and near poor (household income up to 120% or 150% of the national
poverty line). This group is also being targeted by consumer lenders in some
developing countries.

One increasingly popular approach for reaching lower down the income spectrum for
housing finance has been to promote community-based shelter funds and lending
organizations. It is not shown directly above because the lending institutions are often
semi-formal or even informal community based organizations. Many are member-
based effectively with mutual governance structures. UN Habitat’s 2005 report on
Financing Urban Shelter (Chapter 7) highlighted the growth of community-based
shelter funds as a significant trend in housing finance for the poor. While this
approach is relatively new and, as UN Habitat points out, there are as yet few
overviews of its success, there are a number of working examples from the member
organizations of the Shack Dwellers International Network (SDI) in various countries
(such as India and SA among our country sample). Most community funds follow group-
based savings processes to unlock loans for members. Loans are often group-based, as
in traditional microfinance and may enable the building of infrastructure as well as
individual houses. While both the community mobilization around local development
and the discipline of forced savings have been shown to have many benefits,
especially in terms of improved social capital in poor urban communities, the record
of some of these initiatives in terms of credit has often been less favorable.

For example, Ballesteros & Vertido (2004) report on the experience over fifteen years
of the flagship Community Mortgage Program in the Philippines. Under this program,
legally organized associations of up to 300 poor families are provided with mortgage
loans from the state-owned National Home Mortgage Finance Corporation to purchase
land and develop infrastructure and housing on it. By 2003, the program had reached
only 138,871 households; more importantly, the repayment performance on the DCMP
portfolio has been poor, averaging under 80% - far below what is needed for
sustainability. A key problem is how to deal with ‘recalcitrants’—households who
benefit from the project but refuse to become members, with according obligations,
in the community association. The authors find that initial forced savings is correlated
with better performance.

Within the microfinance community more generally, the performance, and therefore
role, of community-managed loan funds (whether for housing or not) is a popular and
controversial issue. Recently, Murray and Rosenberg of CGAP (2006) reviewed the
experience of dozens of community-managed loan funds supported by donors over the
last 15 years, to answer the question, “Which ones work?” They conclude that only
savings-based groups (where there is no external funding for loans) and self help
groups (which start with savings and then leverage bank funding, such as for example


                                                                                     32
V4.0 Draft



the HDFC CBO or NGO clients do) appear viable. In particular, “where loans are
financed by early injection of external funds from donors or governments,
community- managed loan funds (CMLF) projects appear to fail so consistently that
this model of microfinance support is never a prudent gamble” (p.1).

Rapid scaling of community shelter loans funds through external funding is not an easy
answer to extending the reach of housing finance. The important question of how
formal housing microfinance may be scaled from its present limited size is the subject
of another paper at the Symposium.


3.3 Connecting the pieces

As preceding sections have shown, national mortgage and microfinance markets differ
considerably in their reach and depth. Nonetheless, the ‘top-down’ and ‘bottom-up’
views in Sections 2.2 and 3.2 respectively may be combined to produce a generic
income-based profile of the current spread of housing finance instruments, shown in
Figure 7 below.

The Figure shows the conventional reach of mortgage finance in a middle income
country to the top 20-30% of the income range. Housing microfinance then reaches
borrowers in next category--the broad middle of the income distribution, which may
include as many as 50% of households. In this emergent group, unsecured consumer
credit for employed workers and microfinance for self-employed people is growing
fast in many places, creating a credit track record for the first time for some, but also
absorbing repayment capacity and raising risks of over indebtedness for others. For
credit risk management alone, it is therefore important to track the reach and spread
of the different credit instruments at a country level.

Housing microfinance today hardly touches the bottom twenty-five to forty percent of
a population, who are very poor and often dispersed in rural areas which are costly to
service—both with credit and with formal infrastructure. While some specialized
microfinance programs have managed to reach the very poor and destitute, they have
usually done so as part of a structured, subsidized program such as BRAC’s IGVGD.14
The important question of how best to provide housing subsidies to low- and very low-
income households is the subject of another paper at the Symposium.




14
  For more on the limited linkages between microfinance and transfers to the destitute and very poor, see Hashemi
and Rosenberg (2006) CGAP Focus Note No.34: “Graduating the Poorest into Microfinance”


                                                                                                               33
V4.0 Draft




Figure 7: Stylized profile of housing finance instrument usage in developing
countries


      Cumulative % of
      population: 0%

                                                 Mortgage


             20-30%

                             Urban; near
                                poor
             50-75%                                     Housing microloan
    National poverty
    line                  Urban; poor


         60-75%                                             Community-based
                                                            shelter funds
    <50% of National
    poverty line         Rural; very poor


         100%




                                                                               34
V4.0 Draft




4. Conclusions
The housing finance systems of many developing countries, which would less than a
decade ago have been considered missing, fragmented or unstable in Renaud’s
terms15, are today growing fast and becoming more complete in the range of products
on offer. In middle-income countries, it now seems possible, even likely, that
mortgage/GDP ratios of 15% and higher will be common within a decade. This growth
will have vast implications for the financial systems there, and indeed, for the real
economy and society.

Notwithstanding the growth in volumes, this paper has shown how the reach of formal
housing finance is limited in developing countries today: at most, 40% of households in
middle income countries can access mortgages, and formal housing microfinance,
which has the potential to reach the next 30-40% of the income spectrum, is currently
very small in scale in most places, although now growing fast. Considered measures to
push out the mortgage access frontier and expand the scale of home improvement
lending are necessary and important if access to housing finance is to improve
significantly.

This paper has sought to demonstrate the value, and hence the need, for clear
segmentation not of only the housing needs of a population (which is often
undertaken by policy makers) but also of the potential reach of different financing
instruments (which is not). This enables a realistic determination of the limits of
housing finance, and can help focus the targeting (and even blending) of subsidies.

Compiling information on the reach of instruments and the segmentation of the
population on a cross-country basis is no easy task. Indeed, the difficulty and time
taken simply to assemble cross-country data for this one-off analysis supports one
firm conclusion of this paper: there is a need for systematic collection of key housing
finance data. To be really worthwhile, this process has to be ongoing: not only will
the data in this paper soon become dated in rapidly changing markets, but also, the
trends within and across countries are often more important than the comparison of
absolute levels. No information source of this nature or scale of coverage exists
today. Current measures include limited coverage of developing countries: for
example, The Economist’s residential property price index, an important way of
monitoring cross-country price bubbles, currently contains only two developing
countries—China and South Africa. Even though international practitioners have long
recognized the need for such information to inform in-country work, international
financial regulators now recognize this as well as well: “The trend towards
globalization (in housing finance assets), particularly in the investor base, will require
more international information exchange” (CGFS 2006:3).

Two distinct, but related, data sets would be useful in this regard:

15
     See categorization in Renaud (1999)


                                                                                       35
V4.0 Draft




     •   Market level data, containing overall measures pertaining to housing and
         housing finance markets, such as those used in Section 2.I Indeed, this paper
         was intended in part to explore which data provide a useful and consistent
         picture. In EU markets, the European Mortgage Federation is increasingly
         playing this role by publishing useful cross-country data, but there is no
         equivalent for emerging markets.16
     •   Lender or portfolio performance data: there is currently no consistent means of
         collecting data on the size, nature and performance of housing loan books,
         although some industry associations collect and publish data from their
         members. Although rating agencies are the conventional guardians of loan
         portfolio performance data, and are increasingly playing this role in the larger
         emerging economies, there also may be benefit in creating a central repository
         for the collection and aggregation of confidential reports from individual
         housing lenders which focus on the low end of the market at a country,
         regional and global level. In the microcredit sector, The Mix Market, publisher
         of the authoritative MicroBanking Bulletin (see www.themix.org), has played
         this role but has no special categorization or reporting system for low-income
         housing loans.

Better comparable information can inform better policy and even better risk taking by
private housing finance providers. However, better information alone will not stop the
buildup of housing price bubbles, and may not control rampant expectations in some
quarters about the seemingly magical effect of increased housing finance as implied
in the initial quotation by Buckley and Kalarickal. There is today a real risk that the
gathering flood of housing credit turns into a torrent in developing countries, eroding
the very hillside of social and economic stability on which it stands. This would
happen if the increased availability of housing finance translated only or mainly into
higher property prices, enriching the upper and middle classes, but locking out the
global poor from secure or adequate housing. There is no quick financial fix to avoid
this. For one thing, the process of zoning and licensing, which can limit supply of
housing and inflate house prices, has to be addressed; control of this process vests
primarily at a decentralized local level, making it slow to change. But measures which
seek to push outward the boundaries of mortgage markets at the same time as
promoting sustainable growth of non-mortgage solutions will help to connect low
income households to formal housing finance. Unless these connections are made and
strengthened, the increasing flow cannot be diverted and spread to useful effect; and
the torrent becomes unstoppable. In short, success for housing finance today can no
longer be measured by volume, but must be measured by increased access.

In the end, perhaps international housing finance today is a field in the same sense
that microfinance has been recognized as a field for the past ten to fifteen years, and
increasingly, is no longer as it blurs into ‘inclusive financial sectors’. Like
16
  Note that the scale of the task of collection and maintenance of large cross country data sets should not be
underestimated. A previous effort to collect and update city-level data on housing markets, the Global Urban
Observatory, was started by the World Bank in 1990s, and taken over by UN Habitat. However, much of the date is
now old and the site was not even available on-line as at mid-2006


                                                                                                             36
V4.0 Draft



microfinance, international housing finance is worthy of attention for a season at
least so as to create consistent standards and widespread knowledge. These standards
are slowly emerging today, as knowledge and inter-change of ideas increases, in part
through symposia like this one. Then perhaps, international housing finance can be
allowed gracefully to re-submerge into the broader boundaries of retail credit and
financial sector development in general. Housing financiers can then be properly
considered as they should be: neither wizards, who can conjure up rain, nor boosters,
who claim they can, but engineers, who reticulate and plumb new and wider
connections to the housing finance system.




                                                                                  37
V4.0 Draft




References
ACCION (2002) “ACCION Poverty Assessment Framework”, Insight No.1
ACCION (2003) “Poverty Outreach findings: MiBanco, Peru”, Insight No.5
ACCION (2004) “Developing Housing microfinance products in Central America”, Insight No.12
Baumann, T & D. Mitlin (2003) “The South African Homeless people’s Federation—investing in the poor”
    in Small Enterprise Development Journal Vol.14 No.1
Ballasteros, M & D C Vertido (2004) “Can group credit work for housing loans? Some evidence from the
    CMP”, PIDS Policy Notes No.2004-05
Beck, T & A. de la Torre (2006) “The Basic Analytics of Access to Financial services”, Working paper
    March 2006
Berson, D et al (2004) “America’s Home Forecast: the Next Decade for Housing and Mortgage Finance”,
    available via
    http://www.homeownershipalliance.com/studies/pdfs_2004/americas_home_forecast.pdf
Buchholz, T (2002) “Safe at Home: The new role of Housing in the US economy”, paper prepared for
    the Homeownership Alliance, available via
    http://www.homeownershipalliance.com/studies/pdfs_2002/february_safeathome.pdf
Bucks, B., A. Kennekal and Moore (2006) “Recent Changes in US Family Finances: Evidences from 2001
    and 2004 Survey of Consumer Finances”, Federal Reserve Bulletin , available via
    http://www.federalreserve.gov/Pubs/oss/oss2/2004/bull0206.pdf
Buckley, R.M. (2003) “Housing’s Role in Wealth and the Economy”, presentation delivered at the
    Fannie Mae International Housing Finance Symposium Sept 17, Session 1
Buckley, R.M & J. Kalarickal (2005) “Housing Policy in Developing Countries: Conjectures and
    Refutations” World Bank Research Observor
Cities Alliance (2003) “Shelter Finance for the Poor Series Synthesis”, Issue 4v, April 2003, available via
    http://www.citiesalliance.org/citiesalliancehomepage.nsf/6FF87B591473087C0625687E0058053D/
    D5B0C8044038F75986256BCF0063757F?OpenDocument
Committee on the Global Financial System (2006) “Housing Finance in the Global Financial Market”,
    CGFS Papers No 26, available via www.bis.org/cgfs
Collins, D (2006) “Housing and the finances of the poor”, special paper prepared for this conference,
    available via www.financialdiaries.org
Daphnis, F (2005) “Housing Microfinance: Current Issues, Opportunities & Challenges”, presentation at
    IFC/World Bank Housing Finance Seminar March 2005
Devlin, James (2005) “A Detailed Study of Fianncial Exclusion in the UK”, Journal of Consumer Policy
    28:75-108
De Soto, H (2001) The Mystery of Capital, New York: Basic Books
Economist, The (2003) “Whose is the fairest of them all?”, 16 October 2003
ECI Africa (2003) “ Analysis on the Application of Borrowed Funds”, report prepared for the
    MicroFinance Regulatory Council, South Africa, available via www.mfrc.org.za
Fay, M & A. Wellenstein (2005) “Keeping a Roof Over One’s head” Ch 3 in Fay, M (Ed) The Urban Poor
    In Latin America, World Bank
Fay, M & C. Riggeri Laderchi (2005) “Relying on Oneself: Assets of the Poor”, Ch 6 in Fay, M (Ed) The
    Urban Poor In Latin America, World Bank
Field, E & M. Torero (2006) “Do Property Titles increase credit access among the Urban Poor? Evidence
    from a Nationwide Titling Program”, mimeo available from www.tinyurl.xx
Ferguson, B (2004) “An update on Housing Microfinance and its importance to home lenders”,
    Presentation at IUHF Conference, Brussels 23 June 2004
Gwinner, B. (2005) “Microfinanciamiento papa Vivienda: El caso de MiBanco Peru”, presentation
    available via
    http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTFINANCIALSECTOR/EXTHF/0,,contentMD
    K:20803632~menuPK:2178998~pagePK:210058~piPK:210062~theSitePK:396699,00.html
Helms, B (2006) Access for All: Building Inclusive Financial Systems, Washington: IBRD
Honahan, P (2004) Financial Sector Policy and the Poor, World Bank Working Paper No.42




                                                                                                        38
V4.0 Draft



Kuyasa Fund (2005) “Delft Area Housing Needs Analysis”, August 2005, research supported by FinMark
    Trust, available from www.finmarktrust.org.za/research
Lea, M et al (2005) “More than Shelter: Housing as an Instrument of Economic and Social Development:,
    Housing Finance International May 2005
Low, S., Dübel, A. and Sebag-Montefiore, M. (2003). “Study on the Financial
    Integration of European Mortgage Markets”. Study commissioned by the
    European Mortgage Federation to Mercer Oliver Wyman. London/Brussels, available via
    www.hypo.org.
Matul, M, Elena Durmanova & Volodymyr Tounitsky (2006) “Market for Microinsurance in Ukraine: Low-
    Income Households Needs and Market Development Projections”, Mimeo Submitted to
    Microinsurance Centre, commissioned by KfW & Microinsurance Centre
Melzer, I (2006) “The access frontier for housing finance in South Africa: how low can you go?”, report
    prepared for Finmark Trust, available via
    http://www.finmarktrust.org.za/documents/2006/July/AccessFrontierTool.pdf
Mitlin, D (2003) “Finance for Shelter: recent history, future perspectives” in Small Enterprise
    Development Journal Vol.14 No.1
Mortgage Bankers AssociatIon (2005) 2005 MBA Cost Study, available via
    http://www.mbaa.org/ResearchandForecasts/ProductsandSurveys/TheCostStudyIncomeandCostfor
    OriginationandServicingofOne-to-Four-UnitResidentialLoans.htm
Murray, J. & R. Rosenberg (2006) “Community managed loan funds: which ones work?” CGAP Focus
    Note No.36
NAHB (2002) “Housing: the Key to Economic recovery”, accessed via www.nahb.com
OECD (2004) “Housing Marikets, Wealth and the Business Cycle”, Ch.IV in OECD Economic Outlook
Pearson, R & M, Greef (2006) “Causes of Default among Low income Housing Loan Clients”, Report
    commissioned by FinMark Trust, NHFC, RHLF and DBSA, available via www.finmarktrust.org.za
Porteous, D J (2005a) “The Access Frontier as a Tool in Making Markets Work for the Poor”, Paper
    commissioned by DFID, available via www.bankablefrontier.com/publications
Porteous, D J (2005b) “Housing Microfinance: Fashion or Footwear?”, Presentation at 3rd African
    Microfinance Conference, available via http://www.mfrc.co.za/conference/speakers/Porteous.ppt
Prahalad, CK (2004) The Fortune at the Bottom of the Pyramid, Philadelphia: Wharton Press
Renaud, B (1987) “Another Look at Housing Finance in developing Countries”, Cities Vol 4.1, February
Renaud, B (1999) “The Financing of Social Housing in Integrating Financial Markets: A View from
    Developing Countries”, Urban Studies, Vol. 36, No. 4, April 1999, pp. 755-773. Special Issue: Social
    Housing Finance in the European Union.
Renaud, B (2004a) “Housing Finance in a Global Context: Mortgage Market Structure and Housing Price
    Stability” Proceedings of the Seoul International Seminar on Real Estate 2004, Korea Housing
    Association, Seoul, Korea
Renaud, B (2004b) “Mortgage Finance in Emerging Markets: Constraints and Feasible Development
    Paths”, Mimeo v2, November 2004
Rhyne, B (2005) “Meeting Low-Income Housing Needs through Housing Microfinance”, Presentation at
    3rd African Microfinance Conference, available via http://www.mfrc.co.za/conference/speakers
RHLF, NHFC & MFRC (2005) “Housing Microlending: a financial performance analysis”, unpublished
    report
Rutherford, S (2000) The Poor and their Money, Delhi: OUP
Rutherford, S (2002) Money Talks: Conversations with Poor Households in Bangladesh about Managing
    Money, IDPM Working Paper No.45
Rutherford, S (undated) “Uses and users of MFI loans in Bangladesh”, MicroSave Briefing Notes on
    Grameen II No.7, available via www.microsave.org
Rust, K (2003) “Sink or Swim: progress made in the relationship between banks and alternative lenders
    in South Africa’s low income housing finance sector”, Occasional Paper no 9, HFRP< available via:
    http://housingstudies.wits.ac.za/Sink%20or%20Swim.pdf
Segel, A & N. Meghji (2005) “Patrimonio Hoy: A Groundbreaking corporate program to alleviate
    mexico’s housing crisis”, mimeo Harvard Business School
Smith, D “Housing the World’s Poor: The Four Essential Roles of Government”, in Harvard International
    Review 2006



                                                                                                     39
V4.0 Draft



Stickney, C (2006) “Habitat for Humanity LAC”, note prepared for CGAP Housing Microfinance group
Struyk, R (2005) “Home Purchase Affordability and Mortgage Finance“ Ch.5 in Hegedus, J & R. Struyk
    (2005) Housing Finance: New & Old Models in central Europe, Russia and Kazakhstan, available via
    www.mri.hu.
UN Habitat (2005a) Financing Urban Shelter: Global report on Human Settlements, available via
    http://www.unchs.org/pmss/getPage.asp?page=bookView&book=1918
UN Habitat (2005b) “Homeownership through Mortgage Finance”, available via www.unhabitat.org


General Data Sources
Euromonitor International World Consumer Lifestyles Databook, 4th Edition 2005
Jones Lang LaSalle (2006) Real estate transparency index, available via
    http://www.joneslanglasalle.com/en-GB/research/researchabstract?artid=2489
World Bank Group Doing Business, available via http://www.doingbusiness.org/
World Bank, World Development Indicators, accessed via website by country for various years

Countries sources
Armenia
Source: published material, with input and comments from Mher Yergany (Bank Akademie) and Smbat
Nasibyan (Conversebank).

Rabenhorst, C, Struyk et al (2005) Development of a Sustainable Market for Housing Finance:
   Feasibility Study”, prepared for KfW, 2005

Ghana
Source: Jim Hokans (consultant) working with Frank XX in country.
Hokans, J et al (2004) “Strategic Assessment of the Affrodable Housing Sector in Ghana”, report from
    the in-country assessment team, CHF International, available via
    http://www.chfhq.org/content/general/detail/1428/

Honduras
Source: Noah Sawyer (consultant), with advice and input from Mariana Balestrini, BCEO

India
Source: Dev Goel (HDFC)

Deutsche Bank Research (2006) “Building up India: outlook for India’s real estate markets”, available
   via www.dbresearch.com
IUHF (2006) Country fact sheet downloaded from
   http://www.housingfinance.org/Content/ContentIndex.php?Interest=Factsheet&plus=%%
NHB (2004) “Report on Trend and Condition of Housing in India” , available via
   http://www.nhb.org.in/Publications/default.htm

Mexico
Source: Alan Elizondo (SHF)

Babatz, G (2005) “Housing Microfinance: Mexico’s Experience and Attempt to Scale the Program”,
    presentation at the African Microfinance conference, Cape Town 2005
BBVA (2006) Situacion Inmobiliaria, Servicio de estudios Economicos, Agosto, available via
    http://serviciodeestudios.bbva.com/TLBB/tlbb/jsp/sve/america/mexico/pubsect/sitinmob/index.
    jsp#0
JCHS (2004) The State of Mexico’s Housing 2004, available via
    http://www.jchs.harvard.edu/publications/international/som2004.pdf
Klaehn, J, B. Helms & R Deshpande (2005) “Mexico: Country level savings assessment”, available via
    http://microfinancegateway.org/resource_centers/savings/article/27187/



                                                                                                        40
V4.0 Draft




Pakistan
Source: Megmood Begum (HBFC)

Serbia
Source: published materials, with advice and input from Ljubica Gelev, NKOSK; and Mirjana Zakanji
    (ProCredit, Serbia).

Elliot, T, A. Jovic & J. Pupovac (2006) “A Rising Star: The National Mortgage Insurance Corporation of
     Serbia (NKOSK)”, Housing Finance International , forthcoming
Hegedus, J (2002) “Housing Finance in South Eastern Europe”, Metropolitan Research Institute,
     Budapest, available via www.ceemortgagefinance.org
Hegedus, J & R. Struyk (2005) Housing Finance: New & Old Models in central Europe, Russia and
     Kazakhstan, available via www.mri.hu.
Milicevic, G “Mortgage Lending System in Serbia”, Mimeo 2006
Roy, F “Mortgage Lending and Risk management in Serbia”, presentation delivered in Bucharest, 5 April
     2006
Radulovic, A (2006) “Housing Finance and regional Integration: Former Yugoslavia case: could it work”,
     Mimeo July

South Africa
Source: Kecia Rust (FinMark Trust)

Aron, J, J Muellbauer & J Prinsloo (2006) “Estimating household-sector wealth in South Africa”, SARB
    Quarterly Bulletin June 2006
IUHF (2006) Country fact sheet downloaded from
    http://www.housingfinance.org/Content/ContentIndex.php?Interest=Factsheet&plus=%%
Luus, C (2003) “The ABSA residential property market database for South Africa: key data trends and
    implications”, BIS Papers No/21, available via
    http://www.imf.org/external/pubs/ft/reif/2005/eng/index.htm
Settlement Dynamics et al (2005) Research into Housing Supply and Functioning Markets: Final Report,
    Commissioned by Banking Council of SA, available via www.banking.org.za
Shisaka Development Services (2006) Small Scale Landlords, report commissioned by FinMark Trust et
    al, available via
    http://www.finmarktrust.org.za/themes_and_projects/theme_detail/theme_detail.asp?uno=7
Settlement Dynamics & Matthew Nel & Associates (2005) Research into Housing Supply and Functioning
    Markets: Final Report, Prepared for Banking Association of SA, available via
    http://www.banking.org.za/documents/2005/DECEMBER/HousingSupFinal.pdf

Zambia
Bank of Zambia Website: www.boz.com
FinScope Zambia (2006) see www.finscope.co.za




                                                                                                    41
V4.0 Draft




Annex A: MOW mortgage market completeness measure
See Background description in Box E

                                                                                                          Changes from original
                                                            Criteria used                     Weight        MOW Comments
 A      LTV, Borrower & Purpose                                                                   35%   Kept the same
 A1.    Maximum LTV                                                                               10%
        <80%                                                                             0
        80%-89%                                                                          1
        90%-99%                                                                          2
        >100%                                                                            3
 A2     Borrower/ purpose                                                                         25%
        0- no availability; 0.5--limited availability; 1- readily available                             Same
                                                            Young household <30
                                                            Older household >50
                                                            Low equity
        Access criteria                                     Self certified income
                                                            Previously bankrupt
                                                            Credit impaired
                                                            Self employed
                                                            Government sponsored
                                                            Second mortgage
                                                            Overseas holiday home
                                                            Rental
                                                            House equity release
                                                            Shared ownership


 B.     Product                                                                                   50%
        0- no availability; 0.5--limited availability; 1- readily available
                                                            Variable (any)
                                                            Variable (Referenced)
        Rate structure:                                     Fixed for life of loan                      Changed from ‘discounted’
                                                            Capped
                                                            0-5 years
                                                            5-10 years
        Range of fixed term                                 10-20 years
                                                            20+ years
                                                            Amortizing
                                                            Interest only
        Repayment structures:                               Flexible
                                                            Fee free redemption (during fixed period)
                                                                                                        Left out ‘Yield
                                                                                                        maintenance fee on fixed’
                                                                                                        and reweighted category
                                                                                                        accordingly



 C.     Distribution channels                                                                     10%
        1- at least 5% of mortgages distributed through it



                                                                                                                                    42
V4.0 Draft




                                                       In branch
                                                       Tied advisor
        Channels                                       Independent advisor
                                                       Direct



                                                       >80%                                    1
        Concentration: % through primary channel       55%-80%                                 2
                                                       <55%                                    3


 D.     Information quality                                                                5%
        Quantity and quality of information to the consumer
        Score of 1-5, assigned based on answers
        to questions below:
                                                       Are there laws governing                    MOW score in this
                                                       disclosure of information on                category was largely
                                                       mortgage contracts?                         subjective;
                                                                                                   I used these further
                                                       Are mortgage contracts and                  questions to determine
                                                       terms standardized?                         score
                                                       Is borrower education
                                                       available?
                                                       Is borrower education required
                                                       for first time home buyers?


        TOTAL                                                                           100%




Comments:
   1. The MOW framework is a useful one, but does require:
         a. Some common explanation of certain terms used
         b. Some standardization about the difference between ‘readily available’ and ‘somewhat
              available’ mean: certain scores were reduced following further questioning.
   2. Missing product categories relevant to developing countries which were not included:
         a. Mortgages linked to foreign remittances
         b. Mortgages offered in local and a foreign currency option




                                                                                                                            43
        V4.0 Draft



        Table 1: MOW Features per country
                                                                                                                               South
                                                           Armenia         Ghana     Honduras   India     Mexico    Pakistan   Africa    Serbia

A. LTV, Borrower & Purpose
What is the maximum LTV available on mortgages?
<80%                                                          Yes
80%-89%                                                                                          Yes                  Yes                 Yes
90%-99%                                                                                Yes                 Yes
>100%                                                                       Yes                                                 Yes
Are mortgages available to the following (on what basis per score below)                                  General
Young household <30                                          Limited       Limited    Readily   Readily   Readily    Limited   Readily   Limited
Older household >50                                          Limited       Limited    Limited   Readily   Readily     None     Readily   Limited
Low equity                                                    None         Limited     None     Readily   Readily    Limited   Readily    None
Self certified income                                        Limited       Limited    Readily   Limited   Readily    Readily   Limited   Limited
Previously bankrupt                                           None         Limited    Limited   Limited    None       None     Limited    None
Credit impaired                                               None         None       Limited   Limited   Limited     None     Limited    None
Self employed                                                Limited       Limited    Readily   Limited   Readily    Readily   Limited   Limited
Government sponsored                                          None         Limited    Readily   Readily   Readily    Readily   Limited   Limited
Second mortgage                                               None         Limited    Limited   Readily   Readily    Limited   Readily   Limited
Overseas holiday home                                         None         Limited    Readily    None     Limited    Limited    None      None
Rental                                                       Limited       Limited    Readily   None      Limited     None     Readily   Limited
House equity release                                          None         Limited    Limited   Limited    None       None     Readily   Limited
Shared ownership                                              None         Limited     None     Readily   Limited    Readily    None      None


B. Mortgage Product Features
Variable (any)                                               Limited       Limited    Readily   Readily   Limited              Readily   Readily
Variable (Referenced)                                         None         Readily    Readily   Readily   Limited              Readily    None
Fixed                                                        Limited       None       Limited   Readily   Readily               None      None
Variable but capped for life of loan                          None         None        None     Readily   Readily    Readily   Readily    None
0-5 years                                                    Limited       None       Readily   Readily   Readily               None     Readily
5-10 years                                                   Limited       Readily    Readily   Readily   Readily    Readily   Limited   Readily
10-20 years                                                   None         Readily    Readily   Readily   Readily              Readily   Readily
20+ years                                                     None         None       Readily    None     Readily              Limited   Limited
Amortizing                                                   Limited       Readily    Readily   Readily   Readily              Readily   Readily
Interest only                                                 None         None        None     Readily   Limited              Limited    None




                                                                                                                                         44
     V4.0 Draft



                                                                                                                                       South
                                                                        Armenia    Ghana     Honduras   India     Mexico    Pakistan   Africa    Serbia
Flexible                                                                 None      Readily    Limited   Readily   Readily    Readily   Readily   Limited
Fee free redemption (during fixed period)                                Limited   Limited     None     Readily   Readily    Readily    None      None


C. Distribution channels
Is as least 5% of new mortgages originated through the channel below:
In branch                                                                 yes       yes        yes       yes        yes       yes       yes       Yes
Tied advisor                                                                        yes        yes       yes        yes                 yes
Independent advisor                                                                 yes                  yes        yes                 yes
Direct                                                                    yes       yes        yes       yes                                      Yes


What is the % of new mortgages originated by the primary channel:
>80%                                                                      yes       yes        yes                  yes       yes                 Yes
55%-80%                                                                                                  yes                            yes

<55%


D. Information quality
Are there laws governing disclosure of information on mortgage
                                                                           no       yes        yes       yes        yes       yes       yes        No
contracts?
Are mortgage contracts and terms standardized?                             no        no        Yes        no       yes LI     yes       yes      Mostly
Is borrower education available?                                           no       yes      Somewhat    yes        no        yes       yes       Yes
Is borrower education required for first time home buyers?                 no        no         No        no        yes       yes        no        No


     Source: Country correspondents, see Annex for sources




                                                                                                                                                 45

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:0
posted:3/30/2013
language:English
pages:45
dominic.cecilia dominic.cecilia http://
About