10.04.2013 Views

Lisø PhD Dissertation Manuscript - NTNU

Lisø PhD Dissertation Manuscript - NTNU

Lisø PhD Dissertation Manuscript - NTNU

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

768 Nordvik and <strong>Lisø</strong><br />

into the future and their profitability. In other words,<br />

immediate decisions affect the value of real options.<br />

Impacts of climate change on buildings: a<br />

two-dimensional description of the value of<br />

a building<br />

Here we analyse stylized models of the choices made<br />

by the owners under different sets of simplifying<br />

assumptions. A building is a complex asset that can be<br />

described along an almost infinite number dimensions.<br />

We simplify this into a two-dimensional description,<br />

and we proceed by assuming that the value of a building<br />

(V) depends on these two factors. The two factors are<br />

one fixed component W and one variable component, z t .<br />

r<br />

=<br />

t =1<br />

t t t<br />

V Â d v ( W , z )<br />

(1)<br />

where d t is a time dependent discounting factor and v t ()<br />

is a kind of production function.<br />

The production function is assumed to exhibit a<br />

putty-clay structure. A putty-clay production structure<br />

is a structure where the substitutability between production<br />

factors is larger before an investment takes<br />

place than they are after the investment is made. The<br />

putty-clay approach is used in general studies of investment<br />

by, among others, Johansen (1972) and Moene<br />

(1984). Here it means that once the building is completed<br />

the characteristic, W. is fixed. The production<br />

factor z t can be varied according to the technology g t ():<br />

z t = g t (W,z t–1 ,x t–1 ,m t ) (2)<br />

where x t−1 is the uni-dimensional strains (amongst other<br />

things climatic impacts) that the building experience<br />

through period t−1, accumulated up to the start of<br />

period t, and m t is the effort made to increase the value<br />

of the variable production factor at t. One can think of<br />

this effort as maintenance.<br />

The production function g t () plays a crucial role in the<br />

analysis of the choices made by building owners as a<br />

dynamic link between the efforts of any period and the<br />

future performance and need for effort/maintenance of<br />

the building. The costs of operating the building (c t ())<br />

depend on the state of the building and the strains that<br />

the building is exposed to during a given period (e.g.<br />

impacts of different climatic parameters on the everyday<br />

operation of the building, including energy use):<br />

C t = c t (W,z t ,x t ) (3)<br />

The strains a building is exposed to are stochastic. The<br />

outcome of the stochastic process is assumed to be<br />

multinomial distributed over a finite set of outcomes.<br />

The outcomes are uncorrelated over periods. The<br />

probabilities of each of the states that produce outcomes<br />

can change over time. Throughout the rest of the paper,<br />

we will think about the decision maker, who in this<br />

paper is the owner, as being risk neutral. Hence, it is<br />

assumed that the owner maximizes the expected net<br />

present value NV of the building:<br />

T<br />

S<br />

ts t t t t t ts t<br />

NV = ÂÂ p d { v ( W, z ) −c( W , z , x ) −m}<br />

(4)<br />

t = 1 s=<br />

1<br />

In order to enhance the analytical tractability of the<br />

model and to focus on the effects of climate change we<br />

will abstract away the stochastic climate parameters<br />

given different climate change scenarios. The term ‘a<br />

climate change scenario’ here refers to a state, and we<br />

do not allow for any stochastic within each of the states.<br />

Instead of starting out with a very general solution to<br />

the maximization, we begin with some simple cases.<br />

This is done because it enhances the intuition of the<br />

authors and hopefully also the readers. The solutions to<br />

these simple problems will also serve as benchmarks for<br />

the results from analyses of more complex situations.<br />

One particular simplification is that we analyse the<br />

choices of the owner in a three-period setting. The<br />

choice set of the third period consists of only one<br />

element, which we term termination. Throughout the<br />

analysis, we will assume that no action is taken at the<br />

start of period 3 to prepare for the termination. At<br />

the start of the first two periods, the owner first observes<br />

the state of the building, then she chooses her action.<br />

The different models presented will differ in what types<br />

of actions that are contained in the set of possible<br />

actions. After the action is chosen the ‘strain-stochastic’<br />

is realized.<br />

Building maintenance and the risks of future<br />

climate change<br />

Starting in period 1, strains of the preceding period, and<br />

consequently the present state of the building, are<br />

observed, and effort is chosen. In period 2, the owner<br />

observes the strains and chooses an effort. Within this<br />

model the solution to the maximization problem of the<br />

owner will consist of a period 1 effort, and a set of efforts<br />

for each state in period 2: (m 1 , m 21 , m 22 , m 23 , ... , m 2S ).<br />

For simplicity, the termination value in period 3 is<br />

treated as non-stochastic.<br />

1 1 1 1 1 2s 2 2s 2<br />

NV = v ( W, z ) −c( W, z ) − m + p d { v ( W,<br />

z )<br />

2s2 2<br />

−c ( W,<br />

z ) − m +<br />

3 3<br />

dT<br />

S<br />

Â<br />

s=<br />

1<br />

(5)<br />

This way of formulating the problem allows us to<br />

analyse it using traditional tools of static optimization.<br />

Note also that the formulation allows for both the<br />

value of the services produced by the building and the

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

Saved successfully!

Ooh no, something went wrong!