THE AVAILABILITY OF HOUSING
STATISTICS IN IRELAND –
GAPS AND FUTURE NEEDS
Ronan Lyons, Trinity College Dublin
CSO Housing Statistics Seminar
Dublin Castle, 18 October 2016
Lessons from history: the forgotten bubble of 1988
Quarterly change in nominal
house prices
Source: Own work-in-progress
1990I
1989IV
1989III
1989II
1989I
1988IV
1988III
1988II
1988I
1987IV
1987III
1987II
30%
25%
20%
15%
10%
5%
0%
-5%
-10%
Structure
• A History of Housing Statistics
• The New RPPI
• Further Improvements
• Beyond Prices
A brief history of housing prices…
• Valuations (based primarily on rental) started in the mid-19C
• Continued for over a century… (And still going for commercial property)
• Dept of Environment simple averages, from 1969 on
• And quarterly from early 1970s
• The only measure of housing prices until the mid-1990s
• Also other housing statistics, e.g. loan by institution and borrower type
What’s wrong with simple averages?
• The type of properties that trade in
Annual change in Dublin house
prices, by source
busts are structurally different from
those that trade in booms
• Clear insight from DOE series in
2010: implies strong price gains at a
time when prices were falling
-10%
• Above-average income households
-20%
10%
0%
Hedonic (Daft.ie)
Q1 2013
Q1 2012
Q1 2011
Mean (DOE)
Q1 2010
• Location attributes – typically a
Q1 2009
• Property attributes, e.g. type, size, age
collection of markets (e.g. counties)
Q1 2008
-30%
Q1 2006
• Need to control for…
20%
Q1 2007
were less affected
30%
Hedonic (CSO)
A brief history of housing prices…
• Valuations (based primarily on rental) started in the mid-19C
• Continued for over a century… (And still going for commercial property)
• Dept of Environment simple averages, from 1969 on
• And quarterly from early 1970s
• The only measure of housing prices until the mid-1990s
• Also other housing statistics, e.g. loan by institution and borrower type
• PTSB-ESRI: first mix-adjusted index (hedonic price regression)
• Based on one major lender, mortgage-backed transaction prices
• Quarterly from 1996; ended in 2010 – reliance on one lender, problematic
when transaction/mortgage volumes fell 90%+
• Other measures pre-2006 included estate agent valuations/estimates
Convergence in housing price statistics
• 2006 – daft.ie: based on population of online listings
• Including those not on daft.ie; analogous rental price index
• Hedonic price regression, system of ~400 micro-markets; tweaked upon
introduction of CSO v1
• Currently being extended back to 1922 (also rents)
• 2011 – CSO v1: based on population of mortgages (at key banks)
• Hedonic price regression, county-level ‘fixed effects’
• In line with best practice in other countries
• 2012 – myhome.ie: based on online listings
• Hedonic price regression
• Since 2015, method similar to daft.ie/CSO
Structure
• A History of Housing Statistics
• The New RPPI
• Further Improvements
• Beyond Prices
The new CSO index
• A step-change on CSO v1:
• Includes cash purchases; population of all transactions
• Includes property size, age, energy rating (from SEAI BER register)
• Includes location at Census Small Area
• Allows the publication of average transaction prices
Year-on-year change in RPPI: Dublin
30%
CSO v1
CSO v2
Daft.ie
20%
20%
10%
10%
0%
0%
-10%
-10%
-20%
-20%
-30%
-30%
2006 q1
2006 q3
2007 q1
2007 q3
2008 q1
2008 q3
2009 q1
2009 q3
2010 q1
2010 q3
2011 q1
2011 q3
2012 q1
2012 q3
2013 q1
2013 q3
2014 q1
2014 q3
2015 q1
2015 q3
2016 q1
2016 q3
2006 q1
2006 q3
2007 q1
2007 q3
2008 q1
2008 q3
2009 q1
2009 q3
2010 q1
2010 q3
2011 q1
2011 q3
2012 q1
2012 q3
2013 q1
2013 q3
2014 q1
2014 q3
2015 q1
2015 q3
2016 q1
2016 q3
Old and new indices tell similar stories to each other – and to the
Daft.ie index, with differences likely due to weighting
Year-on-year change in RPPI: Ex-Dublin
30%
CSO v1
CSO v2
Daft.ie
Structure
• A History of Housing Statistics
• The New RPPI
• Further Improvements
• Beyond Prices
Suggestions for CSO v3!
• Residential property prices depend on…
• Size: is m2 as recorded in BER accurate? (What about m3?!)
• Age: vintage vs. quality (vs. depreciation)
• Location: schools, transport facilities, green space
• Much of this already bundled into location ‘fixed effect’
• However, single biggest missing factor currently is site size
• One of the first characteristics included in academic research on property
values – garden space, but also option value (mews, corner house)
• OSI PRIME2 database, PRAI Land Registry, EU INSPIRE directive…
Structure
• A History of Housing Statistics
• The New RPPI
• Further Improvements
• Beyond Prices
What do policymakers need to manage the housing system well?
• Register of residency: who lives where?
• Could be linked to land/property tax credits
• Register of beneficial ownership: who owns what?
• Cadastre is standard elsewhere (and was in 1850s Ireland!)
• Register of transactions
• RPPR – but with significantly improved stamp duty form; also bids
• RTB – renters are people too!
• Register of construction projects
• Application – Approval – Commencement – Completion – then occupancy
• Improvements / Change of use / Etc
Thank you!
• Looking forward to comments and questions…