The Great Diversification and its Undoing
GAICIL Insights
Macroeconomic calm can be an accident of composition.1 The Great Diversification and its Undoing, by Vasco Carvalho and Xavier Gabaix, argues that the long swings in economic volatility since the mid-twentieth century trace back to which sectors happened to dominate output. The authors test the hypothesis that macroeconomic fluctuations are primitively the result of many microeconomic shocks, and define fundamental volatility as the volatility that would arise from an economy made entirely of idiosyncratic sectoral or firm-level shocks.
The results reframe recent economic history. Fundamental volatility accounts for the swings in macroeconomic volatility in the major world economies over the past half-century, including the great moderation and its undoing. The great moderation owes much to a decreasing share of manufacturing between 1975 and 1985. The recent rise in volatility is chiefly due to the growth of the financial sector. Fundamental volatility, the authors conclude, has a clear economic foundation, is easy to measure and explains these episodes well, supporting the view that the key to macroeconomic volatility may lie in microeconomic shocks.
The logic is portfolio theory applied to a whole economy. Spread activity across many sectors of similar size and idiosyncratic shocks offset one another. Concentrate it in a few large sectors and their shocks stop averaging out, so aggregate volatility climbs. Structural change, by shifting sector shares, shifts the economy's exposure to micro risk. Nobody votes for the great moderation; it emerges from the mix.
Now run the numbers forward into the age of artificial intelligence. If AI-intensive sectors grow fast and claim a large share of output, the economy's fundamental volatility will hinge increasingly on what happens inside them. Turbulence within AI, swings in investment, sudden changes in technological leadership, could become a major source of aggregate fluctuations. The framework offers a way to measure that exposure in real time from sector shares and sectoral volatility.
The financial-sector episode is the cautionary rehearsal. Finance raised macroeconomic volatility because a large, volatile sector came to account for more of the economy. A single sector's rapid growth changes the risk profile of the whole, even when the sector is innovative and productive. Whether AI repeats that pattern is an empirical question the paper's tools can answer as it unfolds.
None of this makes structural change the enemy. New sectors rise because they raise productivity and living standards. The right response is to understand and manage the exposure: through measurement, robust financial institutions, and policies that support diversification by lowering entry barriers elsewhere. A dynamic economy in which new sectors keep emerging spreads activity across more sectors, which in this framework lowers fundamental volatility. Openness to entry is a stabilizer.
AI could also be a diversifier. As a general-purpose technology, it could raise productivity across many sectors at once, including services that have historically grown slowly. If the gains spread broadly, output may rebalance across sectors and fundamental volatility may fall. The direction depends on a single question: does AI concentrate activity in a few producing sectors, or diffuse productivity improvements throughout the economy? The answer splits the future into two very different macroeconomic regimes.
The paper's method is its gift to policymakers. Fundamental volatility is easy to compute from sectoral data and links stability to the structure of production transparently. In a period of rapid technological change, such a tool flags when the economy is becoming more exposed to particular sectors and when diversification is providing a buffer. It is a dashboard light that can be wired up today.
History delivers a neat symmetry. The great moderation was a gift of structural change, and its undoing was a consequence of structural change too. The AI era will bring structural change on a large scale. This paper equips economists to measure how that change reshapes macroeconomic risk, and points to policies that capture the gains from new sectors while keeping the economy diversified. The next moderation, or the next upheaval, may already be visible in the sector shares.
Introduction
We investigate the hypothesis that macroeconomic fluctuations are primitively the results of many microeconomic shocks. We define fundamental volatility as the volatility that would arise from an economy made entirely of idiosyncratic sectoral or firm-level shocks. Fundamental volatility accounts for the swings in macroeconomic volatility in the major world economies in the past half-century. It accounts for the “great moderation” and its undoing. The initial great moderation is due to a decreasing share of manufacturing between 1975 and 1985. The recent rise of macroeconomic volatility is chiefly due to the growth of the financial sector. (JEL E23, E32, E44)
Takeaway
We have investigated the explanatory power of “fundamental volatility” to understand the swings in macroeconomic volatility, and found it to be quite good. Fundamental volatility explains the great moderation and its undoing. It has a clear economic foundation and has the advantage of being easy to measure. Our findings do support the view that the key to macroeconomic volatility might be found in microeconomic shocks. This is a meaningful, nontrivial empirical result: many other factors (e.g., policy, taxes, globalization) change the structure of the economy, so it is not clear a priori that the microeconomic composition of the economy would have such a large explanatory power.1
Of course, microeconomic shocks need to be enriched by some propagation mechanisms. Their identification might be simplified if we think that microeconomic shocks are the primary factors that are propagated. For instance, we do not deny that monetary shocks may be important. However, they may largely be part of the response to other shocks (e.g., real shocks caused by oil or finance).
Our findings pose the welfare consequences of the microeconomic composition of an economy. In models with financial frictions, a rise in volatility is typically welfare-reducing. Perhaps finance was too big and created too much volatility in the 2000s? Perhaps the oildependent industries were too big and created too much volatility in the 1970s?
In addition, fundamental volatility can serve as an “early warning system” to measure future volatility. In retrospect, the surge in the size of finance in the 2000s could have been used to detect a great source of new macroeconomic volatility. In any case, we think that fundamental microeconomic volatility is a useful theoretical and empirical concept to consider when thinking about the causes and consequences of aggregate fluctuations.
Notes
- Vasco Carvalho & Xavier Gabaix, The Great Diversification and its Undoing, 103 Am. Econ. Rev. 1697 (2013), https://doi.org/10.1257/aer.103.5.1697. Takeaway source: Conclusion, reproduced from a free author or repository copy (working-paper version). ↩
Cite as: Vasco Carvalho & Xavier Gabaix, The Great Diversification and its Undoing, 103 Am. Econ. Rev. 1697 (2013), https://doi.org/10.1257/aer.103.5.1697
Header image: Kelvinhall subway station, Glasgow Subway, Scotland.jpg by Ridiculopathy, CC0, via Wikimedia Commons.
