endogenous depreciation with specific capital

This is even less developed than most of the thoughts I post here, but I've been thinking recently about models of specific capital — macroeconomic models tend to deal with aggregates, so that a factory that produces cars is the same as a stable of machines that produce houses, provided the factory and the machines cost the same amount, and Hayek in particular complained about the importance of limitations on repurposing of capital. This came into my head a couple of weeks ago when my macro professor noted that downturns in the economy tend to be short and sharp, with expansions longer and more gentle, and I noted to myself that specific capital with shocks to demand would produce this pattern. What occurred to me yesterday was that, if different forms of capital have different rates of depreciation, then the aggregate depreciation rate would tend to increase with uncertainty in future demand, i.e. that if you're buying/producing a capital good when you're not sure what demand will look like in 20 years, capital that will depreciate in 20 years looks more attractive compared to capital that will depreciate in 50 years than if there isn't that uncertainty.