My book on the Great Depression is officially published today. Looking back on the long project, I see 6 themes that have been there throughout my research career (which began in 1986, with this project.)
Never reason from a price change. The Great Depression can only be understood by considering both AD and AS shocks. Most previous explanations (not all) focused only on demand-side factors, or only on supply-side factors. It was gold hoarding (a negative demand shock) and artificial attempts to raise wages (a negative supply-shock) that made the Depression "Great". My best journal article (1989 JPE, with Steve Silver) was nothing more than never reason from a price change applied to real wage cyclicality.
It's all about the supply and demand for the medium of account. Gold was the medium of account throughout the Depression, except 1933-34. And even then, gold prices were an indicator of the future expected gold price peg.
Monetary policy shocks are changes in the future expected path of policy. This is where my "long and variable leads" idea comes from. It also explains why Gauti Eggertsson was initially drawn to my gold standard research, in his 2008 AER piece. However, he didn't accept my view that the New Deal wage policies were contractionary.
Auction-style financial and commodity markets are efficient, and provide the best indicator of macro shocks. I looked at short and long term interest rates, bond risk spreads, stock prices, commodity prices, and gold (or forex) prices. Also forward exchange rates. "There is no wait and see", markets immediately provide the optimal forecast of the long run impact of a shock.
It's complicated. A given policy shock can be expansionary in one setting, and contractionary in another. For instance, expectations of dollar depreciation were contractionary under Hoover (when pegged to gold), but expansionary in 1933, when they led to a weaker dollar. Failure to implement Smoot-Hawley was contractionary in October 1929 (when it indicated Congressional disarray) and enactment of Smoot-Hawley was contractionary in June 1930 (when it led to fears of an international trade war.) Higher discount rates were contractionary in 1928, but not in 1931 (when they helped end the wave of gold hoarding).
Musical chairs model. Business cycles are mostly caused by the interaction of sticky nominal wages and nominal shocks. In the book I used falling (wholesale) prices as a proxy for a contractionary nominal shocks, but in practice falling NGDP is better indicator.