The Italian housing market in the crisis:Fresh evidence

OECD STESEG Meeting Paris 10-11 September 2009

Francesco Zollino

1. Importance of investing in information on the housing market

2. The statistical outlook in Italy

3. The Bank of Italy-Tecnoborsa new survey:

4. The new BoI price indicator

5. More on price developments in Italy

Outline

Developments in residential property prices are crucial in economic analysis:

1) changes in house prices affect the business cycle through their impact on (a) households’ wealth and in turn on consumption behaviour

(b) residential investments;

2) sharp price fluctuations impact on financial stability (credit quality, value of collateral...)

3) the functioning of the housing market affects labour mobility

Empirical analysis shows that housing market might contribute to the persistent propagation of shocks that hit the economy

The recent crisis confirms the importance of investing in reliable, complete and (possibly) harmonized statistics on the housing market.

Information available in Italy (as well as in the euro area) is lacking.

This presentation focuses on a recent initiative undertaken in Italy to enhance a timely monitoring of the housing market: a) new survey on market operators;

b) new price indicator

1. The need for a better monitoring the housing market

2. The statistical outlook in ItalyIn Italy information on the housing market is:

• released by many (private and public) institutions/research centres;• not systematic, not timely, not available for a few important dimensions

a) Price statisticsOfficial (NSI) data on property transaction prices are not available yet.

Alternative sources (all pro and cons):

Nomisma: no breakdown in the series but limited geographical coverage

Il Consulente immobiliare (CI): long time series (1965) and broad territorial reach, but frequent changes in the reference sample (breakdown).

Osservatorio del mercato immobiliare (OMI). Broad territorial coverage, but lack of historical depth (since 2002) and data published with a delay of a few months.

All in all (Bank of Italy statistics)

CI is the source that is best suited to analyzing medium-term developments. It is used in BI (Zollino, Muzzicato and Sabbatini, 2008) to compute an index since mid-1960s

OMI is important for the wealth of information provided and territorial coverage. It is used in BI (Cannari and Faiella, 2007) for estimating households’ wealth.

(b) Non-price information

a) Statistics on building permits are released, but only at an annual frequency and with a delay of 1.5 years

b) Information on transactions is available (OMI), but with 3-4 months delay. Furthermore, data are semi-annual (quarterly data only on an experimental basis)

c) Information regarding how long it takes to sell a dwelling, the percentage of discount on the initial price and so on is not collected systematically.

(c) Qualitative (high-frequency) surveys

ISAE monthly surveys. (1) Question on the households’ intention to buy a dwelling (poor leading properties of the answers); (2) business climate of building firms

New initiative to fill the gap for the residential market

Early 2009 Bank of Italy and Tecnoborsa launched a new quarterly survey to collect (timely) information on the residential housing market.

2. The statistical outlook in Italy /2

a) Quarterly survey (since January 2009) that gathers opinions of real-estate agencies on the residential market through a questionnaire (short, simple, structured).

b) Survey outsourced to an external company (Questlab Srl) c) The data are collected in the month following the end of the calendar quarter (Jan, Apr, July, Oct); the reference population consists of real-estate agencies

d) The sampling design is stratified, with a total of 34 strata representative of 4 national macro-areas (North-West; North-East; Centre; South and Islands).

3. New BoI-Tecnoborsa survey: general characteristics

Collect qualitative data useful to fill the informative gap for a set of variables, such as prices, transactions and time for selling, on which quantitative statistics are either available with a considerable lag or missing at all

Main characteristics

Purpose

3. New BoI-Tecnoborsa survey: The methodology

• Initial contact by email; questionnaire filled in through the web (most used) or sent by fax • A leading Confederation of real-estate agents (FIAIP) provided the main list of agencies to extract the sample; strong cooperation to “contact” the agents (information, etc.)

a) Collection of the data

b) Sample design (34 strata) and reference population

• 15 Italian towns with population of 250,000 or more• 15 areas around the towns at letter (a) forming the hinterland• 4 national macro-areas (North-West, North-East, centre, South and Island) excluding the 30 strata above.

Note: each stratum contains a minimum # of units; then, the sample size is large enough to ensure that s.e. are acceptable; the basic # of units per stratum is proportional to the number of transactions recorded in 2006.

The reference population consistes of real-estate agencies (source: Istat)

Each real-estate agency in the sample is assigned a weight giern by the number of firms in the stratum cell to number of firms in the sample

c) Contacts/response rate

Around 3,500 agencies are reached; target 1,500; actual participants around 1,000 (still relatively low response rate)

(a) # of months to sell a dwelling;

(b) unexecuted mandate to sell and new mandate;

(c) main reasons not to renew a mandate;

(e) percentage of sales financed by mortgages and their incidence on transaction prices

New BoI-Tecnoborsa survey: The questionnaire

Section 1. The outlook for transactionsA1. Considering the total number of homes sold by you in the reference quarter, how many

months passed on average between a house being registered with you and its sale (exchange of contracts)? (months)

A2. Compared with the previous quarter (October–December 2008) was the average property selling time (number of months indicated above) in the reference quarter…?

(Much shorter, same, longer, much longer)

A3. Considering the number homes on your books still unsold at the end of the reference quarter (January–March 2009), compared with the end of the previous quarter was that number…?(much smaller, smaller, same, larger, much larger)

A4. Considering the number of homes newly registered with you in the reference quarter, compared with the previous quarter was that number…?(much smaller, smaller, same, larger, much larger)

A5. Regarding non-renewals of selling agreements in the reference quarter, please select the main reasons from the list on the right? (no more than three)

Seller thought the offers received were too low

Owners decided to wait expecting a rise in p No offers were made because buyers thought

the price was too high

Too long since the property was put on the market

Seller thought fees were too high

Buyer had difficulty getting a mortgage

Seller encountered unexpected problems (seizure of goods, separation, etc.)

Other reasons (give details)

A. Considering all the houses sold by you in the reference quarter, how many do you know were bought with a mortgage and for what proportion of p?

|_______| % of homes bought with a mortgage |_______| % of price covered by mortgage|_| Don’t know

Table 1

In the first quarter of 2009 evidence of a slump in the housing market

(a) Price developments in the reference quarter;

(b) Transaction price vis-à-vis offer price

B1. (For agents that sold properties in the reference quarter) For the main type of property sold in the reference quarter, was the average selling price (per sq.m.)?

B1. (For agents that DID NOT sell properties in the reference quarter) Within your territory, in the reference quarter are average selling prices (per sq. m.)?

(much lower, lower, fairly stable, higher, much higher)

B1_1. Can you quantify the change in selling prices as a percentage? (Please also indicate the sign of any change)

B2. Considering the main type of property sold by you in the reference quarter, compared with the seller’s first asking price was the selling price…?

Lower by: more than 30% |_| 20-30% |_| 10-20%|_| 5-10% |_| less than 5% |_| Same (or higher) |_|

Section 2. Prices

y/y vs. q/q percentage changes? Intervals or number? Only quantitative, qualitative, both? Actual price vs. perceptions (answer tend to fall in classes 0, 5, 10, ...)

Table 2

Background issues: how to ask for prices in quantitative terms?

3. New BoI-Tecnoborsa survey: The questionnaire /2

(a) Prices (next quarter);

(b) Number of mandates to sell (next quarter);

(c) General outlook of the residential market in the local market (next quarter) and the national market (next quarter and next 2 years)

C1. As far as concerns the type of property generally sold by you, regardless of recent trends how do you think prices in April–June 2009 will differ from the reference quarter?

(Sharp decrease Decrease Fairly stable Increase Sharp increase)

C1_1. Can you quantify this change in selling prices as a percentage? (Please indicate sign)

C2. In April–June 2009, how do you expect the number of properties newly registered with you to compare with the reference quarter?

(Much smaller Smaller About the same Larger Much larger)

C3. Considering the housing market only in your area, how will the performance in the current quarter (April–June 2009) compare with the previous quarter?

(Worse Same Better)

C4. How do you think the general situation in the housing market throughout the country will develop compared with the present?

In this quarter: worse same betterOver the next two years: worse same better

Section 3. Short-term perspectives

Table 3

Open issues

Difficulty in increasing the response rate (real estate agencies not used to answer questionnaire).

The process of interviewing agencies takes too long (just less than 1 month)

How to collect quantitative information? (prices and transaction)

To be assessed: informative content of the answers, considering that the survey was launched in “exceptional times”

Future developments

Ad hoc section to investigate specific issues

“Confidence indicator” (a few more observations are needed)

3. New BoI-Tecnoborsa survey: Open issues and future developments

4. The new BoI indicator of house prices

The new indicator combines three components that differ by

type of dwelling, geographical coverage and time horizon:

1) prices of new houses in the provincial capitals (from 1966 S1)

2) prices of new and recently-built houses in the provincial capitals (from 1985 S1)

3) prices of new houses in 1400 other municipalities (from 1998 S1)

Each component is calculated at provincial, regional, macro-regional and national

level (also divided in urban and non urban areas)

As in overlapping periods, the three categories of prices differ more in terms of

levels than in dynamics (with trend discrepancies temporarily widening in 2003-05)

i) for years since 1998 we calculate a composite indicator based on prices

recorded in all municipalities (or some 60% of total national stock in dwellings)

ii) for earlier periods, data are based on price changes in provincial capitals, for new

and recently-built dwellings back to 1985 and for the sole new ones back to 1966

The new BoI indicator of house prices /2

Prior to crisis, developments of new indicator are much in line with those published by other research centres (controlling for geographical cover and type of dwelling).

1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

40

60

80

100

120

140

160

180

200

40

60

80

100

120

140

160

180

200

Nomisma

New indicator

new house prices in 13 regional capitals

-4,00

-3,00

-2,00

-1,00

0,00

1,00

2,00

3,00

4,00

2003 2004 2005 2006

Total Capital cities Other municipalities Centre and North South and Islands

Difference in changes (percentage points)

between indicators based on Consulente Immobiliare and OMI

- Based on BoI indicator of house prices, analysis of medium term developments points at four cycles in the Italian housing markets:

- The first cycle, presumably already under way when data collection began in 1965, closed with the peak reached at the end of 1974, when prices surged by more than 30 per cent after five years of broad stability (Figure)

- The second cycle, from the end of 1974 to mid-1981, was characterized by a phase of volatility that was more accentuated around the second oil shock, interrupted by an abrupt rise in prices, that reached a new peak in 1981(Figure) - The third cycle, lasting up to the second half of 1992, began with a gradual downward correction, with prices down in 1986 to the low of the previous cycle, followed by more than 8 per cent per year increase in real terms until the peak in 1992 (Figure)

- In the fourth cycle, likely just closed by the end of 2008, after a seven year fall at a milder pace than in the past, house prices show a brisk upswing (5% per year in real terms) until firstsigns of a slowdown since the end of 2006 (Figure)

5. More on house prices developments in Italy

- Recent developments point to a gradual fading of house inflation in Italy, by a diverging intensity across different sources as transaction volume shrank in latest periods

- According to OMI, house prices in real terms stabilized in the first semester of 2008, and show a slight decline in the second (Figure)

- According to BoI indicator, barely in line with OMI in previous years, house prices were still on the raise in the first semester, with a virtual stagnation in the second, signalling that uncertainty usually surrounding property price increased as the volume recession deepened. In the first semester 2009 prices still keep increasing at a moderate pace.

- Latest figures point to number of transactions keeping plummeting in the first quarter of 2009, totalling an overall drop by 36% since the middle of 2006. (Figure)

5. More on house prices developments in Italy /2

- Based on historical evidence that volume cycles are more timely reactive than prices to changes in fundamentals on housing market, property prices should soften further in the next future.

- However, the contracting volume of sales is likely resulting in a larger margin of error in measuring prices, as it is suggested by the widening gap between different price indicators.

- The qualitative assessment of real-estate agents, according to the Bank of Italy-Tecnoborsa survey, is that the fall in prices had probably begun at the end of the last year (Figure)

- That points at a possible reconciliation between BoI and Territory Agency (OMI) price indicators, as the latter heavily hinges on real estate assessments to validate official records whereas they are poorly reliable on a local basis due to unsatisfactory number of transactions. BoI indicator, as based on raw data from il Consulente Immobiliare, reports only on prices affective paid on the market.

5. More on house prices developments in Italy /3

Back

offers received too

low for seller

expecting prices to rise

no offers made

owing to high price

property on market

for too long

difficulty obtaining mortgage

unexpected problems for seller

other

Total 53.4 24.5 65.9 18.9 50.3 2.6 6.8 Memorandum item:

Q4 2008 52.3 18.4 64.7 16.8 47.0 3.9 7.1

Average time between entry on market and

actual sale

Change with respect to previous quarter (per cent)

(months) shorter same longer Total 6.6 22.0 42.2 35.9

Memorandum item: Q4 2008 6.7 10.8 38.6 50.6

Main reasons for cancelling contract with agent

Property selling times

Percentage of purchases financed with mortgage

Ratio of mortgage to house price

Total 70.4 71.0 Memorandum item: Q4 2008 69.1 68.6

House purchases and mortgages(per cent; reference quarter: January–March 2009)

Table 1(per cent of real-estate agencies; reference quarter: January–March 2009)

Lack of offers due to “too-high asking prices” main reason for contract cancellation, followed by “offers dismissed as too low by sellers” and “buyer’s difficulty obtaining a mortgage”

Selling times stable at around 7 months

Approximately 70% of purchases financed with mortgages (stable)

Ratio of mortgage to house price 71%

Comparison with previous period

decreasing stable increasing Average

(quantitative By geographical area

North-West 58.6 40.6 0.8 North-East 69.3 28.9 1.8 Centre 58.6 40.2 1.2 South and Islands 59.1 37.1 3.8

By resident population Urban areas (over 250,000 inhabitants) 70.4 28.5 1.1 Non-urban areas (under 250,000 inhabitants) 54.4 43.8 1.7

Metropolitan areas (over 500,000 inhabitants)

69.8 28.9 1.3

Non-metropolitan areas (under 500,000 inhabitants)

57.8 40.7 1.6

Total 61.4 37.1 1.5 Memorandum item: Q4 2008 57.6 39.6 2.8

Property selling prices

Back

Table 2

Difference between selling price and seller’s first asking price (1)

same or higher

less than 5%

lower

5% - 10% lower

10% - 20% lower

20% - 30% lower

more than 30%

lower

average reduction

Total 3.7 8.9 36.1 40.6 9.6 1.1 11.8 Memorandum item: Q4 2008 6.1 14.2 44.6 30.5 4.0 0.6 9.5

Evidence of more frequent reductions in prices (balance 59.9 compared to -54.8)...

...also with respect to the seller’s asking price (average reduction increase) Prices 11.8% below the seller’s asking price (compared with 9.5%)Sign that prices might gradually decrease in the next months.

(per cent of real-estate agencies; reference quarter: January–March 2009)

Back

Table 3

Conditions on real-estate agent's market in current quarter (April-

June 2009)

Expected number of newly registered properties in current

quarter (April-June 2009) compared with reference quarter

Expected level of prices in current quarter (April-June

2009) compared with reference quarter (2)

poor normal good lower same higher lower same higher

Total (5) 27.7 51.5 20.8 16.3 55.6 28.1 47.6 49.7 2.7 Memorandum item

Q4 2008 76.6 18.1 5.3 19.8 45.3 35.0 65.0 0.7 34.3

Outlook for the housing market

General situation of the housing market in Italy

Outlook for current quarter compared with reference quarter

Outlook for next two years compared with reference quarter

worse same better worse same better

Total 32.4 56.3 11.4 12.4 22.0 65.6

Memorandum item: Q4 2008 60.8 34.6 4.6 35.0 19.8 45.2

Estate agents are becoming lest pessimistic about the short-term outlook for the housing market, both locally and nationally.

Their medium-term expectations for the market at national level show some improvement on the previous survey.

(per cent of real-estate agencies; reference quarter: January–March 2009)

House prices in Italy

(index 2000=100; constant values based on HICP)

'70 '71 '72 '73 '74 '75 '76 '77 '78 '79 '80 '81 '82 '83 '84 '85 '86 '87 '88 '89 '90 '91 '92 '93 '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08

0

20

40

60

80

100

120

140

160

180

At current values

At constant values

Source: elaborations based on data from Bank of Italy, Istat and Il Consulente immobiliare

back

House prices and transactions in Italy(semi-annual data; percentage changes at constant prices)

Source: elaborations based on data from Bank of Italy and Territorial Agency;

(*) For transactions, 2009-S1 estimate based on 2009-Q1.

back

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009-20

-16

-12

-8

-4

0

4

8

12

-20

-16

-12

- 8

- 4

0

4

8

12

Territorial Agency price indexnumber of salesBank of Italy price index

Italy - Real estate agency assessments of price developments

-100

-80

-60

-40

-20

0

20

40

60

80

100

bal

ance

s b

etw

een

"in

crea

se"

an

d "

dec

lin

e" p

erce

nta

ges

changes in Q4 2008 on previous quarter expected changes in Q1 2009 on previous period

Italy North-West North- East Center South and Islands

Source: Bank of Italy and Tecnoborsa Surveyback