Retrofit Explorer

Neighbourhood emissions trends are often read as a measure of local progress, but they need care. Much of the fall in domestic emissions across Britain since 2010 comes from the decarbonisation of electricity generation, which happens equally in every neighbourhood and owes nothing to local action. The gas series is the more informative one for judging retrofit progress, because it responds to insulation, heating controls, and the replacement of boilers with heat pumps.

As for the postcode chart, but using the whole-neighbourhood totals DESNZ publishes rather than individual postcodes, combined for any larger area shown here: metered consumption multiplied by annual national emissions factors. Source: the sub-national gas and electricity statistics from the Department for Energy Security and Net Zero. Because they come from meter readings these are among the most reliable figures in Carbon & Place, subject to three caveats that apply to every energy chart: totals are not weather-corrected, areas with few meters are suppressed, and unmetered fuels such as oil and solid fuel are absent altogether. See the manual.

Bills make the case for retrofit in terms households recognise, and they connect decarbonisation to the cost of living. Neighbourhoods with high bills are, other things being equal, where efficiency investment does most to reduce fuel poverty, and bill estimates can help target schemes such as the Warm Home Discount, the Energy Company Obligation, and local crisis support. One important caution: a low bill is not necessarily a good sign. Households that cannot afford to heat their homes properly show up here as low consumers, so this chart should be read alongside income, and the Energy bills vs Income map does exactly that.

Bills are estimated by applying average annual unit prices and standing charges (based on the Ofgem price cap and historical tariff data) to metered consumption. Actual bills vary with tariff, payment method, and supplier; the estimate assumes a standard tariff.

A sustained fall in gas consumption is the clearest local signal that fabric improvements and heat electrification are happening. A flat trend suggests little is changing, though it is not proof: gas use also responds to weather, to energy prices, and to how warm people can afford to keep their homes, and a fall driven by rationing heat in a cold home is not a success. Reading this chart alongside the energy bills and income measures helps distinguish the two.

Figures are totals and averages for this area across mains gas meters only. Source: the sub-national gas and electricity statistics from the Department for Energy Security and Net Zero. Because they come from meter readings these are among the most reliable figures in Carbon & Place, subject to three caveats that apply to every energy chart: totals are not weather-corrected, areas with few meters are suppressed, and unmetered fuels such as oil and solid fuel are absent altogether. See the manual.

National electricity consumption per home has fallen for years as lighting and appliances became more efficient, so a neighbourhood where consumption is rising against that trend is unusual and usually indicates early adoption of heat pumps, electric vehicles, or both. These clusters are of practical interest to distribution network operators planning reinforcement, and to councils deciding where charging and heat pump support will find willing households.

Figures are totals and averages for this area; Economy 7 consumption is shown separately where present. Source: the sub-national gas and electricity statistics from the Department for Energy Security and Net Zero. Because they come from meter readings these are among the most reliable figures in Carbon & Place, subject to three caveats that apply to every energy chart: totals are not weather-corrected, areas with few meters are suppressed, and unmetered fuels such as oil and solid fuel are absent altogether. See the manual.

The balance between gas and electricity meters sets the retrofit starting point for an area. Neighbourhoods with few gas meters are off the gas grid, and because the fuels they use instead are both expensive and carbon intensive, they are usually the places where a heat pump improves bills and emissions at the same time. Heavily gas-heated areas face a longer transition that depends on national decisions about the future of the gas network as much as on local programmes.

Meter counts from the DESNZ sub-national statistics, aggregated to the area shown here. Meters do not map one-to-one to households.

EPC bands underpin much of housing policy: minimum standards for rented homes (currently E, with proposals to raise this to C; see the minimum energy efficiency standard), lender targets, and grant eligibility. Concentrations of D to G rated homes mark priority areas for area-based retrofit schemes.

The band comes straight from the certificate, and anything that is not A to G is counted as unknown. Where a home holds several certificates only the most recent is kept, so this is a snapshot rather than a history. Be careful what you read into the band: it scores modelled energy cost per square metre under standard occupancy, not carbon and not measured consumption, so a well-insulated home on an expensive fuel can rank below a leakier home burning gas. The average shown alongside uses the underlying 0 to 100 efficiency score, which is more sensitive than the band to small differences. The manual explains what EPC coverage does and does not include.

Built form shapes what a retrofit programme can realistically offer. Flats usually require agreement from a freeholder or management company and often whole-building works, which is why they are frequently left out of grant schemes. Terraces share party walls and scaffolding, so they suit street-by-street delivery. Detached homes have the most external wall area per occupant to insulate, and also the most room for a heat pump and its pipework. Knowing the local mix is the difference between a programme that can be delivered and one that cannot.

Two certificate fields are pasted together to make each label here: built form (detached, semi-detached, mid-terrace, end-terrace, or enclosed) and property type (house, bungalow, flat, maisonette, or park home). The analysis pipeline then groups the result into eleven categories plus an other bucket, folding the enclosed terraces in with ordinary terraces and treating a bare “House” with no built form as detached. For a count covering every home rather than only certificated ones, use the Dwelling Stock tab.

Tenure determines who is able to act, and through which policy lever. Owner-occupiers decide for themselves and respond mainly to advice, grants, and finance. Private landlords respond chiefly to regulation, in particular the minimum energy efficiency standard for rented homes. Social landlords can plan works across a whole portfolio and are the easiest route to retrofit at scale. The split-incentive problem, where a landlord pays for improvements from which the tenant gains, is the reason the private rented sector so often needs regulation rather than encouragement.

Four categories: owner-occupied, private rented, social rented, and unknown. Tenure is stated at assessment and never updated, so a home sold out of the rented sector keeps its old label until it is next assessed. Two placeholders are mapped to unknown, NO DATA! and the long “not defined — use in the case of a new dwelling” string, and “rental” is normalised to “rented”. Rented homes are over-represented throughout the register, because letting triggers a certificate and continued ownership does not.

Construction age is the best single predictor of how a home is built, and therefore of what it needs. Homes built before about 1919 usually have solid walls, which need internal or external insulation at several times the cost of filling a cavity. Homes built between the wars and the 1980s often have cavities that were never filled, which remain among the cheapest measures available. Homes built since the 1990s were built to progressively tighter Building Regulations and generally need little fabric work. An age profile is therefore the quickest way to size and cost a local retrofit plan.

Thirteen bands plus unknown. The certificate's own age band is used where it exists, with the “England and Wales:” prefix stripped and invalid records set to unknown; where a certificate gives a bare year instead, it is placed into the matching band. Scotland uses different bands entirely and these are mapped onto the England and Wales set, which introduces small errors: a Scottish home recorded as 1992–1998 appears here as 1991–1995, and 1965–1975 as 1967–1975. Treat Scottish age bands as approximate to within a few years.

Fabric-first retrofit, insulating walls, roofs, floors, and windows before changing the heating system, cuts bills and emissions while improving comfort and health. Walls are the largest heat loss surface in most homes, so this data helps councils target area-based schemes (such as ECO flexible eligibility and social housing retrofit programmes) at the streets where poor wall performance is concentrated.

Six categories: the assessor's five-point scale from Very Good to Very Poor, plus other for certificates where the field is missing. The rating is a judgement rather than a measurement — thermal performance is inferred from the construction type and whatever insulation is visible, without opening up the wall — so a cavity filled without a surviving record, or insulation installed badly, will not show. The walls description chart gives the underlying construction.

Wall construction dictates which measure is possible and what it costs. An unfilled cavity can usually be insulated in a day for a few hundred pounds. A solid wall needs insulation fixed to the inside or the outside, typically costing an order of magnitude more, and in the case of internal insulation requiring the room to be cleared and redecorated. The local split between cavity and solid walls therefore drives the cost of any neighbourhood programme more than any other single variable. Solid wall insulation also carries real risks of damp and mould if specified or installed badly, which is why government-funded schemes require a whole-house assessment under the PAS 2035 retrofit standard before work begins; CREDS reviews the technologies available for retrofitting the UK housing stock.

Wall descriptions are free text, cleaned into a controlled vocabulary of about seventy values by the EPC processing code, then grouped by the analysis pipeline into the seven categories plotted here. Two of them, sandstone or limestone and granite or whinstone, are largely Scottish constructions that have no equivalent in the English data. Two things are worth knowing about the source. Scottish certificates describe several elements in one field separated by a pipe character, affecting roughly a quarter of Scottish rows, and only the first is kept, so a mixed-construction home is categorised by its principal wall. And about 9,000 Scottish records carry a roof description in the wall field, an error in the source data, which is set to unknown rather than allowed to pollute the wall categories.

Topping up loft insulation is among the cheapest ways to save energy in an existing home, and the materials generally cost far less than the heat they retain over their life. Uptake is already high, so the value of this chart is in finding the homes where the cheap measure is still missing, which schemes such as the Energy Company Obligation exist to reach. Rooms in the roof and flat roofs are a different matter: both cost considerably more to treat and belong in a whole-house plan rather than a single-measure scheme.

Seven categories: the five-point scale from Very Good to Very Poor, plus dwelling above and other. The dwelling above category is the one to understand first, because in a neighbourhood of flats it will dominate the chart. A home with another dwelling above it has no roof of its own to insulate, so no rating is issued; where the rating is missing and the roof description says another dwelling is above, the analysis pipeline labels it dwelling above rather than leaving it as missing data. A large share in that category is therefore a statement about built form, not about insulation, and only the top-floor homes in a block carry a real rating.

Roof type determines how straightforward insulation will be. A conventional pitched roof with an accessible loft is the simplest and cheapest case. A room in the roof, a flat roof, or a dwelling with another dwelling above it all require different and more expensive approaches. The mix of roof types in a neighbourhood indicates how much of the stock can take the inexpensive measure and how much needs a fuller retrofit.

The cleaned vocabulary holds around fifty-five roof descriptions, but the chart shows six categories: pitched, flat, roof rooms, thatched, dwelling above, and other. Nearly all the detail is lost in the first of those — twenty-seven cleaned values describe a pitched roof at loft insulation depths from a few millimetres to 400 mm, and all of them are counted simply as pitched. If you need the insulation depth, the individual buildings on the Domestic Properties layer retain the full description. Homes with another dwelling above have no roof of their own and are counted as dwelling above; the seven different spellings the two registers use for this are consolidated first. Rooms in the roof are counted separately from ordinary pitched roofs, because they are far more expensive to insulate; thatched roofs take precedence where a certificate mentions both.

Floors are usually the last fabric element worth addressing, because they account for a smaller share of heat loss than walls or roofs and are more disruptive to reach. Suspended timber floors can be insulated at reasonable cost where there is access from beneath; solid floors are rarely worth treating except during a major renovation, when the floor is coming up anyway. Floor data therefore refines an individual whole-house assessment rather than driving an area-based scheme.

Seven categories: the five-point scale from Very Good to Very Poor, plus dwelling below and other. As with roofs, a home with another dwelling beneath it has no exposed floor and receives no rating; where the rating is missing and the floor description says another dwelling is below, it is labelled dwelling below rather than counted as missing. Expect that category to dominate in blocks of flats. Floors are also the fabric element assessors are least able to inspect, being frequently inferred from the age and type of building rather than observed, so the ratings that do exist are the least reliable of the fabric fields.

The distinction between suspended and solid floors decides what can be done. A suspended timber floor can be insulated from the void beneath, or by lifting the boards, at moderate cost and with a useful gain in comfort from reduced draughts. A solid floor can only be insulated by digging out the slab or by building up on top of it, which raises floor levels and affects doors and skirtings, so it is normally done only as part of wider works. This is refinement for a whole-house plan rather than the basis of an area-based scheme.

Nine categories are plotted. Six describe the construction: solid and suspended floors, each split three ways by whether the insulation is absent, limited, or present. A seventh covers floors exposed to external air or to an unheated space. The eighth is dwelling below, for homes with another dwelling beneath them and so no exposed floor at all; the several spellings the two registers use for that case are consolidated into one before counting. The ninth is a residual for anything the rest do not match. Together they account for every certificate in the area.

Full window replacement is an expensive way to save energy, so policy attention focuses on the remaining single glazed homes, where the gains are largest and where conservation area rules may require secondary glazing or slim-profile units instead (Historic England advice). Draught proofing around existing windows and doors delivers much of the comfort benefit at a fraction of the cost.

Six categories: the five-point scale plus other where the field is missing. The rating combines glazing type with the condition of the frames as judged by the assessor, for the dwelling as a whole, so a home with modern glazing at the front and original single glazing at the back is averaged into a single figure.

Most of Britain is already double glazed, so the remaining single glazing tends to be concentrated in particular places rather than spread evenly. Those pockets are often older housing in conservation areas or listed buildings, where replacement windows would need consent and may be refused, and where secondary glazing or slim-profile units are the realistic options (Historic England advice). A cluster on this chart is therefore as likely to indicate a heritage constraint as a simple gap in provision.

Five categories: single, double, triple, secondary, and not defined. One detail is worth knowing, because it is deliberately discarded. Certificates often record when the glazing was installed — before 2002, during or after 2002, or unknown date — and the analysis pipeline strips that qualifier before counting, so all double glazing is counted together regardless of age. Since 2002 marks a significant tightening of glazing standards, two neighbourhoods with identical bars here can have meaningfully different window performance.

Efficiency ratings show where the heating system itself is wasting energy, which is usually a matter of old non-condensing boilers, individual room heaters, or ageing electric storage heating. These are the homes where replacing the system cuts bills most. Note the limit of the measure, though: it rates efficiency rather than carbon, so a well-rated gas boiler still burns a fossil fuel and a poorly rated system on a clean fuel may already be low carbon.

Six categories: the five-point scale plus other. This rates how efficiently the system converts fuel into heat and says nothing about the carbon content of that fuel, which is why a modern gas boiler rates well while burning a fossil fuel, and why an old electric system can rate poorly despite running on an increasingly clean grid. For decarbonisation the fuel and system-description charts are the more useful pair.

The mix of heating systems in an area determines the route its decarbonisation is likely to take. Areas dominated by gas boilers face conversion to heat pumps or connection to a heat network, which is the hardest and slowest of the routes. Areas heated by electric storage heaters are already off gas and can move to modern electric heating or heat pumps without touching the gas network. Buildings on communal systems are natural candidates for heat network zoning under the Heat and Buildings Strategy. The count of heat pumps is the most direct local measure of how far the transition has actually got, though it lags reality because it depends on when each home was last assessed.

This is the messiest field on the certificate. The EPC processing code recognises around 220 distinct single systems, and because England and Wales join multiple systems with commas while a single system description already contains commas, roughly 1,500 compound values exist with no reliable delimiter between their parts. Those are accepted by decomposition into known systems rather than enumerated. The analysis pipeline then reduces all of it to the eight categories here, by matching explicit lists for gas and oil boilers and by keyword for the rest, with exclusions to stop a home being counted twice: room heaters, for instance, excludes anything also mentioning a boiler, storage heating, a heat pump, or a community scheme. Anything unmatched falls into other, which in some areas is a large share.

Main heating fuel divides the stock into groups that different policies reach. Homes on mains gas are the largest group and the hardest to shift, awaiting heat pumps or heat networks. Homes on heating oil, bottled gas, or solid fuel have the highest emissions per unit of heat and the least protection from price rises, since these fuels are bought in bulk at unregulated prices rather than under a capped tariff; they are consequently a priority for Boiler Upgrade Scheme grants and for advice. Electrically heated homes need efficiency measures and help with tariffs rather than a change of fuel, and are over-represented among households in fuel poverty.

Seven fuels are plotted: mains gas, electricity, oil, coal, LPG, biomass, and dual fuel. The raw certificates contain thousands of spellings, which the EPC processing code reduces to about forty by stripping prefixes such as “gas:” and “electricity:” and folding anthracite into coal. The analysis pipeline then removes the community, non-community, and unknown qualifiers, so a home on a district heating scheme is counted under whatever fuel that scheme burns. An eighth other slice catches the fuels outside these seven — biogas, bioethanol, rapeseed oil, waste combustion, and homes recorded as having no heating system at all — so the chart accounts for every certificate rather than quietly dropping the unusual cases.

Heating controls, meaning programmers, room thermostats, thermostatic radiator valves, and zoning, are among the cheapest energy savings available, and unlike insulation they can be fitted in an afternoon without disruption. They also matter more than their cost suggests when converting to a heat pump, which works best running steadily at a lower flow temperature and therefore needs controls suited to that pattern rather than to a boiler. Concentrations of poorly controlled homes are straightforward targets for advice services and small grants.

Six categories: the five-point scale plus other. This rates what was installed on the day of assessment, not whether anyone uses it, which matters more here than for most fields: a programmable thermostat left permanently on the manual setting rates the same as one used well. Controls are also among the cheapest things to change, so this rating goes out of date faster than the fabric ratings do.

This chart shows which combinations of controls are actually installed, and therefore what is missing. A home with a boiler but no room thermostat, or with no thermostatic radiator valves, can usually have them fitted for a modest sum and will save energy immediately. Because the measure is cheap, tenure-neutral, and needs no planning permission, it is one of the few retrofit interventions that can be delivered quickly at scale.

The cleaned vocabulary holds about forty-five combinations of programmer, room thermostat, thermostatic radiator valves, and zoning. The chart shows only the three commonest exactly as recorded — programmer with room thermostat and TRVs, programmer with room thermostat, and programmer with TRVs and bypass — together with time and temperature zone control and an other bucket that absorbs the remaining forty or so combinations. Scottish certificates prefix these descriptions with a group number, which is stripped before matching. Because the three named combinations are matched as exact strings, an unusual variant falls into other even when it describes much the same controls.

Hot water is a modest share of household energy but a universal one, and unlike space heating it does not fall away in summer. Insulating a cylinder and fitting modern controls are inexpensive improvements. Hot water demand also matters for the transition to heat pumps, because a heat pump generally needs a cylinder, and finding space for one in a home that has a combination boiler is a common practical obstacle.

Six categories: the five-point scale plus other. In most homes hot water comes from the main heating system, so this rating tends to track the heating rating closely and adds little on its own. It is more informative where the two diverge, which usually means a separate immersion heater, or a cylinder that is unusually well or badly insulated.

How hot water is provided affects the retrofit route. A home that already has a cylinder can add solar thermal or use the cylinder as flexible storage on a time-of-use tariff, and is straightforward to convert to a heat pump. A home with a combination boiler has no cylinder, and finding space for one is often the deciding practical constraint on whether a heat pump can be installed at all.

Six categories are plotted: five named systems — hot water from the main heating system, an electric immersion heater, a community scheme, an electric instantaneous heater, and a gas multipoint heater — plus a residual for everything else. These are matched as keywords against a cleaned vocabulary of around 110 values, which is large because a single description encodes three things at once: the heat source, whether there is a cylinder thermostat, and whether solar contributes, giving entries such as “community scheme, no cylinder thermostat, plus solar”. Because the vocabulary is so much larger than the five named categories, the residual is a sizeable share in some areas.

Low-energy lighting is the simplest efficiency measure there is, and it is now close to universal in recently assessed homes. That makes this chart more useful as a check on the age of the local EPC sample than as a retrofit priority: a neighbourhood with many poorly rated lighting scores is usually one whose certificates are old, rather than one where people are still using halogen bulbs.

Six categories: the five-point scale plus other. The rating reflects the share of fixed light fittings holding low-energy lamps at the time of assessment. Because LED lighting is now close to universal, this chart has largely stopped measuring lighting and started measuring the age of the local certificates: a neighbourhood with many poor scores is usually one that has not been assessed for a decade, rather than one still running halogens. It is the clearest illustration in this tool of how quickly EPC data ages.

Rooftop solar reduces both bills and demand on the grid, and its uptake is strikingly uneven between neighbourhoods, reflecting roof orientation, tenure, and the capital available to households as much as the resource itself. Councils have several levers: bulk-buying schemes that reduce installation costs, requirements on new development through local plan policy, and installation across their own housing stock. Read the percentages with the coverage caveat in mind, since panels fitted after a home's last certificate do not appear. The Solar Potential layer in this tool shows how much sunlight actually reaches each roof.

A simple yes or no. The certificate records photovoltaic supply as a percentage of roof area, but that is reduced during cleaning to presence or absence: missing values and zero become no, anything else becomes yes. The chart therefore shows how many homes have panels, not how much capacity. Two things push the figure down. Panels fitted after a home's most recent certificate do not appear at all, and installation often follows a purchase, which is exactly when the certificate was issued. Real uptake is higher than shown, probably by a good margin.

Solar water heating is a mature technology that has become a niche one, largely overtaken by photovoltaic panels combined with an immersion diverter, which is cheaper per unit of energy and more flexible in what the electricity can be used for. Existing systems remain worth maintaining for the hot water they provide, but they are seldom the first recommendation for a new installation.

A yes or no flag from the certificate, converted to a logical value during cleaning, with missing values counted as no. The same information also appears inside the hot water description as “plus solar”, and the assessor records the two independently, so they do not always agree. As with photovoltaics, systems installed after a home's last assessment are invisible here.

Band is used to target and means-test support: several energy efficiency and cost-of-living schemes have been restricted to lower bands as a rough proxy for household means. Bands also set council tax income, so the local mix shapes what an authority can fund. Note that the bands rest on 1991 values in England and Scotland, and 2003 in Wales, so they have drifted a long way from current prices.

England and Wales come from the Valuation Office Agency council tax stock statistics (CTSOP), published per LSOA. Scotland comes from the equivalent statistics.gov.scot dwellings-by-band series, published per 2011 Data Zone and apportioned here onto 2022 Data Zones in proportion to postcode counts, so Scottish figures for split or merged zones are estimates rather than counts. Band I exists only in Wales and is empty elsewhere.

Built form decides both what retrofit is physically possible and who has to agree to it. Flats generally need the freeholder's consent and often works to the whole block, an obstacle that has kept them out of most grant schemes despite housing a large share of lower-income households. Terraces suit street-by-street delivery, where survey and scaffolding costs are shared. Detached homes have the most external wall area to insulate and the most room for a heat pump and cylinder. Because this chart counts every dwelling, it gives a fairer picture of the local mix than the equivalent chart drawn from certificates.

From the Valuation Office Agency CTSOP3 table, per LSOA, England and Wales only, 2020 onwards. Type is recorded by the valuer, so conversions and unusual buildings can be classified inconsistently.

Dwelling size drives energy demand and the cost of retrofit more or less directly: a larger home has more wall and roof to insulate and more space to heat. The local size mix also bears on whether an area's housing matches the households who need to live in it, a question that runs through most local plan debates. Note that bedroom counts are a rough proxy for floor area, and that a room recorded as a bedroom by a valuer is not necessarily used as one.

Derived from the same VOA CTSOP3 table by summing bedroom counts across all dwelling types. England and Wales only, 2020 onwards. The count is of rooms the valuer recorded as bedrooms, which need not match how a household actually uses them.

Age sorts neighbourhoods into retrofit packages. Areas of solid-wall housing need the expensive interventions that grant schemes have historically found hardest to deliver at scale, and where those homes are listed or in a conservation area, external changes are constrained as well as costly. Areas built between the wars and the 1980s have cavity walls, and where those cavities are still unfilled they represent the cheapest remaining measures in the stock. This chart covers every dwelling rather than only those with a certificate, so it is the more reliable basis for sizing a programme.

From the Valuation Office Agency CTSOP4 table, per LSOA, England and Wales only, 2020 onwards. The single years from 2009 to 2021 are combined into one band. A substantial number of dwellings have an unknown build period, shown as its own category rather than distributed across the others.

House prices set the financial context for retrofit. In high-value areas, households are more often able to borrow against the property or fund works from savings. In low-value areas the cost of a deep retrofit can exceed the value it adds to the home, which is the clearest economic argument for public subsidy: without it, the work will not happen however sound the case for it. Lenders have begun to reflect energy efficiency in mortgage products, though the evidence on how much of an efficiency premium is capitalised into prices remains mixed.

Prices come from HM Land Registry Price Paid Data (England and Wales), matched to neighbourhoods via UPRN address geocoding. Only actual transactions are recorded, not valuations, so thin markets produce volatile averages, and Scotland is not covered by this dataset.

Transactions matter for retrofit because a sale is the moment when a home is surveyed, finance is arranged, and building work is most likely to be contemplated. Several policy proposals, including minimum efficiency standards at the point of sale and green mortgage products, are designed around that moment. Transaction volumes therefore indicate how often each neighbourhood presents that opportunity: in areas where homes change hands rarely, a sale-triggered policy would take decades to reach the stock.

Transaction counts come from HM Land Registry Price Paid Data, which covers England and Wales only and records residential sales that were sold for value and lodged for registration. A new-build flag distinguishes newly built properties from established ones. Repossessions, however, cannot be identified: they sit in the data's category B alongside buy-to-lets, transfers to non-private individuals, and properties classed as “Other”, and HM Land Registry states that the category does not separately identify those types. Several kinds of ownership change are absent altogether, including transfers not made for value, such as gifts and transfers between family members, leases at a premium with a nominal rent, and deeds transferring more than one property at once. The chart therefore counts sales, and undercounts changes of ownership.