How firms misallocate resources after disasters

When firms expect little help from the state, they over-adapt to shocks at the cost of long-term growth.

Firm behaviour after large shocks reflects both the disruptions they now face and their expectations about institutional response. In contexts where government capacity is limited or perceived to be unreliable, firms may rationally prioritise continuity over efficiency. This is especially true in low-trust environments, where unpredictable state support forces firms to internalise risk that would otherwise be shared or mitigated through public provision.

I present new evidence on how firms adapted following the 2005 Pakistan earthquake, with a particular focus on the vulnerabilities that drove differential impact and the expectations of minimal government aid shaping firm behaviour (Rehman 2025). Using a difference-in-difference methodology on a nationally representative panel of 390 firms, my research explores the immediate disruption, short-term adaptation, and long-term resilience.

The seismic shock damaged stock and reduced sales asymmetrically between. Fragile intermediaries – firms with moderate capital-labour ratios – experienced pronounced declines in sales, while labour-intensive firms and highly capital-intensive firms demonstrated greater resilience. Firms prioritised skilled labour retention, reducing non-production roles, increasing operational hours, and diversifying across markets. There was no evidence of innovation. Building on these findings, I develop a theoretical model of adaptive misallocation: when firms expect little government support, they over-invest in short-term adaptive measures like backup infrastructure, which diverts resources from productivity, weakens public-private trust, and undermines resilience.

Adaptive misallocation occurs when firms, anticipating inadequate policy support, reallocate capital and labour toward self-provisioning – such as investing in backup infrastructure – rather than toward innovation, upgrading, or market expansion. These strategies are individually rational under conditions of policy uncertainty, but collectively inefficient, as they distort the allocation of resources across firms and sectors.

The central contribution of this work lies in linking firm-level adaptation to macroeconomic misallocation through the lens of institutional expectations. In particular, my research focuses on firms that lie between the extremes of the input distribution – those with moderate capital-labour ratios – which are shown to be especially vulnerable to shocks. These fragile intermediaries face limited substitutability across inputs and thus experience both higher exposure and lower adaptability. They bear the highest costs of adaptive misallocation, exhibiting sharper output declines and more persistent productivity losses.

By examining firm responses in the aftermath of the 2005 Pakistan earthquake, I document the heterogeneity of shock responses and model the mechanisms through which weak policy credibility exacerbates resource misallocation. My findings have broader implications for our understanding of firm dynamics under uncertainty, offering a framework to assess why recovery may fail to generate reallocation gains and how institutional credibility can influence post-shock economic trajectories.

Fragile intermediaries: The firms most at risk

The earthquake caused asymmetric firm-level disruptions. Labour-intensive firms maintained sales despite capital stock destruction by flexibly adjusting labour inputs. Capital-intensive firms, with financial reserves and durable assets, were also relatively stable. The worst hit were firms in the middle: fragile intermediaries with moderate capital-labour ratios. These firms suffered the sharpest declines in sales and struggled to adapt due to their dependence on both labour and capital, making them especially vulnerable to supply chain breakdowns.

Using difference-in-differences estimation on a panel of 390 firms, I find that firms within 400 km of the epicentre saw sales fall by 79% due to loss of capital stock. Looking at firms by capital-labour mix shows that the second quartile experienced the steepest losses in output.

The immediate economic hit: Labour and operations

Firm adaptation strategies reflected constrained optimisation. Skilled labour was prioritised for retention, while non-production roles were cut to minimise fixed costs. Firms extended working hours and diversified their geographic market reach to stabilise revenues. But these were defensive, non-innovative responses.

The probability of product innovation declined among affected firms. Rather than reallocating towards technology or upgrading production, firms relied on reallocating labour inputs and operational hours. These changes were made along the extensive margin, focused on sustaining activity rather than improving productivity.

Infrastructure damage compounded the shock. Treated firms were 11.7 percentage points more likely to rely on diesel generators. This shift toward self-provisioning was rational but inefficient, replacing shared infrastructure with firm-level solutions that offered no scale economies.

A model of adaptive misallocation

To formalise these findings, I develop a theoretical model in which firms operate under policy uncertainty. The likelihood of government support is governed by a Markov process, switching between ‘good’ and ‘bad’ states. In this setting, firms optimise by allocating resources between productive investment and adaptive strategies – such as backup infrastructure or labour redundancies.

My model features an augmented production function in which a misallocation term penalises total factor productivity (TFP) as adaptive intensity increases. Firms facing low expectations of state reliability over-invest in resilience, self-insuring against future disruptions. This is not a failure of logic – it is a response to lack of institutional support.

Importantly, fragile intermediaries suffer the most. Their limited input substitutability makes their adaptive investments costlier and more distorting. These firms become trapped in a low-productivity equilibrium, over-adapting while under-investing.

This stands in contrast to standard moral hazard models. Here, inefficiency arises not because firms expect help – but because they expect none. Adaptive misallocation reflects arises from the rational over-preparation for continued abandonment.

Breaking the cycle: Credibility, not generosity

What do these findings imply for policy? Aid alone is insufficient; what matters is whether firms can expect it. Predictable, rules-based disaster response – such as insurance mechanisms and transparent infrastructure commitments – can realign expectations and reduce the incentive to over-adapt.

Policymakers must identify and support fragile intermediaries, whose input structures make them most susceptible to distortion. The credibility of government response is not only a condition for resilience but as a prerequisite for growth.

By modelling how firms respond under institutional uncertainty, my research offers new insight into disaster economics. Firms do not respond solely to physical damage, they also adapt to perceived government failure. To avoid productivity traps, recovery strategies must focus not only on helping firms rebuild, but on restoring their trust in the state.