10.04.2013 Views

Lisø PhD Dissertation Manuscript - NTNU

Lisø PhD Dissertation Manuscript - NTNU

Lisø PhD Dissertation Manuscript - NTNU

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Building economics of climate change<br />

For simplicity, we include maintenance in the conversion<br />

cost function whenever a conversion takes place.<br />

s s<br />

s<br />

V ( M) = v ( W, z ) −c( W,<br />

z ) −m<br />

3<br />

d 3<br />

Mi<br />

+ T all s∈S 2<br />

d<br />

2 2 2 2 2 2<br />

(9b)<br />

V 2s (T)=T 2s all s ∈ S Ti (9c)<br />

The V 2 -functions above are a kind of indirect utility<br />

functions and their values depend on the z and W values<br />

that are chosen in period 1. The choice of action in<br />

period 2 can be seen as a three-step procedure. First,<br />

the state of period 2 is observed. Next, the optimal effort<br />

under the M-strategy and the conversion activity under<br />

the C-strategy are calculated. In the last step, the value<br />

of the building under each of the three strategies are<br />

compared, and the strategy yielding the highest value is<br />

chosen.<br />

Similarly, the owner has three different possible<br />

actions in period 1. The object function for period 1<br />

choice will be somewhat more complex than the object<br />

functions for period 2. There are two reasons for this:<br />

(1) When the choices of period 1 are made, the<br />

future impact of climate change on building<br />

enclosure performance is not known. The value<br />

of the building under each of the possible<br />

actions in period 2 is affected by the choices of<br />

period 1.<br />

(2) The choices of period 1 will partly determine<br />

which actions are optimal in period 2. Hence the<br />

sets S Mi , S Ti and S Ci are affected by choices made<br />

in period 1.<br />

The period 1 value of the building under each of the<br />

strategies can be written as Equations 10a–c. The value<br />

of the building (NV 1 ) in period 1 is given in Equation<br />

10d.<br />

NV( M) = v ( Wˆ, z ) - c ( Wˆ,z)<br />

−m<br />

+ p dV ( C)<br />

1 1 1 1 1<br />

Â<br />

2s 2 2s<br />

CM<br />

s∈S 2s 2 2s 2s 2 2s<br />

 Â<br />

+ p dV ( M) + p dT (10a)<br />

MM TM<br />

s∈S s∈S 1 1 1 1<br />

NV( C) = v ( W, z ) −c( W, z ) −C(<br />

Wˆ, W)<br />

+<br />

2s 2 2s p dV ( C) +<br />

2s 2 2s<br />

p dV ( M)<br />

+<br />

 Â<br />

CC MC<br />

s∈S s∈S Â<br />

2s 2 2s<br />

p dT<br />

(10b)<br />

TC<br />

s∈S NV(T) = T 1 (10c)<br />

NV 1 = max(NV(T), NV(C), NV(M)) (10d)<br />

771<br />

Equations 9a–10c constitute a complex dynamic<br />

stochastic optimization problem. Instead of spelling out<br />

complete and general solutions to this optimization<br />

we will characterize some important properties of the<br />

solution.<br />

The paper is primarily addressing the uncertainty of<br />

future climate affects and the behaviour of owners of<br />

buildings. For this reason, the analysis in the remaining<br />

parts will focus on the choices made in the first period of<br />

the model. Period 2 choices are treated as they are<br />

nested within the period 1 choices.<br />

The maintenance strategy<br />

Previously in the paper we analysed the choice of effort<br />

(m) in the case where no conversions and no termination<br />

took place in the first two periods. The conclusion<br />

from this analysis is altered when taking into consideration<br />

that the owner knows that for some climate<br />

change scenarios she will terminate the building and<br />

for other scenarios, she will choose to rehabilitate or<br />

convert the building.<br />

The first order condition for the choice of effort in<br />

this more general case will be:<br />

1<br />

∂<br />

( − ) (<br />

1<br />

∂<br />

∂ ∂<br />

∂<br />

+<br />

−<br />

∂ ∂<br />

∂<br />

∂<br />

v c z<br />

∂<br />

z z m ∂ ∂<br />

d<br />

1 1<br />

2s<br />

2s<br />

1<br />

2 2sv<br />

c z<br />

1 1 Â p ) a<br />

2s<br />

2s<br />

1<br />

z z m<br />

= 1<br />

s∈S M<br />

s<br />

(11)<br />

As long as the sets S C and S T are non-empty, the<br />

expected marginal return on period 1 effort, for any<br />

level of m 1 , is lower than it is when conversions and<br />

‘early scrapping’ is not a part of the action set. Hence, in<br />

the presence of climate uncertainty a lower level of<br />

effort is put into the maintenance of a building.<br />

Furthermore, the probability of scrapping the building<br />

or converting it in period 2 will be higher if a given<br />

climate change scenario was anticipated than if it comes<br />

as a surprise. If the climate change scenario is not anticipated,<br />

the effort in period 1 will be chosen without<br />

taking the possibility of scrapping or conversion in order<br />

to adapt to a changed climate into consideration. Consequently,<br />

a higher level of effort is chosen, and the state<br />

of the building, measured by z, will be better than if<br />

effort is chosen according to Equation 11. This will<br />

increase the value of V 2 (M) for any s. As a result, for<br />

some s where V 2 (C) or V 2 (T) gives the maximum of<br />

(V 2 (C), V 2 (T), V 2 (M)) when m 1 is chosen according to<br />

Equation 11, V 2 (M) will give the maximum when the<br />

owner did not take climate change into consideration.<br />

In order to enhance the understanding of the consequences<br />

of choosing the M-strategy in period 1,<br />

Equation 10a is rewritten.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!