Fig. [ ]. Source: Saez and Zucman .
A serious potential objection to the consumption reallocation hypothesis would be that new automobiles were less expensive than new houses, thus it probably could not have been the same set of people who preferred purchasing new cars to new houses. However, this objection can be addressed by taking into account that both a large part of new houses and a large part of new cars were purchased on credit.
In 1929, the share of new cars bought on credit reached the level of 61%, a three-percentage point increase over the 1928 level. Based on the data on car installment loans from two major providers (GMAC and CCC) whose total amount rose by 26% in 1929 over 1928, it is possible to suggest that the total amount of such loans rose at a similar rate.11The calculations in this paragraph are based on the data from Olney . This rate of increase in lending is similar to the rate of increase in the new car registrations. At the same time, Gärtner demonstrates that the total mortgage debt almost did not increase in 1929 as compared to 1928, whereas in the preceding three years it was rising by around 7-8% per year.
It remains to deal with the question why the automobile production faced a significant decline in the second half of 1929, and whether it may have been the catalyst for the initial phase of the Great Depression. Although it is difficult to establish it with certainty, similarly to the case with residential construction, it appears that the troubles of the automobile production industry are related more to the satiation of the market for new cars. This hypothesis is bolstered by the fact that the spike in automobile production in 1928-29 followed its temporary decline caused largely by the decision of Ford Motor Company to temporarily close its Ford Model T assembly lines in 1927 to switch to producing the new model A.