Human Capital and Development Accounting Revisited. I quantify the effects on development accounting of allowing for imperfectly substitutable labor services.  To estimate the degree of substitutability between skilled and unskilled labor services in a cross-country setting, it is sufficient to estimate the relative price of skilled labor services, and I develop a novel method for estimating this relative price using international trade data. My method exploits the negative relationship between relative prices of skilled labor services and relative export values in skill-intensive industries. I find an approximately constant elasticity of substitution with a value of about 1.3. When integrating my results into a development accounting exercise, I find that efficiency differences in skilled labor are more important than uniform efficiency differences in explaining world income differences. Under the traditional development accounting assumption of neutral technology differences, the skilled labor efficiency differences reflect human capital quality differences, and human capital differences can explain a majority of world income differences. Relaxing the assumption of neutral technology differences, an alternative explanation is that there are large skill-biased technology differences between rich and poor countries.

Price Level Determination When Tax Payments Are Required in Money. We formalize the idea that the price level can be determined by a requirement that taxes be paid in money. We show that if households have to pay a money tax of a fixed real value and the money supply is constant, there is a unique stationary price level, and a continuum of non-stationary deflationary equilibria. The non-stationary equilibria can be excluded if we introduce an arbitrarily lax borrowing constraint. Thus, in the basic model, tax requirements can uniquely determine the price level. When money has liquidity value, tax requirements can exclude self-fulfilling hyperinflations.

Swedish Unemployment Dynamics. We decompose the sources of unemployment variations into contributions from variations in different labor market flows. We develop a decomposition method that allows for a distinction between permanent and temporary employment and slow convergence to the steady state, and we apply the method to the Swedish labor market for the period 1987-2012. Variations in unemployment are driven to an approximately equal degree by variations in (i) flows from unemployment to employment, (ii) flows from employment to unemployment, and (iii) flows in and out of the labor force. Flows involving temporary contracts account for 44% of unemployment variation, even though temporary workers only constitute 13% of the working-age population. Neglecting out-of-steady-state dynamics leads to an overestimation of the importance of flows involving permanent contracts.

Supply Chain Risk and the Pattern of Trade. This paper analyzes the interaction of supply chain risk and trade patterns. We construct a model where an industry's risk sensitivity is determined by the number of customized components that it uses, and countries with a low supply chain risk specialize in risk-sensitive goods. Based on our theory, we construct an empirical measure of risk sensitivity from input-output tables and customization measures. Using industry-level trade data and a variety of risk proxies, we show that countries with a low supply chain risk disproportionately export risk-sensitive goods.

Random choice theory has traditionally modeled choices over a -nite number of options. This thesis generalizes the literature by studyingthe limiting behavior of choice models as the number of optionsapproach a continuum.The thesis uses the theory of random elds, extreme value theoryand point processes to calculate this limiting behavior. For a numberof distributional assumptions, we can give analytic expressions forthe limiting probability distribution of the characteristics of the bestchoice. In addition, we also outline a straightforward extension to ourtheory which would signicantly relax the distributional assumptionsneeded to derive analytical results.Some examples from commuting research are discussed to illustratepotential applications of the theory.

This chapter deals with probabilistic choice when the number of options is infinite. The choice space is a compact set S⊆R k   and we model choice over S  as a limit of choices over triangular sequences {x n1 ,…,x nn }⊆S  as n→∞  . We employ the theory of random sup measures and show that in the limit when n→∞  , people behave as though they are maximising over a random sup measure. Thus, our results complement Resnick and Roy’s [18] theory of probabilistic choice over infinite sets. They define choice as a maximisation over a stochastic process on S  with upper semi-continuous (usc) paths. This connects to our model as their random usc function can be defined as a sup-derivative of a random sup measure, and their maximisation problem can be transformed into a maximisation problem over this random sup measure. One difference remains though: with our model the limiting random sup measures are independently scattered, without usc paths. A benefit of our model is that we provide a way of connecting the stochastic process in their model with finite case distributional assumptions, which are easier to interpret. In particular, when choices are valued additively with one deterministic and one random part, we explore the importance of the tail behaviour of the random part, and show that the exponential distribution is an important boundary case between heavy-tailed and light-tailed distributions.