<p>This dissertation examines the economic effects of alternative taxation policies on household consumption and investment decisions under uncertainty. A consideration of the alternative approaches to a theory of decision-making under uncertainty (Part 1) reveals that maximisation of expected utility is consistent with rational behaviour in a world of uncertainty. Howeve~, an alternative criterion of minimising the probability that the risky outcome will fall below some critical level (the "safety-:first" principle) may also appear as rational. Consequently, we consider both types of behavioural models.</p> <p>In a single-period framework (Part 2), the particular decision process we consider is that of an individual allocating his initial wealth between riskless assets with a secure rate of return and risky assets with a random rate of return. We then investigate the effects of taxation on risk-taking (both social and private).</p> <p>We observe that a chance-constrained portfolio choice model can be interpreted as a reasonable description of the investor's concern for safety, and that the qualitative results regarding portfolio separation (implying that optimal risky asset ratios are independent of initial wealth) and the effects of taxation (suggesting that a proportional or a lump-sum tax with full loss offsets encourages a movement towards the riskier assets) are the same both for a normal distribution of asset returns and the alternative assumption of a lognormal securities market.</p> <p>Part 2 also investigates the effects of a simple progressive tax schedule (a linear tax with a marginal tax rate which applies both above and below an exemption level) on risk~taking. Assuming only that the investor is a risk-avertor and that the risky asset is superior (or alternatively, assuming only decreasing absolute risk-aversion), we show that linearly progressive taxation of investment income (with full loss offsets) encourages the demand for the risky asset, and that exempting risky capital gains (losses) from taxation discourages (total) risktaking. We further show that a linearly progressive tax on investment income leads to greater risk-taking than a flat rate proportional tax where both of these taxes lead to equal losses of expected utility for the investor, or alternatively, where they yield the same expected revenue.</p> <p>Part 3 considers models of portfolio choice and consumption allocation in an intertemporal context. Apart from taxation of non-asset income and the case where the rate of return on riskless investment is zero, the kind of a priori restrictions we have placed on single-period preferences are no longer sufficient to determine the effects of taxation. However, given the relative nagnitudes of the income elasticities of consumption and of the risky asset demand, assumptions on the riskaversion measures allow us to determine these results. This analyssis, therefore, also indicates the kind of empirical knoftledge that is required in order to meaningfully discuss the implications of alternative taxation policies.</p> <p>We further pursue the framework of intertemporal consumption portfolio allocation to analyze the long debated issue of the differential incidence of a consumption (expenditure) tax rather than an (investment) income tax. We find that under some reasonably interpretable conditions, the differential incidence of a consumption tax is to encourage risktaking and discourage saving more effectively than an investment income tax. This result is in conflict with the general consensus in the literature that an income tax discourages saving as compared to a consumption tax. We, therefore, conclude in this context that with the introduction of uncertainty, the implications of fiscal policy are modified in an important way.</p> / Doctor of Philosophy (PhD)
Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/14078 |
Date | 06 1900 |
Creators | Ahsan, Syed M. |
Contributors | Ahmad, S., Economics |
Source Sets | McMaster University |
Detected Language | English |
Type | thesis |
Page generated in 0.0021 seconds