Essay！！

Guessing Benefits

Richard Martin

University of Victoria

October 16, 2020

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 1 / 19

Recap: no market for environmental quality

The only problem is that there is no market for environmental quality, so no direct way to estimate the demand curve for environmental quality.

So instead we use differences in consumer surplus in related markets as a (partial) measure of the willingness to pay for an improvement in environmental quality. We consider the following 3 techniques:

1 Mitigating expenditures. 2 Hedonic estimation. 3 Surrogate markets.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 2 / 19

Mitigating expenditures

Mitigating expenditures.

Suppose that we had two cities identical in all respects except one city had poor air quality and one city had good air quality. Obviously impossible, but we can probably get close. Winnepeg and Hamilton are quite similar in a lot of respects, but being on the prairies Winnepeg’s air quality is much higher (it is windy and isolated). With mitigating expenditures we look at differences in consumer surplus for goods that can reduce exposure to pollutants: one such good is an air purifier.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 3 / 19

Mitigating expenditures

Air purifier price

Air purifier quantity

Figure 1: Difference in CS due to difference in air quality.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 4 / 19

Mitigating expenditures

Yes, I would be willing to pay more for an air purifier if I lived where air quality is poor, but...

This difference in consumer surplus (for air purifiers) is going to way under-estimate my willingness to pay for a clean environment...Why?

1 Even if we added up the differences in consumer surplus across all possible defensive expenditures, you can’t defend yourself completely from poor environmental quality.

2 It is kind of a pain in the ass to have to defend yourself.

Figure 2: What pollution?

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 5 / 19

Hedonic Estimation

Hedonic estimation can be used when environmental quality has an influence on market prices.

Examples include housing prices and wage rates. Consider housing prices: many other things influence housing prices besides environmental quality. One thing you might consider is matching houses on all observable characteristics, and then looking at how the price difference between matched houses varies with environmental quality.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 6 / 19

Hedonic Estimation

Rather than matching, hedonic estimation utilizes regression analysis to control for the impact that other factors have on housing prices.

Controlling for other influences allows us to isolate the relationship between environmental quality and housing prices.

We will use something called an Additive Model1 to model housing prices in Boston.

Data is from Harrison, D. and Rubinfeld, D.L. (1978) Hedonic prices and the demand for clean air. J. Environ. Economics and Management *5*, 81-102. (but based on 1970 US Census)

1www.stat.cmu.edu/~ryantibs/advmethods/notes/addmodels.pdf Richard Martin (University of Victoria) Chapter 7 October 16, 2020 7 / 19

Hedonic Estimation

Housing Values in Suburbs of Boston

‘crim’ per capita crime rate by town. ‘indus’ proportion of non-retail business acres per town. ‘nox’ nitrogen oxides concentration (parts per 10 million). ‘rm’ average number of rooms per dwelling. ‘dis’ average distance to five Boston employment centres. ‘tax’ full-value property-tax rate per 10,000. ‘ptratio’ pupil-teacher ratio by town. ‘lstat’ lower status of the population (percent). ‘medv’ median value of owner-occupied homes in 1000s.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 8 / 19

Hedonic Estimation

From these supplied variables we define two additional variables: 1 house.price = medv ∗ 1000 (in dollars rather than thousands) 2 air .quality = 1NOx (air quality: inverse of the concentration of NOx)

The average house price in the data is $22,532: seems ridiculously low but $22, 532 × 1.0692549 = $599, 354, which is not far from the average house price in 2019.2

The additive model we fit is given by the following equation, where α is the intercept, s() indicates this is a term to smooth and � is the error term.

house.price = α + s(crim) + s(indus) + s(air .quality) + s(rm)

+s(dis) + s(tax) + s(ptratio) + s(lstat) + �

2https://www.zillow.com/boston-ma/home-values/ Richard Martin (University of Victoria) Chapter 7 October 16, 2020 9 / 19

Hedonic Estimation

0 40 80− 20

00 0

0 10

00 0

30 00

0

crim

ch an

ge p

ric e

0 10 25− 20

00 0

0 10

00 0

30 00

0 indus

ch an

ge p

ric e

1.5 2.5− 20

00 0

0 10

00 0

30 00

0

air.quality

ch an

ge p

ric e

4 6 8− 20

00 0

0 10

00 0

30 00

0

rm

ch an

ge p

ric e

2 6 10− 20

00 0

0 10

00 0

30 00

0

dis

ch an

ge p

ric e

200 500− 20

00 0

0 10

00 0

30 00

0

tax

ch an

ge p

ric e

14 18 22− 20

00 0

0 10

00 0

30 00

0

ptratio

ch an

ge p

ric e

10 30− 20

00 0

0 10

00 0

30 00

0 lstat

ch an

ge p

ric e

Figure 3: Partial effectsRichard Martin (University of Victoria) Chapter 7 October 16, 2020 10 / 19

Hedonic Estimation

If we improve air quality from 1.5 to 2 (a 25% reduction in the concentration of NOx) how much would the price of an otherwise equivalent house increase?

The model predicts price would increase by about $3275, which doesn’t seem like much but remember the average price of a house in Boston in 1970 was $22,532.

Thus the difference in pollution causes a price difference which is about 15%, which is substantial.

The price difference is the capitalized value of the stream of benefits: we typically use the stream value.

If we assume that the stream is a perpetuity and the appropriate discount rate is 7%, the stream value is $3, 275 × .07 = $230year .

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 11 / 19

Hedonic Estimation

1.5 2.0 2.5− 20

00 0

0 10

00 0

20 00

0 30

00 0

Hedonic Price Function

air.quality

H ou

se P

ric e

3275.69

1.5 2.0 2.5

0 20

00 60

00 10

00 0

Demand for Air Quality

air.quality

M W

T P

3212.15

Figure 4: Partial effectsRichard Martin (University of Victoria) Chapter 7 October 16, 2020 12 / 19

Surrogate Goods

We can look at differences in travel costs incurred as a proxy for the value visitors place on a difference in environmental quality.

Consider two identical recreational sites except one has a higher level of environmental quality: the difference in travel costs could indicate how much people value higher environmental quality. Example: consider two similar lakes, one of which has clean water, and one is infested with rock snot:

Figure 5: Rock Snot

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 13 / 19

Surrogate Goods

To do a travel cost analysis we would question visitors as they arrive at the two lakes about the (opportunity) cost of their travel to reach the site.

Travel cost should be per person (group size matters) and per day (duration of visit matters). When would opportunity travel cost deviate from out of pocket cost? We would also want to collect demographic data to attempt to compare apples to apples at the two sites.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 14 / 19

Surrogate Goods

Trip cost

Trips

Figure 6: Difference in travel costs when enviro quality differs.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 15 / 19

Surrogate Goods

Travel cost under-estimates willingness to pay for cleaner environment:

My travel cost is bound by the value I place on the visit not equal to the value I place on the visit.

I would not incur larger travel costs than my value: presumably most visitors are left with some surplus.

Sample-selection bias: we only survey visitors to the sites, which will obviously be a non-random selection of the population.

Biased by non-use and no-use.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 16 / 19

Contingent Valuation

So all the methods of estimating demand for environmental quality underestimate benefits... What to do?

Why not just directly ask people how much they value the environment? What is the maximum you would be willing to pay... Contingent Valuation (CV): ask people how they would act in different (hypothetical) situations. Advantages:

1 can capture both use and non-use values. 2 the other methods we looked at are limited in what can be studied:

with CV you can ask people about anything!

Disadvantages: 1 Hypothetical nature: people can give ridiculous answers with no

consequence. 2 Can’t force people to answer surveys: selection bias. 3 Framing and endowments matter... even when they shouldn’t.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 17 / 19

Contingent Valuation

CV method: 1 Identify population of interest, environmental issue. 2 Selection of random sample from population. 3 Choose how to administer survey (shoe leather, phone, online) 4 Try to establish how much your survey suffers from selection bias. 5 Use the measure of WTP in cost benefit analysis.

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 18 / 19

Contingent Valuation

Experiment to look at framing and endowment effects:

https://web.uvic.ca/~rpmartin/coffee/

Richard Martin (University of Victoria) Chapter 7 October 16, 2020 19 / 19