Category Archives: Micro




Whilst based on the course readings this summary is taken from Varian as the article itself is somewhat complex and doesn’t seem too relevant.


  • Asymmetric information may drive out high quality goods.
  • The seller knows the quality of his product but not the buyer. Lemons refers to the used car market whereby a “lemon” is a used car that is of poor quality and a “plum” is a used car of good quality. Both are offered to the market simultaneously.





Buyer will pay



Seller will sell




  • The quality of the car is not discernable to the buyer although the seller knows what type of car he is selling, hence the asymmetry of information.


  • Assuming buyers are risk neutral they will pay an amount equivalent to the expected utility of the car they receive which in the case of someone risk neutral, should be equal to the expected value of the car (i.e. the value of the average type of car available on the market). If there are 1000 lemons and 1000 plums then the expected value is:


E(U) = 0.5 * U(lemon) + 0.5 * U(plum) = (0.5 * $1000) + (0.5 * $2000) = $1500


How many cars are of what type are sold will depend upon the market price:

P < $750 No cars sold
$750 < P < $1250 Only lemons sold
$1250 < P All cars sold


  • Notice that there is no price at which only plums are sold. This indicates that if there is a market for goods of varying quality and consumers cannot distinguish, then all goods sell at the value of the best quality product.


Consider the following structure:




Buyer will pay



Seller will sell




E(U) = 0.5($2400) + 0.5($1200) = $1800


  • But at this price only the lemons will sell as the value of the average car on the market is below the reservation price of those sellers of plums. As buyers are certain that at this price they will receive a lemon, no plums will sell even though the buyers will pay $400 above their asking price. This has potential to destroy the whole market for used cars as buyers are only willing to pay $1200 for a lemon. This is market failure.
  • Lemon owners put a negative externality on plum owners as they affect market perceptions about the average quality of used cars on the market.




P. Moser

The American Economic Review, Vol. 95, No.4 (Sept., 2005)

A Summary

Principal Research Question Is the direction of innovation determined by patent laws?
Theory A patent offers the inventor exclusive rights to benefits from their invention for a limited period of time. This essentially creates a monopoly for the purpose of encouraging innovation. In the absence of such patents it is unlikely that individuals would invest much in R&D. The longer the patent, the more gains accrue to the inventor, but the larger is the deadweight loss.

Monopolies decide simultaneously on price and output whereas the competitive firm has no control over price. Thus, whilst the competitive industry produces at MR = MC, the monopolist operates where price if greater than marginal cost with a lower level of output. Consumers are typically worse off in such a situation, and we could have a Pareto improvement if the monopolist would supply at competitive prices and quantities. This means there is deadweight loss. This is the price paid for encouraging innovation.

Motivation If innovative activity is motivated by expected profits and if the effectiveness of patent protection varies across industries then countries with no patent protection should specialize in industries which exhibit strong alternative means of protection. Patents may therefore determine the direction of innovation, and this means that the introduction of patents in a country that does not currently have them could incite a change in comparative advantage, as patent protection expands the set of industries in which individuals which to innovate. Any changes in innovative industry could be relevant for determining economic growth.

The relevant alternative means of protection would be secrecy, beating competitors to market, or keeping strict control over assets complimentary to the production/marketing process. Secrecy is probably the most effective.

Data Data from world fair exhibitions. Exhibition data measure innovation whether they are patented or not whereas as patent data only counts those innovation that the inventor decided to patent. Thus exhibition data should be superior.
Method Focuses on data from exhibitors from Northern Europe as there should be small differences only in unobserved characteristics. Denmark and Switzerland did not have patents, and the Netherlands had abolished them when on a free trade drive.
  • Comparisons of US/UK patenting rates are very reveal similar patenting behaviour despite the fact that there were big differences in the laws, and in the cost (much more expensive in UK). Innovators chose to patent in the same broad industry categories, especially machinery. There was a much smaller rate of patenting for scientific instruments, food processing, chemicals etc.
  • Differences in the distribution of innovations across industries varies with length of patent, with increase in the length of short patents being much more significant.
  • Patentless countries show a strong focus on a narrow set of industries particularly scientific instruments.
  • Patentless countries focused on scientific instruments and food processing and less on machinery. Swiss innovators focus on industries where patents are not generally applied for in countries that have strong patent protection.
  • Mining deposits seem to have a strong effect on the level of innovation in mining equipment.
  • Patent laws have strong effect on choice of industry even when GDP, country, education are controlled for, and Switzerland is dropped from the specification.
  • Education has an effect on the level of high tech industry in the country.
  • Size of population seems to have an effect as large economies allow for large scale industry to be profitable.
Interpretation Potential innovators choose simultaneously between industries and their choice may seems to be influenced by patent laws as well as other characteristics of their work environment (e.g. availability of inputs etc.).

Innovators in countries without patent laws innovated where process secrecy was possible. Patent laws thus help to determine how innovators respond to demand from various sectors for innovations and so they also help to determine the supply of knowledge in an economy.

The introduction of patent laws in LDCs may slow rather than aide growth is they are led to compete more directly with innovations from the developed world. Strong patent laws only assist when they encourage new technologies different from those in the developed world.

Robustness Patent laws could be endogenous to the level of innovation. Tries to solve this by constructing a synthetic Switzerland using matching techniques.

The Netherlands abandoned its patent laws. There are not sufficient data points to do country fixed effects, but this change in NL is like a natural experiment. The proportion of Dutch innovations in food processing went from 11 to 37% surpassing textiles as the most prominent sector.

Problems Countries attending the fairs would often compete to demonstrate their technical supremacy in certain industries, so the data may not be an accurate reflection of innovation in any given country. 

Space restrictions, transport costs, and inventions that were easy to copy could mean that certain inventions were underrepresented. However, these problem are somewhat mitigated by the organizers being flexible on space, and the display of blueprints and models to deal with transport and secrecy issues. 



E. P. Lazear

The American Economic Review, Vol 90, No.5 (Dec., 2000)

 A Summary

Principal Research Question Does performance related pay increase worker productivity? They examine this using data from an auto-glass repair/replacement company.
Theory Workers respond to incentives and specifically that paying on the basis of output will induce workers to supply more output.

An hourly wage implies that effective labour is observable i.e. the amount of time spent doing a job is a perfect measure of the amount of labour supplied. However, this is not necessarily the case as effort is not easily observable and thus hours worked could be a poor signal.

An hourly wage also implies that technology is fixed. i.e. inputs and outputs are combined at a fixed rate to create output. However, this is not the case when one of the inputs is human work, as effort can vary.

Motivation The study wants to test theory to see if moving to performance pay, of piece rate pay improves productivity. It is not obvious: if a minimum output level is set which exceeds that which would be chosen voluntarily under piece rate, then output will increase. Similarly, the level chosen under hourly wages may be very high, meaning that if piece rates are introduced with lower minimums, more heterogeneity may be tolerated and this could cause a fall in productivity. 
Data Safelite is the largest auto-glass installer in the US. A new management team instituted a piece rate pay scheme rolled out over time whereby workers were guaranteed $11 p.h. minimum, but they would earn extra based on output. Those who wanted to could thus earn more.
Method OLS with Piece Rate dummy that = 1 when the piece rate scheme was introduced.
Results 44% productivity gain when moved to the piece rate.

Worker dummies are included and this drops to 22%. This represents the pure incentive effect of the scheme holding constant the individual characteristics of the worker. This means the average worker is working 22% harder. The rest of the increase comes from “sorting” i.e. attracting better quality employees. They test the new employees that came on board after the pay scheme, as opposed to those who started before and found indeed that there productivity was significantly higher. The pay scheme also seems to increase tenure, thus improving learning, and hence productivity.

There was an average 10% gain in pay, so the increased productivity was shared with the workers. It is not certain that profits rose, but it seems likely given 44% increase in productivity and only 10% increase in pay.

Variance of output increases.

Interpretation The implication is a switch to performance related pay should increase average levels of output and its variance. This does not mean that profits must rise, as firms choose a compensation scheme based on the costs and benfits of each such scheme. The benfit is a productivity gain, but the costs are monitoring and compliance.

The minimum level of ability of worker does not change, but more able workers are attracted by the pipece rates.

Problems External validity – Piecework requires exactly measurement and this includes costly peoples and machines and types of work conducive to suh monitoring. For many the measurement costs would exceed the benefits. Quality also needs to be capable of monitoring. Managerial/Professional jobs particularly problematic.

Quality – quality may have declined. In this specific case they had a very clear way of monitoring quality, and peer pressure meant that there were large incentives to do good work.



D. Card & A. B. Krueger

The American Economic Review, Vol 84, No.4 (Sept., 1994)

A Summary


Principal Research Question Do rises in the minimum wage reduce employment?
Theory In a competitive labour market, if the government sets a minimum wage that is higher than the equilibrium wage the supply of labour will exceed demand at the higher minimum wage. The increased supply will be complimented by contracted demand and therefore employment levels will fall.
Motivation To be right, and everyone else wrong!!!The rise of minimum wage from $4.25 to $5.05 in NJ provided a good quasi-experiment in that they could do a difference in difference estimation using PN as a control group. They are geographically proximate, and starting wages, meal price, and employment indicators were substantially similar in the pre-treatment survey responses. Additionally, employment in restaurants in the band that paid starting salaries above the (new) minimum wage in both periods fell by an equivalent amount in both NJ and PN. This is significant, as they were paying above the equilibrium price already and so should not have been affected by the wage hike in NJ. Thus they were affected by general economic conditions which appear to have been the same in both states. The distribution of starting wages in the stores was very similar before.

Additionally, as both states were in a recessionary environment with rising unemployment, it is doubtful that any rise in employment found after the wage hike could be attributable to general economic conditions.

Data Data on Fast Food restaurants – they are leading employer of low-wage workers; they comply with regulations; the job requirements are homogenous across restaurants so particular characteristics not an issue; no tips means income levels are easily measurable.They only consider the big chains, 410 of them, in two rounds of survey.
Method Difference in difference with PN as control.
Results Full time employment increased in NJ after the rise. Employment expanded most at low wage stores, and contracted at high wage (those already paying above the minimum).
  • Set employment at 0 for temporarily closed stores.
  • Exclude 35 stores on jersey shore
  • Redefined full time work
  • Exclude stores they called more than twice  – none had major effect
  • They test subgroups to see if demand shocks are making up the rise in employment but they find it is not.
  • They test opening hours, number of cash registers with no significant effect.
  • They see no evidence that non-wage compensation decreased, indeed in NJ the amount of free and subsidized meals given to employees actually rose.
  • They look at macro employment data which showed that NJ employment actually was worse in the period than elsewhere in the US, but teenage employment did better.


Interpretation The results are not compatible with standard theory. This indicates that although firms are price takers in the product market, they have some power in the factor market. If they are facing an upward sloping labour supply curve a rise in wage can mean a rise in emplyment.Ambiguous as to whether the age increase increased full time employment. It would potentially do so because employers want to substitute low wage earners for full-time employees who are probably older and more skilled. Additionally they may be more productive as they have more time to learn by doing.

There is mixed evidence that the wage increase was passed on to customers through product prices. Prices did rise around 3% which would cover the wage increase, but it did so even at stores already paying more than the minimum wage, and they did not rise faster at the stores most affected by the minimum wage. This could be because the market is very competitive.

Problems Reliability of the data is only 0.7 for employment (based on accidentally doing the same interviews twice).They only consider the big chains, who are most likely to be able to pass on prices, and otherwise squeeze suppliers, engage in advertising to boost sales etc. The story could be different at small, independent outfits.

There are external validity issues in other words.  



S.Subramanian & A. Deaton

Journal of Political Economy Vol. 104, No.1 (Feb., 1996) pp.133-162


Principal Research Question How elastic is caloric intake with respect to expenditure?
Theory Elasticity measures the responsiveness of quantities demanded to price changes. It is the % change in quantity divided by the % change in priceIt is calculated as   ΔQd/Q ÷ ΔP/P

In this case we are looking at how caloric intake rises with total household expenditure so     ΔCalories/Calories  ÷  ΔExpenditure/Expenditure

If elasticity >|1| then demand is elastic

If elasticity <|1| then demand is inelastic

Motivation Bouis & Haddad, together with others have claimed that the elasticity of caloric intake with respect to income is close to zero. This goes against the idea that nutrition responds to income and that economic policies that are good for growth do not imply an elimination of hunger. It questions whether real income is a good proxy for thinking about welfare. The promotion of real income would therefore not be conducive to development.
  • National Sample Survey for rural households in Maharashtra, western India.
  • 5,630 households, 10 in each of 563 villages
  • Report expenditure on over 300 items including 149 food items.
  • No income data collected, so total household expenditure is used as the welfare measure.


Method They regress total available calories on the number of meals given to guests, employees and those taken at home to find out how many calories are contained in each type of meal. They then subtract/add those meals given away/received to the total calories available, to create an adjusted figure. If they did not do this the elasticity for richer households would be grossly overstated as they have a large number of available calories as they give away many more meals to employees/guests. When the data are tabulated it becomes clear that the poor spend a lot less money per 1000 calories than the rich. This is because coarse cereals provide a much larger share of caloric intake. As people get richer people substitute between food groups away from cereals toward dairy and meat products etc. which have many fewer calories. Thus although the total food elasticity if 0.772, the price of calories elasticity of 0.32 drives a wedge between the food and calories elasticites. This is due to substitution.

They regress log calorie intake on log expenditure. Non-linearity would be a problem as it is possible that those with insufficient food would have much larger elasticities than those with more income. So they plot the regression line using a grid of 100 different data points. Although it is somewhat steeper at lower expenditures it is fairly close to a straight line.

This method is not appropriate if controlling for other variables, so they also run an OLS regression with additional variables (no. in household etc.)

Results There is no evidence that elasticities are close to zero unlike in Strauss & Thomas’s findings for Brazil. This could be because for all households surveyed wealth is such that calories are still an issue whereas in brazil they were not.Elasticity declines from 0.65 to 0.4 over the range of incomes from lowest to highest.

Once household composition is controlled for other variables are not significant in the OLS regression. Both calorie consumption and price elasticities are around 0.35.

Robustness Various controls added. They check for non-linearity using the specification noted above.
Interpretation Income does constrain caloric consumption but not by very much, at least within the range of household incomes surveyed here. Poor households tend to purchase cheaper calories but also fewer calories overall.
Problems Endogeneity: if hunger caused poverty as well as poverty causing hunger there would be reverse causality problems. The argument is that lack of hunger reduces productivity and thus wage earning capability which prevents calories from  being purchased. Hunger thus creates a “poverty trap”. They cannot rule this out by using e.g. IV regression, but they claim that as the 600 extra calories that are needed to sustain physical work, could be purchased for 4% of the daily wage, that the barriers to sufficient nutrition are not high enough to create a poverty trap.