As I write this in the last few days of 2022, the economy of Canada and many other developed nations appear to be teetering. After a long bull market in Canadian real estate, higher rates have resulted in a rather sharp downturn. Equity prices, particularly on the growthy American side of markets, have seen large losses as well this year.
I’m sitting a suburb where just about everyone has an SUV, often BMWs or Mercedes. But this isn’t a particularly wealthy suburb. Area level statistics tell me that it’s middle class. Having spent a few years of my childhood in Canada before moving away, I remember my father telling me about wanting to buy an SUV as if it was something of a luxury. Although my memory is hazy, I recall getting a similar impression from what little I can recall about TV shows and movies.
People have lots of home equity from astronomical increases in housing prices and this, combined with low interest rates, have financed SUVs across the country. Or so the word on the street goes.
I can’t say I can realte to this way of thinking since I didn’t profit from the real estate bull market. But on a personal level, struggling equity markets have definitely made me a bit more conservative in the way I spend. I have some unrealized losses this year, but my portfolio is much larger than it was in say, mid 2021 when markets were roaring as a result of contributions I made to it.
Why do I hesitate to spend in a bad market? Well, one explanation is that the expected returns might be higher given lower valuations. I’m the kind of person who puts some stock (no pun intended) into the forecastability of cyclically adjusted price to earnings ratios. In other words, I expect markets to do better relative to 2021 and so the opportunity cost of taking a trip to Europe is higher. If I’m being honest, I came up with this explanation on the spot, so clearly that’s not why I’m more reluctant to spend.
But I’m not alone in this phenomenon. I recently read “Stock Market Returns and Consumption” by Maggio, Kermani and Mailjesi, published in 2018 (revised in 2020). It was a Distinguished Paper in the 2021 Dimensional Fund Advisors prize and looks at Swedish stock holdinsg to examine how unrealized capital gains impact household consumption.
The dataset is interesting. It’s limited in that it’s a 1999-2007 Swedish government register. But what I find fascinating is that it has data on all assets held outside of retirement accounts at the end of the tax year. As a Canadian, I can’t imagine a similar dataset existing in Canada. I don’t think the fact that it excludes retirement assets is material either since I imagine the marginal propensity to consume (MPC) due to increases in retirement wealth is very low.
Here are a few takeaways from the paper:
- The MPC of unrealized capital gains varies greatly between group in the wealth distribution. Specifically, the MPC for the lowest 50% of the wealth distribution consume about 23% of unrealized capital gains, while it is 3% for the top 30%.
- For capital gains as a whole, the MPC increases to 53% for the bottom 50% of the distribution, decreasing linearly before reaching 4% for the top 5%.
- Most households spend dividends differently from capital gains. The MPC of dividends is around 30-40% for every group except the top 5%. I find this interesting given the linear decline observed when looking at the MPC for capital gains. People just look at dividends differently, I guess.
The reason for the difference in consumption habits, though, is because of liquidity. This relates to the “buffer-stock model”. Quoting Christopher Carroll’s 1997 paper (Buffer-Stock Saving and the Life Cycle/Permanent Income Hypothesis), individuals with income uncertainty are more likely to consume wealth in the present. I find this intuitive as well; the marginal utility of consuming a dollar in income decreases greatly as household wealth increases.
However, the biggest surprise I found in the paper is that controlling for buffer-stock households (defined as those with financial wealth less than six months of their disposable income), there is basically no difference in in MPC. The sample isn’t robust for the top 10% given that there are relatively few buffer-stock households in that range. However, I did find it neat that per their results, an 80th percentile household’s MPC should be similar to say, a 30th percentile household, controlling for whether they are buffer stock households.
Maybe I’ll read more about household finance in the future. Neat stuff.
Bells and Tails 