Does housing wealth boost start-ups?

I don’t mainly mean tech start-ups but rather the broader concept:

…house price effects work through wealth, liquidity and collateral effects on the propensity to start new firms and expand young ones. Aggregating local effects to the national level, housing market ups and downs play a major role – as transmission channel and driving force – in medium-run fluctuations in young-firm employment shares in recent decades. The great housing bust after 2006 largely drove the cyclical collapse of young-firm activity during the Great Recession, reinforced by a contraction in bank loan supply. As we also show, when the young-firm activity share falls (rises), local employment shifts strongly away from (towards) younger and less-educated workers.

That is from a new paper by Steven J. Davis and John C. Haltiwanger.


"reinforced by a contraction in bank loan supply"

I am interested to see how they demonstrate causality. Did the housing crash cause bank to stop lending which reduced start ups?

Clearly in free lunch economics, lack of consumers with mooney they are willing to spend has nothing to do with banks willingness to loan to businesses to startup or expand production into markets where they will fail to find buyers.

It was totally the 19% interest rates Volcker forced the Fed to impose.

Note, a very common source of finance for small businesses is credit cards that were at something like 23% in the early 80s when mortgages were over 10-15%, but today are 17%, 19%, and up, when mortgaage rates were about 4%.

Starting a business is done with personal loans. Today the biggest source of personal loans is credit cards, and credit caards arefar more easily obtained with ten times the credit limits of the 80s when they were being flogged to people without jobs, eg teenagers, college students.

Given housing was not more valuable, simply the price was simply higher. A house that increased in price from $200,000 to $500,000 does not house two families of 4 instead of one family of 3, it simply has an inflated price. Like a pouund off ground beef at $4 today does not produce 16 times the number of burgers than the 25 cent pound of beef in 1968.

The Free Lunch economics premise that price defines value is pretty bogus.

The Free Lunch premise that supply defines demand, not consumers with money they are willing to spend, is bogus.

So you don't have an answer. Got it.

The mulp does not answer. The mulp does not bow. Do not approach the mulp without averting your eyes. Do not engage the mulp in conversation as you will anger the mulp. The mulp abides.

There are some economists who believe that disruption encourages risk taking, while other economists believe the opposite. For better or worse, we have chosen rising housing prices (and asset prices generally) as integral to risk taking, investment, and economic growth. I'd prefer less reliance on housing and housing prices, but that's the economy we have not the economy we might prefer.

How could it not? If I were going to start a small business, the collateral for investors/creditors would almost certainly include my house.

I have not read the paper itself, so maybe they really have dealt with this in terms of Grainger causality or other methods using proper lags, but the Great Recession came on not too long after 2006, during which both general aggregated demand fell along with local business conditions and housing prices in most places, although the latter went down earlier. It strikes me as being hard to really robustly tease these effects out. In any case, if the effect is clearly there, aside from providing collateral, to the extent that most startups initially focus on their local community for their initial market, the effect may be coming more through expected demand than financial supply, with falling housing prices pretty clearly a bad signal for future local demand.

As some of the other comments have noted, given how highly most variables are correlated across time, can they really identify the causal influence of housing prices?

I've glanced at the body of the paper, they do have reasonably disaggregated (MSA level) panel data so they can compare the differing local fluctuations that different areas experienced. And they throw in some additional variables such as unemployment so that housing prices are not relied upon as the sole driver of the dependent variable.

Is it enough to solve the identification problem? I'd have to look at the paper and data more closely; it looks like they made a good effort at least.

Probably depends on age and wealth position.

Comments for this post are closed