Notes on:

Buera, F. J., Kaboski, J. P., & Townsend, R. M. (2021): From Micro to Macro Development

1 What?

Review on how macro theory and micro empirical research can complement one another to increase knowledge and to improve macro development policy.

2 Why?

Over the last 15 years, there has been a rise of empirical micro using rcts. These methods are limited to a purely empirical strategy, eschewing formal theory and macro-level policy modeling.

Meanwhile, macro models have incorporated rich micro-structure (agent problem, heterogeneity, contacting, market frictions,…). However, they still relied on strong assumptions on functional froms and distribution of unobservables, and on somewhat stylized calibration strategy. Thus, economists still consider them disconected from micro empirical research.

2.1 Efforts to scale up local experiments to macro-level

Field experiments in micro date back at least to Fisher’s work on crops (1921), which in turn are based on experiments started in the 1840s. There is early quasi-experiemental wowrk evaluating theory in algriculture of developing countries (e.g. Schultz 1964). Lucas (2011) forged experiments to assess the impact of monetary policy, with limited success.

Valid natural experiments can produce unbiased estimate of the local average treatment effects (Angrist and Imbens 1995) . Similarly, rcts can produce an internally valid average treatment effects, since randomiztion is an instrument that induces participation across the full sample. Heckman and Vytlacil (2005) introduce the marginal treatment effect (mte) – the infinitesimal counterpart of the late.

Townsend and Urzua (2009) measure the benefit of fostering entrepreneurship by two experiments (a) a lump-sum subsidy to individuals that become entrepreneurs. It required the subsidy to not use the money to finance business as to maintain the monotinicity condition of late estimation. (b) assignment of a lower cost of accessing a financial market (fixed interest rate); this intervention provides an estimate of the iit of financial participation; however, this experiment did not fulfill the conditions needed for late estimations.

3 How?

3.1 Intergrating structural models witth experimental evidence

3.2 Using structural macro models together with small-scale experimental and quasi-experimental evidence

4 And?

Quantitave Heterogeneous-Agent New Keynesian model are ideally suited to connect the dots of micro interventions and, therefore, evaluate (and devise) macro development policies.

Sufficient statistics can be ultilized for ges, but typically sufficient only for a well-defined class of models.

4.1 Case studies

In macro trade, models are built relied on exact hat algebra results and sufficient statistics

Deckle et. al. (2008)

specify a trade model for the world economy in terms of changes from the current equilibrium, avoid the need to assemble proxies for bilateral trade costs or inferring parameters of techonology

Arkolakis et. al. (2012)

show that the welfare implication of a large class of trade model depend on only the trade share and the trade elasticity, very much in the tradition of public finance

Aggregating firm-level distortions:

Baquaee and Farhi (2020)

provide approximate formulas for the aggregate impacts of distortions on tfp and output in ge. There are downsides:

• Frictions must be exogenous
• Doesn’t allow firms to entry

Sraer and Thesmar (2018)

propose a method to use experimental data to analyze the aggregate effects of policies. In particular, propose moments of the distribution of revenue-to-capital as sufficient statistics for aggregate outcomes. Downsides:

• The frictions can be endogenous, but must be homogeneous of degree one with respect to wage and ad.
• Firm-level technologies must be Cobb-Douglas.
• Require long run effects of policies.
• Doesn’t allow firms to entry

5 References