The investment thesis to support the market’s apparent disconnection from [everything], according to some, is that on the basis of a long-term forecast model, the current year and maybe next will hardly show up. The following from Matt Levine summarizes the idea:
It all comes back to the same financial principle. Regardless of the valuation methodology – earnings multiple, revenue multiple, cash flow multiple, some multiple of an accounting fabrication, etc. – these translate to, imply, or are directly underpinned by a financial forecast. This forecast yields a present value calculation in which the terminal or future value – a shortcut for what future markets will bear in perpetuity – makes up the vast majority of the result.
The issue has always been the same, only now even more so… The market thinks it’s “very forward-looking.” But it isn’t:
If we would backtest the idea 30, 50, 100 years ago on the timeline, we’d be likely to find that the subject company no longer exists, or if it does it isn’t the same company that traders bought or sold back when, or maybe it’s now a subsidiary that bears little operating semblance to the original asset.
In other words, the valuation model and all that it implies, is predicated on a forecast in which we might assume some growth or shrinkage, in which the first year or two probably don’t materially matter, as has been pointed out, but in which we are almost certain to be wrong about the nature of the asset in the future to begin with. It isn’t a question of how big or how small it will become, but what actually will it be? In perpetuity… which is a very, very long time.
The discount rate that reduces the guess back to present value, arguably, can take care of the conundrum. But not really:
First off, the terminal value in the model tends to be so disproportionately large (a multiple that magnifies the projection’s final year, by which time the operation has been typically shown to grow), that even on a discounted basis its impact on the overall is massive; and secondly, these days, the discount rate is more than likely low (ref., the risk-free T-yield curve).
Perhaps a more correct approach in what is mainly an academic exercise – because the market’s voice and judgment are what ultimately matter – is to substitute some long-dated option value for the future value in the model. Depending on the business type and profile, its fate may or may not reflect adaptation or survival, let alone leadership or following. And in a time like we’re in now, this great unknown but always possible discounts back to a highly subjective and interpretive guess.
Note *: Borrowed song lyric in the title.