Years ago, when Charles attempted to retire, my mother promptly decided that retirement was the “worst of both worlds”: more husband and less money. She quickly put the kibosh on the idea. The US government is a little more complicated to run than the Gave household, yet health care policy also offers a “worst of both worlds” outcome—profligate spending and rising bureaucracy (from government interference) and eye-watering prices (from the protected private sector). Not that investors in health care stocks seem to care: yesterday they cheered Republican senators’ attempt to water down the House’s “repeal and replace” Obamacare bill. In an otherwise lackluster session for US equities, the S&P health care index rose 1.05%, bringing month-to-date gains to 7.4%, and 17% year-to-date.

Clearly, the continued drop in energy prices—and consequent flattening of yield curves—has meant that it has been a great year to be a “growth at any price” investor (see Figuring Out Where the Ball Will Be). This effect has only been compounded by technology names’ recent outperformance. And yet despite the hype about FAANGS, and how tech is changing the world, health care stocks have so far this year outperformed all other sectors (on an equally-weighted basis).

Looking at yesterday’s performance (and the left hand chart overleaf) it is easy to conclude that health care is the one “Trump trade” that has worked. Indeed, prior to the November election, the general view was that a Hillary Clinton win would bring a new wave of price controls and regulation for the sector. But after this month’s break-out, most investors face the challenging question of whether to jump on the health care bandwagon, or alternatively let that particular train leave the station.

Unfortunately, valuations offer an inconclusive picture. Global health care stocks are trading pretty much in the middle of their historical valuation range, although arguably that is a better relative position than most other growth stocks. Moreover, given the strong momentum currently pushing health care stocks higher, it could be argued that—absent a big change in the macro environment—valuations should continue to move out toward the highs witnessed in mid-2015.

So what change in the macro environment could materialize?

The first obvious threat to all high P/E stocks, including health care, is higher interest rates. However, given the continued collapse in commodity prices, a surge in long rates seems unlikely in the near future. The second threat is an increase in government regulation; hence the selloff before last November’s US election and then yesterday’s rally.

Still, it is hard to not feel at least a little queasy. To quote Herbert Stein, one-time economic adviser to Richard Nixon and Gerald Ford: “If something cannot go on forever, it will stop.” So how long will medical costs be able to grow at more than twice the pace of nominal GDP? Can health care spending exceed 20% of US GDP and 40% of government spending by 2025 and still grow (as currently projected)? The table overleaf comes from a terrific World Bank database and clearly makes the point that one country is a massive spending outlier—the same country where health care stocks have lately been on such a rip!

By combining the worst of both the private and public sector, US health care policy has delivered huge returns for the US health care industry. But can these returns grow from here without triggering a political backlash? Clearly, the Republican Party is busy chasing its own tail, while the Democratic Party seems primed to engage in circular firing squad activity, thereby leaving health care stocks free to expand margins. Do we not ultimately reach a certain limit, at least in the US, constraining how much an economy can spend on health care?

Projecting health care spending into the not-too-distant future makes for scary looking charts. In fact, it can be argued that the imposition of Obamacare stemmed from such charts. President Obama probably looked at the forecast rise in Medicare/Medicaid spending, and other entitlements and thought, “we need to offload this liability!” After all, what was Obamacare, but one big transfer of liability from the government to individuals (who were forced to buy health care insurance).

One effect of this liability transfer was to impose a big hit on the average US consumer’s disposable income. Having to pay for rapidly rising health care costs is crimping spending elsewhere—which, along with the “Amazon effect”—may explain why retail stores are closing at a faster clip today than they did during the 2008-09 crisis. In other words, wouldn’t a rise in US health care spending from current levels carry the seeds of its own demise? If consumers end up having to spend even more on health care, then it is hard to see how the current death spiral in retail, in mall values, and in other struggling assets finds a floor.

That said, it would be reductive only to consider health care stocks with US exposure. Achieving more profit growth in the US may prove politically difficult, but health care is a strong global growth industry. Take China, where in less than 20 years, annual health spending has risen from US$21 per capita to US$420. Or India or Indonesia, where the average citizen spends less on health care each year than the price of a pair of Nike shoes.

Clearly these numbers can only go up. This is in part because as incomes rise, spending on health care tends to soar. But also as countries get richer, diets change with new foodstuffs being consumed (more milk, red meat, sugar, and preservatives). This diet-shift causes individuals to pack on pounds, resulting in far more diabetes. Already, some 415mn people suffer from diabetes globally, projected to rise to 642mn by 2040. Most of the growth will come from emerging markets!

So as health care stocks continue to break out, it probably makes sense to parcel out what is sustainable from what isn’t? I have a hard time believing that US health care providers will be able to extract an ever expanding share of the US economy. Focusing on health care names that allow providers to cut costs (through the likes of robotics, software and outsourcing) or those with a solid footprint in emerging markets makes better longer term sense.

This article originally appeared at Gavekal Research.