1Rental housing in France is characterised by low supply and low mobility, resulting in high rents and greater pressure on social housing in areas with the most dynamic economic growth. Although the government has intervened on a massive scale to address this problem, this intervention is sporadic and, at times, inconsistent.
2While housing is an essential good in which there is a case for intervention when the market produces exclusion, this intervention must be carefully calibrated; otherwise, such intervention can be counterproductive. To date, there has been little evaluation of French housing policy given its high cost (40 billion euros in 2010, of which 16 billion was spent on housing support alone) due, in particular, to difficulties in accessing data.
3French rental housing policy has three main tools: the regulation of relations between landlords and tenants, social housing and the housing support system. The objective of the reforms proposed in this Note is to optimise and align each of these three tools in order to increase mobility, reduce inequalities between regions and strengthen social cohesion.
4In the private rental sector, the authors propose the implementation of flexi-security in housing in the form of greater flexibility in leases, together with a more robust enforceable right to housing. In the first instance, disputes will be heard by joint authorities of representatives of land lords and tenants. They also recommend that the government internalise the cost of non-payments, by acting as a lender of last resort to tenants in order to prevent an increase in non-payments due to the pooling of risk. With regard to rent controls, the authors suggest that a trial be conducted across a number of regions before rolling out this measure on a national level.
5To reduce geographical disparities in social housing, the authors recommend opening up the construction and management of new housing to competition and allocating support according to simple pressure indicators. To promote social diversity, any subsidies provided would be based on spatial segregation indicators. Housing allocation rules would be clarified with the introduction of a points system within each main objective of social housing. Moreover, such an approach would promote mobility by activating rent surcharges calculated according to period of occupancy and income.
6Finally, the Note proposes bringing the housing support system in line with the rest of our redistribution system by making this support part of the income tax system, with rent under a certain threshold becoming tax-deductible. For non-taxpayers, housing support would become a negative tax while for taxpayers it would result in a tax cut. This reform would reduce the difference between the various types of housing, improving the situation of private sector tenants (who have suffered most from rising rents) and reduce the chances of housing support being diverted to owner-landlords.
7In 2011 there were 33.8 million dwellings in France; of these, 28.2 million were main residences. Of these 28.2 million main residences, 16.4 million were owner-occupied, with the remaining 11.8 million occupied by tenants. In this Note, we focus on these 11.8 million rental dwellings, which account for 40% of main residences in France. Of these dwellings, 56% are in the private sector and 44% are social housing. 
8Rental housing in France is characterised by low supply and low mobility, resulting in high rents and greater pressure on social housing in areas with the most dynamic economic growth. Although the government has intervened on a massive scale to address this problem, this intervention has been sporadic and, at times, inconsistent.
9Housing is not a good like any other: it is an absolute necessity. While housing is an essential good in which there is a case for intervention when the market produces exclusion, this intervention must be carefully calibrated; otherwise, such intervention can be counterproductive.
10Housing is also a factor for economic efficiency: a fluid rental market promotes the mobility of labour. A “full home ownership” policy, with its associated onerous transfer costs, on the other hand, does not foster mobility. The non-portable nature of social housing entitlements also hinders mobility.
11Unequal access to housing has resulted in mechanisms, such as housing support, which have failed in part due to their inflationary impact on rents. Conventional redistribution policies would be more effective, encouraging a greater supply of housing and using social housing in the most efficient manner possible.
12It is unfortunate that there has been little evaluation of the cost of French housing policy (40 billion euros in 2010, of which 16 billion was spent on housing support). Data in this area are extremely hard to obtain, limiting the opportunities to acquire expertise and offer a second opinion.
Objectives and instruments of government intervention
13Housing policy has two clear objectives: redistribution; employment and growth.
Objectives and instruments of a redistributive housing policy
14First and foremost, government intervention in housing can be justified on redistributive grounds: the authorities must ensure that low-income households and households in a precarious financial situation have access to decent housing. This fairness objective is reinforced by the effects of housing on children’s health and education.
- A transfer not for those on low incomes;
- Targeted support. For example, landlords receive coupons directly, with the tenant required only to pay the difference between the rent and the value of the coupon. Alternatively, the government reimburses the tenant for part of the rent paid upon presentation of a receipt;
- Social housing provision. With this instrument, the government replaces private landlords to prevent the latter from appropriating financial support.
16Government intervention in redistribution faces two fundamental problems:
- In general, targeted support is inflationary and can be diverted by private suppliers (construction companies and landlords).  However, social housing can be difficult to manage, inefficient, and poorly located, and can reduce mobility and drive out private housing if the supply of housing is rigid. It can also give rise to social tensions as a result of a lack of fairness or transparency.  Chart 1 shows that while housing support is highly redistributive (it is focused on the poorest households), the same is not true of social housing, which is focused on the middle class and has a somewhat diluted effect: in the top five deciles, there are as many households in private rental accommodation as in social housing, which would appear contradictory; 64% of French households are eligible for PLUS  social housing and, once in said housing, can earn above the eligibility threshold and remain there on payment of a reasonable “rent surcharge”.
- While a good redistributive policy focuses government resources on the lowest deciles of the population, it must not result in a concentration of poverty in particular geographical areas, but rather help promote social mix. Therefore, it is crucial to determine whether the objective of social diversity must be pursued within the social housing sector or within a particular district, by having social housing and private housing exist side by side. Paradoxically, France combines social housing that serves groups within the upper quartiles with a high level of segregation in agglomerations. The opposite situation –where the most disadvantaged could live in public housing, and where this social housing is evenly distributed– would be preferable.
Objectives and instruments of a housing policy geared towards employment and growth
17The second objective of housing policy is to promote employment and growth, by improving the mobility of labour and ensuring more reasonable housing costs.
18The instruments used to achieve this objective go beyond redistribution alone: they also include regulation of the private rental sector (in particular the relationship between landlords and tenants) and the tax system.
Housing, mobility, and employment
19In France, there are areas with almost full employment (in 10% of towns and cities, the rate of unemployment for people aged 25-49 is below 5.7%) and others with high unemployment (in 10% of town and cities, the rate of unemployment for the same group is over 17.1%).  For young people the situation is more acute, with unemployment for 15-24 year olds at over 38% in 10% of towns and cities, and below 13% in 10% of towns and cities. While geographical mobility should reduce some of these differences, this is not happening for a number of reasons related in particular to insufficient fluidity in rental markets and high transfer costs upon the purchase of property.
Financial benefits conferred by housing allowances (APL) and low-income housing (in euros)
Financial benefits conferred by housing allowances (APL) and low-income housing (in euros)
Tenant households in urban areas according to type of housing (in thousands of households)
Tenant households in urban areas according to type of housing (in thousands of households)
20There is a dual connection between housing and unemployment. On the one hand, people on fixed-term contracts or who are seeking employement (and who tend to be young people) find it difficult to secure housing due to the strong reluctance of landlords to run the risk of non-payment; on the other hand, the rigidity of the housing market limits employment prospects by limiting geographic mobility. And yet, the housing occupancy status is not totally separate from mobility, which is twice as high in the private rental sector as in the social housing sector, where mobility tends to fall. 
21Thus, policies to encourage home ownership and the development of social housing could have a negative impact on mobility and employment.  Clearly, the objective is to determine whether the status of the occupier determines immobility, or whether mobility determines a household’s particular status (with the most mobile choosing to rent in the private sector). While there is disagreement among experts as to the nature of the causal relationship and the precise mechanisms involved, all other factors being equal people who own their home outright and social housing tenants tend to remain unemployed for longer than those who are paying off a mortgage (since they have to repay a loan) and private sector tenants.  In short, the factors that hinder mobility differ between homeowners and social housing tenants: the former face significant transaction costs and tax disincentives,  while the latter cannot transfer their rights from one property to another and are not given priority in the allocation of social housing in their destination. 
Housing and competitiveness
22Housing costs, which account for a large proportion of the budgets of households buying their own home and private sector tenants, indirectly penalise businesses “obliged” to increase the remuneration paid to their employees to keep up with increases in housing costs. The link between housing costs and competitiveness is highlighted in two recent studies. 
The private rental sector
23A lease agreement resembles a financial agreement similar to a debt contracted by the tenant: the landlord lends their dwelling with immediate effect (thus foregoing use of the same with immediate effect) against a promise of a future financial return. As is the case with the financial agreement, there is an element of uncertainty attached to the payment of rent: the risk of non-payment is the equivalent of the risk of default, except for the fact that the landlord is in principle assured of recovering his/her asset, even if it has been damaged. Managing the risk of non-payment of rent is a central issue for the operation of the private rental market. In this regard, France is considered to provide tenants with strong protection. The Access to Housing and Urban Renovation Law (Loi Accès au logement et urbanisme rénové, or ALUR) has introduced fundamental changes, with a provision for compulsory insurance (the Universal Rent Guarantee (Garantie universelle des loyers, or GUL) paid for in equal measure by landlord and tenant. Here, we will examine ways to improve the functioning of the rental housing market, with a focus on the relationship between landlords and tenants, while housing support policy will be addressed in the last part of this Note.
The problem of non-payment
24The problem of non-payment can be addressed in two ways:
- Deposits: according to the report by Baïetto-Beysson and Béguin (2008),  the proportion of tenants required to provide a deposit when renting a property rose from 25% in 1996 to 50% in 2006. This proportion has probably risen further since 2006.  The need to provide a deposit prevents the most vulnerable from accessing the private rental sector, and reduces their mobility. Yet in the event of non-payment, it is difficult to recover rent not paid from the guarantor. Thus, the role of the deposit is essentially a preventative one (with the tenant assuming a commitment to a close friend or relative), and is used by landlords as a signal used to select tenants;
- Insurance: the alternative to the deposit is the purchase of insurance by the tenant or landlord. The first system designed to cover the risks faced by landlords was the Guarantee for Unpaid Rent (Garantie des loyers impayés, or GLI) first offered by insurance companies in the 1980s. This is a standard insurance policy whereby the landlord pays a premium equal to between 2 and 4% of rent for cover against the risk of non-payment and receives compensation to cover legal costs in the event of a dispute. However, under the terms of this insurance, landlords cannot rent their property to groups deemed to pose a risk: indeed, insurance companies stipulate that the tenant must have an income equal to at least three times their rent and have a permanent employment contract. Some 700,000 dwellings are covered by GLIs (14% of the private housing sector).  To alleviate this problem, various inclusive guarantee mechanisms have been put in place, such as the Loca-pass and the tenancy agreement scheme (Garantie des risques locatifs, or GRL), which cover only a small proportion of leases.
25In general terms, insurance gives rise to moral hazard (the tenant may be encouraged to stop paying their rent) and anti-selection (only “problematic” tenants take out insurance). Moral hazard can arise with the signing of the lease itself (where the tenant intentionally chooses a property that is more expensive than the one they would have chosen if they had no insurance). There are two possible insurance models:
- Rent insurance based on the home loan insurance model, which limits moral hazard (only well-defined risks that are beyond the control of the tenant –death, unemployment, illness– are covered). However, anti-selection remains and landlords will continue to want to protect themselves against risks not covered by engaging in statistical discrimination in selecting their tenants;
- The cooperative management of risks based on the unemployment insurance model, which eliminates anti-selection (owner-landlords no longer select tenants) but reinforces moral hazard, which would result in sharp increases in non-payment without specific attempts being made to stem such increases. The GUL described in the ALUR law is based on this model.
For flexi-security in housing
26The GUL, which is compulsory, is financed by an insurance premium equal to around 1% of rent which is paid in equal measure by the landlord and the tenant. It covers all risks of non-payment that arise for whatever reason. In order for this reform to bear fruit in terms of access to private rental accommodation, while at the same time avoiding an increase in nonpayment that would render the system financially unstable, it must be followed through to its logical conclusion, creating a genuine flexi-security in housing based on three avenues: a relaxation of rules relating to leases; an enforceable right to housing (DALO) that is attached to the tenant rather than to the dwelling, with the introduction of joint management of the sector; and more robust management of non-payment.
Relaxation of rules relating to leases
27There is considerable scope for the simplification of leases in a number of areas, such as making it easier to end a lease when landlords wish to recover their property, while at the same time preserving rules regarding periods of notice. One option which could free up housing could be to align the default end of leases with a specific time of year, such as the end of August (just as seasonal leases run from Saturday to Saturday). Of course, it would be possible to suspend and resume a lease at any time, but this coordination would improve the allocation of housing. Greater flexibility could also be achieved by authorising leases of variable duration (one year, two years), rather than stipulating a three-year lease.
28One simple measure to reduce the risk of damage to a property or neighbourhood disturbances would be to allow the landlord to visit the property once a year –two months before the anniversary date of the end of the lease– to prepare an inventory of the property. It is essential that should the landlord choose to exercise this right, said inventory must be prepared in the presence of a professional, certified third party, and paid for by the landlord. If the landlord notes a serious deterioration in the premises –this concept is similar to that of gross negligence in the case of dismissal–, he or she will then have the right to give the tenant notice. Naturally, the tenant will have a right of appeal.
29Recommendation 1. Reinforce the fluidity of the housing market by broadening the range of reasons for the termination of a lease, allowing the landlord to inspect the property, aligning the ends of lease agreements with a specific period of the year, and encouraging leases of flexible duration as agreed between the tenant and the landlord. In return, DALO would be rendered effective by focusing this right on social housing.
Joint handling of disputes between tenants and landlords
30Disputes between landlords and tenants are common and can create mistrust. These disputes must be handled by a joint body such as that in place in Quebec, where one such institution has the power to terminate a lease.  The idea would be to put in place a structure, which we would call the Régie du logement (rental board), which would bring together all existing structures (such as rental cost observatories). Landlord and tenant representatives would sit on this body and resolve rent disputes under a certain threshold. Decisions reached by the board would be binding in the first instance as is the case for labour courts helping speed up cases and relieve congestion in the courts  and reducing costs. Within such a structure, the government would have a minority representation. The ideal principle to ensure that it is as representative as possible is that of elections held to appoint representatives of landlords and tenants. 
31Recommendation 2. Establish joint representation in rental housing, by developing a network of joint bodies under a Régie du logement, a network financed by the rent tax currently provided for in the ALUR law. Stipulate that all disputes under a certain threshold be heard by the Régie du logement, in particular those relating to the termination of leases due to non-payment.
The recovery of unpaid amounts
32As with any insurance policy, the GUL increases the chances of non-payment. Even so, the current situation is far from satisfactory: landlords bear the cost of non-payments, either directly or indirectly (via insurance premiums). Yet the government has no direct interest in speeding up procedures, since it does not meet the cost of non-payments.  This situation exacerbates landlords’ fears of a long wait before achieving a return on the property rented. The supply of housing available falls, resulting in higher rents and exacerbating the problem of non-payment.
33The situation is inefficient, representing as it does a compromise between the Quebec solution, whereby the costs of non-payments are met by the private sector but disputes are resolved in a prompt manner, and a situation in which the government slows down procedures but ultimately assumes responsibility for non-payments for which it is also be liable. With the government having decided to pool insurance via the GUL, we believe the pooling approach should be followed through to its logical conclusion, ensuring that the government internalises the costs associated with non-payments. This can be achieved by the government purchasing all debts associated with non-payments at a discount from the Régie du logement.
34Recommendation 3. The Régie du logement would intervene in all rental disputes, in particular to control rent increases deemed excessive and organise the GUL; this would dispose of debts associated with non-payment to government services, which would recover amounts due, in exchange for a transfer of funds from the government to the Régie.
35Another advantage of flexi-security would be that it allows landlords to manage lease agreements themselves, reducing transaction costs and, therefore, rents. Disputes, which risk becoming more widespread given the recently-introduced avenues for recourse and rent controls, will be referred to the Régie du logement.
The setting of rents
36The issue of rent controls has been at the heart of numerous debates since the shelving of the law of 1948, which has been described as a first-generation rent control mechanism. Second-generation rent controls govern changes in rent between leases: in the event of a change of tenant, a landlord cannot increase the rent on their property by more than the increase in a particular reference index. The decree of July 2012 implemented in densely-populated areas is part of this approach. Third-generation controls relate to the current system (except in densely-populated areas): rent is controlled for the duration of the lease, but freely reassessed when there is a change of tenant. The ALUR bill provides for more robust control of rents in densely-populated areas, with control now exercised in relation to changes in rent from one lease to the next but also in relation to a median level of rent. This type of control is often presented with reference to the system in place in Germany. However, it differs from the German system in a number of ways (Box 1).
1. Rent controls in Germany
In Germany, limits apply to increases in rent during the course of a lease (since leases are open-ended). If a landlord finds that the rent set for their tenant differs from the market rate, they cannot increase the rent of the tenant by more than 20% over three years. In France, the ALUR law aims to keep rents at a reasonable level in densely-populated areas, including from one lease to the next.
In Germany, a tenant can appeal to a judge if they believe that their rent is excessive (i.e. more than 20% above the rent paid for identical accommodation) and can prove that it was the only accommodation available at the time the lease was signed. This dual condition considerably reduces the chances of an appeal.
37Past experience and international examples would indicate that in practice, rent controls reduce the quality of housing: once the (reduced) quality of housing is taken into account, rents remain the same. At once, the yield on investment in the property diminishes, prompting the government to offer tax breaks to landlords –an additional drain on the public purse. Experience also suggests that second-generation controls tend to reduce mobility.
38The method used to control rents as described in the bill combines a standard in terms of space and of time. Yet crosssection regulation of rents (a given year) is extremely difficult, as information available on the rental market remains patchy. While the collection and publication of information on rents offsets a serious shortfall in this area, the size of the samples envisaged (for example, 5,000 observations on Paris, 25,000 on the Paris area) means that the large number of relevant characteristics within a particular district cannot be taken into account. 
39Recommendation 4. Rent controls, in their planned form, could give rise to inefficiencies in the private rental sector. Before any measure is implemented on a wide scale, trials must be carried out in pilot areas.
40Another form of intervention is tax exemptions targeted at the low end of the rental market. This approach would seek to encourage low rents, rather than prevent rent increases (cf. Dufflot system). A policy that consists of targeting rents below the median rent (in this case, 20% below the median rent), however, is difficult to implement as the median rent is difficult to measure. It can also encourage lower quality housing (noise, exposure to pollution). Rents will be 20% below the median, but in fact they will be at (or even above) market levels, given their characteristics as observed by government authorities. 
The social rental sector
41The French social rental sector is highly complex due to the multiplicity of agents, but also as a result of a degree of ambiguity in its objectives. With its waiting lists, regional inequalities, and low mobility, the current situation of the French social rental sector is unsatisfactory. In this section, we propose a number of approaches in order to move forward in these three areas.
Ambiguities in the low-income housing sector
42The social housing sector is organised into two large networks of similar sizes: the federation of 272 public offices for low-income housing (OPH), which operates under public law and manages 2.3 million homes; and the network of 261 housing associations (ESH), which operates under private law and manages 2.1 million homes. In addition to these two large networks there are 250 structures, including cooperatives, which manage all other social housing. Thus, the supply of low-income housing is managed by no fewer than 783 organisers, or close to eight per department on average. This proliferation of structures, which is associated with the history of the low-income housing movement, now poses problems when it comes to optimising construction, occupation, and mobility. In particular, it does not allow visible reductions to be made in the geographic disparities seen in France (Box 2).
2. The main characteristics of social housing
On average, low-income housing bodies are profitable, even if the mean reflects a certain heterogeneity. OPH (public offices of low-income housing) produce a margin to the tune of 10% of rent, while ESH (housing associations) produce a margin of 12%. [a]
In the social housing sector, the gross rate of nonpayment is estimated at 4.4% (compared with 1.4% in the private sector). [b] However, due to the different systems in place, including Loca-pass (which covers a quarter of new dwellings that have become social housing in the last decade), the financial consequences for social housing landlords are few: 0.7% in the housing association sector, and 1.1% in the OPH sector. [c]
Wide geographic disparities
The French social housing sector is characterised by wide disparities in tenant turnover and occupancy rates. In the Paris region, the rate of tenant turnover is less than 5%; in Provence-Alpes-Côte d’Azur, it is just 5.3%, while in Languedoc-Roussillon and on Corsica the rate is 6.7%. In these regions, the whole population is replaced after around every 20 years. In other metropolitan regions, the rate of tenant turnover is between 8% and 10%. [d]
Some regions have high vacancy rates, reaching up to 8% in OPHs in Alsace and between 5% and 7% of social housing in Champagne-Ardenne, Bourgogne, Franche-Comté, and Limousin. These disparities cannot be off-set by higher rates of construction in regions with a high population density.
43The relatively broad targeting of the social sector in terms of income (cf. Graph) is also a product of history. At present, three arguments can be made to justify this approach:
- Social mix: as seen above, the diversity argument is not fully valid unless social housing is not integrated into the city and organised in isolation. At present, social mix is seen more as the coexistence of social housing and the private rental sector within the same district;
- Political acceptability: since social housing is, to a large extent, financed from a public savings account which attracts a low interest rate (Livret A), there is a political logic to a large proportion of the population having theoretical access to it.  However, this theoretical possibility could also lead to frustration: in 2010, just 14.5% of French households were in low-income housing, while in theory 64% were eligible for PLUS accommodation;
- Financial equilibrium: with 10 to 20% of funds for the (re)construction of social housing coming from self-financing,  low-income housing bodies must create a margin; hence the decision to diversify the profile of their tenants and, in particular, to include among their clients a fraction of tenants from the middle class, with more regular income than the lower class. This income-based approach contradicts the social approach. The more ambitious the objectives assigned to social housing landlords in terms of low-income housing production (for example, 150,000 homes per year), the more prevalent this income-based approach will be relative to the social objective.
Improving the performance of social housing
Social and first-level social housing
44Should more social housing be built? In view of the long waiting lists that currently exist, this may seem a surprising question. It is less surprising when one considers that social housing tends to crowd out private housing  in a general context of a rationing of supply. The two sectors in effect compete for land, while in terms of demand a household in the social housing sector frees up a home in the private sector. 
45Thus, the objective of expending social housing supply could be to reduce its cost for the lower and middle classes, in order to reduce expenditure on housing. Rents are effectively reduced for households in receipt of social housing. However, for those living in private rented accommodation, the fall in rents (which is associated with the release of housing in this sector) is only temporary, with lower returns to the landlord resulting in a reduction in investment in rental properties.
46The main objective of expending social housing supply should therefore be to offer a decent alternative to those who cannot rent in the private sector. Social housing provides accommodation to just one in two disadvantaged households, a surprising situation, although the proportion of disadvantaged households living in social housing has risen considerably since the 1970s. 
47The ALUR bill (Chapter 5) clearly sets out the responsibility of social housing landlords vis-à-vis the enforceable right to housing. The resources of bodies should be returned to this marginalised population for PLAI/PLUS  accommodation, while the expansion in the supply of social housing allocated to the lower and middle classes, which is a profitable activity, could be opened up further (see infra).
48Thus, the risk is one of greater spatial segregation. The appropriate measure by which to evaluate this risk is not the municipality, unless it merges with the town or city. Otherwise, the relevant level of assessment is the agglomeration (a town and its surrounding suburbs). To avert this risk, the location of new first-level social housing should be left in the hands of inter-communal bodies and government subsidies based on transparent, meaningful spatial segregation indicators.  Housing grants for social housing would be reserved for intercommunal bodies that submit a programme for the demolition, sale, and construction of first-level social housing that reduces these segregation indicators. The social mix programme could also be encouraged by the introduction of private housing in former joint development zones.
49Recommendation 5. In areas with a shortage of first-level social housing (PLAI and PLUS), build such housing using the resources of low-income housing bodies, preferably in mixed-purpose buildings. Reserve government subsidies for projects that reduce spatial segregation, as measured using a transparent indicator at the level of the agglomeration, and replace article 55 of the SRU law with this policy to foster social mix at the same level.
50We have seen that social mix within the low-income housing sector allows housing bodies to balance their accounts. A second reform avenue is therefore the development of social housing for the lower and middle classes (PLS/PLI),  for which there are great differences between geographical areas in terms of access. One way to reduce differences in this “profitable” segment of the social housing sector would be to have existing structures –low-income housing bodies and private landlords– compete for new housing, in particular in densely-populated areas. In practice, local authorities that wish to implement a low-income housing programme would launch an invitation to tender aimed at all landlords and developers in France, who would compete to produce and manage the programme under the rental conditions set out in the invitation to tender for the duration of its repayment period (40 to 50 years). At the end of this period, the status of the property may change, be sold, or become part of the private rental sector, as in Germany. Each participant in the invitation to tender could receive loans from the Caisse des Dépôts for the whole programme. This system would attract funds from the private sector, generating a return on the output value of the asset, without using government resources (with some exceptions) other than the opportunity cost of the credit lent by the Caisse des Dépôts. 
51Recommendation 6. For the construction of social housing (PLS, PLI), have low-income housing bodies and private structures compete on a national level for the production and management of new housing, in particular in densely-populated areas. Allow landlords, if they so desire, to use housing for other purposes after a period of 40 to 50 years with the repayment of the subsidy received at the end of the period, where applicable.
Coordinate regional readjustment
52Having organisations compete on a national level will not reduce regional imbalances without coordination at a national level. This coordination must be guided by a simple principle: at a given level of income, a household should no longer be eligible to receive social housing in one area over another. This spatial equity can be pursued by developing an indicator to measure the tension between demand for, and supply of, housing for each objective (social or first-level social housing). For first-level social housing, one simple indicator is the relationship between the rate of poverty and the proportion of PLAI/PLUS households. For social housing, demand can be measured by using the number of months’ income that a household on the median disposable income (median for the agglomeration or urban area) must allocate to buy one square metre of housing.  This indicator of tension in the private sector could be related to the proportion of homes in the rental sector.
53These two indicators of tension have several advantages when coordinating the territorial dimension of social housing policy over the current zoning policy. First of all, these indicators are continuous (they avoid the threshold effect); secondly, they cannot be manipulated by local agents at the level of an agglomeration or urban area; and finally, they can record demographic, economic, and migratory changes from year to year, as well as distributive shocks that affect an area and can lead to changes in demand for social housing.
54Once objective criteria have been established to decide between different zones in their demand for social housing, mechanisms to ensure funding priority to social housing in the most densely-populated areas must still be put in place. Competition across France on a specific programme is one answer for social housing, but not for unprofitable first-level social housing. To achieve a readjustment, we are calling not for the “nationalisation” of low-income housing bodies, but for a more federal approach.
55One system, which was put in place in 2011, consists of taking part of the financial means of landlords who do not use their own funds, to redirect them to densely-populated areas. This system is to be replaced by a pooled fund, as provided for in the agreement reached between the government and the Union sociale pour l’habitat in July 2013. The contributions of landlords will be calculated as a function of their shareholding in the rental social housing guarantee fund, rents received, and number of dwellings. In return, support provided to landlords will depend on the number and location of new homes built. Landlords in densely-populated areas may also receive support for the redevelopment/destruction of obsolete housing stock. While it is too early to determine the genuinely redistributive nature of such an equalisation system, the progress made has been significant.
56Recommendation 7. Use transparent indicators of tension to ensure equalisation between low-income housing offices and allocate subsidies across regions. Put in place instruments to evaluate the pooled fund for low-income housing bodies.
Methods of allocation and mobility
57The rules for the allocation of social housing are not very transparent and are seen as random, and even fraught with suspicions of favouritism. Rules for allocation and the management of rents must allow social housing to play its role as a safety net in terms of housing, while at the same time encouraging greater mobility within and out of the social housing sector. One solution is to adopt the one-stop shop system at an inter-communal body level that includes all landlords, such as that adopted in Rennes: the allocation commission brings together all landlords operating in an inter-communal area. To match applicants and vacant housing effectively, information on the preferences of both parties (applicants and landlords) must be available:
- Applicants: at present, information to be provided by an applicant essentially relates to their household and income, not their preferences. Thus, an applicant may refuse an offer of social housing, in which case the allocation procedure is drawn out. The system implemented in Delft, which was subsequently rolled out across the Low Countries and is now in place in the United Kingdom (Choice Based Lettings), consists of providing applicants with a description of vacant housing and its characteristics and the rent on these properties (similar to a real estate advertisement). Applicants must express an interest in at least one dwelling in order for their application to be taken into consideration. If their application is accepted (on the grounds that they are considered priority applicants), they are matched with the property, reducing the number of return trips. While we believe such a system should be trialled before being rolled out on a large scale, the feedback on results in other countries has been very positive;
- Landlords: under the current system, each financier (government, local authority, landlord) is entitled to a quota (statutory rights) determined as a function of their financial contribution. Each financier arrives with their list of priorities (hence a feeling that the system is very opaque), with priority rules varying from one financier to another. In Rennes and its surroundings, a transparent points-based system has been trialled with a single form to be completed, and all quotas abandoned. This system can be amended if it is too difficult to prioritise all demands: these being the product of approaches that differ according to whether they relate to social housing, first-level social housing, or on the grounds, to be reiterated, of reducing commuting distances. Thus, what we propose is a points system within each of these three categories, with housing to be allocated across these three categories using a discretionary quota approach.
58A commission that brings together all stakeholders (landlords and financiers) at an inter-communal level (coordinated by regional authorities) would deliberate at each meeting on the share of vacant social housing to be allocated according to each criterion as a function of demographic, migratory, social, and economic contingencies, and display this allocation for the meeting that will determine allocation. Statutory rights would be abandoned. On the other hand, a contingency of around 10% would be reserved for emergency situations (domestic violence, insalubrity, etc.) assessed on a case-by-case basis.
59For housing that has not been allocated after several meetings, a national auction could be held, in order to attract people from other regions (via associations).
60Recommendation 8. For the allocation of social housing, adopt the one-stop shop system at an inter-communal level and include all landlords. Have applicants express their preferences in terms of vacant housing. Make the choices of landlords more explicit in the form of a points system within each main objective of social housing, while retaining a small quota for emergency situations.
61In the Paris region, the average period of occupation of social housing is close to 20 years. This immobility results in waiting lists that are particularly detrimental to young people. The difference between social housing rents and rents in the private sector is such that households allocated low-income housing have no incentive to leave it. In general, rent is independent of the situation of the tenant (as in the private sector), except when the income of the tenant is more than 20% above the threshold for receipt of low-income housing. We propose that this rent surcharge be applied as soon as the tenant exceeds the income threshold and that the period of occupancy be taken into account, to reaffirm the principle that social housing is a right that expires once the recipient exceeds the eligibility threshold. After a significant period of occupancy in densely-populated areas, rent surcharges would be applied as a function of their period of occupancy, taking the age and income of the occupants into account.
62Recommendation 9. In order not to perpetuate inequalities between households on similar incomes, depending on whether or not they have access to social housing, implement rent surcharges calculated as a function of income but also of the period of occupancy, in order to increase mobility within the social housing sector.
Incorporate housing support Incorporate into a broader housing redistributive support policy
63Housing for low-income households in France is based on two instruments: social housing and housing support. In reality, this dual system results in great inequality between households with the same level of income, depending on who has or does not have access to social housing. This horizontal inequality results not only in frustration, but also in a degree of immobility within the social sector. To remedy this situation, we propose, in addition to rent surcharges calculated as a function of the period of occupancy of a social dwelling (cf. Proposal 9), a rethink of the housing support system.
- APL, or individualised housing assistance, which is applied irrespective of the composition or age of the occupants. This assistance is provided to a particular sector of housing that includes, in addition to certain homeowners, ordinary housing and sheltered accommodation the object of an agreement between the government and the landlord;
- Housing Allowance (AL), which covers Family Housing Allowance (ALF) and Social Housing Allowance (ALS). For the most part, ALF is given to individuals who live in housing that does not entitle them to APL. ALS is given to households that do not qualify for APL or ALF.
65While the method used to calculate this support is very complex,  it is the same method used for all housing allocations: it puts housing support halfway between allocated subsidies and global transfers such as the active solidarity income (RSA). In many cases, rent is in excess of the threshold; however, support will depend solely on income, in particular in the unregulated sector.
66Housing support must have more than an equity objective, for three reasons. The first reason is that there are different types of inequality (vertical –between households with different incomes– and horizontal –between households with the same income but which are treated differently). Another reason is that the tax system (not just grants and subsidies) must also be taken into account. The third reason is that the price of housing cannot be considered as independent of the subsidy and tax system. Four objectives must be pursued in order to bring about an overhaul of the system:
- Equity between social housing and the private rental sector: the first objective would not exist if all households wishing to access to social housing could do so. Living in social housing could then be a matter of choice, and the difference in rents would to some extent be offset by differences in the characteristics of dwellings. This is not the case in densely-populated areas, where there are long waiting lists for social housing, resulting in rationing. Thus, two households on identical incomes that wish to access social housing find themselves in very different situations in terms of well-being, depending on whether or not they have access to this housing. Housing support should reduce this difference, since the support provided depends on rent paid. However, in densely-populated areas the threshold is very quickly reached, which means that it is not possible to offset the difference in rent.
- Reduce the inflationary effects of subsidies: several studies have shown that housing support, as it is currently designed and paid, is appropriated (at least in the short-term) by landlords. Support will increase demand, and therefore the equilibrium price, in particular if supply is rigid. However, the fact that housing support can be paid directly to landlords exacerbates this appro priation: the landlord can advertise rent net of subsidy and increase the gross rent payable as those increase. This practice appears to be a reality in particular in the rent of rooms to students;
- Ensure coherence across all grants to low-income groups. There are three mechanisms to provide assistance to low-income households in France: the Revenu social d’activité (RSA), or social income; the Prime pour l’emploi (PPE), a tax credit program; and housing support. Around 11 billion euros were spent on the first two of these mechanisms in 2011,  compared with 16 billion on housing support. While France is not the only country that combines support mechanisms for those on low incomes, it may seem unique that the Housing Allowance should be the focus of our support system. Greater coherence and a simplification of the methods used to calculate entitlements are therefore desirable, together with a rationalisation of management processes;
- Fiscal neutrality between landlords and tenants: this objective is far from being achieved for main residences. In particular, income tax gives landlords preferential treatment over tenants in relation to the main residence.  In this Note, what is important to us is not so much the consequence to financing the economy, as it is the fiscal bias in favour of home ownership. If the rate of home ownership was 30% as in Switzerland, this would not be of consequence. But with home ownership at 58%, this bias is of greater concern in a country with no shortage of rigidities in other areas.
Reform for four objectives
67We suggest that these four objectives be addressed using a single tax mechanism, accompanied by consistency in, and simplification of, the formula used to calculate housing support.
68Recommendation 10. Integrate housing support into the income tax mechanism. For households that pay income tax, deduct rent paid by tenants from their taxable income, up to a limit determined by the composition of the household. For households that do not pay income tax, turn housing support into a negative tax, managed by family benefits funds.
69Where possible, foregone income tax receipts will be offset by a moderate increase in property taxes assessed on the net commercial value of outstanding loans. This proposal is based on principles of social justice that aim to include irreducible expenditures in the calculation of one’s ability to pay income tax.
70For households that do not pay income tax, support will be provided in the form of monthly advances from the family allowances fund and adjusted by the tax office at yearend. It should be noted that the formula for calculating housing allowance will be clarified, with only a marginal effect on subsidy profiles. 
71This simple measure can improve the situation as regards the four objectives stated above:
- Equity between social housing and the private rental sector: this measure reduces differences in how households in the social housing sector and those in the private sector are treated, since the difference between private rent paid and housing support is tax-deductible. Naturally, the level of payments depends on the threshold. Neutrality can be reinforced if the threshold depends on the status of the occupier;
- Reducing the inflationary impact of support: for households that pay income tax, the extent to which rent is subsidised becomes a function of the marginal rate of tax. Unless the tenant shows their landlord their tax return, the latter has no way of knowing the implicit rate of the subsidy on rent. This introduces greater balance in negotiations between the two parties, for example, when renewing the lease; it should also reduce the appropriation of support by landlords;
- Ensure coherence across all support to low-income groups: once housing allocations have been taxed, the way is clear to integrate them into a legible and coherent support system for low-income households;
- Neutrality between landlords and tenants: there is strong opposition to the proposal to tax implicit rents,  in part due to the increase in house prices, which results in an unexpected increase in the tax burden. One alternative (but not equivalent) measure is to deduct rent from tenants’ taxable income.
72It should be noted that in the United States, rent can be deducted from taxable income for a limited period (six months) to encourage professional mobility.
73This reform increases tenants’ purchasing power. However, it will only be of benefit if accompanied by a policy that loosens the various constraints on supply, as recommended in CAE Note no 2. In this regard, the provision that entrusts the design of the local town development plan (PLU) to the inter-communal body in the ALUR bill is to be welcomed as is the development of Greater Paris.
Aix-Marseille School of Economics and EHESS, Member of the Conseil d’analyse économique.
Sciences-Po Paris, LIEPP, Member of the Conseil d’analyse économique.
Statistics drawn from the Compte Logement 2011. On property ownership, see Trannoy A. and É. Wasmer (2013a): “Comment modérer les prix de l’immobilier?”, Les Notes du CAE, no 2, February.
In general, rent controls are not seen as an appropriate instrument, since they are inconsistent with the objective of efficiency.
See Fack G. (2005): “Pourquoi les ménages pauvres paient-ils des loyers de plus en plus élevés?”, Économie et Statistique, vol. 381, no 1, pp. 17-40.Online
Cf. Algan Y. and P. Cahuc (2007): “La société de défiance: comment le modèle social français s’autodétruit”, Opuscule du CEPREMAP, no 9, Éditions Rue d’Ulm.
PLUS (Prêt locatif à usage social), a subsidised loan for the construction or acquisition of low-cost housing, is the main tool for financing social housing. Cf. Union Sociale pour l’Habitat (USH) (2013): “Les plafonds de ressources”, Fiches Thématiques, no 6.
Calculations by authors based on Pôle emploi, DARES (jobseekers registered with the Job Centre required and to actively search for employement) and INSEE (RP 2009 supplementary utilisation) data.
Between 2005 and 2007, the annual rate of turnover was 3-4% for owner-occupier properties and 9.5% for social housing tenants, compared with 18% for private sector tenants. See Commissariat Général au Développement Durable (GGDD) (2009): “La mobilité résidentielle progresse dans le parc locative privé et diminue dans le parc social”, Le Point Sur, no 27.
Oswald highlights a close correlation between home ownership and the rate of unemployment, cf. Oswald A.J. (1996): “A Conjecture on the Explanation for High Unemployment in the Industrialized Nations: Part 1”, University of Warwick Economics Research Paper, no 475.
See Costes N. and S. El Kasmi (2013): “Les freins à la mobilité résidentielle pénalisent-ils la qualité de l’appariement sur le marché du travail?”, Trésor-Éco, no 116, Direction générale du Trésor.
See Trannoy A. and É. Wasmer (2013b): “La politique du logement en France” (CAE Working paper) and Artus P., A. Bozio and C. García-Peñalosa (2013): “La fiscalité des revenus du capital”, Les Notes du CAE, no 9, September.
This immobility has been protected by the rule on maintenance of the premises (and not of housing on a local or national level) since 1948.
Egert B. and R. Kierzenkowski (2010): “Exports and Property Prices in France: Are They Connected?”, OECD Economic Department Working Paper, no 759 and Askenazy P. (2013): “Capital Prices and Eurozone Competitiveness Differentials”, Forschungsinstitut zur Zukunft der Arbeit Discussion Paper Series, no 7219.
Baïetto-Beysson S. and B. Béguin (2008): Faciliter l’accès au logement en sécurisant bailleurs et locataires: place d’un système assurantiel, Report published at the request of Christine Lagarde, Minister for the Economy, Industry, and Employment, and Christine Boutin, Minister for Housing and Cities, La Documentation française.
See Ministère du Logement (2013): Étude d’impact du projet de loi ALUR.
Cf. Baïetto-Beysson and Béguin (2008), op.cit.
Housing Regulations of Quebec (2006): Annual Report 2004-2005, Bibliothèque nationale du Québec.
3% of lease agreements result in legal action. Ministère du Logement (2013), op.cit.
Electoral roles will be prepared using tax returns (property tax returns for landlords, housing tax receipts for tenants).
Wasmer É. (2007): “Analyse économique du marché du logement locatif”, Revue Économique, vol. 58, no 6, pp. 1247-1264.
See OLAP (2012): Dossier de l’OLAP, no 25. This can be significant, since an article of the ALUR bill government that the reference rent measure can be used by the tenant once they are in their homes to contest their rent. To illustrate the potential deviation of a rent control mechanism, we have simulated a hedonic regression equation in which only the surface, number or rooms, and the arrondissement of Paris (the variables used by OLAP to explain private rents) were incorporated as explanatory variables. At least 30% of households living in low-income housing paid rent equal to more than 20% above the reference median rent.
Indeed, assessments of hedonistic rents indicate that with three or four characteristics below the mean (poor soundproofing, ground floor apartments, proximity to arterial roads or any other nuisance), rent can very easily be 20% below other rents: thus, the rent is offered at the market price but within a quality range that is at least 20% below other accommodation.
In political economy, a policy can only be targeted at a very large proportion of the population, or it will lose political support. This is illustrated in the generalist nature of social housing, as with family policy. See De Donder P. and J. Hindriks (1998): “The Political Economy of Targeting”, Public Choice, vol. 95, no 1-2, pp. 177-200.
On average, the subsidised loans provided by the Caisse des Dépôts cover 71% of finance and local authorities no more than 10%, cf. Ministère de l’Egalité et des Territoires (2013): Bilan des logements aidés année 2012.
See, for example, Chapelle G. (2013): Does Social Housing Crowd Out Private Construction?, Mimeo Sciences-Po Paris, or M.D. Eriksen and S.S. Rosenthal (2010): “Crowd Out Effects of Place-Based Subsidized Rental Housing: New Evidence from the LIHTC Program”, Journal of Public Economics, vol. 94, no 11, pp. 953-966.Online
However, one additional unit of social housing “frees up” less than one unit of private housing, since an increase in the supply of social housing may prompt recipients to no longer live together, attract new local populations, or allow households to increase in size.
See Houard N. and H. Sainte Marie (2013): Concertation sur les attributions de logements sociaux, Work group report for Cécile Duflot, Ministère de l’Egalité, des Territoires et du Logement, La Documentation Française, Fgure 1, p. 10.
PLAI: Subsidised rental loans for integration (one-third of households were eligible in 2008); PLUS: A subsidised loan for the construction or acquisition of low-cost housing (64% of households were eligible), cf. Union Sociale pour l’Habitat (USH) (2013) op.cit.
This point is developed in Trannoy and Wasmer (2013b) op.cit. Article 55 of the SRU law (which imposed upon each village and town the obligation to have at least 20% of housing allocated for social housing) is an indirect tool and too large for this objective of social mix. Moreover, to date it has not been effective, see Bono P.H., R. Davidson and A.Trannoy (2012): “Analyse contrefactuelle de l’article 55 de la loi SRU sur la production de logements sociaux”, AMSE Working Paper, no 2013-05. The reinforcement of penalties should not have a massive effect, since they do not exceed 5% of the budget of a village or town.
PLS: Low-rental housing loans (84% of French households were eligible for these loans in 2008); PLI: Intermediate rental loans (90% of French households were eligible for these loans in 2008), cf. Ministère du Logement (2009): Question no DL55, PLF 2009.
Tax conditions would be the same for all participants (3.6 billion euros in tax breaks for the low-income housing sector in 2011, according to the Compte Logement 2011).
Courtesy of the Bases Notaires, the details of purchase prices are better known than rents.
Commissariat Général au Développement Durable (CGDD) (2012): Compte du Logement 2011, 1st. results for 2012, Service de l’Observation et des Statistiques.
See Box 7 in Trannoy and Wasmer (2013b) op.cit. and ministère de l’Égalité, des territoires et du Logement (2012): Éléments de calcul des aides personnelles au logement.
Senate Finance Committee (2011): Rapport sur la loi de finances 2012 and Caisse Nationale des Allocations Familiales (2012): “Étude sur le Revenu de solidarité active”, Dossier d’Étude de la CNAF, no 156.
See Artus, Bozio, and García-Peñalosa (2013) op.cit.
For more information, see Trannoy and Wasmer (2013b) op.cit.
Artus, Bozio, and García-Peñalosa (2013) op.cit.