risk and uncertainty in decision making Flashcards
(33 cards)
what are the attitudes to risk?
- risk neutral - ignore possible variations in outcome and concerned only with the most likely. if a venture has a 60% chance of earning $100, and a 40% chance of losing $20, then the expected value is $52 and a risk neutral investor would be willing to pay $52 to make that investment.
- risk averse - prefer alternative with least variation associated with it. If a venture has a 60% chance of earning $100, and a 40% chance of losing $20, then the expected value is $52 and a risk averse investor would pay less than $52 for that investment, in order to compensate for the risk.
- risk seeker - only concerned with best possible outcome regardless of likelihood of occurrence
how is utility an important aspect of risk/uncertainty?
asis of theory is that individual’s attitude to certain risk profiles will depend on the amount of money involved. for example, most people would accept a bet on toss of a coin if outcome were that they would win $6 if it came heads and pay $4 if it was tails. if stakes were then $6000 and $4000, average person might think twice and reject the bet. utility theory attaches weights to sums of money involved - tailor made to individual’s attitude towards winning and losing certain sums of money.
what is an exact vs empirical probability?
this is an exact probability, involving full list of all outcomes and counting exact number of outcomes that constitute the event. rare that you can find all the outcomes in practice.
taking a sample of components rather than whole population = empirical probability. essentially approximations to the true but unobtainable exact probabilities.
what conditions must be met for an approximation on probability to be valid?
sample must be representative
larger samples are more accurate and preferable, considering time and money available.
how is probability found from an unsuitable sample?
when a suitable sample is not available, an empirical probability cannot be found. so instead a subjective probability can be estimated based on judgement and experience. not entirely reliable but can be useful.
what are the addition rules of probability?
P(A U B) = P(A) + P(B) − P(A ∩ B)
when A and B can both be true
P(A U B) = P(A) + P(B)
when only A or B can be true at one time.
what are the basic multiplicative rules of probability?
P(A ∩ B) = P(A) × P(B|A)
generally
P(A ∩ B) = P(A) × P(B)
if A and B are statistically independent.
what is important to remember when applying empirical probabilities?
when applying empirical probabilities you are assuming those having prob applied are ‘typical’. if you know they are not then probability should be revised.
when is an EV of an investment NOT helpful?
if there’s a 50/50 chance of success/failure, EVs aren’t very helpful as you will literally never get the EV - you’ll either get the lower or higher outcome. conclusions will depend on utility the investor places on the amounts involved. unlikely that many will risk the loss of failure if the cash outflow is high.
define discrete probability distribution
can only assume certain values regardless of level of precision to which it is measured. e.g. errors made will always be 1,2,3 never 2.7. distributions are fairly easy to construct
what is the EV criterion?
just pick project with highest EV, however consider whether you can afford the potential loss and adjust for this.
what are the limitations of EV criterion?
uses estimated data - analysis can never be more reliable than the estimations upon which it is based
data is often simplified for use - very limited distributions used when more complex ones or even continuous distributions may be better.
not relevant for one-off decisions - analysis of EV gives best average profit in long run where there are many repeats, so use is questionable for one off.
takes no account of attitude to risk
what is important to consider when using decision trees?
only as good as information it contains. main difficulty is obtaining accurate predictions of probabilities that undermine the uncertainty.
what are the shape conventions in decision tree analysis?
squares are decisions, circles are probabilities
what is the value of perfect information?
One feature of uncertainty in business situations is that it is often possible to reduce it or even eliminate it. For example, market research can be used to determine, with a reasonable degree of accuracy, what the demand for a new product will be. However, market research costs money and the decision maker has to decide whether it is worth paying for the research in order to reduce or eliminate the uncertainty of product demand.
value of perfect info = expected profit with info - expected profit without
what are some benefits of decision trees?
force the decision maker to consider the logical sequence of events. complex probs broken down into easier to handle sections. financial outcomes and probs shown separately. make sure only relevant costs are included!!
what should be considered when dealing with decision tree type problems?
TVOM - should be inc in calculations if lasting more than a year. options like selling a patent bring immediate income whereas keeping the patent and developing the product is spread over time.
assumes risk neutrality - some decision makers don’t choose options which give best EV because of risk attitude.
sensitivity analysis - depends very much on values of probabilities in tree. can use probability as a variable too, e.g. at what probability of x occurring would the decision change?
oversimplification - to make tree mgeable, situation often simplified. makes it appear far more discrete than really is. much more likely in practice that outcomes would form near continuous range of inflows and outflows. cannot show on a decision tree, so usually simplified.
what is std dev? how can it be helpful?
std dev = dispersion of a probability, so effectively measures risk associated with range of outcomes. two probability distributions with different expected values don’t have directly comparable std deviations. can overcome this with coefficients of variation, which is std dev / EV. can then compare risk directly.
what is the maximin criterion?
The maximin criterion suggests that a decision maker should select the alternative that offers the least unattractive worst outcome. The decision maker assumes that the worst possible outcome will always occur and therefore he/she should select the largest payoff under these circumstances. This means that the decision maker would choose the alternative that maximises the minimum profit.
what is the maximax criterion?
On the other hand, the maximax criterion looks at the best possible results and therefore the decision maker assumes that the best payoff will occur, i.e. maximise the maximum profit.
what is the regret criterion?
Once a decision maker selects a criteria that turns out not to be the best he/she will regret the selection of not having chosen another alternative when he/she had the opportunity. This is known as the ‘regret criterion’.
what are the pros and cons of EV?
pros - takes risk into account, info reduced to single number so easier decisions, relatively simple
cons - subjective probabilities usually used, EV is merely a weighted average and has little meaning for one off projects, gives risk neutral decision, may not correspond to any actual possible outcomes.
utility theory can help with the last point.
what is VaR?
measure of how market value of an asset or portfolio of assets is likely to decrease over certain time - the holding period - under normal market conditions. usually this is 1-10 days.
how is VaR measured?
measured using normal distribution theory. typically used by investment banks to measure mkt risk of their asset portfolios. amount of risk to be lost from investment under usual conditions over given holding period, at a particular confidence level. usually 95 or 99%.
