Impact of information letters on participation in Payments for Environmental Services

In this project, we sent letters to eligible farmers (located near zones where PES are available) informing them of their eligibility, of the type of program available to them (practices subsidized and subsidy amount) and providing them with the contact information to caseworkers in charge of helping them with their application. We also varied the content of the letter, adding testimonies of farmers already benefiting from PES and the ease of response by adding easy to use pre-stamped return letters. Finally, we test for the existence of a social multiplier in the numbers of letters sent by varying the proportion of farmers in a given commune receiving the letter.

2018-01-01

2018-05-15

All outcomes are measured during the 2018 season
1/ Overall take up of PES contracts
2/ Overall area under PES
3/ Uptake of "hard core" PES or easy PES

All outcomes are measured during the 2018 season
1/ Overall take up of PES contracts (0/1 dummy)
2/ Overall area and/or length under PES (we have to convert some length units in area units, for example by assuming a constant width (3m) for hedgerows)
3/ Uptake of "hard core" PES or easy PES (hard core and easy PES are coded beforehand and independently by the CR HdF, this indicator measures whether the PES is more or less stringent)

1/ Contact taken with caseworkers

1/ Contact taken with caseworkers: measures whether a caseworker has been contacted by the farmer (0/1)

We randomly select 10205 eligible farmers to receive the letters and 10169 to be in the control group.
Among the treated farmers, 5104 receive only the basic letter informing them of their eligibility, of the subsidies available and of the contact information of the caseworker; 3401 receive additionally a testimony of farmers already benefiting from the subsidy; 1700 receive additionally the easy to use pre-stamped reply letter.
There is also a first stage where communes are selected randomly in order to study possible social multiplier effects of the letters. 5427 treated farmers and 5430 control farmers are included in communes where only 50% of the eligible farmers are treated; 3448 control farmers are included in communes where noone receives the treatment; 3482 treated farmers are located in communes where every eligible farmer is treated.
There are also 1294 treated and 1293 controls that are located in communes only participating in the second stage of the experiment (their communes were too small to study diffusion effects but they could be included in the main effect specification).
Stratification was done by PES zone and by commune. That is we selected first communes in PES zones, and then farmers within communes.

Randomization done in office by computer.
The approach used was to classify farmers within strata by the value of a uniformly drawn random variable and to assign them in cycle to the different treatments.

There is a two step randomization process, with first communes randomly drawn within PES zones and then farmers drawn within communes.

Yes

542 communes.

20474 farmers.

In the final sample, we have 325 communes where 50% of eligible farmers receive a letter, 107 super control communes where no farmer receives a letter and 110 super treated communes where all eligible farmers receive a letter.
The design can also be analyzed directly by comparing treated and control farmers: 10205 treated and 10169 control. Among the treated farmers, 5104 receive only the basic letter informing them of their eligibility, of the subsidies available and of the contact information of the caseworker; 3401 receive additionally a testimony of farmers already benefiting from the subsidy; 1700 receive additionally the easy to use pre-stamped reply letter.

Power calculation has been conducted using randomization inference. MDE for 80% power, 2sided t-test, 5% size (in percentage points): - Direct effect of letter (all treatments merged, no diffusion): 0.75 p.p. - Effect of simple (bare) letter: 0.98 p.p. - Effect of letter with testimonies: 1.02 p.p. - Effect of letter with testimonies and easy to use reply: 1.45 p.p. - Effect of testimonies vs bare letter: 1.10 p.p. - Effect of testimonies + easy response vs bare letter: 1.45 p.p. - Effect of testimonies + easy response vs testimonies: 1.52 p.p. - Effect of treatment on the untreated in communes with 50% treated (vs control): 1.3 p.p. - Effect of saturated effect vs control: 1.55 p.p. - Effect of saturated vs 50% treated: 1.35p.p.

PreAnalysisPlan